* [PATCH 1/3 v3] gdb/progspace: Add reverse safe iterator
2024-05-03 22:58 [PATCH 0/3] On-demand debuginfo downloading Aaron Merey
@ 2024-05-03 22:59 ` Aaron Merey
2024-05-17 14:09 ` [PING][PATCH " Aaron Merey
2024-05-03 22:59 ` [PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading Aaron Merey
2024-05-03 22:59 ` [PATCH 3/3 v6] gdb/debuginfod: Add .debug_line downloading Aaron Merey
2 siblings, 1 reply; 18+ messages in thread
From: Aaron Merey @ 2024-05-03 22:59 UTC (permalink / raw)
To: gdb-patches; +Cc: aburgess, Aaron Merey
This is a reposting of v3 from
https://sourceware.org/pipermail/gdb-patches/2024-January/205955.html
This patch changes progspace objfile_list insertion so that separate
debug objfiles are placed into the list after the parent objfile,
instead of before. Additionally qf_require_partial_symbols now returns
a safe_range.
These changes are intended to prepare gdb for on-demand debuginfo
downloading and the downloading of .gdb_index sections.
With on-demand downloading enabled, gdb might need to delete a
.gdb_index quick_symbol_functions from a parent objfile while looping
the objfile's list of quick_symbol_functions because the separate
debug objfile has just been downloaded. The use of a safe_range
prevents this removal from causing iterator invalidation.
gdb might also download a debuginfo file during symtab expansion.
In this case an objfile will be added to the current progspace's
objfiles_list during iteration over the list (for example, in
iterate_over_symtabs). We want these loops to also iterate over
newly downloaded objfiles. So objfiles need to be inserted into
objfiles_list after their parent since it is during the search of
the parent objfile for some symbol or filename that the separate
debug objfile might be downloaded.
To facilitate the safe deletion of objfiles, this patch also adds
basic_safe_reverse_range and basic_safe_reverse_iterator. This allows
objfiles to be removed from the objfiles_list in a loop without iterator
invalidation.
If a forward safe iterator were to be used, the deletion of an
objfile could invalidate the safe iterator's reference to the next
objfile in the objfiles_list. This can happen when the deletion
of an objfile causes the deletion of a separate debug objfile that
happens to the be next element in the objfiles_list.
The standard reverse iterator is not suitable for safe objfile deletion.
In order to safely delete the first objfile in the objfiles_list, the
standard reverse iterator's underlying begin iterator would have to be
decremented, resulting in undefined behavior.
A small change was also made to a testcase in py-objfile.exp to
account for the new placement of separate debug objfiles in
objfiles_list.
---
gdb/jit.c | 7 +-
gdb/objfiles.c | 14 +--
gdb/objfiles.h | 17 +++-
gdb/progspace.c | 19 +++-
gdb/progspace.h | 31 ++++---
gdb/symfile-debug.c | 34 +++----
gdb/testsuite/gdb.python/py-objfile.exp | 2 +-
gdbsupport/safe-iterator.h | 115 ++++++++++++++++++++++++
8 files changed, 195 insertions(+), 44 deletions(-)
diff --git a/gdb/jit.c b/gdb/jit.c
index 797be95a8da..90f9159fa9d 100644
--- a/gdb/jit.c
+++ b/gdb/jit.c
@@ -1240,11 +1240,10 @@ jit_breakpoint_re_set (void)
static void
jit_inferior_exit_hook (struct inferior *inf)
{
- for (objfile *objf : current_program_space->objfiles_safe ())
+ current_program_space->unlink_objfiles_if ([&] (const objfile *objf)
{
- if (objf->jited_data != nullptr && objf->jited_data->addr != 0)
- objf->unlink ();
- }
+ return (objf->jited_data != nullptr) && (objf->jited_data->addr != 0);
+ });
}
void
diff --git a/gdb/objfiles.c b/gdb/objfiles.c
index ba88ed1bc41..70017c7fb32 100644
--- a/gdb/objfiles.c
+++ b/gdb/objfiles.c
@@ -469,6 +469,12 @@ objfile::unlink ()
current_program_space->remove_objfile (this);
}
+qf_safe_range
+objfile::qf_safe ()
+{
+ return qf_safe_range (qf_range (qf.begin (), qf.end ()));
+}
+
/* Free all separate debug objfile of OBJFILE, but don't free OBJFILE
itself. */
@@ -792,14 +798,12 @@ have_full_symbols (void)
void
objfile_purge_solibs (void)
{
- for (objfile *objf : current_program_space->objfiles_safe ())
+ current_program_space->unlink_objfiles_if ([&] (const objfile *objf)
{
/* We assume that the solib package has been purged already, or will
be soon. */
-
- if (!(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED))
- objf->unlink ();
- }
+ return !(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED);
+ });
}
diff --git a/gdb/objfiles.h b/gdb/objfiles.h
index 8b8b7182e87..1c8f79a2826 100644
--- a/gdb/objfiles.h
+++ b/gdb/objfiles.h
@@ -370,6 +370,12 @@ class separate_debug_iterator
typedef iterator_range<separate_debug_iterator> separate_debug_range;
+/* See objfile::qf_safe. */
+
+using qf_list = std::forward_list<quick_symbol_functions_up>;
+using qf_range = iterator_range<qf_list::iterator>;
+using qf_safe_range = basic_safe_range<qf_range>;
+
/* Sections in an objfile. The section offsets are stored in the
OBJFILE. */
@@ -767,8 +773,15 @@ struct objfile
const struct sym_fns *sf = nullptr;
/* The "quick" (aka partial) symbol functions for this symbol
- reader. */
- std::forward_list<quick_symbol_functions_up> qf;
+ reader. Many quick_symbol_functions methods may result
+ in the deletion of a quick_symbol_functions from this
+ qf_list. It is recommended that qf_safe be used to iterate
+ over the qf_list. */
+ qf_list qf;
+
+ /* Returns an iterable object that allows for safe deletion during
+ iteration. See gdbsupport/safe-iterator.h. */
+ qf_safe_range qf_safe ();
/* Per objfile data-pointers required by other GDB modules. */
diff --git a/gdb/progspace.c b/gdb/progspace.c
index 0deca3f2376..b1b2dfb30c8 100644
--- a/gdb/progspace.c
+++ b/gdb/progspace.c
@@ -140,19 +140,19 @@ program_space::free_all_objfiles ()
void
program_space::add_objfile (std::unique_ptr<objfile> &&objfile,
- struct objfile *before)
+ struct objfile *after)
{
- if (before == nullptr)
+ if (after == nullptr)
objfiles_list.push_back (std::move (objfile));
else
{
auto iter = std::find_if (objfiles_list.begin (), objfiles_list.end (),
[=] (const std::unique_ptr<::objfile> &objf)
{
- return objf.get () == before;
+ return objf.get () == after;
});
gdb_assert (iter != objfiles_list.end ());
- objfiles_list.insert (iter, std::move (objfile));
+ objfiles_list.insert (++iter, std::move (objfile));
}
}
@@ -181,6 +181,17 @@ program_space::remove_objfile (struct objfile *objfile)
/* See progspace.h. */
+void
+program_space::unlink_objfiles_if
+ (gdb::function_view<bool (const objfile *objfile)> predicate)
+{
+ for (auto &it : objfiles_safe ())
+ if (predicate (it.get ()))
+ it->unlink ();
+}
+
+/* See progspace.h. */
+
struct objfile *
program_space::objfile_for_address (CORE_ADDR address)
{
diff --git a/gdb/progspace.h b/gdb/progspace.h
index bbf54efa07a..4d5ac991849 100644
--- a/gdb/progspace.h
+++ b/gdb/progspace.h
@@ -250,28 +250,32 @@ struct program_space
unwrapping_objfile_iterator (objfiles_list.end ()));
}
- using objfiles_safe_range = basic_safe_range<objfiles_range>;
+ using objfiles_it_range = iterator_range<objfile_list::iterator>;
+ using objfiles_safe_reverse_range
+ = basic_safe_reverse_range<objfiles_it_range>;
/* An iterable object that can be used to iterate over all objfiles.
The basic use is in a foreach, like:
for (objfile *objf : pspace->objfiles_safe ()) { ... }
- This variant uses a basic_safe_iterator so that objfiles can be
- deleted during iteration. */
- objfiles_safe_range objfiles_safe ()
+ This variant uses a basic_safe_reverse_iterator so that objfiles
+ can be deleted during iteration.
+
+ The use of a reverse iterator helps ensure that separate debug
+ objfiles are deleted before their parent objfile. This prevents
+ iterator invalidation due to the deletion of a parent objfile. */
+ objfiles_safe_reverse_range objfiles_safe ()
{
- return objfiles_safe_range
- (objfiles_range
- (unwrapping_objfile_iterator (objfiles_list.begin ()),
- unwrapping_objfile_iterator (objfiles_list.end ())));
+ return objfiles_safe_reverse_range
+ (objfiles_it_range (objfiles_list.begin (), objfiles_list.end ()));
}
- /* Add OBJFILE to the list of objfiles, putting it just before
- BEFORE. If BEFORE is nullptr, it will go at the end of the
+ /* Add OBJFILE to the list of objfiles, putting it just after
+ AFTER. If AFTER is nullptr, it will go at the end of the
list. */
void add_objfile (std::unique_ptr<objfile> &&objfile,
- struct objfile *before);
+ struct objfile *after);
/* Remove OBJFILE from the list of objfiles. */
void remove_objfile (struct objfile *objfile);
@@ -286,6 +290,11 @@ struct program_space
/* Free all the objfiles associated with this program space. */
void free_all_objfiles ();
+ /* Unlink all objfiles associated with this program space for which
+ PREDICATE evaluates to true. */
+ void unlink_objfiles_if
+ (gdb::function_view<bool (const objfile *objfile)> predicate);
+
/* Return the objfile containing ADDRESS, or nullptr if the address
is outside all objfiles in this progspace. */
struct objfile *objfile_for_address (CORE_ADDR address);
diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c
index 86c7010b2d2..330a520dca6 100644
--- a/gdb/symfile-debug.c
+++ b/gdb/symfile-debug.c
@@ -84,7 +84,7 @@ objfile::has_partial_symbols ()
them, then that is an indication that they are in fact available. Without
this function the symbols may have been already read in but they also may
not be present in this objfile. */
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
{
retval = iter->has_symbols (this);
if (retval)
@@ -107,7 +107,7 @@ objfile::has_unexpanded_symtabs ()
objfile_debug_name (this));
bool result = false;
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
{
if (iter->has_unexpanded_symtabs (this))
{
@@ -132,7 +132,7 @@ objfile::find_last_source_symtab ()
gdb_printf (gdb_stdlog, "qf->find_last_source_symtab (%s)\n",
objfile_debug_name (this));
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
{
retval = iter->find_last_source_symtab (this);
if (retval != nullptr)
@@ -165,7 +165,7 @@ objfile::forget_cached_source_info ()
}
}
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
iter->forget_cached_source_info (this);
}
@@ -212,7 +212,7 @@ objfile::map_symtabs_matching_filename
return result;
};
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
{
if (!iter->expand_symtabs_matching (this,
match_one_filename,
@@ -275,7 +275,7 @@ objfile::lookup_symbol (block_enum kind, const lookup_name_info &name,
return true;
};
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
{
if (!iter->expand_symtabs_matching (this,
nullptr,
@@ -305,7 +305,7 @@ objfile::print_stats (bool print_bcache)
gdb_printf (gdb_stdlog, "qf->print_stats (%s, %d)\n",
objfile_debug_name (this), print_bcache);
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
iter->print_stats (this, print_bcache);
}
@@ -316,7 +316,7 @@ objfile::dump ()
gdb_printf (gdb_stdlog, "qf->dump (%s)\n",
objfile_debug_name (this));
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
iter->dump (this);
}
@@ -331,7 +331,7 @@ objfile::expand_symtabs_for_function (const char *func_name)
lookup_name_info base_lookup (func_name, symbol_name_match_type::FULL);
lookup_name_info lookup_name = base_lookup.make_ignore_params ();
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
iter->expand_symtabs_matching (this,
nullptr,
&lookup_name,
@@ -349,7 +349,7 @@ objfile::expand_all_symtabs ()
gdb_printf (gdb_stdlog, "qf->expand_all_symtabs (%s)\n",
objfile_debug_name (this));
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
iter->expand_all_symtabs (this);
}
@@ -367,7 +367,7 @@ objfile::expand_symtabs_with_fullname (const char *fullname)
return filename_cmp (basenames ? basename : fullname, filename) == 0;
};
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
iter->expand_symtabs_matching (this,
file_matcher,
nullptr,
@@ -399,7 +399,7 @@ objfile::expand_symtabs_matching
host_address_to_string (&expansion_notify),
domain_name (domain).c_str ());
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
if (!iter->expand_symtabs_matching (this, file_matcher, lookup_name,
symbol_matcher, expansion_notify,
search_flags, domain))
@@ -424,7 +424,7 @@ objfile::find_pc_sect_compunit_symtab (struct bound_minimal_symbol msymbol,
host_address_to_string (section),
warn_if_readin);
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
{
retval = iter->find_pc_sect_compunit_symtab (this, msymbol, pc, section,
warn_if_readin);
@@ -452,7 +452,7 @@ objfile::map_symbol_filenames (gdb::function_view<symbol_filename_ftype> fun,
objfile_debug_name (this),
need_fullname);
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
iter->map_symbol_filenames (this, fun, need_fullname);
}
@@ -464,7 +464,7 @@ objfile::compute_main_name ()
"qf->compute_main_name (%s)\n",
objfile_debug_name (this));
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
iter->compute_main_name (this);
}
@@ -478,7 +478,7 @@ objfile::find_compunit_symtab_by_address (CORE_ADDR address)
hex_string (address));
struct compunit_symtab *result = NULL;
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
{
result = iter->find_compunit_symtab_by_address (this, address);
if (result != nullptr)
@@ -503,7 +503,7 @@ objfile::lookup_global_symbol_language (const char *name,
enum language result = language_unknown;
*symbol_found_p = false;
- for (const auto &iter : qf)
+ for (const auto &iter : qf_safe ())
{
result = iter->lookup_global_symbol_language (this, name, domain,
symbol_found_p);
diff --git a/gdb/testsuite/gdb.python/py-objfile.exp b/gdb/testsuite/gdb.python/py-objfile.exp
index 2f5b7752b43..8460c300de0 100644
--- a/gdb/testsuite/gdb.python/py-objfile.exp
+++ b/gdb/testsuite/gdb.python/py-objfile.exp
@@ -135,7 +135,7 @@ gdb_test "p main" "= {<text variable, no debug info>} $hex <main>" \
gdb_py_test_silent_cmd "python objfile.add_separate_debug_file(\"${binfile}\")" \
"Add separate debug file file" 1
-gdb_py_test_silent_cmd "python sep_objfile = gdb.objfiles()\[0\]" \
+gdb_py_test_silent_cmd "python sep_objfile = gdb.objfiles()\[1\]" \
"Get separate debug info objfile" 1
gdb_test "python print (sep_objfile.owner.filename)" "${testfile}2" \
diff --git a/gdbsupport/safe-iterator.h b/gdbsupport/safe-iterator.h
index f69f3896758..be23cf7920a 100644
--- a/gdbsupport/safe-iterator.h
+++ b/gdbsupport/safe-iterator.h
@@ -136,4 +136,119 @@ class basic_safe_range
Range m_range;
};
+/* A reverse basic_safe_iterator. See basic_safe_iterator for intended use. */
+
+template<typename Iterator>
+class basic_safe_reverse_iterator
+{
+public:
+ typedef basic_safe_reverse_iterator self_type;
+ typedef typename Iterator::value_type value_type;
+ typedef typename Iterator::reference reference;
+ typedef typename Iterator::pointer pointer;
+ typedef typename Iterator::iterator_category iterator_category;
+ typedef typename Iterator::difference_type difference_type;
+
+ /* Construct the iterator using ARG, and construct the end iterator
+ using ARG2. ARG and ARG2 should be forward iterators, typically
+ from begin and end methods, respectively.
+
+ For example if ARG1 is created with container.begin and ARG2 is
+ is created with container.end, then the basic_safe_reverse_iterator
+ will traverse from the last element in the container to the first
+ element in the container. */
+ template<typename Arg>
+ explicit basic_safe_reverse_iterator (Arg &&arg, Arg &&arg2)
+ : m_begin (std::forward<Arg> (arg)),
+ m_end (std::forward<Arg> (arg2)),
+ m_it (m_end),
+ m_next (m_end)
+ {
+ /* M_IT and M_NEXT are initialized as one-past-end. Set M_IT to point
+ to the last element and set M_NEXT to point to the second last element,
+ if such elements exist. */
+ if (m_it != m_begin)
+ {
+ --m_it;
+
+ if (m_it != m_begin)
+ {
+ --m_next;
+ --m_next;
+ }
+ }
+ }
+
+ typename std::invoke_result<decltype(&Iterator::operator*), Iterator>::type
+ operator* () const
+ { return *m_it; }
+
+ self_type &operator++ ()
+ {
+ m_it = m_next;
+
+ if (m_it != m_end)
+ {
+ /* Use M_BEGIN only if we sure that it is valid. */
+ if (m_it == m_begin)
+ m_next = m_end;
+ else
+ --m_next;
+ }
+
+ return *this;
+ }
+
+ bool operator== (const self_type &other) const
+ { return m_it == other.m_it; }
+
+ bool operator!= (const self_type &other) const
+ { return m_it != other.m_it; }
+
+private:
+ /* The first element. */
+ Iterator m_begin {};
+
+ /* A one-past-end iterator. */
+ Iterator m_end {};
+
+ /* The current element. */
+ Iterator m_it {};
+
+ /* The next element. Always one element ahead of M_IT. */
+ Iterator m_next {};
+};
+
+/* A range adapter that wraps a forward range, and then returns
+ safe reverse iterators wrapping the original range's iterators. */
+
+template<typename Range>
+class basic_safe_reverse_range
+{
+public:
+
+ typedef basic_safe_reverse_iterator<typename Range::iterator> iterator;
+
+ /* RANGE must be a forward range. basic_safe_reverse_iterators
+ will be used to traverse the forward range from the last element
+ to the first. */
+ explicit basic_safe_reverse_range (Range range)
+ : m_range (range)
+ {
+ }
+
+ iterator begin ()
+ {
+ return iterator (m_range.begin (), m_range.end ());
+ }
+
+ iterator end ()
+ {
+ return iterator (m_range.end (), m_range.end ());
+ }
+
+private:
+
+ Range m_range;
+};
#endif /* COMMON_SAFE_ITERATOR_H */
--
2.43.0
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING][PATCH 1/3 v3] gdb/progspace: Add reverse safe iterator
2024-05-03 22:59 ` [PATCH 1/3 v3] gdb/progspace: Add reverse safe iterator Aaron Merey
@ 2024-05-17 14:09 ` Aaron Merey
2024-05-31 20:58 ` [PING*2][PATCH " Aaron Merey
0 siblings, 1 reply; 18+ messages in thread
From: Aaron Merey @ 2024-05-17 14:09 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
>
> This is a reposting of v3 from
> https://sourceware.org/pipermail/gdb-patches/2024-January/205955.html
>
> This patch changes progspace objfile_list insertion so that separate
> debug objfiles are placed into the list after the parent objfile,
> instead of before. Additionally qf_require_partial_symbols now returns
> a safe_range.
>
> These changes are intended to prepare gdb for on-demand debuginfo
> downloading and the downloading of .gdb_index sections.
>
> With on-demand downloading enabled, gdb might need to delete a
> .gdb_index quick_symbol_functions from a parent objfile while looping
> the objfile's list of quick_symbol_functions because the separate
> debug objfile has just been downloaded. The use of a safe_range
> prevents this removal from causing iterator invalidation.
>
> gdb might also download a debuginfo file during symtab expansion.
> In this case an objfile will be added to the current progspace's
> objfiles_list during iteration over the list (for example, in
> iterate_over_symtabs). We want these loops to also iterate over
> newly downloaded objfiles. So objfiles need to be inserted into
> objfiles_list after their parent since it is during the search of
> the parent objfile for some symbol or filename that the separate
> debug objfile might be downloaded.
>
> To facilitate the safe deletion of objfiles, this patch also adds
> basic_safe_reverse_range and basic_safe_reverse_iterator. This allows
> objfiles to be removed from the objfiles_list in a loop without iterator
> invalidation.
>
> If a forward safe iterator were to be used, the deletion of an
> objfile could invalidate the safe iterator's reference to the next
> objfile in the objfiles_list. This can happen when the deletion
> of an objfile causes the deletion of a separate debug objfile that
> happens to the be next element in the objfiles_list.
>
> The standard reverse iterator is not suitable for safe objfile deletion.
> In order to safely delete the first objfile in the objfiles_list, the
> standard reverse iterator's underlying begin iterator would have to be
> decremented, resulting in undefined behavior.
>
> A small change was also made to a testcase in py-objfile.exp to
> account for the new placement of separate debug objfiles in
> objfiles_list.
> ---
> gdb/jit.c | 7 +-
> gdb/objfiles.c | 14 +--
> gdb/objfiles.h | 17 +++-
> gdb/progspace.c | 19 +++-
> gdb/progspace.h | 31 ++++---
> gdb/symfile-debug.c | 34 +++----
> gdb/testsuite/gdb.python/py-objfile.exp | 2 +-
> gdbsupport/safe-iterator.h | 115 ++++++++++++++++++++++++
> 8 files changed, 195 insertions(+), 44 deletions(-)
>
> diff --git a/gdb/jit.c b/gdb/jit.c
> index 797be95a8da..90f9159fa9d 100644
> --- a/gdb/jit.c
> +++ b/gdb/jit.c
> @@ -1240,11 +1240,10 @@ jit_breakpoint_re_set (void)
> static void
> jit_inferior_exit_hook (struct inferior *inf)
> {
> - for (objfile *objf : current_program_space->objfiles_safe ())
> + current_program_space->unlink_objfiles_if ([&] (const objfile *objf)
> {
> - if (objf->jited_data != nullptr && objf->jited_data->addr != 0)
> - objf->unlink ();
> - }
> + return (objf->jited_data != nullptr) && (objf->jited_data->addr != 0);
> + });
> }
>
> void
> diff --git a/gdb/objfiles.c b/gdb/objfiles.c
> index ba88ed1bc41..70017c7fb32 100644
> --- a/gdb/objfiles.c
> +++ b/gdb/objfiles.c
> @@ -469,6 +469,12 @@ objfile::unlink ()
> current_program_space->remove_objfile (this);
> }
>
> +qf_safe_range
> +objfile::qf_safe ()
> +{
> + return qf_safe_range (qf_range (qf.begin (), qf.end ()));
> +}
> +
> /* Free all separate debug objfile of OBJFILE, but don't free OBJFILE
> itself. */
>
> @@ -792,14 +798,12 @@ have_full_symbols (void)
> void
> objfile_purge_solibs (void)
> {
> - for (objfile *objf : current_program_space->objfiles_safe ())
> + current_program_space->unlink_objfiles_if ([&] (const objfile *objf)
> {
> /* We assume that the solib package has been purged already, or will
> be soon. */
> -
> - if (!(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED))
> - objf->unlink ();
> - }
> + return !(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED);
> + });
> }
>
>
> diff --git a/gdb/objfiles.h b/gdb/objfiles.h
> index 8b8b7182e87..1c8f79a2826 100644
> --- a/gdb/objfiles.h
> +++ b/gdb/objfiles.h
> @@ -370,6 +370,12 @@ class separate_debug_iterator
>
> typedef iterator_range<separate_debug_iterator> separate_debug_range;
>
> +/* See objfile::qf_safe. */
> +
> +using qf_list = std::forward_list<quick_symbol_functions_up>;
> +using qf_range = iterator_range<qf_list::iterator>;
> +using qf_safe_range = basic_safe_range<qf_range>;
> +
> /* Sections in an objfile. The section offsets are stored in the
> OBJFILE. */
>
> @@ -767,8 +773,15 @@ struct objfile
> const struct sym_fns *sf = nullptr;
>
> /* The "quick" (aka partial) symbol functions for this symbol
> - reader. */
> - std::forward_list<quick_symbol_functions_up> qf;
> + reader. Many quick_symbol_functions methods may result
> + in the deletion of a quick_symbol_functions from this
> + qf_list. It is recommended that qf_safe be used to iterate
> + over the qf_list. */
> + qf_list qf;
> +
> + /* Returns an iterable object that allows for safe deletion during
> + iteration. See gdbsupport/safe-iterator.h. */
> + qf_safe_range qf_safe ();
>
> /* Per objfile data-pointers required by other GDB modules. */
>
> diff --git a/gdb/progspace.c b/gdb/progspace.c
> index 0deca3f2376..b1b2dfb30c8 100644
> --- a/gdb/progspace.c
> +++ b/gdb/progspace.c
> @@ -140,19 +140,19 @@ program_space::free_all_objfiles ()
>
> void
> program_space::add_objfile (std::unique_ptr<objfile> &&objfile,
> - struct objfile *before)
> + struct objfile *after)
> {
> - if (before == nullptr)
> + if (after == nullptr)
> objfiles_list.push_back (std::move (objfile));
> else
> {
> auto iter = std::find_if (objfiles_list.begin (), objfiles_list.end (),
> [=] (const std::unique_ptr<::objfile> &objf)
> {
> - return objf.get () == before;
> + return objf.get () == after;
> });
> gdb_assert (iter != objfiles_list.end ());
> - objfiles_list.insert (iter, std::move (objfile));
> + objfiles_list.insert (++iter, std::move (objfile));
> }
> }
>
> @@ -181,6 +181,17 @@ program_space::remove_objfile (struct objfile *objfile)
>
> /* See progspace.h. */
>
> +void
> +program_space::unlink_objfiles_if
> + (gdb::function_view<bool (const objfile *objfile)> predicate)
> +{
> + for (auto &it : objfiles_safe ())
> + if (predicate (it.get ()))
> + it->unlink ();
> +}
> +
> +/* See progspace.h. */
> +
> struct objfile *
> program_space::objfile_for_address (CORE_ADDR address)
> {
> diff --git a/gdb/progspace.h b/gdb/progspace.h
> index bbf54efa07a..4d5ac991849 100644
> --- a/gdb/progspace.h
> +++ b/gdb/progspace.h
> @@ -250,28 +250,32 @@ struct program_space
> unwrapping_objfile_iterator (objfiles_list.end ()));
> }
>
> - using objfiles_safe_range = basic_safe_range<objfiles_range>;
> + using objfiles_it_range = iterator_range<objfile_list::iterator>;
> + using objfiles_safe_reverse_range
> + = basic_safe_reverse_range<objfiles_it_range>;
>
> /* An iterable object that can be used to iterate over all objfiles.
> The basic use is in a foreach, like:
>
> for (objfile *objf : pspace->objfiles_safe ()) { ... }
>
> - This variant uses a basic_safe_iterator so that objfiles can be
> - deleted during iteration. */
> - objfiles_safe_range objfiles_safe ()
> + This variant uses a basic_safe_reverse_iterator so that objfiles
> + can be deleted during iteration.
> +
> + The use of a reverse iterator helps ensure that separate debug
> + objfiles are deleted before their parent objfile. This prevents
> + iterator invalidation due to the deletion of a parent objfile. */
> + objfiles_safe_reverse_range objfiles_safe ()
> {
> - return objfiles_safe_range
> - (objfiles_range
> - (unwrapping_objfile_iterator (objfiles_list.begin ()),
> - unwrapping_objfile_iterator (objfiles_list.end ())));
> + return objfiles_safe_reverse_range
> + (objfiles_it_range (objfiles_list.begin (), objfiles_list.end ()));
> }
>
> - /* Add OBJFILE to the list of objfiles, putting it just before
> - BEFORE. If BEFORE is nullptr, it will go at the end of the
> + /* Add OBJFILE to the list of objfiles, putting it just after
> + AFTER. If AFTER is nullptr, it will go at the end of the
> list. */
> void add_objfile (std::unique_ptr<objfile> &&objfile,
> - struct objfile *before);
> + struct objfile *after);
>
> /* Remove OBJFILE from the list of objfiles. */
> void remove_objfile (struct objfile *objfile);
> @@ -286,6 +290,11 @@ struct program_space
> /* Free all the objfiles associated with this program space. */
> void free_all_objfiles ();
>
> + /* Unlink all objfiles associated with this program space for which
> + PREDICATE evaluates to true. */
> + void unlink_objfiles_if
> + (gdb::function_view<bool (const objfile *objfile)> predicate);
> +
> /* Return the objfile containing ADDRESS, or nullptr if the address
> is outside all objfiles in this progspace. */
> struct objfile *objfile_for_address (CORE_ADDR address);
> diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c
> index 86c7010b2d2..330a520dca6 100644
> --- a/gdb/symfile-debug.c
> +++ b/gdb/symfile-debug.c
> @@ -84,7 +84,7 @@ objfile::has_partial_symbols ()
> them, then that is an indication that they are in fact available. Without
> this function the symbols may have been already read in but they also may
> not be present in this objfile. */
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> {
> retval = iter->has_symbols (this);
> if (retval)
> @@ -107,7 +107,7 @@ objfile::has_unexpanded_symtabs ()
> objfile_debug_name (this));
>
> bool result = false;
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> {
> if (iter->has_unexpanded_symtabs (this))
> {
> @@ -132,7 +132,7 @@ objfile::find_last_source_symtab ()
> gdb_printf (gdb_stdlog, "qf->find_last_source_symtab (%s)\n",
> objfile_debug_name (this));
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> {
> retval = iter->find_last_source_symtab (this);
> if (retval != nullptr)
> @@ -165,7 +165,7 @@ objfile::forget_cached_source_info ()
> }
> }
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> iter->forget_cached_source_info (this);
> }
>
> @@ -212,7 +212,7 @@ objfile::map_symtabs_matching_filename
> return result;
> };
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> {
> if (!iter->expand_symtabs_matching (this,
> match_one_filename,
> @@ -275,7 +275,7 @@ objfile::lookup_symbol (block_enum kind, const lookup_name_info &name,
> return true;
> };
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> {
> if (!iter->expand_symtabs_matching (this,
> nullptr,
> @@ -305,7 +305,7 @@ objfile::print_stats (bool print_bcache)
> gdb_printf (gdb_stdlog, "qf->print_stats (%s, %d)\n",
> objfile_debug_name (this), print_bcache);
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> iter->print_stats (this, print_bcache);
> }
>
> @@ -316,7 +316,7 @@ objfile::dump ()
> gdb_printf (gdb_stdlog, "qf->dump (%s)\n",
> objfile_debug_name (this));
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> iter->dump (this);
> }
>
> @@ -331,7 +331,7 @@ objfile::expand_symtabs_for_function (const char *func_name)
> lookup_name_info base_lookup (func_name, symbol_name_match_type::FULL);
> lookup_name_info lookup_name = base_lookup.make_ignore_params ();
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> iter->expand_symtabs_matching (this,
> nullptr,
> &lookup_name,
> @@ -349,7 +349,7 @@ objfile::expand_all_symtabs ()
> gdb_printf (gdb_stdlog, "qf->expand_all_symtabs (%s)\n",
> objfile_debug_name (this));
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> iter->expand_all_symtabs (this);
> }
>
> @@ -367,7 +367,7 @@ objfile::expand_symtabs_with_fullname (const char *fullname)
> return filename_cmp (basenames ? basename : fullname, filename) == 0;
> };
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> iter->expand_symtabs_matching (this,
> file_matcher,
> nullptr,
> @@ -399,7 +399,7 @@ objfile::expand_symtabs_matching
> host_address_to_string (&expansion_notify),
> domain_name (domain).c_str ());
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> if (!iter->expand_symtabs_matching (this, file_matcher, lookup_name,
> symbol_matcher, expansion_notify,
> search_flags, domain))
> @@ -424,7 +424,7 @@ objfile::find_pc_sect_compunit_symtab (struct bound_minimal_symbol msymbol,
> host_address_to_string (section),
> warn_if_readin);
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> {
> retval = iter->find_pc_sect_compunit_symtab (this, msymbol, pc, section,
> warn_if_readin);
> @@ -452,7 +452,7 @@ objfile::map_symbol_filenames (gdb::function_view<symbol_filename_ftype> fun,
> objfile_debug_name (this),
> need_fullname);
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> iter->map_symbol_filenames (this, fun, need_fullname);
> }
>
> @@ -464,7 +464,7 @@ objfile::compute_main_name ()
> "qf->compute_main_name (%s)\n",
> objfile_debug_name (this));
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> iter->compute_main_name (this);
> }
>
> @@ -478,7 +478,7 @@ objfile::find_compunit_symtab_by_address (CORE_ADDR address)
> hex_string (address));
>
> struct compunit_symtab *result = NULL;
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> {
> result = iter->find_compunit_symtab_by_address (this, address);
> if (result != nullptr)
> @@ -503,7 +503,7 @@ objfile::lookup_global_symbol_language (const char *name,
> enum language result = language_unknown;
> *symbol_found_p = false;
>
> - for (const auto &iter : qf)
> + for (const auto &iter : qf_safe ())
> {
> result = iter->lookup_global_symbol_language (this, name, domain,
> symbol_found_p);
> diff --git a/gdb/testsuite/gdb.python/py-objfile.exp b/gdb/testsuite/gdb.python/py-objfile.exp
> index 2f5b7752b43..8460c300de0 100644
> --- a/gdb/testsuite/gdb.python/py-objfile.exp
> +++ b/gdb/testsuite/gdb.python/py-objfile.exp
> @@ -135,7 +135,7 @@ gdb_test "p main" "= {<text variable, no debug info>} $hex <main>" \
> gdb_py_test_silent_cmd "python objfile.add_separate_debug_file(\"${binfile}\")" \
> "Add separate debug file file" 1
>
> -gdb_py_test_silent_cmd "python sep_objfile = gdb.objfiles()\[0\]" \
> +gdb_py_test_silent_cmd "python sep_objfile = gdb.objfiles()\[1\]" \
> "Get separate debug info objfile" 1
>
> gdb_test "python print (sep_objfile.owner.filename)" "${testfile}2" \
> diff --git a/gdbsupport/safe-iterator.h b/gdbsupport/safe-iterator.h
> index f69f3896758..be23cf7920a 100644
> --- a/gdbsupport/safe-iterator.h
> +++ b/gdbsupport/safe-iterator.h
> @@ -136,4 +136,119 @@ class basic_safe_range
> Range m_range;
> };
>
> +/* A reverse basic_safe_iterator. See basic_safe_iterator for intended use. */
> +
> +template<typename Iterator>
> +class basic_safe_reverse_iterator
> +{
> +public:
> + typedef basic_safe_reverse_iterator self_type;
> + typedef typename Iterator::value_type value_type;
> + typedef typename Iterator::reference reference;
> + typedef typename Iterator::pointer pointer;
> + typedef typename Iterator::iterator_category iterator_category;
> + typedef typename Iterator::difference_type difference_type;
> +
> + /* Construct the iterator using ARG, and construct the end iterator
> + using ARG2. ARG and ARG2 should be forward iterators, typically
> + from begin and end methods, respectively.
> +
> + For example if ARG1 is created with container.begin and ARG2 is
> + is created with container.end, then the basic_safe_reverse_iterator
> + will traverse from the last element in the container to the first
> + element in the container. */
> + template<typename Arg>
> + explicit basic_safe_reverse_iterator (Arg &&arg, Arg &&arg2)
> + : m_begin (std::forward<Arg> (arg)),
> + m_end (std::forward<Arg> (arg2)),
> + m_it (m_end),
> + m_next (m_end)
> + {
> + /* M_IT and M_NEXT are initialized as one-past-end. Set M_IT to point
> + to the last element and set M_NEXT to point to the second last element,
> + if such elements exist. */
> + if (m_it != m_begin)
> + {
> + --m_it;
> +
> + if (m_it != m_begin)
> + {
> + --m_next;
> + --m_next;
> + }
> + }
> + }
> +
> + typename std::invoke_result<decltype(&Iterator::operator*), Iterator>::type
> + operator* () const
> + { return *m_it; }
> +
> + self_type &operator++ ()
> + {
> + m_it = m_next;
> +
> + if (m_it != m_end)
> + {
> + /* Use M_BEGIN only if we sure that it is valid. */
> + if (m_it == m_begin)
> + m_next = m_end;
> + else
> + --m_next;
> + }
> +
> + return *this;
> + }
> +
> + bool operator== (const self_type &other) const
> + { return m_it == other.m_it; }
> +
> + bool operator!= (const self_type &other) const
> + { return m_it != other.m_it; }
> +
> +private:
> + /* The first element. */
> + Iterator m_begin {};
> +
> + /* A one-past-end iterator. */
> + Iterator m_end {};
> +
> + /* The current element. */
> + Iterator m_it {};
> +
> + /* The next element. Always one element ahead of M_IT. */
> + Iterator m_next {};
> +};
> +
> +/* A range adapter that wraps a forward range, and then returns
> + safe reverse iterators wrapping the original range's iterators. */
> +
> +template<typename Range>
> +class basic_safe_reverse_range
> +{
> +public:
> +
> + typedef basic_safe_reverse_iterator<typename Range::iterator> iterator;
> +
> + /* RANGE must be a forward range. basic_safe_reverse_iterators
> + will be used to traverse the forward range from the last element
> + to the first. */
> + explicit basic_safe_reverse_range (Range range)
> + : m_range (range)
> + {
> + }
> +
> + iterator begin ()
> + {
> + return iterator (m_range.begin (), m_range.end ());
> + }
> +
> + iterator end ()
> + {
> + return iterator (m_range.end (), m_range.end ());
> + }
> +
> +private:
> +
> + Range m_range;
> +};
> #endif /* COMMON_SAFE_ITERATOR_H */
> --
> 2.43.0
>
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING*2][PATCH 1/3 v3] gdb/progspace: Add reverse safe iterator
2024-05-17 14:09 ` [PING][PATCH " Aaron Merey
@ 2024-05-31 20:58 ` Aaron Merey
2024-06-07 20:48 ` [PING*3][PATCH " Aaron Merey
2024-06-14 18:23 ` [PING*2][PATCH " Aaron Merey
0 siblings, 2 replies; 18+ messages in thread
From: Aaron Merey @ 2024-05-31 20:58 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 17, 2024 at 10:09 AM Aaron Merey <amerey@redhat.com> wrote:
>
> Ping
>
> Thanks,
> Aaron
>
> On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
> >
> > This is a reposting of v3 from
> > https://sourceware.org/pipermail/gdb-patches/2024-January/205955.html
> >
> > This patch changes progspace objfile_list insertion so that separate
> > debug objfiles are placed into the list after the parent objfile,
> > instead of before. Additionally qf_require_partial_symbols now returns
> > a safe_range.
> >
> > These changes are intended to prepare gdb for on-demand debuginfo
> > downloading and the downloading of .gdb_index sections.
> >
> > With on-demand downloading enabled, gdb might need to delete a
> > .gdb_index quick_symbol_functions from a parent objfile while looping
> > the objfile's list of quick_symbol_functions because the separate
> > debug objfile has just been downloaded. The use of a safe_range
> > prevents this removal from causing iterator invalidation.
> >
> > gdb might also download a debuginfo file during symtab expansion.
> > In this case an objfile will be added to the current progspace's
> > objfiles_list during iteration over the list (for example, in
> > iterate_over_symtabs). We want these loops to also iterate over
> > newly downloaded objfiles. So objfiles need to be inserted into
> > objfiles_list after their parent since it is during the search of
> > the parent objfile for some symbol or filename that the separate
> > debug objfile might be downloaded.
> >
> > To facilitate the safe deletion of objfiles, this patch also adds
> > basic_safe_reverse_range and basic_safe_reverse_iterator. This allows
> > objfiles to be removed from the objfiles_list in a loop without iterator
> > invalidation.
> >
> > If a forward safe iterator were to be used, the deletion of an
> > objfile could invalidate the safe iterator's reference to the next
> > objfile in the objfiles_list. This can happen when the deletion
> > of an objfile causes the deletion of a separate debug objfile that
> > happens to the be next element in the objfiles_list.
> >
> > The standard reverse iterator is not suitable for safe objfile deletion.
> > In order to safely delete the first objfile in the objfiles_list, the
> > standard reverse iterator's underlying begin iterator would have to be
> > decremented, resulting in undefined behavior.
> >
> > A small change was also made to a testcase in py-objfile.exp to
> > account for the new placement of separate debug objfiles in
> > objfiles_list.
> > ---
> > gdb/jit.c | 7 +-
> > gdb/objfiles.c | 14 +--
> > gdb/objfiles.h | 17 +++-
> > gdb/progspace.c | 19 +++-
> > gdb/progspace.h | 31 ++++---
> > gdb/symfile-debug.c | 34 +++----
> > gdb/testsuite/gdb.python/py-objfile.exp | 2 +-
> > gdbsupport/safe-iterator.h | 115 ++++++++++++++++++++++++
> > 8 files changed, 195 insertions(+), 44 deletions(-)
> >
> > diff --git a/gdb/jit.c b/gdb/jit.c
> > index 797be95a8da..90f9159fa9d 100644
> > --- a/gdb/jit.c
> > +++ b/gdb/jit.c
> > @@ -1240,11 +1240,10 @@ jit_breakpoint_re_set (void)
> > static void
> > jit_inferior_exit_hook (struct inferior *inf)
> > {
> > - for (objfile *objf : current_program_space->objfiles_safe ())
> > + current_program_space->unlink_objfiles_if ([&] (const objfile *objf)
> > {
> > - if (objf->jited_data != nullptr && objf->jited_data->addr != 0)
> > - objf->unlink ();
> > - }
> > + return (objf->jited_data != nullptr) && (objf->jited_data->addr != 0);
> > + });
> > }
> >
> > void
> > diff --git a/gdb/objfiles.c b/gdb/objfiles.c
> > index ba88ed1bc41..70017c7fb32 100644
> > --- a/gdb/objfiles.c
> > +++ b/gdb/objfiles.c
> > @@ -469,6 +469,12 @@ objfile::unlink ()
> > current_program_space->remove_objfile (this);
> > }
> >
> > +qf_safe_range
> > +objfile::qf_safe ()
> > +{
> > + return qf_safe_range (qf_range (qf.begin (), qf.end ()));
> > +}
> > +
> > /* Free all separate debug objfile of OBJFILE, but don't free OBJFILE
> > itself. */
> >
> > @@ -792,14 +798,12 @@ have_full_symbols (void)
> > void
> > objfile_purge_solibs (void)
> > {
> > - for (objfile *objf : current_program_space->objfiles_safe ())
> > + current_program_space->unlink_objfiles_if ([&] (const objfile *objf)
> > {
> > /* We assume that the solib package has been purged already, or will
> > be soon. */
> > -
> > - if (!(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED))
> > - objf->unlink ();
> > - }
> > + return !(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED);
> > + });
> > }
> >
> >
> > diff --git a/gdb/objfiles.h b/gdb/objfiles.h
> > index 8b8b7182e87..1c8f79a2826 100644
> > --- a/gdb/objfiles.h
> > +++ b/gdb/objfiles.h
> > @@ -370,6 +370,12 @@ class separate_debug_iterator
> >
> > typedef iterator_range<separate_debug_iterator> separate_debug_range;
> >
> > +/* See objfile::qf_safe. */
> > +
> > +using qf_list = std::forward_list<quick_symbol_functions_up>;
> > +using qf_range = iterator_range<qf_list::iterator>;
> > +using qf_safe_range = basic_safe_range<qf_range>;
> > +
> > /* Sections in an objfile. The section offsets are stored in the
> > OBJFILE. */
> >
> > @@ -767,8 +773,15 @@ struct objfile
> > const struct sym_fns *sf = nullptr;
> >
> > /* The "quick" (aka partial) symbol functions for this symbol
> > - reader. */
> > - std::forward_list<quick_symbol_functions_up> qf;
> > + reader. Many quick_symbol_functions methods may result
> > + in the deletion of a quick_symbol_functions from this
> > + qf_list. It is recommended that qf_safe be used to iterate
> > + over the qf_list. */
> > + qf_list qf;
> > +
> > + /* Returns an iterable object that allows for safe deletion during
> > + iteration. See gdbsupport/safe-iterator.h. */
> > + qf_safe_range qf_safe ();
> >
> > /* Per objfile data-pointers required by other GDB modules. */
> >
> > diff --git a/gdb/progspace.c b/gdb/progspace.c
> > index 0deca3f2376..b1b2dfb30c8 100644
> > --- a/gdb/progspace.c
> > +++ b/gdb/progspace.c
> > @@ -140,19 +140,19 @@ program_space::free_all_objfiles ()
> >
> > void
> > program_space::add_objfile (std::unique_ptr<objfile> &&objfile,
> > - struct objfile *before)
> > + struct objfile *after)
> > {
> > - if (before == nullptr)
> > + if (after == nullptr)
> > objfiles_list.push_back (std::move (objfile));
> > else
> > {
> > auto iter = std::find_if (objfiles_list.begin (), objfiles_list.end (),
> > [=] (const std::unique_ptr<::objfile> &objf)
> > {
> > - return objf.get () == before;
> > + return objf.get () == after;
> > });
> > gdb_assert (iter != objfiles_list.end ());
> > - objfiles_list.insert (iter, std::move (objfile));
> > + objfiles_list.insert (++iter, std::move (objfile));
> > }
> > }
> >
> > @@ -181,6 +181,17 @@ program_space::remove_objfile (struct objfile *objfile)
> >
> > /* See progspace.h. */
> >
> > +void
> > +program_space::unlink_objfiles_if
> > + (gdb::function_view<bool (const objfile *objfile)> predicate)
> > +{
> > + for (auto &it : objfiles_safe ())
> > + if (predicate (it.get ()))
> > + it->unlink ();
> > +}
> > +
> > +/* See progspace.h. */
> > +
> > struct objfile *
> > program_space::objfile_for_address (CORE_ADDR address)
> > {
> > diff --git a/gdb/progspace.h b/gdb/progspace.h
> > index bbf54efa07a..4d5ac991849 100644
> > --- a/gdb/progspace.h
> > +++ b/gdb/progspace.h
> > @@ -250,28 +250,32 @@ struct program_space
> > unwrapping_objfile_iterator (objfiles_list.end ()));
> > }
> >
> > - using objfiles_safe_range = basic_safe_range<objfiles_range>;
> > + using objfiles_it_range = iterator_range<objfile_list::iterator>;
> > + using objfiles_safe_reverse_range
> > + = basic_safe_reverse_range<objfiles_it_range>;
> >
> > /* An iterable object that can be used to iterate over all objfiles.
> > The basic use is in a foreach, like:
> >
> > for (objfile *objf : pspace->objfiles_safe ()) { ... }
> >
> > - This variant uses a basic_safe_iterator so that objfiles can be
> > - deleted during iteration. */
> > - objfiles_safe_range objfiles_safe ()
> > + This variant uses a basic_safe_reverse_iterator so that objfiles
> > + can be deleted during iteration.
> > +
> > + The use of a reverse iterator helps ensure that separate debug
> > + objfiles are deleted before their parent objfile. This prevents
> > + iterator invalidation due to the deletion of a parent objfile. */
> > + objfiles_safe_reverse_range objfiles_safe ()
> > {
> > - return objfiles_safe_range
> > - (objfiles_range
> > - (unwrapping_objfile_iterator (objfiles_list.begin ()),
> > - unwrapping_objfile_iterator (objfiles_list.end ())));
> > + return objfiles_safe_reverse_range
> > + (objfiles_it_range (objfiles_list.begin (), objfiles_list.end ()));
> > }
> >
> > - /* Add OBJFILE to the list of objfiles, putting it just before
> > - BEFORE. If BEFORE is nullptr, it will go at the end of the
> > + /* Add OBJFILE to the list of objfiles, putting it just after
> > + AFTER. If AFTER is nullptr, it will go at the end of the
> > list. */
> > void add_objfile (std::unique_ptr<objfile> &&objfile,
> > - struct objfile *before);
> > + struct objfile *after);
> >
> > /* Remove OBJFILE from the list of objfiles. */
> > void remove_objfile (struct objfile *objfile);
> > @@ -286,6 +290,11 @@ struct program_space
> > /* Free all the objfiles associated with this program space. */
> > void free_all_objfiles ();
> >
> > + /* Unlink all objfiles associated with this program space for which
> > + PREDICATE evaluates to true. */
> > + void unlink_objfiles_if
> > + (gdb::function_view<bool (const objfile *objfile)> predicate);
> > +
> > /* Return the objfile containing ADDRESS, or nullptr if the address
> > is outside all objfiles in this progspace. */
> > struct objfile *objfile_for_address (CORE_ADDR address);
> > diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c
> > index 86c7010b2d2..330a520dca6 100644
> > --- a/gdb/symfile-debug.c
> > +++ b/gdb/symfile-debug.c
> > @@ -84,7 +84,7 @@ objfile::has_partial_symbols ()
> > them, then that is an indication that they are in fact available. Without
> > this function the symbols may have been already read in but they also may
> > not be present in this objfile. */
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > {
> > retval = iter->has_symbols (this);
> > if (retval)
> > @@ -107,7 +107,7 @@ objfile::has_unexpanded_symtabs ()
> > objfile_debug_name (this));
> >
> > bool result = false;
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > {
> > if (iter->has_unexpanded_symtabs (this))
> > {
> > @@ -132,7 +132,7 @@ objfile::find_last_source_symtab ()
> > gdb_printf (gdb_stdlog, "qf->find_last_source_symtab (%s)\n",
> > objfile_debug_name (this));
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > {
> > retval = iter->find_last_source_symtab (this);
> > if (retval != nullptr)
> > @@ -165,7 +165,7 @@ objfile::forget_cached_source_info ()
> > }
> > }
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > iter->forget_cached_source_info (this);
> > }
> >
> > @@ -212,7 +212,7 @@ objfile::map_symtabs_matching_filename
> > return result;
> > };
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > {
> > if (!iter->expand_symtabs_matching (this,
> > match_one_filename,
> > @@ -275,7 +275,7 @@ objfile::lookup_symbol (block_enum kind, const lookup_name_info &name,
> > return true;
> > };
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > {
> > if (!iter->expand_symtabs_matching (this,
> > nullptr,
> > @@ -305,7 +305,7 @@ objfile::print_stats (bool print_bcache)
> > gdb_printf (gdb_stdlog, "qf->print_stats (%s, %d)\n",
> > objfile_debug_name (this), print_bcache);
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > iter->print_stats (this, print_bcache);
> > }
> >
> > @@ -316,7 +316,7 @@ objfile::dump ()
> > gdb_printf (gdb_stdlog, "qf->dump (%s)\n",
> > objfile_debug_name (this));
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > iter->dump (this);
> > }
> >
> > @@ -331,7 +331,7 @@ objfile::expand_symtabs_for_function (const char *func_name)
> > lookup_name_info base_lookup (func_name, symbol_name_match_type::FULL);
> > lookup_name_info lookup_name = base_lookup.make_ignore_params ();
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > iter->expand_symtabs_matching (this,
> > nullptr,
> > &lookup_name,
> > @@ -349,7 +349,7 @@ objfile::expand_all_symtabs ()
> > gdb_printf (gdb_stdlog, "qf->expand_all_symtabs (%s)\n",
> > objfile_debug_name (this));
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > iter->expand_all_symtabs (this);
> > }
> >
> > @@ -367,7 +367,7 @@ objfile::expand_symtabs_with_fullname (const char *fullname)
> > return filename_cmp (basenames ? basename : fullname, filename) == 0;
> > };
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > iter->expand_symtabs_matching (this,
> > file_matcher,
> > nullptr,
> > @@ -399,7 +399,7 @@ objfile::expand_symtabs_matching
> > host_address_to_string (&expansion_notify),
> > domain_name (domain).c_str ());
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > if (!iter->expand_symtabs_matching (this, file_matcher, lookup_name,
> > symbol_matcher, expansion_notify,
> > search_flags, domain))
> > @@ -424,7 +424,7 @@ objfile::find_pc_sect_compunit_symtab (struct bound_minimal_symbol msymbol,
> > host_address_to_string (section),
> > warn_if_readin);
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > {
> > retval = iter->find_pc_sect_compunit_symtab (this, msymbol, pc, section,
> > warn_if_readin);
> > @@ -452,7 +452,7 @@ objfile::map_symbol_filenames (gdb::function_view<symbol_filename_ftype> fun,
> > objfile_debug_name (this),
> > need_fullname);
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > iter->map_symbol_filenames (this, fun, need_fullname);
> > }
> >
> > @@ -464,7 +464,7 @@ objfile::compute_main_name ()
> > "qf->compute_main_name (%s)\n",
> > objfile_debug_name (this));
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > iter->compute_main_name (this);
> > }
> >
> > @@ -478,7 +478,7 @@ objfile::find_compunit_symtab_by_address (CORE_ADDR address)
> > hex_string (address));
> >
> > struct compunit_symtab *result = NULL;
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > {
> > result = iter->find_compunit_symtab_by_address (this, address);
> > if (result != nullptr)
> > @@ -503,7 +503,7 @@ objfile::lookup_global_symbol_language (const char *name,
> > enum language result = language_unknown;
> > *symbol_found_p = false;
> >
> > - for (const auto &iter : qf)
> > + for (const auto &iter : qf_safe ())
> > {
> > result = iter->lookup_global_symbol_language (this, name, domain,
> > symbol_found_p);
> > diff --git a/gdb/testsuite/gdb.python/py-objfile.exp b/gdb/testsuite/gdb.python/py-objfile.exp
> > index 2f5b7752b43..8460c300de0 100644
> > --- a/gdb/testsuite/gdb.python/py-objfile.exp
> > +++ b/gdb/testsuite/gdb.python/py-objfile.exp
> > @@ -135,7 +135,7 @@ gdb_test "p main" "= {<text variable, no debug info>} $hex <main>" \
> > gdb_py_test_silent_cmd "python objfile.add_separate_debug_file(\"${binfile}\")" \
> > "Add separate debug file file" 1
> >
> > -gdb_py_test_silent_cmd "python sep_objfile = gdb.objfiles()\[0\]" \
> > +gdb_py_test_silent_cmd "python sep_objfile = gdb.objfiles()\[1\]" \
> > "Get separate debug info objfile" 1
> >
> > gdb_test "python print (sep_objfile.owner.filename)" "${testfile}2" \
> > diff --git a/gdbsupport/safe-iterator.h b/gdbsupport/safe-iterator.h
> > index f69f3896758..be23cf7920a 100644
> > --- a/gdbsupport/safe-iterator.h
> > +++ b/gdbsupport/safe-iterator.h
> > @@ -136,4 +136,119 @@ class basic_safe_range
> > Range m_range;
> > };
> >
> > +/* A reverse basic_safe_iterator. See basic_safe_iterator for intended use. */
> > +
> > +template<typename Iterator>
> > +class basic_safe_reverse_iterator
> > +{
> > +public:
> > + typedef basic_safe_reverse_iterator self_type;
> > + typedef typename Iterator::value_type value_type;
> > + typedef typename Iterator::reference reference;
> > + typedef typename Iterator::pointer pointer;
> > + typedef typename Iterator::iterator_category iterator_category;
> > + typedef typename Iterator::difference_type difference_type;
> > +
> > + /* Construct the iterator using ARG, and construct the end iterator
> > + using ARG2. ARG and ARG2 should be forward iterators, typically
> > + from begin and end methods, respectively.
> > +
> > + For example if ARG1 is created with container.begin and ARG2 is
> > + is created with container.end, then the basic_safe_reverse_iterator
> > + will traverse from the last element in the container to the first
> > + element in the container. */
> > + template<typename Arg>
> > + explicit basic_safe_reverse_iterator (Arg &&arg, Arg &&arg2)
> > + : m_begin (std::forward<Arg> (arg)),
> > + m_end (std::forward<Arg> (arg2)),
> > + m_it (m_end),
> > + m_next (m_end)
> > + {
> > + /* M_IT and M_NEXT are initialized as one-past-end. Set M_IT to point
> > + to the last element and set M_NEXT to point to the second last element,
> > + if such elements exist. */
> > + if (m_it != m_begin)
> > + {
> > + --m_it;
> > +
> > + if (m_it != m_begin)
> > + {
> > + --m_next;
> > + --m_next;
> > + }
> > + }
> > + }
> > +
> > + typename std::invoke_result<decltype(&Iterator::operator*), Iterator>::type
> > + operator* () const
> > + { return *m_it; }
> > +
> > + self_type &operator++ ()
> > + {
> > + m_it = m_next;
> > +
> > + if (m_it != m_end)
> > + {
> > + /* Use M_BEGIN only if we sure that it is valid. */
> > + if (m_it == m_begin)
> > + m_next = m_end;
> > + else
> > + --m_next;
> > + }
> > +
> > + return *this;
> > + }
> > +
> > + bool operator== (const self_type &other) const
> > + { return m_it == other.m_it; }
> > +
> > + bool operator!= (const self_type &other) const
> > + { return m_it != other.m_it; }
> > +
> > +private:
> > + /* The first element. */
> > + Iterator m_begin {};
> > +
> > + /* A one-past-end iterator. */
> > + Iterator m_end {};
> > +
> > + /* The current element. */
> > + Iterator m_it {};
> > +
> > + /* The next element. Always one element ahead of M_IT. */
> > + Iterator m_next {};
> > +};
> > +
> > +/* A range adapter that wraps a forward range, and then returns
> > + safe reverse iterators wrapping the original range's iterators. */
> > +
> > +template<typename Range>
> > +class basic_safe_reverse_range
> > +{
> > +public:
> > +
> > + typedef basic_safe_reverse_iterator<typename Range::iterator> iterator;
> > +
> > + /* RANGE must be a forward range. basic_safe_reverse_iterators
> > + will be used to traverse the forward range from the last element
> > + to the first. */
> > + explicit basic_safe_reverse_range (Range range)
> > + : m_range (range)
> > + {
> > + }
> > +
> > + iterator begin ()
> > + {
> > + return iterator (m_range.begin (), m_range.end ());
> > + }
> > +
> > + iterator end ()
> > + {
> > + return iterator (m_range.end (), m_range.end ());
> > + }
> > +
> > +private:
> > +
> > + Range m_range;
> > +};
> > #endif /* COMMON_SAFE_ITERATOR_H */
> > --
> > 2.43.0
> >
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING*3][PATCH 1/3 v3] gdb/progspace: Add reverse safe iterator
2024-05-31 20:58 ` [PING*2][PATCH " Aaron Merey
@ 2024-06-07 20:48 ` Aaron Merey
2024-06-14 18:23 ` [PING*2][PATCH " Aaron Merey
1 sibling, 0 replies; 18+ messages in thread
From: Aaron Merey @ 2024-06-07 20:48 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 31, 2024 at 4:58 PM Aaron Merey <amerey@redhat.com> wrote:
>
> Ping
>
> Thanks,
> Aaron
>
> On Fri, May 17, 2024 at 10:09 AM Aaron Merey <amerey@redhat.com> wrote:
> >
> > Ping
> >
> > Thanks,
> > Aaron
> >
> > On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
> > >
> > > This is a reposting of v3 from
> > > https://sourceware.org/pipermail/gdb-patches/2024-January/205955.html
> > >
> > > This patch changes progspace objfile_list insertion so that separate
> > > debug objfiles are placed into the list after the parent objfile,
> > > instead of before. Additionally qf_require_partial_symbols now returns
> > > a safe_range.
> > >
> > > These changes are intended to prepare gdb for on-demand debuginfo
> > > downloading and the downloading of .gdb_index sections.
> > >
> > > With on-demand downloading enabled, gdb might need to delete a
> > > .gdb_index quick_symbol_functions from a parent objfile while looping
> > > the objfile's list of quick_symbol_functions because the separate
> > > debug objfile has just been downloaded. The use of a safe_range
> > > prevents this removal from causing iterator invalidation.
> > >
> > > gdb might also download a debuginfo file during symtab expansion.
> > > In this case an objfile will be added to the current progspace's
> > > objfiles_list during iteration over the list (for example, in
> > > iterate_over_symtabs). We want these loops to also iterate over
> > > newly downloaded objfiles. So objfiles need to be inserted into
> > > objfiles_list after their parent since it is during the search of
> > > the parent objfile for some symbol or filename that the separate
> > > debug objfile might be downloaded.
> > >
> > > To facilitate the safe deletion of objfiles, this patch also adds
> > > basic_safe_reverse_range and basic_safe_reverse_iterator. This allows
> > > objfiles to be removed from the objfiles_list in a loop without iterator
> > > invalidation.
> > >
> > > If a forward safe iterator were to be used, the deletion of an
> > > objfile could invalidate the safe iterator's reference to the next
> > > objfile in the objfiles_list. This can happen when the deletion
> > > of an objfile causes the deletion of a separate debug objfile that
> > > happens to the be next element in the objfiles_list.
> > >
> > > The standard reverse iterator is not suitable for safe objfile deletion.
> > > In order to safely delete the first objfile in the objfiles_list, the
> > > standard reverse iterator's underlying begin iterator would have to be
> > > decremented, resulting in undefined behavior.
> > >
> > > A small change was also made to a testcase in py-objfile.exp to
> > > account for the new placement of separate debug objfiles in
> > > objfiles_list.
> > > ---
> > > gdb/jit.c | 7 +-
> > > gdb/objfiles.c | 14 +--
> > > gdb/objfiles.h | 17 +++-
> > > gdb/progspace.c | 19 +++-
> > > gdb/progspace.h | 31 ++++---
> > > gdb/symfile-debug.c | 34 +++----
> > > gdb/testsuite/gdb.python/py-objfile.exp | 2 +-
> > > gdbsupport/safe-iterator.h | 115 ++++++++++++++++++++++++
> > > 8 files changed, 195 insertions(+), 44 deletions(-)
> > >
> > > diff --git a/gdb/jit.c b/gdb/jit.c
> > > index 797be95a8da..90f9159fa9d 100644
> > > --- a/gdb/jit.c
> > > +++ b/gdb/jit.c
> > > @@ -1240,11 +1240,10 @@ jit_breakpoint_re_set (void)
> > > static void
> > > jit_inferior_exit_hook (struct inferior *inf)
> > > {
> > > - for (objfile *objf : current_program_space->objfiles_safe ())
> > > + current_program_space->unlink_objfiles_if ([&] (const objfile *objf)
> > > {
> > > - if (objf->jited_data != nullptr && objf->jited_data->addr != 0)
> > > - objf->unlink ();
> > > - }
> > > + return (objf->jited_data != nullptr) && (objf->jited_data->addr != 0);
> > > + });
> > > }
> > >
> > > void
> > > diff --git a/gdb/objfiles.c b/gdb/objfiles.c
> > > index ba88ed1bc41..70017c7fb32 100644
> > > --- a/gdb/objfiles.c
> > > +++ b/gdb/objfiles.c
> > > @@ -469,6 +469,12 @@ objfile::unlink ()
> > > current_program_space->remove_objfile (this);
> > > }
> > >
> > > +qf_safe_range
> > > +objfile::qf_safe ()
> > > +{
> > > + return qf_safe_range (qf_range (qf.begin (), qf.end ()));
> > > +}
> > > +
> > > /* Free all separate debug objfile of OBJFILE, but don't free OBJFILE
> > > itself. */
> > >
> > > @@ -792,14 +798,12 @@ have_full_symbols (void)
> > > void
> > > objfile_purge_solibs (void)
> > > {
> > > - for (objfile *objf : current_program_space->objfiles_safe ())
> > > + current_program_space->unlink_objfiles_if ([&] (const objfile *objf)
> > > {
> > > /* We assume that the solib package has been purged already, or will
> > > be soon. */
> > > -
> > > - if (!(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED))
> > > - objf->unlink ();
> > > - }
> > > + return !(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED);
> > > + });
> > > }
> > >
> > >
> > > diff --git a/gdb/objfiles.h b/gdb/objfiles.h
> > > index 8b8b7182e87..1c8f79a2826 100644
> > > --- a/gdb/objfiles.h
> > > +++ b/gdb/objfiles.h
> > > @@ -370,6 +370,12 @@ class separate_debug_iterator
> > >
> > > typedef iterator_range<separate_debug_iterator> separate_debug_range;
> > >
> > > +/* See objfile::qf_safe. */
> > > +
> > > +using qf_list = std::forward_list<quick_symbol_functions_up>;
> > > +using qf_range = iterator_range<qf_list::iterator>;
> > > +using qf_safe_range = basic_safe_range<qf_range>;
> > > +
> > > /* Sections in an objfile. The section offsets are stored in the
> > > OBJFILE. */
> > >
> > > @@ -767,8 +773,15 @@ struct objfile
> > > const struct sym_fns *sf = nullptr;
> > >
> > > /* The "quick" (aka partial) symbol functions for this symbol
> > > - reader. */
> > > - std::forward_list<quick_symbol_functions_up> qf;
> > > + reader. Many quick_symbol_functions methods may result
> > > + in the deletion of a quick_symbol_functions from this
> > > + qf_list. It is recommended that qf_safe be used to iterate
> > > + over the qf_list. */
> > > + qf_list qf;
> > > +
> > > + /* Returns an iterable object that allows for safe deletion during
> > > + iteration. See gdbsupport/safe-iterator.h. */
> > > + qf_safe_range qf_safe ();
> > >
> > > /* Per objfile data-pointers required by other GDB modules. */
> > >
> > > diff --git a/gdb/progspace.c b/gdb/progspace.c
> > > index 0deca3f2376..b1b2dfb30c8 100644
> > > --- a/gdb/progspace.c
> > > +++ b/gdb/progspace.c
> > > @@ -140,19 +140,19 @@ program_space::free_all_objfiles ()
> > >
> > > void
> > > program_space::add_objfile (std::unique_ptr<objfile> &&objfile,
> > > - struct objfile *before)
> > > + struct objfile *after)
> > > {
> > > - if (before == nullptr)
> > > + if (after == nullptr)
> > > objfiles_list.push_back (std::move (objfile));
> > > else
> > > {
> > > auto iter = std::find_if (objfiles_list.begin (), objfiles_list.end (),
> > > [=] (const std::unique_ptr<::objfile> &objf)
> > > {
> > > - return objf.get () == before;
> > > + return objf.get () == after;
> > > });
> > > gdb_assert (iter != objfiles_list.end ());
> > > - objfiles_list.insert (iter, std::move (objfile));
> > > + objfiles_list.insert (++iter, std::move (objfile));
> > > }
> > > }
> > >
> > > @@ -181,6 +181,17 @@ program_space::remove_objfile (struct objfile *objfile)
> > >
> > > /* See progspace.h. */
> > >
> > > +void
> > > +program_space::unlink_objfiles_if
> > > + (gdb::function_view<bool (const objfile *objfile)> predicate)
> > > +{
> > > + for (auto &it : objfiles_safe ())
> > > + if (predicate (it.get ()))
> > > + it->unlink ();
> > > +}
> > > +
> > > +/* See progspace.h. */
> > > +
> > > struct objfile *
> > > program_space::objfile_for_address (CORE_ADDR address)
> > > {
> > > diff --git a/gdb/progspace.h b/gdb/progspace.h
> > > index bbf54efa07a..4d5ac991849 100644
> > > --- a/gdb/progspace.h
> > > +++ b/gdb/progspace.h
> > > @@ -250,28 +250,32 @@ struct program_space
> > > unwrapping_objfile_iterator (objfiles_list.end ()));
> > > }
> > >
> > > - using objfiles_safe_range = basic_safe_range<objfiles_range>;
> > > + using objfiles_it_range = iterator_range<objfile_list::iterator>;
> > > + using objfiles_safe_reverse_range
> > > + = basic_safe_reverse_range<objfiles_it_range>;
> > >
> > > /* An iterable object that can be used to iterate over all objfiles.
