* [PATCH 01/13] Add a new overload of gdb_bfd_sections
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 16:34 ` Simon Marchi
2020-08-28 16:23 ` [PATCH 02/13] Use gdb_bfd_sections in core_target_open Tom Tromey
` (14 subsequent siblings)
15 siblings, 1 reply; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This adds a new overload of gdb_bfd_sections, that accepts a
gdb_bfd_ref_ptr. This also fixes the formatting of the existing
function, since I happened to notice it was mildly off.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* gdb_bfd.h (gdb_bfd_sections): New overload. Fix formatting of
existing function.
---
gdb/ChangeLog | 5 +++++
gdb/gdb_bfd.h | 10 ++++++++--
2 files changed, 13 insertions(+), 2 deletions(-)
diff --git a/gdb/gdb_bfd.h b/gdb/gdb_bfd.h
index 0f450567d50..12330970873 100644
--- a/gdb/gdb_bfd.h
+++ b/gdb/gdb_bfd.h
@@ -205,10 +205,16 @@ bool gdb_bfd_get_full_section_contents (bfd *abfd, asection *section,
using gdb_bfd_section_iterator = next_iterator<asection>;
using gdb_bfd_section_range = next_adapter<asection, gdb_bfd_section_iterator>;
-static inline
-gdb_bfd_section_range gdb_bfd_sections (bfd *abfd)
+static inline gdb_bfd_section_range
+gdb_bfd_sections (bfd *abfd)
{
return gdb_bfd_section_range (abfd->sections);
}
+static inline gdb_bfd_section_range
+gdb_bfd_sections (const gdb_bfd_ref_ptr &abfd)
+{
+ return gdb_bfd_section_range (abfd->sections);
+};
+
#endif /* GDB_BFD_H */
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH 01/13] Add a new overload of gdb_bfd_sections
2020-08-28 16:23 ` [PATCH 01/13] Add a new overload of gdb_bfd_sections Tom Tromey
@ 2020-08-28 16:34 ` Simon Marchi
2020-08-28 19:25 ` Tom Tromey
0 siblings, 1 reply; 25+ messages in thread
From: Simon Marchi @ 2020-08-28 16:34 UTC (permalink / raw)
To: Tom Tromey, gdb-patches
On 2020-08-28 12:23 p.m., Tom Tromey wrote:
> This adds a new overload of gdb_bfd_sections, that accepts a
> gdb_bfd_ref_ptr. This also fixes the formatting of the existing
> function, since I happened to notice it was mildly off.
Ah thanks for fixing the formatting. I'm confusing different styles for
different code bases I'm working with at the moment :).
>
> gdb/ChangeLog
> 2020-08-28 Tom Tromey <tom@tromey.com>
>
> * gdb_bfd.h (gdb_bfd_sections): New overload. Fix formatting of
> existing function.
> ---
> gdb/ChangeLog | 5 +++++
> gdb/gdb_bfd.h | 10 ++++++++--
> 2 files changed, 13 insertions(+), 2 deletions(-)
>
> diff --git a/gdb/gdb_bfd.h b/gdb/gdb_bfd.h
> index 0f450567d50..12330970873 100644
> --- a/gdb/gdb_bfd.h
> +++ b/gdb/gdb_bfd.h
> @@ -205,10 +205,16 @@ bool gdb_bfd_get_full_section_contents (bfd *abfd, asection *section,
> using gdb_bfd_section_iterator = next_iterator<asection>;
> using gdb_bfd_section_range = next_adapter<asection, gdb_bfd_section_iterator>;
>
> -static inline
> -gdb_bfd_section_range gdb_bfd_sections (bfd *abfd)
> +static inline gdb_bfd_section_range
> +gdb_bfd_sections (bfd