> > > The basic use is in a foreach, like:
> > >
> > > for (objfile *objf : pspace->objfiles_safe ()) { ... }
> > >
> > > - This variant uses a basic_safe_iterator so that objfiles can be
> > > - deleted during iteration. */
> > > - objfiles_safe_range objfiles_safe ()
> > > + This variant uses a basic_safe_reverse_iterator so that objfiles
> > > + can be deleted during iteration.
> > > +
> > > + The use of a reverse iterator helps ensure that separate debug
> > > + objfiles are deleted before their parent objfile. This prevents
> > > + iterator invalidation due to the deletion of a parent objfile. */
> > > + objfiles_safe_reverse_range objfiles_safe ()
> > > {
> > > - return objfiles_safe_range
> > > - (objfiles_range
> > > - (unwrapping_objfile_iterator (objfiles_list.begin ()),
> > > - unwrapping_objfile_iterator (objfiles_list.end ())));
> > > + return objfiles_safe_reverse_range
> > > + (objfiles_it_range (objfiles_list.begin (), objfiles_list.end ()));
> > > }
> > >
> > > - /* Add OBJFILE to the list of objfiles, putting it just before
> > > - BEFORE. If BEFORE is nullptr, it will go at the end of the
> > > + /* Add OBJFILE to the list of objfiles, putting it just after
> > > + AFTER. If AFTER is nullptr, it will go at the end of the
> > > list. */
> > > void add_objfile (std::unique_ptr<objfile> &&objfile,
> > > - struct objfile *before);
> > > + struct objfile *after);
> > >
> > > /* Remove OBJFILE from the list of objfiles. */
> > > void remove_objfile (struct objfile *objfile);
> > > @@ -286,6 +290,11 @@ struct program_space
> > > /* Free all the objfiles associated with this program space. */
> > > void free_all_objfiles ();
> > >
> > > + /* Unlink all objfiles associated with this program space for which
> > > + PREDICATE evaluates to true. */
> > > + void unlink_objfiles_if
> > > + (gdb::function_view<bool (const objfile *objfile)> predicate);
> > > +
> > > /* Return the objfile containing ADDRESS, or nullptr if the address
> > > is outside all objfiles in this progspace. */
> > > struct objfile *objfile_for_address (CORE_ADDR address);
> > > diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c
> > > index 86c7010b2d2..330a520dca6 100644
> > > --- a/gdb/symfile-debug.c
> > > +++ b/gdb/symfile-debug.c
> > > @@ -84,7 +84,7 @@ objfile::has_partial_symbols ()
> > > them, then that is an indication that they are in fact available. Without
> > > this function the symbols may have been already read in but they also may
> > > not be present in this objfile. */
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > retval = iter->has_symbols (this);
> > > if (retval)
> > > @@ -107,7 +107,7 @@ objfile::has_unexpanded_symtabs ()
> > > objfile_debug_name (this));
> > >
> > > bool result = false;
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > if (iter->has_unexpanded_symtabs (this))
> > > {
> > > @@ -132,7 +132,7 @@ objfile::find_last_source_symtab ()
> > > gdb_printf (gdb_stdlog, "qf->find_last_source_symtab (%s)\n",
> > > objfile_debug_name (this));
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > retval = iter->find_last_source_symtab (this);
> > > if (retval != nullptr)
> > > @@ -165,7 +165,7 @@ objfile::forget_cached_source_info ()
> > > }
> > > }
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->forget_cached_source_info (this);
> > > }
> > >
> > > @@ -212,7 +212,7 @@ objfile::map_symtabs_matching_filename
> > > return result;
> > > };
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > if (!iter->expand_symtabs_matching (this,
> > > match_one_filename,
> > > @@ -275,7 +275,7 @@ objfile::lookup_symbol (block_enum kind, const lookup_name_info &name,
> > > return true;
> > > };
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > if (!iter->expand_symtabs_matching (this,
> > > nullptr,
> > > @@ -305,7 +305,7 @@ objfile::print_stats (bool print_bcache)
> > > gdb_printf (gdb_stdlog, "qf->print_stats (%s, %d)\n",
> > > objfile_debug_name (this), print_bcache);
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->print_stats (this, print_bcache);
> > > }
> > >
> > > @@ -316,7 +316,7 @@ objfile::dump ()
> > > gdb_printf (gdb_stdlog, "qf->dump (%s)\n",
> > > objfile_debug_name (this));
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->dump (this);
> > > }
> > >
> > > @@ -331,7 +331,7 @@ objfile::expand_symtabs_for_function (const char *func_name)
> > > lookup_name_info base_lookup (func_name, symbol_name_match_type::FULL);
> > > lookup_name_info lookup_name = base_lookup.make_ignore_params ();
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->expand_symtabs_matching (this,
> > > nullptr,
> > > &lookup_name,
> > > @@ -349,7 +349,7 @@ objfile::expand_all_symtabs ()
> > > gdb_printf (gdb_stdlog, "qf->expand_all_symtabs (%s)\n",
> > > objfile_debug_name (this));
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->expand_all_symtabs (this);
> > > }
> > >
> > > @@ -367,7 +367,7 @@ objfile::expand_symtabs_with_fullname (const char *fullname)
> > > return filename_cmp (basenames ? basename : fullname, filename) == 0;
> > > };
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->expand_symtabs_matching (this,
> > > file_matcher,
> > > nullptr,
> > > @@ -399,7 +399,7 @@ objfile::expand_symtabs_matching
> > > host_address_to_string (&expansion_notify),
> > > domain_name (domain).c_str ());
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > if (!iter->expand_symtabs_matching (this, file_matcher, lookup_name,
> > > symbol_matcher, expansion_notify,
> > > search_flags, domain))
> > > @@ -424,7 +424,7 @@ objfile::find_pc_sect_compunit_symtab (struct bound_minimal_symbol msymbol,
> > > host_address_to_string (section),
> > > warn_if_readin);
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > retval = iter->find_pc_sect_compunit_symtab (this, msymbol, pc, section,
> > > warn_if_readin);
> > > @@ -452,7 +452,7 @@ objfile::map_symbol_filenames (gdb::function_view<symbol_filename_ftype> fun,
> > > objfile_debug_name (this),
> > > need_fullname);
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->map_symbol_filenames (this, fun, need_fullname);
> > > }
> > >
> > > @@ -464,7 +464,7 @@ objfile::compute_main_name ()
> > > "qf->compute_main_name (%s)\n",
> > > objfile_debug_name (this));
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->compute_main_name (this);
> > > }
> > >
> > > @@ -478,7 +478,7 @@ objfile::find_compunit_symtab_by_address (CORE_ADDR address)
> > > hex_string (address));
> > >
> > > struct compunit_symtab *result = NULL;
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > result = iter->find_compunit_symtab_by_address (this, address);
> > > if (result != nullptr)
> > > @@ -503,7 +503,7 @@ objfile::lookup_global_symbol_language (const char *name,
> > > enum language result = language_unknown;
> > > *symbol_found_p = false;
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > result = iter->lookup_global_symbol_language (this, name, domain,
> > > symbol_found_p);
> > > diff --git a/gdb/testsuite/gdb.python/py-objfile.exp b/gdb/testsuite/gdb.python/py-objfile.exp
> > > index 2f5b7752b43..8460c300de0 100644
> > > --- a/gdb/testsuite/gdb.python/py-objfile.exp
> > > +++ b/gdb/testsuite/gdb.python/py-objfile.exp
> > > @@ -135,7 +135,7 @@ gdb_test "p main" "= {<text variable, no debug info>} $hex <main>" \
> > > gdb_py_test_silent_cmd "python objfile.add_separate_debug_file(\"${binfile}\")" \
> > > "Add separate debug file file" 1
> > >
> > > -gdb_py_test_silent_cmd "python sep_objfile = gdb.objfiles()\[0\]" \
> > > +gdb_py_test_silent_cmd "python sep_objfile = gdb.objfiles()\[1\]" \
> > > "Get separate debug info objfile" 1
> > >
> > > gdb_test "python print (sep_objfile.owner.filename)" "${testfile}2" \
> > > diff --git a/gdbsupport/safe-iterator.h b/gdbsupport/safe-iterator.h
> > > index f69f3896758..be23cf7920a 100644
> > > --- a/gdbsupport/safe-iterator.h
> > > +++ b/gdbsupport/safe-iterator.h
> > > @@ -136,4 +136,119 @@ class basic_safe_range
> > > Range m_range;
> > > };
> > >
> > > +/* A reverse basic_safe_iterator. See basic_safe_iterator for intended use. */
> > > +
> > > +template<typename Iterator>
> > > +class basic_safe_reverse_iterator
> > > +{
> > > +public:
> > > + typedef basic_safe_reverse_iterator self_type;
> > > + typedef typename Iterator::value_type value_type;
> > > + typedef typename Iterator::reference reference;
> > > + typedef typename Iterator::pointer pointer;
> > > + typedef typename Iterator::iterator_category iterator_category;
> > > + typedef typename Iterator::difference_type difference_type;
> > > +
> > > + /* Construct the iterator using ARG, and construct the end iterator
> > > + using ARG2. ARG and ARG2 should be forward iterators, typically
> > > + from begin and end methods, respectively.
> > > +
> > > + For example if ARG1 is created with container.begin and ARG2 is
> > > + is created with container.end, then the basic_safe_reverse_iterator
> > > + will traverse from the last element in the container to the first
> > > + element in the container. */
> > > + template<typename Arg>
> > > + explicit basic_safe_reverse_iterator (Arg &&arg, Arg &&arg2)
> > > + : m_begin (std::forward<Arg> (arg)),
> > > + m_end (std::forward<Arg> (arg2)),
> > > + m_it (m_end),
> > > + m_next (m_end)
> > > + {
> > > + /* M_IT and M_NEXT are initialized as one-past-end. Set M_IT to point
> > > + to the last element and set M_NEXT to point to the second last element,
> > > + if such elements exist. */
> > > + if (m_it != m_begin)
> > > + {
> > > + --m_it;
> > > +
> > > + if (m_it != m_begin)
> > > + {
> > > + --m_next;
> > > + --m_next;
> > > + }
> > > + }
> > > + }
> > > +
> > > + typename std::invoke_result<decltype(&Iterator::operator*), Iterator>::type
> > > + operator* () const
> > > + { return *m_it; }
> > > +
> > > + self_type &operator++ ()
> > > + {
> > > + m_it = m_next;
> > > +
> > > + if (m_it != m_end)
> > > + {
> > > + /* Use M_BEGIN only if we sure that it is valid. */
> > > + if (m_it == m_begin)
> > > + m_next = m_end;
> > > + else
> > > + --m_next;
> > > + }
> > > +
> > > + return *this;
> > > + }
> > > +
> > > + bool operator== (const self_type &other) const
> > > + { return m_it == other.m_it; }
> > > +
> > > + bool operator!= (const self_type &other) const
> > > + { return m_it != other.m_it; }
> > > +
> > > +private:
> > > + /* The first element. */
> > > + Iterator m_begin {};
> > > +
> > > + /* A one-past-end iterator. */
> > > + Iterator m_end {};
> > > +
> > > + /* The current element. */
> > > + Iterator m_it {};
> > > +
> > > + /* The next element. Always one element ahead of M_IT. */
> > > + Iterator m_next {};
> > > +};
> > > +
> > > +/* A range adapter that wraps a forward range, and then returns
> > > + safe reverse iterators wrapping the original range's iterators. */
> > > +
> > > +template<typename Range>
> > > +class basic_safe_reverse_range
> > > +{
> > > +public:
> > > +
> > > + typedef basic_safe_reverse_iterator<typename Range::iterator> iterator;
> > > +
> > > + /* RANGE must be a forward range. basic_safe_reverse_iterators
> > > + will be used to traverse the forward range from the last element
> > > + to the first. */
> > > + explicit basic_safe_reverse_range (Range range)
> > > + : m_range (range)
> > > + {
> > > + }
> > > +
> > > + iterator begin ()
> > > + {
> > > + return iterator (m_range.begin (), m_range.end ());
> > > + }
> > > +
> > > + iterator end ()
> > > + {
> > > + return iterator (m_range.end (), m_range.end ());
> > > + }
> > > +
> > > +private:
> > > +
> > > + Range m_range;
> > > +};
> > > #endif /* COMMON_SAFE_ITERATOR_H */
> > > --
> > > 2.43.0
> > >
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PING*2][PATCH 1/3 v3] gdb/progspace: Add reverse safe iterator
2024-05-31 20:58 ` [PING*2][PATCH " Aaron Merey
2024-06-07 20:48 ` [PING*3][PATCH " Aaron Merey
@ 2024-06-14 18:23 ` Aaron Merey
1 sibling, 0 replies; 18+ messages in thread
From: Aaron Merey @ 2024-06-14 18:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 31, 2024 at 4:58 PM Aaron Merey <amerey@redhat.com> wrote:
>
> Ping
>
> Thanks,
> Aaron
>
> On Fri, May 17, 2024 at 10:09 AM Aaron Merey <amerey@redhat.com> wrote:
> >
> > Ping
> >
> > Thanks,
> > Aaron
> >
> > On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
> > >
> > > This is a reposting of v3 from
> > > https://sourceware.org/pipermail/gdb-patches/2024-January/205955.html
> > >
> > > This patch changes progspace objfile_list insertion so that separate
> > > debug objfiles are placed into the list after the parent objfile,
> > > instead of before. Additionally qf_require_partial_symbols now returns
> > > a safe_range.
> > >
> > > These changes are intended to prepare gdb for on-demand debuginfo
> > > downloading and the downloading of .gdb_index sections.
> > >
> > > With on-demand downloading enabled, gdb might need to delete a
> > > .gdb_index quick_symbol_functions from a parent objfile while looping
> > > the objfile's list of quick_symbol_functions because the separate
> > > debug objfile has just been downloaded. The use of a safe_range
> > > prevents this removal from causing iterator invalidation.
> > >
> > > gdb might also download a debuginfo file during symtab expansion.
> > > In this case an objfile will be added to the current progspace's
> > > objfiles_list during iteration over the list (for example, in
> > > iterate_over_symtabs). We want these loops to also iterate over
> > > newly downloaded objfiles. So objfiles need to be inserted into
> > > objfiles_list after their parent since it is during the search of
> > > the parent objfile for some symbol or filename that the separate
> > > debug objfile might be downloaded.
> > >
> > > To facilitate the safe deletion of objfiles, this patch also adds
> > > basic_safe_reverse_range and basic_safe_reverse_iterator. This allows
> > > objfiles to be removed from the objfiles_list in a loop without iterator
> > > invalidation.
> > >
> > > If a forward safe iterator were to be used, the deletion of an
> > > objfile could invalidate the safe iterator's reference to the next
> > > objfile in the objfiles_list. This can happen when the deletion
> > > of an objfile causes the deletion of a separate debug objfile that
> > > happens to the be next element in the objfiles_list.
> > >
> > > The standard reverse iterator is not suitable for safe objfile deletion.
> > > In order to safely delete the first objfile in the objfiles_list, the
> > > standard reverse iterator's underlying begin iterator would have to be
> > > decremented, resulting in undefined behavior.
> > >
> > > A small change was also made to a testcase in py-objfile.exp to
> > > account for the new placement of separate debug objfiles in
> > > objfiles_list.
> > > ---
> > > gdb/jit.c | 7 +-
> > > gdb/objfiles.c | 14 +--
> > > gdb/objfiles.h | 17 +++-
> > > gdb/progspace.c | 19 +++-
> > > gdb/progspace.h | 31 ++++---
> > > gdb/symfile-debug.c | 34 +++----
> > > gdb/testsuite/gdb.python/py-objfile.exp | 2 +-
> > > gdbsupport/safe-iterator.h | 115 ++++++++++++++++++++++++
> > > 8 files changed, 195 insertions(+), 44 deletions(-)
> > >
> > > diff --git a/gdb/jit.c b/gdb/jit.c
> > > index 797be95a8da..90f9159fa9d 100644
> > > --- a/gdb/jit.c
> > > +++ b/gdb/jit.c
> > > @@ -1240,11 +1240,10 @@ jit_breakpoint_re_set (void)
> > > static void
> > > jit_inferior_exit_hook (struct inferior *inf)
> > > {
> > > - for (objfile *objf : current_program_space->objfiles_safe ())
> > > + current_program_space->unlink_objfiles_if ([&] (const objfile *objf)
> > > {
> > > - if (objf->jited_data != nullptr && objf->jited_data->addr != 0)
> > > - objf->unlink ();
> > > - }
> > > + return (objf->jited_data != nullptr) && (objf->jited_data->addr != 0);
> > > + });
> > > }
> > >
> > > void
> > > diff --git a/gdb/objfiles.c b/gdb/objfiles.c
> > > index ba88ed1bc41..70017c7fb32 100644
> > > --- a/gdb/objfiles.c
> > > +++ b/gdb/objfiles.c
> > > @@ -469,6 +469,12 @@ objfile::unlink ()
> > > current_program_space->remove_objfile (this);
> > > }
> > >
> > > +qf_safe_range
> > > +objfile::qf_safe ()
> > > +{
> > > + return qf_safe_range (qf_range (qf.begin (), qf.end ()));
> > > +}
> > > +
> > > /* Free all separate debug objfile of OBJFILE, but don't free OBJFILE
> > > itself. */
> > >
> > > @@ -792,14 +798,12 @@ have_full_symbols (void)
> > > void
> > > objfile_purge_solibs (void)
> > > {
> > > - for (objfile *objf : current_program_space->objfiles_safe ())
> > > + current_program_space->unlink_objfiles_if ([&] (const objfile *objf)
> > > {
> > > /* We assume that the solib package has been purged already, or will
> > > be soon. */
> > > -
> > > - if (!(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED))
> > > - objf->unlink ();
> > > - }
> > > + return !(objf->flags & OBJF_USERLOADED) && (objf->flags & OBJF_SHARED);
> > > + });
> > > }
> > >
> > >
> > > diff --git a/gdb/objfiles.h b/gdb/objfiles.h
> > > index 8b8b7182e87..1c8f79a2826 100644
> > > --- a/gdb/objfiles.h
> > > +++ b/gdb/objfiles.h
> > > @@ -370,6 +370,12 @@ class separate_debug_iterator
> > >
> > > typedef iterator_range<separate_debug_iterator> separate_debug_range;
> > >
> > > +/* See objfile::qf_safe. */
> > > +
> > > +using qf_list = std::forward_list<quick_symbol_functions_up>;
> > > +using qf_range = iterator_range<qf_list::iterator>;
> > > +using qf_safe_range = basic_safe_range<qf_range>;
> > > +
> > > /* Sections in an objfile. The section offsets are stored in the
> > > OBJFILE. */
> > >
> > > @@ -767,8 +773,15 @@ struct objfile
> > > const struct sym_fns *sf = nullptr;
> > >
> > > /* The "quick" (aka partial) symbol functions for this symbol
> > > - reader. */
> > > - std::forward_list<quick_symbol_functions_up> qf;
> > > + reader. Many quick_symbol_functions methods may result
> > > + in the deletion of a quick_symbol_functions from this
> > > + qf_list. It is recommended that qf_safe be used to iterate
> > > + over the qf_list. */
> > > + qf_list qf;
> > > +
> > > + /* Returns an iterable object that allows for safe deletion during
> > > + iteration. See gdbsupport/safe-iterator.h. */
> > > + qf_safe_range qf_safe ();
> > >
> > > /* Per objfile data-pointers required by other GDB modules. */
> > >
> > > diff --git a/gdb/progspace.c b/gdb/progspace.c
> > > index 0deca3f2376..b1b2dfb30c8 100644
> > > --- a/gdb/progspace.c
> > > +++ b/gdb/progspace.c
> > > @@ -140,19 +140,19 @@ program_space::free_all_objfiles ()
> > >
> > > void
> > > program_space::add_objfile (std::unique_ptr<objfile> &&objfile,
> > > - struct objfile *before)
> > > + struct objfile *after)
> > > {
> > > - if (before == nullptr)
> > > + if (after == nullptr)
> > > objfiles_list.push_back (std::move (objfile));
> > > else
> > > {
> > > auto iter = std::find_if (objfiles_list.begin (), objfiles_list.end (),
> > > [=] (const std::unique_ptr<::objfile> &objf)
> > > {
> > > - return objf.get () == before;
> > > + return objf.get () == after;
> > > });
> > > gdb_assert (iter != objfiles_list.end ());
> > > - objfiles_list.insert (iter, std::move (objfile));
> > > + objfiles_list.insert (++iter, std::move (objfile));
> > > }
> > > }
> > >
> > > @@ -181,6 +181,17 @@ program_space::remove_objfile (struct objfile *objfile)
> > >
> > > /* See progspace.h. */
> > >
> > > +void
> > > +program_space::unlink_objfiles_if
> > > + (gdb::function_view<bool (const objfile *objfile)> predicate)
> > > +{
> > > + for (auto &it : objfiles_safe ())
> > > + if (predicate (it.get ()))
> > > + it->unlink ();
> > > +}
> > > +
> > > +/* See progspace.h. */
> > > +
> > > struct objfile *
> > > program_space::objfile_for_address (CORE_ADDR address)
> > > {
> > > diff --git a/gdb/progspace.h b/gdb/progspace.h
> > > index bbf54efa07a..4d5ac991849 100644
> > > --- a/gdb/progspace.h
> > > +++ b/gdb/progspace.h
> > > @@ -250,28 +250,32 @@ struct program_space
> > > unwrapping_objfile_iterator (objfiles_list.end ()));
> > > }
> > >
> > > - using objfiles_safe_range = basic_safe_range<objfiles_range>;
> > > + using objfiles_it_range = iterator_range<objfile_list::iterator>;
> > > + using objfiles_safe_reverse_range
> > > + = basic_safe_reverse_range<objfiles_it_range>;
> > >
> > > /* An iterable object that can be used to iterate over all objfiles.
> > > The basic use is in a foreach, like:
> > >
> > > for (objfile *objf : pspace->objfiles_safe ()) { ... }
> > >
> > > - This variant uses a basic_safe_iterator so that objfiles can be
> > > - deleted during iteration. */
> > > - objfiles_safe_range objfiles_safe ()
> > > + This variant uses a basic_safe_reverse_iterator so that objfiles
> > > + can be deleted during iteration.
> > > +
> > > + The use of a reverse iterator helps ensure that separate debug
> > > + objfiles are deleted before their parent objfile. This prevents
> > > + iterator invalidation due to the deletion of a parent objfile. */
> > > + objfiles_safe_reverse_range objfiles_safe ()
> > > {
> > > - return objfiles_safe_range
> > > - (objfiles_range
> > > - (unwrapping_objfile_iterator (objfiles_list.begin ()),
> > > - unwrapping_objfile_iterator (objfiles_list.end ())));
> > > + return objfiles_safe_reverse_range
> > > + (objfiles_it_range (objfiles_list.begin (), objfiles_list.end ()));
> > > }
> > >
> > > - /* Add OBJFILE to the list of objfiles, putting it just before
> > > - BEFORE. If BEFORE is nullptr, it will go at the end of the
> > > + /* Add OBJFILE to the list of objfiles, putting it just after
> > > + AFTER. If AFTER is nullptr, it will go at the end of the
> > > list. */
> > > void add_objfile (std::unique_ptr<objfile> &&objfile,
> > > - struct objfile *before);
> > > + struct objfile *after);
> > >
> > > /* Remove OBJFILE from the list of objfiles. */
> > > void remove_objfile (struct objfile *objfile);
> > > @@ -286,6 +290,11 @@ struct program_space
> > > /* Free all the objfiles associated with this program space. */
> > > void free_all_objfiles ();
> > >
> > > + /* Unlink all objfiles associated with this program space for which
> > > + PREDICATE evaluates to true. */
> > > + void unlink_objfiles_if
> > > + (gdb::function_view<bool (const objfile *objfile)> predicate);
> > > +
> > > /* Return the objfile containing ADDRESS, or nullptr if the address
> > > is outside all objfiles in this progspace. */
> > > struct objfile *objfile_for_address (CORE_ADDR address);
> > > diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c
> > > index 86c7010b2d2..330a520dca6 100644
> > > --- a/gdb/symfile-debug.c
> > > +++ b/gdb/symfile-debug.c
> > > @@ -84,7 +84,7 @@ objfile::has_partial_symbols ()
> > > them, then that is an indication that they are in fact available. Without
> > > this function the symbols may have been already read in but they also may
> > > not be present in this objfile. */
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > retval = iter->has_symbols (this);
> > > if (retval)
> > > @@ -107,7 +107,7 @@ objfile::has_unexpanded_symtabs ()
> > > objfile_debug_name (this));
> > >
> > > bool result = false;
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > if (iter->has_unexpanded_symtabs (this))
> > > {
> > > @@ -132,7 +132,7 @@ objfile::find_last_source_symtab ()
> > > gdb_printf (gdb_stdlog, "qf->find_last_source_symtab (%s)\n",
> > > objfile_debug_name (this));
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > retval = iter->find_last_source_symtab (this);
> > > if (retval != nullptr)
> > > @@ -165,7 +165,7 @@ objfile::forget_cached_source_info ()
> > > }
> > > }
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->forget_cached_source_info (this);
> > > }
> > >
> > > @@ -212,7 +212,7 @@ objfile::map_symtabs_matching_filename
> > > return result;
> > > };
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > if (!iter->expand_symtabs_matching (this,
> > > match_one_filename,
> > > @@ -275,7 +275,7 @@ objfile::lookup_symbol (block_enum kind, const lookup_name_info &name,
> > > return true;
> > > };
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > if (!iter->expand_symtabs_matching (this,
> > > nullptr,
> > > @@ -305,7 +305,7 @@ objfile::print_stats (bool print_bcache)
> > > gdb_printf (gdb_stdlog, "qf->print_stats (%s, %d)\n",
> > > objfile_debug_name (this), print_bcache);
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->print_stats (this, print_bcache);
> > > }
> > >
> > > @@ -316,7 +316,7 @@ objfile::dump ()
> > > gdb_printf (gdb_stdlog, "qf->dump (%s)\n",
> > > objfile_debug_name (this));
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->dump (this);
> > > }
> > >
> > > @@ -331,7 +331,7 @@ objfile::expand_symtabs_for_function (const char *func_name)
> > > lookup_name_info base_lookup (func_name, symbol_name_match_type::FULL);
> > > lookup_name_info lookup_name = base_lookup.make_ignore_params ();
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->expand_symtabs_matching (this,
> > > nullptr,
> > > &lookup_name,
> > > @@ -349,7 +349,7 @@ objfile::expand_all_symtabs ()
> > > gdb_printf (gdb_stdlog, "qf->expand_all_symtabs (%s)\n",
> > > objfile_debug_name (this));
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->expand_all_symtabs (this);
> > > }
> > >
> > > @@ -367,7 +367,7 @@ objfile::expand_symtabs_with_fullname (const char *fullname)
> > > return filename_cmp (basenames ? basename : fullname, filename) == 0;
> > > };
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->expand_symtabs_matching (this,
> > > file_matcher,
> > > nullptr,
> > > @@ -399,7 +399,7 @@ objfile::expand_symtabs_matching
> > > host_address_to_string (&expansion_notify),
> > > domain_name (domain).c_str ());
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > if (!iter->expand_symtabs_matching (this, file_matcher, lookup_name,
> > > symbol_matcher, expansion_notify,
> > > search_flags, domain))
> > > @@ -424,7 +424,7 @@ objfile::find_pc_sect_compunit_symtab (struct bound_minimal_symbol msymbol,
> > > host_address_to_string (section),
> > > warn_if_readin);
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > retval = iter->find_pc_sect_compunit_symtab (this, msymbol, pc, section,
> > > warn_if_readin);
> > > @@ -452,7 +452,7 @@ objfile::map_symbol_filenames (gdb::function_view<symbol_filename_ftype> fun,
> > > objfile_debug_name (this),
> > > need_fullname);
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->map_symbol_filenames (this, fun, need_fullname);
> > > }
> > >
> > > @@ -464,7 +464,7 @@ objfile::compute_main_name ()
> > > "qf->compute_main_name (%s)\n",
> > > objfile_debug_name (this));
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > iter->compute_main_name (this);
> > > }
> > >
> > > @@ -478,7 +478,7 @@ objfile::find_compunit_symtab_by_address (CORE_ADDR address)
> > > hex_string (address));
> > >
> > > struct compunit_symtab *result = NULL;
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > result = iter->find_compunit_symtab_by_address (this, address);
> > > if (result != nullptr)
> > > @@ -503,7 +503,7 @@ objfile::lookup_global_symbol_language (const char *name,
> > > enum language result = language_unknown;
> > > *symbol_found_p = false;
> > >
> > > - for (const auto &iter : qf)
> > > + for (const auto &iter : qf_safe ())
> > > {
> > > result = iter->lookup_global_symbol_language (this, name, domain,
> > > symbol_found_p);
> > > diff --git a/gdb/testsuite/gdb.python/py-objfile.exp b/gdb/testsuite/gdb.python/py-objfile.exp
> > > index 2f5b7752b43..8460c300de0 100644
> > > --- a/gdb/testsuite/gdb.python/py-objfile.exp
> > > +++ b/gdb/testsuite/gdb.python/py-objfile.exp
> > > @@ -135,7 +135,7 @@ gdb_test "p main" "= {<text variable, no debug info>} $hex <main>" \
> > > gdb_py_test_silent_cmd "python objfile.add_separate_debug_file(\"${binfile}\")" \
> > > "Add separate debug file file" 1
> > >
> > > -gdb_py_test_silent_cmd "python sep_objfile = gdb.objfiles()\[0\]" \
> > > +gdb_py_test_silent_cmd "python sep_objfile = gdb.objfiles()\[1\]" \
> > > "Get separate debug info objfile" 1
> > >
> > > gdb_test "python print (sep_objfile.owner.filename)" "${testfile}2" \
> > > diff --git a/gdbsupport/safe-iterator.h b/gdbsupport/safe-iterator.h
> > > index f69f3896758..be23cf7920a 100644
> > > --- a/gdbsupport/safe-iterator.h
> > > +++ b/gdbsupport/safe-iterator.h
> > > @@ -136,4 +136,119 @@ class basic_safe_range
> > > Range m_range;
> > > };
> > >
> > > +/* A reverse basic_safe_iterator. See basic_safe_iterator for intended use. */
> > > +
> > > +template<typename Iterator>
> > > +class basic_safe_reverse_iterator
> > > +{
> > > +public:
> > > + typedef basic_safe_reverse_iterator self_type;
> > > + typedef typename Iterator::value_type value_type;
> > > + typedef typename Iterator::reference reference;
> > > + typedef typename Iterator::pointer pointer;
> > > + typedef typename Iterator::iterator_category iterator_category;
> > > + typedef typename Iterator::difference_type difference_type;
> > > +
> > > + /* Construct the iterator using ARG, and construct the end iterator
> > > + using ARG2. ARG and ARG2 should be forward iterators, typically
> > > + from begin and end methods, respectively.
> > > +
> > > + For example if ARG1 is created with container.begin and ARG2 is
> > > + is created with container.end, then the basic_safe_reverse_iterator
> > > + will traverse from the last element in the container to the first
> > > + element in the container. */
> > > + template<typename Arg>
> > > + explicit basic_safe_reverse_iterator (Arg &&arg, Arg &&arg2)
> > > + : m_begin (std::forward<Arg> (arg)),
> > > + m_end (std::forward<Arg> (arg2)),
> > > + m_it (m_end),
> > > + m_next (m_end)
> > > + {
> > > + /* M_IT and M_NEXT are initialized as one-past-end. Set M_IT to point
> > > + to the last element and set M_NEXT to point to the second last element,
> > > + if such elements exist. */
> > > + if (m_it != m_begin)
> > > + {
> > > + --m_it;
> > > +
> > > + if (m_it != m_begin)
> > > + {
> > > + --m_next;
> > > + --m_next;
> > > + }
> > > + }
> > > + }
> > > +
> > > + typename std::invoke_result<decltype(&Iterator::operator*), Iterator>::type
> > > + operator* () const
> > > + { return *m_it; }
> > > +
> > > + self_type &operator++ ()
> > > + {
> > > + m_it = m_next;
> > > +
> > > + if (m_it != m_end)
> > > + {
> > > + /* Use M_BEGIN only if we sure that it is valid. */
> > > + if (m_it == m_begin)
> > > + m_next = m_end;
> > > + else
> > > + --m_next;
> > > + }
> > > +
> > > + return *this;
> > > + }
> > > +
> > > + bool operator== (const self_type &other) const
> > > + { return m_it == other.m_it; }
> > > +
> > > + bool operator!= (const self_type &other) const
> > > + { return m_it != other.m_it; }
> > > +
> > > +private:
> > > + /* The first element. */
> > > + Iterator m_begin {};
> > > +
> > > + /* A one-past-end iterator. */
> > > + Iterator m_end {};
> > > +
> > > + /* The current element. */
> > > + Iterator m_it {};
> > > +
> > > + /* The next element. Always one element ahead of M_IT. */
> > > + Iterator m_next {};
> > > +};
> > > +
> > > +/* A range adapter that wraps a forward range, and then returns
> > > + safe reverse iterators wrapping the original range's iterators. */
> > > +
> > > +template<typename Range>
> > > +class basic_safe_reverse_range
> > > +{
> > > +public:
> > > +
> > > + typedef basic_safe_reverse_iterator<typename Range::iterator> iterator;
> > > +
> > > + /* RANGE must be a forward range. basic_safe_reverse_iterators
> > > + will be used to traverse the forward range from the last element
> > > + to the first. */
> > > + explicit basic_safe_reverse_range (Range range)
> > > + : m_range (range)
> > > + {
> > > + }
> > > +
> > > + iterator begin ()
> > > + {
> > > + return iterator (m_range.begin (), m_range.end ());
> > > + }
> > > +
> > > + iterator end ()
> > > + {
> > > + return iterator (m_range.end (), m_range.end ());
> > > + }
> > > +
> > > +private:
> > > +
> > > + Range m_range;
> > > +};
> > > #endif /* COMMON_SAFE_ITERATOR_H */
> > > --
> > > 2.43.0
> > >
^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading
2024-05-03 22:58 [PATCH 0/3] On-demand debuginfo downloading Aaron Merey
2024-05-03 22:59 ` [PATCH 1/3 v3] gdb/progspace: Add reverse safe iterator Aaron Merey
@ 2024-05-03 22:59 ` Aaron Merey
2024-05-17 14:10 ` [PING][PATCH " Aaron Merey
2024-06-17 10:45 ` [PATCH " Andrew Burgess
2024-05-03 22:59 ` [PATCH 3/3 v6] gdb/debuginfod: Add .debug_line downloading Aaron Merey
2 siblings, 2 replies; 18+ messages in thread
From: Aaron Merey @ 2024-05-03 22:59 UTC (permalink / raw)
To: gdb-patches; +Cc: aburgess, Aaron Merey
v6 fixes a comment for dwarf2_has_separate_index and merges
lookup_gdb_index_debuginfod with lookup_gdb_index.
At the beginning of a session, gdb may attempt to download debuginfo
for all shared libraries associated with the process or core file
being debugged. This can be a waste of time and storage space when much
of the debuginfo ends up not being used during the session.
To reduce the gdb's startup latency and to download only the debuginfo
that is really needed, this patch adds on-demand downloading of debuginfo.
'set debuginfo enabled on' now causes gdb to attempt to download a .gdb_index
for each shared library instead of its full debuginfo. Each corresponding
separate debuginfo will be deferred until gdb needs to expand symtabs
associated with the debuginfo's index.
Because these indices are significantly smaller than their corresponding
debuginfo, this generally reduces the total amount of data gdb downloads.
Reductions of 80%-95% have been observed when debugging large GUI programs.
(gdb) set debuginfod enabled on
(gdb) start
Downloading section .gdb_index for /lib64/libcurl.so.4
[...]
1826 client->server_mhandle = curl_multi_init ();
(gdb) step
Downloading separate debug info for /lib64/libcurl.so.4
Downloading separate debug info for [libcurl dwz]
Downloading source file /usr/src/debug/curl-7.85.0-6.fc37.x86_64/build-full/lib/../../lib/multi.c
curl_multi_init () at ../../lib/multi.c:457
457 {
(gdb)
Some of the key functions below include dwarf2_has_separate_index which
downloads the separate .gdb_index. If successful, the shared library
objfile owns the index until the separate debug objfile is downloaded
or confirmed to not be available.
read_full_dwarf_from_debuginfod downloads the full debuginfo and
initializes the separate debug objfile. It is called by functions
such as dwarf2_gdb_index::expand_symtabs_matching and
dwarf2_base_index_functions::find_pc_sect_compunit_symtab when symtab
expansion is required.
---
gdb/dwarf2/frame.c | 13 ++
gdb/dwarf2/frame.h | 4 +
gdb/dwarf2/index-cache.c | 22 ++-
gdb/dwarf2/index-cache.h | 11 ++
gdb/dwarf2/public.h | 5 +
gdb/dwarf2/read-gdb-index.c | 123 +++++++++++++---
gdb/dwarf2/read.c | 149 ++++++++++++++++++-
gdb/dwarf2/read.h | 10 ++
gdb/dwarf2/section.c | 3 +-
gdb/elfread.c | 3 +-
gdb/frame.c | 27 +++-
gdb/objfile-flags.h | 4 +
gdb/objfiles.h | 20 +++
gdb/quick-symbol.h | 4 +
gdb/symfile-debug.c | 11 ++
gdb/symfile.c | 12 +-
gdb/symtab.c | 18 ++-
gdb/testsuite/gdb.debuginfod/libsection1.c | 43 ++++++
gdb/testsuite/gdb.debuginfod/libsection2.c | 37 +++++
gdb/testsuite/gdb.debuginfod/section.c | 29 ++++
gdb/testsuite/gdb.debuginfod/section.exp | 157 +++++++++++++++++++++
gdb/testsuite/lib/debuginfod-support.exp | 25 +++-
gdb/testsuite/lib/gdb.exp | 8 +-
23 files changed, 702 insertions(+), 36 deletions(-)
create mode 100644 gdb/testsuite/gdb.debuginfod/libsection1.c
create mode 100644 gdb/testsuite/gdb.debuginfod/libsection2.c
create mode 100644 gdb/testsuite/gdb.debuginfod/section.c
create mode 100644 gdb/testsuite/gdb.debuginfod/section.exp
diff --git a/gdb/dwarf2/frame.c b/gdb/dwarf2/frame.c
index 9ebf3ac4ac4..074417765f4 100644
--- a/gdb/dwarf2/frame.c
+++ b/gdb/dwarf2/frame.c
@@ -1621,6 +1621,19 @@ set_comp_unit (struct objfile *objfile, struct comp_unit *unit)
return dwarf2_frame_bfd_data.set (abfd, unit);
}
+/* See frame.h. */
+
+void
+dwarf2_clear_frame_data (struct objfile *objfile)
+{
+ bfd *abfd = objfile->obfd.get ();
+
+ if (gdb_bfd_requires_relocations (abfd))
+ dwarf2_frame_objfile_data.clear (objfile);
+ else
+ dwarf2_frame_bfd_data.clear (abfd);
+}
+
/* Find the FDE for *PC. Return a pointer to the FDE, and store the
initial location associated with it into *PC. */
diff --git a/gdb/dwarf2/frame.h b/gdb/dwarf2/frame.h
index 2167310fbdf..423a9355b8a 100644
--- a/gdb/dwarf2/frame.h
+++ b/gdb/dwarf2/frame.h
@@ -237,6 +237,10 @@ void dwarf2_append_unwinders (struct gdbarch *gdbarch);
extern const struct frame_base *
dwarf2_frame_base_sniffer (const frame_info_ptr &this_frame);
+/* Delete OBJFILEs comp_unit. */
+
+extern void dwarf2_clear_frame_data (struct objfile * objfile);
+
/* Compute the DWARF CFA for a frame. */
CORE_ADDR dwarf2_frame_cfa (const frame_info_ptr &this_frame);
diff --git a/gdb/dwarf2/index-cache.c b/gdb/dwarf2/index-cache.c
index 1720e925ee6..7096a3ea078 100644
--- a/gdb/dwarf2/index-cache.c
+++ b/gdb/dwarf2/index-cache.c
@@ -219,14 +219,22 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
/* Compute where we would expect a gdb index file for this build id to be. */
std::string filename = make_index_filename (build_id, INDEX4_SUFFIX);
+ return lookup_gdb_index (filename.c_str (), resource);
+}
+
+/* See index-cache.h. */
+
+gdb::array_view<const gdb_byte>
+index_cache::lookup_gdb_index (const char *filename,
+ std::unique_ptr<index_cache_resource> *resource)
+{
try
{
- index_cache_debug ("trying to read %s",
- filename.c_str ());
+ index_cache_debug ("trying to read %s", filename);
/* Try to map that file. */
index_cache_resource_mmap *mmap_resource
- = new index_cache_resource_mmap (filename.c_str ());
+ = new index_cache_resource_mmap (filename);
/* Yay, it worked! Hand the resource to the caller. */
resource->reset (mmap_resource);
@@ -238,7 +246,7 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
catch (const gdb_exception_error &except)
{
index_cache_debug ("couldn't read %s: %s",
- filename.c_str (), except.what ());
+ filename, except.what ());
}
return {};
@@ -255,6 +263,12 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
return {};
}
+gdb::array_view<const gdb_byte>
+index_cache::lookup_gdb_index_debuginfod (const char *index_path,
+ std::unique_ptr<index_cache_resource> *resource)
+{
+ return {};
+}
#endif
/* See dwarf-index-cache.h. */
diff --git a/gdb/dwarf2/index-cache.h b/gdb/dwarf2/index-cache.h
index 95f217ed961..f42e6918ce2 100644
--- a/gdb/dwarf2/index-cache.h
+++ b/gdb/dwarf2/index-cache.h
@@ -93,6 +93,17 @@ class index_cache
lookup_gdb_index (const bfd_build_id *build_id,
std::unique_ptr<index_cache_resource> *resource);
+ /* Look for an index file located at INDEX_PATH. If found, return the
+ contents as an array_view and store the underlying resources (allocated
+ memory, mapped file, etc) in RESOURCE. The returned array_view is valid
+ as long as RESOURCE is not destroyed.
+
+ If no matching index file is found, return an empty array view. This
+ function does not exit early if the index cache has not been enabled. */
+ gdb::array_view<const gdb_byte>
+ lookup_gdb_index (const char *filename,
+ std::unique_ptr<index_cache_resource> *resource);
+
/* Return the number of cache hits. */
unsigned int n_hits () const
{ return m_n_hits; }
diff --git a/gdb/dwarf2/public.h b/gdb/dwarf2/public.h
index bc419ff208f..6523854386c 100644
--- a/gdb/dwarf2/public.h
+++ b/gdb/dwarf2/public.h
@@ -44,4 +44,9 @@ extern bool dwarf2_initialize_objfile
extern void dwarf2_build_frame_info (struct objfile *);
+/* Query debuginfod for the .gdb_index associated with OBJFILE.
+ Used to defer separate debuginfo downloading until necessary. */
+
+extern bool dwarf2_has_separate_index (struct objfile *);
+
#endif /* DWARF2_PUBLIC_H */
diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
index e743ce13978..97b8137aabf 100644
--- a/gdb/dwarf2/read-gdb-index.c
+++ b/gdb/dwarf2/read-gdb-index.c
@@ -139,6 +139,8 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
gdb.dwarf2/gdb-index.exp testcase. */
void dump (struct objfile *objfile) override;
+ /* Calls do_expand_symtabs_matching and triggers debuginfo downloading
+ if necessary. */
bool expand_symtabs_matching
(struct objfile *objfile,
gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
@@ -147,8 +149,59 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
block_search_flags search_flags,
domain_search_flags domain) override;
+
+ /* Expand symtabs matching a given symbol or file. */
+ bool do_expand_symtabs_matching
+ (struct objfile *objfile,
+ gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
+ const lookup_name_info *lookup_name,
+ gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
+ gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
+ block_search_flags search_flags,
+ domain_search_flags domain);
+
+ /* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
+ debuginfo if necessary. */
+ void expand_all_symtabs (struct objfile *objfile) override;
+
+ /* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
+ debuginfo if necessary. */
+ struct symtab *find_last_source_symtab (struct objfile *objfile) override;
};
+void
+dwarf2_gdb_index::expand_all_symtabs (struct objfile *objfile)
+{
+ try
+ {
+ dwarf2_base_index_functions::expand_all_symtabs (objfile);
+ }
+ catch (const gdb_exception &e)
+ {
+ if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
+ exception_print (gdb_stderr, e);
+ else
+ read_full_dwarf_from_debuginfod (objfile, this);
+ }
+}
+
+struct symtab *
+dwarf2_gdb_index::find_last_source_symtab (struct objfile *objfile)
+{
+ try
+ {
+ return dwarf2_base_index_functions::find_last_source_symtab (objfile);
+ }
+ catch (const gdb_exception &e)
+ {
+ if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
+ exception_print (gdb_stderr, e);
+ else
+ read_full_dwarf_from_debuginfod (objfile, this);
+ return nullptr;
+ }
+}
+
/* This dumps minimal information about the index.
It is called via "mt print objfiles".
One use is to verify .gdb_index has been loaded by the
@@ -264,7 +317,7 @@ dw2_expand_marked_cus
}
bool
-dwarf2_gdb_index::expand_symtabs_matching
+dwarf2_gdb_index::do_expand_symtabs_matching
(struct objfile *objfile,
gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
const lookup_name_info *lookup_name,
@@ -312,6 +365,38 @@ dwarf2_gdb_index::expand_symtabs_matching
return result;
}
+bool
+dwarf2_gdb_index::expand_symtabs_matching
+ (struct objfile *objfile,
+ gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
+ const lookup_name_info *lookup_name,
+ gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
+ gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
+ block_search_flags search_flags,
+ domain_search_flags domain)
+{
+ if ((objfile->flags & OBJF_READNEVER) != 0)
+ return false;
+
+ try
+ {
+ return do_expand_symtabs_matching (objfile, file_matcher, lookup_name,
+ symbol_matcher, expansion_notify,
+ search_flags, domain);
+ }
+ catch (const gdb_exception &e)
+ {
+ if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
+ {
+ exception_print (gdb_stderr, e);
+ return false;
+ }
+
+ read_full_dwarf_from_debuginfod (objfile, this);
+ return true;
+ }
+}
+
quick_symbol_functions_up
mapped_gdb_index::make_quick_functions () const
{
@@ -647,28 +732,32 @@ dwarf2_read_gdb_index
/* If there is a .dwz file, read it so we can get its CU list as
well. */
- dwz = dwarf2_get_dwz_file (per_bfd);
- if (dwz != NULL)
+ if (get_gdb_index_contents_dwz != nullptr)
{
mapped_gdb_index dwz_map;
const gdb_byte *dwz_types_ignore;
offset_type dwz_types_elements_ignore;
+ dwz = dwarf2_get_dwz_file (per_bfd);
- gdb::array_view<const gdb_byte> dwz_index_content
- = get_gdb_index_contents_dwz (objfile, dwz);
-
- if (dwz_index_content.empty ())
- return 0;
-
- if (!read_gdb_index_from_buffer (bfd_get_filename (dwz->dwz_bfd.get ()),
- 1, dwz_index_content, &dwz_map,
- &dwz_list, &dwz_list_elements,
- &dwz_types_ignore,
- &dwz_types_elements_ignore))
+ if (dwz != nullptr)
{
- warning (_("could not read '.gdb_index' section from %s; skipping"),
- bfd_get_filename (dwz->dwz_bfd.get ()));
- return 0;
+ gdb::array_view<const gdb_byte> dwz_index_content
+ = get_gdb_index_contents_dwz (objfile, dwz);
+
+ if (dwz_index_content.empty ())
+ return 0;
+
+ if (!read_gdb_index_from_buffer (bfd_get_filename
+ (dwz->dwz_bfd.get ()),
+ 1, dwz_index_content, &dwz_map,
+ &dwz_list, &dwz_list_elements,
+ &dwz_types_ignore,
+ &dwz_types_elements_ignore))
+ {
+ warning (_("could not read '.gdb_index' section from %s; skipping"),
+ bfd_get_filename (dwz->dwz_bfd.get ()));
+ return 0;
+ }
}
}
diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
index 00e8608f9ce..4461c76347e 100644
--- a/gdb/dwarf2/read.c
+++ b/gdb/dwarf2/read.c
@@ -33,6 +33,7 @@
#include "dwarf2/attribute.h"
#include "dwarf2/comp-unit-head.h"
#include "dwarf2/cu.h"
+#include "dwarf2/frame.h"
#include "dwarf2/index-cache.h"
#include "dwarf2/index-common.h"
#include "dwarf2/leb.h"
@@ -96,6 +97,8 @@
#include "gdbsupport/thread-pool.h"
#include "run-on-main-thread.h"
#include "dwarf2/parent-map.h"
+#include "inferior.h"
+#include "debuginfod-support.h"
/* When == 1, print basic high level tracing messages.
When > 1, be more verbose.
@@ -3020,7 +3023,7 @@ dwarf2_base_index_functions::find_per_cu (dwarf2_per_bfd *per_bfd,
}
struct compunit_symtab *
-dwarf2_base_index_functions::find_pc_sect_compunit_symtab
+dwarf2_base_index_functions::do_find_pc_sect_compunit_symtab
(struct objfile *objfile,
struct bound_minimal_symbol msymbol,
CORE_ADDR pc,
@@ -3051,6 +3054,32 @@ dwarf2_base_index_functions::find_pc_sect_compunit_symtab
return result;
}
+struct compunit_symtab *
+dwarf2_base_index_functions::find_pc_sect_compunit_symtab
+ (struct objfile *objfile,
+ struct bound_minimal_symbol msymbol,
+ CORE_ADDR pc,
+ struct obj_section *section,
+ int warn_if_readin)
+{
+ if ((objfile->flags & OBJF_READNEVER) != 0)
+ return nullptr;
+
+ try
+ {
+ return do_find_pc_sect_compunit_symtab (objfile, msymbol, pc,
+ section, warn_if_readin);
+ }
+ catch (const gdb_exception &e)
+ {
+ if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
+ exception_print (gdb_stderr, e);
+ else
+ read_full_dwarf_from_debuginfod (objfile, this);
+ return nullptr;
+ }
+}
+
void
dwarf2_base_index_functions::map_symbol_filenames
(struct objfile *objfile,
@@ -3207,6 +3236,30 @@ get_gdb_index_contents_from_cache_dwz (objfile *obj, dwz_file *dwz)
return global_index_cache.lookup_gdb_index (build_id, &dwz->index_cache_res);
}
+/* Query debuginfod for the .gdb_index matching OBJFILE's build-id. Return the
+ contents if successful. */
+
+static gdb::array_view<const gdb_byte>
+get_gdb_index_contents_from_debuginfod (objfile *objfile, dwarf2_per_bfd *per_bfd)
+{
+ const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
+ if (build_id == nullptr)
+ return {};
+
+ gdb::unique_xmalloc_ptr<char> index_path;
+ scoped_fd fd = debuginfod_section_query (build_id->data, build_id->size,
+ bfd_get_filename
+ (objfile->obfd.get ()),
+ ".gdb_index",
+ &index_path);
+ if (fd.get () < 0)
+ return {};
+
+ return global_index_cache.lookup_gdb_index
+ (index_path.get (), &per_bfd->index_cache_res);
+}
+
+
static void start_debug_info_reader (dwarf2_per_objfile *);
/* See dwarf2/public.h. */
@@ -3216,11 +3269,13 @@ dwarf2_initialize_objfile (struct objfile *objfile,
const struct dwarf2_debug_sections *names,
bool can_copy)
{
- if (!dwarf2_has_info (objfile, names, can_copy))
+ if (!dwarf2_has_info (objfile, names, can_copy)
+ && (objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
return false;
dwarf2_per_objfile *per_objfile = get_dwarf2_per_objfile (objfile);
dwarf2_per_bfd *per_bfd = per_objfile->per_bfd;
+ bool separate_index = false;
dwarf_read_debug_printf ("called");
@@ -3256,6 +3311,15 @@ dwarf2_initialize_objfile (struct objfile *objfile,
dwarf_read_debug_printf ("found gdb index from cache");
global_index_cache.hit ();
}
+ /* Try to read just a separately downloaded gdb index. */
+ else if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0
+ && dwarf2_read_gdb_index (per_objfile,
+ get_gdb_index_contents_from_debuginfod,
+ nullptr))
+ {
+ dwarf_read_debug_printf ("found .gdb_index from debuginfod");
+ separate_index = true;
+ }
else
{
global_index_cache.miss ();
@@ -3267,8 +3331,89 @@ dwarf2_initialize_objfile (struct objfile *objfile,
if (dwarf_synchronous)
per_bfd->index_table->wait_completely ();
objfile->qf.push_front (per_bfd->index_table->make_quick_functions ());
+
+ if (separate_index)
+ objfile->qf.begin ()->get ()->from_separate_index = true;
+ }
+
+ return true;
+}
+
+/* See read.h. */
+
+void
+read_full_dwarf_from_debuginfod (struct objfile *objfile,
+ dwarf2_base_index_functions *fncs)
+{
+ gdb_assert ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
+
+ SCOPE_EXIT { objfile->remove_deferred_status (); };
+
+ const struct bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
+ if (build_id == nullptr)
+ return;
+
+ const char *filename = bfd_get_filename (objfile->obfd.get ());
+ gdb::unique_xmalloc_ptr<char> symfile_path;
+ scoped_fd fd;
+
+ fd = debuginfod_debuginfo_query (build_id->data, build_id->size,
+ filename, &symfile_path);
+ if (fd.get () < 0)
+ return;
+
+ /* Separate debuginfo successfully retrieved from server. */
+ gdb_bfd_ref_ptr debug_bfd = symfile_bfd_open (symfile_path.get ());
+ if (debug_bfd == nullptr
+ || !build_id_verify (debug_bfd.get (), build_id->size, build_id->data))
+ {
+ warning (_("File \"%s\" from debuginfod cannot be opened as bfd"),
+ filename);
+ return;
}
+ /* Clear frame data so it can be recalculated using DWARF. */
+ dwarf2_clear_frame_data (objfile);
+
+ /* This may also trigger a dwz download. */
+ symbol_file_add_separate (debug_bfd, symfile_path.get (),
+ current_inferior ()->symfile_flags, objfile);
+}
+
+/* See public.h. */
+
+bool
+dwarf2_has_separate_index (struct objfile *objfile)
+{
+ if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0)
+ return true;
+ if ((objfile->flags & OBJF_MAINLINE) != 0)
+ return false;
+ if (!IS_DIR_SEPARATOR (*objfile_filename (objfile)))
+ return false;
+
+ const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
+
+ if (build_id == nullptr)
+ return false;
+
+ gdb::unique_xmalloc_ptr<char> index_path;
+ scoped_fd fd = debuginfod_section_query (build_id->data,
+ build_id->size,
+ bfd_get_filename
+ (objfile->obfd.get ()),
+ ".gdb_index",
+ &index_path);
+
+ if (fd.get () < 0)
+ return false;
+
+ /* We found a separate .gdb_index file so a separate debuginfo file
+ should exist, but we don't want to download it until necessary.
+ Attach the index to this objfile and defer the debuginfo download
+ until gdb needs to expand symtabs referenced by the index. */
+ objfile->flags |= OBJF_DOWNLOAD_DEFERRED;
+ dwarf2_initialize_objfile (objfile);
return true;
}
diff --git a/gdb/dwarf2/read.h b/gdb/dwarf2/read.h
index 73def88c4c0..099d3d65c27 100644
--- a/gdb/dwarf2/read.h
+++ b/gdb/dwarf2/read.h
@@ -860,6 +860,10 @@ struct dwarf2_base_index_functions : public quick_symbol_functions
CORE_ADDR pc, struct obj_section *section, int warn_if_readin)
override;
+ struct compunit_symtab *do_find_pc_sect_compunit_symtab
+ (struct objfile *objfile, struct bound_minimal_symbol msymbol,
+ CORE_ADDR pc, struct obj_section *section, int warn_if_readin);
+
struct compunit_symtab *find_compunit_symtab_by_address
(struct objfile *objfile, CORE_ADDR address) override
{
@@ -932,4 +936,10 @@ extern void create_all_units (dwarf2_per_objfile *per_objfile);
extern htab_up create_quick_file_names_table (unsigned int nr_initial_entries);
+/* If OBJFILE contains information from a separately downloaded .gdb_index,
+ attempt to download the full debuginfo. */
+
+extern void read_full_dwarf_from_debuginfod (struct objfile *,
+ dwarf2_base_index_functions *);
+
#endif /* DWARF2READ_H */
diff --git a/gdb/dwarf2/section.c b/gdb/dwarf2/section.c
index 6aab0b9aee9..481f37351f4 100644
--- a/gdb/dwarf2/section.c
+++ b/gdb/dwarf2/section.c
@@ -53,7 +53,8 @@ dwarf2_section_info::get_bfd_owner () const
section = get_containing_section ();
gdb_assert (!section->is_virtual);
}
- gdb_assert (section->s.section != nullptr);
+ if (section->s.section == nullptr)
+ error (_("Can't find owner of DWARF section."));
return section->s.section->owner;
}
diff --git a/gdb/elfread.c b/gdb/elfread.c
index 7a6a8cadced..0f59bb1a58d 100644
--- a/gdb/elfread.c
+++ b/gdb/elfread.c
@@ -1217,7 +1217,8 @@ elf_symfile_read_dwarf2 (struct objfile *objfile,
&& objfile->separate_debug_objfile_backlink == NULL)
{
if (objfile->find_and_add_separate_symbol_file (symfile_flags))
- gdb_assert (objfile->separate_debug_objfile != nullptr);
+ gdb_assert (objfile->separate_debug_objfile != nullptr
+ || (objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
else
has_dwarf2 = false;
}
diff --git a/gdb/frame.c b/gdb/frame.c
index 43efac07d95..1718ab5086c 100644
--- a/gdb/frame.c
+++ b/gdb/frame.c
@@ -1778,14 +1778,10 @@ restore_selected_frame (frame_id frame_id, int frame_level)
selected_frame = nullptr;
}
-/* Lookup the frame_info object for the selected frame FRAME_ID /
- FRAME_LEVEL and cache the result.
-
- If FRAME_LEVEL > 0 and the originally selected frame isn't found,
- warn and select the innermost (current) frame. */
+/* See lookup_selected_frame. */
static void
-lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
+lookup_selected_frame_1 (struct frame_id &a_frame_id, int frame_level)
{
frame_info_ptr frame = NULL;
int count;
@@ -1857,6 +1853,25 @@ lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
}
}
+/* Lookup the frame_info object for the selected frame FRAME_ID /
+ FRAME_LEVEL and cache the result.
+
+ If FRAME_LEVEL > 0 and the originally selected frame isn't found,
+ warn and select the innermost (current) frame. */
+
+static void
+lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
+{
+ lookup_selected_frame_1 (selected_frame_id, selected_frame_level);
+
+ /* It is possible for lookup_selected_frame_1 to cause a new objfile
+ to be loaded. However some objfile observers may choose to clear
+ selected_frame when an objfile is loaded. Work around this by
+ calling lookup_selected_frame_1 again if the first call failed. */
+ if (selected_frame == nullptr)
+ lookup_selected_frame_1 (selected_frame_id, selected_frame_level);
+}
+
bool
has_stack_frames ()
{
diff --git a/gdb/objfile-flags.h b/gdb/objfile-flags.h
index 43d888e2a5a..08e4a75a9f2 100644
--- a/gdb/objfile-flags.h
+++ b/gdb/objfile-flags.h
@@ -56,6 +56,10 @@ enum objfile_flag : unsigned
/* User requested that we do not read this objfile's symbolic
information. */
OBJF_READNEVER = 1 << 6,
+
+ /* A separate .gdb_index has been downloaded for this objfile.
+ Debuginfo for this objfile can be downloaded when required. */
+ OBJF_DOWNLOAD_DEFERRED = 1 << 7,
};
DEF_ENUM_FLAGS_TYPE (enum objfile_flag, objfile_flags);
diff --git a/gdb/objfiles.h b/gdb/objfiles.h
index 1c8f79a2826..8c8f27f87f1 100644
--- a/gdb/objfiles.h
+++ b/gdb/objfiles.h
@@ -628,6 +628,26 @@ struct objfile
domain_search_flags domain,
bool *symbol_found_p);
+ /* Indicate that the acquisition of this objfile's separate debug objfile
+ is no longer deferred. Used when the debug objfile has been acquired
+ or could not be found. */
+ void remove_deferred_status ()
+ {
+ flags &= ~OBJF_DOWNLOAD_DEFERRED;
+
+ /* Remove quick_symbol_functions derived from a separately downloaded
+ index. If available the separate debug objfile's index will be used
+ instead, since that objfile actually contains the symbols and CUs
+ referenced in the index.
+
+ No more than one element of qf should have from_separate_index set
+ to true. */
+ qf.remove_if ([&] (const quick_symbol_functions_up &qf_up)
+ {
+ return qf_up->from_separate_index;
+ });
+ }
+
/* Return the relocation offset applied to SECTION. */
CORE_ADDR section_offset (bfd_section *section) const
{
diff --git a/gdb/quick-symbol.h b/gdb/quick-symbol.h
index 676c3ede373..8b91b7765d5 100644
--- a/gdb/quick-symbol.h
+++ b/gdb/quick-symbol.h
@@ -191,6 +191,10 @@ struct quick_symbol_functions
virtual void compute_main_name (struct objfile *objfile)
{
}
+
+ /* True if this quick_symbol_functions is derived from a separately
+ downloaded index. */
+ bool from_separate_index = false;
};
typedef std::unique_ptr<quick_symbol_functions> quick_symbol_functions_up;
diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c
index 330a520dca6..8099d464081 100644
--- a/gdb/symfile-debug.c
+++ b/gdb/symfile-debug.c
@@ -36,6 +36,7 @@
#include "cli/cli-style.h"
#include "build-id.h"
#include "debuginfod-support.h"
+#include "dwarf2/public.h"
/* We need to save a pointer to the real symbol functions.
Plus, the debug versions are malloc'd because we have to NULL out the
@@ -610,6 +611,16 @@ objfile::find_and_add_separate_symbol_file (symfile_add_flags symfile_flags)
= simple_find_and_open_separate_symbol_file
(this, find_separate_debug_file_by_debuglink, &warnings);
+ /* Attempt to download only an index from the separate debug info.
+ As with debuginfod_find_and_open_separate_symbol_file, only attempt
+ this once. */
+ if (debug_bfd == nullptr && attempt == 0
+ && dwarf2_has_separate_index (this))
+ {
+ has_dwarf2 = true;
+ break;
+ }
+
/* Only try debuginfod on the first attempt. Sure, we could imagine
an extension that somehow adds the required debug info to the
debuginfod server but, at least for now, we don't support this
diff --git a/gdb/symfile.c b/gdb/symfile.c
index 7ebb07d5394..35c4a3b78e9 100644
--- a/gdb/symfile.c
+++ b/gdb/symfile.c
@@ -989,6 +989,10 @@ syms_from_objfile (struct objfile *objfile,
static void
finish_new_objfile (struct objfile *objfile, symfile_add_flags add_flags)
{
+ struct objfile *parent = objfile->separate_debug_objfile_backlink;
+ bool was_deferred
+ = (parent != nullptr) && ((parent->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
+
/* If this is the main symbol file we have to clean up all users of the
old main symbol file. Otherwise it is sufficient to fixup all the
breakpoints that may have been redefined by this symbol file. */
@@ -999,7 +1003,7 @@ finish_new_objfile (struct objfile *objfile, symfile_add_flags add_flags)
clear_symtab_users (add_flags);
}
- else if ((add_flags & SYMFILE_DEFER_BP_RESET) == 0)
+ else if ((add_flags & SYMFILE_DEFER_BP_RESET) == 0 && !was_deferred)
{
breakpoint_re_set ();
}
@@ -1120,6 +1124,12 @@ symbol_file_add_with_addrs (const gdb_bfd_ref_ptr &abfd, const char *name,
if (objfile->sf != nullptr)
finish_new_objfile (objfile, add_flags);
+ /* Remove deferred status now in case any observers trigger symtab
+ expansion. Otherwise gdb might try to read parent for psymbols
+ when it should read the separate debug objfile instead. */
+ if (parent != nullptr && ((parent->flags & OBJF_DOWNLOAD_DEFERRED) != 0))
+ parent->remove_deferred_status ();
+
gdb::observers::new_objfile.notify (objfile);
return objfile;
diff --git a/gdb/symtab.c b/gdb/symtab.c
index 58648a8779d..26f1b1f631e 100644
--- a/gdb/symtab.c
+++ b/gdb/symtab.c
@@ -2956,14 +2956,30 @@ find_pc_sect_compunit_symtab (CORE_ADDR pc, struct obj_section *section)
if (best_cust != NULL)
return best_cust;
+ int warn_if_readin = 1;
+
/* Not found in symtabs, search the "quick" symtabs (e.g. psymtabs). */
for (objfile *objf : current_program_space->objfiles ())
{
+ bool was_deferred = (objf->flags & OBJF_DOWNLOAD_DEFERRED) != 0;
+
struct compunit_symtab *result
- = objf->find_pc_sect_compunit_symtab (msymbol, pc, section, 1);
+ = objf->find_pc_sect_compunit_symtab (msymbol, pc, section,
+ warn_if_readin);
+
if (result != NULL)
return result;
+
+ /* If OBJF's separate debug info was just acquired, disable
+ warn_if_readin for the next iteration of this loop. This prevents
+ a spurious warning in case an observer already triggered expansion
+ of the separate debug objfile's symtabs. */
+ if (was_deferred && objf->separate_debug_objfile != nullptr
+ && (objf->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
+ warn_if_readin = 0;
+ else if (warn_if_readin == 0)
+ warn_if_readin = 1;
}
return NULL;
diff --git a/gdb/testsuite/gdb.debuginfod/libsection1.c b/gdb/testsuite/gdb.debuginfod/libsection1.c
new file mode 100644
index 00000000000..e4f05db0a79
--- /dev/null
+++ b/gdb/testsuite/gdb.debuginfod/libsection1.c
@@ -0,0 +1,43 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+ Copyright 2023-2024 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>. */
+
+#include <stdio.h>
+#include <pthread.h>
+#include <unistd.h>
+
+extern void libsection2_test ();
+extern void *libsection2_thread_test (void *);
+
+static volatile int flag = 0;
+
+void
+libsection1_test ()
+{
+ pthread_t thr;
+
+ printf ("In libsection1\n");
+ libsection2_test ();
+
+ pthread_create (&thr, NULL, libsection2_thread_test, (void *) &flag);
+
+ /* Give the new thread a chance to actually enter libsection2_thread_test. */
+ while (!flag)
+ ;
+
+ printf ("Cancelling thread\n");
+ pthread_cancel (thr);
+}
diff --git a/gdb/testsuite/gdb.debuginfod/libsection2.c b/gdb/testsuite/gdb.debuginfod/libsection2.c
new file mode 100644
index 00000000000..3c82808e5fe
--- /dev/null
+++ b/gdb/testsuite/gdb.debuginfod/libsection2.c
@@ -0,0 +1,37 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+ Copyright 2023-2024 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>. */
+
+#include <stdio.h>
+#include <pthread.h>
+
+void
+libsection2_test ()
+{
+ printf ("In libsection2\n");
+}
+
+void *
+libsection2_thread_test (void *arg)
+{
+ int *flag = (int *) arg;
+ printf ("In thread test\n");
+
+ while (1)
+ *flag = 1;
+
+ return NULL;
+}
diff --git a/gdb/testsuite/gdb.debuginfod/section.c b/gdb/testsuite/gdb.debuginfod/section.c
new file mode 100644
index 00000000000..fa9ac536a6e
--- /dev/null
+++ b/gdb/testsuite/gdb.debuginfod/section.c
@@ -0,0 +1,29 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+ Copyright 2023-2024 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>. */
+
+#include <stdio.h>
+
+extern void libsection1_test ();
+
+int
+main ()
+{
+ libsection1_test ();
+ printf ("in section exec\n");
+
+ return 0;
+}
diff --git a/gdb/testsuite/gdb.debuginfod/section.exp b/gdb/testsuite/gdb.debuginfod/section.exp
new file mode 100644
index 00000000000..35086d15705
--- /dev/null
+++ b/gdb/testsuite/gdb.debuginfod/section.exp
@@ -0,0 +1,157 @@
+# Copyright 2023-2024 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+# Test debuginfod functionality
+
+standard_testfile
+
+load_lib debuginfod-support.exp
+
+clean_restart
+require allow_debuginfod_tests
+require allow_debuginfod_section_downloads
+
+# BINFILE calls a function from LIB_SL1.
+set libfile1 "libsection1"
+set libsrc1 $srcdir/$subdir/$libfile1.c
+set lib_sl1 [standard_output_file $libfile1.sl]
+
+# LIB_SL1 calls functions from LIB_SL2.
+set libfile2 "libsection2"
+set libsrc2 $srcdir/$subdir/$libfile2.c
+set lib_sl2 [standard_output_file $libfile2.sl]
+
+# Build LIB_SL2, LIB_SL1 and BINFILE.
+if { [build_executable "build $libfile2" $lib_sl2 $libsrc2 \
+ {debug build-id shlib}] != 0 } {
+ return -1
+}
+
+if { [build_executable "build $libfile1" $lib_sl1 $libsrc1 \
+ [list debug build-id shlib_pthreads shlib=$lib_sl2]] != 0 } {
+ return -1
+}
+
+if { [build_executable "build executable" $binfile $srcfile \
+ [list debug build-id shlib=$lib_sl1 shlib=$lib_sl2]] != 0 } {
+ return -1
+}
+
+# Make sure libsection1 and libsection2 contain .gdb_index.
+if { [ensure_gdb_index $lib_sl1 "" "libsection1"] != 1 } {
+ untested "failed to add .gdb_index to $libfile1"
+ return -1
+}
+
+if { [ensure_gdb_index $lib_sl2 "" "libsection2"] != 1 } {
+ untested "failed to add .gdb_index to $libfile2"
+ return -1
+}
+
+# Strip solib debuginfo into separate files.
+if { [gdb_gnu_strip_debug $lib_sl1 ""] != 0} {
+ fail "strip $lib_sl1 debuginfo"
+ return -1
+}
+
+if { [gdb_gnu_strip_debug $lib_sl2 ""] != 0} {
+ fail "strip $lib_sl2 debuginfo"
+ return -1
+}
+
+# Move debuginfo files into directory that debuginfod will serve from.
+set debugdir [standard_output_file "debug"]
+set debuginfo_sl1 [standard_output_file $libfile1.sl.debug]
+set debuginfo_sl2 [standard_output_file $libfile2.sl.debug]
+
+file mkdir $debugdir
+file rename -force $debuginfo_sl1 $debugdir
+file rename -force $debuginfo_sl2 $debugdir
+
+# Restart GDB and clear the debuginfod client cache. Then load BINFILE into
+# GDB and start running it. Match output with pattern RES and use TESTNAME
+# as the test name.
+proc_with_prefix clean_restart_with_prompt { binfile testname } {
+ global cache
+
+ # Delete client cache so debuginfo downloads again.
+ file delete -force $cache
+ clean_restart
+
+ gdb_test_no_output "set debuginfod enabled on" \
+ "clean_restart enable $testname"
+ gdb_load $binfile
+
+ runto_main
+}
+
+# Tests with no debuginfod server running.
+proc_with_prefix no_url { } {
+ global binfile libfile1 libfile2
+
+ gdb_load $binfile
+ if {![runto_main]} {
+ return
+ }
+
+ # Check that no section is downloaded and no debuginfo is found.
+ gdb_test "info sharedlibrary" ".*Yes \\(\\*\\).*$libfile1.*" \
+ "found no url lib1"
+ gdb_test "info sharedlibrary" ".*Yes \\(\\*\\).*$libfile2.*" \
+ "found no url lib2"
+}
+
+# Tests with a debuginfod server running.
+proc_with_prefix local_url { } {
+ global binfile
+ global libsrc1 lib_sl1 libfile1
+ global libsrc2 lib_sl2 libfile2
+ global debugdir db
+
+ set url [start_debuginfod $db $debugdir]
+ if { $url == "" } {
+ unresolved "failed to start debuginfod server"
+ return
+ }
+
+ # Point GDB to the server.
+ setenv DEBUGINFOD_URLS $url
+
+ clean_restart_with_prompt $binfile "index"
+
+ # Download debuginfo when stepping into a function.
+ set res ".*separate debug info for $lib_sl1.*\"In ${libfile1}\\\\n\".*"
+ gdb_test "step" $res "step"
+
+ clean_restart_with_prompt $binfile "break"
+
+ # Download debuginfo when setting a breakpoint.
+ set res "Download.*separate debug info for $lib_sl2.*"
+ gdb_test "br libsection2_test" $res "break set"
+
+ # Hit the breakpoint.
+ set res ".*Breakpoint 2, libsection2_test.*\"In ${libfile2}\\\\n\".*"
+ gdb_test "c" $res "break continue"
+}
+
+# Create CACHE and DB directories ready for debuginfod to use.
+prepare_for_debuginfod cache db
+
+with_debuginfod_env $cache {
+ no_url
+ local_url
+}
+
+stop_debuginfod
diff --git a/gdb/testsuite/lib/debuginfod-support.exp b/gdb/testsuite/lib/debuginfod-support.exp
index 0096448567e..4d8bdcdc7cb 100644
--- a/gdb/testsuite/lib/debuginfod-support.exp
+++ b/gdb/testsuite/lib/debuginfod-support.exp
@@ -113,6 +113,8 @@ proc with_debuginfod_env { cache body } {
proc start_debuginfod { db debugdir } {
global debuginfod_spawn_id spawn_id
+ set logfile [standard_output_file "server_log"]
+
# Find an unused port.
set port 7999
set found false
@@ -127,7 +129,8 @@ proc start_debuginfod { db debugdir } {
set old_spawn_id $spawn_id
}
- spawn debuginfod -vvvv -d $db -p $port -F $debugdir
+ spawn sh -c "debuginfod -vvvv -d $db -p $port -F $debugdir 2>&1 \
+ | tee $logfile"
set debuginfod_spawn_id $spawn_id
if { [info exists old_spawn_id] } {
@@ -194,3 +197,23 @@ proc stop_debuginfod { } {
unset debuginfod_spawn_id
}
}
+
+# Return true if gdb is configured to download ELF/DWARF sections from
+# debuginfod servers. Otherwise return false.
+proc allow_debuginfod_section_downloads { } {
+ set cmd "maint set debuginfod download-sections on"
+ set msg "enable section downloads"
+
+ gdb_test_multiple $cmd $msg {
+ -re -wrap ".*not compiled into GDB.*" {
+ return false
+ }
+ -re -wrap "^" {
+ return true
+ }
+ -re -wrap "" {
+ fail "$gdb_test_name (unexpected output)"
+ return false
+ }
+ }
+}
diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
index 5fb81f6e28d..a3680cb6371 100644
--- a/gdb/testsuite/lib/gdb.exp
+++ b/gdb/testsuite/lib/gdb.exp
@@ -9515,10 +9515,14 @@ proc get_index_type { objfile { testname "" } } {
# STYLE controls which style of index to add, if needed. The empty
# string (the default) means .gdb_index; "-dwarf-5" means .debug_names.
-proc ensure_gdb_index { binfile {style ""} } {
+proc ensure_gdb_index { binfile {style ""} {testname_prefix ""} } {
set testfile [file tail $binfile]
-
set test "check if index present"
+
+ if { $testname_prefix != "" } {
+ set test "$testname_prefix $test"
+ }
+
set index_type [get_index_type $testfile $test]
if { $index_type eq "gdb" || $index_type eq "dwarf5" } {
--
2.43.0
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING][PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading
2024-05-03 22:59 ` [PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading Aaron Merey
@ 2024-05-17 14:10 ` Aaron Merey
2024-05-31 20:58 ` [PING*2][PATCH " Aaron Merey
2024-06-17 10:45 ` [PATCH " Andrew Burgess
1 sibling, 1 reply; 18+ messages in thread
From: Aaron Merey @ 2024-05-17 14:10 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
>
> v6 fixes a comment for dwarf2_has_separate_index and merges
> lookup_gdb_index_debuginfod with lookup_gdb_index.
>
> At the beginning of a session, gdb may attempt to download debuginfo
> for all shared libraries associated with the process or core file
> being debugged. This can be a waste of time and storage space when much
> of the debuginfo ends up not being used during the session.
>
> To reduce the gdb's startup latency and to download only the debuginfo
> that is really needed, this patch adds on-demand downloading of debuginfo.
>
> 'set debuginfo enabled on' now causes gdb to attempt to download a .gdb_index
> for each shared library instead of its full debuginfo. Each corresponding
> separate debuginfo will be deferred until gdb needs to expand symtabs
> associated with the debuginfo's index.
>
> Because these indices are significantly smaller than their corresponding
> debuginfo, this generally reduces the total amount of data gdb downloads.
> Reductions of 80%-95% have been observed when debugging large GUI programs.
>
> (gdb) set debuginfod enabled on
> (gdb) start
> Downloading section .gdb_index for /lib64/libcurl.so.4
> [...]
> 1826 client->server_mhandle = curl_multi_init ();
> (gdb) step
> Downloading separate debug info for /lib64/libcurl.so.4
> Downloading separate debug info for [libcurl dwz]
> Downloading source file /usr/src/debug/curl-7.85.0-6.fc37.x86_64/build-full/lib/../../lib/multi.c
> curl_multi_init () at ../../lib/multi.c:457
> 457 {
> (gdb)
>
> Some of the key functions below include dwarf2_has_separate_index which
> downloads the separate .gdb_index. If successful, the shared library
> objfile owns the index until the separate debug objfile is downloaded
> or confirmed to not be available.
>
> read_full_dwarf_from_debuginfod downloads the full debuginfo and
> initializes the separate debug objfile. It is called by functions
> such as dwarf2_gdb_index::expand_symtabs_matching and
> dwarf2_base_index_functions::find_pc_sect_compunit_symtab when symtab
> expansion is required.
> ---
> gdb/dwarf2/frame.c | 13 ++
> gdb/dwarf2/frame.h | 4 +
> gdb/dwarf2/index-cache.c | 22 ++-
> gdb/dwarf2/index-cache.h | 11 ++
> gdb/dwarf2/public.h | 5 +
> gdb/dwarf2/read-gdb-index.c | 123 +++++++++++++---
> gdb/dwarf2/read.c | 149 ++++++++++++++++++-
> gdb/dwarf2/read.h | 10 ++
> gdb/dwarf2/section.c | 3 +-
> gdb/elfread.c | 3 +-
> gdb/frame.c | 27 +++-
> gdb/objfile-flags.h | 4 +
> gdb/objfiles.h | 20 +++
> gdb/quick-symbol.h | 4 +
> gdb/symfile-debug.c | 11 ++
> gdb/symfile.c | 12 +-
> gdb/symtab.c | 18 ++-
> gdb/testsuite/gdb.debuginfod/libsection1.c | 43 ++++++
> gdb/testsuite/gdb.debuginfod/libsection2.c | 37 +++++
> gdb/testsuite/gdb.debuginfod/section.c | 29 ++++
> gdb/testsuite/gdb.debuginfod/section.exp | 157 +++++++++++++++++++++
> gdb/testsuite/lib/debuginfod-support.exp | 25 +++-
> gdb/testsuite/lib/gdb.exp | 8 +-
> 23 files changed, 702 insertions(+), 36 deletions(-)
> create mode 100644 gdb/testsuite/gdb.debuginfod/libsection1.c
> create mode 100644 gdb/testsuite/gdb.debuginfod/libsection2.c
> create mode 100644 gdb/testsuite/gdb.debuginfod/section.c
> create mode 100644 gdb/testsuite/gdb.debuginfod/section.exp
>
> diff --git a/gdb/dwarf2/frame.c b/gdb/dwarf2/frame.c
> index 9ebf3ac4ac4..074417765f4 100644
> --- a/gdb/dwarf2/frame.c
> +++ b/gdb/dwarf2/frame.c
> @@ -1621,6 +1621,19 @@ set_comp_unit (struct objfile *objfile, struct comp_unit *unit)
> return dwarf2_frame_bfd_data.set (abfd, unit);
> }
>
> +/* See frame.h. */
> +
> +void
> +dwarf2_clear_frame_data (struct objfile *objfile)
> +{
> + bfd *abfd = objfile->obfd.get ();
> +
> + if (gdb_bfd_requires_relocations (abfd))
> + dwarf2_frame_objfile_data.clear (objfile);
> + else
> + dwarf2_frame_bfd_data.clear (abfd);
> +}
> +
> /* Find the FDE for *PC. Return a pointer to the FDE, and store the
> initial location associated with it into *PC. */
>
> diff --git a/gdb/dwarf2/frame.h b/gdb/dwarf2/frame.h
> index 2167310fbdf..423a9355b8a 100644
> --- a/gdb/dwarf2/frame.h
> +++ b/gdb/dwarf2/frame.h
> @@ -237,6 +237,10 @@ void dwarf2_append_unwinders (struct gdbarch *gdbarch);
> extern const struct frame_base *
> dwarf2_frame_base_sniffer (const frame_info_ptr &this_frame);
>
> +/* Delete OBJFILEs comp_unit. */
> +
> +extern void dwarf2_clear_frame_data (struct objfile * objfile);
> +
> /* Compute the DWARF CFA for a frame. */
>
> CORE_ADDR dwarf2_frame_cfa (const frame_info_ptr &this_frame);
> diff --git a/gdb/dwarf2/index-cache.c b/gdb/dwarf2/index-cache.c
> index 1720e925ee6..7096a3ea078 100644
> --- a/gdb/dwarf2/index-cache.c
> +++ b/gdb/dwarf2/index-cache.c
> @@ -219,14 +219,22 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> /* Compute where we would expect a gdb index file for this build id to be. */
> std::string filename = make_index_filename (build_id, INDEX4_SUFFIX);
>
> + return lookup_gdb_index (filename.c_str (), resource);
> +}
> +
> +/* See index-cache.h. */
> +
> +gdb::array_view<const gdb_byte>
> +index_cache::lookup_gdb_index (const char *filename,
> + std::unique_ptr<index_cache_resource> *resource)
> +{
> try
> {
> - index_cache_debug ("trying to read %s",
> - filename.c_str ());
> + index_cache_debug ("trying to read %s", filename);
>
> /* Try to map that file. */
> index_cache_resource_mmap *mmap_resource
> - = new index_cache_resource_mmap (filename.c_str ());
> + = new index_cache_resource_mmap (filename);
>
> /* Yay, it worked! Hand the resource to the caller. */
> resource->reset (mmap_resource);
> @@ -238,7 +246,7 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> catch (const gdb_exception_error &except)
> {
> index_cache_debug ("couldn't read %s: %s",
> - filename.c_str (), except.what ());
> + filename, except.what ());
> }
>
> return {};
> @@ -255,6 +263,12 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> return {};
> }
>
> +gdb::array_view<const gdb_byte>
> +index_cache::lookup_gdb_index_debuginfod (const char *index_path,
> + std::unique_ptr<index_cache_resource> *resource)
> +{
> + return {};
> +}
> #endif
>
> /* See dwarf-index-cache.h. */
> diff --git a/gdb/dwarf2/index-cache.h b/gdb/dwarf2/index-cache.h
> index 95f217ed961..f42e6918ce2 100644
> --- a/gdb/dwarf2/index-cache.h
> +++ b/gdb/dwarf2/index-cache.h
> @@ -93,6 +93,17 @@ class index_cache
> lookup_gdb_index (const bfd_build_id *build_id,
> std::unique_ptr<index_cache_resource> *resource);
>
> + /* Look for an index file located at INDEX_PATH. If found, return the
> + contents as an array_view and store the underlying resources (allocated
> + memory, mapped file, etc) in RESOURCE. The returned array_view is valid
> + as long as RESOURCE is not destroyed.
> +
> + If no matching index file is found, return an empty array view. This
> + function does not exit early if the index cache has not been enabled. */
> + gdb::array_view<const gdb_byte>
> + lookup_gdb_index (const char *filename,
> + std::unique_ptr<index_cache_resource> *resource);
> +
> /* Return the number of cache hits. */
> unsigned int n_hits () const
> { return m_n_hits; }
> diff --git a/gdb/dwarf2/public.h b/gdb/dwarf2/public.h
> index bc419ff208f..6523854386c 100644
> --- a/gdb/dwarf2/public.h
> +++ b/gdb/dwarf2/public.h
> @@ -44,4 +44,9 @@ extern bool dwarf2_initialize_objfile
>
> extern void dwarf2_build_frame_info (struct objfile *);
>
> +/* Query debuginfod for the .gdb_index associated with OBJFILE.
> + Used to defer separate debuginfo downloading until necessary. */
> +
> +extern bool dwarf2_has_separate_index (struct objfile *);
> +
> #endif /* DWARF2_PUBLIC_H */
> diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
> index e743ce13978..97b8137aabf 100644
> --- a/gdb/dwarf2/read-gdb-index.c
> +++ b/gdb/dwarf2/read-gdb-index.c
> @@ -139,6 +139,8 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> gdb.dwarf2/gdb-index.exp testcase. */
> void dump (struct objfile *objfile) override;
>
> + /* Calls do_expand_symtabs_matching and triggers debuginfo downloading
> + if necessary. */
> bool expand_symtabs_matching
> (struct objfile *objfile,
> gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> @@ -147,8 +149,59 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> block_search_flags search_flags,
> domain_search_flags domain) override;
> +
> + /* Expand symtabs matching a given symbol or file. */
> + bool do_expand_symtabs_matching
> + (struct objfile *objfile,
> + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> + const lookup_name_info *lookup_name,
> + gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
> + gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> + block_search_flags search_flags,
> + domain_search_flags domain);
> +
> + /* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
> + debuginfo if necessary. */
> + void expand_all_symtabs (struct objfile *objfile) override;
> +
> + /* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
> + debuginfo if necessary. */
> + struct symtab *find_last_source_symtab (struct objfile *objfile) override;
> };
>
> +void
> +dwarf2_gdb_index::expand_all_symtabs (struct objfile *objfile)
> +{
> + try
> + {
> + dwarf2_base_index_functions::expand_all_symtabs (objfile);
> + }
> + catch (const gdb_exception &e)
> + {
> + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> + exception_print (gdb_stderr, e);
> + else
> + read_full_dwarf_from_debuginfod (objfile, this);
> + }
> +}
> +
> +struct symtab *
> +dwarf2_gdb_index::find_last_source_symtab (struct objfile *objfile)
> +{
> + try
> + {
> + return dwarf2_base_index_functions::find_last_source_symtab (objfile);
> + }
> + catch (const gdb_exception &e)
> + {
> + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> + exception_print (gdb_stderr, e);
> + else
> + read_full_dwarf_from_debuginfod (objfile, this);
> + return nullptr;
> + }
> +}
> +
> /* This dumps minimal information about the index.
> It is called via "mt print objfiles".
> One use is to verify .gdb_index has been loaded by the
> @@ -264,7 +317,7 @@ dw2_expand_marked_cus
> }
>
> bool
> -dwarf2_gdb_index::expand_symtabs_matching
> +dwarf2_gdb_index::do_expand_symtabs_matching
> (struct objfile *objfile,
> gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> const lookup_name_info *lookup_name,
> @@ -312,6 +365,38 @@ dwarf2_gdb_index::expand_symtabs_matching
> return result;
> }
>
> +bool
> +dwarf2_gdb_index::expand_symtabs_matching
> + (struct objfile *objfile,
> + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> + const lookup_name_info *lookup_name,
> + gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
> + gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> + block_search_flags search_flags,
> + domain_search_flags domain)
> +{
> + if ((objfile->flags & OBJF_READNEVER) != 0)
> + return false;
> +
> + try
> + {
> + return do_expand_symtabs_matching (objfile, file_matcher, lookup_name,
> + symbol_matcher, expansion_notify,
> + search_flags, domain);
> + }
> + catch (const gdb_exception &e)
> + {
> + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> + {
> + exception_print (gdb_stderr, e);
> + return false;
> + }
> +
> + read_full_dwarf_from_debuginfod (objfile, this);
> + return true;
> + }
> +}
> +
> quick_symbol_functions_up
> mapped_gdb_index::make_quick_functions () const
> {
> @@ -647,28 +732,32 @@ dwarf2_read_gdb_index
>
> /* If there is a .dwz file, read it so we can get its CU list as
> well. */
> - dwz = dwarf2_get_dwz_file (per_bfd);
> - if (dwz != NULL)
> + if (get_gdb_index_contents_dwz != nullptr)
> {
> mapped_gdb_index dwz_map;
> const gdb_byte *dwz_types_ignore;
> offset_type dwz_types_elements_ignore;
> + dwz = dwarf2_get_dwz_file (per_bfd);
>
> - gdb::array_view<const gdb_byte> dwz_index_content
> - = get_gdb_index_contents_dwz (objfile, dwz);
> -
> - if (dwz_index_content.empty ())
> - return 0;
> -
> - if (!read_gdb_index_from_buffer (bfd_get_filename (dwz->dwz_bfd.get ()),
> - 1, dwz_index_content, &dwz_map,
> - &dwz_list, &dwz_list_elements,
> - &dwz_types_ignore,
> - &dwz_types_elements_ignore))
> + if (dwz != nullptr)
> {
> - warning (_("could not read '.gdb_index' section from %s; skipping"),
> - bfd_get_filename (dwz->dwz_bfd.get ()));
> - return 0;
> + gdb::array_view<const gdb_byte> dwz_index_content
> + = get_gdb_index_contents_dwz (objfile, dwz);
> +
> + if (dwz_index_content.empty ())
> + return 0;
> +
> + if (!read_gdb_index_from_buffer (bfd_get_filename
> + (dwz->dwz_bfd.get ()),
> + 1, dwz_index_content, &dwz_map,
> + &dwz_list, &dwz_list_elements,
> + &dwz_types_ignore,
> + &dwz_types_elements_ignore))
> + {
> + warning (_("could not read '.gdb_index' section from %s; skipping"),
> + bfd_get_filename (dwz->dwz_bfd.get ()));
> + return 0;
> + }
> }
> }
>
> diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
> index 00e8608f9ce..4461c76347e 100644
> --- a/gdb/dwarf2/read.c
> +++ b/gdb/dwarf2/read.c
> @@ -33,6 +33,7 @@
> #include "dwarf2/attribute.h"
> #include "dwarf2/comp-unit-head.h"
> #include "dwarf2/cu.h"
> +#include "dwarf2/frame.h"
> #include "dwarf2/index-cache.h"
> #include "dwarf2/index-common.h"
> #include "dwarf2/leb.h"
> @@ -96,6 +97,8 @@
> #include "gdbsupport/thread-pool.h"
> #include "run-on-main-thread.h"
> #include "dwarf2/parent-map.h"
> +#include "inferior.h"
> +#include "debuginfod-support.h"
>
> /* When == 1, print basic high level tracing messages.
> When > 1, be more verbose.
> @@ -3020,7 +3023,7 @@ dwarf2_base_index_functions::find_per_cu (dwarf2_per_bfd *per_bfd,
> }
>
> struct compunit_symtab *
> -dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> +dwarf2_base_index_functions::do_find_pc_sect_compunit_symtab
> (struct objfile *objfile,
> struct bound_minimal_symbol msymbol,
> CORE_ADDR pc,
> @@ -3051,6 +3054,32 @@ dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> return result;
> }
>
> +struct compunit_symtab *
> +dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> + (struct objfile *objfile,
> + struct bound_minimal_symbol msymbol,
> + CORE_ADDR pc,
> + struct obj_section *section,
> + int warn_if_readin)
> +{
> + if ((objfile->flags & OBJF_READNEVER) != 0)
> + return nullptr;
> +
> + try
> + {
> + return do_find_pc_sect_compunit_symtab (objfile, msymbol, pc,
> + section, warn_if_readin);
> + }
> + catch (const gdb_exception &e)
> + {
> + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> + exception_print (gdb_stderr, e);
> + else
> + read_full_dwarf_from_debuginfod (objfile, this);
> + return nullptr;
> + }
> +}
> +
> void
> dwarf2_base_index_functions::map_symbol_filenames
> (struct objfile *objfile,
> @@ -3207,6 +3236,30 @@ get_gdb_index_contents_from_cache_dwz (objfile *obj, dwz_file *dwz)
> return global_index_cache.lookup_gdb_index (build_id, &dwz->index_cache_res);
> }
>
> +/* Query debuginfod for the .gdb_index matching OBJFILE's build-id. Return the
> + contents if successful. */
> +
> +static gdb::array_view<const gdb_byte>
> +get_gdb_index_contents_from_debuginfod (objfile *objfile, dwarf2_per_bfd *per_bfd)
> +{
> + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> + if (build_id == nullptr)
> + return {};
> +
> + gdb::unique_xmalloc_ptr<char> index_path;
> + scoped_fd fd = debuginfod_section_query (build_id->data, build_id->size,
> + bfd_get_filename
> + (objfile->obfd.get ()),
> + ".gdb_index",
> + &index_path);
> + if (fd.get () < 0)
> + return {};
> +
> + return global_index_cache.lookup_gdb_index
> + (index_path.get (), &per_bfd->index_cache_res);
> +}
> +
> +
> static void start_debug_info_reader (dwarf2_per_objfile *);
>
> /* See dwarf2/public.h. */
> @@ -3216,11 +3269,13 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> const struct dwarf2_debug_sections *names,
> bool can_copy)
> {
> - if (!dwarf2_has_info (objfile, names, can_copy))
> + if (!dwarf2_has_info (objfile, names, can_copy)
> + && (objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> return false;
>
> dwarf2_per_objfile *per_objfile = get_dwarf2_per_objfile (objfile);
> dwarf2_per_bfd *per_bfd = per_objfile->per_bfd;
> + bool separate_index = false;
>
> dwarf_read_debug_printf ("called");
>
> @@ -3256,6 +3311,15 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> dwarf_read_debug_printf ("found gdb index from cache");
> global_index_cache.hit ();
> }
> + /* Try to read just a separately downloaded gdb index. */
> + else if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0
> + && dwarf2_read_gdb_index (per_objfile,
> + get_gdb_index_contents_from_debuginfod,
> + nullptr))
> + {
> + dwarf_read_debug_printf ("found .gdb_index from debuginfod");
> + separate_index = true;
> + }
> else
> {
> global_index_cache.miss ();
> @@ -3267,8 +3331,89 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> if (dwarf_synchronous)
> per_bfd->index_table->wait_completely ();
> objfile->qf.push_front (per_bfd->index_table->make_quick_functions ());
> +
> + if (separate_index)
> + objfile->qf.begin ()->get ()->from_separate_index = true;
> + }
> +
> + return true;
> +}
> +
> +/* See read.h. */
> +
> +void
> +read_full_dwarf_from_debuginfod (struct objfile *objfile,
> + dwarf2_base_index_functions *fncs)
> +{
> + gdb_assert ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> +
> + SCOPE_EXIT { objfile->remove_deferred_status (); };
> +
> + const struct bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> + if (build_id == nullptr)
> + return;
> +
> + const char *filename = bfd_get_filename (objfile->obfd.get ());
> + gdb::unique_xmalloc_ptr<char> symfile_path;
> + scoped_fd fd;
> +
> + fd = debuginfod_debuginfo_query (build_id->data, build_id->size,
> + filename, &symfile_path);
> + if (fd.get () < 0)
> + return;
> +
> + /* Separate debuginfo successfully retrieved from server. */
> + gdb_bfd_ref_ptr debug_bfd = symfile_bfd_open (symfile_path.get ());
> + if (debug_bfd == nullptr
> + || !build_id_verify (debug_bfd.get (), build_id->size, build_id->data))
> + {
> + warning (_("File \"%s\" from debuginfod cannot be opened as bfd"),
> + filename);
> + return;
> }
>
> + /* Clear frame data so it can be recalculated using DWARF. */
> + dwarf2_clear_frame_data (objfile);
> +
> + /* This may also trigger a dwz download. */
> + symbol_file_add_separate (debug_bfd, symfile_path.get (),
> + current_inferior ()->symfile_flags, objfile);
> +}
> +
> +/* See public.h. */
> +
> +bool
> +dwarf2_has_separate_index (struct objfile *objfile)
> +{
> + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0)
> + return true;
> + if ((objfile->flags & OBJF_MAINLINE) != 0)
> + return false;
> + if (!IS_DIR_SEPARATOR (*objfile_filename (objfile)))
> + return false;
> +
> + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> +
> + if (build_id == nullptr)
> + return false;
> +
> + gdb::unique_xmalloc_ptr<char> index_path;
> + scoped_fd fd = debuginfod_section_query (build_id->data,
> + build_id->size,
> + bfd_get_filename
> + (objfile->obfd.get ()),
> + ".gdb_index",
> + &index_path);
> +
> + if (fd.get () < 0)
> + return false;
> +
> + /* We found a separate .gdb_index file so a separate debuginfo file
> + should exist, but we don't want to download it until necessary.
> + Attach the index to this objfile and defer the debuginfo download
> + until gdb needs to expand symtabs referenced by the index. */
> + objfile->flags |= OBJF_DOWNLOAD_DEFERRED;
> + dwarf2_initialize_objfile (objfile);
> return true;
> }
>
> diff --git a/gdb/dwarf2/read.h b/gdb/dwarf2/read.h
> index 73def88c4c0..099d3d65c27 100644
> --- a/gdb/dwarf2/read.h
> +++ b/gdb/dwarf2/read.h
> @@ -860,6 +860,10 @@ struct dwarf2_base_index_functions : public quick_symbol_functions
> CORE_ADDR pc, struct obj_section *section, int warn_if_readin)
> override;
>
> + struct compunit_symtab *do_find_pc_sect_compunit_symtab
> + (struct objfile *objfile, struct bound_minimal_symbol msymbol,
> + CORE_ADDR pc, struct obj_section *section, int warn_if_readin);
> +
> struct compunit_symtab *find_compunit_symtab_by_address
> (struct objfile *objfile, CORE_ADDR address) override
> {
> @@ -932,4 +936,10 @@ extern void create_all_units (dwarf2_per_objfile *per_objfile);
>
> extern htab_up create_quick_file_names_table (unsigned int nr_initial_entries);
>
> +/* If OBJFILE contains information from a separately downloaded .gdb_index,
> + attempt to download the full debuginfo. */
> +
> +extern void read_full_dwarf_from_debuginfod (struct objfile *,
> + dwarf2_base_index_functions *);
> +
> #endif /* DWARF2READ_H */
> diff --git a/gdb/dwarf2/section.c b/gdb/dwarf2/section.c
> index 6aab0b9aee9..481f37351f4 100644
> --- a/gdb/dwarf2/section.c
> +++ b/gdb/dwarf2/section.c
> @@ -53,7 +53,8 @@ dwarf2_section_info::get_bfd_owner () const
> section = get_containing_section ();
> gdb_assert (!section->is_virtual);
> }
> - gdb_assert (section->s.section != nullptr);
> + if (section->s.section == nullptr)
> + error (_("Can't find owner of DWARF section."));
> return section->s.section->owner;
> }
>
> diff --git a/gdb/elfread.c b/gdb/elfread.c
> index 7a6a8cadced..0f59bb1a58d 100644
> --- a/gdb/elfread.c
> +++ b/gdb/elfread.c
> @@ -1217,7 +1217,8 @@ elf_symfile_read_dwarf2 (struct objfile *objfile,
> && objfile->separate_debug_objfile_backlink == NULL)
> {
> if (objfile->find_and_add_separate_symbol_file (symfile_flags))
> - gdb_assert (objfile->separate_debug_objfile != nullptr);
> + gdb_assert (objfile->separate_debug_objfile != nullptr
> + || (objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> else
> has_dwarf2 = false;
> }
> diff --git a/gdb/frame.c b/gdb/frame.c
> index 43efac07d95..1718ab5086c 100644
> --- a/gdb/frame.c
> +++ b/gdb/frame.c
> @@ -1778,14 +1778,10 @@ restore_selected_frame (frame_id frame_id, int frame_level)
> selected_frame = nullptr;
> }
>
> -/* Lookup the frame_info object for the selected frame FRAME_ID /
> - FRAME_LEVEL and cache the result.
> -
> - If FRAME_LEVEL > 0 and the originally selected frame isn't found,
> - warn and select the innermost (current) frame. */
> +/* See lookup_selected_frame. */
>
> static void
> -lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> +lookup_selected_frame_1 (struct frame_id &a_frame_id, int frame_level)
> {
> frame_info_ptr frame = NULL;
> int count;
> @@ -1857,6 +1853,25 @@ lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> }
> }
>
> +/* Lookup the frame_info object for the selected frame FRAME_ID /
> + FRAME_LEVEL and cache the result.
> +
> + If FRAME_LEVEL > 0 and the originally selected frame isn't found,
> + warn and select the innermost (current) frame. */
> +
> +static void
> +lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> +{
> + lookup_selected_frame_1 (selected_frame_id, selected_frame_level);
> +
> + /* It is possible for lookup_selected_frame_1 to cause a new objfile
> + to be loaded. However some objfile observers may choose to clear
> + selected_frame when an objfile is loaded. Work around this by
> + calling lookup_selected_frame_1 again if the first call failed. */
> + if (selected_frame == nullptr)
> + lookup_selected_frame_1 (selected_frame_id, selected_frame_level);
> +}
> +
> bool
> has_stack_frames ()
> {
> diff --git a/gdb/objfile-flags.h b/gdb/objfile-flags.h
> index 43d888e2a5a..08e4a75a9f2 100644
> --- a/gdb/objfile-flags.h
> +++ b/gdb/objfile-flags.h
> @@ -56,6 +56,10 @@ enum objfile_flag : unsigned
> /* User requested that we do not read this objfile's symbolic
> information. */
> OBJF_READNEVER = 1 << 6,
> +
> + /* A separate .gdb_index has been downloaded for this objfile.
> + Debuginfo for this objfile can be downloaded when required. */
> + OBJF_DOWNLOAD_DEFERRED = 1 << 7,
> };
>
> DEF_ENUM_FLAGS_TYPE (enum objfile_flag, objfile_flags);
> diff --git a/gdb/objfiles.h b/gdb/objfiles.h
> index 1c8f79a2826..8c8f27f87f1 100644
> --- a/gdb/objfiles.h
> +++ b/gdb/objfiles.h
> @@ -628,6 +628,26 @@ struct objfile
> domain_search_flags domain,
> bool *symbol_found_p);
>
> + /* Indicate that the acquisition of this objfile's separate debug objfile
> + is no longer deferred. Used when the debug objfile has been acquired
> + or could not be found. */
> + void remove_deferred_status ()
> + {
> + flags &= ~OBJF_DOWNLOAD_DEFERRED;
> +
> + /* Remove quick_symbol_functions derived from a separately downloaded
> + index. If available the separate debug objfile's index will be used
> + instead, since that objfile actually contains the symbols and CUs
> + referenced in the index.
> +
> + No more than one element of qf should have from_separate_index set
> + to true. */
> + qf.remove_if ([&] (const quick_symbol_functions_up &qf_up)
> + {
> + return qf_up->from_separate_index;
> + });
> + }
> +
> /* Return the relocation offset applied to SECTION. */
> CORE_ADDR section_offset (bfd_section *section) const
> {
> diff --git a/gdb/quick-symbol.h b/gdb/quick-symbol.h
> index 676c3ede373..8b91b7765d5 100644
> --- a/gdb/quick-symbol.h
> +++ b/gdb/quick-symbol.h
> @@ -191,6 +191,10 @@ struct quick_symbol_functions
> virtual void compute_main_name (struct objfile *objfile)
> {
> }
> +
> + /* True if this quick_symbol_functions is derived from a separately
> + downloaded index. */
> + bool from_separate_index = false;
> };
>
> typedef std::unique_ptr<quick_symbol_functions> quick_symbol_functions_up;
> diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c
> index 330a520dca6..8099d464081 100644
> --- a/gdb/symfile-debug.c
> +++ b/gdb/symfile-debug.c
> @@ -36,6 +36,7 @@
> #include "cli/cli-style.h"
> #include "build-id.h"
> #include "debuginfod-support.h"
> +#include "dwarf2/public.h"
>
> /* We need to save a pointer to the real symbol functions.
> Plus, the debug versions are malloc'd because we have to NULL out the
> @@ -610,6 +611,16 @@ objfile::find_and_add_separate_symbol_file (symfile_add_flags symfile_flags)
> = simple_find_and_open_separate_symbol_file
> (this, find_separate_debug_file_by_debuglink, &warnings);
>
> + /* Attempt to download only an index from the separate debug info.
> + As with debuginfod_find_and_open_separate_symbol_file, only attempt
> + this once. */
> + if (debug_bfd == nullptr && attempt == 0
> + && dwarf2_has_separate_index (this))
> + {
> + has_dwarf2 = true;
> + break;
> + }
> +
> /* Only try debuginfod on the first attempt. Sure, we could imagine
> an extension that somehow adds the required debug info to the
> debuginfod server but, at least for now, we don't support this
> diff --git a/gdb/symfile.c b/gdb/symfile.c
> index 7ebb07d5394..35c4a3b78e9 100644
> --- a/gdb/symfile.c
> +++ b/gdb/symfile.c
> @@ -989,6 +989,10 @@ syms_from_objfile (struct objfile *objfile,
> static void
> finish_new_objfile (struct objfile *objfile, symfile_add_flags add_flags)
> {
> + struct objfile *parent = objfile->separate_debug_objfile_backlink;
> + bool was_deferred
> + = (parent != nullptr) && ((parent->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> +
> /* If this is the main symbol file we have to clean up all users of the
> old main symbol file. Otherwise it is sufficient to fixup all the
> breakpoints that may have been redefined by this symbol file. */
> @@ -999,7 +1003,7 @@ finish_new_objfile (struct objfile *objfile, symfile_add_flags add_flags)
>
> clear_symtab_users (add_flags);
> }
> - else if ((add_flags & SYMFILE_DEFER_BP_RESET) == 0)
> + else if ((add_flags & SYMFILE_DEFER_BP_RESET) == 0 && !was_deferred)
> {
> breakpoint_re_set ();
> }
> @@ -1120,6 +1124,12 @@ symbol_file_add_with_addrs (const gdb_bfd_ref_ptr &abfd, const char *name,
> if (objfile->sf != nullptr)
> finish_new_objfile (objfile, add_flags);
>
> + /* Remove deferred status now in case any observers trigger symtab
> + expansion. Otherwise gdb might try to read parent for psymbols
> + when it should read the separate debug objfile instead. */
> + if (parent != nullptr && ((parent->flags & OBJF_DOWNLOAD_DEFERRED) != 0))
> + parent->remove_deferred_status ();
> +
> gdb::observers::new_objfile.notify (objfile);
>
> return objfile;
> diff --git a/gdb/symtab.c b/gdb/symtab.c
> index 58648a8779d..26f1b1f631e 100644
> --- a/gdb/symtab.c
> +++ b/gdb/symtab.c
> @@ -2956,14 +2956,30 @@ find_pc_sect_compunit_symtab (CORE_ADDR pc, struct obj_section *section)
> if (best_cust != NULL)
> return best_cust;
>
> + int warn_if_readin = 1;
> +
> /* Not found in symtabs, search the "quick" symtabs (e.g. psymtabs). */
>
> for (objfile *objf : current_program_space->objfiles ())
> {
> + bool was_deferred = (objf->flags & OBJF_DOWNLOAD_DEFERRED) != 0;
> +
> struct compunit_symtab *result
> - = objf->find_pc_sect_compunit_symtab (msymbol, pc, section, 1);
> + = objf->find_pc_sect_compunit_symtab (msymbol, pc, section,
> + warn_if_readin);
> +
> if (result != NULL)
> return result;
> +
> + /* If OBJF's separate debug info was just acquired, disable
> + warn_if_readin for the next iteration of this loop. This prevents
> + a spurious warning in case an observer already triggered expansion
> + of the separate debug objfile's symtabs. */
> + if (was_deferred && objf->separate_debug_objfile != nullptr
> + && (objf->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> + warn_if_readin = 0;
> + else if (warn_if_readin == 0)
> + warn_if_readin = 1;
> }
>
> return NULL;
> diff --git a/gdb/testsuite/gdb.debuginfod/libsection1.c b/gdb/testsuite/gdb.debuginfod/libsection1.c
> new file mode 100644
> index 00000000000..e4f05db0a79
> --- /dev/null
> +++ b/gdb/testsuite/gdb.debuginfod/libsection1.c
> @@ -0,0 +1,43 @@
> +/* This testcase is part of GDB, the GNU debugger.
> +
> + Copyright 2023-2024 Free Software Foundation, Inc.
> +
> + This program is free software; you can redistribute it and/or modify
> + it under the terms of the GNU General Public License as published by
> + the Free Software Foundation; either version 3 of the License, or
> + (at your option) any later version.
> +
> + This program is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + GNU General Public License for more details.
> +
> + You should have received a copy of the GNU General Public License
> + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> +
> +#include <stdio.h>
> +#include <pthread.h>
> +#include <unistd.h>
> +
> +extern void libsection2_test ();
> +extern void *libsection2_thread_test (void *);
> +
> +static volatile int flag = 0;
> +
> +void
> +libsection1_test ()
> +{
> + pthread_t thr;
> +
> + printf ("In libsection1\n");
> + libsection2_test ();
> +
> + pthread_create (&thr, NULL, libsection2_thread_test, (void *) &flag);
> +
> + /* Give the new thread a chance to actually enter libsection2_thread_test. */
> + while (!flag)
> + ;
> +
> + printf ("Cancelling thread\n");
> + pthread_cancel (thr);
> +}
> diff --git a/gdb/testsuite/gdb.debuginfod/libsection2.c b/gdb/testsuite/gdb.debuginfod/libsection2.c
> new file mode 100644
> index 00000000000..3c82808e5fe
> --- /dev/null
> +++ b/gdb/testsuite/gdb.debuginfod/libsection2.c
> @@ -0,0 +1,37 @@
> +/* This testcase is part of GDB, the GNU debugger.
> +
> + Copyright 2023-2024 Free Software Foundation, Inc.
> +
> + This program is free software; you can redistribute it and/or modify
> + it under the terms of the GNU General Public License as published by
> + the Free Software Foundation; either version 3 of the License, or
> + (at your option) any later version.
> +
> + This program is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + GNU General Public License for more details.
> +
> + You should have received a copy of the GNU General Public License
> + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> +
> +#include <stdio.h>
> +#include <pthread.h>
> +
> +void
> +libsection2_test ()
> +{
> + printf ("In libsection2\n");
> +}
> +
> +void *
> +libsection2_thread_test (void *arg)
> +{
> + int *flag = (int *) arg;
> + printf ("In thread test\n");
> +
> + while (1)
> + *flag = 1;
> +
> + return NULL;
> +}
> diff --git a/gdb/testsuite/gdb.debuginfod/section.c b/gdb/testsuite/gdb.debuginfod/section.c
> new file mode 100644
> index 00000000000..fa9ac536a6e
> --- /dev/null
> +++ b/gdb/testsuite/gdb.debuginfod/section.c
> @@ -0,0 +1,29 @@
> +/* This testcase is part of GDB, the GNU debugger.
> +
> + Copyright 2023-2024 Free Software Foundation, Inc.
> +
> + This program is free software; you can redistribute it and/or modify
> + it under the terms of the GNU General Public License as published by
> + the Free Software Foundation; either version 3 of the License, or
> + (at your option) any later version.
> +
> + This program is distributed in the hope that it will be useful,
> + but WITHOUT ANY WARRANTY; without even the implied warranty of
> + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + GNU General Public License for more details.
> +
> + You should have received a copy of the GNU General Public License
> + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> +
> +#include <stdio.h>
> +
> +extern void libsection1_test ();
> +
> +int
> +main ()
> +{
> + libsection1_test ();
> + printf ("in section exec\n");
> +
> + return 0;
> +}
> diff --git a/gdb/testsuite/gdb.debuginfod/section.exp b/gdb/testsuite/gdb.debuginfod/section.exp
> new file mode 100644
> index 00000000000..35086d15705
> --- /dev/null
> +++ b/gdb/testsuite/gdb.debuginfod/section.exp
> @@ -0,0 +1,157 @@
> +# Copyright 2023-2024 Free Software Foundation, Inc.
> +
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# This program is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program. If not, see <http://www.gnu.org/licenses/>.
> +
> +# Test debuginfod functionality
> +
> +standard_testfile
> +
> +load_lib debuginfod-support.exp
> +
> +clean_restart
> +require allow_debuginfod_tests
> +require allow_debuginfod_section_downloads
> +
> +# BINFILE calls a function from LIB_SL1.
> +set libfile1 "libsection1"
> +set libsrc1 $srcdir/$subdir/$libfile1.c
> +set lib_sl1 [standard_output_file $libfile1.sl]
> +
> +# LIB_SL1 calls functions from LIB_SL2.
> +set libfile2 "libsection2"
> +set libsrc2 $srcdir/$subdir/$libfile2.c
> +set lib_sl2 [standard_output_file $libfile2.sl]
> +
> +# Build LIB_SL2, LIB_SL1 and BINFILE.
> +if { [build_executable "build $libfile2" $lib_sl2 $libsrc2 \
> + {debug build-id shlib}] != 0 } {
> + return -1
> +}
> +
> +if { [build_executable "build $libfile1" $lib_sl1 $libsrc1 \
> + [list debug build-id shlib_pthreads shlib=$lib_sl2]] != 0 } {
> + return -1
> +}
> +
> +if { [build_executable "build executable" $binfile $srcfile \
> + [list debug build-id shlib=$lib_sl1 shlib=$lib_sl2]] != 0 } {
> + return -1
> +}
> +
> +# Make sure libsection1 and libsection2 contain .gdb_index.
> +if { [ensure_gdb_index $lib_sl1 "" "libsection1"] != 1 } {
> + untested "failed to add .gdb_index to $libfile1"
> + return -1
> +}
> +
> +if { [ensure_gdb_index $lib_sl2 "" "libsection2"] != 1 } {
> + untested "failed to add .gdb_index to $libfile2"
> + return -1
> +}
> +
> +# Strip solib debuginfo into separate files.
> +if { [gdb_gnu_strip_debug $lib_sl1 ""] != 0} {
> + fail "strip $lib_sl1 debuginfo"
> + return -1
> +}
> +
> +if { [gdb_gnu_strip_debug $lib_sl2 ""] != 0} {
> + fail "strip $lib_sl2 debuginfo"
> + return -1
> +}
> +
> +# Move debuginfo files into directory that debuginfod will serve from.
> +set debugdir [standard_output_file "debug"]
> +set debuginfo_sl1 [standard_output_file $libfile1.sl.debug]
> +set debuginfo_sl2 [standard_output_file $libfile2.sl.debug]
> +
> +file mkdir $debugdir
> +file rename -force $debuginfo_sl1 $debugdir
> +file rename -force $debuginfo_sl2 $debugdir
> +
> +# Restart GDB and clear the debuginfod client cache. Then load BINFILE into
> +# GDB and start running it. Match output with pattern RES and use TESTNAME
> +# as the test name.
> +proc_with_prefix clean_restart_with_prompt { binfile testname } {
> + global cache
> +
> + # Delete client cache so debuginfo downloads again.
> + file delete -force $cache
> + clean_restart
> +
> + gdb_test_no_output "set debuginfod enabled on" \
> + "clean_restart enable $testname"
> + gdb_load $binfile
> +
> + runto_main
> +}
> +
> +# Tests with no debuginfod server running.
> +proc_with_prefix no_url { } {
> + global binfile libfile1 libfile2
> +
> + gdb_load $binfile
> + if {![runto_main]} {
> + return
> + }
> +
> + # Check that no section is downloaded and no debuginfo is found.
> + gdb_test "info sharedlibrary" ".*Yes \\(\\*\\).*$libfile1.*" \
> + "found no url lib1"
> + gdb_test "info sharedlibrary" ".*Yes \\(\\*\\).*$libfile2.*" \
> + "found no url lib2"
> +}
> +
> +# Tests with a debuginfod server running.
> +proc_with_prefix local_url { } {
> + global binfile
> + global libsrc1 lib_sl1 libfile1
> + global libsrc2 lib_sl2 libfile2
> + global debugdir db
> +
> + set url [start_debuginfod $db $debugdir]
> + if { $url == "" } {
> + unresolved "failed to start debuginfod server"
> + return
> + }
> +
> + # Point GDB to the server.
> + setenv DEBUGINFOD_URLS $url
> +
> + clean_restart_with_prompt $binfile "index"
> +
> + # Download debuginfo when stepping into a function.
> + set res ".*separate debug info for $lib_sl1.*\"In ${libfile1}\\\\n\".*"
> + gdb_test "step" $res "step"
> +
> + clean_restart_with_prompt $binfile "break"
> +
> + # Download debuginfo when setting a breakpoint.
> + set res "Download.*separate debug info for $lib_sl2.*"
> + gdb_test "br libsection2_test" $res "break set"
> +
> + # Hit the breakpoint.
> + set res ".*Breakpoint 2, libsection2_test.*\"In ${libfile2}\\\\n\".*"
> + gdb_test "c" $res "break continue"
> +}
> +
> +# Create CACHE and DB directories ready for debuginfod to use.
> +prepare_for_debuginfod cache db
> +
> +with_debuginfod_env $cache {
> + no_url
> + local_url
> +}
> +
> +stop_debuginfod
> diff --git a/gdb/testsuite/lib/debuginfod-support.exp b/gdb/testsuite/lib/debuginfod-support.exp
> index 0096448567e..4d8bdcdc7cb 100644
> --- a/gdb/testsuite/lib/debuginfod-support.exp
> +++ b/gdb/testsuite/lib/debuginfod-support.exp
> @@ -113,6 +113,8 @@ proc with_debuginfod_env { cache body } {
> proc start_debuginfod { db debugdir } {
> global debuginfod_spawn_id spawn_id
>
> + set logfile [standard_output_file "server_log"]
> +
> # Find an unused port.
> set port 7999
> set found false
> @@ -127,7 +129,8 @@ proc start_debuginfod { db debugdir } {
> set old_spawn_id $spawn_id
> }
>
> - spawn debuginfod -vvvv -d $db -p $port -F $debugdir
> + spawn sh -c "debuginfod -vvvv -d $db -p $port -F $debugdir 2>&1 \
> + | tee $logfile"
> set debuginfod_spawn_id $spawn_id
>
> if { [info exists old_spawn_id] } {
> @@ -194,3 +197,23 @@ proc stop_debuginfod { } {
> unset debuginfod_spawn_id
> }
> }
> +
> +# Return true if gdb is configured to download ELF/DWARF sections from
> +# debuginfod servers. Otherwise return false.
> +proc allow_debuginfod_section_downloads { } {
> + set cmd "maint set debuginfod download-sections on"
> + set msg "enable section downloads"
> +
> + gdb_test_multiple $cmd $msg {
> + -re -wrap ".*not compiled into GDB.*" {
> + return false
> + }
> + -re -wrap "^" {
> + return true
> + }
> + -re -wrap "" {
> + fail "$gdb_test_name (unexpected output)"
> + return false
> + }
> + }
> +}
> diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
> index 5fb81f6e28d..a3680cb6371 100644
> --- a/gdb/testsuite/lib/gdb.exp
> +++ b/gdb/testsuite/lib/gdb.exp
> @@ -9515,10 +9515,14 @@ proc get_index_type { objfile { testname "" } } {
> # STYLE controls which style of index to add, if needed. The empty
> # string (the default) means .gdb_index; "-dwarf-5" means .debug_names.
>
> -proc ensure_gdb_index { binfile {style ""} } {
> +proc ensure_gdb_index { binfile {style ""} {testname_prefix ""} } {
> set testfile [file tail $binfile]
> -
> set test "check if index present"
> +
> + if { $testname_prefix != "" } {
> + set test "$testname_prefix $test"
> + }
> +
> set index_type [get_index_type $testfile $test]
>
> if { $index_type eq "gdb" || $index_type eq "dwarf5" } {
> --
> 2.43.0
>
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING*2][PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading
2024-05-17 14:10 ` [PING][PATCH " Aaron Merey
@ 2024-05-31 20:58 ` Aaron Merey
2024-06-07 20:48 ` [PING*3][PATCH " Aaron Merey
2024-06-14 18:24 ` Aaron Merey
0 siblings, 2 replies; 18+ messages in thread
From: Aaron Merey @ 2024-05-31 20:58 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 17, 2024 at 10:10 AM Aaron Merey <amerey@redhat.com> wrote:
>
> Ping
>
> Thanks,
> Aaron
>
> On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
> >
> > v6 fixes a comment for dwarf2_has_separate_index and merges
> > lookup_gdb_index_debuginfod with lookup_gdb_index.
> >
> > At the beginning of a session, gdb may attempt to download debuginfo
> > for all shared libraries associated with the process or core file
> > being debugged. This can be a waste of time and storage space when much
> > of the debuginfo ends up not being used during the session.
> >
> > To reduce the gdb's startup latency and to download only the debuginfo
> > that is really needed, this patch adds on-demand downloading of debuginfo.
> >
> > 'set debuginfo enabled on' now causes gdb to attempt to download a .gdb_index
> > for each shared library instead of its full debuginfo. Each corresponding
> > separate debuginfo will be deferred until gdb needs to expand symtabs
> > associated with the debuginfo's index.
> >
> > Because these indices are significantly smaller than their corresponding
> > debuginfo, this generally reduces the total amount of data gdb downloads.
> > Reductions of 80%-95% have been observed when debugging large GUI programs.
> >
> > (gdb) set debuginfod enabled on
> > (gdb) start
> > Downloading section .gdb_index for /lib64/libcurl.so.4
> > [...]
> > 1826 client->server_mhandle = curl_multi_init ();
> > (gdb) step
> > Downloading separate debug info for /lib64/libcurl.so.4
> > Downloading separate debug info for [libcurl dwz]
> > Downloading source file /usr/src/debug/curl-7.85.0-6.fc37.x86_64/build-full/lib/../../lib/multi.c
> > curl_multi_init () at ../../lib/multi.c:457
> > 457 {
> > (gdb)
> >
> > Some of the key functions below include dwarf2_has_separate_index which
> > downloads the separate .gdb_index. If successful, the shared library
> > objfile owns the index until the separate debug objfile is downloaded
> > or confirmed to not be available.
> >
> > read_full_dwarf_from_debuginfod downloads the full debuginfo and
> > initializes the separate debug objfile. It is called by functions
> > such as dwarf2_gdb_index::expand_symtabs_matching and
> > dwarf2_base_index_functions::find_pc_sect_compunit_symtab when symtab
> > expansion is required.
> > ---
> > gdb/dwarf2/frame.c | 13 ++
> > gdb/dwarf2/frame.h | 4 +
> > gdb/dwarf2/index-cache.c | 22 ++-
> > gdb/dwarf2/index-cache.h | 11 ++
> > gdb/dwarf2/public.h | 5 +
> > gdb/dwarf2/read-gdb-index.c | 123 +++++++++++++---
> > gdb/dwarf2/read.c | 149 ++++++++++++++++++-
> > gdb/dwarf2/read.h | 10 ++
> > gdb/dwarf2/section.c | 3 +-
> > gdb/elfread.c | 3 +-
> > gdb/frame.c | 27 +++-
> > gdb/objfile-flags.h | 4 +
> > gdb/objfiles.h | 20 +++
> > gdb/quick-symbol.h | 4 +
> > gdb/symfile-debug.c | 11 ++
> > gdb/symfile.c | 12 +-
> > gdb/symtab.c | 18 ++-
> > gdb/testsuite/gdb.debuginfod/libsection1.c | 43 ++++++
> > gdb/testsuite/gdb.debuginfod/libsection2.c | 37 +++++
> > gdb/testsuite/gdb.debuginfod/section.c | 29 ++++
> > gdb/testsuite/gdb.debuginfod/section.exp | 157 +++++++++++++++++++++
> > gdb/testsuite/lib/debuginfod-support.exp | 25 +++-
> > gdb/testsuite/lib/gdb.exp | 8 +-
> > 23 files changed, 702 insertions(+), 36 deletions(-)
> > create mode 100644 gdb/testsuite/gdb.debuginfod/libsection1.c
> > create mode 100644 gdb/testsuite/gdb.debuginfod/libsection2.c
> > create mode 100644 gdb/testsuite/gdb.debuginfod/section.c
> > create mode 100644 gdb/testsuite/gdb.debuginfod/section.exp
> >
> > diff --git a/gdb/dwarf2/frame.c b/gdb/dwarf2/frame.c
> > index 9ebf3ac4ac4..074417765f4 100644
> > --- a/gdb/dwarf2/frame.c
> > +++ b/gdb/dwarf2/frame.c
> > @@ -1621,6 +1621,19 @@ set_comp_unit (struct objfile *objfile, struct comp_unit *unit)
> > return dwarf2_frame_bfd_data.set (abfd, unit);
> > }
> >
> > +/* See frame.h. */
> > +
> > +void
> > +dwarf2_clear_frame_data (struct objfile *objfile)
> > +{
> > + bfd *abfd = objfile->obfd.get ();
> > +
> > + if (gdb_bfd_requires_relocations (abfd))
> > + dwarf2_frame_objfile_data.clear (objfile);
> > + else
> > + dwarf2_frame_bfd_data.clear (abfd);
> > +}
> > +
> > /* Find the FDE for *PC. Return a pointer to the FDE, and store the
> > initial location associated with it into *PC. */
> >
> > diff --git a/gdb/dwarf2/frame.h b/gdb/dwarf2/frame.h
> > index 2167310fbdf..423a9355b8a 100644
> > --- a/gdb/dwarf2/frame.h
> > +++ b/gdb/dwarf2/frame.h
> > @@ -237,6 +237,10 @@ void dwarf2_append_unwinders (struct gdbarch *gdbarch);
> > extern const struct frame_base *
> > dwarf2_frame_base_sniffer (const frame_info_ptr &this_frame);
> >
> > +/* Delete OBJFILEs comp_unit. */
> > +
> > +extern void dwarf2_clear_frame_data (struct objfile * objfile);
> > +
> > /* Compute the DWARF CFA for a frame. */
> >
> > CORE_ADDR dwarf2_frame_cfa (const frame_info_ptr &this_frame);
> > diff --git a/gdb/dwarf2/index-cache.c b/gdb/dwarf2/index-cache.c
> > index 1720e925ee6..7096a3ea078 100644
> > --- a/gdb/dwarf2/index-cache.c
> > +++ b/gdb/dwarf2/index-cache.c
> > @@ -219,14 +219,22 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> > /* Compute where we would expect a gdb index file for this build id to be. */
> > std::string filename = make_index_filename (build_id, INDEX4_SUFFIX);
> >
> > + return lookup_gdb_index (filename.c_str (), resource);
> > +}
> > +
> > +/* See index-cache.h. */
> > +
> > +gdb::array_view<const gdb_byte>
> > +index_cache::lookup_gdb_index (const char *filename,
> > + std::unique_ptr<index_cache_resource> *resource)
> > +{
> > try
> > {
> > - index_cache_debug ("trying to read %s",
> > - filename.c_str ());
> > + index_cache_debug ("trying to read %s", filename);
> >
> > /* Try to map that file. */
> > index_cache_resource_mmap *mmap_resource
> > - = new index_cache_resource_mmap (filename.c_str ());
> > + = new index_cache_resource_mmap (filename);
> >
> > /* Yay, it worked! Hand the resource to the caller. */
> > resource->reset (mmap_resource);
> > @@ -238,7 +246,7 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> > catch (const gdb_exception_error &except)
> > {
> > index_cache_debug ("couldn't read %s: %s",
> > - filename.c_str (), except.what ());
> > + filename, except.what ());
> > }
> >
> > return {};
> > @@ -255,6 +263,12 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> > return {};
> > }
> >
> > +gdb::array_view<const gdb_byte>
> > +index_cache::lookup_gdb_index_debuginfod (const char *index_path,
> > + std::unique_ptr<index_cache_resource> *resource)
> > +{
> > + return {};
> > +}
> > #endif
> >
> > /* See dwarf-index-cache.h. */
> > diff --git a/gdb/dwarf2/index-cache.h b/gdb/dwarf2/index-cache.h
> > index 95f217ed961..f42e6918ce2 100644
> > --- a/gdb/dwarf2/index-cache.h
> > +++ b/gdb/dwarf2/index-cache.h
> > @@ -93,6 +93,17 @@ class index_cache
> > lookup_gdb_index (const bfd_build_id *build_id,
> > std::unique_ptr<index_cache_resource> *resource);
> >
> > + /* Look for an index file located at INDEX_PATH. If found, return the
> > + contents as an array_view and store the underlying resources (allocated
> > + memory, mapped file, etc) in RESOURCE. The returned array_view is valid
> > + as long as RESOURCE is not destroyed.
> > +
> > + If no matching index file is found, return an empty array view. This
> > + function does not exit early if the index cache has not been enabled. */
> > + gdb::array_view<const gdb_byte>
> > + lookup_gdb_index (const char *filename,
> > + std::unique_ptr<index_cache_resource> *resource);
> > +
> > /* Return the number of cache hits. */
> > unsigned int n_hits () const
> > { return m_n_hits; }
> > diff --git a/gdb/dwarf2/public.h b/gdb/dwarf2/public.h
> > index bc419ff208f..6523854386c 100644
> > --- a/gdb/dwarf2/public.h
> > +++ b/gdb/dwarf2/public.h
> > @@ -44,4 +44,9 @@ extern bool dwarf2_initialize_objfile
> >
> > extern void dwarf2_build_frame_info (struct objfile *);
> >
> > +/* Query debuginfod for the .gdb_index associated with OBJFILE.
> > + Used to defer separate debuginfo downloading until necessary. */
> > +
> > +extern bool dwarf2_has_separate_index (struct objfile *);
> > +
> > #endif /* DWARF2_PUBLIC_H */
> > diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
> > index e743ce13978..97b8137aabf 100644
> > --- a/gdb/dwarf2/read-gdb-index.c
> > +++ b/gdb/dwarf2/read-gdb-index.c
> > @@ -139,6 +139,8 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > gdb.dwarf2/gdb-index.exp testcase. */
> > void dump (struct objfile *objfile) override;
> >
> > + /* Calls do_expand_symtabs_matching and triggers debuginfo downloading
> > + if necessary. */
> > bool expand_symtabs_matching
> > (struct objfile *objfile,
> > gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > @@ -147,8 +149,59 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> > block_search_flags search_flags,
> > domain_search_flags domain) override;
> > +
> > + /* Expand symtabs matching a given symbol or file. */
> > + bool do_expand_symtabs_matching
> > + (struct objfile *objfile,
> > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > + const lookup_name_info *lookup_name,
> > + gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
> > + gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> > + block_search_flags search_flags,
> > + domain_search_flags domain);
> > +
> > + /* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
> > + debuginfo if necessary. */
> > + void expand_all_symtabs (struct objfile *objfile) override;
> > +
> > + /* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
> > + debuginfo if necessary. */
> > + struct symtab *find_last_source_symtab (struct objfile *objfile) override;
> > };
> >
> > +void
> > +dwarf2_gdb_index::expand_all_symtabs (struct objfile *objfile)
> > +{
> > + try
> > + {
> > + dwarf2_base_index_functions::expand_all_symtabs (objfile);
> > + }
> > + catch (const gdb_exception &e)
> > + {
> > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > + exception_print (gdb_stderr, e);
> > + else
> > + read_full_dwarf_from_debuginfod (objfile, this);
> > + }
> > +}
> > +
> > +struct symtab *
> > +dwarf2_gdb_index::find_last_source_symtab (struct objfile *objfile)
> > +{
> > + try
> > + {
> > + return dwarf2_base_index_functions::find_last_source_symtab (objfile);
> > + }
> > + catch (const gdb_exception &e)
> > + {
> > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > + exception_print (gdb_stderr, e);
> > + else
> > + read_full_dwarf_from_debuginfod (objfile, this);
> > + return nullptr;
> > + }
> > +}
> > +
> > /* This dumps minimal information about the index.
> > It is called via "mt print objfiles".
> > One use is to verify .gdb_index has been loaded by the
> > @@ -264,7 +317,7 @@ dw2_expand_marked_cus
> > }
> >
> > bool
> > -dwarf2_gdb_index::expand_symtabs_matching
> > +dwarf2_gdb_index::do_expand_symtabs_matching
> > (struct objfile *objfile,
> > gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > const lookup_name_info *lookup_name,
> > @@ -312,6 +365,38 @@ dwarf2_gdb_index::expand_symtabs_matching
> > return result;
> > }
> >
> > +bool
> > +dwarf2_gdb_index::expand_symtabs_matching
> > + (struct objfile *objfile,
> > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > + const lookup_name_info *lookup_name,
> > + gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
> > + gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> > + block_search_flags search_flags,
> > + domain_search_flags domain)
> > +{
> > + if ((objfile->flags & OBJF_READNEVER) != 0)
> > + return false;
> > +
> > + try
> > + {
> > + return do_expand_symtabs_matching (objfile, file_matcher, lookup_name,
> > + symbol_matcher, expansion_notify,
> > + search_flags, domain);
> > + }
> > + catch (const gdb_exception &e)
> > + {
> > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > + {
> > + exception_print (gdb_stderr, e);
> > + return false;
> > + }
> > +
> > + read_full_dwarf_from_debuginfod (objfile, this);
> > + return true;
> > + }
> > +}
> > +
> > quick_symbol_functions_up
> > mapped_gdb_index::make_quick_functions () const
> > {
> > @@ -647,28 +732,32 @@ dwarf2_read_gdb_index
> >
> > /* If there is a .dwz file, read it so we can get its CU list as
> > well. */
> > - dwz = dwarf2_get_dwz_file (per_bfd);
> > - if (dwz != NULL)
> > + if (get_gdb_index_contents_dwz != nullptr)
> > {
> > mapped_gdb_index dwz_map;
> > const gdb_byte *dwz_types_ignore;
> > offset_type dwz_types_elements_ignore;
> > + dwz = dwarf2_get_dwz_file (per_bfd);
> >
> > - gdb::array_view<const gdb_byte> dwz_index_content
> > - = get_gdb_index_contents_dwz (objfile, dwz);
> > -
> > - if (dwz_index_content.empty ())
> > - return 0;
> > -
> > - if (!read_gdb_index_from_buffer (bfd_get_filename (dwz->dwz_bfd.get ()),
> > - 1, dwz_index_content, &dwz_map,
> > - &dwz_list, &dwz_list_elements,
> > - &dwz_types_ignore,
> > - &dwz_types_elements_ignore))
> > + if (dwz != nullptr)
> > {
> > - warning (_("could not read '.gdb_index' section from %s; skipping"),
> > - bfd_get_filename (dwz->dwz_bfd.get ()));
> > - return 0;
> > + gdb::array_view<const gdb_byte> dwz_index_content
> > + = get_gdb_index_contents_dwz (objfile, dwz);
> > +
> > + if (dwz_index_content.empty ())
> > + return 0;
> > +
> > + if (!read_gdb_index_from_buffer (bfd_get_filename
> > + (dwz->dwz_bfd.get ()),
> > + 1, dwz_index_content, &dwz_map,
> > + &dwz_list, &dwz_list_elements,
> > + &dwz_types_ignore,
> > + &dwz_types_elements_ignore))
> > + {
> > + warning (_("could not read '.gdb_index' section from %s; skipping"),
> > + bfd_get_filename (dwz->dwz_bfd.get ()));
> > + return 0;
> > + }
> > }
> > }
> >
> > diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
> > index 00e8608f9ce..4461c76347e 100644
> > --- a/gdb/dwarf2/read.c
> > +++ b/gdb/dwarf2/read.c
> > @@ -33,6 +33,7 @@
> > #include "dwarf2/attribute.h"
> > #include "dwarf2/comp-unit-head.h"
> > #include "dwarf2/cu.h"
> > +#include "dwarf2/frame.h"
> > #include "dwarf2/index-cache.h"
> > #include "dwarf2/index-common.h"
> > #include "dwarf2/leb.h"
> > @@ -96,6 +97,8 @@
> > #include "gdbsupport/thread-pool.h"
> > #include "run-on-main-thread.h"
> > #include "dwarf2/parent-map.h"
> > +#include "inferior.h"
> > +#include "debuginfod-support.h"
> >
> > /* When == 1, print basic high level tracing messages.
> > When > 1, be more verbose.
> > @@ -3020,7 +3023,7 @@ dwarf2_base_index_functions::find_per_cu (dwarf2_per_bfd *per_bfd,
> > }
> >
> > struct compunit_symtab *
> > -dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> > +dwarf2_base_index_functions::do_find_pc_sect_compunit_symtab
> > (struct objfile *objfile,
> > struct bound_minimal_symbol msymbol,
> > CORE_ADDR pc,
> > @@ -3051,6 +3054,32 @@ dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> > return result;
> > }
> >
> > +struct compunit_symtab *
> > +dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> > + (struct objfile *objfile,
> > + struct bound_minimal_symbol msymbol,
> > + CORE_ADDR pc,
> > + struct obj_section *section,
> > + int warn_if_readin)
> > +{
> > + if ((objfile->flags & OBJF_READNEVER) != 0)
> > + return nullptr;
> > +
> > + try
> > + {
> > + return do_find_pc_sect_compunit_symtab (objfile, msymbol, pc,
> > + section, warn_if_readin);
> > + }
> > + catch (const gdb_exception &e)
> > + {
> > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > + exception_print (gdb_stderr, e);
> > + else
> > + read_full_dwarf_from_debuginfod (objfile, this);
> > + return nullptr;
> > + }
> > +}
> > +
> > void
> > dwarf2_base_index_functions::map_symbol_filenames
> > (struct objfile *objfile,
> > @@ -3207,6 +3236,30 @@ get_gdb_index_contents_from_cache_dwz (objfile *obj, dwz_file *dwz)
> > return global_index_cache.lookup_gdb_index (build_id, &dwz->index_cache_res);
> > }
> >
> > +/* Query debuginfod for the .gdb_index matching OBJFILE's build-id. Return the
> > + contents if successful. */
> > +
> > +static gdb::array_view<const gdb_byte>
> > +get_gdb_index_contents_from_debuginfod (objfile *objfile, dwarf2_per_bfd *per_bfd)
> > +{
> > + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > + if (build_id == nullptr)
> > + return {};
> > +
> > + gdb::unique_xmalloc_ptr<char> index_path;
> > + scoped_fd fd = debuginfod_section_query (build_id->data, build_id->size,
> > + bfd_get_filename
> > + (objfile->obfd.get ()),
> > + ".gdb_index",
> > + &index_path);
> > + if (fd.get () < 0)
> > + return {};
> > +
> > + return global_index_cache.lookup_gdb_index
> > + (index_path.get (), &per_bfd->index_cache_res);
> > +}
> > +
> > +
> > static void start_debug_info_reader (dwarf2_per_objfile *);
> >
> > /* See dwarf2/public.h. */
> > @@ -3216,11 +3269,13 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> > const struct dwarf2_debug_sections *names,
> > bool can_copy)
> > {
> > - if (!dwarf2_has_info (objfile, names, can_copy))
> > + if (!dwarf2_has_info (objfile, names, can_copy)
> > + && (objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > return false;
> >
> > dwarf2_per_objfile *per_objfile = get_dwarf2_per_objfile (objfile);
> > dwarf2_per_bfd *per_bfd = per_objfile->per_bfd;
> > + bool separate_index = false;
> >
> > dwarf_read_debug_printf ("called");
> >
> > @@ -3256,6 +3311,15 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> > dwarf_read_debug_printf ("found gdb index from cache");
> > global_index_cache.hit ();
> > }
> > + /* Try to read just a separately downloaded gdb index. */
> > + else if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0
> > + && dwarf2_read_gdb_index (per_objfile,
> > + get_gdb_index_contents_from_debuginfod,
> > + nullptr))
> > + {
> > + dwarf_read_debug_printf ("found .gdb_index from debuginfod");
> > + separate_index = true;
> > + }
> > else
> > {
> > global_index_cache.miss ();
> > @@ -3267,8 +3331,89 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> > if (dwarf_synchronous)
> > per_bfd->index_table->wait_completely ();
> > objfile->qf.push_front (per_bfd->index_table->make_quick_functions ());
> > +
> > + if (separate_index)
> > + objfile->qf.begin ()->get ()->from_separate_index = true;
> > + }
> > +
> > + return true;
> > +}
> > +
> > +/* See read.h. */
> > +
> > +void
> > +read_full_dwarf_from_debuginfod (struct objfile *objfile,
> > + dwarf2_base_index_functions *fncs)
> > +{
> > + gdb_assert ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> > +
> > + SCOPE_EXIT { objfile->remove_deferred_status (); };
> > +
> > + const struct bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > + if (build_id == nullptr)
> > + return;
> > +
> > + const char *filename = bfd_get_filename (objfile->obfd.get ());
> > + gdb::unique_xmalloc_ptr<char> symfile_path;
> > + scoped_fd fd;
> > +
> > + fd = debuginfod_debuginfo_query (build_id->data, build_id->size,
> > + filename, &symfile_path);
> > + if (fd.get () < 0)
> > + return;
> > +
> > + /* Separate debuginfo successfully retrieved from server. */
> > + gdb_bfd_ref_ptr debug_bfd = symfile_bfd_open (symfile_path.get ());
> > + if (debug_bfd == nullptr
> > + || !build_id_verify (debug_bfd.get (), build_id->size, build_id->data))
> > + {
> > + warning (_("File \"%s\" from debuginfod cannot be opened as bfd"),
> > + filename);
> > + return;
> > }
> >
> > + /* Clear frame data so it can be recalculated using DWARF. */
> > + dwarf2_clear_frame_data (objfile);
> > +
> > + /* This may also trigger a dwz download. */
> > + symbol_file_add_separate (debug_bfd, symfile_path.get (),
> > + current_inferior ()->symfile_flags, objfile);
> > +}
> > +
> > +/* See public.h. */
> > +
> > +bool
> > +dwarf2_has_separate_index (struct objfile *objfile)
> > +{
> > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0)
> > + return true;
> > + if ((objfile->flags & OBJF_MAINLINE) != 0)
> > + return false;
> > + if (!IS_DIR_SEPARATOR (*objfile_filename (objfile)))
> > + return false;
> > +
> > + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > +
> > + if (build_id == nullptr)
> > + return false;
> > +
> > + gdb::unique_xmalloc_ptr<char> index_path;
> > + scoped_fd fd = debuginfod_section_query (build_id->data,
> > + build_id->size,
> > + bfd_get_filename
> > + (objfile->obfd.get ()),
> > + ".gdb_index",
> > + &index_path);
> > +
> > + if (fd.get () < 0)
> > + return false;
> > +
> > + /* We found a separate .gdb_index file so a separate debuginfo file
> > + should exist, but we don't want to download it until necessary.
> > + Attach the index to this objfile and defer the debuginfo download
> > + until gdb needs to expand symtabs referenced by the index. */
> > + objfile->flags |= OBJF_DOWNLOAD_DEFERRED;
> > + dwarf2_initialize_objfile (objfile);
> > return true;
> > }
> >
> > diff --git a/gdb/dwarf2/read.h b/gdb/dwarf2/read.h
> > index 73def88c4c0..099d3d65c27 100644
> > --- a/gdb/dwarf2/read.h
> > +++ b/gdb/dwarf2/read.h
> > @@ -860,6 +860,10 @@ struct dwarf2_base_index_functions : public quick_symbol_functions
> > CORE_ADDR pc, struct obj_section *section, int warn_if_readin)
> > override;
> >
> > + struct compunit_symtab *do_find_pc_sect_compunit_symtab
> > + (struct objfile *objfile, struct bound_minimal_symbol msymbol,
> > + CORE_ADDR pc, struct obj_section *section, int warn_if_readin);
> > +
> > struct compunit_symtab *find_compunit_symtab_by_address
> > (struct objfile *objfile, CORE_ADDR address) override
> > {
> > @@ -932,4 +936,10 @@ extern void create_all_units (dwarf2_per_objfile *per_objfile);
> >
> > extern htab_up create_quick_file_names_table (unsigned int nr_initial_entries);
> >
> > +/* If OBJFILE contains information from a separately downloaded .gdb_index,
> > + attempt to download the full debuginfo. */
> > +
> > +extern void read_full_dwarf_from_debuginfod (struct objfile *,
> > + dwarf2_base_index_functions *);
> > +
> > #endif /* DWARF2READ_H */
> > diff --git a/gdb/dwarf2/section.c b/gdb/dwarf2/section.c
> > index 6aab0b9aee9..481f37351f4 100644
> > --- a/gdb/dwarf2/section.c
> > +++ b/gdb/dwarf2/section.c
> > @@ -53,7 +53,8 @@ dwarf2_section_info::get_bfd_owner () const
> > section = get_containing_section ();
> > gdb_assert (!section->is_virtual);
> > }
> > - gdb_assert (section->s.section != nullptr);
> > + if (section->s.section == nullptr)
> > + error (_("Can't find owner of DWARF section."));
> > return section->s.section->owner;
> > }
> >
> > diff --git a/gdb/elfread.c b/gdb/elfread.c
> > index 7a6a8cadced..0f59bb1a58d 100644
> > --- a/gdb/elfread.c
> > +++ b/gdb/elfread.c
> > @@ -1217,7 +1217,8 @@ elf_symfile_read_dwarf2 (struct objfile *objfile,
> > && objfile->separate_debug_objfile_backlink == NULL)
> > {
> > if (objfile->find_and_add_separate_symbol_file (symfile_flags))
> > - gdb_assert (objfile->separate_debug_objfile != nullptr);
> > + gdb_assert (objfile->separate_debug_objfile != nullptr
> > + || (objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> > else
> > has_dwarf2 = false;
> > }
> > diff --git a/gdb/frame.c b/gdb/frame.c
> > index 43efac07d95..1718ab5086c 100644
> > --- a/gdb/frame.c
> > +++ b/gdb/frame.c
> > @@ -1778,14 +1778,10 @@ restore_selected_frame (frame_id frame_id, int frame_level)
> > selected_frame = nullptr;
> > }
> >
> > -/* Lookup the frame_info object for the selected frame FRAME_ID /
> > - FRAME_LEVEL and cache the result.
> > -
> > - If FRAME_LEVEL > 0 and the originally selected frame isn't found,
> > - warn and select the innermost (current) frame. */
> > +/* See lookup_selected_frame. */
> >
> > static void
> > -lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> > +lookup_selected_frame_1 (struct frame_id &a_frame_id, int frame_level)
> > {
> > frame_info_ptr frame = NULL;
> > int count;
> > @@ -1857,6 +1853,25 @@ lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> > }
> > }
> >
> > +/* Lookup the frame_info object for the selected frame FRAME_ID /
> > + FRAME_LEVEL and cache the result.
> > +
> > + If FRAME_LEVEL > 0 and the originally selected frame isn't found,
> > + warn and select the innermost (current) frame. */
> > +
> > +static void
> > +lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> > +{
> > + lookup_selected_frame_1 (selected_frame_id, selected_frame_level);
> > +
> > + /* It is possible for lookup_selected_frame_1 to cause a new objfile
> > + to be loaded. However some objfile observers may choose to clear
> > + selected_frame when an objfile is loaded. Work around this by
> > + calling lookup_selected_frame_1 again if the first call failed. */
> > + if (selected_frame == nullptr)
> > + lookup_selected_frame_1 (selected_frame_id, selected_frame_level);
> > +}
> > +
> > bool
> > has_stack_frames ()
> > {
> > diff --git a/gdb/objfile-flags.h b/gdb/objfile-flags.h
> > index 43d888e2a5a..08e4a75a9f2 100644
> > --- a/gdb/objfile-flags.h
> > +++ b/gdb/objfile-flags.h
> > @@ -56,6 +56,10 @@ enum objfile_flag : unsigned
> > /* User requested that we do not read this objfile's symbolic
> > information. */
> > OBJF_READNEVER = 1 << 6,
> > +
> > + /* A separate .gdb_index has been downloaded for this objfile.
> > + Debuginfo for this objfile can be downloaded when required. */
> > + OBJF_DOWNLOAD_DEFERRED = 1 << 7,
> > };
> >
> > DEF_ENUM_FLAGS_TYPE (enum objfile_flag, objfile_flags);
> > diff --git a/gdb/objfiles.h b/gdb/objfiles.h
> > index 1c8f79a2826..8c8f27f87f1 100644
> > --- a/gdb/objfiles.h
> > +++ b/gdb/objfiles.h
> > @@ -628,6 +628,26 @@ struct objfile
> > domain_search_flags domain,
> > bool *symbol_found_p);
> >
> > + /* Indicate that the acquisition of this objfile's separate debug objfile
> > + is no longer deferred. Used when the debug objfile has been acquired
> > + or could not be found. */
> > + void remove_deferred_status ()
> > + {
> > + flags &= ~OBJF_DOWNLOAD_DEFERRED;
> > +
> > + /* Remove quick_symbol_functions derived from a separately downloaded
> > + index. If available the separate debug objfile's index will be used
> > + instead, since that objfile actually contains the symbols and CUs
> > + referenced in the index.
> > +
> > + No more than one element of qf should have from_separate_index set
> > + to true. */
> > + qf.remove_if ([&] (const quick_symbol_functions_up &qf_up)
> > + {
> > + return qf_up->from_separate_index;
> > + });
> > + }
> > +
> > /* Return the relocation offset applied to SECTION. */
> > CORE_ADDR section_offset (bfd_section *section) const
> > {
> > diff --git a/gdb/quick-symbol.h b/gdb/quick-symbol.h
> > index 676c3ede373..8b91b7765d5 100644
> > --- a/gdb/quick-symbol.h
> > +++ b/gdb/quick-symbol.h
> > @@ -191,6 +191,10 @@ struct quick_symbol_functions
> > virtual void compute_main_name (struct objfile *objfile)
> > {
> > }
> > +
> > + /* True if this quick_symbol_functions is derived from a separately
> > + downloaded index. */
> > + bool from_separate_index = false;
> > };
> >
> > typedef std::unique_ptr<quick_symbol_functions> quick_symbol_functions_up;
> > diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c
> > index 330a520dca6..8099d464081 100644
> > --- a/gdb/symfile-debug.c
> > +++ b/gdb/symfile-debug.c
> > @@ -36,6 +36,7 @@
> > #include "cli/cli-style.h"
> > #include "build-id.h"
> > #include "debuginfod-support.h"
> > +#include "dwarf2/public.h"
> >
> > /* We need to save a pointer to the real symbol functions.
> > Plus, the debug versions are malloc'd because we have to NULL out the
> > @@ -610,6 +611,16 @@ objfile::find_and_add_separate_symbol_file (symfile_add_flags symfile_flags)
> > = simple_find_and_open_separate_symbol_file
> > (this, find_separate_debug_file_by_debuglink, &warnings);
> >
> > + /* Attempt to download only an index from the separate debug info.
> > + As with debuginfod_find_and_open_separate_symbol_file, only attempt
> > + this once. */
> > + if (debug_bfd == nullptr && attempt == 0
> > + && dwarf2_has_separate_index (this))
> > + {
> > + has_dwarf2 = true;
> > + break;
> > + }
> > +
> > /* Only try debuginfod on the first attempt. Sure, we could imagine
> > an extension that somehow adds the required debug info to the
> > debuginfod server but, at least for now, we don't support this
> > diff --git a/gdb/symfile.c b/gdb/symfile.c
> > index 7ebb07d5394..35c4a3b78e9 100644
> > --- a/gdb/symfile.c
> > +++ b/gdb/symfile.c
> > @@ -989,6 +989,10 @@ syms_from_objfile (struct objfile *objfile,
> > static void
> > finish_new_objfile (struct objfile *objfile, symfile_add_flags add_flags)
> > {
> > + struct objfile *parent = objfile->separate_debug_objfile_backlink;
> > + bool was_deferred
> > + = (parent != nullptr) && ((parent->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> > +
> > /* If this is the main symbol file we have to clean up all users of the
> > old main symbol file. Otherwise it is sufficient to fixup all the
> > breakpoints that may have been redefined by this symbol file. */
> > @@ -999,7 +1003,7 @@ finish_new_objfile (struct objfile *objfile, symfile_add_flags add_flags)
> >
> > clear_symtab_users (add_flags);
> > }
> > - else if ((add_flags & SYMFILE_DEFER_BP_RESET) == 0)
> > + else if ((add_flags & SYMFILE_DEFER_BP_RESET) == 0 && !was_deferred)
> > {
> > breakpoint_re_set ();
> > }
> > @@ -1120,6 +1124,12 @@ symbol_file_add_with_addrs (const gdb_bfd_ref_ptr &abfd, const char *name,
> > if (objfile->sf != nullptr)
> > finish_new_objfile (objfile, add_flags);
> >
> > + /* Remove deferred status now in case any observers trigger symtab
> > + expansion. Otherwise gdb might try to read parent for psymbols
> > + when it should read the separate debug objfile instead. */
> > + if (parent != nullptr && ((parent->flags & OBJF_DOWNLOAD_DEFERRED) != 0))
> > + parent->remove_deferred_status ();
> > +
> > gdb::observers::new_objfile.notify (objfile);
> >
> > return objfile;
> > diff --git a/gdb/symtab.c b/gdb/symtab.c
> > index 58648a8779d..26f1b1f631e 100644
> > --- a/gdb/symtab.c
> > +++ b/gdb/symtab.c
> > @@ -2956,14 +2956,30 @@ find_pc_sect_compunit_symtab (CORE_ADDR pc, struct obj_section *section)
> > if (best_cust != NULL)
> > return best_cust;
> >
> > + int warn_if_readin = 1;
> > +
> > /* Not found in symtabs, search the "quick" symtabs (e.g. psymtabs). */
> >
> > for (objfile *objf : current_program_space->objfiles ())
> > {
> > + bool was_deferred = (objf->flags & OBJF_DOWNLOAD_DEFERRED) != 0;
> > +
> > struct compunit_symtab *result
> > - = objf->find_pc_sect_compunit_symtab (msymbol, pc, section, 1);
> > + = objf->find_pc_sect_compunit_symtab (msymbol, pc, section,
> > + warn_if_readin);
> > +
> > if (result != NULL)
> > return result;
> > +
> > + /* If OBJF's separate debug info was just acquired, disable
> > + warn_if_readin for the next iteration of this loop. This prevents
> > + a spurious warning in case an observer already triggered expansion
> > + of the separate debug objfile's symtabs. */
> > + if (was_deferred && objf->separate_debug_objfile != nullptr
> > + && (objf->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > + warn_if_readin = 0;
> > + else if (warn_if_readin == 0)
> > + warn_if_readin = 1;
> > }
> >
> > return NULL;
> > diff --git a/gdb/testsuite/gdb.debuginfod/libsection1.c b/gdb/testsuite/gdb.debuginfod/libsection1.c
> > new file mode 100644
> > index 00000000000..e4f05db0a79
> > --- /dev/null
> > +++ b/gdb/testsuite/gdb.debuginfod/libsection1.c
> > @@ -0,0 +1,43 @@
> > +/* This testcase is part of GDB, the GNU debugger.
> > +
> > + Copyright 2023-2024 Free Software Foundation, Inc.
> > +
> > + This program is free software; you can redistribute it and/or modify
> > + it under the terms of the GNU General Public License as published by
> > + the Free Software Foundation; either version 3 of the License, or
> > + (at your option) any later version.
> > +
> > + This program is distributed in the hope that it will be useful,
> > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > + GNU General Public License for more details.
> > +
> > + You should have received a copy of the GNU General Public License
> > + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> > +
> > +#include <stdio.h>
> > +#include <pthread.h>
> > +#include <unistd.h>
> > +
> > +extern void libsection2_test ();
> > +extern void *libsection2_thread_test (void *);
> > +
> > +static volatile int flag = 0;
> > +
> > +void
> > +libsection1_test ()
> > +{
> > + pthread_t thr;
> > +
> > + printf ("In libsection1\n");
> > + libsection2_test ();
> > +
> > + pthread_create (&thr, NULL, libsection2_thread_test, (void *) &flag);
> > +
> > + /* Give the new thread a chance to actually enter libsection2_thread_test. */
> > + while (!flag)
> > + ;
> > +
> > + printf ("Cancelling thread\n");
> > + pthread_cancel (thr);
> > +}
> > diff --git a/gdb/testsuite/gdb.debuginfod/libsection2.c b/gdb/testsuite/gdb.debuginfod/libsection2.c
> > new file mode 100644
> > index 00000000000..3c82808e5fe
> > --- /dev/null
> > +++ b/gdb/testsuite/gdb.debuginfod/libsection2.c
> > @@ -0,0 +1,37 @@
> > +/* This testcase is part of GDB, the GNU debugger.
> > +
> > + Copyright 2023-2024 Free Software Foundation, Inc.
> > +
> > + This program is free software; you can redistribute it and/or modify
> > + it under the terms of the GNU General Public License as published by
> > + the Free Software Foundation; either version 3 of the License, or
> > + (at your option) any later version.
> > +
> > + This program is distributed in the hope that it will be useful,
> > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > + GNU General Public License for more details.
> > +
> > + You should have received a copy of the GNU General Public License
> > + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> > +
> > +#include <stdio.h>
> > +#include <pthread.h>
> > +
> > +void
> > +libsection2_test ()
> > +{
> > + printf ("In libsection2\n");
> > +}
> > +
> > +void *
> > +libsection2_thread_test (void *arg)
> > +{
> > + int *flag = (int *) arg;
> > + printf ("In thread test\n");
> > +
> > + while (1)
> > + *flag = 1;
> > +
> > + return NULL;
> > +}
> > diff --git a/gdb/testsuite/gdb.debuginfod/section.c b/gdb/testsuite/gdb.debuginfod/section.c
> > new file mode 100644
> > index 00000000000..fa9ac536a6e
> > --- /dev/null
> > +++ b/gdb/testsuite/gdb.debuginfod/section.c
> > @@ -0,0 +1,29 @@
> > +/* This testcase is part of GDB, the GNU debugger.
> > +
> > + Copyright 2023-2024 Free Software Foundation, Inc.
> > +
> > + This program is free software; you can redistribute it and/or modify
> > + it under the terms of the GNU General Public License as published by
> > + the Free Software Foundation; either version 3 of the License, or
> > + (at your option) any later version.
> > +
> > + This program is distributed in the hope that it will be useful,
> > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > + GNU General Public License for more details.
> > +
> > + You should have received a copy of the GNU General Public License
> > + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> > +
> > +#include <stdio.h>
> > +
> > +extern void libsection1_test ();
> > +
> > +int
> > +main ()
> > +{
> > + libsection1_test ();
> > + printf ("in section exec\n");
> > +
> > + return 0;
> > +}
> > diff --git a/gdb/testsuite/gdb.debuginfod/section.exp b/gdb/testsuite/gdb.debuginfod/section.exp
> > new file mode 100644
> > index 00000000000..35086d15705
> > --- /dev/null
> > +++ b/gdb/testsuite/gdb.debuginfod/section.exp
> > @@ -0,0 +1,157 @@
> > +# Copyright 2023-2024 Free Software Foundation, Inc.
> > +
> > +# This program is free software; you can redistribute it and/or modify
> > +# it under the terms of the GNU General Public License as published by
> > +# the Free Software Foundation; either version 3 of the License, or
> > +# (at your option) any later version.
> > +#
> > +# This program is distributed in the hope that it will be useful,
> > +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> > +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > +# GNU General Public License for more details.
> > +#
> > +# You should have received a copy of the GNU General Public License
> > +# along with this program. If not, see <http://www.gnu.org/licenses/>.
> > +
> > +# Test debuginfod functionality
> > +
> > +standard_testfile
> > +
> > +load_lib debuginfod-support.exp
> > +
> > +clean_restart
> > +require allow_debuginfod_tests
> > +require allow_debuginfod_section_downloads
> > +
> > +# BINFILE calls a function from LIB_SL1.
> > +set libfile1 "libsection1"
> > +set libsrc1 $srcdir/$subdir/$libfile1.c
> > +set lib_sl1 [standard_output_file $libfile1.sl]
> > +
> > +# LIB_SL1 calls functions from LIB_SL2.
> > +set libfile2 "libsection2"
> > +set libsrc2 $srcdir/$subdir/$libfile2.c
> > +set lib_sl2 [standard_output_file $libfile2.sl]
> > +
> > +# Build LIB_SL2, LIB_SL1 and BINFILE.
> > +if { [build_executable "build $libfile2" $lib_sl2 $libsrc2 \
> > + {debug build-id shlib}] != 0 } {
> > + return -1
> > +}
> > +
> > +if { [build_executable "build $libfile1" $lib_sl1 $libsrc1 \
> > + [list debug build-id shlib_pthreads shlib=$lib_sl2]] != 0 } {
> > + return -1
> > +}
> > +
> > +if { [build_executable "build executable" $binfile $srcfile \
> > + [list debug build-id shlib=$lib_sl1 shlib=$lib_sl2]] != 0 } {
> > + return -1
> > +}
> > +
> > +# Make sure libsection1 and libsection2 contain .gdb_index.
> > +if { [ensure_gdb_index $lib_sl1 "" "libsection1"] != 1 } {
> > + untested "failed to add .gdb_index to $libfile1"
> > + return -1
> > +}
> > +
> > +if { [ensure_gdb_index $lib_sl2 "" "libsection2"] != 1 } {
> > + untested "failed to add .gdb_index to $libfile2"
> > + return -1
> > +}
> > +
> > +# Strip solib debuginfo into separate files.
> > +if { [gdb_gnu_strip_debug $lib_sl1 ""] != 0} {
> > + fail "strip $lib_sl1 debuginfo"
> > + return -1
> > +}
> > +
> > +if { [gdb_gnu_strip_debug $lib_sl2 ""] != 0} {
> > + fail "strip $lib_sl2 debuginfo"
> > + return -1
> > +}
> > +
> > +# Move debuginfo files into directory that debuginfod will serve from.
> > +set debugdir [standard_output_file "debug"]
> > +set debuginfo_sl1 [standard_output_file $libfile1.sl.debug]
> > +set debuginfo_sl2 [standard_output_file $libfile2.sl.debug]
> > +
> > +file mkdir $debugdir
> > +file rename -force $debuginfo_sl1 $debugdir
> > +file rename -force $debuginfo_sl2 $debugdir
> > +
> > +# Restart GDB and clear the debuginfod client cache. Then load BINFILE into
> > +# GDB and start running it. Match output with pattern RES and use TESTNAME
> > +# as the test name.
> > +proc_with_prefix clean_restart_with_prompt { binfile testname } {
> > + global cache
> > +
> > + # Delete client cache so debuginfo downloads again.
> > + file delete -force $cache
> > + clean_restart
> > +
> > + gdb_test_no_output "set debuginfod enabled on" \
> > + "clean_restart enable $testname"
> > + gdb_load $binfile
> > +
> > + runto_main
> > +}
> > +
> > +# Tests with no debuginfod server running.
> > +proc_with_prefix no_url { } {
> > + global binfile libfile1 libfile2
> > +
> > + gdb_load $binfile
> > + if {![runto_main]} {
> > + return
> > + }
> > +
> > + # Check that no section is downloaded and no debuginfo is found.
> > + gdb_test "info sharedlibrary" ".*Yes \\(\\*\\).*$libfile1.*" \
> > + "found no url lib1"
> > + gdb_test "info sharedlibrary" ".*Yes \\(\\*\\).*$libfile2.*" \
> > + "found no url lib2"
> > +}
> > +
> > +# Tests with a debuginfod server running.
> > +proc_with_prefix local_url { } {
> > + global binfile
> > + global libsrc1 lib_sl1 libfile1
> > + global libsrc2 lib_sl2 libfile2
> > + global debugdir db
> > +
> > + set url [start_debuginfod $db $debugdir]
> > + if { $url == "" } {
> > + unresolved "failed to start debuginfod server"
> > + return
> > + }
> > +
> > + # Point GDB to the server.
> > + setenv DEBUGINFOD_URLS $url
> > +
> > + clean_restart_with_prompt $binfile "index"
> > +
> > + # Download debuginfo when stepping into a function.
> > + set res ".*separate debug info for $lib_sl1.*\"In ${libfile1}\\\\n\".*"
> > + gdb_test "step" $res "step"
> > +
> > + clean_restart_with_prompt $binfile "break"
> > +
> > + # Download debuginfo when setting a breakpoint.
> > + set res "Download.*separate debug info for $lib_sl2.*"
> > + gdb_test "br libsection2_test" $res "break set"
> > +
> > + # Hit the breakpoint.
> > + set res ".*Breakpoint 2, libsection2_test.*\"In ${libfile2}\\\\n\".*"
> > + gdb_test "c" $res "break continue"
> > +}
> > +
> > +# Create CACHE and DB directories ready for debuginfod to use.
> > +prepare_for_debuginfod cache db
> > +
> > +with_debuginfod_env $cache {
> > + no_url
> > + local_url
> > +}
> > +
> > +stop_debuginfod
> > diff --git a/gdb/testsuite/lib/debuginfod-support.exp b/gdb/testsuite/lib/debuginfod-support.exp
> > index 0096448567e..4d8bdcdc7cb 100644
> > --- a/gdb/testsuite/lib/debuginfod-support.exp
> > +++ b/gdb/testsuite/lib/debuginfod-support.exp
> > @@ -113,6 +113,8 @@ proc with_debuginfod_env { cache body } {
> > proc start_debuginfod { db debugdir } {
> > global debuginfod_spawn_id spawn_id
> >
> > + set logfile [standard_output_file "server_log"]
> > +
> > # Find an unused port.
> > set port 7999
> > set found false
> > @@ -127,7 +129,8 @@ proc start_debuginfod { db debugdir } {
> > set old_spawn_id $spawn_id
> > }
> >
> > - spawn debuginfod -vvvv -d $db -p $port -F $debugdir
> > + spawn sh -c "debuginfod -vvvv -d $db -p $port -F $debugdir 2>&1 \
> > + | tee $logfile"
> > set debuginfod_spawn_id $spawn_id
> >
> > if { [info exists old_spawn_id] } {
> > @@ -194,3 +197,23 @@ proc stop_debuginfod { } {
> > unset debuginfod_spawn_id
> > }
> > }
> > +
> > +# Return true if gdb is configured to download ELF/DWARF sections from
> > +# debuginfod servers. Otherwise return false.
> > +proc allow_debuginfod_section_downloads { } {
> > + set cmd "maint set debuginfod download-sections on"
> > + set msg "enable section downloads"
> > +
> > + gdb_test_multiple $cmd $msg {
> > + -re -wrap ".*not compiled into GDB.*" {
> > + return false
> > + }
> > + -re -wrap "^" {
> > + return true
> > + }
> > + -re -wrap "" {
> > + fail "$gdb_test_name (unexpected output)"
> > + return false
> > + }
> > + }
> > +}
> > diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
> > index 5fb81f6e28d..a3680cb6371 100644
> > --- a/gdb/testsuite/lib/gdb.exp
> > +++ b/gdb/testsuite/lib/gdb.exp
> > @@ -9515,10 +9515,14 @@ proc get_index_type { objfile { testname "" } } {
> > # STYLE controls which style of index to add, if needed. The empty
> > # string (the default) means .gdb_index; "-dwarf-5" means .debug_names.
> >
> > -proc ensure_gdb_index { binfile {style ""} } {
> > +proc ensure_gdb_index { binfile {style ""} {testname_prefix ""} } {
> > set testfile [file tail $binfile]
> > -
> > set test "check if index present"
> > +
> > + if { $testname_prefix != "" } {
> > + set test "$testname_prefix $test"
> > + }
> > +
> > set index_type [get_index_type $testfile $test]
> >
> > if { $index_type eq "gdb" || $index_type eq "dwarf5" } {
> > --
> > 2.43.0
> >
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING*3][PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading
2024-05-31 20:58 ` [PING*2][PATCH " Aaron Merey
@ 2024-06-07 20:48 ` Aaron Merey
2024-06-14 18:24 ` Aaron Merey
1 sibling, 0 replies; 18+ messages in thread
From: Aaron Merey @ 2024-06-07 20:48 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 31, 2024 at 4:58 PM Aaron Merey <amerey@redhat.com> wrote:
>
> Ping
>
> Thanks,
> Aaron
>
> On Fri, May 17, 2024 at 10:10 AM Aaron Merey <amerey@redhat.com> wrote:
> >
> > Ping
> >
> > Thanks,
> > Aaron
> >
> > On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
> > >
> > > v6 fixes a comment for dwarf2_has_separate_index and merges
> > > lookup_gdb_index_debuginfod with lookup_gdb_index.
> > >
> > > At the beginning of a session, gdb may attempt to download debuginfo
> > > for all shared libraries associated with the process or core file
> > > being debugged. This can be a waste of time and storage space when much
> > > of the debuginfo ends up not being used during the session.
> > >
> > > To reduce the gdb's startup latency and to download only the debuginfo
> > > that is really needed, this patch adds on-demand downloading of debuginfo.
> > >
> > > 'set debuginfo enabled on' now causes gdb to attempt to download a .gdb_index
> > > for each shared library instead of its full debuginfo. Each corresponding
> > > separate debuginfo will be deferred until gdb needs to expand symtabs
> > > associated with the debuginfo's index.
> > >
> > > Because these indices are significantly smaller than their corresponding
> > > debuginfo, this generally reduces the total amount of data gdb downloads.
> > > Reductions of 80%-95% have been observed when debugging large GUI programs.
> > >
> > > (gdb) set debuginfod enabled on
> > > (gdb) start
> > > Downloading section .gdb_index for /lib64/libcurl.so.4
> > > [...]
> > > 1826 client->server_mhandle = curl_multi_init ();
> > > (gdb) step
> > > Downloading separate debug info for /lib64/libcurl.so.4
> > > Downloading separate debug info for [libcurl dwz]
> > > Downloading source file /usr/src/debug/curl-7.85.0-6.fc37.x86_64/build-full/lib/../../lib/multi.c
> > > curl_multi_init () at ../../lib/multi.c:457
> > > 457 {
> > > (gdb)
> > >
> > > Some of the key functions below include dwarf2_has_separate_index which
> > > downloads the separate .gdb_index. If successful, the shared library
> > > objfile owns the index until the separate debug objfile is downloaded
> > > or confirmed to not be available.
> > >
> > > read_full_dwarf_from_debuginfod downloads the full debuginfo and
> > > initializes the separate debug objfile. It is called by functions
> > > such as dwarf2_gdb_index::expand_symtabs_matching and
> > > dwarf2_base_index_functions::find_pc_sect_compunit_symtab when symtab
> > > expansion is required.
> > > ---
> > > gdb/dwarf2/frame.c | 13 ++
> > > gdb/dwarf2/frame.h | 4 +
> > > gdb/dwarf2/index-cache.c | 22 ++-
> > > gdb/dwarf2/index-cache.h | 11 ++
> > > gdb/dwarf2/public.h | 5 +
> > > gdb/dwarf2/read-gdb-index.c | 123 +++++++++++++---
> > > gdb/dwarf2/read.c | 149 ++++++++++++++++++-
> > > gdb/dwarf2/read.h | 10 ++
> > > gdb/dwarf2/section.c | 3 +-
> > > gdb/elfread.c | 3 +-
> > > gdb/frame.c | 27 +++-
> > > gdb/objfile-flags.h | 4 +
> > > gdb/objfiles.h | 20 +++
> > > gdb/quick-symbol.h | 4 +
> > > gdb/symfile-debug.c | 11 ++
> > > gdb/symfile.c | 12 +-
> > > gdb/symtab.c | 18 ++-
> > > gdb/testsuite/gdb.debuginfod/libsection1.c | 43 ++++++
> > > gdb/testsuite/gdb.debuginfod/libsection2.c | 37 +++++
> > > gdb/testsuite/gdb.debuginfod/section.c | 29 ++++
> > > gdb/testsuite/gdb.debuginfod/section.exp | 157 +++++++++++++++++++++
> > > gdb/testsuite/lib/debuginfod-support.exp | 25 +++-
> > > gdb/testsuite/lib/gdb.exp | 8 +-
> > > 23 files changed, 702 insertions(+), 36 deletions(-)
> > > create mode 100644 gdb/testsuite/gdb.debuginfod/libsection1.c
> > > create mode 100644 gdb/testsuite/gdb.debuginfod/libsection2.c
> > > create mode 100644 gdb/testsuite/gdb.debuginfod/section.c
> > > create mode 100644 gdb/testsuite/gdb.debuginfod/section.exp
> > >
> > > diff --git a/gdb/dwarf2/frame.c b/gdb/dwarf2/frame.c
> > > index 9ebf3ac4ac4..074417765f4 100644
> > > --- a/gdb/dwarf2/frame.c
> > > +++ b/gdb/dwarf2/frame.c
> > > @@ -1621,6 +1621,19 @@ set_comp_unit (struct objfile *objfile, struct comp_unit *unit)
> > > return dwarf2_frame_bfd_data.set (abfd, unit);
> > > }
> > >
> > > +/* See frame.h. */
> > > +
> > > +void
> > > +dwarf2_clear_frame_data (struct objfile *objfile)
> > > +{
> > > + bfd *abfd = objfile->obfd.get ();
> > > +
> > > + if (gdb_bfd_requires_relocations (abfd))
> > > + dwarf2_frame_objfile_data.clear (objfile);
> > > + else
> > > + dwarf2_frame_bfd_data.clear (abfd);
> > > +}
> > > +
> > > /* Find the FDE for *PC. Return a pointer to the FDE, and store the
> > > initial location associated with it into *PC. */
> > >
> > > diff --git a/gdb/dwarf2/frame.h b/gdb/dwarf2/frame.h
> > > index 2167310fbdf..423a9355b8a 100644
> > > --- a/gdb/dwarf2/frame.h
> > > +++ b/gdb/dwarf2/frame.h
> > > @@ -237,6 +237,10 @@ void dwarf2_append_unwinders (struct gdbarch *gdbarch);
> > > extern const struct frame_base *
> > > dwarf2_frame_base_sniffer (const frame_info_ptr &this_frame);
> > >
> > > +/* Delete OBJFILEs comp_unit. */
> > > +
> > > +extern void dwarf2_clear_frame_data (struct objfile * objfile);
> > > +
> > > /* Compute the DWARF CFA for a frame. */
> > >
> > > CORE_ADDR dwarf2_frame_cfa (const frame_info_ptr &this_frame);
> > > diff --git a/gdb/dwarf2/index-cache.c b/gdb/dwarf2/index-cache.c
> > > index 1720e925ee6..7096a3ea078 100644
> > > --- a/gdb/dwarf2/index-cache.c
> > > +++ b/gdb/dwarf2/index-cache.c
> > > @@ -219,14 +219,22 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> > > /* Compute where we would expect a gdb index file for this build id to be. */
> > > std::string filename = make_index_filename (build_id, INDEX4_SUFFIX);
> > >
> > > + return lookup_gdb_index (filename.c_str (), resource);
> > > +}
> > > +
> > > +/* See index-cache.h. */
> > > +
> > > +gdb::array_view<const gdb_byte>
> > > +index_cache::lookup_gdb_index (const char *filename,
> > > + std::unique_ptr<index_cache_resource> *resource)
> > > +{
> > > try
> > > {
> > > - index_cache_debug ("trying to read %s",
> > > - filename.c_str ());
> > > + index_cache_debug ("trying to read %s", filename);
> > >
> > > /* Try to map that file. */
> > > index_cache_resource_mmap *mmap_resource
> > > - = new index_cache_resource_mmap (filename.c_str ());
> > > + = new index_cache_resource_mmap (filename);
> > >
> > > /* Yay, it worked! Hand the resource to the caller. */
> > > resource->reset (mmap_resource);
> > > @@ -238,7 +246,7 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> > > catch (const gdb_exception_error &except)
> > > {
> > > index_cache_debug ("couldn't read %s: %s",
> > > - filename.c_str (), except.what ());
> > > + filename, except.what ());
> > > }
> > >
> > > return {};
> > > @@ -255,6 +263,12 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> > > return {};
> > > }
> > >
> > > +gdb::array_view<const gdb_byte>
> > > +index_cache::lookup_gdb_index_debuginfod (const char *index_path,
> > > + std::unique_ptr<index_cache_resource> *resource)
> > > +{
> > > + return {};
> > > +}
> > > #endif
> > >
> > > /* See dwarf-index-cache.h. */
> > > diff --git a/gdb/dwarf2/index-cache.h b/gdb/dwarf2/index-cache.h
> > > index 95f217ed961..f42e6918ce2 100644
> > > --- a/gdb/dwarf2/index-cache.h
> > > +++ b/gdb/dwarf2/index-cache.h
> > > @@ -93,6 +93,17 @@ class index_cache
> > > lookup_gdb_index (const bfd_build_id *build_id,
> > > std::unique_ptr<index_cache_resource> *resource);
> > >
> > > + /* Look for an index file located at INDEX_PATH. If found, return the
> > > + contents as an array_view and store the underlying resources (allocated
> > > + memory, mapped file, etc) in RESOURCE. The returned array_view is valid
> > > + as long as RESOURCE is not destroyed.
> > > +
> > > + If no matching index file is found, return an empty array view. This
> > > + function does not exit early if the index cache has not been enabled. */
> > > + gdb::array_view<const gdb_byte>
> > > + lookup_gdb_index (const char *filename,
> > > + std::unique_ptr<index_cache_resource> *resource);
> > > +
> > > /* Return the number of cache hits. */
> > > unsigned int n_hits () const
> > > { return m_n_hits; }
> > > diff --git a/gdb/dwarf2/public.h b/gdb/dwarf2/public.h
> > > index bc419ff208f..6523854386c 100644
> > > --- a/gdb/dwarf2/public.h
> > > +++ b/gdb/dwarf2/public.h
> > > @@ -44,4 +44,9 @@ extern bool dwarf2_initialize_objfile
> > >
> > > extern void dwarf2_build_frame_info (struct objfile *);
> > >
> > > +/* Query debuginfod for the .gdb_index associated with OBJFILE.
> > > + Used to defer separate debuginfo downloading until necessary. */
> > > +
> > > +extern bool dwarf2_has_separate_index (struct objfile *);
> > > +
> > > #endif /* DWARF2_PUBLIC_H */
> > > diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
> > > index e743ce13978..97b8137aabf 100644
> > > --- a/gdb/dwarf2/read-gdb-index.c
> > > +++ b/gdb/dwarf2/read-gdb-index.c
> > > @@ -139,6 +139,8 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > > gdb.dwarf2/gdb-index.exp testcase. */
> > > void dump (struct objfile *objfile) override;
> > >
> > > + /* Calls do_expand_symtabs_matching and triggers debuginfo downloading
> > > + if necessary. */
> > > bool expand_symtabs_matching
> > > (struct objfile *objfile,
> > > gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > > @@ -147,8 +149,59 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > > gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> > > block_search_flags search_flags,
> > > domain_search_flags domain) override;
> > > +
> > > + /* Expand symtabs matching a given symbol or file. */
> > > + bool do_expand_symtabs_matching
> > > + (struct objfile *objfile,
> > > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > > + const lookup_name_info *lookup_name,
> > > + gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
> > > + gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> > > + block_search_flags search_flags,
> > > + domain_search_flags domain);
> > > +
> > > + /* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
> > > + debuginfo if necessary. */
> > > + void expand_all_symtabs (struct objfile *objfile) override;
> > > +
> > > + /* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
> > > + debuginfo if necessary. */
> > > + struct symtab *find_last_source_symtab (struct objfile *objfile) override;
> > > };
> > >
> > > +void
> > > +dwarf2_gdb_index::expand_all_symtabs (struct objfile *objfile)
> > > +{
> > > + try
> > > + {
> > > + dwarf2_base_index_functions::expand_all_symtabs (objfile);
> > > + }
> > > + catch (const gdb_exception &e)
> > > + {
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + exception_print (gdb_stderr, e);
> > > + else
> > > + read_full_dwarf_from_debuginfod (objfile, this);
> > > + }
> > > +}
> > > +
> > > +struct symtab *
> > > +dwarf2_gdb_index::find_last_source_symtab (struct objfile *objfile)
> > > +{
> > > + try
> > > + {
> > > + return dwarf2_base_index_functions::find_last_source_symtab (objfile);
> > > + }
> > > + catch (const gdb_exception &e)
> > > + {
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + exception_print (gdb_stderr, e);
> > > + else
> > > + read_full_dwarf_from_debuginfod (objfile, this);
> > > + return nullptr;
> > > + }
> > > +}
> > > +
> > > /* This dumps minimal information about the index.
> > > It is called via "mt print objfiles".
> > > One use is to verify .gdb_index has been loaded by the
> > > @@ -264,7 +317,7 @@ dw2_expand_marked_cus
> > > }
> > >
> > > bool
> > > -dwarf2_gdb_index::expand_symtabs_matching
> > > +dwarf2_gdb_index::do_expand_symtabs_matching
> > > (struct objfile *objfile,
> > > gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > > const lookup_name_info *lookup_name,
> > > @@ -312,6 +365,38 @@ dwarf2_gdb_index::expand_symtabs_matching
> > > return result;
> > > }
> > >
> > > +bool
> > > +dwarf2_gdb_index::expand_symtabs_matching
> > > + (struct objfile *objfile,
> > > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > > + const lookup_name_info *lookup_name,
> > > + gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
> > > + gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> > > + block_search_flags search_flags,
> > > + domain_search_flags domain)
> > > +{
> > > + if ((objfile->flags & OBJF_READNEVER) != 0)
> > > + return false;
> > > +
> > > + try
> > > + {
> > > + return do_expand_symtabs_matching (objfile, file_matcher, lookup_name,
> > > + symbol_matcher, expansion_notify,
> > > + search_flags, domain);
> > > + }
> > > + catch (const gdb_exception &e)
> > > + {
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + {
> > > + exception_print (gdb_stderr, e);
> > > + return false;
> > > + }
> > > +
> > > + read_full_dwarf_from_debuginfod (objfile, this);
> > > + return true;
> > > + }
> > > +}
> > > +
> > > quick_symbol_functions_up
> > > mapped_gdb_index::make_quick_functions () const
> > > {
> > > @@ -647,28 +732,32 @@ dwarf2_read_gdb_index
> > >
> > > /* If there is a .dwz file, read it so we can get its CU list as
> > > well. */
> > > - dwz = dwarf2_get_dwz_file (per_bfd);
> > > - if (dwz != NULL)
> > > + if (get_gdb_index_contents_dwz != nullptr)
> > > {
> > > mapped_gdb_index dwz_map;
> > > const gdb_byte *dwz_types_ignore;
> > > offset_type dwz_types_elements_ignore;
> > > + dwz = dwarf2_get_dwz_file (per_bfd);
> > >
> > > - gdb::array_view<const gdb_byte> dwz_index_content
> > > - = get_gdb_index_contents_dwz (objfile, dwz);
> > > -
> > > - if (dwz_index_content.empty ())
> > > - return 0;
> > > -
> > > - if (!read_gdb_index_from_buffer (bfd_get_filename (dwz->dwz_bfd.get ()),
> > > - 1, dwz_index_content, &dwz_map,
> > > - &dwz_list, &dwz_list_elements,
> > > - &dwz_types_ignore,
> > > - &dwz_types_elements_ignore))
> > > + if (dwz != nullptr)
> > > {
> > > - warning (_("could not read '.gdb_index' section from %s; skipping"),
> > > - bfd_get_filename (dwz->dwz_bfd.get ()));
> > > - return 0;
> > > + gdb::array_view<const gdb_byte> dwz_index_content
> > > + = get_gdb_index_contents_dwz (objfile, dwz);
> > > +
> > > + if (dwz_index_content.empty ())
> > > + return 0;
> > > +
> > > + if (!read_gdb_index_from_buffer (bfd_get_filename
> > > + (dwz->dwz_bfd.get ()),
> > > + 1, dwz_index_content, &dwz_map,
> > > + &dwz_list, &dwz_list_elements,
> > > + &dwz_types_ignore,
> > > + &dwz_types_elements_ignore))
> > > + {
> > > + warning (_("could not read '.gdb_index' section from %s; skipping"),
> > > + bfd_get_filename (dwz->dwz_bfd.get ()));
> > > + return 0;
> > > + }
> > > }
> > > }
> > >
> > > diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
> > > index 00e8608f9ce..4461c76347e 100644
> > > --- a/gdb/dwarf2/read.c
> > > +++ b/gdb/dwarf2/read.c
> > > @@ -33,6 +33,7 @@
> > > #include "dwarf2/attribute.h"
> > > #include "dwarf2/comp-unit-head.h"
> > > #include "dwarf2/cu.h"
> > > +#include "dwarf2/frame.h"
> > > #include "dwarf2/index-cache.h"
> > > #include "dwarf2/index-common.h"
> > > #include "dwarf2/leb.h"
> > > @@ -96,6 +97,8 @@
> > > #include "gdbsupport/thread-pool.h"
> > > #include "run-on-main-thread.h"
> > > #include "dwarf2/parent-map.h"
> > > +#include "inferior.h"
> > > +#include "debuginfod-support.h"
> > >
> > > /* When == 1, print basic high level tracing messages.
> > > When > 1, be more verbose.
> > > @@ -3020,7 +3023,7 @@ dwarf2_base_index_functions::find_per_cu (dwarf2_per_bfd *per_bfd,
> > > }
> > >
> > > struct compunit_symtab *
> > > -dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> > > +dwarf2_base_index_functions::do_find_pc_sect_compunit_symtab
> > > (struct objfile *objfile,
> > > struct bound_minimal_symbol msymbol,
> > > CORE_ADDR pc,
> > > @@ -3051,6 +3054,32 @@ dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> > > return result;
> > > }
> > >
> > > +struct compunit_symtab *
> > > +dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> > > + (struct objfile *objfile,
> > > + struct bound_minimal_symbol msymbol,
> > > + CORE_ADDR pc,
> > > + struct obj_section *section,
> > > + int warn_if_readin)
> > > +{
> > > + if ((objfile->flags & OBJF_READNEVER) != 0)
> > > + return nullptr;
> > > +
> > > + try
> > > + {
> > > + return do_find_pc_sect_compunit_symtab (objfile, msymbol, pc,
> > > + section, warn_if_readin);
> > > + }
> > > + catch (const gdb_exception &e)
> > > + {
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + exception_print (gdb_stderr, e);
> > > + else
> > > + read_full_dwarf_from_debuginfod (objfile, this);
> > > + return nullptr;
> > > + }
> > > +}
> > > +
> > > void
> > > dwarf2_base_index_functions::map_symbol_filenames
> > > (struct objfile *objfile,
> > > @@ -3207,6 +3236,30 @@ get_gdb_index_contents_from_cache_dwz (objfile *obj, dwz_file *dwz)
> > > return global_index_cache.lookup_gdb_index (build_id, &dwz->index_cache_res);
> > > }
> > >
> > > +/* Query debuginfod for the .gdb_index matching OBJFILE's build-id. Return the
> > > + contents if successful. */
> > > +
> > > +static gdb::array_view<const gdb_byte>
> > > +get_gdb_index_contents_from_debuginfod (objfile *objfile, dwarf2_per_bfd *per_bfd)
> > > +{
> > > + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > > + if (build_id == nullptr)
> > > + return {};
> > > +
> > > + gdb::unique_xmalloc_ptr<char> index_path;
> > > + scoped_fd fd = debuginfod_section_query (build_id->data, build_id->size,
> > > + bfd_get_filename
> > > + (objfile->obfd.get ()),
> > > + ".gdb_index",
> > > + &index_path);
> > > + if (fd.get () < 0)
> > > + return {};
> > > +
> > > + return global_index_cache.lookup_gdb_index
> > > + (index_path.get (), &per_bfd->index_cache_res);
> > > +}
> > > +
> > > +
> > > static void start_debug_info_reader (dwarf2_per_objfile *);
> > >
> > > /* See dwarf2/public.h. */
> > > @@ -3216,11 +3269,13 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> > > const struct dwarf2_debug_sections *names,
> > > bool can_copy)
> > > {
> > > - if (!dwarf2_has_info (objfile, names, can_copy))
> > > + if (!dwarf2_has_info (objfile, names, can_copy)
> > > + && (objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > return false;
> > >
> > > dwarf2_per_objfile *per_objfile = get_dwarf2_per_objfile (objfile);
> > > dwarf2_per_bfd *per_bfd = per_objfile->per_bfd;
> > > + bool separate_index = false;
> > >
> > > dwarf_read_debug_printf ("called");
> > >
> > > @@ -3256,6 +3311,15 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> > > dwarf_read_debug_printf ("found gdb index from cache");
> > > global_index_cache.hit ();
> > > }
> > > + /* Try to read just a separately downloaded gdb index. */
> > > + else if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0
> > > + && dwarf2_read_gdb_index (per_objfile,
> > > + get_gdb_index_contents_from_debuginfod,
> > > + nullptr))
> > > + {
> > > + dwarf_read_debug_printf ("found .gdb_index from debuginfod");
> > > + separate_index = true;
> > > + }
> > > else
> > > {
> > > global_index_cache.miss ();
> > > @@ -3267,8 +3331,89 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> > > if (dwarf_synchronous)
> > > per_bfd->index_table->wait_completely ();
> > > objfile->qf.push_front (per_bfd->index_table->make_quick_functions ());
> > > +
> > > + if (separate_index)
> > > + objfile->qf.begin ()->get ()->from_separate_index = true;
> > > + }
> > > +
> > > + return true;
> > > +}
> > > +
> > > +/* See read.h. */
> > > +
> > > +void
> > > +read_full_dwarf_from_debuginfod (struct objfile *objfile,
> > > + dwarf2_base_index_functions *fncs)
> > > +{
> > > + gdb_assert ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> > > +
> > > + SCOPE_EXIT { objfile->remove_deferred_status (); };
> > > +
> > > + const struct bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > > + if (build_id == nullptr)
> > > + return;
> > > +
> > > + const char *filename = bfd_get_filename (objfile->obfd.get ());
> > > + gdb::unique_xmalloc_ptr<char> symfile_path;
> > > + scoped_fd fd;
> > > +
> > > + fd = debuginfod_debuginfo_query (build_id->data, build_id->size,
> > > + filename, &symfile_path);
> > > + if (fd.get () < 0)
> > > + return;
> > > +
> > > + /* Separate debuginfo successfully retrieved from server. */
> > > + gdb_bfd_ref_ptr debug_bfd = symfile_bfd_open (symfile_path.get ());
> > > + if (debug_bfd == nullptr
> > > + || !build_id_verify (debug_bfd.get (), build_id->size, build_id->data))
> > > + {
> > > + warning (_("File \"%s\" from debuginfod cannot be opened as bfd"),
> > > + filename);
> > > + return;
> > > }
> > >
> > > + /* Clear frame data so it can be recalculated using DWARF. */
> > > + dwarf2_clear_frame_data (objfile);
> > > +
> > > + /* This may also trigger a dwz download. */
> > > + symbol_file_add_separate (debug_bfd, symfile_path.get (),
> > > + current_inferior ()->symfile_flags, objfile);
> > > +}
> > > +
> > > +/* See public.h. */
> > > +
> > > +bool
> > > +dwarf2_has_separate_index (struct objfile *objfile)
> > > +{
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0)
> > > + return true;
> > > + if ((objfile->flags & OBJF_MAINLINE) != 0)
> > > + return false;
> > > + if (!IS_DIR_SEPARATOR (*objfile_filename (objfile)))
> > > + return false;
> > > +
> > > + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > > +
> > > + if (build_id == nullptr)
> > > + return false;
> > > +
> > > + gdb::unique_xmalloc_ptr<char> index_path;
> > > + scoped_fd fd = debuginfod_section_query (build_id->data,
> > > + build_id->size,
> > > + bfd_get_filename
> > > + (objfile->obfd.get ()),
> > > + ".gdb_index",
> > > + &index_path);
> > > +
> > > + if (fd.get () < 0)
> > > + return false;
> > > +
> > > + /* We found a separate .gdb_index file so a separate debuginfo file
> > > + should exist, but we don't want to download it until necessary.
> > > + Attach the index to this objfile and defer the debuginfo download
> > > + until gdb needs to expand symtabs referenced by the index. */
> > > + objfile->flags |= OBJF_DOWNLOAD_DEFERRED;
> > > + dwarf2_initialize_objfile (objfile);
> > > return true;
> > > }
> > >
> > > diff --git a/gdb/dwarf2/read.h b/gdb/dwarf2/read.h
> > > index 73def88c4c0..099d3d65c27 100644
> > > --- a/gdb/dwarf2/read.h
> > > +++ b/gdb/dwarf2/read.h
> > > @@ -860,6 +860,10 @@ struct dwarf2_base_index_functions : public quick_symbol_functions
> > > CORE_ADDR pc, struct obj_section *section, int warn_if_readin)
> > > override;
> > >
> > > + struct compunit_symtab *do_find_pc_sect_compunit_symtab
> > > + (struct objfile *objfile, struct bound_minimal_symbol msymbol,
> > > + CORE_ADDR pc, struct obj_section *section, int warn_if_readin);
> > > +
> > > struct compunit_symtab *find_compunit_symtab_by_address
> > > (struct objfile *objfile, CORE_ADDR address) override
> > > {
> > > @@ -932,4 +936,10 @@ extern void create_all_units (dwarf2_per_objfile *per_objfile);
> > >
> > > extern htab_up create_quick_file_names_table (unsigned int nr_initial_entries);
> > >
> > > +/* If OBJFILE contains information from a separately downloaded .gdb_index,
> > > + attempt to download the full debuginfo. */
> > > +
> > > +extern void read_full_dwarf_from_debuginfod (struct objfile *,
> > > + dwarf2_base_index_functions *);
> > > +
> > > #endif /* DWARF2READ_H */
> > > diff --git a/gdb/dwarf2/section.c b/gdb/dwarf2/section.c
> > > index 6aab0b9aee9..481f37351f4 100644
> > > --- a/gdb/dwarf2/section.c
> > > +++ b/gdb/dwarf2/section.c
> > > @@ -53,7 +53,8 @@ dwarf2_section_info::get_bfd_owner () const
> > > section = get_containing_section ();
> > > gdb_assert (!section->is_virtual);
> > > }
> > > - gdb_assert (section->s.section != nullptr);
> > > + if (section->s.section == nullptr)
> > > + error (_("Can't find owner of DWARF section."));
> > > return section->s.section->owner;
> > > }
> > >
> > > diff --git a/gdb/elfread.c b/gdb/elfread.c
> > > index 7a6a8cadced..0f59bb1a58d 100644
> > > --- a/gdb/elfread.c
> > > +++ b/gdb/elfread.c
> > > @@ -1217,7 +1217,8 @@ elf_symfile_read_dwarf2 (struct objfile *objfile,
> > > && objfile->separate_debug_objfile_backlink == NULL)
> > > {
> > > if (objfile->find_and_add_separate_symbol_file (symfile_flags))
> > > - gdb_assert (objfile->separate_debug_objfile != nullptr);
> > > + gdb_assert (objfile->separate_debug_objfile != nullptr
> > > + || (objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> > > else
> > > has_dwarf2 = false;
> > > }
> > > diff --git a/gdb/frame.c b/gdb/frame.c
> > > index 43efac07d95..1718ab5086c 100644
> > > --- a/gdb/frame.c
> > > +++ b/gdb/frame.c
> > > @@ -1778,14 +1778,10 @@ restore_selected_frame (frame_id frame_id, int frame_level)
> > > selected_frame = nullptr;
> > > }
> > >
> > > -/* Lookup the frame_info object for the selected frame FRAME_ID /
> > > - FRAME_LEVEL and cache the result.
> > > -
> > > - If FRAME_LEVEL > 0 and the originally selected frame isn't found,
> > > - warn and select the innermost (current) frame. */
> > > +/* See lookup_selected_frame. */
> > >
> > > static void
> > > -lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> > > +lookup_selected_frame_1 (struct frame_id &a_frame_id, int frame_level)
> > > {
> > > frame_info_ptr frame = NULL;
> > > int count;
> > > @@ -1857,6 +1853,25 @@ lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> > > }
> > > }
> > >
> > > +/* Lookup the frame_info object for the selected frame FRAME_ID /
> > > + FRAME_LEVEL and cache the result.
> > > +
> > > + If FRAME_LEVEL > 0 and the originally selected frame isn't found,
> > > + warn and select the innermost (current) frame. */
> > > +
> > > +static void
> > > +lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> > > +{
> > > + lookup_selected_frame_1 (selected_frame_id, selected_frame_level);
> > > +
> > > + /* It is possible for lookup_selected_frame_1 to cause a new objfile
> > > + to be loaded. However some objfile observers may choose to clear
> > > + selected_frame when an objfile is loaded. Work around this by
> > > + calling lookup_selected_frame_1 again if the first call failed. */
> > > + if (selected_frame == nullptr)
> > > + lookup_selected_frame_1 (selected_frame_id, selected_frame_level);
> > > +}
> > > +
> > > bool
> > > has_stack_frames ()
> > > {
> > > diff --git a/gdb/objfile-flags.h b/gdb/objfile-flags.h
> > > index 43d888e2a5a..08e4a75a9f2 100644
> > > --- a/gdb/objfile-flags.h
> > > +++ b/gdb/objfile-flags.h
> > > @@ -56,6 +56,10 @@ enum objfile_flag : unsigned
> > > /* User requested that we do not read this objfile's symbolic
> > > information. */
> > > OBJF_READNEVER = 1 << 6,
> > > +
> > > + /* A separate .gdb_index has been downloaded for this objfile.
> > > + Debuginfo for this objfile can be downloaded when required. */
> > > + OBJF_DOWNLOAD_DEFERRED = 1 << 7,
> > > };
> > >
> > > DEF_ENUM_FLAGS_TYPE (enum objfile_flag, objfile_flags);
> > > diff --git a/gdb/objfiles.h b/gdb/objfiles.h
> > > index 1c8f79a2826..8c8f27f87f1 100644
> > > --- a/gdb/objfiles.h
> > > +++ b/gdb/objfiles.h
> > > @@ -628,6 +628,26 @@ struct objfile
> > > domain_search_flags domain,
> > > bool *symbol_found_p);
> > >
> > > + /* Indicate that the acquisition of this objfile's separate debug objfile
> > > + is no longer deferred. Used when the debug objfile has been acquired
> > > + or could not be found. */
> > > + void remove_deferred_status ()
> > > + {
> > > + flags &= ~OBJF_DOWNLOAD_DEFERRED;
> > > +
> > > + /* Remove quick_symbol_functions derived from a separately downloaded
> > > + index. If available the separate debug objfile's index will be used
> > > + instead, since that objfile actually contains the symbols and CUs
> > > + referenced in the index.
> > > +
> > > + No more than one element of qf should have from_separate_index set
> > > + to true. */
> > > + qf.remove_if ([&] (const quick_symbol_functions_up &qf_up)
> > > + {
> > > + return qf_up->from_separate_index;
> > > + });
> > > + }
> > > +
> > > /* Return the relocation offset applied to SECTION. */
> > > CORE_ADDR section_offset (bfd_section *section) const
> > > {
> > > diff --git a/gdb/quick-symbol.h b/gdb/quick-symbol.h
> > > index 676c3ede373..8b91b7765d5 100644
> > > --- a/gdb/quick-symbol.h
> > > +++ b/gdb/quick-symbol.h
> > > @@ -191,6 +191,10 @@ struct quick_symbol_functions
> > > virtual void compute_main_name (struct objfile *objfile)
> > > {
> > > }
> > > +
> > > + /* True if this quick_symbol_functions is derived from a separately
> > > + downloaded index. */
> > > + bool from_separate_index = false;
> > > };
> > >
> > > typedef std::unique_ptr<quick_symbol_functions> quick_symbol_functions_up;
> > > diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c
> > > index 330a520dca6..8099d464081 100644
> > > --- a/gdb/symfile-debug.c
> > > +++ b/gdb/symfile-debug.c
> > > @@ -36,6 +36,7 @@
> > > #include "cli/cli-style.h"
> > > #include "build-id.h"
> > > #include "debuginfod-support.h"
> > > +#include "dwarf2/public.h"
> > >
> > > /* We need to save a pointer to the real symbol functions.
> > > Plus, the debug versions are malloc'd because we have to NULL out the
> > > @@ -610,6 +611,16 @@ objfile::find_and_add_separate_symbol_file (symfile_add_flags symfile_flags)
> > > = simple_find_and_open_separate_symbol_file
> > > (this, find_separate_debug_file_by_debuglink, &warnings);
> > >
> > > + /* Attempt to download only an index from the separate debug info.
> > > + As with debuginfod_find_and_open_separate_symbol_file, only attempt
> > > + this once. */
> > > + if (debug_bfd == nullptr && attempt == 0
> > > + && dwarf2_has_separate_index (this))
> > > + {
> > > + has_dwarf2 = true;
> > > + break;
> > > + }
> > > +
> > > /* Only try debuginfod on the first attempt. Sure, we could imagine
> > > an extension that somehow adds the required debug info to the
> > > debuginfod server but, at least for now, we don't support this
> > > diff --git a/gdb/symfile.c b/gdb/symfile.c
> > > index 7ebb07d5394..35c4a3b78e9 100644
> > > --- a/gdb/symfile.c
> > > +++ b/gdb/symfile.c
> > > @@ -989,6 +989,10 @@ syms_from_objfile (struct objfile *objfile,
> > > static void
> > > finish_new_objfile (struct objfile *objfile, symfile_add_flags add_flags)
> > > {
> > > + struct objfile *parent = objfile->separate_debug_objfile_backlink;
> > > + bool was_deferred
> > > + = (parent != nullptr) && ((parent->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> > > +
> > > /* If this is the main symbol file we have to clean up all users of the
> > > old main symbol file. Otherwise it is sufficient to fixup all the
> > > breakpoints that may have been redefined by this symbol file. */
> > > @@ -999,7 +1003,7 @@ finish_new_objfile (struct objfile *objfile, symfile_add_flags add_flags)
> > >
> > > clear_symtab_users (add_flags);
> > > }
> > > - else if ((add_flags & SYMFILE_DEFER_BP_RESET) == 0)
> > > + else if ((add_flags & SYMFILE_DEFER_BP_RESET) == 0 && !was_deferred)
> > > {
> > > breakpoint_re_set ();
> > > }
> > > @@ -1120,6 +1124,12 @@ symbol_file_add_with_addrs (const gdb_bfd_ref_ptr &abfd, const char *name,
> > > if (objfile->sf != nullptr)
> > > finish_new_objfile (objfile, add_flags);
> > >
> > > + /* Remove deferred status now in case any observers trigger symtab
> > > + expansion. Otherwise gdb might try to read parent for psymbols
> > > + when it should read the separate debug objfile instead. */
> > > + if (parent != nullptr && ((parent->flags & OBJF_DOWNLOAD_DEFERRED) != 0))
> > > + parent->remove_deferred_status ();
> > > +
> > > gdb::observers::new_objfile.notify (objfile);
> > >
> > > return objfile;
> > > diff --git a/gdb/symtab.c b/gdb/symtab.c
> > > index 58648a8779d..26f1b1f631e 100644
> > > --- a/gdb/symtab.c
> > > +++ b/gdb/symtab.c
> > > @@ -2956,14 +2956,30 @@ find_pc_sect_compunit_symtab (CORE_ADDR pc, struct obj_section *section)
> > > if (best_cust != NULL)
> > > return best_cust;
> > >
> > > + int warn_if_readin = 1;
> > > +
> > > /* Not found in symtabs, search the "quick" symtabs (e.g. psymtabs). */
> > >
> > > for (objfile *objf : current_program_space->objfiles ())
> > > {
> > > + bool was_deferred = (objf->flags & OBJF_DOWNLOAD_DEFERRED) != 0;
> > > +
> > > struct compunit_symtab *result
> > > - = objf->find_pc_sect_compunit_symtab (msymbol, pc, section, 1);
> > > + = objf->find_pc_sect_compunit_symtab (msymbol, pc, section,
> > > + warn_if_readin);
> > > +
> > > if (result != NULL)
> > > return result;
> > > +
> > > + /* If OBJF's separate debug info was just acquired, disable
> > > + warn_if_readin for the next iteration of this loop. This prevents
> > > + a spurious warning in case an observer already triggered expansion
> > > + of the separate debug objfile's symtabs. */
> > > + if (was_deferred && objf->separate_debug_objfile != nullptr
> > > + && (objf->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + warn_if_readin = 0;
> > > + else if (warn_if_readin == 0)
> > > + warn_if_readin = 1;
> > > }
> > >
> > > return NULL;
> > > diff --git a/gdb/testsuite/gdb.debuginfod/libsection1.c b/gdb/testsuite/gdb.debuginfod/libsection1.c
> > > new file mode 100644
> > > index 00000000000..e4f05db0a79
> > > --- /dev/null
> > > +++ b/gdb/testsuite/gdb.debuginfod/libsection1.c
> > > @@ -0,0 +1,43 @@
> > > +/* This testcase is part of GDB, the GNU debugger.
> > > +
> > > + Copyright 2023-2024 Free Software Foundation, Inc.
> > > +
> > > + This program is free software; you can redistribute it and/or modify
> > > + it under the terms of the GNU General Public License as published by
> > > + the Free Software Foundation; either version 3 of the License, or
> > > + (at your option) any later version.
> > > +
> > > + This program is distributed in the hope that it will be useful,
> > > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > > + GNU General Public License for more details.
> > > +
> > > + You should have received a copy of the GNU General Public License
> > > + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> > > +
> > > +#include <stdio.h>
> > > +#include <pthread.h>
> > > +#include <unistd.h>
> > > +
> > > +extern void libsection2_test ();
> > > +extern void *libsection2_thread_test (void *);
> > > +
> > > +static volatile int flag = 0;
> > > +
> > > +void
> > > +libsection1_test ()
> > > +{
> > > + pthread_t thr;
> > > +
> > > + printf ("In libsection1\n");
> > > + libsection2_test ();
> > > +
> > > + pthread_create (&thr, NULL, libsection2_thread_test, (void *) &flag);
> > > +
> > > + /* Give the new thread a chance to actually enter libsection2_thread_test. */
> > > + while (!flag)
> > > + ;
> > > +
> > > + printf ("Cancelling thread\n");
> > > + pthread_cancel (thr);
> > > +}
> > > diff --git a/gdb/testsuite/gdb.debuginfod/libsection2.c b/gdb/testsuite/gdb.debuginfod/libsection2.c
> > > new file mode 100644
> > > index 00000000000..3c82808e5fe
> > > --- /dev/null
> > > +++ b/gdb/testsuite/gdb.debuginfod/libsection2.c
> > > @@ -0,0 +1,37 @@
> > > +/* This testcase is part of GDB, the GNU debugger.
> > > +
> > > + Copyright 2023-2024 Free Software Foundation, Inc.
> > > +
> > > + This program is free software; you can redistribute it and/or modify
> > > + it under the terms of the GNU General Public License as published by
> > > + the Free Software Foundation; either version 3 of the License, or
> > > + (at your option) any later version.
> > > +
> > > + This program is distributed in the hope that it will be useful,
> > > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > > + GNU General Public License for more details.
> > > +
> > > + You should have received a copy of the GNU General Public License
> > > + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> > > +
> > > +#include <stdio.h>
> > > +#include <pthread.h>
> > > +
> > > +void
> > > +libsection2_test ()
> > > +{
> > > + printf ("In libsection2\n");
> > > +}
> > > +
> > > +void *
> > > +libsection2_thread_test (void *arg)
> > > +{
> > > + int *flag = (int *) arg;
> > > + printf ("In thread test\n");
> > > +
> > > + while (1)
> > > + *flag = 1;
> > > +
> > > + return NULL;
> > > +}
> > > diff --git a/gdb/testsuite/gdb.debuginfod/section.c b/gdb/testsuite/gdb.debuginfod/section.c
> > > new file mode 100644
> > > index 00000000000..fa9ac536a6e
> > > --- /dev/null
> > > +++ b/gdb/testsuite/gdb.debuginfod/section.c
> > > @@ -0,0 +1,29 @@
> > > +/* This testcase is part of GDB, the GNU debugger.
> > > +
> > > + Copyright 2023-2024 Free Software Foundation, Inc.
> > > +
> > > + This program is free software; you can redistribute it and/or modify
> > > + it under the terms of the GNU General Public License as published by
> > > + the Free Software Foundation; either version 3 of the License, or
> > > + (at your option) any later version.
> > > +
> > > + This program is distributed in the hope that it will be useful,
> > > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > > + GNU General Public License for more details.
> > > +
> > > + You should have received a copy of the GNU General Public License
> > > + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> > > +
> > > +#include <stdio.h>
> > > +
> > > +extern void libsection1_test ();
> > > +
> > > +int
> > > +main ()
> > > +{
> > > + libsection1_test ();
> > > + printf ("in section exec\n");
> > > +
> > > + return 0;
> > > +}
> > > diff --git a/gdb/testsuite/gdb.debuginfod/section.exp b/gdb/testsuite/gdb.debuginfod/section.exp
> > > new file mode 100644
> > > index 00000000000..35086d15705
> > > --- /dev/null
> > > +++ b/gdb/testsuite/gdb.debuginfod/section.exp
> > > @@ -0,0 +1,157 @@
> > > +# Copyright 2023-2024 Free Software Foundation, Inc.
> > > +
> > > +# This program is free software; you can redistribute it and/or modify
> > > +# it under the terms of the GNU General Public License as published by
> > > +# the Free Software Foundation; either version 3 of the License, or
> > > +# (at your option) any later version.
> > > +#
> > > +# This program is distributed in the hope that it will be useful,
> > > +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> > > +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > > +# GNU General Public License for more details.
> > > +#
> > > +# You should have received a copy of the GNU General Public License
> > > +# along with this program. If not, see <http://www.gnu.org/licenses/>.
> > > +
> > > +# Test debuginfod functionality
> > > +
> > > +standard_testfile
> > > +
> > > +load_lib debuginfod-support.exp
> > > +
> > > +clean_restart
> > > +require allow_debuginfod_tests
> > > +require allow_debuginfod_section_downloads
> > > +
> > > +# BINFILE calls a function from LIB_SL1.
> > > +set libfile1 "libsection1"
> > > +set libsrc1 $srcdir/$subdir/$libfile1.c
> > > +set lib_sl1 [standard_output_file $libfile1.sl]
> > > +
> > > +# LIB_SL1 calls functions from LIB_SL2.
> > > +set libfile2 "libsection2"
> > > +set libsrc2 $srcdir/$subdir/$libfile2.c
> > > +set lib_sl2 [standard_output_file $libfile2.sl]
> > > +
> > > +# Build LIB_SL2, LIB_SL1 and BINFILE.
> > > +if { [build_executable "build $libfile2" $lib_sl2 $libsrc2 \
> > > + {debug build-id shlib}] != 0 } {
> > > + return -1
> > > +}
> > > +
> > > +if { [build_executable "build $libfile1" $lib_sl1 $libsrc1 \
> > > + [list debug build-id shlib_pthreads shlib=$lib_sl2]] != 0 } {
> > > + return -1
> > > +}
> > > +
> > > +if { [build_executable "build executable" $binfile $srcfile \
> > > + [list debug build-id shlib=$lib_sl1 shlib=$lib_sl2]] != 0 } {
> > > + return -1
> > > +}
> > > +
> > > +# Make sure libsection1 and libsection2 contain .gdb_index.
> > > +if { [ensure_gdb_index $lib_sl1 "" "libsection1"] != 1 } {
> > > + untested "failed to add .gdb_index to $libfile1"
> > > + return -1
> > > +}
> > > +
> > > +if { [ensure_gdb_index $lib_sl2 "" "libsection2"] != 1 } {
> > > + untested "failed to add .gdb_index to $libfile2"
> > > + return -1
> > > +}
> > > +
> > > +# Strip solib debuginfo into separate files.
> > > +if { [gdb_gnu_strip_debug $lib_sl1 ""] != 0} {
> > > + fail "strip $lib_sl1 debuginfo"
> > > + return -1
> > > +}
> > > +
> > > +if { [gdb_gnu_strip_debug $lib_sl2 ""] != 0} {
> > > + fail "strip $lib_sl2 debuginfo"
> > > + return -1
> > > +}
> > > +
> > > +# Move debuginfo files into directory that debuginfod will serve from.
> > > +set debugdir [standard_output_file "debug"]
> > > +set debuginfo_sl1 [standard_output_file $libfile1.sl.debug]
> > > +set debuginfo_sl2 [standard_output_file $libfile2.sl.debug]
> > > +
> > > +file mkdir $debugdir
> > > +file rename -force $debuginfo_sl1 $debugdir
> > > +file rename -force $debuginfo_sl2 $debugdir
> > > +
> > > +# Restart GDB and clear the debuginfod client cache. Then load BINFILE into
> > > +# GDB and start running it. Match output with pattern RES and use TESTNAME
> > > +# as the test name.
> > > +proc_with_prefix clean_restart_with_prompt { binfile testname } {
> > > + global cache
> > > +
> > > + # Delete client cache so debuginfo downloads again.
> > > + file delete -force $cache
> > > + clean_restart
> > > +
> > > + gdb_test_no_output "set debuginfod enabled on" \
> > > + "clean_restart enable $testname"
> > > + gdb_load $binfile
> > > +
> > > + runto_main
> > > +}
> > > +
> > > +# Tests with no debuginfod server running.
> > > +proc_with_prefix no_url { } {
> > > + global binfile libfile1 libfile2
> > > +
> > > + gdb_load $binfile
> > > + if {![runto_main]} {
> > > + return
> > > + }
> > > +
> > > + # Check that no section is downloaded and no debuginfo is found.
> > > + gdb_test "info sharedlibrary" ".*Yes \\(\\*\\).*$libfile1.*" \
> > > + "found no url lib1"
> > > + gdb_test "info sharedlibrary" ".*Yes \\(\\*\\).*$libfile2.*" \
> > > + "found no url lib2"
> > > +}
> > > +
> > > +# Tests with a debuginfod server running.
> > > +proc_with_prefix local_url { } {
> > > + global binfile
> > > + global libsrc1 lib_sl1 libfile1
> > > + global libsrc2 lib_sl2 libfile2
> > > + global debugdir db
> > > +
> > > + set url [start_debuginfod $db $debugdir]
> > > + if { $url == "" } {
> > > + unresolved "failed to start debuginfod server"
> > > + return
> > > + }
> > > +
> > > + # Point GDB to the server.
> > > + setenv DEBUGINFOD_URLS $url
> > > +
> > > + clean_restart_with_prompt $binfile "index"
> > > +
> > > + # Download debuginfo when stepping into a function.
> > > + set res ".*separate debug info for $lib_sl1.*\"In ${libfile1}\\\\n\".*"
> > > + gdb_test "step" $res "step"
> > > +
> > > + clean_restart_with_prompt $binfile "break"
> > > +
> > > + # Download debuginfo when setting a breakpoint.
> > > + set res "Download.*separate debug info for $lib_sl2.*"
> > > + gdb_test "br libsection2_test" $res "break set"
> > > +
> > > + # Hit the breakpoint.
> > > + set res ".*Breakpoint 2, libsection2_test.*\"In ${libfile2}\\\\n\".*"
> > > + gdb_test "c" $res "break continue"
> > > +}
> > > +
> > > +# Create CACHE and DB directories ready for debuginfod to use.
> > > +prepare_for_debuginfod cache db
> > > +
> > > +with_debuginfod_env $cache {
> > > + no_url
> > > + local_url
> > > +}
> > > +
> > > +stop_debuginfod
> > > diff --git a/gdb/testsuite/lib/debuginfod-support.exp b/gdb/testsuite/lib/debuginfod-support.exp
> > > index 0096448567e..4d8bdcdc7cb 100644
> > > --- a/gdb/testsuite/lib/debuginfod-support.exp
> > > +++ b/gdb/testsuite/lib/debuginfod-support.exp
> > > @@ -113,6 +113,8 @@ proc with_debuginfod_env { cache body } {
> > > proc start_debuginfod { db debugdir } {
> > > global debuginfod_spawn_id spawn_id
> > >
> > > + set logfile [standard_output_file "server_log"]
> > > +
> > > # Find an unused port.
> > > set port 7999
> > > set found false
> > > @@ -127,7 +129,8 @@ proc start_debuginfod { db debugdir } {
> > > set old_spawn_id $spawn_id
> > > }
> > >
> > > - spawn debuginfod -vvvv -d $db -p $port -F $debugdir
> > > + spawn sh -c "debuginfod -vvvv -d $db -p $port -F $debugdir 2>&1 \
> > > + | tee $logfile"
> > > set debuginfod_spawn_id $spawn_id
> > >
> > > if { [info exists old_spawn_id] } {
> > > @@ -194,3 +197,23 @@ proc stop_debuginfod { } {
> > > unset debuginfod_spawn_id
> > > }
> > > }
> > > +
> > > +# Return true if gdb is configured to download ELF/DWARF sections from
> > > +# debuginfod servers. Otherwise return false.
> > > +proc allow_debuginfod_section_downloads { } {
> > > + set cmd "maint set debuginfod download-sections on"
> > > + set msg "enable section downloads"
> > > +
> > > + gdb_test_multiple $cmd $msg {
> > > + -re -wrap ".*not compiled into GDB.*" {
> > > + return false
> > > + }
> > > + -re -wrap "^" {
> > > + return true
> > > + }
> > > + -re -wrap "" {
> > > + fail "$gdb_test_name (unexpected output)"
> > > + return false
> > > + }
> > > + }
> > > +}
> > > diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
> > > index 5fb81f6e28d..a3680cb6371 100644
> > > --- a/gdb/testsuite/lib/gdb.exp
> > > +++ b/gdb/testsuite/lib/gdb.exp
> > > @@ -9515,10 +9515,14 @@ proc get_index_type { objfile { testname "" } } {
> > > # STYLE controls which style of index to add, if needed. The empty
> > > # string (the default) means .gdb_index; "-dwarf-5" means .debug_names.
> > >
> > > -proc ensure_gdb_index { binfile {style ""} } {
> > > +proc ensure_gdb_index { binfile {style ""} {testname_prefix ""} } {
> > > set testfile [file tail $binfile]
> > > -
> > > set test "check if index present"
> > > +
> > > + if { $testname_prefix != "" } {
> > > + set test "$testname_prefix $test"
> > > + }
> > > +
> > > set index_type [get_index_type $testfile $test]
> > >
> > > if { $index_type eq "gdb" || $index_type eq "dwarf5" } {
> > > --
> > > 2.43.0
> > >
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING*3][PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading
2024-05-31 20:58 ` [PING*2][PATCH " Aaron Merey
2024-06-07 20:48 ` [PING*3][PATCH " Aaron Merey
@ 2024-06-14 18:24 ` Aaron Merey
1 sibling, 0 replies; 18+ messages in thread
From: Aaron Merey @ 2024-06-14 18:24 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 31, 2024 at 4:58 PM Aaron Merey <amerey@redhat.com> wrote:
>
> Ping
>
> Thanks,
> Aaron
>
> On Fri, May 17, 2024 at 10:10 AM Aaron Merey <amerey@redhat.com> wrote:
> >
> > Ping
> >
> > Thanks,
> > Aaron
> >
> > On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
> > >
> > > v6 fixes a comment for dwarf2_has_separate_index and merges
> > > lookup_gdb_index_debuginfod with lookup_gdb_index.
> > >
> > > At the beginning of a session, gdb may attempt to download debuginfo
> > > for all shared libraries associated with the process or core file
> > > being debugged. This can be a waste of time and storage space when much
> > > of the debuginfo ends up not being used during the session.
> > >
> > > To reduce the gdb's startup latency and to download only the debuginfo
> > > that is really needed, this patch adds on-demand downloading of debuginfo.
> > >
> > > 'set debuginfo enabled on' now causes gdb to attempt to download a .gdb_index
> > > for each shared library instead of its full debuginfo. Each corresponding
> > > separate debuginfo will be deferred until gdb needs to expand symtabs
> > > associated with the debuginfo's index.
> > >
> > > Because these indices are significantly smaller than their corresponding
> > > debuginfo, this generally reduces the total amount of data gdb downloads.
> > > Reductions of 80%-95% have been observed when debugging large GUI programs.
> > >
> > > (gdb) set debuginfod enabled on
> > > (gdb) start
> > > Downloading section .gdb_index for /lib64/libcurl.so.4
> > > [...]
> > > 1826 client->server_mhandle = curl_multi_init ();
> > > (gdb) step
> > > Downloading separate debug info for /lib64/libcurl.so.4
> > > Downloading separate debug info for [libcurl dwz]
> > > Downloading source file /usr/src/debug/curl-7.85.0-6.fc37.x86_64/build-full/lib/../../lib/multi.c
> > > curl_multi_init () at ../../lib/multi.c:457
> > > 457 {
> > > (gdb)
> > >
> > > Some of the key functions below include dwarf2_has_separate_index which
> > > downloads the separate .gdb_index. If successful, the shared library
> > > objfile owns the index until the separate debug objfile is downloaded
> > > or confirmed to not be available.
> > >
> > > read_full_dwarf_from_debuginfod downloads the full debuginfo and
> > > initializes the separate debug objfile. It is called by functions
> > > such as dwarf2_gdb_index::expand_symtabs_matching and
> > > dwarf2_base_index_functions::find_pc_sect_compunit_symtab when symtab
> > > expansion is required.
> > > ---
> > > gdb/dwarf2/frame.c | 13 ++
> > > gdb/dwarf2/frame.h | 4 +
> > > gdb/dwarf2/index-cache.c | 22 ++-
> > > gdb/dwarf2/index-cache.h | 11 ++
> > > gdb/dwarf2/public.h | 5 +
> > > gdb/dwarf2/read-gdb-index.c | 123 +++++++++++++---
> > > gdb/dwarf2/read.c | 149 ++++++++++++++++++-
> > > gdb/dwarf2/read.h | 10 ++
> > > gdb/dwarf2/section.c | 3 +-
> > > gdb/elfread.c | 3 +-
> > > gdb/frame.c | 27 +++-
> > > gdb/objfile-flags.h | 4 +
> > > gdb/objfiles.h | 20 +++
> > > gdb/quick-symbol.h | 4 +
> > > gdb/symfile-debug.c | 11 ++
> > > gdb/symfile.c | 12 +-
> > > gdb/symtab.c | 18 ++-
> > > gdb/testsuite/gdb.debuginfod/libsection1.c | 43 ++++++
> > > gdb/testsuite/gdb.debuginfod/libsection2.c | 37 +++++
> > > gdb/testsuite/gdb.debuginfod/section.c | 29 ++++
> > > gdb/testsuite/gdb.debuginfod/section.exp | 157 +++++++++++++++++++++
> > > gdb/testsuite/lib/debuginfod-support.exp | 25 +++-
> > > gdb/testsuite/lib/gdb.exp | 8 +-
> > > 23 files changed, 702 insertions(+), 36 deletions(-)
> > > create mode 100644 gdb/testsuite/gdb.debuginfod/libsection1.c
> > > create mode 100644 gdb/testsuite/gdb.debuginfod/libsection2.c
> > > create mode 100644 gdb/testsuite/gdb.debuginfod/section.c
> > > create mode 100644 gdb/testsuite/gdb.debuginfod/section.exp
> > >
> > > diff --git a/gdb/dwarf2/frame.c b/gdb/dwarf2/frame.c
> > > index 9ebf3ac4ac4..074417765f4 100644
> > > --- a/gdb/dwarf2/frame.c
> > > +++ b/gdb/dwarf2/frame.c
> > > @@ -1621,6 +1621,19 @@ set_comp_unit (struct objfile *objfile, struct comp_unit *unit)
> > > return dwarf2_frame_bfd_data.set (abfd, unit);
> > > }
> > >
> > > +/* See frame.h. */
> > > +
> > > +void
> > > +dwarf2_clear_frame_data (struct objfile *objfile)
> > > +{
> > > + bfd *abfd = objfile->obfd.get ();
> > > +
> > > + if (gdb_bfd_requires_relocations (abfd))
> > > + dwarf2_frame_objfile_data.clear (objfile);
> > > + else
> > > + dwarf2_frame_bfd_data.clear (abfd);
> > > +}
> > > +
> > > /* Find the FDE for *PC. Return a pointer to the FDE, and store the
> > > initial location associated with it into *PC. */
> > >
> > > diff --git a/gdb/dwarf2/frame.h b/gdb/dwarf2/frame.h
> > > index 2167310fbdf..423a9355b8a 100644
> > > --- a/gdb/dwarf2/frame.h
> > > +++ b/gdb/dwarf2/frame.h
> > > @@ -237,6 +237,10 @@ void dwarf2_append_unwinders (struct gdbarch *gdbarch);
> > > extern const struct frame_base *
> > > dwarf2_frame_base_sniffer (const frame_info_ptr &this_frame);
> > >
> > > +/* Delete OBJFILEs comp_unit. */
> > > +
> > > +extern void dwarf2_clear_frame_data (struct objfile * objfile);
> > > +
> > > /* Compute the DWARF CFA for a frame. */
> > >
> > > CORE_ADDR dwarf2_frame_cfa (const frame_info_ptr &this_frame);
> > > diff --git a/gdb/dwarf2/index-cache.c b/gdb/dwarf2/index-cache.c
> > > index 1720e925ee6..7096a3ea078 100644
> > > --- a/gdb/dwarf2/index-cache.c
> > > +++ b/gdb/dwarf2/index-cache.c
> > > @@ -219,14 +219,22 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> > > /* Compute where we would expect a gdb index file for this build id to be. */
> > > std::string filename = make_index_filename (build_id, INDEX4_SUFFIX);
> > >
> > > + return lookup_gdb_index (filename.c_str (), resource);
> > > +}
> > > +
> > > +/* See index-cache.h. */
> > > +
> > > +gdb::array_view<const gdb_byte>
> > > +index_cache::lookup_gdb_index (const char *filename,
> > > + std::unique_ptr<index_cache_resource> *resource)
> > > +{
> > > try
> > > {
> > > - index_cache_debug ("trying to read %s",
> > > - filename.c_str ());
> > > + index_cache_debug ("trying to read %s", filename);
> > >
> > > /* Try to map that file. */
> > > index_cache_resource_mmap *mmap_resource
> > > - = new index_cache_resource_mmap (filename.c_str ());
> > > + = new index_cache_resource_mmap (filename);
> > >
> > > /* Yay, it worked! Hand the resource to the caller. */
> > > resource->reset (mmap_resource);
> > > @@ -238,7 +246,7 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> > > catch (const gdb_exception_error &except)
> > > {
> > > index_cache_debug ("couldn't read %s: %s",
> > > - filename.c_str (), except.what ());
> > > + filename, except.what ());
> > > }
> > >
> > > return {};
> > > @@ -255,6 +263,12 @@ index_cache::lookup_gdb_index (const bfd_build_id *build_id,
> > > return {};
> > > }
> > >
> > > +gdb::array_view<const gdb_byte>
> > > +index_cache::lookup_gdb_index_debuginfod (const char *index_path,
> > > + std::unique_ptr<index_cache_resource> *resource)
> > > +{
> > > + return {};
> > > +}
> > > #endif
> > >
> > > /* See dwarf-index-cache.h. */
> > > diff --git a/gdb/dwarf2/index-cache.h b/gdb/dwarf2/index-cache.h
> > > index 95f217ed961..f42e6918ce2 100644
> > > --- a/gdb/dwarf2/index-cache.h
> > > +++ b/gdb/dwarf2/index-cache.h
> > > @@ -93,6 +93,17 @@ class index_cache
> > > lookup_gdb_index (const bfd_build_id *build_id,
> > > std::unique_ptr<index_cache_resource> *resource);
> > >
> > > + /* Look for an index file located at INDEX_PATH. If found, return the
> > > + contents as an array_view and store the underlying resources (allocated
> > > + memory, mapped file, etc) in RESOURCE. The returned array_view is valid
> > > + as long as RESOURCE is not destroyed.
> > > +
> > > + If no matching index file is found, return an empty array view. This
> > > + function does not exit early if the index cache has not been enabled. */
> > > + gdb::array_view<const gdb_byte>
> > > + lookup_gdb_index (const char *filename,
> > > + std::unique_ptr<index_cache_resource> *resource);
> > > +
> > > /* Return the number of cache hits. */
> > > unsigned int n_hits () const
> > > { return m_n_hits; }
> > > diff --git a/gdb/dwarf2/public.h b/gdb/dwarf2/public.h
> > > index bc419ff208f..6523854386c 100644
> > > --- a/gdb/dwarf2/public.h
> > > +++ b/gdb/dwarf2/public.h
> > > @@ -44,4 +44,9 @@ extern bool dwarf2_initialize_objfile
> > >
> > > extern void dwarf2_build_frame_info (struct objfile *);
> > >
> > > +/* Query debuginfod for the .gdb_index associated with OBJFILE.
> > > + Used to defer separate debuginfo downloading until necessary. */
> > > +
> > > +extern bool dwarf2_has_separate_index (struct objfile *);
> > > +
> > > #endif /* DWARF2_PUBLIC_H */
> > > diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
> > > index e743ce13978..97b8137aabf 100644
> > > --- a/gdb/dwarf2/read-gdb-index.c
> > > +++ b/gdb/dwarf2/read-gdb-index.c
> > > @@ -139,6 +139,8 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > > gdb.dwarf2/gdb-index.exp testcase. */
> > > void dump (struct objfile *objfile) override;
> > >
> > > + /* Calls do_expand_symtabs_matching and triggers debuginfo downloading
> > > + if necessary. */
> > > bool expand_symtabs_matching
> > > (struct objfile *objfile,
> > > gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > > @@ -147,8 +149,59 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > > gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> > > block_search_flags search_flags,
> > > domain_search_flags domain) override;
> > > +
> > > + /* Expand symtabs matching a given symbol or file. */
> > > + bool do_expand_symtabs_matching
> > > + (struct objfile *objfile,
> > > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > > + const lookup_name_info *lookup_name,
> > > + gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
> > > + gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> > > + block_search_flags search_flags,
> > > + domain_search_flags domain);
> > > +
> > > + /* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
> > > + debuginfo if necessary. */
> > > + void expand_all_symtabs (struct objfile *objfile) override;
> > > +
> > > + /* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
> > > + debuginfo if necessary. */
> > > + struct symtab *find_last_source_symtab (struct objfile *objfile) override;
> > > };
> > >
> > > +void
> > > +dwarf2_gdb_index::expand_all_symtabs (struct objfile *objfile)
> > > +{
> > > + try
> > > + {
> > > + dwarf2_base_index_functions::expand_all_symtabs (objfile);
> > > + }
> > > + catch (const gdb_exception &e)
> > > + {
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + exception_print (gdb_stderr, e);
> > > + else
> > > + read_full_dwarf_from_debuginfod (objfile, this);
> > > + }
> > > +}
> > > +
> > > +struct symtab *
> > > +dwarf2_gdb_index::find_last_source_symtab (struct objfile *objfile)
> > > +{
> > > + try
> > > + {
> > > + return dwarf2_base_index_functions::find_last_source_symtab (objfile);
> > > + }
> > > + catch (const gdb_exception &e)
> > > + {
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + exception_print (gdb_stderr, e);
> > > + else
> > > + read_full_dwarf_from_debuginfod (objfile, this);
> > > + return nullptr;
> > > + }
> > > +}
> > > +
> > > /* This dumps minimal information about the index.
> > > It is called via "mt print objfiles".
> > > One use is to verify .gdb_index has been loaded by the
> > > @@ -264,7 +317,7 @@ dw2_expand_marked_cus
> > > }
> > >
> > > bool
> > > -dwarf2_gdb_index::expand_symtabs_matching
> > > +dwarf2_gdb_index::do_expand_symtabs_matching
> > > (struct objfile *objfile,
> > > gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > > const lookup_name_info *lookup_name,
> > > @@ -312,6 +365,38 @@ dwarf2_gdb_index::expand_symtabs_matching
> > > return result;
> > > }
> > >
> > > +bool
> > > +dwarf2_gdb_index::expand_symtabs_matching
> > > + (struct objfile *objfile,
> > > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> > > + const lookup_name_info *lookup_name,
> > > + gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
> > > + gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> > > + block_search_flags search_flags,
> > > + domain_search_flags domain)
> > > +{
> > > + if ((objfile->flags & OBJF_READNEVER) != 0)
> > > + return false;
> > > +
> > > + try
> > > + {
> > > + return do_expand_symtabs_matching (objfile, file_matcher, lookup_name,
> > > + symbol_matcher, expansion_notify,
> > > + search_flags, domain);
> > > + }
> > > + catch (const gdb_exception &e)
> > > + {
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + {
> > > + exception_print (gdb_stderr, e);
> > > + return false;
> > > + }
> > > +
> > > + read_full_dwarf_from_debuginfod (objfile, this);
> > > + return true;
> > > + }
> > > +}
> > > +
> > > quick_symbol_functions_up
> > > mapped_gdb_index::make_quick_functions () const
> > > {
> > > @@ -647,28 +732,32 @@ dwarf2_read_gdb_index
> > >
> > > /* If there is a .dwz file, read it so we can get its CU list as
> > > well. */
> > > - dwz = dwarf2_get_dwz_file (per_bfd);
> > > - if (dwz != NULL)
> > > + if (get_gdb_index_contents_dwz != nullptr)
> > > {
> > > mapped_gdb_index dwz_map;
> > > const gdb_byte *dwz_types_ignore;
> > > offset_type dwz_types_elements_ignore;
> > > + dwz = dwarf2_get_dwz_file (per_bfd);
> > >
> > > - gdb::array_view<const gdb_byte> dwz_index_content
> > > - = get_gdb_index_contents_dwz (objfile, dwz);
> > > -
> > > - if (dwz_index_content.empty ())
> > > - return 0;
> > > -
> > > - if (!read_gdb_index_from_buffer (bfd_get_filename (dwz->dwz_bfd.get ()),
> > > - 1, dwz_index_content, &dwz_map,
> > > - &dwz_list, &dwz_list_elements,
> > > - &dwz_types_ignore,
> > > - &dwz_types_elements_ignore))
> > > + if (dwz != nullptr)
> > > {
> > > - warning (_("could not read '.gdb_index' section from %s; skipping"),
> > > - bfd_get_filename (dwz->dwz_bfd.get ()));
> > > - return 0;
> > > + gdb::array_view<const gdb_byte> dwz_index_content
> > > + = get_gdb_index_contents_dwz (objfile, dwz);
> > > +
> > > + if (dwz_index_content.empty ())
> > > + return 0;
> > > +
> > > + if (!read_gdb_index_from_buffer (bfd_get_filename
> > > + (dwz->dwz_bfd.get ()),
> > > + 1, dwz_index_content, &dwz_map,
> > > + &dwz_list, &dwz_list_elements,
> > > + &dwz_types_ignore,
> > > + &dwz_types_elements_ignore))
> > > + {
> > > + warning (_("could not read '.gdb_index' section from %s; skipping"),
> > > + bfd_get_filename (dwz->dwz_bfd.get ()));
> > > + return 0;
> > > + }
> > > }
> > > }
> > >
> > > diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
> > > index 00e8608f9ce..4461c76347e 100644
> > > --- a/gdb/dwarf2/read.c
> > > +++ b/gdb/dwarf2/read.c
> > > @@ -33,6 +33,7 @@
> > > #include "dwarf2/attribute.h"
> > > #include "dwarf2/comp-unit-head.h"
> > > #include "dwarf2/cu.h"
> > > +#include "dwarf2/frame.h"
> > > #include "dwarf2/index-cache.h"
> > > #include "dwarf2/index-common.h"
> > > #include "dwarf2/leb.h"
> > > @@ -96,6 +97,8 @@
> > > #include "gdbsupport/thread-pool.h"
> > > #include "run-on-main-thread.h"
> > > #include "dwarf2/parent-map.h"
> > > +#include "inferior.h"
> > > +#include "debuginfod-support.h"
> > >
> > > /* When == 1, print basic high level tracing messages.
> > > When > 1, be more verbose.
> > > @@ -3020,7 +3023,7 @@ dwarf2_base_index_functions::find_per_cu (dwarf2_per_bfd *per_bfd,
> > > }
> > >
> > > struct compunit_symtab *
> > > -dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> > > +dwarf2_base_index_functions::do_find_pc_sect_compunit_symtab
> > > (struct objfile *objfile,
> > > struct bound_minimal_symbol msymbol,
> > > CORE_ADDR pc,
> > > @@ -3051,6 +3054,32 @@ dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> > > return result;
> > > }
> > >
> > > +struct compunit_symtab *
> > > +dwarf2_base_index_functions::find_pc_sect_compunit_symtab
> > > + (struct objfile *objfile,
> > > + struct bound_minimal_symbol msymbol,
> > > + CORE_ADDR pc,
> > > + struct obj_section *section,
> > > + int warn_if_readin)
> > > +{
> > > + if ((objfile->flags & OBJF_READNEVER) != 0)
> > > + return nullptr;
> > > +
> > > + try
> > > + {
> > > + return do_find_pc_sect_compunit_symtab (objfile, msymbol, pc,
> > > + section, warn_if_readin);
> > > + }
> > > + catch (const gdb_exception &e)
> > > + {
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + exception_print (gdb_stderr, e);
> > > + else
> > > + read_full_dwarf_from_debuginfod (objfile, this);
> > > + return nullptr;
> > > + }
> > > +}
> > > +
> > > void
> > > dwarf2_base_index_functions::map_symbol_filenames
> > > (struct objfile *objfile,
> > > @@ -3207,6 +3236,30 @@ get_gdb_index_contents_from_cache_dwz (objfile *obj, dwz_file *dwz)
> > > return global_index_cache.lookup_gdb_index (build_id, &dwz->index_cache_res);
> > > }
> > >
> > > +/* Query debuginfod for the .gdb_index matching OBJFILE's build-id. Return the
> > > + contents if successful. */
> > > +
> > > +static gdb::array_view<const gdb_byte>
> > > +get_gdb_index_contents_from_debuginfod (objfile *objfile, dwarf2_per_bfd *per_bfd)
> > > +{
> > > + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > > + if (build_id == nullptr)
> > > + return {};
> > > +
> > > + gdb::unique_xmalloc_ptr<char> index_path;
> > > + scoped_fd fd = debuginfod_section_query (build_id->data, build_id->size,
> > > + bfd_get_filename
> > > + (objfile->obfd.get ()),
> > > + ".gdb_index",
> > > + &index_path);
> > > + if (fd.get () < 0)
> > > + return {};
> > > +
> > > + return global_index_cache.lookup_gdb_index
> > > + (index_path.get (), &per_bfd->index_cache_res);
> > > +}
> > > +
> > > +
> > > static void start_debug_info_reader (dwarf2_per_objfile *);
> > >
> > > /* See dwarf2/public.h. */
> > > @@ -3216,11 +3269,13 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> > > const struct dwarf2_debug_sections *names,
> > > bool can_copy)
> > > {
> > > - if (!dwarf2_has_info (objfile, names, can_copy))
> > > + if (!dwarf2_has_info (objfile, names, can_copy)
> > > + && (objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > return false;
> > >
> > > dwarf2_per_objfile *per_objfile = get_dwarf2_per_objfile (objfile);
> > > dwarf2_per_bfd *per_bfd = per_objfile->per_bfd;
> > > + bool separate_index = false;
> > >
> > > dwarf_read_debug_printf ("called");
> > >
> > > @@ -3256,6 +3311,15 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> > > dwarf_read_debug_printf ("found gdb index from cache");
> > > global_index_cache.hit ();
> > > }
> > > + /* Try to read just a separately downloaded gdb index. */
> > > + else if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0
> > > + && dwarf2_read_gdb_index (per_objfile,
> > > + get_gdb_index_contents_from_debuginfod,
> > > + nullptr))
> > > + {
> > > + dwarf_read_debug_printf ("found .gdb_index from debuginfod");
> > > + separate_index = true;
> > > + }
> > > else
> > > {
> > > global_index_cache.miss ();
> > > @@ -3267,8 +3331,89 @@ dwarf2_initialize_objfile (struct objfile *objfile,
> > > if (dwarf_synchronous)
> > > per_bfd->index_table->wait_completely ();
> > > objfile->qf.push_front (per_bfd->index_table->make_quick_functions ());
> > > +
> > > + if (separate_index)
> > > + objfile->qf.begin ()->get ()->from_separate_index = true;
> > > + }
> > > +
> > > + return true;
> > > +}
> > > +
> > > +/* See read.h. */
> > > +
> > > +void
> > > +read_full_dwarf_from_debuginfod (struct objfile *objfile,
> > > + dwarf2_base_index_functions *fncs)
> > > +{
> > > + gdb_assert ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> > > +
> > > + SCOPE_EXIT { objfile->remove_deferred_status (); };
> > > +
> > > + const struct bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > > + if (build_id == nullptr)
> > > + return;
> > > +
> > > + const char *filename = bfd_get_filename (objfile->obfd.get ());
> > > + gdb::unique_xmalloc_ptr<char> symfile_path;
> > > + scoped_fd fd;
> > > +
> > > + fd = debuginfod_debuginfo_query (build_id->data, build_id->size,
> > > + filename, &symfile_path);
> > > + if (fd.get () < 0)
> > > + return;
> > > +
> > > + /* Separate debuginfo successfully retrieved from server. */
> > > + gdb_bfd_ref_ptr debug_bfd = symfile_bfd_open (symfile_path.get ());
> > > + if (debug_bfd == nullptr
> > > + || !build_id_verify (debug_bfd.get (), build_id->size, build_id->data))
> > > + {
> > > + warning (_("File \"%s\" from debuginfod cannot be opened as bfd"),
> > > + filename);
> > > + return;
> > > }
> > >
> > > + /* Clear frame data so it can be recalculated using DWARF. */
> > > + dwarf2_clear_frame_data (objfile);
> > > +
> > > + /* This may also trigger a dwz download. */
> > > + symbol_file_add_separate (debug_bfd, symfile_path.get (),
> > > + current_inferior ()->symfile_flags, objfile);
> > > +}
> > > +
> > > +/* See public.h. */
> > > +
> > > +bool
> > > +dwarf2_has_separate_index (struct objfile *objfile)
> > > +{
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0)
> > > + return true;
> > > + if ((objfile->flags & OBJF_MAINLINE) != 0)
> > > + return false;
> > > + if (!IS_DIR_SEPARATOR (*objfile_filename (objfile)))
> > > + return false;
> > > +
> > > + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > > +
> > > + if (build_id == nullptr)
> > > + return false;
> > > +
> > > + gdb::unique_xmalloc_ptr<char> index_path;
> > > + scoped_fd fd = debuginfod_section_query (build_id->data,
> > > + build_id->size,
> > > + bfd_get_filename
> > > + (objfile->obfd.get ()),
> > > + ".gdb_index",
> > > + &index_path);
> > > +
> > > + if (fd.get () < 0)
> > > + return false;
> > > +
> > > + /* We found a separate .gdb_index file so a separate debuginfo file
> > > + should exist, but we don't want to download it until necessary.
> > > + Attach the index to this objfile and defer the debuginfo download
> > > + until gdb needs to expand symtabs referenced by the index. */
> > > + objfile->flags |= OBJF_DOWNLOAD_DEFERRED;
> > > + dwarf2_initialize_objfile (objfile);
> > > return true;
> > > }
> > >
> > > diff --git a/gdb/dwarf2/read.h b/gdb/dwarf2/read.h
> > > index 73def88c4c0..099d3d65c27 100644
> > > --- a/gdb/dwarf2/read.h
> > > +++ b/gdb/dwarf2/read.h
> > > @@ -860,6 +860,10 @@ struct dwarf2_base_index_functions : public quick_symbol_functions
> > > CORE_ADDR pc, struct obj_section *section, int warn_if_readin)
> > > override;
> > >
> > > + struct compunit_symtab *do_find_pc_sect_compunit_symtab
> > > + (struct objfile *objfile, struct bound_minimal_symbol msymbol,
> > > + CORE_ADDR pc, struct obj_section *section, int warn_if_readin);
> > > +
> > > struct compunit_symtab *find_compunit_symtab_by_address
> > > (struct objfile *objfile, CORE_ADDR address) override
> > > {
> > > @@ -932,4 +936,10 @@ extern void create_all_units (dwarf2_per_objfile *per_objfile);
> > >
> > > extern htab_up create_quick_file_names_table (unsigned int nr_initial_entries);
> > >
> > > +/* If OBJFILE contains information from a separately downloaded .gdb_index,
> > > + attempt to download the full debuginfo. */
> > > +
> > > +extern void read_full_dwarf_from_debuginfod (struct objfile *,
> > > + dwarf2_base_index_functions *);
> > > +
> > > #endif /* DWARF2READ_H */
> > > diff --git a/gdb/dwarf2/section.c b/gdb/dwarf2/section.c
> > > index 6aab0b9aee9..481f37351f4 100644
> > > --- a/gdb/dwarf2/section.c
> > > +++ b/gdb/dwarf2/section.c
> > > @@ -53,7 +53,8 @@ dwarf2_section_info::get_bfd_owner () const
> > > section = get_containing_section ();
> > > gdb_assert (!section->is_virtual);
> > > }
> > > - gdb_assert (section->s.section != nullptr);
> > > + if (section->s.section == nullptr)
> > > + error (_("Can't find owner of DWARF section."));
> > > return section->s.section->owner;
> > > }
> > >
> > > diff --git a/gdb/elfread.c b/gdb/elfread.c
> > > index 7a6a8cadced..0f59bb1a58d 100644
> > > --- a/gdb/elfread.c
> > > +++ b/gdb/elfread.c
> > > @@ -1217,7 +1217,8 @@ elf_symfile_read_dwarf2 (struct objfile *objfile,
> > > && objfile->separate_debug_objfile_backlink == NULL)
> > > {
> > > if (objfile->find_and_add_separate_symbol_file (symfile_flags))
> > > - gdb_assert (objfile->separate_debug_objfile != nullptr);
> > > + gdb_assert (objfile->separate_debug_objfile != nullptr
> > > + || (objfile->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> > > else
> > > has_dwarf2 = false;
> > > }
> > > diff --git a/gdb/frame.c b/gdb/frame.c
> > > index 43efac07d95..1718ab5086c 100644
> > > --- a/gdb/frame.c
> > > +++ b/gdb/frame.c
> > > @@ -1778,14 +1778,10 @@ restore_selected_frame (frame_id frame_id, int frame_level)
> > > selected_frame = nullptr;
> > > }
> > >
> > > -/* Lookup the frame_info object for the selected frame FRAME_ID /
> > > - FRAME_LEVEL and cache the result.
> > > -
> > > - If FRAME_LEVEL > 0 and the originally selected frame isn't found,
> > > - warn and select the innermost (current) frame. */
> > > +/* See lookup_selected_frame. */
> > >
> > > static void
> > > -lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> > > +lookup_selected_frame_1 (struct frame_id &a_frame_id, int frame_level)
> > > {
> > > frame_info_ptr frame = NULL;
> > > int count;
> > > @@ -1857,6 +1853,25 @@ lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> > > }
> > > }
> > >
> > > +/* Lookup the frame_info object for the selected frame FRAME_ID /
> > > + FRAME_LEVEL and cache the result.
> > > +
> > > + If FRAME_LEVEL > 0 and the originally selected frame isn't found,
> > > + warn and select the innermost (current) frame. */
> > > +
> > > +static void
> > > +lookup_selected_frame (struct frame_id a_frame_id, int frame_level)
> > > +{
> > > + lookup_selected_frame_1 (selected_frame_id, selected_frame_level);
> > > +
> > > + /* It is possible for lookup_selected_frame_1 to cause a new objfile
> > > + to be loaded. However some objfile observers may choose to clear
> > > + selected_frame when an objfile is loaded. Work around this by
> > > + calling lookup_selected_frame_1 again if the first call failed. */
> > > + if (selected_frame == nullptr)
> > > + lookup_selected_frame_1 (selected_frame_id, selected_frame_level);
> > > +}
> > > +
> > > bool
> > > has_stack_frames ()
> > > {
> > > diff --git a/gdb/objfile-flags.h b/gdb/objfile-flags.h
> > > index 43d888e2a5a..08e4a75a9f2 100644
> > > --- a/gdb/objfile-flags.h
> > > +++ b/gdb/objfile-flags.h
> > > @@ -56,6 +56,10 @@ enum objfile_flag : unsigned
> > > /* User requested that we do not read this objfile's symbolic
> > > information. */
> > > OBJF_READNEVER = 1 << 6,
> > > +
> > > + /* A separate .gdb_index has been downloaded for this objfile.
> > > + Debuginfo for this objfile can be downloaded when required. */
> > > + OBJF_DOWNLOAD_DEFERRED = 1 << 7,
> > > };
> > >
> > > DEF_ENUM_FLAGS_TYPE (enum objfile_flag, objfile_flags);
> > > diff --git a/gdb/objfiles.h b/gdb/objfiles.h
> > > index 1c8f79a2826..8c8f27f87f1 100644
> > > --- a/gdb/objfiles.h
> > > +++ b/gdb/objfiles.h
> > > @@ -628,6 +628,26 @@ struct objfile
> > > domain_search_flags domain,
> > > bool *symbol_found_p);
> > >
> > > + /* Indicate that the acquisition of this objfile's separate debug objfile
> > > + is no longer deferred. Used when the debug objfile has been acquired
> > > + or could not be found. */
> > > + void remove_deferred_status ()
> > > + {
> > > + flags &= ~OBJF_DOWNLOAD_DEFERRED;
> > > +
> > > + /* Remove quick_symbol_functions derived from a separately downloaded
> > > + index. If available the separate debug objfile's index will be used
> > > + instead, since that objfile actually contains the symbols and CUs
> > > + referenced in the index.
> > > +
> > > + No more than one element of qf should have from_separate_index set
> > > + to true. */
> > > + qf.remove_if ([&] (const quick_symbol_functions_up &qf_up)
> > > + {
> > > + return qf_up->from_separate_index;
> > > + });
> > > + }
> > > +
> > > /* Return the relocation offset applied to SECTION. */
> > > CORE_ADDR section_offset (bfd_section *section) const
> > > {
> > > diff --git a/gdb/quick-symbol.h b/gdb/quick-symbol.h
> > > index 676c3ede373..8b91b7765d5 100644
> > > --- a/gdb/quick-symbol.h
> > > +++ b/gdb/quick-symbol.h
> > > @@ -191,6 +191,10 @@ struct quick_symbol_functions
> > > virtual void compute_main_name (struct objfile *objfile)
> > > {
> > > }
> > > +
> > > + /* True if this quick_symbol_functions is derived from a separately
> > > + downloaded index. */
> > > + bool from_separate_index = false;
> > > };
> > >
> > > typedef std::unique_ptr<quick_symbol_functions> quick_symbol_functions_up;
> > > diff --git a/gdb/symfile-debug.c b/gdb/symfile-debug.c
> > > index 330a520dca6..8099d464081 100644
> > > --- a/gdb/symfile-debug.c
> > > +++ b/gdb/symfile-debug.c
> > > @@ -36,6 +36,7 @@
> > > #include "cli/cli-style.h"
> > > #include "build-id.h"
> > > #include "debuginfod-support.h"
> > > +#include "dwarf2/public.h"
> > >
> > > /* We need to save a pointer to the real symbol functions.
> > > Plus, the debug versions are malloc'd because we have to NULL out the
> > > @@ -610,6 +611,16 @@ objfile::find_and_add_separate_symbol_file (symfile_add_flags symfile_flags)
> > > = simple_find_and_open_separate_symbol_file
> > > (this, find_separate_debug_file_by_debuglink, &warnings);
> > >
> > > + /* Attempt to download only an index from the separate debug info.
> > > + As with debuginfod_find_and_open_separate_symbol_file, only attempt
> > > + this once. */
> > > + if (debug_bfd == nullptr && attempt == 0
> > > + && dwarf2_has_separate_index (this))
> > > + {
> > > + has_dwarf2 = true;
> > > + break;
> > > + }
> > > +
> > > /* Only try debuginfod on the first attempt. Sure, we could imagine
> > > an extension that somehow adds the required debug info to the
> > > debuginfod server but, at least for now, we don't support this
> > > diff --git a/gdb/symfile.c b/gdb/symfile.c
> > > index 7ebb07d5394..35c4a3b78e9 100644
> > > --- a/gdb/symfile.c
> > > +++ b/gdb/symfile.c
> > > @@ -989,6 +989,10 @@ syms_from_objfile (struct objfile *objfile,
> > > static void
> > > finish_new_objfile (struct objfile *objfile, symfile_add_flags add_flags)
> > > {
> > > + struct objfile *parent = objfile->separate_debug_objfile_backlink;
> > > + bool was_deferred
> > > + = (parent != nullptr) && ((parent->flags & OBJF_DOWNLOAD_DEFERRED) != 0);
> > > +
> > > /* If this is the main symbol file we have to clean up all users of the
> > > old main symbol file. Otherwise it is sufficient to fixup all the
> > > breakpoints that may have been redefined by this symbol file. */
> > > @@ -999,7 +1003,7 @@ finish_new_objfile (struct objfile *objfile, symfile_add_flags add_flags)
> > >
> > > clear_symtab_users (add_flags);
> > > }
> > > - else if ((add_flags & SYMFILE_DEFER_BP_RESET) == 0)
> > > + else if ((add_flags & SYMFILE_DEFER_BP_RESET) == 0 && !was_deferred)
> > > {
> > > breakpoint_re_set ();
> > > }
> > > @@ -1120,6 +1124,12 @@ symbol_file_add_with_addrs (const gdb_bfd_ref_ptr &abfd, const char *name,
> > > if (objfile->sf != nullptr)
> > > finish_new_objfile (objfile, add_flags);
> > >
> > > + /* Remove deferred status now in case any observers trigger symtab
> > > + expansion. Otherwise gdb might try to read parent for psymbols
> > > + when it should read the separate debug objfile instead. */
> > > + if (parent != nullptr && ((parent->flags & OBJF_DOWNLOAD_DEFERRED) != 0))
> > > + parent->remove_deferred_status ();
> > > +
> > > gdb::observers::new_objfile.notify (objfile);
> > >
> > > return objfile;
> > > diff --git a/gdb/symtab.c b/gdb/symtab.c
> > > index 58648a8779d..26f1b1f631e 100644
> > > --- a/gdb/symtab.c
> > > +++ b/gdb/symtab.c
> > > @@ -2956,14 +2956,30 @@ find_pc_sect_compunit_symtab (CORE_ADDR pc, struct obj_section *section)
> > > if (best_cust != NULL)
> > > return best_cust;
> > >
> > > + int warn_if_readin = 1;
> > > +
> > > /* Not found in symtabs, search the "quick" symtabs (e.g. psymtabs). */
> > >
> > > for (objfile *objf : current_program_space->objfiles ())
> > > {
> > > + bool was_deferred = (objf->flags & OBJF_DOWNLOAD_DEFERRED) != 0;
> > > +
> > > struct compunit_symtab *result
> > > - = objf->find_pc_sect_compunit_symtab (msymbol, pc, section, 1);
> > > + = objf->find_pc_sect_compunit_symtab (msymbol, pc, section,
> > > + warn_if_readin);
> > > +
> > > if (result != NULL)
> > > return result;
> > > +
> > > + /* If OBJF's separate debug info was just acquired, disable
> > > + warn_if_readin for the next iteration of this loop. This prevents
> > > + a spurious warning in case an observer already triggered expansion
> > > + of the separate debug objfile's symtabs. */
> > > + if (was_deferred && objf->separate_debug_objfile != nullptr
> > > + && (objf->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + warn_if_readin = 0;
> > > + else if (warn_if_readin == 0)
> > > + warn_if_readin = 1;
> > > }
> > >
> > > return NULL;
> > > diff --git a/gdb/testsuite/gdb.debuginfod/libsection1.c b/gdb/testsuite/gdb.debuginfod/libsection1.c
> > > new file mode 100644
> > > index 00000000000..e4f05db0a79
> > > --- /dev/null
> > > +++ b/gdb/testsuite/gdb.debuginfod/libsection1.c
> > > @@ -0,0 +1,43 @@
> > > +/* This testcase is part of GDB, the GNU debugger.
> > > +
> > > + Copyright 2023-2024 Free Software Foundation, Inc.
> > > +
> > > + This program is free software; you can redistribute it and/or modify
> > > + it under the terms of the GNU General Public License as published by
> > > + the Free Software Foundation; either version 3 of the License, or
> > > + (at your option) any later version.
> > > +
> > > + This program is distributed in the hope that it will be useful,
> > > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > > + GNU General Public License for more details.
> > > +
> > > + You should have received a copy of the GNU General Public License
> > > + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> > > +
> > > +#include <stdio.h>
> > > +#include <pthread.h>
> > > +#include <unistd.h>
> > > +
> > > +extern void libsection2_test ();
> > > +extern void *libsection2_thread_test (void *);
> > > +
> > > +static volatile int flag = 0;
> > > +
> > > +void
> > > +libsection1_test ()
> > > +{
> > > + pthread_t thr;
> > > +
> > > + printf ("In libsection1\n");
> > > + libsection2_test ();
> > > +
> > > + pthread_create (&thr, NULL, libsection2_thread_test, (void *) &flag);
> > > +
> > > + /* Give the new thread a chance to actually enter libsection2_thread_test. */
> > > + while (!flag)
> > > + ;
> > > +
> > > + printf ("Cancelling thread\n");
> > > + pthread_cancel (thr);
> > > +}
> > > diff --git a/gdb/testsuite/gdb.debuginfod/libsection2.c b/gdb/testsuite/gdb.debuginfod/libsection2.c
> > > new file mode 100644
> > > index 00000000000..3c82808e5fe
> > > --- /dev/null
> > > +++ b/gdb/testsuite/gdb.debuginfod/libsection2.c
> > > @@ -0,0 +1,37 @@
> > > +/* This testcase is part of GDB, the GNU debugger.
> > > +
> > > + Copyright 2023-2024 Free Software Foundation, Inc.
> > > +
> > > + This program is free software; you can redistribute it and/or modify
> > > + it under the terms of the GNU General Public License as published by
> > > + the Free Software Foundation; either version 3 of the License, or
> > > + (at your option) any later version.
> > > +
> > > + This program is distributed in the hope that it will be useful,
> > > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > > + GNU General Public License for more details.
> > > +
> > > + You should have received a copy of the GNU General Public License
> > > + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> > > +
> > > +#include <stdio.h>
> > > +#include <pthread.h>
> > > +
> > > +void
> > > +libsection2_test ()
> > > +{
> > > + printf ("In libsection2\n");
> > > +}
> > > +
> > > +void *
> > > +libsection2_thread_test (void *arg)
> > > +{
> > > + int *flag = (int *) arg;
> > > + printf ("In thread test\n");
> > > +
> > > + while (1)
> > > + *flag = 1;
> > > +
> > > + return NULL;
> > > +}
> > > diff --git a/gdb/testsuite/gdb.debuginfod/section.c b/gdb/testsuite/gdb.debuginfod/section.c
> > > new file mode 100644
> > > index 00000000000..fa9ac536a6e
> > > --- /dev/null
> > > +++ b/gdb/testsuite/gdb.debuginfod/section.c
> > > @@ -0,0 +1,29 @@
> > > +/* This testcase is part of GDB, the GNU debugger.
> > > +
> > > + Copyright 2023-2024 Free Software Foundation, Inc.
> > > +
> > > + This program is free software; you can redistribute it and/or modify
> > > + it under the terms of the GNU General Public License as published by
> > > + the Free Software Foundation; either version 3 of the License, or
> > > + (at your option) any later version.
> > > +
> > > + This program is distributed in the hope that it will be useful,
> > > + but WITHOUT ANY WARRANTY; without even the implied warranty of
> > > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > > + GNU General Public License for more details.
> > > +
> > > + You should have received a copy of the GNU General Public License
> > > + along with this program. If not, see <http://www.gnu.org/licenses/>. */
> > > +
> > > +#include <stdio.h>
> > > +
> > > +extern void libsection1_test ();
> > > +
> > > +int
> > > +main ()
> > > +{
> > > + libsection1_test ();
> > > + printf ("in section exec\n");
> > > +
> > > + return 0;
> > > +}
> > > diff --git a/gdb/testsuite/gdb.debuginfod/section.exp b/gdb/testsuite/gdb.debuginfod/section.exp
> > > new file mode 100644
> > > index 00000000000..35086d15705
> > > --- /dev/null
> > > +++ b/gdb/testsuite/gdb.debuginfod/section.exp
> > > @@ -0,0 +1,157 @@
> > > +# Copyright 2023-2024 Free Software Foundation, Inc.
> > > +
> > > +# This program is free software; you can redistribute it and/or modify
> > > +# it under the terms of the GNU General Public License as published by
> > > +# the Free Software Foundation; either version 3 of the License, or
> > > +# (at your option) any later version.
> > > +#
> > > +# This program is distributed in the hope that it will be useful,
> > > +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> > > +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> > > +# GNU General Public License for more details.
> > > +#
> > > +# You should have received a copy of the GNU General Public License
> > > +# along with this program. If not, see <http://www.gnu.org/licenses/>.
> > > +
> > > +# Test debuginfod functionality
> > > +
> > > +standard_testfile
> > > +
> > > +load_lib debuginfod-support.exp
> > > +
> > > +clean_restart
> > > +require allow_debuginfod_tests
> > > +require allow_debuginfod_section_downloads
> > > +
> > > +# BINFILE calls a function from LIB_SL1.
> > > +set libfile1 "libsection1"
> > > +set libsrc1 $srcdir/$subdir/$libfile1.c
> > > +set lib_sl1 [standard_output_file $libfile1.sl]
> > > +
> > > +# LIB_SL1 calls functions from LIB_SL2.
> > > +set libfile2 "libsection2"
> > > +set libsrc2 $srcdir/$subdir/$libfile2.c
> > > +set lib_sl2 [standard_output_file $libfile2.sl]
> > > +
> > > +# Build LIB_SL2, LIB_SL1 and BINFILE.
> > > +if { [build_executable "build $libfile2" $lib_sl2 $libsrc2 \
> > > + {debug build-id shlib}] != 0 } {
> > > + return -1
> > > +}
> > > +
> > > +if { [build_executable "build $libfile1" $lib_sl1 $libsrc1 \
> > > + [list debug build-id shlib_pthreads shlib=$lib_sl2]] != 0 } {
> > > + return -1
> > > +}
> > > +
> > > +if { [build_executable "build executable" $binfile $srcfile \
> > > + [list debug build-id shlib=$lib_sl1 shlib=$lib_sl2]] != 0 } {
> > > + return -1
> > > +}
> > > +
> > > +# Make sure libsection1 and libsection2 contain .gdb_index.
> > > +if { [ensure_gdb_index $lib_sl1 "" "libsection1"] != 1 } {
> > > + untested "failed to add .gdb_index to $libfile1"
> > > + return -1
> > > +}
> > > +
> > > +if { [ensure_gdb_index $lib_sl2 "" "libsection2"] != 1 } {
> > > + untested "failed to add .gdb_index to $libfile2"
> > > + return -1
> > > +}
> > > +
> > > +# Strip solib debuginfo into separate files.
> > > +if { [gdb_gnu_strip_debug $lib_sl1 ""] != 0} {
> > > + fail "strip $lib_sl1 debuginfo"
> > > + return -1
> > > +}
> > > +
> > > +if { [gdb_gnu_strip_debug $lib_sl2 ""] != 0} {
> > > + fail "strip $lib_sl2 debuginfo"
> > > + return -1
> > > +}
> > > +
> > > +# Move debuginfo files into directory that debuginfod will serve from.
> > > +set debugdir [standard_output_file "debug"]
> > > +set debuginfo_sl1 [standard_output_file $libfile1.sl.debug]
> > > +set debuginfo_sl2 [standard_output_file $libfile2.sl.debug]
> > > +
> > > +file mkdir $debugdir
> > > +file rename -force $debuginfo_sl1 $debugdir
> > > +file rename -force $debuginfo_sl2 $debugdir
> > > +
> > > +# Restart GDB and clear the debuginfod client cache. Then load BINFILE into
> > > +# GDB and start running it. Match output with pattern RES and use TESTNAME
> > > +# as the test name.
> > > +proc_with_prefix clean_restart_with_prompt { binfile testname } {
> > > + global cache
> > > +
> > > + # Delete client cache so debuginfo downloads again.
> > > + file delete -force $cache
> > > + clean_restart
> > > +
> > > + gdb_test_no_output "set debuginfod enabled on" \
> > > + "clean_restart enable $testname"
> > > + gdb_load $binfile
> > > +
> > > + runto_main
> > > +}
> > > +
> > > +# Tests with no debuginfod server running.
> > > +proc_with_prefix no_url { } {
> > > + global binfile libfile1 libfile2
> > > +
> > > + gdb_load $binfile
> > > + if {![runto_main]} {
> > > + return
> > > + }
> > > +
> > > + # Check that no section is downloaded and no debuginfo is found.
> > > + gdb_test "info sharedlibrary" ".*Yes \\(\\*\\).*$libfile1.*" \
> > > + "found no url lib1"
> > > + gdb_test "info sharedlibrary" ".*Yes \\(\\*\\).*$libfile2.*" \
> > > + "found no url lib2"
> > > +}
> > > +
> > > +# Tests with a debuginfod server running.
> > > +proc_with_prefix local_url { } {
> > > + global binfile
> > > + global libsrc1 lib_sl1 libfile1
> > > + global libsrc2 lib_sl2 libfile2
> > > + global debugdir db
> > > +
> > > + set url [start_debuginfod $db $debugdir]
> > > + if { $url == "" } {
> > > + unresolved "failed to start debuginfod server"
> > > + return
> > > + }
> > > +
> > > + # Point GDB to the server.
> > > + setenv DEBUGINFOD_URLS $url
> > > +
> > > + clean_restart_with_prompt $binfile "index"
> > > +
> > > + # Download debuginfo when stepping into a function.
> > > + set res ".*separate debug info for $lib_sl1.*\"In ${libfile1}\\\\n\".*"
> > > + gdb_test "step" $res "step"
> > > +
> > > + clean_restart_with_prompt $binfile "break"
> > > +
> > > + # Download debuginfo when setting a breakpoint.
> > > + set res "Download.*separate debug info for $lib_sl2.*"
> > > + gdb_test "br libsection2_test" $res "break set"
> > > +
> > > + # Hit the breakpoint.
> > > + set res ".*Breakpoint 2, libsection2_test.*\"In ${libfile2}\\\\n\".*"
> > > + gdb_test "c" $res "break continue"
> > > +}
> > > +
> > > +# Create CACHE and DB directories ready for debuginfod to use.
> > > +prepare_for_debuginfod cache db
> > > +
> > > +with_debuginfod_env $cache {
> > > + no_url
> > > + local_url
> > > +}
> > > +
> > > +stop_debuginfod
> > > diff --git a/gdb/testsuite/lib/debuginfod-support.exp b/gdb/testsuite/lib/debuginfod-support.exp
> > > index 0096448567e..4d8bdcdc7cb 100644
> > > --- a/gdb/testsuite/lib/debuginfod-support.exp
> > > +++ b/gdb/testsuite/lib/debuginfod-support.exp
> > > @@ -113,6 +113,8 @@ proc with_debuginfod_env { cache body } {
> > > proc start_debuginfod { db debugdir } {
> > > global debuginfod_spawn_id spawn_id
> > >
> > > + set logfile [standard_output_file "server_log"]
> > > +
> > > # Find an unused port.
> > > set port 7999
> > > set found false
> > > @@ -127,7 +129,8 @@ proc start_debuginfod { db debugdir } {
> > > set old_spawn_id $spawn_id
> > > }
> > >
> > > - spawn debuginfod -vvvv -d $db -p $port -F $debugdir
> > > + spawn sh -c "debuginfod -vvvv -d $db -p $port -F $debugdir 2>&1 \
> > > + | tee $logfile"
> > > set debuginfod_spawn_id $spawn_id
> > >
> > > if { [info exists old_spawn_id] } {
> > > @@ -194,3 +197,23 @@ proc stop_debuginfod { } {
> > > unset debuginfod_spawn_id
> > > }
> > > }
> > > +
> > > +# Return true if gdb is configured to download ELF/DWARF sections from
> > > +# debuginfod servers. Otherwise return false.
> > > +proc allow_debuginfod_section_downloads { } {
> > > + set cmd "maint set debuginfod download-sections on"
> > > + set msg "enable section downloads"
> > > +
> > > + gdb_test_multiple $cmd $msg {
> > > + -re -wrap ".*not compiled into GDB.*" {
> > > + return false
> > > + }
> > > + -re -wrap "^" {
> > > + return true
> > > + }
> > > + -re -wrap "" {
> > > + fail "$gdb_test_name (unexpected output)"
> > > + return false
> > > + }
> > > + }
> > > +}
> > > diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
> > > index 5fb81f6e28d..a3680cb6371 100644
> > > --- a/gdb/testsuite/lib/gdb.exp
> > > +++ b/gdb/testsuite/lib/gdb.exp
> > > @@ -9515,10 +9515,14 @@ proc get_index_type { objfile { testname "" } } {
> > > # STYLE controls which style of index to add, if needed. The empty
> > > # string (the default) means .gdb_index; "-dwarf-5" means .debug_names.
> > >
> > > -proc ensure_gdb_index { binfile {style ""} } {
> > > +proc ensure_gdb_index { binfile {style ""} {testname_prefix ""} } {
> > > set testfile [file tail $binfile]
> > > -
> > > set test "check if index present"
> > > +
> > > + if { $testname_prefix != "" } {
> > > + set test "$testname_prefix $test"
> > > + }
> > > +
> > > set index_type [get_index_type $testfile $test]
> > >
> > > if { $index_type eq "gdb" || $index_type eq "dwarf5" } {
> > > --
> > > 2.43.0
> > >
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading
2024-05-03 22:59 ` [PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading Aaron Merey
2024-05-17 14:10 ` [PING][PATCH " Aaron Merey
@ 2024-06-17 10:45 ` Andrew Burgess
2024-06-17 19:57 ` Aaron Merey
1 sibling, 1 reply; 18+ messages in thread
From: Andrew Burgess @ 2024-06-17 10:45 UTC (permalink / raw)
To: Aaron Merey, gdb-patches; +Cc: Aaron Merey
Aaron,
Thanks for continuing to work on this, and sorry it's taken me so long
to get back to looking at this patch.
Aaron Merey <amerey@redhat.com> writes:
> v6 fixes a comment for dwarf2_has_separate_index and merges
> lookup_gdb_index_debuginfod with lookup_gdb_index.
>
> At the beginning of a session, gdb may attempt to download debuginfo
> for all shared libraries associated with the process or core file
> being debugged. This can be a waste of time and storage space when much
> of the debuginfo ends up not being used during the session.
>
> To reduce the gdb's startup latency and to download only the debuginfo
> that is really needed, this patch adds on-demand downloading of debuginfo.
>
> 'set debuginfo enabled on' now causes gdb to attempt to download a .gdb_index
> for each shared library instead of its full debuginfo. Each corresponding
> separate debuginfo will be deferred until gdb needs to expand symtabs
> associated with the debuginfo's index.
>
> Because these indices are significantly smaller than their corresponding
> debuginfo, this generally reduces the total amount of data gdb downloads.
> Reductions of 80%-95% have been observed when debugging large GUI programs.
>
> (gdb) set debuginfod enabled on
> (gdb) start
> Downloading section .gdb_index for /lib64/libcurl.so.4
> [...]
> 1826 client->server_mhandle = curl_multi_init ();
> (gdb) step
> Downloading separate debug info for /lib64/libcurl.so.4
> Downloading separate debug info for [libcurl dwz]
> Downloading source file /usr/src/debug/curl-7.85.0-6.fc37.x86_64/build-full/lib/../../lib/multi.c
> curl_multi_init () at ../../lib/multi.c:457
> 457 {
> (gdb)
>
> Some of the key functions below include dwarf2_has_separate_index which
> downloads the separate .gdb_index. If successful, the shared library
> objfile owns the index until the separate debug objfile is downloaded
> or confirmed to not be available.
>
> read_full_dwarf_from_debuginfod downloads the full debuginfo and
> initializes the separate debug objfile. It is called by functions
> such as dwarf2_gdb_index::expand_symtabs_matching and
> dwarf2_base_index_functions::find_pc_sect_compunit_symtab when symtab
> expansion is required.
> ---
> diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
> index e743ce13978..97b8137aabf 100644
> --- a/gdb/dwarf2/read-gdb-index.c
> +++ b/gdb/dwarf2/read-gdb-index.c
> @@ -139,6 +139,8 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> gdb.dwarf2/gdb-index.exp testcase. */
> void dump (struct objfile *objfile) override;
>
> + /* Calls do_expand_symtabs_matching and triggers debuginfo downloading
> + if necessary. */
> bool expand_symtabs_matching
> (struct objfile *objfile,
> gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> @@ -147,8 +149,59 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> block_search_flags search_flags,
> domain_search_flags domain) override;
> +
> + /* Expand symtabs matching a given symbol or file. */
> + bool do_expand_symtabs_matching
> + (struct objfile *objfile,
> + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher,
> + const lookup_name_info *lookup_name,
> + gdb::function_view<expand_symtabs_symbol_matcher_ftype> symbol_matcher,
> + gdb::function_view<expand_symtabs_exp_notify_ftype> expansion_notify,
> + block_search_flags search_flags,
> + domain_search_flags domain);
> +
> + /* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
> + debuginfo if necessary. */
> + void expand_all_symtabs (struct objfile *objfile) override;
> +
> + /* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
> + debuginfo if necessary. */
> + struct symtab *find_last_source_symtab (struct objfile *objfile) override;
> };
>
> +void
> +dwarf2_gdb_index::expand_all_symtabs (struct objfile *objfile)
> +{
> + try
> + {
> + dwarf2_base_index_functions::expand_all_symtabs (objfile);
> + }
> + catch (const gdb_exception &e)
> + {
> + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> + exception_print (gdb_stderr, e);
> + else
> + read_full_dwarf_from_debuginfod (objfile, this);
I know this isn't what you want to hear, but I really don't think we
should move forward with this approach.
Every piece of C++ advice I've ever read always says the same thing: use
exceptions for error handling, not as an alternative control flow.
Now only is it confusing to folk who will assume that try/catch is all
about handling errors, but in this exact case the objfile might well
have deferred debug information, but some other exception might have
been thrown, in which case we're going to swallow the real error and
just read the deferred debug. If we did want this to work reliably we'd
need a unique exception type which could be thrown and caught that is
_only_ thrown when we want deferred debug to be loaded...
...and by that point, you might has well move the deferred debug loading
to the point where the exception is being thrown.
Which is exactly the route I think we need to go down.
If I understand what's going on correctly you're hooking around the code
which searches the index. Normally (in the non debuginfod case) we
parse the index, then when a request arrives for some debug information
we'll scan the index, and if we get a hit we expand the relevant CU's
debug information.
This would seem to imply there's a two stage process, (1) a search of
the index, and (2) an expansion phase if we get a hit.
In my head, an ideal solution would parameterise this process, stage
(1), the search remains common, but what we do when we get a hit
changes. In the "normal" case we expand the CU's we already have on
hand, and in the "debuginfod" case we'll download (and expand/parse) the
full debug info.
Did you consider an approach like this at all? If so what issues did
you encounter? If not, would you be willing to consider this approach?
For the wider GDB maintainer / contributor community, am I being too
harsh here? Is my reluctance to use exceptions for (what I see as)
control flow unreasonable? If I'm in the minority here then I'm happy
to stand aside.
Thanks,
Andrew
^ permalink raw reply [flat|nested] 18+ messages in thread* Re: [PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading
2024-06-17 10:45 ` [PATCH " Andrew Burgess
@ 2024-06-17 19:57 ` Aaron Merey
0 siblings, 0 replies; 18+ messages in thread
From: Aaron Merey @ 2024-06-17 19:57 UTC (permalink / raw)
To: Andrew Burgess; +Cc: gdb-patches
Hi Andrew,
On Mon, Jun 17, 2024 at 6:45 AM Andrew Burgess <aburgess@redhat.com> wrote:
>
> > +void
> > +dwarf2_gdb_index::expand_all_symtabs (struct objfile *objfile)
> > +{
> > + try
> > + {
> > + dwarf2_base_index_functions::expand_all_symtabs (objfile);
> > + }
> > + catch (const gdb_exception &e)
> > + {
> > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > + exception_print (gdb_stderr, e);
> > + else
> > + read_full_dwarf_from_debuginfod (objfile, this);
>
> I know this isn't what you want to hear, but I really don't think we
> should move forward with this approach.
>
> Every piece of C++ advice I've ever read always says the same thing: use
> exceptions for error handling, not as an alternative control flow.
>
> Now only is it confusing to folk who will assume that try/catch is all
> about handling errors, but in this exact case the objfile might well
> have deferred debug information, but some other exception might have
> been thrown, in which case we're going to swallow the real error and
> just read the deferred debug. If we did want this to work reliably we'd
> need a unique exception type which could be thrown and caught that is
> _only_ thrown when we want deferred debug to be loaded...
If any error occurs during CU expansion, I think gdb should attempt to
download deferred debuginfo since there is a chance that the error
will be bypassed when the expansion is performed with full debuginfo
available.
There is of course a chance that gdb downloads the debuginfo
unnecessarily and the error occurs again. But I think we should weigh
this risk of unnecessary downloading against the current approach
(download all debuginfo immediately), which typically results in many
more unnecessary downloads.
If we want gdb to try downloading deferred debuginfo for any error
here then it makes sense to keep the current exception-based
approach.
>
> ...and by that point, you might has well move the deferred debug loading
> to the point where the exception is being thrown.
>
> Which is exactly the route I think we need to go down.
>
> If I understand what's going on correctly you're hooking around the code
> which searches the index. Normally (in the non debuginfod case) we
> parse the index, then when a request arrives for some debug information
> we'll scan the index, and if we get a hit we expand the relevant CU's
> debug information.
>
> This would seem to imply there's a two stage process, (1) a search of
> the index, and (2) an expansion phase if we get a hit.
>
> In my head, an ideal solution would parameterise this process, stage
> (1), the search remains common, but what we do when we get a hit
> changes. In the "normal" case we expand the CU's we already have on
> hand, and in the "debuginfod" case we'll download (and expand/parse) the
> full debug info.
>
> Did you consider an approach like this at all? If so what issues did
> you encounter? If not, would you be willing to consider this approach?
Part of the rationale for the current approach was to fallback to
deferred debuginfo downloading for any error in the hopes that
retrying with full debuginfo bypasses the error.
If the consensus is that the exception-based approach should change
then I am willing to rewrite the patch. Although I think there is
some benefit to merging the patch as-is, even if we decide to
change parts of it later. The performance improvements are
very significant and I think users would appreciate having
this sooner rather than later. Plus the changes being proposed
are implementation details that are more or less hidden to users
(I have not run into a situation where an unrelated error triggers
an unnecessary deferred debuginfo download). Since gdb 15.1
branched recently, I think now is a good time for users to
experiment with this feature well before the next gdb release.
>
> For the wider GDB maintainer / contributor community, am I being too
> harsh here? Is my reluctance to use exceptions for (what I see as)
> control flow unreasonable? If I'm in the minority here then I'm happy
> to stand aside.
Aaron
^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH 3/3 v6] gdb/debuginfod: Add .debug_line downloading
2024-05-03 22:58 [PATCH 0/3] On-demand debuginfo downloading Aaron Merey
2024-05-03 22:59 ` [PATCH 1/3 v3] gdb/progspace: Add reverse safe iterator Aaron Merey
2024-05-03 22:59 ` [PATCH 2/3 v6] gdb/debuginfod: Support on-demand debuginfo downloading Aaron Merey
@ 2024-05-03 22:59 ` Aaron Merey
2024-05-17 14:10 ` [PING][PATCH " Aaron Merey
2 siblings, 1 reply; 18+ messages in thread
From: Aaron Merey @ 2024-05-03 22:59 UTC (permalink / raw)
To: gdb-patches; +Cc: aburgess, Aaron Merey
This is a repost of v6 from
https://sourceware.org/pipermail/gdb-patches/2024-January/205956.html
ELF/DWARF section downloading allows gdb to download .gdb_index files in
order to defer full debuginfo downloads. However .gdb_index does not
contain any information regarding source filenames. When a gdb command
includes a filename argument (ex. 'break main.c:50'), this results in
the mass downloading of all deferred debuginfo so that gdb can search the
debuginfo for matching source filenames. This can result in unnecessary
downloads.
To improve this, have gdb instead download each debuginfo's .debug_line
(and .debug_line_str if using DWARF5) when executing these commands.
Download full debuginfo only when its .debug_line contains a matching
filename.
Since the combined size of .debug_line and .debug_line_str is only about
1% the size of the corresponding debuginfo, significant time can be saved
by checking these sections before choosing to download an entire debuginfo.
This patch also redirects stdout and stderr of the debuginfod server
used by testsuite/gdb.debuginfod tests to a server_log standard output
file. While adding tests for this patch I ran into an issue where the
test server would block when logging to stderr, presumably because the
stderr buffer filled up and wasn't being read from. Redirecting the
log to a file fixes this and also makes the server log more accessible
when debugging test failures.
---
gdb/cli-out.c | 11 +-
gdb/completer.c | 18 +-
| 161 ++++++++++++------
| 10 ++
gdb/dwarf2/read-gdb-index.c | 60 +++++++
gdb/dwarf2/read.c | 208 +++++++++++++++++++++++
gdb/dwarf2/read.h | 37 ++++
gdb/mi/mi-out.c | 9 +-
gdb/testsuite/gdb.debuginfod/section.exp | 27 +++
gdb/ui-out.c | 3 +
gdb/ui-out.h | 20 +++
11 files changed, 511 insertions(+), 53 deletions(-)
diff --git a/gdb/cli-out.c b/gdb/cli-out.c
index 1c303f09662..d65b69ebe01 100644
--- a/gdb/cli-out.c
+++ b/gdb/cli-out.c
@@ -306,16 +306,23 @@ cli_ui_out::do_progress_notify (const std::string &msg,
if (info.state == progress_update::START)
{
+ std::string prefix;
+ if (cur_prefix_state == prefix_state_t::NEWLINE_NEEDED)
+ {
+ prefix = "\n";
+ cur_prefix_state = prefix_state_t::NEWLINE_PRINTED;
+ }
+
if (stream->isatty ()
&& current_ui->input_interactive_p ()
&& chars_per_line >= MIN_CHARS_PER_LINE)
{
- gdb_printf (stream, "%s\n", msg.c_str ());
+ gdb_printf (stream, "%s\n", (prefix + msg).c_str ());
info.state = progress_update::BAR;
}
else
{
- gdb_printf (stream, "%s...\n", msg.c_str ());
+ gdb_printf (stream, "%s...\n", (prefix + msg).c_str ());
info.state = progress_update::WORKING;
}
}
diff --git a/gdb/completer.c b/gdb/completer.c
index f1f44109bdc..708081cf280 100644
--- a/gdb/completer.c
+++ b/gdb/completer.c
@@ -1354,6 +1354,10 @@ complete_line_internal_1 (completion_tracker &tracker,
{
/* We've recognized a full command. */
+ /* Disable pagination since responding to the pagination prompt
+ overwrites rl_line_buffer. */
+ scoped_restore pag_restore = make_scoped_restore (&pagination_enabled, false);
+
if (p == tmp_command + point)
{
/* There is no non-whitespace in the line beyond the
@@ -1453,7 +1457,8 @@ complete_line_internal_1 (completion_tracker &tracker,
}
/* Wrapper around complete_line_internal_1 to handle
- MAX_COMPLETIONS_REACHED_ERROR. */
+ MAX_COMPLETIONS_REACHED_ERROR and possible progress update
+ interactions. */
static void
complete_line_internal (completion_tracker &tracker,
@@ -1461,6 +1466,11 @@ complete_line_internal (completion_tracker &tracker,
const char *line_buffer, int point,
complete_line_internal_reason reason)
{
+ scoped_restore restore_prefix_state
+ = make_scoped_restore
+ (&cur_prefix_state,
+ ui_out::progress_update::prefix_state::NEWLINE_NEEDED);
+
try
{
complete_line_internal_1 (tracker, text, line_buffer, point, reason);
@@ -1470,6 +1480,12 @@ complete_line_internal (completion_tracker &tracker,
if (except.error != MAX_COMPLETIONS_REACHED_ERROR)
throw;
}
+
+ /* If progress update messages printed, then the text being completed
+ needs to be printed again. */
+ if (cur_prefix_state
+ == ui_out::progress_update::prefix_state::NEWLINE_PRINTED)
+ rl_forced_update_display ();
}
/* See completer.h. */
--git a/gdb/dwarf2/line-header.c b/gdb/dwarf2/line-header.c
index eddb2ef7ae8..1d24386c0eb 100644
--- a/gdb/dwarf2/line-header.c
+++ b/gdb/dwarf2/line-header.c
@@ -101,12 +101,15 @@ read_checked_initial_length_and_offset (bfd *abfd, const gdb_byte *buf,
{
LONGEST length = read_initial_length (abfd, buf, bytes_read);
- gdb_assert (cu_header->initial_length_size == 4
- || cu_header->initial_length_size == 8
- || cu_header->initial_length_size == 12);
+ if (cu_header != nullptr)
+ {
+ gdb_assert (cu_header->initial_length_size == 4
+ || cu_header->initial_length_size == 8
+ || cu_header->initial_length_size == 12);
- if (cu_header->initial_length_size != *bytes_read)
- complaint (_("intermixed 32-bit and 64-bit DWARF sections"));
+ if (cu_header->initial_length_size != *bytes_read)
+ complaint (_("intermixed 32-bit and 64-bit DWARF sections"));
+ }
*offset_size = (*bytes_read == 4) ? 4 : 8;
return length;
@@ -115,21 +118,27 @@ read_checked_initial_length_and_offset (bfd *abfd, const gdb_byte *buf,
/* Read directory or file name entry format, starting with byte of
format count entries, ULEB128 pairs of entry formats, ULEB128 of
entries count and the entries themselves in the described entry
- format. */
+ format.
+
+ .debug_line and .debug_line_str sections are stored in LINE_BUFP
+ and LINE_STR_DATA respectively. */
static void
-read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
- const gdb_byte **bufp, struct line_header *lh,
- unsigned int offset_size,
- void (*callback) (struct line_header *lh,
- const char *name,
- dir_index d_index,
- unsigned int mod_time,
- unsigned int length))
+read_formatted_entries
+ (bfd *abfd, const gdb_byte **line_bufp,
+ const gdb::array_view<const gdb_byte> line_str_data,
+ struct line_header *lh,
+ unsigned int offset_size,
+ void (*callback) (struct line_header *lh,
+ const char *name,
+ dir_index d_index,
+ unsigned int mod_time,
+ unsigned int length))
{
gdb_byte format_count, formati;
ULONGEST data_count, datai;
- const gdb_byte *buf = *bufp;
+ const gdb_byte *buf = *line_bufp;
+ const gdb_byte *str_buf = line_str_data.data ();
const gdb_byte *format_header_data;
unsigned int bytes_read;
@@ -156,7 +165,7 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
ULONGEST content_type = read_unsigned_leb128 (abfd, format, &bytes_read);
format += bytes_read;
- ULONGEST form = read_unsigned_leb128 (abfd, format, &bytes_read);
+ ULONGEST form = read_unsigned_leb128 (abfd, format, &bytes_read);
format += bytes_read;
std::optional<const char *> string;
@@ -171,12 +180,24 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
case DW_FORM_line_strp:
{
- const char *str
- = per_objfile->read_line_string (buf, offset_size);
+ if (line_str_data.empty ())
+ warning (_("Dwarf Error: DW_FORM_line_strp used without " \
+ "required section"));
+
+ if (line_str_data.size () <= offset_size)
+ warning (_("Dwarf Error: DW_FORM_line_strp pointing outside " \
+ "of section .debug_line_str"));
+
+ ULONGEST str_offset = read_offset (abfd, buf, offset_size);
+ const char *str;
+ if (str_buf[str_offset] == '\0')
+ str = nullptr;
+ else
+ str = (const char *) (str_buf + str_offset);
string.emplace (str);
buf += offset_size;
+ break;
}
- break;
case DW_FORM_data1:
uint.emplace (read_1_byte (abfd, buf));
@@ -247,28 +268,30 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
callback (lh, fe.name, fe.d_index, fe.mod_time, fe.length);
}
- *bufp = buf;
+ *line_bufp = buf;
}
/* See line-header.h. */
line_header_up
-dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
- dwarf2_per_objfile *per_objfile,
- struct dwarf2_section_info *section,
- const struct comp_unit_head *cu_header,
- const char *comp_dir)
+dwarf_decode_line_header (bfd *abfd,
+ gdb::array_view<const gdb_byte> line_data,
+ gdb::array_view<const gdb_byte> line_str_data,
+ const gdb_byte **debug_line_ptr,
+ bool is_dwz,
+ const struct comp_unit_head *cu_header,
+ const char *comp_dir)
{
- const gdb_byte *line_ptr;
+ const gdb_byte *line_ptr, *buf;
unsigned int bytes_read, offset_size;
int i;
const char *cur_dir, *cur_file;
- bfd *abfd = section->get_bfd_owner ();
+ buf = *debug_line_ptr;
/* Make sure that at least there's room for the total_length field.
That could be 12 bytes long, but we're just going to fudge that. */
- if (to_underlying (sect_off) + 4 >= section->size)
+ if (buf + 4 >= line_data.data () + line_data.size ())
{
dwarf2_statement_list_fits_in_line_number_section_complaint ();
return 0;
@@ -276,24 +299,26 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
line_header_up lh (new line_header (comp_dir));
- lh->sect_off = sect_off;
+ lh->sect_off = (sect_offset) (buf - line_data.data ());
lh->offset_in_dwz = is_dwz;
- line_ptr = section->buffer + to_underlying (sect_off);
+ line_ptr = buf;
/* Read in the header. */
LONGEST unit_length
- = read_checked_initial_length_and_offset (abfd, line_ptr, cu_header,
+ = read_checked_initial_length_and_offset (abfd, buf, cu_header,
&bytes_read, &offset_size);
- line_ptr += bytes_read;
- const gdb_byte *start_here = line_ptr;
+ line_ptr += bytes_read;
- if (line_ptr + unit_length > (section->buffer + section->size))
+ if (line_ptr + unit_length > buf + line_data.size ())
{
dwarf2_statement_list_fits_in_line_number_section_complaint ();
return 0;
}
+
+ const gdb_byte *start_here = line_ptr;
+
lh->statement_program_end = start_here + unit_length;
lh->version = read_2_bytes (abfd, line_ptr);
line_ptr += 2;
@@ -302,7 +327,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
/* This is a version we don't understand. The format could have
changed in ways we don't handle properly so just punt. */
complaint (_("unsupported version in .debug_line section"));
- return NULL;
+ return nullptr;
}
if (lh->version >= 5)
{
@@ -319,13 +344,14 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
complaint (_("unsupported segment selector size %u "
"in .debug_line section"),
segment_selector_size);
- return NULL;
+ return nullptr;
}
}
LONGEST header_length = read_offset (abfd, line_ptr, offset_size);
line_ptr += offset_size;
lh->statement_program_start = line_ptr + header_length;
+
lh->minimum_instruction_length = read_1_byte (abfd, line_ptr);
line_ptr += 1;
@@ -346,12 +372,16 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
lh->default_is_stmt = read_1_byte (abfd, line_ptr);
line_ptr += 1;
+
lh->line_base = read_1_signed_byte (abfd, line_ptr);
line_ptr += 1;
+
lh->line_range = read_1_byte (abfd, line_ptr);
line_ptr += 1;
+
lh->opcode_base = read_1_byte (abfd, line_ptr);
line_ptr += 1;
+
lh->standard_opcode_lengths.reset (new unsigned char[lh->opcode_base]);
lh->standard_opcode_lengths[0] = 1; /* This should never be used anyway. */
@@ -364,21 +394,23 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
if (lh->version >= 5)
{
/* Read directory table. */
- read_formatted_entries (per_objfile, abfd, &line_ptr, lh.get (),
- offset_size,
- [] (struct line_header *header, const char *name,
- dir_index d_index, unsigned int mod_time,
- unsigned int length)
+ read_formatted_entries
+ (abfd, &line_ptr, line_str_data,
+ lh.get (), offset_size,
+ [] (struct line_header *header, const char *name,
+ dir_index d_index, unsigned int mod_time,
+ unsigned int length)
{
header->add_include_dir (name);
});
/* Read file name table. */
- read_formatted_entries (per_objfile, abfd, &line_ptr, lh.get (),
- offset_size,
- [] (struct line_header *header, const char *name,
- dir_index d_index, unsigned int mod_time,
- unsigned int length)
+ read_formatted_entries
+ (abfd, &line_ptr, line_str_data,
+ lh.get (), offset_size,
+ [] (struct line_header *header, const char *name,
+ dir_index d_index, unsigned int mod_time,
+ unsigned int length)
{
header->add_file_name (name, d_index, mod_time, length);
});
@@ -386,7 +418,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
else
{
/* Read directory table. */
- while ((cur_dir = read_direct_string (abfd, line_ptr, &bytes_read)) != NULL)
+ while ((cur_dir = read_direct_string (abfd, line_ptr, &bytes_read)) != nullptr)
{
line_ptr += bytes_read;
lh->add_include_dir (cur_dir);
@@ -394,7 +426,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
line_ptr += bytes_read;
/* Read file name table. */
- while ((cur_file = read_direct_string (abfd, line_ptr, &bytes_read)) != NULL)
+ while ((cur_file = read_direct_string (abfd, line_ptr, &bytes_read)) != nullptr)
{
unsigned int mod_time, length;
dir_index d_index;
@@ -412,9 +444,40 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
line_ptr += bytes_read;
}
- if (line_ptr > (section->buffer + section->size))
+ if (line_ptr > (buf + line_data.size ()))
complaint (_("line number info header doesn't "
"fit in `.debug_line' section"));
+ *debug_line_ptr += unit_length + offset_size;
return lh;
}
+
+line_header_up
+dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
+ dwarf2_per_objfile *per_objfile,
+ struct dwarf2_section_info *section,
+ const struct comp_unit_head *cu_header,
+ const char *comp_dir)
+{
+ struct objfile *objfile = per_objfile->objfile;
+ struct dwarf2_per_bfd *per_bfd = per_objfile->per_bfd;
+
+ /* Read .debug_line. */
+ dwarf2_section_info *line_sec = &per_bfd->line;
+ bfd_size_type line_size = line_sec->get_size (objfile);
+
+ gdb::array_view<const gdb_byte> line (line_sec->buffer, line_size);
+
+ /* Read .debug_line_str. */
+ dwarf2_section_info *line_str_sec = &per_bfd->line_str;
+ bfd_size_type line_str_size = line_str_sec->get_size (objfile);
+
+ gdb::array_view<const gdb_byte> line_str (line_str_sec->buffer,
+ line_str_size);
+
+ const gdb_byte *line_ptr = line.data () + to_underlying (sect_off);
+
+ return dwarf_decode_line_header
+ (per_bfd->obfd, line, line_str, &line_ptr,
+ is_dwz, cu_header, comp_dir);
+}
--git a/gdb/dwarf2/line-header.h b/gdb/dwarf2/line-header.h
index c068dff70a3..45011882321 100644
--- a/gdb/dwarf2/line-header.h
+++ b/gdb/dwarf2/line-header.h
@@ -217,4 +217,14 @@ extern line_header_up dwarf_decode_line_header
struct dwarf2_section_info *section, const struct comp_unit_head *cu_header,
const char *comp_dir);
+/* Like above but the .debug_line and .debug_line_str are stored in
+ LINE_DATA and LINE_STR_DATA. *DEBUG_LINE_PTR should point to a
+ statement program header within LINE_DATA. */
+
+extern line_header_up dwarf_decode_line_header
+ (bfd *abfd, gdb::array_view<const gdb_byte> line_data,
+ gdb::array_view<const gdb_byte> line_str_data,
+ const gdb_byte **debug_line_ptr, bool is_dwz,
+ const comp_unit_head *cu_header, const char *comp_dir);
+
#endif /* DWARF2_LINE_HEADER_H */
diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
index 97b8137aabf..c52099d5f6a 100644
--- a/gdb/dwarf2/read-gdb-index.c
+++ b/gdb/dwarf2/read-gdb-index.c
@@ -131,6 +131,9 @@ struct mapped_gdb_index final : public mapped_index_base
}
};
+struct mapped_debug_line;
+typedef std::unique_ptr<mapped_debug_line> mapped_debug_line_up;
+
struct dwarf2_gdb_index : public dwarf2_base_index_functions
{
/* This dumps minimal information about the index.
@@ -160,6 +163,15 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
block_search_flags search_flags,
domain_search_flags domain);
+ /* If OBJFILE's debuginfo download has been deferred, use a mapped_debug_line
+ to generate filenames.
+
+ Otherwise call dwarf2_base_index_functions::map_symbol_filenames. */
+
+ void map_symbol_filenames (struct objfile *objfile,
+ gdb::function_view<symbol_filename_ftype> fun,
+ bool need_fullname) override;
+
/* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
debuginfo if necessary. */
void expand_all_symtabs (struct objfile *objfile) override;
@@ -167,6 +179,15 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
/* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
debuginfo if necessary. */
struct symtab *find_last_source_symtab (struct objfile *objfile) override;
+
+ /* Filename information related to this .gdb_index. */
+ mapped_debug_line_up mdl;
+
+ /* Return true if any of the filenames in this .gdb_index's .debug_line
+ mapping match FILE_MATCHER. Initializes the mapping if necessary. */
+ bool filename_in_debug_line
+ (objfile *objfile,
+ gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher);
};
void
@@ -202,6 +223,30 @@ dwarf2_gdb_index::find_last_source_symtab (struct objfile *objfile)
}
}
+void
+dwarf2_gdb_index::map_symbol_filenames
+ (struct objfile *objfile,
+ gdb::function_view<symbol_filename_ftype> fun,
+ bool need_fullname)
+{
+ try
+ {
+ dwarf2_base_index_functions::map_symbol_filenames (objfile, fun,
+ need_fullname);
+ }
+ catch (const gdb_exception &e)
+ {
+ if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
+ exception_print (gdb_stderr, e);
+ else
+ {
+ if (mdl == nullptr)
+ mdl.reset (new mapped_debug_line (objfile));
+ mdl->map_filenames (fun, need_fullname);
+ }
+ }
+}
+
/* This dumps minimal information about the index.
It is called via "mt print objfiles".
One use is to verify .gdb_index has been loaded by the
@@ -365,6 +410,17 @@ dwarf2_gdb_index::do_expand_symtabs_matching
return result;
}
+bool
+dwarf2_gdb_index::filename_in_debug_line
+ (objfile *objfile,
+ gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
+{
+ if (mdl == nullptr)
+ mdl.reset (new mapped_debug_line (objfile));
+
+ return mdl->contains_matching_filename (file_matcher);
+}
+
bool
dwarf2_gdb_index::expand_symtabs_matching
(struct objfile *objfile,
@@ -392,6 +448,10 @@ dwarf2_gdb_index::expand_symtabs_matching
return false;
}
+ if (file_matcher != nullptr
+ && !filename_in_debug_line (objfile, file_matcher))
+ return true;
+
read_full_dwarf_from_debuginfod (objfile, this);
return true;
}
diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
index 4461c76347e..ca219c9b910 100644
--- a/gdb/dwarf2/read.c
+++ b/gdb/dwarf2/read.c
@@ -81,6 +81,7 @@
#include <optional>
#include "gdbsupport/underlying.h"
#include "gdbsupport/hash_enum.h"
+#include "gdbsupport/scoped_mmap.h"
#include "filename-seen-cache.h"
#include "producer.h"
#include <fcntl.h>
@@ -1967,6 +1968,213 @@ dw2_get_file_names (dwarf2_per_cu_data *this_cu,
return this_cu->file_names;
}
+#if !HAVE_SYS_MMAN_H
+
+bool
+mapped_debug_line::contains_matching_filename
+ (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
+{
+ return false;
+}
+
+gdb::array_view<const gdb_byte>
+mapped_debug_line::read_debug_line_separate
+ (const char *filename, std::unique_ptr<index_cache_resource> *resource)
+{
+ return {};
+}
+
+bool
+mapped_debug_line::read_debug_line_from_debuginfod (objfile *objfile)
+{
+ return false;
+}
+
+void
+mapped_debug_line::map_filenames
+ (gdb::function_view<symbol_filename_ftype> fun,
+ bool need_fullname)
+{
+ return;
+}
+
+#else /* !HAVE_SYS_MMAN_H */
+
+struct line_resource_mmap final : public index_cache_resource
+{
+ /* Try to mmap FILENAME. Throw an exception on failure, including if the
+ file doesn't exist. */
+ line_resource_mmap (const char *filename)
+ : mapping (mmap_file (filename))
+ {}
+
+ scoped_mmap mapping;
+};
+
+/* See read.h. */
+
+bool
+mapped_debug_line::contains_matching_filename
+ (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
+{
+ for (line_header_up &lh : line_headers)
+ for (file_entry &fe : lh->file_names ())
+ {
+ const char *filename = fe.name;
+
+ if (file_matcher (fe.name, false))
+ return true;
+
+ bool basename_match = file_matcher (lbasename (fe.name), true);
+
+ if (!basenames_may_differ && !basename_match)
+ continue;
+
+ /* DW_AT_comp_dir is not explicitly mentioned in the .debug_line
+ until DWARF5. Since we don't have access to the CU at this
+ point we just check for a partial match on the filename.
+ If there is a match, the full debuginfo will be downloaded
+ ane the match will be re-evalute with DW_AT_comp_dir. */
+ if (lh->version < 5 && fe.d_index == 0)
+ return basename_match;
+
+ const char *dirname = fe.include_dir (&*lh);
+ std::string fullname;
+
+ if (dirname == nullptr || IS_ABSOLUTE_PATH (filename))
+ fullname = filename;
+ else
+ fullname = std::string (dirname) + SLASH_STRING + filename;
+
+ gdb::unique_xmalloc_ptr<char> rewritten
+ = rewrite_source_path (fullname.c_str ());
+ if (rewritten != nullptr)
+ fullname = rewritten.release ();
+
+ if (file_matcher (fullname.c_str (), false))
+ return true;
+ }
+
+ return false;
+}
+
+/* See read.h. */
+
+void
+mapped_debug_line::map_filenames
+ (gdb::function_view<symbol_filename_ftype> fun,
+ bool need_fullname)
+{
+ for (line_header_up &lh : line_headers)
+ for (file_entry &fe : lh->file_names ())
+ {
+ const char *filename = fe.name;
+
+ if (!need_fullname)
+ {
+ fun (filename, nullptr);
+ continue;
+ }
+
+ const char *dirname = fe.include_dir (&*lh);
+ std::string fullname;
+
+ if (dirname == nullptr || IS_ABSOLUTE_PATH (filename))
+ fullname = filename;
+ else
+ fullname = std::string (dirname) + SLASH_STRING + filename;
+
+ gdb::unique_xmalloc_ptr<char> rewritten
+ = rewrite_source_path (fullname.c_str ());
+ if (rewritten != nullptr)
+ fullname = rewritten.release ();
+
+ fun (filename, fullname.c_str ());
+ }
+}
+
+/* See read.h. */
+
+gdb::array_view<const gdb_byte>
+mapped_debug_line::read_debug_line_separate
+ (const char *filename, std::unique_ptr<index_cache_resource> *resource)
+{
+ if (filename == nullptr)
+ return {};
+
+ try
+ {
+ line_resource_mmap *mmap_resource
+ = new line_resource_mmap (filename);
+
+ resource->reset (mmap_resource);
+
+ return gdb::array_view<const gdb_byte>
+ ((const gdb_byte *) mmap_resource->mapping.get (),
+ mmap_resource->mapping.size ());
+ }
+ catch (const gdb_exception &except)
+ {
+ exception_print (gdb_stderr, except);
+ }
+
+ return {};
+}
+
+/* See read.h. */
+
+bool
+mapped_debug_line::read_debug_line_from_debuginfod (objfile *objfile)
+{
+ const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
+ if (build_id == nullptr)
+ return false;
+
+ gdb::unique_xmalloc_ptr<char> line_path;
+ scoped_fd line_fd = debuginfod_section_query (build_id->data,
+ build_id->size,
+ bfd_get_filename
+ (objfile->obfd.get ()),
+ ".debug_line",
+ &line_path);
+
+ if (line_fd.get () < 0)
+ return false;
+
+ gdb::unique_xmalloc_ptr<char> line_str_path;
+ scoped_fd line_str_fd = debuginfod_section_query (build_id->data,
+ build_id->size,
+ bfd_get_filename
+ (objfile->obfd.get ()),
+ ".debug_line_str",
+ &line_str_path);
+
+ line_data = read_debug_line_separate (line_path.get (), &line_resource);
+ line_str_data = read_debug_line_separate (line_str_path.get (),
+ &line_str_resource);
+
+ const gdb_byte *line_ptr = line_data.data ();
+
+ while (line_ptr < line_data.data () + line_data.size ())
+ {
+ line_header_up lh
+ = dwarf_decode_line_header (objfile->obfd.get (),
+ line_data, line_str_data,
+ &line_ptr, false,
+ nullptr, nullptr);
+ line_headers.emplace_back (std::move (lh));
+ }
+
+ return true;
+}
+#endif /* !HAVE_SYS_MMAN_H */
+
+mapped_debug_line::mapped_debug_line (objfile *objfile)
+{
+ if (!read_debug_line_from_debuginfod (objfile))
+ line_headers.clear ();
+}
+
/* A helper for the "quick" functions which computes and caches the
real path for a given file name from the line table. */
diff --git a/gdb/dwarf2/read.h b/gdb/dwarf2/read.h
index 099d3d65c27..19d19a6edc1 100644
--- a/gdb/dwarf2/read.h
+++ b/gdb/dwarf2/read.h
@@ -33,6 +33,7 @@
#include "gdbsupport/hash_enum.h"
#include "gdbsupport/function-view.h"
#include "gdbsupport/packed.h"
+#include "dwarf2/line-header.h"
/* Hold 'maintenance (set|show) dwarf' commands. */
extern struct cmd_list_element *set_dwarf_cmdlist;
@@ -942,4 +943,40 @@ extern htab_up create_quick_file_names_table (unsigned int nr_initial_entries);
extern void read_full_dwarf_from_debuginfod (struct objfile *,
dwarf2_base_index_functions *);
+struct mapped_debug_line
+{
+ mapped_debug_line (objfile *objfile);
+
+ /* Return true if any of the mapped .debug_line's filenames match
+ FILE_MATCHER. */
+
+ bool contains_matching_filename
+ (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher);
+
+ /* Call FUN with each filename in this mapped .debug_line. Include
+ each file's fullname if NEED_FULLNAME is true. */
+
+ void map_filenames (gdb::function_view<symbol_filename_ftype> fun,
+ bool need_fullname);
+
+private:
+ std::vector<line_header_up> line_headers;
+
+ gdb::array_view<const gdb_byte> line_data;
+ gdb::array_view<const gdb_byte> line_str_data;
+
+ std::unique_ptr<index_cache_resource> line_resource;
+ std::unique_ptr<index_cache_resource> line_str_resource;
+
+ /* Download the .debug_line and .debug_line_str associated with OBJFILE
+ and populate line_headers. */
+
+ bool read_debug_line_from_debuginfod (objfile *objfile);
+
+ /* Initialize line_data and line_str_data with the .debug_line and
+ .debug_line_str downloaded read_debug_line_from_debuginfod. */
+
+ gdb::array_view<const gdb_byte> read_debug_line_separate
+ (const char *filename, std::unique_ptr<index_cache_resource> *resource);
+};
#endif /* DWARF2READ_H */
diff --git a/gdb/mi/mi-out.c b/gdb/mi/mi-out.c
index ff93d2cd448..8ad2c7f4497 100644
--- a/gdb/mi/mi-out.c
+++ b/gdb/mi/mi-out.c
@@ -277,7 +277,14 @@ mi_ui_out::do_progress_notify (const std::string &msg, const char *unit,
if (info.state == progress_update::START)
{
- gdb_printf ("%s...\n", msg.c_str ());
+ std::string prefix;
+ if (cur_prefix_state == prefix_state_t::NEWLINE_NEEDED)
+ {
+ prefix = "\n";
+ cur_prefix_state = prefix_state_t::NEWLINE_PRINTED;
+ }
+
+ gdb_printf ("%s...\n", (prefix + msg).c_str ());
info.state = progress_update::WORKING;
}
}
diff --git a/gdb/testsuite/gdb.debuginfod/section.exp b/gdb/testsuite/gdb.debuginfod/section.exp
index 35086d15705..a9dad927696 100644
--- a/gdb/testsuite/gdb.debuginfod/section.exp
+++ b/gdb/testsuite/gdb.debuginfod/section.exp
@@ -18,6 +18,7 @@
standard_testfile
load_lib debuginfod-support.exp
+load_lib completion-support.exp
clean_restart
require allow_debuginfod_tests
@@ -144,6 +145,32 @@ proc_with_prefix local_url { } {
# Hit the breakpoint.
set res ".*Breakpoint 2, libsection2_test.*\"In ${libfile2}\\\\n\".*"
gdb_test "c" $res "break continue"
+
+ clean_restart_with_prompt $binfile "line 1"
+
+ # List source file using .debug_line download.
+ set res ".*\.debug_line.*$lib_sl1.*21.*extern void libsection2_test.*"
+ gdb_test "list $libsrc1:21" $res "line 1 list"
+
+ clean_restart_with_prompt $binfile "line 2"
+
+ # Set breakpoint using .debug_line download.
+ set res ".*section \.debug_line for $lib_sl1.*Breakpoint 2 at.*$libsrc1.*"
+ gdb_test "br $libsrc1:41" $res "line 2 br"
+
+ # Continue to breakpoint.
+ set res "Breakpoint 2, libsection1_test.*\"Cancelling thread\\\\n\".*"
+ gdb_test "c" $res "line 2 continue"
+
+ # Check that download progress message is correctly formatted
+ # when printing threads.
+ set res ".*separate debug info for $lib_sl2\.\.\.\r\n.* 2 Thread.*"
+ gdb_test "info thr" $res "line thread"
+
+ clean_restart_with_prompt $binfile "autocomplete"
+
+ # Download debuginfo during autocompletion.
+ test_gdb_complete_tab_unique "br lib" ".*Downloading separate debug.*" ""
}
# Create CACHE and DB directories ready for debuginfod to use.
diff --git a/gdb/ui-out.c b/gdb/ui-out.c
index b3a2fb7f4e5..52d36d5bbf8 100644
--- a/gdb/ui-out.c
+++ b/gdb/ui-out.c
@@ -31,6 +31,9 @@
#include <memory>
#include <string>
+/* Current state of newline prefixing for progress update messages. */
+prefix_state_t cur_prefix_state = prefix_state_t::NEWLINE_OFF;
+
namespace {
/* A header of a ui_out_table. */
diff --git a/gdb/ui-out.h b/gdb/ui-out.h
index 5b6ddd55063..7e885cb1d38 100644
--- a/gdb/ui-out.h
+++ b/gdb/ui-out.h
@@ -296,6 +296,21 @@ class ui_out
BAR
};
+ /* Used to communicate the status of a newline prefix for the next progress
+ update message. */
+ enum prefix_state
+ {
+ /* Do not modify the next progress update message. */
+ NEWLINE_OFF,
+
+ /* The next progress update message should include a newline prefix. */
+ NEWLINE_NEEDED,
+
+ /* A newline prefix was included in a debuginfod progress update
+ message. */
+ NEWLINE_PRINTED
+ };
+
/* SHOULD_PRINT indicates whether something should be printed for a tty. */
progress_update ()
{
@@ -393,6 +408,11 @@ class ui_out
ui_out_level *current_level () const;
};
+typedef ui_out::progress_update::prefix_state prefix_state_t;
+
+/* Current state of the newline prefix. */
+extern prefix_state_t cur_prefix_state;
+
/* Start a new tuple or list on construction, and end it on
destruction. Normally this is used via the typedefs
ui_out_emit_tuple and ui_out_emit_list. */
--
2.43.0
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING][PATCH 3/3 v6] gdb/debuginfod: Add .debug_line downloading
2024-05-03 22:59 ` [PATCH 3/3 v6] gdb/debuginfod: Add .debug_line downloading Aaron Merey
@ 2024-05-17 14:10 ` Aaron Merey
2024-05-31 20:58 ` [PING*2][PATCH " Aaron Merey
0 siblings, 1 reply; 18+ messages in thread
From: Aaron Merey @ 2024-05-17 14:10 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
>
> This is a repost of v6 from
> https://sourceware.org/pipermail/gdb-patches/2024-January/205956.html
>
> ELF/DWARF section downloading allows gdb to download .gdb_index files in
> order to defer full debuginfo downloads. However .gdb_index does not
> contain any information regarding source filenames. When a gdb command
> includes a filename argument (ex. 'break main.c:50'), this results in
> the mass downloading of all deferred debuginfo so that gdb can search the
> debuginfo for matching source filenames. This can result in unnecessary
> downloads.
>
> To improve this, have gdb instead download each debuginfo's .debug_line
> (and .debug_line_str if using DWARF5) when executing these commands.
> Download full debuginfo only when its .debug_line contains a matching
> filename.
>
> Since the combined size of .debug_line and .debug_line_str is only about
> 1% the size of the corresponding debuginfo, significant time can be saved
> by checking these sections before choosing to download an entire debuginfo.
>
> This patch also redirects stdout and stderr of the debuginfod server
> used by testsuite/gdb.debuginfod tests to a server_log standard output
> file. While adding tests for this patch I ran into an issue where the
> test server would block when logging to stderr, presumably because the
> stderr buffer filled up and wasn't being read from. Redirecting the
> log to a file fixes this and also makes the server log more accessible
> when debugging test failures.
> ---
> gdb/cli-out.c | 11 +-
> gdb/completer.c | 18 +-
> gdb/dwarf2/line-header.c | 161 ++++++++++++------
> gdb/dwarf2/line-header.h | 10 ++
> gdb/dwarf2/read-gdb-index.c | 60 +++++++
> gdb/dwarf2/read.c | 208 +++++++++++++++++++++++
> gdb/dwarf2/read.h | 37 ++++
> gdb/mi/mi-out.c | 9 +-
> gdb/testsuite/gdb.debuginfod/section.exp | 27 +++
> gdb/ui-out.c | 3 +
> gdb/ui-out.h | 20 +++
> 11 files changed, 511 insertions(+), 53 deletions(-)
>
> diff --git a/gdb/cli-out.c b/gdb/cli-out.c
> index 1c303f09662..d65b69ebe01 100644
> --- a/gdb/cli-out.c
> +++ b/gdb/cli-out.c
> @@ -306,16 +306,23 @@ cli_ui_out::do_progress_notify (const std::string &msg,
>
> if (info.state == progress_update::START)
> {
> + std::string prefix;
> + if (cur_prefix_state == prefix_state_t::NEWLINE_NEEDED)
> + {
> + prefix = "\n";
> + cur_prefix_state = prefix_state_t::NEWLINE_PRINTED;
> + }
> +
> if (stream->isatty ()
> && current_ui->input_interactive_p ()
> && chars_per_line >= MIN_CHARS_PER_LINE)
> {
> - gdb_printf (stream, "%s\n", msg.c_str ());
> + gdb_printf (stream, "%s\n", (prefix + msg).c_str ());
> info.state = progress_update::BAR;
> }
> else
> {
> - gdb_printf (stream, "%s...\n", msg.c_str ());
> + gdb_printf (stream, "%s...\n", (prefix + msg).c_str ());
> info.state = progress_update::WORKING;
> }
> }
> diff --git a/gdb/completer.c b/gdb/completer.c
> index f1f44109bdc..708081cf280 100644
> --- a/gdb/completer.c
> +++ b/gdb/completer.c
> @@ -1354,6 +1354,10 @@ complete_line_internal_1 (completion_tracker &tracker,
> {
> /* We've recognized a full command. */
>
> + /* Disable pagination since responding to the pagination prompt
> + overwrites rl_line_buffer. */
> + scoped_restore pag_restore = make_scoped_restore (&pagination_enabled, false);
> +
> if (p == tmp_command + point)
> {
> /* There is no non-whitespace in the line beyond the
> @@ -1453,7 +1457,8 @@ complete_line_internal_1 (completion_tracker &tracker,
> }
>
> /* Wrapper around complete_line_internal_1 to handle
> - MAX_COMPLETIONS_REACHED_ERROR. */
> + MAX_COMPLETIONS_REACHED_ERROR and possible progress update
> + interactions. */
>
> static void
> complete_line_internal (completion_tracker &tracker,
> @@ -1461,6 +1466,11 @@ complete_line_internal (completion_tracker &tracker,
> const char *line_buffer, int point,
> complete_line_internal_reason reason)
> {
> + scoped_restore restore_prefix_state
> + = make_scoped_restore
> + (&cur_prefix_state,
> + ui_out::progress_update::prefix_state::NEWLINE_NEEDED);
> +
> try
> {
> complete_line_internal_1 (tracker, text, line_buffer, point, reason);
> @@ -1470,6 +1480,12 @@ complete_line_internal (completion_tracker &tracker,
> if (except.error != MAX_COMPLETIONS_REACHED_ERROR)
> throw;
> }
> +
> + /* If progress update messages printed, then the text being completed
> + needs to be printed again. */
> + if (cur_prefix_state
> + == ui_out::progress_update::prefix_state::NEWLINE_PRINTED)
> + rl_forced_update_display ();
> }
>
> /* See completer.h. */
> diff --git a/gdb/dwarf2/line-header.c b/gdb/dwarf2/line-header.c
> index eddb2ef7ae8..1d24386c0eb 100644
> --- a/gdb/dwarf2/line-header.c
> +++ b/gdb/dwarf2/line-header.c
> @@ -101,12 +101,15 @@ read_checked_initial_length_and_offset (bfd *abfd, const gdb_byte *buf,
> {
> LONGEST length = read_initial_length (abfd, buf, bytes_read);
>
> - gdb_assert (cu_header->initial_length_size == 4
> - || cu_header->initial_length_size == 8
> - || cu_header->initial_length_size == 12);
> + if (cu_header != nullptr)
> + {
> + gdb_assert (cu_header->initial_length_size == 4
> + || cu_header->initial_length_size == 8
> + || cu_header->initial_length_size == 12);
>
> - if (cu_header->initial_length_size != *bytes_read)
> - complaint (_("intermixed 32-bit and 64-bit DWARF sections"));
> + if (cu_header->initial_length_size != *bytes_read)
> + complaint (_("intermixed 32-bit and 64-bit DWARF sections"));
> + }
>
> *offset_size = (*bytes_read == 4) ? 4 : 8;
> return length;
> @@ -115,21 +118,27 @@ read_checked_initial_length_and_offset (bfd *abfd, const gdb_byte *buf,
> /* Read directory or file name entry format, starting with byte of
> format count entries, ULEB128 pairs of entry formats, ULEB128 of
> entries count and the entries themselves in the described entry
> - format. */
> + format.
> +
> + .debug_line and .debug_line_str sections are stored in LINE_BUFP
> + and LINE_STR_DATA respectively. */
>
> static void
> -read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> - const gdb_byte **bufp, struct line_header *lh,
> - unsigned int offset_size,
> - void (*callback) (struct line_header *lh,
> - const char *name,
> - dir_index d_index,
> - unsigned int mod_time,
> - unsigned int length))
> +read_formatted_entries
> + (bfd *abfd, const gdb_byte **line_bufp,
> + const gdb::array_view<const gdb_byte> line_str_data,
> + struct line_header *lh,
> + unsigned int offset_size,
> + void (*callback) (struct line_header *lh,
> + const char *name,
> + dir_index d_index,
> + unsigned int mod_time,
> + unsigned int length))
> {
> gdb_byte format_count, formati;
> ULONGEST data_count, datai;
> - const gdb_byte *buf = *bufp;
> + const gdb_byte *buf = *line_bufp;
> + const gdb_byte *str_buf = line_str_data.data ();
> const gdb_byte *format_header_data;
> unsigned int bytes_read;
>
> @@ -156,7 +165,7 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> ULONGEST content_type = read_unsigned_leb128 (abfd, format, &bytes_read);
> format += bytes_read;
>
> - ULONGEST form = read_unsigned_leb128 (abfd, format, &bytes_read);
> + ULONGEST form = read_unsigned_leb128 (abfd, format, &bytes_read);
> format += bytes_read;
>
> std::optional<const char *> string;
> @@ -171,12 +180,24 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
>
> case DW_FORM_line_strp:
> {
> - const char *str
> - = per_objfile->read_line_string (buf, offset_size);
> + if (line_str_data.empty ())
> + warning (_("Dwarf Error: DW_FORM_line_strp used without " \
> + "required section"));
> +
> + if (line_str_data.size () <= offset_size)
> + warning (_("Dwarf Error: DW_FORM_line_strp pointing outside " \
> + "of section .debug_line_str"));
> +
> + ULONGEST str_offset = read_offset (abfd, buf, offset_size);
> + const char *str;
> + if (str_buf[str_offset] == '\0')
> + str = nullptr;
> + else
> + str = (const char *) (str_buf + str_offset);
> string.emplace (str);
> buf += offset_size;
> + break;
> }
> - break;
>
> case DW_FORM_data1:
> uint.emplace (read_1_byte (abfd, buf));
> @@ -247,28 +268,30 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> callback (lh, fe.name, fe.d_index, fe.mod_time, fe.length);
> }
>
> - *bufp = buf;
> + *line_bufp = buf;
> }
>
> /* See line-header.h. */
>
> line_header_up
> -dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> - dwarf2_per_objfile *per_objfile,
> - struct dwarf2_section_info *section,
> - const struct comp_unit_head *cu_header,
> - const char *comp_dir)
> +dwarf_decode_line_header (bfd *abfd,
> + gdb::array_view<const gdb_byte> line_data,
> + gdb::array_view<const gdb_byte> line_str_data,
> + const gdb_byte **debug_line_ptr,
> + bool is_dwz,
> + const struct comp_unit_head *cu_header,
> + const char *comp_dir)
> {
> - const gdb_byte *line_ptr;
> + const gdb_byte *line_ptr, *buf;
> unsigned int bytes_read, offset_size;
> int i;
> const char *cur_dir, *cur_file;
>
> - bfd *abfd = section->get_bfd_owner ();
> + buf = *debug_line_ptr;
>
> /* Make sure that at least there's room for the total_length field.
> That could be 12 bytes long, but we're just going to fudge that. */
> - if (to_underlying (sect_off) + 4 >= section->size)
> + if (buf + 4 >= line_data.data () + line_data.size ())
> {
> dwarf2_statement_list_fits_in_line_number_section_complaint ();
> return 0;
> @@ -276,24 +299,26 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
>
> line_header_up lh (new line_header (comp_dir));
>
> - lh->sect_off = sect_off;
> + lh->sect_off = (sect_offset) (buf - line_data.data ());
> lh->offset_in_dwz = is_dwz;
>
> - line_ptr = section->buffer + to_underlying (sect_off);
> + line_ptr = buf;
>
> /* Read in the header. */
> LONGEST unit_length
> - = read_checked_initial_length_and_offset (abfd, line_ptr, cu_header,
> + = read_checked_initial_length_and_offset (abfd, buf, cu_header,
> &bytes_read, &offset_size);
> - line_ptr += bytes_read;
>
> - const gdb_byte *start_here = line_ptr;
> + line_ptr += bytes_read;
>
> - if (line_ptr + unit_length > (section->buffer + section->size))
> + if (line_ptr + unit_length > buf + line_data.size ())
> {
> dwarf2_statement_list_fits_in_line_number_section_complaint ();
> return 0;
> }
> +
> + const gdb_byte *start_here = line_ptr;
> +
> lh->statement_program_end = start_here + unit_length;
> lh->version = read_2_bytes (abfd, line_ptr);
> line_ptr += 2;
> @@ -302,7 +327,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> /* This is a version we don't understand. The format could have
> changed in ways we don't handle properly so just punt. */
> complaint (_("unsupported version in .debug_line section"));
> - return NULL;
> + return nullptr;
> }
> if (lh->version >= 5)
> {
> @@ -319,13 +344,14 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> complaint (_("unsupported segment selector size %u "
> "in .debug_line section"),
> segment_selector_size);
> - return NULL;
> + return nullptr;
> }
> }
>
> LONGEST header_length = read_offset (abfd, line_ptr, offset_size);
> line_ptr += offset_size;
> lh->statement_program_start = line_ptr + header_length;
> +
> lh->minimum_instruction_length = read_1_byte (abfd, line_ptr);
> line_ptr += 1;
>
> @@ -346,12 +372,16 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
>
> lh->default_is_stmt = read_1_byte (abfd, line_ptr);
> line_ptr += 1;
> +
> lh->line_base = read_1_signed_byte (abfd, line_ptr);
> line_ptr += 1;
> +
> lh->line_range = read_1_byte (abfd, line_ptr);
> line_ptr += 1;
> +
> lh->opcode_base = read_1_byte (abfd, line_ptr);
> line_ptr += 1;
> +
> lh->standard_opcode_lengths.reset (new unsigned char[lh->opcode_base]);
>
> lh->standard_opcode_lengths[0] = 1; /* This should never be used anyway. */
> @@ -364,21 +394,23 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> if (lh->version >= 5)
> {
> /* Read directory table. */
> - read_formatted_entries (per_objfile, abfd, &line_ptr, lh.get (),
> - offset_size,
> - [] (struct line_header *header, const char *name,
> - dir_index d_index, unsigned int mod_time,
> - unsigned int length)
> + read_formatted_entries
> + (abfd, &line_ptr, line_str_data,
> + lh.get (), offset_size,
> + [] (struct line_header *header, const char *name,
> + dir_index d_index, unsigned int mod_time,
> + unsigned int length)
> {
> header->add_include_dir (name);
> });
>
> /* Read file name table. */
> - read_formatted_entries (per_objfile, abfd, &line_ptr, lh.get (),
> - offset_size,
> - [] (struct line_header *header, const char *name,
> - dir_index d_index, unsigned int mod_time,
> - unsigned int length)
> + read_formatted_entries
> + (abfd, &line_ptr, line_str_data,
> + lh.get (), offset_size,
> + [] (struct line_header *header, const char *name,
> + dir_index d_index, unsigned int mod_time,
> + unsigned int length)
> {
> header->add_file_name (name, d_index, mod_time, length);
> });
> @@ -386,7 +418,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> else
> {
> /* Read directory table. */
> - while ((cur_dir = read_direct_string (abfd, line_ptr, &bytes_read)) != NULL)
> + while ((cur_dir = read_direct_string (abfd, line_ptr, &bytes_read)) != nullptr)
> {
> line_ptr += bytes_read;
> lh->add_include_dir (cur_dir);
> @@ -394,7 +426,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> line_ptr += bytes_read;
>
> /* Read file name table. */
> - while ((cur_file = read_direct_string (abfd, line_ptr, &bytes_read)) != NULL)
> + while ((cur_file = read_direct_string (abfd, line_ptr, &bytes_read)) != nullptr)
> {
> unsigned int mod_time, length;
> dir_index d_index;
> @@ -412,9 +444,40 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> line_ptr += bytes_read;
> }
>
> - if (line_ptr > (section->buffer + section->size))
> + if (line_ptr > (buf + line_data.size ()))
> complaint (_("line number info header doesn't "
> "fit in `.debug_line' section"));
>
> + *debug_line_ptr += unit_length + offset_size;
> return lh;
> }
> +
> +line_header_up
> +dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> + dwarf2_per_objfile *per_objfile,
> + struct dwarf2_section_info *section,
> + const struct comp_unit_head *cu_header,
> + const char *comp_dir)
> +{
> + struct objfile *objfile = per_objfile->objfile;
> + struct dwarf2_per_bfd *per_bfd = per_objfile->per_bfd;
> +
> + /* Read .debug_line. */
> + dwarf2_section_info *line_sec = &per_bfd->line;
> + bfd_size_type line_size = line_sec->get_size (objfile);
> +
> + gdb::array_view<const gdb_byte> line (line_sec->buffer, line_size);
> +
> + /* Read .debug_line_str. */
> + dwarf2_section_info *line_str_sec = &per_bfd->line_str;
> + bfd_size_type line_str_size = line_str_sec->get_size (objfile);
> +
> + gdb::array_view<const gdb_byte> line_str (line_str_sec->buffer,
> + line_str_size);
> +
> + const gdb_byte *line_ptr = line.data () + to_underlying (sect_off);
> +
> + return dwarf_decode_line_header
> + (per_bfd->obfd, line, line_str, &line_ptr,
> + is_dwz, cu_header, comp_dir);
> +}
> diff --git a/gdb/dwarf2/line-header.h b/gdb/dwarf2/line-header.h
> index c068dff70a3..45011882321 100644
> --- a/gdb/dwarf2/line-header.h
> +++ b/gdb/dwarf2/line-header.h
> @@ -217,4 +217,14 @@ extern line_header_up dwarf_decode_line_header
> struct dwarf2_section_info *section, const struct comp_unit_head *cu_header,
> const char *comp_dir);
>
> +/* Like above but the .debug_line and .debug_line_str are stored in
> + LINE_DATA and LINE_STR_DATA. *DEBUG_LINE_PTR should point to a
> + statement program header within LINE_DATA. */
> +
> +extern line_header_up dwarf_decode_line_header
> + (bfd *abfd, gdb::array_view<const gdb_byte> line_data,
> + gdb::array_view<const gdb_byte> line_str_data,
> + const gdb_byte **debug_line_ptr, bool is_dwz,
> + const comp_unit_head *cu_header, const char *comp_dir);
> +
> #endif /* DWARF2_LINE_HEADER_H */
> diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
> index 97b8137aabf..c52099d5f6a 100644
> --- a/gdb/dwarf2/read-gdb-index.c
> +++ b/gdb/dwarf2/read-gdb-index.c
> @@ -131,6 +131,9 @@ struct mapped_gdb_index final : public mapped_index_base
> }
> };
>
> +struct mapped_debug_line;
> +typedef std::unique_ptr<mapped_debug_line> mapped_debug_line_up;
> +
> struct dwarf2_gdb_index : public dwarf2_base_index_functions
> {
> /* This dumps minimal information about the index.
> @@ -160,6 +163,15 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> block_search_flags search_flags,
> domain_search_flags domain);
>
> + /* If OBJFILE's debuginfo download has been deferred, use a mapped_debug_line
> + to generate filenames.
> +
> + Otherwise call dwarf2_base_index_functions::map_symbol_filenames. */
> +
> + void map_symbol_filenames (struct objfile *objfile,
> + gdb::function_view<symbol_filename_ftype> fun,
> + bool need_fullname) override;
> +
> /* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
> debuginfo if necessary. */
> void expand_all_symtabs (struct objfile *objfile) override;
> @@ -167,6 +179,15 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> /* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
> debuginfo if necessary. */
> struct symtab *find_last_source_symtab (struct objfile *objfile) override;
> +
> + /* Filename information related to this .gdb_index. */
> + mapped_debug_line_up mdl;
> +
> + /* Return true if any of the filenames in this .gdb_index's .debug_line
> + mapping match FILE_MATCHER. Initializes the mapping if necessary. */
> + bool filename_in_debug_line
> + (objfile *objfile,
> + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher);
> };
>
> void
> @@ -202,6 +223,30 @@ dwarf2_gdb_index::find_last_source_symtab (struct objfile *objfile)
> }
> }
>
> +void
> +dwarf2_gdb_index::map_symbol_filenames
> + (struct objfile *objfile,
> + gdb::function_view<symbol_filename_ftype> fun,
> + bool need_fullname)
> +{
> + try
> + {
> + dwarf2_base_index_functions::map_symbol_filenames (objfile, fun,
> + need_fullname);
> + }
> + catch (const gdb_exception &e)
> + {
> + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> + exception_print (gdb_stderr, e);
> + else
> + {
> + if (mdl == nullptr)
> + mdl.reset (new mapped_debug_line (objfile));
> + mdl->map_filenames (fun, need_fullname);
> + }
> + }
> +}
> +
> /* This dumps minimal information about the index.
> It is called via "mt print objfiles".
> One use is to verify .gdb_index has been loaded by the
> @@ -365,6 +410,17 @@ dwarf2_gdb_index::do_expand_symtabs_matching
> return result;
> }
>
> +bool
> +dwarf2_gdb_index::filename_in_debug_line
> + (objfile *objfile,
> + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> +{
> + if (mdl == nullptr)
> + mdl.reset (new mapped_debug_line (objfile));
> +
> + return mdl->contains_matching_filename (file_matcher);
> +}
> +
> bool
> dwarf2_gdb_index::expand_symtabs_matching
> (struct objfile *objfile,
> @@ -392,6 +448,10 @@ dwarf2_gdb_index::expand_symtabs_matching
> return false;
> }
>
> + if (file_matcher != nullptr
> + && !filename_in_debug_line (objfile, file_matcher))
> + return true;
> +
> read_full_dwarf_from_debuginfod (objfile, this);
> return true;
> }
> diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
> index 4461c76347e..ca219c9b910 100644
> --- a/gdb/dwarf2/read.c
> +++ b/gdb/dwarf2/read.c
> @@ -81,6 +81,7 @@
> #include <optional>
> #include "gdbsupport/underlying.h"
> #include "gdbsupport/hash_enum.h"
> +#include "gdbsupport/scoped_mmap.h"
> #include "filename-seen-cache.h"
> #include "producer.h"
> #include <fcntl.h>
> @@ -1967,6 +1968,213 @@ dw2_get_file_names (dwarf2_per_cu_data *this_cu,
> return this_cu->file_names;
> }
>
> +#if !HAVE_SYS_MMAN_H
> +
> +bool
> +mapped_debug_line::contains_matching_filename
> + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> +{
> + return false;
> +}
> +
> +gdb::array_view<const gdb_byte>
> +mapped_debug_line::read_debug_line_separate
> + (const char *filename, std::unique_ptr<index_cache_resource> *resource)
> +{
> + return {};
> +}
> +
> +bool
> +mapped_debug_line::read_debug_line_from_debuginfod (objfile *objfile)
> +{
> + return false;
> +}
> +
> +void
> +mapped_debug_line::map_filenames
> + (gdb::function_view<symbol_filename_ftype> fun,
> + bool need_fullname)
> +{
> + return;
> +}
> +
> +#else /* !HAVE_SYS_MMAN_H */
> +
> +struct line_resource_mmap final : public index_cache_resource
> +{
> + /* Try to mmap FILENAME. Throw an exception on failure, including if the
> + file doesn't exist. */
> + line_resource_mmap (const char *filename)
> + : mapping (mmap_file (filename))
> + {}
> +
> + scoped_mmap mapping;
> +};
> +
> +/* See read.h. */
> +
> +bool
> +mapped_debug_line::contains_matching_filename
> + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> +{
> + for (line_header_up &lh : line_headers)
> + for (file_entry &fe : lh->file_names ())
> + {
> + const char *filename = fe.name;
> +
> + if (file_matcher (fe.name, false))
> + return true;
> +
> + bool basename_match = file_matcher (lbasename (fe.name), true);
> +
> + if (!basenames_may_differ && !basename_match)
> + continue;
> +
> + /* DW_AT_comp_dir is not explicitly mentioned in the .debug_line
> + until DWARF5. Since we don't have access to the CU at this
> + point we just check for a partial match on the filename.
> + If there is a match, the full debuginfo will be downloaded
> + ane the match will be re-evalute with DW_AT_comp_dir. */
> + if (lh->version < 5 && fe.d_index == 0)
> + return basename_match;
> +
> + const char *dirname = fe.include_dir (&*lh);
> + std::string fullname;
> +
> + if (dirname == nullptr || IS_ABSOLUTE_PATH (filename))
> + fullname = filename;
> + else
> + fullname = std::string (dirname) + SLASH_STRING + filename;
> +
> + gdb::unique_xmalloc_ptr<char> rewritten
> + = rewrite_source_path (fullname.c_str ());
> + if (rewritten != nullptr)
> + fullname = rewritten.release ();
> +
> + if (file_matcher (fullname.c_str (), false))
> + return true;
> + }
> +
> + return false;
> +}
> +
> +/* See read.h. */
> +
> +void
> +mapped_debug_line::map_filenames
> + (gdb::function_view<symbol_filename_ftype> fun,
> + bool need_fullname)
> +{
> + for (line_header_up &lh : line_headers)
> + for (file_entry &fe : lh->file_names ())
> + {
> + const char *filename = fe.name;
> +
> + if (!need_fullname)
> + {
> + fun (filename, nullptr);
> + continue;
> + }
> +
> + const char *dirname = fe.include_dir (&*lh);
> + std::string fullname;
> +
> + if (dirname == nullptr || IS_ABSOLUTE_PATH (filename))
> + fullname = filename;
> + else
> + fullname = std::string (dirname) + SLASH_STRING + filename;
> +
> + gdb::unique_xmalloc_ptr<char> rewritten
> + = rewrite_source_path (fullname.c_str ());
> + if (rewritten != nullptr)
> + fullname = rewritten.release ();
> +
> + fun (filename, fullname.c_str ());
> + }
> +}
> +
> +/* See read.h. */
> +
> +gdb::array_view<const gdb_byte>
> +mapped_debug_line::read_debug_line_separate
> + (const char *filename, std::unique_ptr<index_cache_resource> *resource)
> +{
> + if (filename == nullptr)
> + return {};
> +
> + try
> + {
> + line_resource_mmap *mmap_resource
> + = new line_resource_mmap (filename);
> +
> + resource->reset (mmap_resource);
> +
> + return gdb::array_view<const gdb_byte>
> + ((const gdb_byte *) mmap_resource->mapping.get (),
> + mmap_resource->mapping.size ());
> + }
> + catch (const gdb_exception &except)
> + {
> + exception_print (gdb_stderr, except);
> + }
> +
> + return {};
> +}
> +
> +/* See read.h. */
> +
> +bool
> +mapped_debug_line::read_debug_line_from_debuginfod (objfile *objfile)
> +{
> + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> + if (build_id == nullptr)
> + return false;
> +
> + gdb::unique_xmalloc_ptr<char> line_path;
> + scoped_fd line_fd = debuginfod_section_query (build_id->data,
> + build_id->size,
> + bfd_get_filename
> + (objfile->obfd.get ()),
> + ".debug_line",
> + &line_path);
> +
> + if (line_fd.get () < 0)
> + return false;
> +
> + gdb::unique_xmalloc_ptr<char> line_str_path;
> + scoped_fd line_str_fd = debuginfod_section_query (build_id->data,
> + build_id->size,
> + bfd_get_filename
> + (objfile->obfd.get ()),
> + ".debug_line_str",
> + &line_str_path);
> +
> + line_data = read_debug_line_separate (line_path.get (), &line_resource);
> + line_str_data = read_debug_line_separate (line_str_path.get (),
> + &line_str_resource);
> +
> + const gdb_byte *line_ptr = line_data.data ();
> +
> + while (line_ptr < line_data.data () + line_data.size ())
> + {
> + line_header_up lh
> + = dwarf_decode_line_header (objfile->obfd.get (),
> + line_data, line_str_data,
> + &line_ptr, false,
> + nullptr, nullptr);
> + line_headers.emplace_back (std::move (lh));
> + }
> +
> + return true;
> +}
> +#endif /* !HAVE_SYS_MMAN_H */
> +
> +mapped_debug_line::mapped_debug_line (objfile *objfile)
> +{
> + if (!read_debug_line_from_debuginfod (objfile))
> + line_headers.clear ();
> +}
> +
> /* A helper for the "quick" functions which computes and caches the
> real path for a given file name from the line table. */
>
> diff --git a/gdb/dwarf2/read.h b/gdb/dwarf2/read.h
> index 099d3d65c27..19d19a6edc1 100644
> --- a/gdb/dwarf2/read.h
> +++ b/gdb/dwarf2/read.h
> @@ -33,6 +33,7 @@
> #include "gdbsupport/hash_enum.h"
> #include "gdbsupport/function-view.h"
> #include "gdbsupport/packed.h"
> +#include "dwarf2/line-header.h"
>
> /* Hold 'maintenance (set|show) dwarf' commands. */
> extern struct cmd_list_element *set_dwarf_cmdlist;
> @@ -942,4 +943,40 @@ extern htab_up create_quick_file_names_table (unsigned int nr_initial_entries);
> extern void read_full_dwarf_from_debuginfod (struct objfile *,
> dwarf2_base_index_functions *);
>
> +struct mapped_debug_line
> +{
> + mapped_debug_line (objfile *objfile);
> +
> + /* Return true if any of the mapped .debug_line's filenames match
> + FILE_MATCHER. */
> +
> + bool contains_matching_filename
> + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher);
> +
> + /* Call FUN with each filename in this mapped .debug_line. Include
> + each file's fullname if NEED_FULLNAME is true. */
> +
> + void map_filenames (gdb::function_view<symbol_filename_ftype> fun,
> + bool need_fullname);
> +
> +private:
> + std::vector<line_header_up> line_headers;
> +
> + gdb::array_view<const gdb_byte> line_data;
> + gdb::array_view<const gdb_byte> line_str_data;
> +
> + std::unique_ptr<index_cache_resource> line_resource;
> + std::unique_ptr<index_cache_resource> line_str_resource;
> +
> + /* Download the .debug_line and .debug_line_str associated with OBJFILE
> + and populate line_headers. */
> +
> + bool read_debug_line_from_debuginfod (objfile *objfile);
> +
> + /* Initialize line_data and line_str_data with the .debug_line and
> + .debug_line_str downloaded read_debug_line_from_debuginfod. */
> +
> + gdb::array_view<const gdb_byte> read_debug_line_separate
> + (const char *filename, std::unique_ptr<index_cache_resource> *resource);
> +};
> #endif /* DWARF2READ_H */
> diff --git a/gdb/mi/mi-out.c b/gdb/mi/mi-out.c
> index ff93d2cd448..8ad2c7f4497 100644
> --- a/gdb/mi/mi-out.c
> +++ b/gdb/mi/mi-out.c
> @@ -277,7 +277,14 @@ mi_ui_out::do_progress_notify (const std::string &msg, const char *unit,
>
> if (info.state == progress_update::START)
> {
> - gdb_printf ("%s...\n", msg.c_str ());
> + std::string prefix;
> + if (cur_prefix_state == prefix_state_t::NEWLINE_NEEDED)
> + {
> + prefix = "\n";
> + cur_prefix_state = prefix_state_t::NEWLINE_PRINTED;
> + }
> +
> + gdb_printf ("%s...\n", (prefix + msg).c_str ());
> info.state = progress_update::WORKING;
> }
> }
> diff --git a/gdb/testsuite/gdb.debuginfod/section.exp b/gdb/testsuite/gdb.debuginfod/section.exp
> index 35086d15705..a9dad927696 100644
> --- a/gdb/testsuite/gdb.debuginfod/section.exp
> +++ b/gdb/testsuite/gdb.debuginfod/section.exp
> @@ -18,6 +18,7 @@
> standard_testfile
>
> load_lib debuginfod-support.exp
> +load_lib completion-support.exp
>
> clean_restart
> require allow_debuginfod_tests
> @@ -144,6 +145,32 @@ proc_with_prefix local_url { } {
> # Hit the breakpoint.
> set res ".*Breakpoint 2, libsection2_test.*\"In ${libfile2}\\\\n\".*"
> gdb_test "c" $res "break continue"
> +
> + clean_restart_with_prompt $binfile "line 1"
> +
> + # List source file using .debug_line download.
> + set res ".*\.debug_line.*$lib_sl1.*21.*extern void libsection2_test.*"
> + gdb_test "list $libsrc1:21" $res "line 1 list"
> +
> + clean_restart_with_prompt $binfile "line 2"
> +
> + # Set breakpoint using .debug_line download.
> + set res ".*section \.debug_line for $lib_sl1.*Breakpoint 2 at.*$libsrc1.*"
> + gdb_test "br $libsrc1:41" $res "line 2 br"
> +
> + # Continue to breakpoint.
> + set res "Breakpoint 2, libsection1_test.*\"Cancelling thread\\\\n\".*"
> + gdb_test "c" $res "line 2 continue"
> +
> + # Check that download progress message is correctly formatted
> + # when printing threads.
> + set res ".*separate debug info for $lib_sl2\.\.\.\r\n.* 2 Thread.*"
> + gdb_test "info thr" $res "line thread"
> +
> + clean_restart_with_prompt $binfile "autocomplete"
> +
> + # Download debuginfo during autocompletion.
> + test_gdb_complete_tab_unique "br lib" ".*Downloading separate debug.*" ""
> }
>
> # Create CACHE and DB directories ready for debuginfod to use.
> diff --git a/gdb/ui-out.c b/gdb/ui-out.c
> index b3a2fb7f4e5..52d36d5bbf8 100644
> --- a/gdb/ui-out.c
> +++ b/gdb/ui-out.c
> @@ -31,6 +31,9 @@
> #include <memory>
> #include <string>
>
> +/* Current state of newline prefixing for progress update messages. */
> +prefix_state_t cur_prefix_state = prefix_state_t::NEWLINE_OFF;
> +
> namespace {
>
> /* A header of a ui_out_table. */
> diff --git a/gdb/ui-out.h b/gdb/ui-out.h
> index 5b6ddd55063..7e885cb1d38 100644
> --- a/gdb/ui-out.h
> +++ b/gdb/ui-out.h
> @@ -296,6 +296,21 @@ class ui_out
> BAR
> };
>
> + /* Used to communicate the status of a newline prefix for the next progress
> + update message. */
> + enum prefix_state
> + {
> + /* Do not modify the next progress update message. */
> + NEWLINE_OFF,
> +
> + /* The next progress update message should include a newline prefix. */
> + NEWLINE_NEEDED,
> +
> + /* A newline prefix was included in a debuginfod progress update
> + message. */
> + NEWLINE_PRINTED
> + };
> +
> /* SHOULD_PRINT indicates whether something should be printed for a tty. */
> progress_update ()
> {
> @@ -393,6 +408,11 @@ class ui_out
> ui_out_level *current_level () const;
> };
>
> +typedef ui_out::progress_update::prefix_state prefix_state_t;
> +
> +/* Current state of the newline prefix. */
> +extern prefix_state_t cur_prefix_state;
> +
> /* Start a new tuple or list on construction, and end it on
> destruction. Normally this is used via the typedefs
> ui_out_emit_tuple and ui_out_emit_list. */
> --
> 2.43.0
>
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING*2][PATCH 3/3 v6] gdb/debuginfod: Add .debug_line downloading
2024-05-17 14:10 ` [PING][PATCH " Aaron Merey
@ 2024-05-31 20:58 ` Aaron Merey
2024-06-07 20:49 ` [PING*3][PATCH " Aaron Merey
2024-06-14 18:24 ` Aaron Merey
0 siblings, 2 replies; 18+ messages in thread
From: Aaron Merey @ 2024-05-31 20:58 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 17, 2024 at 10:10 AM Aaron Merey <amerey@redhat.com> wrote:
>
> Ping
>
> Thanks,
> Aaron
>
> On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
> >
> > This is a repost of v6 from
> > https://sourceware.org/pipermail/gdb-patches/2024-January/205956.html
> >
> > ELF/DWARF section downloading allows gdb to download .gdb_index files in
> > order to defer full debuginfo downloads. However .gdb_index does not
> > contain any information regarding source filenames. When a gdb command
> > includes a filename argument (ex. 'break main.c:50'), this results in
> > the mass downloading of all deferred debuginfo so that gdb can search the
> > debuginfo for matching source filenames. This can result in unnecessary
> > downloads.
> >
> > To improve this, have gdb instead download each debuginfo's .debug_line
> > (and .debug_line_str if using DWARF5) when executing these commands.
> > Download full debuginfo only when its .debug_line contains a matching
> > filename.
> >
> > Since the combined size of .debug_line and .debug_line_str is only about
> > 1% the size of the corresponding debuginfo, significant time can be saved
> > by checking these sections before choosing to download an entire debuginfo.
> >
> > This patch also redirects stdout and stderr of the debuginfod server
> > used by testsuite/gdb.debuginfod tests to a server_log standard output
> > file. While adding tests for this patch I ran into an issue where the
> > test server would block when logging to stderr, presumably because the
> > stderr buffer filled up and wasn't being read from. Redirecting the
> > log to a file fixes this and also makes the server log more accessible
> > when debugging test failures.
> > ---
> > gdb/cli-out.c | 11 +-
> > gdb/completer.c | 18 +-
> > gdb/dwarf2/line-header.c | 161 ++++++++++++------
> > gdb/dwarf2/line-header.h | 10 ++
> > gdb/dwarf2/read-gdb-index.c | 60 +++++++
> > gdb/dwarf2/read.c | 208 +++++++++++++++++++++++
> > gdb/dwarf2/read.h | 37 ++++
> > gdb/mi/mi-out.c | 9 +-
> > gdb/testsuite/gdb.debuginfod/section.exp | 27 +++
> > gdb/ui-out.c | 3 +
> > gdb/ui-out.h | 20 +++
> > 11 files changed, 511 insertions(+), 53 deletions(-)
> >
> > diff --git a/gdb/cli-out.c b/gdb/cli-out.c
> > index 1c303f09662..d65b69ebe01 100644
> > --- a/gdb/cli-out.c
> > +++ b/gdb/cli-out.c
> > @@ -306,16 +306,23 @@ cli_ui_out::do_progress_notify (const std::string &msg,
> >
> > if (info.state == progress_update::START)
> > {
> > + std::string prefix;
> > + if (cur_prefix_state == prefix_state_t::NEWLINE_NEEDED)
> > + {
> > + prefix = "\n";
> > + cur_prefix_state = prefix_state_t::NEWLINE_PRINTED;
> > + }
> > +
> > if (stream->isatty ()
> > && current_ui->input_interactive_p ()
> > && chars_per_line >= MIN_CHARS_PER_LINE)
> > {
> > - gdb_printf (stream, "%s\n", msg.c_str ());
> > + gdb_printf (stream, "%s\n", (prefix + msg).c_str ());
> > info.state = progress_update::BAR;
> > }
> > else
> > {
> > - gdb_printf (stream, "%s...\n", msg.c_str ());
> > + gdb_printf (stream, "%s...\n", (prefix + msg).c_str ());
> > info.state = progress_update::WORKING;
> > }
> > }
> > diff --git a/gdb/completer.c b/gdb/completer.c
> > index f1f44109bdc..708081cf280 100644
> > --- a/gdb/completer.c
> > +++ b/gdb/completer.c
> > @@ -1354,6 +1354,10 @@ complete_line_internal_1 (completion_tracker &tracker,
> > {
> > /* We've recognized a full command. */
> >
> > + /* Disable pagination since responding to the pagination prompt
> > + overwrites rl_line_buffer. */
> > + scoped_restore pag_restore = make_scoped_restore (&pagination_enabled, false);
> > +
> > if (p == tmp_command + point)
> > {
> > /* There is no non-whitespace in the line beyond the
> > @@ -1453,7 +1457,8 @@ complete_line_internal_1 (completion_tracker &tracker,
> > }
> >
> > /* Wrapper around complete_line_internal_1 to handle
> > - MAX_COMPLETIONS_REACHED_ERROR. */
> > + MAX_COMPLETIONS_REACHED_ERROR and possible progress update
> > + interactions. */
> >
> > static void
> > complete_line_internal (completion_tracker &tracker,
> > @@ -1461,6 +1466,11 @@ complete_line_internal (completion_tracker &tracker,
> > const char *line_buffer, int point,
> > complete_line_internal_reason reason)
> > {
> > + scoped_restore restore_prefix_state
> > + = make_scoped_restore
> > + (&cur_prefix_state,
> > + ui_out::progress_update::prefix_state::NEWLINE_NEEDED);
> > +
> > try
> > {
> > complete_line_internal_1 (tracker, text, line_buffer, point, reason);
> > @@ -1470,6 +1480,12 @@ complete_line_internal (completion_tracker &tracker,
> > if (except.error != MAX_COMPLETIONS_REACHED_ERROR)
> > throw;
> > }
> > +
> > + /* If progress update messages printed, then the text being completed
> > + needs to be printed again. */
> > + if (cur_prefix_state
> > + == ui_out::progress_update::prefix_state::NEWLINE_PRINTED)
> > + rl_forced_update_display ();
> > }
> >
> > /* See completer.h. */
> > diff --git a/gdb/dwarf2/line-header.c b/gdb/dwarf2/line-header.c
> > index eddb2ef7ae8..1d24386c0eb 100644
> > --- a/gdb/dwarf2/line-header.c
> > +++ b/gdb/dwarf2/line-header.c
> > @@ -101,12 +101,15 @@ read_checked_initial_length_and_offset (bfd *abfd, const gdb_byte *buf,
> > {
> > LONGEST length = read_initial_length (abfd, buf, bytes_read);
> >
> > - gdb_assert (cu_header->initial_length_size == 4
> > - || cu_header->initial_length_size == 8
> > - || cu_header->initial_length_size == 12);
> > + if (cu_header != nullptr)
> > + {
> > + gdb_assert (cu_header->initial_length_size == 4
> > + || cu_header->initial_length_size == 8
> > + || cu_header->initial_length_size == 12);
> >
> > - if (cu_header->initial_length_size != *bytes_read)
> > - complaint (_("intermixed 32-bit and 64-bit DWARF sections"));
> > + if (cu_header->initial_length_size != *bytes_read)
> > + complaint (_("intermixed 32-bit and 64-bit DWARF sections"));
> > + }
> >
> > *offset_size = (*bytes_read == 4) ? 4 : 8;
> > return length;
> > @@ -115,21 +118,27 @@ read_checked_initial_length_and_offset (bfd *abfd, const gdb_byte *buf,
> > /* Read directory or file name entry format, starting with byte of
> > format count entries, ULEB128 pairs of entry formats, ULEB128 of
> > entries count and the entries themselves in the described entry
> > - format. */
> > + format.
> > +
> > + .debug_line and .debug_line_str sections are stored in LINE_BUFP
> > + and LINE_STR_DATA respectively. */
> >
> > static void
> > -read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > - const gdb_byte **bufp, struct line_header *lh,
> > - unsigned int offset_size,
> > - void (*callback) (struct line_header *lh,
> > - const char *name,
> > - dir_index d_index,
> > - unsigned int mod_time,
> > - unsigned int length))
> > +read_formatted_entries
> > + (bfd *abfd, const gdb_byte **line_bufp,
> > + const gdb::array_view<const gdb_byte> line_str_data,
> > + struct line_header *lh,
> > + unsigned int offset_size,
> > + void (*callback) (struct line_header *lh,
> > + const char *name,
> > + dir_index d_index,
> > + unsigned int mod_time,
> > + unsigned int length))
> > {
> > gdb_byte format_count, formati;
> > ULONGEST data_count, datai;
> > - const gdb_byte *buf = *bufp;
> > + const gdb_byte *buf = *line_bufp;
> > + const gdb_byte *str_buf = line_str_data.data ();
> > const gdb_byte *format_header_data;
> > unsigned int bytes_read;
> >
> > @@ -156,7 +165,7 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > ULONGEST content_type = read_unsigned_leb128 (abfd, format, &bytes_read);
> > format += bytes_read;
> >
> > - ULONGEST form = read_unsigned_leb128 (abfd, format, &bytes_read);
> > + ULONGEST form = read_unsigned_leb128 (abfd, format, &bytes_read);
> > format += bytes_read;
> >
> > std::optional<const char *> string;
> > @@ -171,12 +180,24 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> >
> > case DW_FORM_line_strp:
> > {
> > - const char *str
> > - = per_objfile->read_line_string (buf, offset_size);
> > + if (line_str_data.empty ())
> > + warning (_("Dwarf Error: DW_FORM_line_strp used without " \
> > + "required section"));
> > +
> > + if (line_str_data.size () <= offset_size)
> > + warning (_("Dwarf Error: DW_FORM_line_strp pointing outside " \
> > + "of section .debug_line_str"));
> > +
> > + ULONGEST str_offset = read_offset (abfd, buf, offset_size);
> > + const char *str;
> > + if (str_buf[str_offset] == '\0')
> > + str = nullptr;
> > + else
> > + str = (const char *) (str_buf + str_offset);
> > string.emplace (str);
> > buf += offset_size;
> > + break;
> > }
> > - break;
> >
> > case DW_FORM_data1:
> > uint.emplace (read_1_byte (abfd, buf));
> > @@ -247,28 +268,30 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > callback (lh, fe.name, fe.d_index, fe.mod_time, fe.length);
> > }
> >
> > - *bufp = buf;
> > + *line_bufp = buf;
> > }
> >
> > /* See line-header.h. */
> >
> > line_header_up
> > -dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > - dwarf2_per_objfile *per_objfile,
> > - struct dwarf2_section_info *section,
> > - const struct comp_unit_head *cu_header,
> > - const char *comp_dir)
> > +dwarf_decode_line_header (bfd *abfd,
> > + gdb::array_view<const gdb_byte> line_data,
> > + gdb::array_view<const gdb_byte> line_str_data,
> > + const gdb_byte **debug_line_ptr,
> > + bool is_dwz,
> > + const struct comp_unit_head *cu_header,
> > + const char *comp_dir)
> > {
> > - const gdb_byte *line_ptr;
> > + const gdb_byte *line_ptr, *buf;
> > unsigned int bytes_read, offset_size;
> > int i;
> > const char *cur_dir, *cur_file;
> >
> > - bfd *abfd = section->get_bfd_owner ();
> > + buf = *debug_line_ptr;
> >
> > /* Make sure that at least there's room for the total_length field.
> > That could be 12 bytes long, but we're just going to fudge that. */
> > - if (to_underlying (sect_off) + 4 >= section->size)
> > + if (buf + 4 >= line_data.data () + line_data.size ())
> > {
> > dwarf2_statement_list_fits_in_line_number_section_complaint ();
> > return 0;
> > @@ -276,24 +299,26 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> >
> > line_header_up lh (new line_header (comp_dir));
> >
> > - lh->sect_off = sect_off;
> > + lh->sect_off = (sect_offset) (buf - line_data.data ());
> > lh->offset_in_dwz = is_dwz;
> >
> > - line_ptr = section->buffer + to_underlying (sect_off);
> > + line_ptr = buf;
> >
> > /* Read in the header. */
> > LONGEST unit_length
> > - = read_checked_initial_length_and_offset (abfd, line_ptr, cu_header,
> > + = read_checked_initial_length_and_offset (abfd, buf, cu_header,
> > &bytes_read, &offset_size);
> > - line_ptr += bytes_read;
> >
> > - const gdb_byte *start_here = line_ptr;
> > + line_ptr += bytes_read;
> >
> > - if (line_ptr + unit_length > (section->buffer + section->size))
> > + if (line_ptr + unit_length > buf + line_data.size ())
> > {
> > dwarf2_statement_list_fits_in_line_number_section_complaint ();
> > return 0;
> > }
> > +
> > + const gdb_byte *start_here = line_ptr;
> > +
> > lh->statement_program_end = start_here + unit_length;
> > lh->version = read_2_bytes (abfd, line_ptr);
> > line_ptr += 2;
> > @@ -302,7 +327,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > /* This is a version we don't understand. The format could have
> > changed in ways we don't handle properly so just punt. */
> > complaint (_("unsupported version in .debug_line section"));
> > - return NULL;
> > + return nullptr;
> > }
> > if (lh->version >= 5)
> > {
> > @@ -319,13 +344,14 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > complaint (_("unsupported segment selector size %u "
> > "in .debug_line section"),
> > segment_selector_size);
> > - return NULL;
> > + return nullptr;
> > }
> > }
> >
> > LONGEST header_length = read_offset (abfd, line_ptr, offset_size);
> > line_ptr += offset_size;
> > lh->statement_program_start = line_ptr + header_length;
> > +
> > lh->minimum_instruction_length = read_1_byte (abfd, line_ptr);
> > line_ptr += 1;
> >
> > @@ -346,12 +372,16 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> >
> > lh->default_is_stmt = read_1_byte (abfd, line_ptr);
> > line_ptr += 1;
> > +
> > lh->line_base = read_1_signed_byte (abfd, line_ptr);
> > line_ptr += 1;
> > +
> > lh->line_range = read_1_byte (abfd, line_ptr);
> > line_ptr += 1;
> > +
> > lh->opcode_base = read_1_byte (abfd, line_ptr);
> > line_ptr += 1;
> > +
> > lh->standard_opcode_lengths.reset (new unsigned char[lh->opcode_base]);
> >
> > lh->standard_opcode_lengths[0] = 1; /* This should never be used anyway. */
> > @@ -364,21 +394,23 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > if (lh->version >= 5)
> > {
> > /* Read directory table. */
> > - read_formatted_entries (per_objfile, abfd, &line_ptr, lh.get (),
> > - offset_size,
> > - [] (struct line_header *header, const char *name,
> > - dir_index d_index, unsigned int mod_time,
> > - unsigned int length)
> > + read_formatted_entries
> > + (abfd, &line_ptr, line_str_data,
> > + lh.get (), offset_size,
> > + [] (struct line_header *header, const char *name,
> > + dir_index d_index, unsigned int mod_time,
> > + unsigned int length)
> > {
> > header->add_include_dir (name);
> > });
> >
> > /* Read file name table. */
> > - read_formatted_entries (per_objfile, abfd, &line_ptr, lh.get (),
> > - offset_size,
> > - [] (struct line_header *header, const char *name,
> > - dir_index d_index, unsigned int mod_time,
> > - unsigned int length)
> > + read_formatted_entries
> > + (abfd, &line_ptr, line_str_data,
> > + lh.get (), offset_size,
> > + [] (struct line_header *header, const char *name,
> > + dir_index d_index, unsigned int mod_time,
> > + unsigned int length)
> > {
> > header->add_file_name (name, d_index, mod_time, length);
> > });
> > @@ -386,7 +418,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > else
> > {
> > /* Read directory table. */
> > - while ((cur_dir = read_direct_string (abfd, line_ptr, &bytes_read)) != NULL)
> > + while ((cur_dir = read_direct_string (abfd, line_ptr, &bytes_read)) != nullptr)
> > {
> > line_ptr += bytes_read;
> > lh->add_include_dir (cur_dir);
> > @@ -394,7 +426,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > line_ptr += bytes_read;
> >
> > /* Read file name table. */
> > - while ((cur_file = read_direct_string (abfd, line_ptr, &bytes_read)) != NULL)
> > + while ((cur_file = read_direct_string (abfd, line_ptr, &bytes_read)) != nullptr)
> > {
> > unsigned int mod_time, length;
> > dir_index d_index;
> > @@ -412,9 +444,40 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > line_ptr += bytes_read;
> > }
> >
> > - if (line_ptr > (section->buffer + section->size))
> > + if (line_ptr > (buf + line_data.size ()))
> > complaint (_("line number info header doesn't "
> > "fit in `.debug_line' section"));
> >
> > + *debug_line_ptr += unit_length + offset_size;
> > return lh;
> > }
> > +
> > +line_header_up
> > +dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > + dwarf2_per_objfile *per_objfile,
> > + struct dwarf2_section_info *section,
> > + const struct comp_unit_head *cu_header,
> > + const char *comp_dir)
> > +{
> > + struct objfile *objfile = per_objfile->objfile;
> > + struct dwarf2_per_bfd *per_bfd = per_objfile->per_bfd;
> > +
> > + /* Read .debug_line. */
> > + dwarf2_section_info *line_sec = &per_bfd->line;
> > + bfd_size_type line_size = line_sec->get_size (objfile);
> > +
> > + gdb::array_view<const gdb_byte> line (line_sec->buffer, line_size);
> > +
> > + /* Read .debug_line_str. */
> > + dwarf2_section_info *line_str_sec = &per_bfd->line_str;
> > + bfd_size_type line_str_size = line_str_sec->get_size (objfile);
> > +
> > + gdb::array_view<const gdb_byte> line_str (line_str_sec->buffer,
> > + line_str_size);
> > +
> > + const gdb_byte *line_ptr = line.data () + to_underlying (sect_off);
> > +
> > + return dwarf_decode_line_header
> > + (per_bfd->obfd, line, line_str, &line_ptr,
> > + is_dwz, cu_header, comp_dir);
> > +}
> > diff --git a/gdb/dwarf2/line-header.h b/gdb/dwarf2/line-header.h
> > index c068dff70a3..45011882321 100644
> > --- a/gdb/dwarf2/line-header.h
> > +++ b/gdb/dwarf2/line-header.h
> > @@ -217,4 +217,14 @@ extern line_header_up dwarf_decode_line_header
> > struct dwarf2_section_info *section, const struct comp_unit_head *cu_header,
> > const char *comp_dir);
> >
> > +/* Like above but the .debug_line and .debug_line_str are stored in
> > + LINE_DATA and LINE_STR_DATA. *DEBUG_LINE_PTR should point to a
> > + statement program header within LINE_DATA. */
> > +
> > +extern line_header_up dwarf_decode_line_header
> > + (bfd *abfd, gdb::array_view<const gdb_byte> line_data,
> > + gdb::array_view<const gdb_byte> line_str_data,
> > + const gdb_byte **debug_line_ptr, bool is_dwz,
> > + const comp_unit_head *cu_header, const char *comp_dir);
> > +
> > #endif /* DWARF2_LINE_HEADER_H */
> > diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
> > index 97b8137aabf..c52099d5f6a 100644
> > --- a/gdb/dwarf2/read-gdb-index.c
> > +++ b/gdb/dwarf2/read-gdb-index.c
> > @@ -131,6 +131,9 @@ struct mapped_gdb_index final : public mapped_index_base
> > }
> > };
> >
> > +struct mapped_debug_line;
> > +typedef std::unique_ptr<mapped_debug_line> mapped_debug_line_up;
> > +
> > struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > {
> > /* This dumps minimal information about the index.
> > @@ -160,6 +163,15 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > block_search_flags search_flags,
> > domain_search_flags domain);
> >
> > + /* If OBJFILE's debuginfo download has been deferred, use a mapped_debug_line
> > + to generate filenames.
> > +
> > + Otherwise call dwarf2_base_index_functions::map_symbol_filenames. */
> > +
> > + void map_symbol_filenames (struct objfile *objfile,
> > + gdb::function_view<symbol_filename_ftype> fun,
> > + bool need_fullname) override;
> > +
> > /* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
> > debuginfo if necessary. */
> > void expand_all_symtabs (struct objfile *objfile) override;
> > @@ -167,6 +179,15 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > /* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
> > debuginfo if necessary. */
> > struct symtab *find_last_source_symtab (struct objfile *objfile) override;
> > +
> > + /* Filename information related to this .gdb_index. */
> > + mapped_debug_line_up mdl;
> > +
> > + /* Return true if any of the filenames in this .gdb_index's .debug_line
> > + mapping match FILE_MATCHER. Initializes the mapping if necessary. */
> > + bool filename_in_debug_line
> > + (objfile *objfile,
> > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher);
> > };
> >
> > void
> > @@ -202,6 +223,30 @@ dwarf2_gdb_index::find_last_source_symtab (struct objfile *objfile)
> > }
> > }
> >
> > +void
> > +dwarf2_gdb_index::map_symbol_filenames
> > + (struct objfile *objfile,
> > + gdb::function_view<symbol_filename_ftype> fun,
> > + bool need_fullname)
> > +{
> > + try
> > + {
> > + dwarf2_base_index_functions::map_symbol_filenames (objfile, fun,
> > + need_fullname);
> > + }
> > + catch (const gdb_exception &e)
> > + {
> > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > + exception_print (gdb_stderr, e);
> > + else
> > + {
> > + if (mdl == nullptr)
> > + mdl.reset (new mapped_debug_line (objfile));
> > + mdl->map_filenames (fun, need_fullname);
> > + }
> > + }
> > +}
> > +
> > /* This dumps minimal information about the index.
> > It is called via "mt print objfiles".
> > One use is to verify .gdb_index has been loaded by the
> > @@ -365,6 +410,17 @@ dwarf2_gdb_index::do_expand_symtabs_matching
> > return result;
> > }
> >
> > +bool
> > +dwarf2_gdb_index::filename_in_debug_line
> > + (objfile *objfile,
> > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> > +{
> > + if (mdl == nullptr)
> > + mdl.reset (new mapped_debug_line (objfile));
> > +
> > + return mdl->contains_matching_filename (file_matcher);
> > +}
> > +
> > bool
> > dwarf2_gdb_index::expand_symtabs_matching
> > (struct objfile *objfile,
> > @@ -392,6 +448,10 @@ dwarf2_gdb_index::expand_symtabs_matching
> > return false;
> > }
> >
> > + if (file_matcher != nullptr
> > + && !filename_in_debug_line (objfile, file_matcher))
> > + return true;
> > +
> > read_full_dwarf_from_debuginfod (objfile, this);
> > return true;
> > }
> > diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
> > index 4461c76347e..ca219c9b910 100644
> > --- a/gdb/dwarf2/read.c
> > +++ b/gdb/dwarf2/read.c
> > @@ -81,6 +81,7 @@
> > #include <optional>
> > #include "gdbsupport/underlying.h"
> > #include "gdbsupport/hash_enum.h"
> > +#include "gdbsupport/scoped_mmap.h"
> > #include "filename-seen-cache.h"
> > #include "producer.h"
> > #include <fcntl.h>
> > @@ -1967,6 +1968,213 @@ dw2_get_file_names (dwarf2_per_cu_data *this_cu,
> > return this_cu->file_names;
> > }
> >
> > +#if !HAVE_SYS_MMAN_H
> > +
> > +bool
> > +mapped_debug_line::contains_matching_filename
> > + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> > +{
> > + return false;
> > +}
> > +
> > +gdb::array_view<const gdb_byte>
> > +mapped_debug_line::read_debug_line_separate
> > + (const char *filename, std::unique_ptr<index_cache_resource> *resource)
> > +{
> > + return {};
> > +}
> > +
> > +bool
> > +mapped_debug_line::read_debug_line_from_debuginfod (objfile *objfile)
> > +{
> > + return false;
> > +}
> > +
> > +void
> > +mapped_debug_line::map_filenames
> > + (gdb::function_view<symbol_filename_ftype> fun,
> > + bool need_fullname)
> > +{
> > + return;
> > +}
> > +
> > +#else /* !HAVE_SYS_MMAN_H */
> > +
> > +struct line_resource_mmap final : public index_cache_resource
> > +{
> > + /* Try to mmap FILENAME. Throw an exception on failure, including if the
> > + file doesn't exist. */
> > + line_resource_mmap (const char *filename)
> > + : mapping (mmap_file (filename))
> > + {}
> > +
> > + scoped_mmap mapping;
> > +};
> > +
> > +/* See read.h. */
> > +
> > +bool
> > +mapped_debug_line::contains_matching_filename
> > + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> > +{
> > + for (line_header_up &lh : line_headers)
> > + for (file_entry &fe : lh->file_names ())
> > + {
> > + const char *filename = fe.name;
> > +
> > + if (file_matcher (fe.name, false))
> > + return true;
> > +
> > + bool basename_match = file_matcher (lbasename (fe.name), true);
> > +
> > + if (!basenames_may_differ && !basename_match)
> > + continue;
> > +
> > + /* DW_AT_comp_dir is not explicitly mentioned in the .debug_line
> > + until DWARF5. Since we don't have access to the CU at this
> > + point we just check for a partial match on the filename.
> > + If there is a match, the full debuginfo will be downloaded
> > + ane the match will be re-evalute with DW_AT_comp_dir. */
> > + if (lh->version < 5 && fe.d_index == 0)
> > + return basename_match;
> > +
> > + const char *dirname = fe.include_dir (&*lh);
> > + std::string fullname;
> > +
> > + if (dirname == nullptr || IS_ABSOLUTE_PATH (filename))
> > + fullname = filename;
> > + else
> > + fullname = std::string (dirname) + SLASH_STRING + filename;
> > +
> > + gdb::unique_xmalloc_ptr<char> rewritten
> > + = rewrite_source_path (fullname.c_str ());
> > + if (rewritten != nullptr)
> > + fullname = rewritten.release ();
> > +
> > + if (file_matcher (fullname.c_str (), false))
> > + return true;
> > + }
> > +
> > + return false;
> > +}
> > +
> > +/* See read.h. */
> > +
> > +void
> > +mapped_debug_line::map_filenames
> > + (gdb::function_view<symbol_filename_ftype> fun,
> > + bool need_fullname)
> > +{
> > + for (line_header_up &lh : line_headers)
> > + for (file_entry &fe : lh->file_names ())
> > + {
> > + const char *filename = fe.name;
> > +
> > + if (!need_fullname)
> > + {
> > + fun (filename, nullptr);
> > + continue;
> > + }
> > +
> > + const char *dirname = fe.include_dir (&*lh);
> > + std::string fullname;
> > +
> > + if (dirname == nullptr || IS_ABSOLUTE_PATH (filename))
> > + fullname = filename;
> > + else
> > + fullname = std::string (dirname) + SLASH_STRING + filename;
> > +
> > + gdb::unique_xmalloc_ptr<char> rewritten
> > + = rewrite_source_path (fullname.c_str ());
> > + if (rewritten != nullptr)
> > + fullname = rewritten.release ();
> > +
> > + fun (filename, fullname.c_str ());
> > + }
> > +}
> > +
> > +/* See read.h. */
> > +
> > +gdb::array_view<const gdb_byte>
> > +mapped_debug_line::read_debug_line_separate
> > + (const char *filename, std::unique_ptr<index_cache_resource> *resource)
> > +{
> > + if (filename == nullptr)
> > + return {};
> > +
> > + try
> > + {
> > + line_resource_mmap *mmap_resource
> > + = new line_resource_mmap (filename);
> > +
> > + resource->reset (mmap_resource);
> > +
> > + return gdb::array_view<const gdb_byte>
> > + ((const gdb_byte *) mmap_resource->mapping.get (),
> > + mmap_resource->mapping.size ());
> > + }
> > + catch (const gdb_exception &except)
> > + {
> > + exception_print (gdb_stderr, except);
> > + }
> > +
> > + return {};
> > +}
> > +
> > +/* See read.h. */
> > +
> > +bool
> > +mapped_debug_line::read_debug_line_from_debuginfod (objfile *objfile)
> > +{
> > + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > + if (build_id == nullptr)
> > + return false;
> > +
> > + gdb::unique_xmalloc_ptr<char> line_path;
> > + scoped_fd line_fd = debuginfod_section_query (build_id->data,
> > + build_id->size,
> > + bfd_get_filename
> > + (objfile->obfd.get ()),
> > + ".debug_line",
> > + &line_path);
> > +
> > + if (line_fd.get () < 0)
> > + return false;
> > +
> > + gdb::unique_xmalloc_ptr<char> line_str_path;
> > + scoped_fd line_str_fd = debuginfod_section_query (build_id->data,
> > + build_id->size,
> > + bfd_get_filename
> > + (objfile->obfd.get ()),
> > + ".debug_line_str",
> > + &line_str_path);
> > +
> > + line_data = read_debug_line_separate (line_path.get (), &line_resource);
> > + line_str_data = read_debug_line_separate (line_str_path.get (),
> > + &line_str_resource);
> > +
> > + const gdb_byte *line_ptr = line_data.data ();
> > +
> > + while (line_ptr < line_data.data () + line_data.size ())
> > + {
> > + line_header_up lh
> > + = dwarf_decode_line_header (objfile->obfd.get (),
> > + line_data, line_str_data,
> > + &line_ptr, false,
> > + nullptr, nullptr);
> > + line_headers.emplace_back (std::move (lh));
> > + }
> > +
> > + return true;
> > +}
> > +#endif /* !HAVE_SYS_MMAN_H */
> > +
> > +mapped_debug_line::mapped_debug_line (objfile *objfile)
> > +{
> > + if (!read_debug_line_from_debuginfod (objfile))
> > + line_headers.clear ();
> > +}
> > +
> > /* A helper for the "quick" functions which computes and caches the
> > real path for a given file name from the line table. */
> >
> > diff --git a/gdb/dwarf2/read.h b/gdb/dwarf2/read.h
> > index 099d3d65c27..19d19a6edc1 100644
> > --- a/gdb/dwarf2/read.h
> > +++ b/gdb/dwarf2/read.h
> > @@ -33,6 +33,7 @@
> > #include "gdbsupport/hash_enum.h"
> > #include "gdbsupport/function-view.h"
> > #include "gdbsupport/packed.h"
> > +#include "dwarf2/line-header.h"
> >
> > /* Hold 'maintenance (set|show) dwarf' commands. */
> > extern struct cmd_list_element *set_dwarf_cmdlist;
> > @@ -942,4 +943,40 @@ extern htab_up create_quick_file_names_table (unsigned int nr_initial_entries);
> > extern void read_full_dwarf_from_debuginfod (struct objfile *,
> > dwarf2_base_index_functions *);
> >
> > +struct mapped_debug_line
> > +{
> > + mapped_debug_line (objfile *objfile);
> > +
> > + /* Return true if any of the mapped .debug_line's filenames match
> > + FILE_MATCHER. */
> > +
> > + bool contains_matching_filename
> > + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher);
> > +
> > + /* Call FUN with each filename in this mapped .debug_line. Include
> > + each file's fullname if NEED_FULLNAME is true. */
> > +
> > + void map_filenames (gdb::function_view<symbol_filename_ftype> fun,
> > + bool need_fullname);
> > +
> > +private:
> > + std::vector<line_header_up> line_headers;
> > +
> > + gdb::array_view<const gdb_byte> line_data;
> > + gdb::array_view<const gdb_byte> line_str_data;
> > +
> > + std::unique_ptr<index_cache_resource> line_resource;
> > + std::unique_ptr<index_cache_resource> line_str_resource;
> > +
> > + /* Download the .debug_line and .debug_line_str associated with OBJFILE
> > + and populate line_headers. */
> > +
> > + bool read_debug_line_from_debuginfod (objfile *objfile);
> > +
> > + /* Initialize line_data and line_str_data with the .debug_line and
> > + .debug_line_str downloaded read_debug_line_from_debuginfod. */
> > +
> > + gdb::array_view<const gdb_byte> read_debug_line_separate
> > + (const char *filename, std::unique_ptr<index_cache_resource> *resource);
> > +};
> > #endif /* DWARF2READ_H */
> > diff --git a/gdb/mi/mi-out.c b/gdb/mi/mi-out.c
> > index ff93d2cd448..8ad2c7f4497 100644
> > --- a/gdb/mi/mi-out.c
> > +++ b/gdb/mi/mi-out.c
> > @@ -277,7 +277,14 @@ mi_ui_out::do_progress_notify (const std::string &msg, const char *unit,
> >
> > if (info.state == progress_update::START)
> > {
> > - gdb_printf ("%s...\n", msg.c_str ());
> > + std::string prefix;
> > + if (cur_prefix_state == prefix_state_t::NEWLINE_NEEDED)
> > + {
> > + prefix = "\n";
> > + cur_prefix_state = prefix_state_t::NEWLINE_PRINTED;
> > + }
> > +
> > + gdb_printf ("%s...\n", (prefix + msg).c_str ());
> > info.state = progress_update::WORKING;
> > }
> > }
> > diff --git a/gdb/testsuite/gdb.debuginfod/section.exp b/gdb/testsuite/gdb.debuginfod/section.exp
> > index 35086d15705..a9dad927696 100644
> > --- a/gdb/testsuite/gdb.debuginfod/section.exp
> > +++ b/gdb/testsuite/gdb.debuginfod/section.exp
> > @@ -18,6 +18,7 @@
> > standard_testfile
> >
> > load_lib debuginfod-support.exp
> > +load_lib completion-support.exp
> >
> > clean_restart
> > require allow_debuginfod_tests
> > @@ -144,6 +145,32 @@ proc_with_prefix local_url { } {
> > # Hit the breakpoint.
> > set res ".*Breakpoint 2, libsection2_test.*\"In ${libfile2}\\\\n\".*"
> > gdb_test "c" $res "break continue"
> > +
> > + clean_restart_with_prompt $binfile "line 1"
> > +
> > + # List source file using .debug_line download.
> > + set res ".*\.debug_line.*$lib_sl1.*21.*extern void libsection2_test.*"
> > + gdb_test "list $libsrc1:21" $res "line 1 list"
> > +
> > + clean_restart_with_prompt $binfile "line 2"
> > +
> > + # Set breakpoint using .debug_line download.
> > + set res ".*section \.debug_line for $lib_sl1.*Breakpoint 2 at.*$libsrc1.*"
> > + gdb_test "br $libsrc1:41" $res "line 2 br"
> > +
> > + # Continue to breakpoint.
> > + set res "Breakpoint 2, libsection1_test.*\"Cancelling thread\\\\n\".*"
> > + gdb_test "c" $res "line 2 continue"
> > +
> > + # Check that download progress message is correctly formatted
> > + # when printing threads.
> > + set res ".*separate debug info for $lib_sl2\.\.\.\r\n.* 2 Thread.*"
> > + gdb_test "info thr" $res "line thread"
> > +
> > + clean_restart_with_prompt $binfile "autocomplete"
> > +
> > + # Download debuginfo during autocompletion.
> > + test_gdb_complete_tab_unique "br lib" ".*Downloading separate debug.*" ""
> > }
> >
> > # Create CACHE and DB directories ready for debuginfod to use.
> > diff --git a/gdb/ui-out.c b/gdb/ui-out.c
> > index b3a2fb7f4e5..52d36d5bbf8 100644
> > --- a/gdb/ui-out.c
> > +++ b/gdb/ui-out.c
> > @@ -31,6 +31,9 @@
> > #include <memory>
> > #include <string>
> >
> > +/* Current state of newline prefixing for progress update messages. */
> > +prefix_state_t cur_prefix_state = prefix_state_t::NEWLINE_OFF;
> > +
> > namespace {
> >
> > /* A header of a ui_out_table. */
> > diff --git a/gdb/ui-out.h b/gdb/ui-out.h
> > index 5b6ddd55063..7e885cb1d38 100644
> > --- a/gdb/ui-out.h
> > +++ b/gdb/ui-out.h
> > @@ -296,6 +296,21 @@ class ui_out
> > BAR
> > };
> >
> > + /* Used to communicate the status of a newline prefix for the next progress
> > + update message. */
> > + enum prefix_state
> > + {
> > + /* Do not modify the next progress update message. */
> > + NEWLINE_OFF,
> > +
> > + /* The next progress update message should include a newline prefix. */
> > + NEWLINE_NEEDED,
> > +
> > + /* A newline prefix was included in a debuginfod progress update
> > + message. */
> > + NEWLINE_PRINTED
> > + };
> > +
> > /* SHOULD_PRINT indicates whether something should be printed for a tty. */
> > progress_update ()
> > {
> > @@ -393,6 +408,11 @@ class ui_out
> > ui_out_level *current_level () const;
> > };
> >
> > +typedef ui_out::progress_update::prefix_state prefix_state_t;
> > +
> > +/* Current state of the newline prefix. */
> > +extern prefix_state_t cur_prefix_state;
> > +
> > /* Start a new tuple or list on construction, and end it on
> > destruction. Normally this is used via the typedefs
> > ui_out_emit_tuple and ui_out_emit_list. */
> > --
> > 2.43.0
> >
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING*3][PATCH 3/3 v6] gdb/debuginfod: Add .debug_line downloading
2024-05-31 20:58 ` [PING*2][PATCH " Aaron Merey
@ 2024-06-07 20:49 ` Aaron Merey
2024-06-14 18:24 ` Aaron Merey
1 sibling, 0 replies; 18+ messages in thread
From: Aaron Merey @ 2024-06-07 20:49 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 31, 2024 at 4:58 PM Aaron Merey <amerey@redhat.com> wrote:
>
> Ping
>
> Thanks,
> Aaron
>
> On Fri, May 17, 2024 at 10:10 AM Aaron Merey <amerey@redhat.com> wrote:
> >
> > Ping
> >
> > Thanks,
> > Aaron
> >
> > On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
> > >
> > > This is a repost of v6 from
> > > https://sourceware.org/pipermail/gdb-patches/2024-January/205956.html
> > >
> > > ELF/DWARF section downloading allows gdb to download .gdb_index files in
> > > order to defer full debuginfo downloads. However .gdb_index does not
> > > contain any information regarding source filenames. When a gdb command
> > > includes a filename argument (ex. 'break main.c:50'), this results in
> > > the mass downloading of all deferred debuginfo so that gdb can search the
> > > debuginfo for matching source filenames. This can result in unnecessary
> > > downloads.
> > >
> > > To improve this, have gdb instead download each debuginfo's .debug_line
> > > (and .debug_line_str if using DWARF5) when executing these commands.
> > > Download full debuginfo only when its .debug_line contains a matching
> > > filename.
> > >
> > > Since the combined size of .debug_line and .debug_line_str is only about
> > > 1% the size of the corresponding debuginfo, significant time can be saved
> > > by checking these sections before choosing to download an entire debuginfo.
> > >
> > > This patch also redirects stdout and stderr of the debuginfod server
> > > used by testsuite/gdb.debuginfod tests to a server_log standard output
> > > file. While adding tests for this patch I ran into an issue where the
> > > test server would block when logging to stderr, presumably because the
> > > stderr buffer filled up and wasn't being read from. Redirecting the
> > > log to a file fixes this and also makes the server log more accessible
> > > when debugging test failures.
> > > ---
> > > gdb/cli-out.c | 11 +-
> > > gdb/completer.c | 18 +-
> > > gdb/dwarf2/line-header.c | 161 ++++++++++++------
> > > gdb/dwarf2/line-header.h | 10 ++
> > > gdb/dwarf2/read-gdb-index.c | 60 +++++++
> > > gdb/dwarf2/read.c | 208 +++++++++++++++++++++++
> > > gdb/dwarf2/read.h | 37 ++++
> > > gdb/mi/mi-out.c | 9 +-
> > > gdb/testsuite/gdb.debuginfod/section.exp | 27 +++
> > > gdb/ui-out.c | 3 +
> > > gdb/ui-out.h | 20 +++
> > > 11 files changed, 511 insertions(+), 53 deletions(-)
> > >
> > > diff --git a/gdb/cli-out.c b/gdb/cli-out.c
> > > index 1c303f09662..d65b69ebe01 100644
> > > --- a/gdb/cli-out.c
> > > +++ b/gdb/cli-out.c
> > > @@ -306,16 +306,23 @@ cli_ui_out::do_progress_notify (const std::string &msg,
> > >
> > > if (info.state == progress_update::START)
> > > {
> > > + std::string prefix;
> > > + if (cur_prefix_state == prefix_state_t::NEWLINE_NEEDED)
> > > + {
> > > + prefix = "\n";
> > > + cur_prefix_state = prefix_state_t::NEWLINE_PRINTED;
> > > + }
> > > +
> > > if (stream->isatty ()
> > > && current_ui->input_interactive_p ()
> > > && chars_per_line >= MIN_CHARS_PER_LINE)
> > > {
> > > - gdb_printf (stream, "%s\n", msg.c_str ());
> > > + gdb_printf (stream, "%s\n", (prefix + msg).c_str ());
> > > info.state = progress_update::BAR;
> > > }
> > > else
> > > {
> > > - gdb_printf (stream, "%s...\n", msg.c_str ());
> > > + gdb_printf (stream, "%s...\n", (prefix + msg).c_str ());
> > > info.state = progress_update::WORKING;
> > > }
> > > }
> > > diff --git a/gdb/completer.c b/gdb/completer.c
> > > index f1f44109bdc..708081cf280 100644
> > > --- a/gdb/completer.c
> > > +++ b/gdb/completer.c
> > > @@ -1354,6 +1354,10 @@ complete_line_internal_1 (completion_tracker &tracker,
> > > {
> > > /* We've recognized a full command. */
> > >
> > > + /* Disable pagination since responding to the pagination prompt
> > > + overwrites rl_line_buffer. */
> > > + scoped_restore pag_restore = make_scoped_restore (&pagination_enabled, false);
> > > +
> > > if (p == tmp_command + point)
> > > {
> > > /* There is no non-whitespace in the line beyond the
> > > @@ -1453,7 +1457,8 @@ complete_line_internal_1 (completion_tracker &tracker,
> > > }
> > >
> > > /* Wrapper around complete_line_internal_1 to handle
> > > - MAX_COMPLETIONS_REACHED_ERROR. */
> > > + MAX_COMPLETIONS_REACHED_ERROR and possible progress update
> > > + interactions. */
> > >
> > > static void
> > > complete_line_internal (completion_tracker &tracker,
> > > @@ -1461,6 +1466,11 @@ complete_line_internal (completion_tracker &tracker,
> > > const char *line_buffer, int point,
> > > complete_line_internal_reason reason)
> > > {
> > > + scoped_restore restore_prefix_state
> > > + = make_scoped_restore
> > > + (&cur_prefix_state,
> > > + ui_out::progress_update::prefix_state::NEWLINE_NEEDED);
> > > +
> > > try
> > > {
> > > complete_line_internal_1 (tracker, text, line_buffer, point, reason);
> > > @@ -1470,6 +1480,12 @@ complete_line_internal (completion_tracker &tracker,
> > > if (except.error != MAX_COMPLETIONS_REACHED_ERROR)
> > > throw;
> > > }
> > > +
> > > + /* If progress update messages printed, then the text being completed
> > > + needs to be printed again. */
> > > + if (cur_prefix_state
> > > + == ui_out::progress_update::prefix_state::NEWLINE_PRINTED)
> > > + rl_forced_update_display ();
> > > }
> > >
> > > /* See completer.h. */
> > > diff --git a/gdb/dwarf2/line-header.c b/gdb/dwarf2/line-header.c
> > > index eddb2ef7ae8..1d24386c0eb 100644
> > > --- a/gdb/dwarf2/line-header.c
> > > +++ b/gdb/dwarf2/line-header.c
> > > @@ -101,12 +101,15 @@ read_checked_initial_length_and_offset (bfd *abfd, const gdb_byte *buf,
> > > {
> > > LONGEST length = read_initial_length (abfd, buf, bytes_read);
> > >
> > > - gdb_assert (cu_header->initial_length_size == 4
> > > - || cu_header->initial_length_size == 8
> > > - || cu_header->initial_length_size == 12);
> > > + if (cu_header != nullptr)
> > > + {
> > > + gdb_assert (cu_header->initial_length_size == 4
> > > + || cu_header->initial_length_size == 8
> > > + || cu_header->initial_length_size == 12);
> > >
> > > - if (cu_header->initial_length_size != *bytes_read)
> > > - complaint (_("intermixed 32-bit and 64-bit DWARF sections"));
> > > + if (cu_header->initial_length_size != *bytes_read)
> > > + complaint (_("intermixed 32-bit and 64-bit DWARF sections"));
> > > + }
> > >
> > > *offset_size = (*bytes_read == 4) ? 4 : 8;
> > > return length;
> > > @@ -115,21 +118,27 @@ read_checked_initial_length_and_offset (bfd *abfd, const gdb_byte *buf,
> > > /* Read directory or file name entry format, starting with byte of
> > > format count entries, ULEB128 pairs of entry formats, ULEB128 of
> > > entries count and the entries themselves in the described entry
> > > - format. */
> > > + format.
> > > +
> > > + .debug_line and .debug_line_str sections are stored in LINE_BUFP
> > > + and LINE_STR_DATA respectively. */
> > >
> > > static void
> > > -read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > > - const gdb_byte **bufp, struct line_header *lh,
> > > - unsigned int offset_size,
> > > - void (*callback) (struct line_header *lh,
> > > - const char *name,
> > > - dir_index d_index,
> > > - unsigned int mod_time,
> > > - unsigned int length))
> > > +read_formatted_entries
> > > + (bfd *abfd, const gdb_byte **line_bufp,
> > > + const gdb::array_view<const gdb_byte> line_str_data,
> > > + struct line_header *lh,
> > > + unsigned int offset_size,
> > > + void (*callback) (struct line_header *lh,
> > > + const char *name,
> > > + dir_index d_index,
> > > + unsigned int mod_time,
> > > + unsigned int length))
> > > {
> > > gdb_byte format_count, formati;
> > > ULONGEST data_count, datai;
> > > - const gdb_byte *buf = *bufp;
> > > + const gdb_byte *buf = *line_bufp;
> > > + const gdb_byte *str_buf = line_str_data.data ();
> > > const gdb_byte *format_header_data;
> > > unsigned int bytes_read;
> > >
> > > @@ -156,7 +165,7 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > > ULONGEST content_type = read_unsigned_leb128 (abfd, format, &bytes_read);
> > > format += bytes_read;
> > >
> > > - ULONGEST form = read_unsigned_leb128 (abfd, format, &bytes_read);
> > > + ULONGEST form = read_unsigned_leb128 (abfd, format, &bytes_read);
> > > format += bytes_read;
> > >
> > > std::optional<const char *> string;
> > > @@ -171,12 +180,24 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > >
> > > case DW_FORM_line_strp:
> > > {
> > > - const char *str
> > > - = per_objfile->read_line_string (buf, offset_size);
> > > + if (line_str_data.empty ())
> > > + warning (_("Dwarf Error: DW_FORM_line_strp used without " \
> > > + "required section"));
> > > +
> > > + if (line_str_data.size () <= offset_size)
> > > + warning (_("Dwarf Error: DW_FORM_line_strp pointing outside " \
> > > + "of section .debug_line_str"));
> > > +
> > > + ULONGEST str_offset = read_offset (abfd, buf, offset_size);
> > > + const char *str;
> > > + if (str_buf[str_offset] == '\0')
> > > + str = nullptr;
> > > + else
> > > + str = (const char *) (str_buf + str_offset);
> > > string.emplace (str);
> > > buf += offset_size;
> > > + break;
> > > }
> > > - break;
> > >
> > > case DW_FORM_data1:
> > > uint.emplace (read_1_byte (abfd, buf));
> > > @@ -247,28 +268,30 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > > callback (lh, fe.name, fe.d_index, fe.mod_time, fe.length);
> > > }
> > >
> > > - *bufp = buf;
> > > + *line_bufp = buf;
> > > }
> > >
> > > /* See line-header.h. */
> > >
> > > line_header_up
> > > -dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > - dwarf2_per_objfile *per_objfile,
> > > - struct dwarf2_section_info *section,
> > > - const struct comp_unit_head *cu_header,
> > > - const char *comp_dir)
> > > +dwarf_decode_line_header (bfd *abfd,
> > > + gdb::array_view<const gdb_byte> line_data,
> > > + gdb::array_view<const gdb_byte> line_str_data,
> > > + const gdb_byte **debug_line_ptr,
> > > + bool is_dwz,
> > > + const struct comp_unit_head *cu_header,
> > > + const char *comp_dir)
> > > {
> > > - const gdb_byte *line_ptr;
> > > + const gdb_byte *line_ptr, *buf;
> > > unsigned int bytes_read, offset_size;
> > > int i;
> > > const char *cur_dir, *cur_file;
> > >
> > > - bfd *abfd = section->get_bfd_owner ();
> > > + buf = *debug_line_ptr;
> > >
> > > /* Make sure that at least there's room for the total_length field.
> > > That could be 12 bytes long, but we're just going to fudge that. */
> > > - if (to_underlying (sect_off) + 4 >= section->size)
> > > + if (buf + 4 >= line_data.data () + line_data.size ())
> > > {
> > > dwarf2_statement_list_fits_in_line_number_section_complaint ();
> > > return 0;
> > > @@ -276,24 +299,26 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > >
> > > line_header_up lh (new line_header (comp_dir));
> > >
> > > - lh->sect_off = sect_off;
> > > + lh->sect_off = (sect_offset) (buf - line_data.data ());
> > > lh->offset_in_dwz = is_dwz;
> > >
> > > - line_ptr = section->buffer + to_underlying (sect_off);
> > > + line_ptr = buf;
> > >
> > > /* Read in the header. */
> > > LONGEST unit_length
> > > - = read_checked_initial_length_and_offset (abfd, line_ptr, cu_header,
> > > + = read_checked_initial_length_and_offset (abfd, buf, cu_header,
> > > &bytes_read, &offset_size);
> > > - line_ptr += bytes_read;
> > >
> > > - const gdb_byte *start_here = line_ptr;
> > > + line_ptr += bytes_read;
> > >
> > > - if (line_ptr + unit_length > (section->buffer + section->size))
> > > + if (line_ptr + unit_length > buf + line_data.size ())
> > > {
> > > dwarf2_statement_list_fits_in_line_number_section_complaint ();
> > > return 0;
> > > }
> > > +
> > > + const gdb_byte *start_here = line_ptr;
> > > +
> > > lh->statement_program_end = start_here + unit_length;
> > > lh->version = read_2_bytes (abfd, line_ptr);
> > > line_ptr += 2;
> > > @@ -302,7 +327,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > /* This is a version we don't understand. The format could have
> > > changed in ways we don't handle properly so just punt. */
> > > complaint (_("unsupported version in .debug_line section"));
> > > - return NULL;
> > > + return nullptr;
> > > }
> > > if (lh->version >= 5)
> > > {
> > > @@ -319,13 +344,14 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > complaint (_("unsupported segment selector size %u "
> > > "in .debug_line section"),
> > > segment_selector_size);
> > > - return NULL;
> > > + return nullptr;
> > > }
> > > }
> > >
> > > LONGEST header_length = read_offset (abfd, line_ptr, offset_size);
> > > line_ptr += offset_size;
> > > lh->statement_program_start = line_ptr + header_length;
> > > +
> > > lh->minimum_instruction_length = read_1_byte (abfd, line_ptr);
> > > line_ptr += 1;
> > >
> > > @@ -346,12 +372,16 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > >
> > > lh->default_is_stmt = read_1_byte (abfd, line_ptr);
> > > line_ptr += 1;
> > > +
> > > lh->line_base = read_1_signed_byte (abfd, line_ptr);
> > > line_ptr += 1;
> > > +
> > > lh->line_range = read_1_byte (abfd, line_ptr);
> > > line_ptr += 1;
> > > +
> > > lh->opcode_base = read_1_byte (abfd, line_ptr);
> > > line_ptr += 1;
> > > +
> > > lh->standard_opcode_lengths.reset (new unsigned char[lh->opcode_base]);
> > >
> > > lh->standard_opcode_lengths[0] = 1; /* This should never be used anyway. */
> > > @@ -364,21 +394,23 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > if (lh->version >= 5)
> > > {
> > > /* Read directory table. */
> > > - read_formatted_entries (per_objfile, abfd, &line_ptr, lh.get (),
> > > - offset_size,
> > > - [] (struct line_header *header, const char *name,
> > > - dir_index d_index, unsigned int mod_time,
> > > - unsigned int length)
> > > + read_formatted_entries
> > > + (abfd, &line_ptr, line_str_data,
> > > + lh.get (), offset_size,
> > > + [] (struct line_header *header, const char *name,
> > > + dir_index d_index, unsigned int mod_time,
> > > + unsigned int length)
> > > {
> > > header->add_include_dir (name);
> > > });
> > >
> > > /* Read file name table. */
> > > - read_formatted_entries (per_objfile, abfd, &line_ptr, lh.get (),
> > > - offset_size,
> > > - [] (struct line_header *header, const char *name,
> > > - dir_index d_index, unsigned int mod_time,
> > > - unsigned int length)
> > > + read_formatted_entries
> > > + (abfd, &line_ptr, line_str_data,
> > > + lh.get (), offset_size,
> > > + [] (struct line_header *header, const char *name,
> > > + dir_index d_index, unsigned int mod_time,
> > > + unsigned int length)
> > > {
> > > header->add_file_name (name, d_index, mod_time, length);
> > > });
> > > @@ -386,7 +418,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > else
> > > {
> > > /* Read directory table. */
> > > - while ((cur_dir = read_direct_string (abfd, line_ptr, &bytes_read)) != NULL)
> > > + while ((cur_dir = read_direct_string (abfd, line_ptr, &bytes_read)) != nullptr)
> > > {
> > > line_ptr += bytes_read;
> > > lh->add_include_dir (cur_dir);
> > > @@ -394,7 +426,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > line_ptr += bytes_read;
> > >
> > > /* Read file name table. */
> > > - while ((cur_file = read_direct_string (abfd, line_ptr, &bytes_read)) != NULL)
> > > + while ((cur_file = read_direct_string (abfd, line_ptr, &bytes_read)) != nullptr)
> > > {
> > > unsigned int mod_time, length;
> > > dir_index d_index;
> > > @@ -412,9 +444,40 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > line_ptr += bytes_read;
> > > }
> > >
> > > - if (line_ptr > (section->buffer + section->size))
> > > + if (line_ptr > (buf + line_data.size ()))
> > > complaint (_("line number info header doesn't "
> > > "fit in `.debug_line' section"));
> > >
> > > + *debug_line_ptr += unit_length + offset_size;
> > > return lh;
> > > }
> > > +
> > > +line_header_up
> > > +dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > + dwarf2_per_objfile *per_objfile,
> > > + struct dwarf2_section_info *section,
> > > + const struct comp_unit_head *cu_header,
> > > + const char *comp_dir)
> > > +{
> > > + struct objfile *objfile = per_objfile->objfile;
> > > + struct dwarf2_per_bfd *per_bfd = per_objfile->per_bfd;
> > > +
> > > + /* Read .debug_line. */
> > > + dwarf2_section_info *line_sec = &per_bfd->line;
> > > + bfd_size_type line_size = line_sec->get_size (objfile);
> > > +
> > > + gdb::array_view<const gdb_byte> line (line_sec->buffer, line_size);
> > > +
> > > + /* Read .debug_line_str. */
> > > + dwarf2_section_info *line_str_sec = &per_bfd->line_str;
> > > + bfd_size_type line_str_size = line_str_sec->get_size (objfile);
> > > +
> > > + gdb::array_view<const gdb_byte> line_str (line_str_sec->buffer,
> > > + line_str_size);
> > > +
> > > + const gdb_byte *line_ptr = line.data () + to_underlying (sect_off);
> > > +
> > > + return dwarf_decode_line_header
> > > + (per_bfd->obfd, line, line_str, &line_ptr,
> > > + is_dwz, cu_header, comp_dir);
> > > +}
> > > diff --git a/gdb/dwarf2/line-header.h b/gdb/dwarf2/line-header.h
> > > index c068dff70a3..45011882321 100644
> > > --- a/gdb/dwarf2/line-header.h
> > > +++ b/gdb/dwarf2/line-header.h
> > > @@ -217,4 +217,14 @@ extern line_header_up dwarf_decode_line_header
> > > struct dwarf2_section_info *section, const struct comp_unit_head *cu_header,
> > > const char *comp_dir);
> > >
> > > +/* Like above but the .debug_line and .debug_line_str are stored in
> > > + LINE_DATA and LINE_STR_DATA. *DEBUG_LINE_PTR should point to a
> > > + statement program header within LINE_DATA. */
> > > +
> > > +extern line_header_up dwarf_decode_line_header
> > > + (bfd *abfd, gdb::array_view<const gdb_byte> line_data,
> > > + gdb::array_view<const gdb_byte> line_str_data,
> > > + const gdb_byte **debug_line_ptr, bool is_dwz,
> > > + const comp_unit_head *cu_header, const char *comp_dir);
> > > +
> > > #endif /* DWARF2_LINE_HEADER_H */
> > > diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
> > > index 97b8137aabf..c52099d5f6a 100644
> > > --- a/gdb/dwarf2/read-gdb-index.c
> > > +++ b/gdb/dwarf2/read-gdb-index.c
> > > @@ -131,6 +131,9 @@ struct mapped_gdb_index final : public mapped_index_base
> > > }
> > > };
> > >
> > > +struct mapped_debug_line;
> > > +typedef std::unique_ptr<mapped_debug_line> mapped_debug_line_up;
> > > +
> > > struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > > {
> > > /* This dumps minimal information about the index.
> > > @@ -160,6 +163,15 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > > block_search_flags search_flags,
> > > domain_search_flags domain);
> > >
> > > + /* If OBJFILE's debuginfo download has been deferred, use a mapped_debug_line
> > > + to generate filenames.
> > > +
> > > + Otherwise call dwarf2_base_index_functions::map_symbol_filenames. */
> > > +
> > > + void map_symbol_filenames (struct objfile *objfile,
> > > + gdb::function_view<symbol_filename_ftype> fun,
> > > + bool need_fullname) override;
> > > +
> > > /* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
> > > debuginfo if necessary. */
> > > void expand_all_symtabs (struct objfile *objfile) override;
> > > @@ -167,6 +179,15 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > > /* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
> > > debuginfo if necessary. */
> > > struct symtab *find_last_source_symtab (struct objfile *objfile) override;
> > > +
> > > + /* Filename information related to this .gdb_index. */
> > > + mapped_debug_line_up mdl;
> > > +
> > > + /* Return true if any of the filenames in this .gdb_index's .debug_line
> > > + mapping match FILE_MATCHER. Initializes the mapping if necessary. */
> > > + bool filename_in_debug_line
> > > + (objfile *objfile,
> > > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher);
> > > };
> > >
> > > void
> > > @@ -202,6 +223,30 @@ dwarf2_gdb_index::find_last_source_symtab (struct objfile *objfile)
> > > }
> > > }
> > >
> > > +void
> > > +dwarf2_gdb_index::map_symbol_filenames
> > > + (struct objfile *objfile,
> > > + gdb::function_view<symbol_filename_ftype> fun,
> > > + bool need_fullname)
> > > +{
> > > + try
> > > + {
> > > + dwarf2_base_index_functions::map_symbol_filenames (objfile, fun,
> > > + need_fullname);
> > > + }
> > > + catch (const gdb_exception &e)
> > > + {
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + exception_print (gdb_stderr, e);
> > > + else
> > > + {
> > > + if (mdl == nullptr)
> > > + mdl.reset (new mapped_debug_line (objfile));
> > > + mdl->map_filenames (fun, need_fullname);
> > > + }
> > > + }
> > > +}
> > > +
> > > /* This dumps minimal information about the index.
> > > It is called via "mt print objfiles".
> > > One use is to verify .gdb_index has been loaded by the
> > > @@ -365,6 +410,17 @@ dwarf2_gdb_index::do_expand_symtabs_matching
> > > return result;
> > > }
> > >
> > > +bool
> > > +dwarf2_gdb_index::filename_in_debug_line
> > > + (objfile *objfile,
> > > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> > > +{
> > > + if (mdl == nullptr)
> > > + mdl.reset (new mapped_debug_line (objfile));
> > > +
> > > + return mdl->contains_matching_filename (file_matcher);
> > > +}
> > > +
> > > bool
> > > dwarf2_gdb_index::expand_symtabs_matching
> > > (struct objfile *objfile,
> > > @@ -392,6 +448,10 @@ dwarf2_gdb_index::expand_symtabs_matching
> > > return false;
> > > }
> > >
> > > + if (file_matcher != nullptr
> > > + && !filename_in_debug_line (objfile, file_matcher))
> > > + return true;
> > > +
> > > read_full_dwarf_from_debuginfod (objfile, this);
> > > return true;
> > > }
> > > diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
> > > index 4461c76347e..ca219c9b910 100644
> > > --- a/gdb/dwarf2/read.c
> > > +++ b/gdb/dwarf2/read.c
> > > @@ -81,6 +81,7 @@
> > > #include <optional>
> > > #include "gdbsupport/underlying.h"
> > > #include "gdbsupport/hash_enum.h"
> > > +#include "gdbsupport/scoped_mmap.h"
> > > #include "filename-seen-cache.h"
> > > #include "producer.h"
> > > #include <fcntl.h>
> > > @@ -1967,6 +1968,213 @@ dw2_get_file_names (dwarf2_per_cu_data *this_cu,
> > > return this_cu->file_names;
> > > }
> > >
> > > +#if !HAVE_SYS_MMAN_H
> > > +
> > > +bool
> > > +mapped_debug_line::contains_matching_filename
> > > + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> > > +{
> > > + return false;
> > > +}
> > > +
> > > +gdb::array_view<const gdb_byte>
> > > +mapped_debug_line::read_debug_line_separate
> > > + (const char *filename, std::unique_ptr<index_cache_resource> *resource)
> > > +{
> > > + return {};
> > > +}
> > > +
> > > +bool
> > > +mapped_debug_line::read_debug_line_from_debuginfod (objfile *objfile)
> > > +{
> > > + return false;
> > > +}
> > > +
> > > +void
> > > +mapped_debug_line::map_filenames
> > > + (gdb::function_view<symbol_filename_ftype> fun,
> > > + bool need_fullname)
> > > +{
> > > + return;
> > > +}
> > > +
> > > +#else /* !HAVE_SYS_MMAN_H */
> > > +
> > > +struct line_resource_mmap final : public index_cache_resource
> > > +{
> > > + /* Try to mmap FILENAME. Throw an exception on failure, including if the
> > > + file doesn't exist. */
> > > + line_resource_mmap (const char *filename)
> > > + : mapping (mmap_file (filename))
> > > + {}
> > > +
> > > + scoped_mmap mapping;
> > > +};
> > > +
> > > +/* See read.h. */
> > > +
> > > +bool
> > > +mapped_debug_line::contains_matching_filename
> > > + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> > > +{
> > > + for (line_header_up &lh : line_headers)
> > > + for (file_entry &fe : lh->file_names ())
> > > + {
> > > + const char *filename = fe.name;
> > > +
> > > + if (file_matcher (fe.name, false))
> > > + return true;
> > > +
> > > + bool basename_match = file_matcher (lbasename (fe.name), true);
> > > +
> > > + if (!basenames_may_differ && !basename_match)
> > > + continue;
> > > +
> > > + /* DW_AT_comp_dir is not explicitly mentioned in the .debug_line
> > > + until DWARF5. Since we don't have access to the CU at this
> > > + point we just check for a partial match on the filename.
> > > + If there is a match, the full debuginfo will be downloaded
> > > + ane the match will be re-evalute with DW_AT_comp_dir. */
> > > + if (lh->version < 5 && fe.d_index == 0)
> > > + return basename_match;
> > > +
> > > + const char *dirname = fe.include_dir (&*lh);
> > > + std::string fullname;
> > > +
> > > + if (dirname == nullptr || IS_ABSOLUTE_PATH (filename))
> > > + fullname = filename;
> > > + else
> > > + fullname = std::string (dirname) + SLASH_STRING + filename;
> > > +
> > > + gdb::unique_xmalloc_ptr<char> rewritten
> > > + = rewrite_source_path (fullname.c_str ());
> > > + if (rewritten != nullptr)
> > > + fullname = rewritten.release ();
> > > +
> > > + if (file_matcher (fullname.c_str (), false))
> > > + return true;
> > > + }
> > > +
> > > + return false;
> > > +}
> > > +
> > > +/* See read.h. */
> > > +
> > > +void
> > > +mapped_debug_line::map_filenames
> > > + (gdb::function_view<symbol_filename_ftype> fun,
> > > + bool need_fullname)
> > > +{
> > > + for (line_header_up &lh : line_headers)
> > > + for (file_entry &fe : lh->file_names ())
> > > + {
> > > + const char *filename = fe.name;
> > > +
> > > + if (!need_fullname)
> > > + {
> > > + fun (filename, nullptr);
> > > + continue;
> > > + }
> > > +
> > > + const char *dirname = fe.include_dir (&*lh);
> > > + std::string fullname;
> > > +
> > > + if (dirname == nullptr || IS_ABSOLUTE_PATH (filename))
> > > + fullname = filename;
> > > + else
> > > + fullname = std::string (dirname) + SLASH_STRING + filename;
> > > +
> > > + gdb::unique_xmalloc_ptr<char> rewritten
> > > + = rewrite_source_path (fullname.c_str ());
> > > + if (rewritten != nullptr)
> > > + fullname = rewritten.release ();
> > > +
> > > + fun (filename, fullname.c_str ());
> > > + }
> > > +}
> > > +
> > > +/* See read.h. */
> > > +
> > > +gdb::array_view<const gdb_byte>
> > > +mapped_debug_line::read_debug_line_separate
> > > + (const char *filename, std::unique_ptr<index_cache_resource> *resource)
> > > +{
> > > + if (filename == nullptr)
> > > + return {};
> > > +
> > > + try
> > > + {
> > > + line_resource_mmap *mmap_resource
> > > + = new line_resource_mmap (filename);
> > > +
> > > + resource->reset (mmap_resource);
> > > +
> > > + return gdb::array_view<const gdb_byte>
> > > + ((const gdb_byte *) mmap_resource->mapping.get (),
> > > + mmap_resource->mapping.size ());
> > > + }
> > > + catch (const gdb_exception &except)
> > > + {
> > > + exception_print (gdb_stderr, except);
> > > + }
> > > +
> > > + return {};
> > > +}
> > > +
> > > +/* See read.h. */
> > > +
> > > +bool
> > > +mapped_debug_line::read_debug_line_from_debuginfod (objfile *objfile)
> > > +{
> > > + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > > + if (build_id == nullptr)
> > > + return false;
> > > +
> > > + gdb::unique_xmalloc_ptr<char> line_path;
> > > + scoped_fd line_fd = debuginfod_section_query (build_id->data,
> > > + build_id->size,
> > > + bfd_get_filename
> > > + (objfile->obfd.get ()),
> > > + ".debug_line",
> > > + &line_path);
> > > +
> > > + if (line_fd.get () < 0)
> > > + return false;
> > > +
> > > + gdb::unique_xmalloc_ptr<char> line_str_path;
> > > + scoped_fd line_str_fd = debuginfod_section_query (build_id->data,
> > > + build_id->size,
> > > + bfd_get_filename
> > > + (objfile->obfd.get ()),
> > > + ".debug_line_str",
> > > + &line_str_path);
> > > +
> > > + line_data = read_debug_line_separate (line_path.get (), &line_resource);
> > > + line_str_data = read_debug_line_separate (line_str_path.get (),
> > > + &line_str_resource);
> > > +
> > > + const gdb_byte *line_ptr = line_data.data ();
> > > +
> > > + while (line_ptr < line_data.data () + line_data.size ())
> > > + {
> > > + line_header_up lh
> > > + = dwarf_decode_line_header (objfile->obfd.get (),
> > > + line_data, line_str_data,
> > > + &line_ptr, false,
> > > + nullptr, nullptr);
> > > + line_headers.emplace_back (std::move (lh));
> > > + }
> > > +
> > > + return true;
> > > +}
> > > +#endif /* !HAVE_SYS_MMAN_H */
> > > +
> > > +mapped_debug_line::mapped_debug_line (objfile *objfile)
> > > +{
> > > + if (!read_debug_line_from_debuginfod (objfile))
> > > + line_headers.clear ();
> > > +}
> > > +
> > > /* A helper for the "quick" functions which computes and caches the
> > > real path for a given file name from the line table. */
> > >
> > > diff --git a/gdb/dwarf2/read.h b/gdb/dwarf2/read.h
> > > index 099d3d65c27..19d19a6edc1 100644
> > > --- a/gdb/dwarf2/read.h
> > > +++ b/gdb/dwarf2/read.h
> > > @@ -33,6 +33,7 @@
> > > #include "gdbsupport/hash_enum.h"
> > > #include "gdbsupport/function-view.h"
> > > #include "gdbsupport/packed.h"
> > > +#include "dwarf2/line-header.h"
> > >
> > > /* Hold 'maintenance (set|show) dwarf' commands. */
> > > extern struct cmd_list_element *set_dwarf_cmdlist;
> > > @@ -942,4 +943,40 @@ extern htab_up create_quick_file_names_table (unsigned int nr_initial_entries);
> > > extern void read_full_dwarf_from_debuginfod (struct objfile *,
> > > dwarf2_base_index_functions *);
> > >
> > > +struct mapped_debug_line
> > > +{
> > > + mapped_debug_line (objfile *objfile);
> > > +
> > > + /* Return true if any of the mapped .debug_line's filenames match
> > > + FILE_MATCHER. */
> > > +
> > > + bool contains_matching_filename
> > > + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher);
> > > +
> > > + /* Call FUN with each filename in this mapped .debug_line. Include
> > > + each file's fullname if NEED_FULLNAME is true. */
> > > +
> > > + void map_filenames (gdb::function_view<symbol_filename_ftype> fun,
> > > + bool need_fullname);
> > > +
> > > +private:
> > > + std::vector<line_header_up> line_headers;
> > > +
> > > + gdb::array_view<const gdb_byte> line_data;
> > > + gdb::array_view<const gdb_byte> line_str_data;
> > > +
> > > + std::unique_ptr<index_cache_resource> line_resource;
> > > + std::unique_ptr<index_cache_resource> line_str_resource;
> > > +
> > > + /* Download the .debug_line and .debug_line_str associated with OBJFILE
> > > + and populate line_headers. */
> > > +
> > > + bool read_debug_line_from_debuginfod (objfile *objfile);
> > > +
> > > + /* Initialize line_data and line_str_data with the .debug_line and
> > > + .debug_line_str downloaded read_debug_line_from_debuginfod. */
> > > +
> > > + gdb::array_view<const gdb_byte> read_debug_line_separate
> > > + (const char *filename, std::unique_ptr<index_cache_resource> *resource);
> > > +};
> > > #endif /* DWARF2READ_H */
> > > diff --git a/gdb/mi/mi-out.c b/gdb/mi/mi-out.c
> > > index ff93d2cd448..8ad2c7f4497 100644
> > > --- a/gdb/mi/mi-out.c
> > > +++ b/gdb/mi/mi-out.c
> > > @@ -277,7 +277,14 @@ mi_ui_out::do_progress_notify (const std::string &msg, const char *unit,
> > >
> > > if (info.state == progress_update::START)
> > > {
> > > - gdb_printf ("%s...\n", msg.c_str ());
> > > + std::string prefix;
> > > + if (cur_prefix_state == prefix_state_t::NEWLINE_NEEDED)
> > > + {
> > > + prefix = "\n";
> > > + cur_prefix_state = prefix_state_t::NEWLINE_PRINTED;
> > > + }
> > > +
> > > + gdb_printf ("%s...\n", (prefix + msg).c_str ());
> > > info.state = progress_update::WORKING;
> > > }
> > > }
> > > diff --git a/gdb/testsuite/gdb.debuginfod/section.exp b/gdb/testsuite/gdb.debuginfod/section.exp
> > > index 35086d15705..a9dad927696 100644
> > > --- a/gdb/testsuite/gdb.debuginfod/section.exp
> > > +++ b/gdb/testsuite/gdb.debuginfod/section.exp
> > > @@ -18,6 +18,7 @@
> > > standard_testfile
> > >
> > > load_lib debuginfod-support.exp
> > > +load_lib completion-support.exp
> > >
> > > clean_restart
> > > require allow_debuginfod_tests
> > > @@ -144,6 +145,32 @@ proc_with_prefix local_url { } {
> > > # Hit the breakpoint.
> > > set res ".*Breakpoint 2, libsection2_test.*\"In ${libfile2}\\\\n\".*"
> > > gdb_test "c" $res "break continue"
> > > +
> > > + clean_restart_with_prompt $binfile "line 1"
> > > +
> > > + # List source file using .debug_line download.
> > > + set res ".*\.debug_line.*$lib_sl1.*21.*extern void libsection2_test.*"
> > > + gdb_test "list $libsrc1:21" $res "line 1 list"
> > > +
> > > + clean_restart_with_prompt $binfile "line 2"
> > > +
> > > + # Set breakpoint using .debug_line download.
> > > + set res ".*section \.debug_line for $lib_sl1.*Breakpoint 2 at.*$libsrc1.*"
> > > + gdb_test "br $libsrc1:41" $res "line 2 br"
> > > +
> > > + # Continue to breakpoint.
> > > + set res "Breakpoint 2, libsection1_test.*\"Cancelling thread\\\\n\".*"
> > > + gdb_test "c" $res "line 2 continue"
> > > +
> > > + # Check that download progress message is correctly formatted
> > > + # when printing threads.
> > > + set res ".*separate debug info for $lib_sl2\.\.\.\r\n.* 2 Thread.*"
> > > + gdb_test "info thr" $res "line thread"
> > > +
> > > + clean_restart_with_prompt $binfile "autocomplete"
> > > +
> > > + # Download debuginfo during autocompletion.
> > > + test_gdb_complete_tab_unique "br lib" ".*Downloading separate debug.*" ""
> > > }
> > >
> > > # Create CACHE and DB directories ready for debuginfod to use.
> > > diff --git a/gdb/ui-out.c b/gdb/ui-out.c
> > > index b3a2fb7f4e5..52d36d5bbf8 100644
> > > --- a/gdb/ui-out.c
> > > +++ b/gdb/ui-out.c
> > > @@ -31,6 +31,9 @@
> > > #include <memory>
> > > #include <string>
> > >
> > > +/* Current state of newline prefixing for progress update messages. */
> > > +prefix_state_t cur_prefix_state = prefix_state_t::NEWLINE_OFF;
> > > +
> > > namespace {
> > >
> > > /* A header of a ui_out_table. */
> > > diff --git a/gdb/ui-out.h b/gdb/ui-out.h
> > > index 5b6ddd55063..7e885cb1d38 100644
> > > --- a/gdb/ui-out.h
> > > +++ b/gdb/ui-out.h
> > > @@ -296,6 +296,21 @@ class ui_out
> > > BAR
> > > };
> > >
> > > + /* Used to communicate the status of a newline prefix for the next progress
> > > + update message. */
> > > + enum prefix_state
> > > + {
> > > + /* Do not modify the next progress update message. */
> > > + NEWLINE_OFF,
> > > +
> > > + /* The next progress update message should include a newline prefix. */
> > > + NEWLINE_NEEDED,
> > > +
> > > + /* A newline prefix was included in a debuginfod progress update
> > > + message. */
> > > + NEWLINE_PRINTED
> > > + };
> > > +
> > > /* SHOULD_PRINT indicates whether something should be printed for a tty. */
> > > progress_update ()
> > > {
> > > @@ -393,6 +408,11 @@ class ui_out
> > > ui_out_level *current_level () const;
> > > };
> > >
> > > +typedef ui_out::progress_update::prefix_state prefix_state_t;
> > > +
> > > +/* Current state of the newline prefix. */
> > > +extern prefix_state_t cur_prefix_state;
> > > +
> > > /* Start a new tuple or list on construction, and end it on
> > > destruction. Normally this is used via the typedefs
> > > ui_out_emit_tuple and ui_out_emit_list. */
> > > --
> > > 2.43.0
> > >
^ permalink raw reply [flat|nested] 18+ messages in thread* [PING*3][PATCH 3/3 v6] gdb/debuginfod: Add .debug_line downloading
2024-05-31 20:58 ` [PING*2][PATCH " Aaron Merey
2024-06-07 20:49 ` [PING*3][PATCH " Aaron Merey
@ 2024-06-14 18:24 ` Aaron Merey
1 sibling, 0 replies; 18+ messages in thread
From: Aaron Merey @ 2024-06-14 18:24 UTC (permalink / raw)
To: gdb-patches; +Cc: Andrew Burgess
Ping
Thanks,
Aaron
On Fri, May 31, 2024 at 4:58 PM Aaron Merey <amerey@redhat.com> wrote:
>
> Ping
>
> Thanks,
> Aaron
>
> On Fri, May 17, 2024 at 10:10 AM Aaron Merey <amerey@redhat.com> wrote:
> >
> > Ping
> >
> > Thanks,
> > Aaron
> >
> > On Fri, May 3, 2024 at 6:59 PM Aaron Merey <amerey@redhat.com> wrote:
> > >
> > > This is a repost of v6 from
> > > https://sourceware.org/pipermail/gdb-patches/2024-January/205956.html
> > >
> > > ELF/DWARF section downloading allows gdb to download .gdb_index files in
> > > order to defer full debuginfo downloads. However .gdb_index does not
> > > contain any information regarding source filenames. When a gdb command
> > > includes a filename argument (ex. 'break main.c:50'), this results in
> > > the mass downloading of all deferred debuginfo so that gdb can search the
> > > debuginfo for matching source filenames. This can result in unnecessary
> > > downloads.
> > >
> > > To improve this, have gdb instead download each debuginfo's .debug_line
> > > (and .debug_line_str if using DWARF5) when executing these commands.
> > > Download full debuginfo only when its .debug_line contains a matching
> > > filename.
> > >
> > > Since the combined size of .debug_line and .debug_line_str is only about
> > > 1% the size of the corresponding debuginfo, significant time can be saved
> > > by checking these sections before choosing to download an entire debuginfo.
> > >
> > > This patch also redirects stdout and stderr of the debuginfod server
> > > used by testsuite/gdb.debuginfod tests to a server_log standard output
> > > file. While adding tests for this patch I ran into an issue where the
> > > test server would block when logging to stderr, presumably because the
> > > stderr buffer filled up and wasn't being read from. Redirecting the
> > > log to a file fixes this and also makes the server log more accessible
> > > when debugging test failures.
> > > ---
> > > gdb/cli-out.c | 11 +-
> > > gdb/completer.c | 18 +-
> > > gdb/dwarf2/line-header.c | 161 ++++++++++++------
> > > gdb/dwarf2/line-header.h | 10 ++
> > > gdb/dwarf2/read-gdb-index.c | 60 +++++++
> > > gdb/dwarf2/read.c | 208 +++++++++++++++++++++++
> > > gdb/dwarf2/read.h | 37 ++++
> > > gdb/mi/mi-out.c | 9 +-
> > > gdb/testsuite/gdb.debuginfod/section.exp | 27 +++
> > > gdb/ui-out.c | 3 +
> > > gdb/ui-out.h | 20 +++
> > > 11 files changed, 511 insertions(+), 53 deletions(-)
> > >
> > > diff --git a/gdb/cli-out.c b/gdb/cli-out.c
> > > index 1c303f09662..d65b69ebe01 100644
> > > --- a/gdb/cli-out.c
> > > +++ b/gdb/cli-out.c
> > > @@ -306,16 +306,23 @@ cli_ui_out::do_progress_notify (const std::string &msg,
> > >
> > > if (info.state == progress_update::START)
> > > {
> > > + std::string prefix;
> > > + if (cur_prefix_state == prefix_state_t::NEWLINE_NEEDED)
> > > + {
> > > + prefix = "\n";
> > > + cur_prefix_state = prefix_state_t::NEWLINE_PRINTED;
> > > + }
> > > +
> > > if (stream->isatty ()
> > > && current_ui->input_interactive_p ()
> > > && chars_per_line >= MIN_CHARS_PER_LINE)
> > > {
> > > - gdb_printf (stream, "%s\n", msg.c_str ());
> > > + gdb_printf (stream, "%s\n", (prefix + msg).c_str ());
> > > info.state = progress_update::BAR;
> > > }
> > > else
> > > {
> > > - gdb_printf (stream, "%s...\n", msg.c_str ());
> > > + gdb_printf (stream, "%s...\n", (prefix + msg).c_str ());
> > > info.state = progress_update::WORKING;
> > > }
> > > }
> > > diff --git a/gdb/completer.c b/gdb/completer.c
> > > index f1f44109bdc..708081cf280 100644
> > > --- a/gdb/completer.c
> > > +++ b/gdb/completer.c
> > > @@ -1354,6 +1354,10 @@ complete_line_internal_1 (completion_tracker &tracker,
> > > {
> > > /* We've recognized a full command. */
> > >
> > > + /* Disable pagination since responding to the pagination prompt
> > > + overwrites rl_line_buffer. */
> > > + scoped_restore pag_restore = make_scoped_restore (&pagination_enabled, false);
> > > +
> > > if (p == tmp_command + point)
> > > {
> > > /* There is no non-whitespace in the line beyond the
> > > @@ -1453,7 +1457,8 @@ complete_line_internal_1 (completion_tracker &tracker,
> > > }
> > >
> > > /* Wrapper around complete_line_internal_1 to handle
> > > - MAX_COMPLETIONS_REACHED_ERROR. */
> > > + MAX_COMPLETIONS_REACHED_ERROR and possible progress update
> > > + interactions. */
> > >
> > > static void
> > > complete_line_internal (completion_tracker &tracker,
> > > @@ -1461,6 +1466,11 @@ complete_line_internal (completion_tracker &tracker,
> > > const char *line_buffer, int point,
> > > complete_line_internal_reason reason)
> > > {
> > > + scoped_restore restore_prefix_state
> > > + = make_scoped_restore
> > > + (&cur_prefix_state,
> > > + ui_out::progress_update::prefix_state::NEWLINE_NEEDED);
> > > +
> > > try
> > > {
> > > complete_line_internal_1 (tracker, text, line_buffer, point, reason);
> > > @@ -1470,6 +1480,12 @@ complete_line_internal (completion_tracker &tracker,
> > > if (except.error != MAX_COMPLETIONS_REACHED_ERROR)
> > > throw;
> > > }
> > > +
> > > + /* If progress update messages printed, then the text being completed
> > > + needs to be printed again. */
> > > + if (cur_prefix_state
> > > + == ui_out::progress_update::prefix_state::NEWLINE_PRINTED)
> > > + rl_forced_update_display ();
> > > }
> > >
> > > /* See completer.h. */
> > > diff --git a/gdb/dwarf2/line-header.c b/gdb/dwarf2/line-header.c
> > > index eddb2ef7ae8..1d24386c0eb 100644
> > > --- a/gdb/dwarf2/line-header.c
> > > +++ b/gdb/dwarf2/line-header.c
> > > @@ -101,12 +101,15 @@ read_checked_initial_length_and_offset (bfd *abfd, const gdb_byte *buf,
> > > {
> > > LONGEST length = read_initial_length (abfd, buf, bytes_read);
> > >
> > > - gdb_assert (cu_header->initial_length_size == 4
> > > - || cu_header->initial_length_size == 8
> > > - || cu_header->initial_length_size == 12);
> > > + if (cu_header != nullptr)
> > > + {
> > > + gdb_assert (cu_header->initial_length_size == 4
> > > + || cu_header->initial_length_size == 8
> > > + || cu_header->initial_length_size == 12);
> > >
> > > - if (cu_header->initial_length_size != *bytes_read)
> > > - complaint (_("intermixed 32-bit and 64-bit DWARF sections"));
> > > + if (cu_header->initial_length_size != *bytes_read)
> > > + complaint (_("intermixed 32-bit and 64-bit DWARF sections"));
> > > + }
> > >
> > > *offset_size = (*bytes_read == 4) ? 4 : 8;
> > > return length;
> > > @@ -115,21 +118,27 @@ read_checked_initial_length_and_offset (bfd *abfd, const gdb_byte *buf,
> > > /* Read directory or file name entry format, starting with byte of
> > > format count entries, ULEB128 pairs of entry formats, ULEB128 of
> > > entries count and the entries themselves in the described entry
> > > - format. */
> > > + format.
> > > +
> > > + .debug_line and .debug_line_str sections are stored in LINE_BUFP
> > > + and LINE_STR_DATA respectively. */
> > >
> > > static void
> > > -read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > > - const gdb_byte **bufp, struct line_header *lh,
> > > - unsigned int offset_size,
> > > - void (*callback) (struct line_header *lh,
> > > - const char *name,
> > > - dir_index d_index,
> > > - unsigned int mod_time,
> > > - unsigned int length))
> > > +read_formatted_entries
> > > + (bfd *abfd, const gdb_byte **line_bufp,
> > > + const gdb::array_view<const gdb_byte> line_str_data,
> > > + struct line_header *lh,
> > > + unsigned int offset_size,
> > > + void (*callback) (struct line_header *lh,
> > > + const char *name,
> > > + dir_index d_index,
> > > + unsigned int mod_time,
> > > + unsigned int length))
> > > {
> > > gdb_byte format_count, formati;
> > > ULONGEST data_count, datai;
> > > - const gdb_byte *buf = *bufp;
> > > + const gdb_byte *buf = *line_bufp;
> > > + const gdb_byte *str_buf = line_str_data.data ();
> > > const gdb_byte *format_header_data;
> > > unsigned int bytes_read;
> > >
> > > @@ -156,7 +165,7 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > > ULONGEST content_type = read_unsigned_leb128 (abfd, format, &bytes_read);
> > > format += bytes_read;
> > >
> > > - ULONGEST form = read_unsigned_leb128 (abfd, format, &bytes_read);
> > > + ULONGEST form = read_unsigned_leb128 (abfd, format, &bytes_read);
> > > format += bytes_read;
> > >
> > > std::optional<const char *> string;
> > > @@ -171,12 +180,24 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > >
> > > case DW_FORM_line_strp:
> > > {
> > > - const char *str
> > > - = per_objfile->read_line_string (buf, offset_size);
> > > + if (line_str_data.empty ())
> > > + warning (_("Dwarf Error: DW_FORM_line_strp used without " \
> > > + "required section"));
> > > +
> > > + if (line_str_data.size () <= offset_size)
> > > + warning (_("Dwarf Error: DW_FORM_line_strp pointing outside " \
> > > + "of section .debug_line_str"));
> > > +
> > > + ULONGEST str_offset = read_offset (abfd, buf, offset_size);
> > > + const char *str;
> > > + if (str_buf[str_offset] == '\0')
> > > + str = nullptr;
> > > + else
> > > + str = (const char *) (str_buf + str_offset);
> > > string.emplace (str);
> > > buf += offset_size;
> > > + break;
> > > }
> > > - break;
> > >
> > > case DW_FORM_data1:
> > > uint.emplace (read_1_byte (abfd, buf));
> > > @@ -247,28 +268,30 @@ read_formatted_entries (dwarf2_per_objfile *per_objfile, bfd *abfd,
> > > callback (lh, fe.name, fe.d_index, fe.mod_time, fe.length);
> > > }
> > >
> > > - *bufp = buf;
> > > + *line_bufp = buf;
> > > }
> > >
> > > /* See line-header.h. */
> > >
> > > line_header_up
> > > -dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > - dwarf2_per_objfile *per_objfile,
> > > - struct dwarf2_section_info *section,
> > > - const struct comp_unit_head *cu_header,
> > > - const char *comp_dir)
> > > +dwarf_decode_line_header (bfd *abfd,
> > > + gdb::array_view<const gdb_byte> line_data,
> > > + gdb::array_view<const gdb_byte> line_str_data,
> > > + const gdb_byte **debug_line_ptr,
> > > + bool is_dwz,
> > > + const struct comp_unit_head *cu_header,
> > > + const char *comp_dir)
> > > {
> > > - const gdb_byte *line_ptr;
> > > + const gdb_byte *line_ptr, *buf;
> > > unsigned int bytes_read, offset_size;
> > > int i;
> > > const char *cur_dir, *cur_file;
> > >
> > > - bfd *abfd = section->get_bfd_owner ();
> > > + buf = *debug_line_ptr;
> > >
> > > /* Make sure that at least there's room for the total_length field.
> > > That could be 12 bytes long, but we're just going to fudge that. */
> > > - if (to_underlying (sect_off) + 4 >= section->size)
> > > + if (buf + 4 >= line_data.data () + line_data.size ())
> > > {
> > > dwarf2_statement_list_fits_in_line_number_section_complaint ();
> > > return 0;
> > > @@ -276,24 +299,26 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > >
> > > line_header_up lh (new line_header (comp_dir));
> > >
> > > - lh->sect_off = sect_off;
> > > + lh->sect_off = (sect_offset) (buf - line_data.data ());
> > > lh->offset_in_dwz = is_dwz;
> > >
> > > - line_ptr = section->buffer + to_underlying (sect_off);
> > > + line_ptr = buf;
> > >
> > > /* Read in the header. */
> > > LONGEST unit_length
> > > - = read_checked_initial_length_and_offset (abfd, line_ptr, cu_header,
> > > + = read_checked_initial_length_and_offset (abfd, buf, cu_header,
> > > &bytes_read, &offset_size);
> > > - line_ptr += bytes_read;
> > >
> > > - const gdb_byte *start_here = line_ptr;
> > > + line_ptr += bytes_read;
> > >
> > > - if (line_ptr + unit_length > (section->buffer + section->size))
> > > + if (line_ptr + unit_length > buf + line_data.size ())
> > > {
> > > dwarf2_statement_list_fits_in_line_number_section_complaint ();
> > > return 0;
> > > }
> > > +
> > > + const gdb_byte *start_here = line_ptr;
> > > +
> > > lh->statement_program_end = start_here + unit_length;
> > > lh->version = read_2_bytes (abfd, line_ptr);
> > > line_ptr += 2;
> > > @@ -302,7 +327,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > /* This is a version we don't understand. The format could have
> > > changed in ways we don't handle properly so just punt. */
> > > complaint (_("unsupported version in .debug_line section"));
> > > - return NULL;
> > > + return nullptr;
> > > }
> > > if (lh->version >= 5)
> > > {
> > > @@ -319,13 +344,14 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > complaint (_("unsupported segment selector size %u "
> > > "in .debug_line section"),
> > > segment_selector_size);
> > > - return NULL;
> > > + return nullptr;
> > > }
> > > }
> > >
> > > LONGEST header_length = read_offset (abfd, line_ptr, offset_size);
> > > line_ptr += offset_size;
> > > lh->statement_program_start = line_ptr + header_length;
> > > +
> > > lh->minimum_instruction_length = read_1_byte (abfd, line_ptr);
> > > line_ptr += 1;
> > >
> > > @@ -346,12 +372,16 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > >
> > > lh->default_is_stmt = read_1_byte (abfd, line_ptr);
> > > line_ptr += 1;
> > > +
> > > lh->line_base = read_1_signed_byte (abfd, line_ptr);
> > > line_ptr += 1;
> > > +
> > > lh->line_range = read_1_byte (abfd, line_ptr);
> > > line_ptr += 1;
> > > +
> > > lh->opcode_base = read_1_byte (abfd, line_ptr);
> > > line_ptr += 1;
> > > +
> > > lh->standard_opcode_lengths.reset (new unsigned char[lh->opcode_base]);
> > >
> > > lh->standard_opcode_lengths[0] = 1; /* This should never be used anyway. */
> > > @@ -364,21 +394,23 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > if (lh->version >= 5)
> > > {
> > > /* Read directory table. */
> > > - read_formatted_entries (per_objfile, abfd, &line_ptr, lh.get (),
> > > - offset_size,
> > > - [] (struct line_header *header, const char *name,
> > > - dir_index d_index, unsigned int mod_time,
> > > - unsigned int length)
> > > + read_formatted_entries
> > > + (abfd, &line_ptr, line_str_data,
> > > + lh.get (), offset_size,
> > > + [] (struct line_header *header, const char *name,
> > > + dir_index d_index, unsigned int mod_time,
> > > + unsigned int length)
> > > {
> > > header->add_include_dir (name);
> > > });
> > >
> > > /* Read file name table. */
> > > - read_formatted_entries (per_objfile, abfd, &line_ptr, lh.get (),
> > > - offset_size,
> > > - [] (struct line_header *header, const char *name,
> > > - dir_index d_index, unsigned int mod_time,
> > > - unsigned int length)
> > > + read_formatted_entries
> > > + (abfd, &line_ptr, line_str_data,
> > > + lh.get (), offset_size,
> > > + [] (struct line_header *header, const char *name,
> > > + dir_index d_index, unsigned int mod_time,
> > > + unsigned int length)
> > > {
> > > header->add_file_name (name, d_index, mod_time, length);
> > > });
> > > @@ -386,7 +418,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > else
> > > {
> > > /* Read directory table. */
> > > - while ((cur_dir = read_direct_string (abfd, line_ptr, &bytes_read)) != NULL)
> > > + while ((cur_dir = read_direct_string (abfd, line_ptr, &bytes_read)) != nullptr)
> > > {
> > > line_ptr += bytes_read;
> > > lh->add_include_dir (cur_dir);
> > > @@ -394,7 +426,7 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > line_ptr += bytes_read;
> > >
> > > /* Read file name table. */
> > > - while ((cur_file = read_direct_string (abfd, line_ptr, &bytes_read)) != NULL)
> > > + while ((cur_file = read_direct_string (abfd, line_ptr, &bytes_read)) != nullptr)
> > > {
> > > unsigned int mod_time, length;
> > > dir_index d_index;
> > > @@ -412,9 +444,40 @@ dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > line_ptr += bytes_read;
> > > }
> > >
> > > - if (line_ptr > (section->buffer + section->size))
> > > + if (line_ptr > (buf + line_data.size ()))
> > > complaint (_("line number info header doesn't "
> > > "fit in `.debug_line' section"));
> > >
> > > + *debug_line_ptr += unit_length + offset_size;
> > > return lh;
> > > }
> > > +
> > > +line_header_up
> > > +dwarf_decode_line_header (sect_offset sect_off, bool is_dwz,
> > > + dwarf2_per_objfile *per_objfile,
> > > + struct dwarf2_section_info *section,
> > > + const struct comp_unit_head *cu_header,
> > > + const char *comp_dir)
> > > +{
> > > + struct objfile *objfile = per_objfile->objfile;
> > > + struct dwarf2_per_bfd *per_bfd = per_objfile->per_bfd;
> > > +
> > > + /* Read .debug_line. */
> > > + dwarf2_section_info *line_sec = &per_bfd->line;
> > > + bfd_size_type line_size = line_sec->get_size (objfile);
> > > +
> > > + gdb::array_view<const gdb_byte> line (line_sec->buffer, line_size);
> > > +
> > > + /* Read .debug_line_str. */
> > > + dwarf2_section_info *line_str_sec = &per_bfd->line_str;
> > > + bfd_size_type line_str_size = line_str_sec->get_size (objfile);
> > > +
> > > + gdb::array_view<const gdb_byte> line_str (line_str_sec->buffer,
> > > + line_str_size);
> > > +
> > > + const gdb_byte *line_ptr = line.data () + to_underlying (sect_off);
> > > +
> > > + return dwarf_decode_line_header
> > > + (per_bfd->obfd, line, line_str, &line_ptr,
> > > + is_dwz, cu_header, comp_dir);
> > > +}
> > > diff --git a/gdb/dwarf2/line-header.h b/gdb/dwarf2/line-header.h
> > > index c068dff70a3..45011882321 100644
> > > --- a/gdb/dwarf2/line-header.h
> > > +++ b/gdb/dwarf2/line-header.h
> > > @@ -217,4 +217,14 @@ extern line_header_up dwarf_decode_line_header
> > > struct dwarf2_section_info *section, const struct comp_unit_head *cu_header,
> > > const char *comp_dir);
> > >
> > > +/* Like above but the .debug_line and .debug_line_str are stored in
> > > + LINE_DATA and LINE_STR_DATA. *DEBUG_LINE_PTR should point to a
> > > + statement program header within LINE_DATA. */
> > > +
> > > +extern line_header_up dwarf_decode_line_header
> > > + (bfd *abfd, gdb::array_view<const gdb_byte> line_data,
> > > + gdb::array_view<const gdb_byte> line_str_data,
> > > + const gdb_byte **debug_line_ptr, bool is_dwz,
> > > + const comp_unit_head *cu_header, const char *comp_dir);
> > > +
> > > #endif /* DWARF2_LINE_HEADER_H */
> > > diff --git a/gdb/dwarf2/read-gdb-index.c b/gdb/dwarf2/read-gdb-index.c
> > > index 97b8137aabf..c52099d5f6a 100644
> > > --- a/gdb/dwarf2/read-gdb-index.c
> > > +++ b/gdb/dwarf2/read-gdb-index.c
> > > @@ -131,6 +131,9 @@ struct mapped_gdb_index final : public mapped_index_base
> > > }
> > > };
> > >
> > > +struct mapped_debug_line;
> > > +typedef std::unique_ptr<mapped_debug_line> mapped_debug_line_up;
> > > +
> > > struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > > {
> > > /* This dumps minimal information about the index.
> > > @@ -160,6 +163,15 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > > block_search_flags search_flags,
> > > domain_search_flags domain);
> > >
> > > + /* If OBJFILE's debuginfo download has been deferred, use a mapped_debug_line
> > > + to generate filenames.
> > > +
> > > + Otherwise call dwarf2_base_index_functions::map_symbol_filenames. */
> > > +
> > > + void map_symbol_filenames (struct objfile *objfile,
> > > + gdb::function_view<symbol_filename_ftype> fun,
> > > + bool need_fullname) override;
> > > +
> > > /* Calls dwarf2_base_index_functions::expand_all_symtabs and downloads
> > > debuginfo if necessary. */
> > > void expand_all_symtabs (struct objfile *objfile) override;
> > > @@ -167,6 +179,15 @@ struct dwarf2_gdb_index : public dwarf2_base_index_functions
> > > /* Calls dwarf2_base_index_functions::find_last_source_symtab and downloads
> > > debuginfo if necessary. */
> > > struct symtab *find_last_source_symtab (struct objfile *objfile) override;
> > > +
> > > + /* Filename information related to this .gdb_index. */
> > > + mapped_debug_line_up mdl;
> > > +
> > > + /* Return true if any of the filenames in this .gdb_index's .debug_line
> > > + mapping match FILE_MATCHER. Initializes the mapping if necessary. */
> > > + bool filename_in_debug_line
> > > + (objfile *objfile,
> > > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher);
> > > };
> > >
> > > void
> > > @@ -202,6 +223,30 @@ dwarf2_gdb_index::find_last_source_symtab (struct objfile *objfile)
> > > }
> > > }
> > >
> > > +void
> > > +dwarf2_gdb_index::map_symbol_filenames
> > > + (struct objfile *objfile,
> > > + gdb::function_view<symbol_filename_ftype> fun,
> > > + bool need_fullname)
> > > +{
> > > + try
> > > + {
> > > + dwarf2_base_index_functions::map_symbol_filenames (objfile, fun,
> > > + need_fullname);
> > > + }
> > > + catch (const gdb_exception &e)
> > > + {
> > > + if ((objfile->flags & OBJF_DOWNLOAD_DEFERRED) == 0)
> > > + exception_print (gdb_stderr, e);
> > > + else
> > > + {
> > > + if (mdl == nullptr)
> > > + mdl.reset (new mapped_debug_line (objfile));
> > > + mdl->map_filenames (fun, need_fullname);
> > > + }
> > > + }
> > > +}
> > > +
> > > /* This dumps minimal information about the index.
> > > It is called via "mt print objfiles".
> > > One use is to verify .gdb_index has been loaded by the
> > > @@ -365,6 +410,17 @@ dwarf2_gdb_index::do_expand_symtabs_matching
> > > return result;
> > > }
> > >
> > > +bool
> > > +dwarf2_gdb_index::filename_in_debug_line
> > > + (objfile *objfile,
> > > + gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> > > +{
> > > + if (mdl == nullptr)
> > > + mdl.reset (new mapped_debug_line (objfile));
> > > +
> > > + return mdl->contains_matching_filename (file_matcher);
> > > +}
> > > +
> > > bool
> > > dwarf2_gdb_index::expand_symtabs_matching
> > > (struct objfile *objfile,
> > > @@ -392,6 +448,10 @@ dwarf2_gdb_index::expand_symtabs_matching
> > > return false;
> > > }
> > >
> > > + if (file_matcher != nullptr
> > > + && !filename_in_debug_line (objfile, file_matcher))
> > > + return true;
> > > +
> > > read_full_dwarf_from_debuginfod (objfile, this);
> > > return true;
> > > }
> > > diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
> > > index 4461c76347e..ca219c9b910 100644
> > > --- a/gdb/dwarf2/read.c
> > > +++ b/gdb/dwarf2/read.c
> > > @@ -81,6 +81,7 @@
> > > #include <optional>
> > > #include "gdbsupport/underlying.h"
> > > #include "gdbsupport/hash_enum.h"
> > > +#include "gdbsupport/scoped_mmap.h"
> > > #include "filename-seen-cache.h"
> > > #include "producer.h"
> > > #include <fcntl.h>
> > > @@ -1967,6 +1968,213 @@ dw2_get_file_names (dwarf2_per_cu_data *this_cu,
> > > return this_cu->file_names;
> > > }
> > >
> > > +#if !HAVE_SYS_MMAN_H
> > > +
> > > +bool
> > > +mapped_debug_line::contains_matching_filename
> > > + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> > > +{
> > > + return false;
> > > +}
> > > +
> > > +gdb::array_view<const gdb_byte>
> > > +mapped_debug_line::read_debug_line_separate
> > > + (const char *filename, std::unique_ptr<index_cache_resource> *resource)
> > > +{
> > > + return {};
> > > +}
> > > +
> > > +bool
> > > +mapped_debug_line::read_debug_line_from_debuginfod (objfile *objfile)
> > > +{
> > > + return false;
> > > +}
> > > +
> > > +void
> > > +mapped_debug_line::map_filenames
> > > + (gdb::function_view<symbol_filename_ftype> fun,
> > > + bool need_fullname)
> > > +{
> > > + return;
> > > +}
> > > +
> > > +#else /* !HAVE_SYS_MMAN_H */
> > > +
> > > +struct line_resource_mmap final : public index_cache_resource
> > > +{
> > > + /* Try to mmap FILENAME. Throw an exception on failure, including if the
> > > + file doesn't exist. */
> > > + line_resource_mmap (const char *filename)
> > > + : mapping (mmap_file (filename))
> > > + {}
> > > +
> > > + scoped_mmap mapping;
> > > +};
> > > +
> > > +/* See read.h. */
> > > +
> > > +bool
> > > +mapped_debug_line::contains_matching_filename
> > > + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher)
> > > +{
> > > + for (line_header_up &lh : line_headers)
> > > + for (file_entry &fe : lh->file_names ())
> > > + {
> > > + const char *filename = fe.name;
> > > +
> > > + if (file_matcher (fe.name, false))
> > > + return true;
> > > +
> > > + bool basename_match = file_matcher (lbasename (fe.name), true);
> > > +
> > > + if (!basenames_may_differ && !basename_match)
> > > + continue;
> > > +
> > > + /* DW_AT_comp_dir is not explicitly mentioned in the .debug_line
> > > + until DWARF5. Since we don't have access to the CU at this
> > > + point we just check for a partial match on the filename.
> > > + If there is a match, the full debuginfo will be downloaded
> > > + ane the match will be re-evalute with DW_AT_comp_dir. */
> > > + if (lh->version < 5 && fe.d_index == 0)
> > > + return basename_match;
> > > +
> > > + const char *dirname = fe.include_dir (&*lh);
> > > + std::string fullname;
> > > +
> > > + if (dirname == nullptr || IS_ABSOLUTE_PATH (filename))
> > > + fullname = filename;
> > > + else
> > > + fullname = std::string (dirname) + SLASH_STRING + filename;
> > > +
> > > + gdb::unique_xmalloc_ptr<char> rewritten
> > > + = rewrite_source_path (fullname.c_str ());
> > > + if (rewritten != nullptr)
> > > + fullname = rewritten.release ();
> > > +
> > > + if (file_matcher (fullname.c_str (), false))
> > > + return true;
> > > + }
> > > +
> > > + return false;
> > > +}
> > > +
> > > +/* See read.h. */
> > > +
> > > +void
> > > +mapped_debug_line::map_filenames
> > > + (gdb::function_view<symbol_filename_ftype> fun,
> > > + bool need_fullname)
> > > +{
> > > + for (line_header_up &lh : line_headers)
> > > + for (file_entry &fe : lh->file_names ())
> > > + {
> > > + const char *filename = fe.name;
> > > +
> > > + if (!need_fullname)
> > > + {
> > > + fun (filename, nullptr);
> > > + continue;
> > > + }
> > > +
> > > + const char *dirname = fe.include_dir (&*lh);
> > > + std::string fullname;
> > > +
> > > + if (dirname == nullptr || IS_ABSOLUTE_PATH (filename))
> > > + fullname = filename;
> > > + else
> > > + fullname = std::string (dirname) + SLASH_STRING + filename;
> > > +
> > > + gdb::unique_xmalloc_ptr<char> rewritten
> > > + = rewrite_source_path (fullname.c_str ());
> > > + if (rewritten != nullptr)
> > > + fullname = rewritten.release ();
> > > +
> > > + fun (filename, fullname.c_str ());
> > > + }
> > > +}
> > > +
> > > +/* See read.h. */
> > > +
> > > +gdb::array_view<const gdb_byte>
> > > +mapped_debug_line::read_debug_line_separate
> > > + (const char *filename, std::unique_ptr<index_cache_resource> *resource)
> > > +{
> > > + if (filename == nullptr)
> > > + return {};
> > > +
> > > + try
> > > + {
> > > + line_resource_mmap *mmap_resource
> > > + = new line_resource_mmap (filename);
> > > +
> > > + resource->reset (mmap_resource);
> > > +
> > > + return gdb::array_view<const gdb_byte>
> > > + ((const gdb_byte *) mmap_resource->mapping.get (),
> > > + mmap_resource->mapping.size ());
> > > + }
> > > + catch (const gdb_exception &except)
> > > + {
> > > + exception_print (gdb_stderr, except);
> > > + }
> > > +
> > > + return {};
> > > +}
> > > +
> > > +/* See read.h. */
> > > +
> > > +bool
> > > +mapped_debug_line::read_debug_line_from_debuginfod (objfile *objfile)
> > > +{
> > > + const bfd_build_id *build_id = build_id_bfd_get (objfile->obfd.get ());
> > > + if (build_id == nullptr)
> > > + return false;
> > > +
> > > + gdb::unique_xmalloc_ptr<char> line_path;
> > > + scoped_fd line_fd = debuginfod_section_query (build_id->data,
> > > + build_id->size,
> > > + bfd_get_filename
> > > + (objfile->obfd.get ()),
> > > + ".debug_line",
> > > + &line_path);
> > > +
> > > + if (line_fd.get () < 0)
> > > + return false;
> > > +
> > > + gdb::unique_xmalloc_ptr<char> line_str_path;
> > > + scoped_fd line_str_fd = debuginfod_section_query (build_id->data,
> > > + build_id->size,
> > > + bfd_get_filename
> > > + (objfile->obfd.get ()),
> > > + ".debug_line_str",
> > > + &line_str_path);
> > > +
> > > + line_data = read_debug_line_separate (line_path.get (), &line_resource);
> > > + line_str_data = read_debug_line_separate (line_str_path.get (),
> > > + &line_str_resource);
> > > +
> > > + const gdb_byte *line_ptr = line_data.data ();
> > > +
> > > + while (line_ptr < line_data.data () + line_data.size ())
> > > + {
> > > + line_header_up lh
> > > + = dwarf_decode_line_header (objfile->obfd.get (),
> > > + line_data, line_str_data,
> > > + &line_ptr, false,
> > > + nullptr, nullptr);
> > > + line_headers.emplace_back (std::move (lh));
> > > + }
> > > +
> > > + return true;
> > > +}
> > > +#endif /* !HAVE_SYS_MMAN_H */
> > > +
> > > +mapped_debug_line::mapped_debug_line (objfile *objfile)
> > > +{
> > > + if (!read_debug_line_from_debuginfod (objfile))
> > > + line_headers.clear ();
> > > +}
> > > +
> > > /* A helper for the "quick" functions which computes and caches the
> > > real path for a given file name from the line table. */
> > >
> > > diff --git a/gdb/dwarf2/read.h b/gdb/dwarf2/read.h
> > > index 099d3d65c27..19d19a6edc1 100644
> > > --- a/gdb/dwarf2/read.h
> > > +++ b/gdb/dwarf2/read.h
> > > @@ -33,6 +33,7 @@
> > > #include "gdbsupport/hash_enum.h"
> > > #include "gdbsupport/function-view.h"
> > > #include "gdbsupport/packed.h"
> > > +#include "dwarf2/line-header.h"
> > >
> > > /* Hold 'maintenance (set|show) dwarf' commands. */
> > > extern struct cmd_list_element *set_dwarf_cmdlist;
> > > @@ -942,4 +943,40 @@ extern htab_up create_quick_file_names_table (unsigned int nr_initial_entries);
> > > extern void read_full_dwarf_from_debuginfod (struct objfile *,
> > > dwarf2_base_index_functions *);
> > >
> > > +struct mapped_debug_line
> > > +{
> > > + mapped_debug_line (objfile *objfile);
> > > +
> > > + /* Return true if any of the mapped .debug_line's filenames match
> > > + FILE_MATCHER. */
> > > +
> > > + bool contains_matching_filename
> > > + (gdb::function_view<expand_symtabs_file_matcher_ftype> file_matcher);
> > > +
> > > + /* Call FUN with each filename in this mapped .debug_line. Include
> > > + each file's fullname if NEED_FULLNAME is true. */
> > > +
> > > + void map_filenames (gdb::function_view<symbol_filename_ftype> fun,
> > > + bool need_fullname);
> > > +
> > > +private:
> > > + std::vector<line_header_up> line_headers;
> > > +
> > > + gdb::array_view<const gdb_byte> line_data;
> > > + gdb::array_view<const gdb_byte> line_str_data;
> > > +
> > > + std::unique_ptr<index_cache_resource> line_resource;
> > > + std::unique_ptr<index_cache_resource> line_str_resource;
> > > +
> > > + /* Download the .debug_line and .debug_line_str associated with OBJFILE
> > > + and populate line_headers. */
> > > +
> > > + bool read_debug_line_from_debuginfod (objfile *objfile);
> > > +
> > > + /* Initialize line_data and line_str_data with the .debug_line and
> > > + .debug_line_str downloaded read_debug_line_from_debuginfod. */
> > > +
> > > + gdb::array_view<const gdb_byte> read_debug_line_separate
> > > + (const char *filename, std::unique_ptr<index_cache_resource> *resource);
> > > +};
> > > #endif /* DWARF2READ_H */
> > > diff --git a/gdb/mi/mi-out.c b/gdb/mi/mi-out.c
> > > index ff93d2cd448..8ad2c7f4497 100644
> > > --- a/gdb/mi/mi-out.c
> > > +++ b/gdb/mi/mi-out.c
> > > @@ -277,7 +277,14 @@ mi_ui_out::do_progress_notify (const std::string &msg, const char *unit,
> > >
> > > if (info.state == progress_update::START)
> > > {
> > > - gdb_printf ("%s...\n", msg.c_str ());
> > > + std::string prefix;
> > > + if (cur_prefix_state == prefix_state_t::NEWLINE_NEEDED)
> > > + {
> > > + prefix = "\n";
> > > + cur_prefix_state = prefix_state_t::NEWLINE_PRINTED;
> > > + }
> > > +
> > > + gdb_printf ("%s...\n", (prefix + msg).c_str ());
> > > info.state = progress_update::WORKING;
> > > }
> > > }
> > > diff --git a/gdb/testsuite/gdb.debuginfod/section.exp b/gdb/testsuite/gdb.debuginfod/section.exp
> > > index 35086d15705..a9dad927696 100644
> > > --- a/gdb/testsuite/gdb.debuginfod/section.exp
> > > +++ b/gdb/testsuite/gdb.debuginfod/section.exp
> > > @@ -18,6 +18,7 @@
> > > standard_testfile
> > >
> > > load_lib debuginfod-support.exp
> > > +load_lib completion-support.exp
> > >
> > > clean_restart
> > > require allow_debuginfod_tests
> > > @@ -144,6 +145,32 @@ proc_with_prefix local_url { } {
> > > # Hit the breakpoint.
> > > set res ".*Breakpoint 2, libsection2_test.*\"In ${libfile2}\\\\n\".*"
> > > gdb_test "c" $res "break continue"
> > > +
> > > + clean_restart_with_prompt $binfile "line 1"
> > > +
> > > + # List source file using .debug_line download.
> > > + set res ".*\.debug_line.*$lib_sl1.*21.*extern void libsection2_test.*"
> > > + gdb_test "list $libsrc1:21" $res "line 1 list"
> > > +
> > > + clean_restart_with_prompt $binfile "line 2"
> > > +
> > > + # Set breakpoint using .debug_line download.
> > > + set res ".*section \.debug_line for $lib_sl1.*Breakpoint 2 at.*$libsrc1.*"
> > > + gdb_test "br $libsrc1:41" $res "line 2 br"
> > > +
> > > + # Continue to breakpoint.
> > > + set res "Breakpoint 2, libsection1_test.*\"Cancelling thread\\\\n\".*"
> > > + gdb_test "c" $res "line 2 continue"
> > > +
> > > + # Check that download progress message is correctly formatted
> > > + # when printing threads.
> > > + set res ".*separate debug info for $lib_sl2\.\.\.\r\n.* 2 Thread.*"
> > > + gdb_test "info thr" $res "line thread"
> > > +
> > > + clean_restart_with_prompt $binfile "autocomplete"
> > > +
> > > + # Download debuginfo during autocompletion.
> > > + test_gdb_complete_tab_unique "br lib" ".*Downloading separate debug.*" ""
> > > }
> > >
> > > # Create CACHE and DB directories ready for debuginfod to use.
> > > diff --git a/gdb/ui-out.c b/gdb/ui-out.c
> > > index b3a2fb7f4e5..52d36d5bbf8 100644
> > > --- a/gdb/ui-out.c
> > > +++ b/gdb/ui-out.c
> > > @@ -31,6 +31,9 @@
> > > #include <memory>
> > > #include <string>
> > >
> > > +/* Current state of newline prefixing for progress update messages. */
> > > +prefix_state_t cur_prefix_state = prefix_state_t::NEWLINE_OFF;
> > > +
> > > namespace {
> > >
> > > /* A header of a ui_out_table. */
> > > diff --git a/gdb/ui-out.h b/gdb/ui-out.h
> > > index 5b6ddd55063..7e885cb1d38 100644
> > > --- a/gdb/ui-out.h
> > > +++ b/gdb/ui-out.h
> > > @@ -296,6 +296,21 @@ class ui_out
> > > BAR
> > > };
> > >
> > > + /* Used to communicate the status of a newline prefix for the next progress
> > > + update message. */
> > > + enum prefix_state
> > > + {
> > > + /* Do not modify the next progress update message. */
> > > + NEWLINE_OFF,
> > > +
> > > + /* The next progress update message should include a newline prefix. */
> > > + NEWLINE_NEEDED,
> > > +
> > > + /* A newline prefix was included in a debuginfod progress update
> > > + message. */
> > > + NEWLINE_PRINTED
> > > + };
> > > +
> > > /* SHOULD_PRINT indicates whether something should be printed for a tty. */
> > > progress_update ()
> > > {
> > > @@ -393,6 +408,11 @@ class ui_out
> > > ui_out_level *current_level () const;
> > > };
> > >
> > > +typedef ui_out::progress_update::prefix_state prefix_state_t;
> > > +
> > > +/* Current state of the newline prefix. */
> > > +extern prefix_state_t cur_prefix_state;
> > > +
> > > /* Start a new tuple or list on construction, and end it on
> > > destruction. Normally this is used via the typedefs
> > > ui_out_emit_tuple and ui_out_emit_list. */
> > > --
> > > 2.43.0
> > >
^ permalink raw reply [flat|nested] 18+ messages in thread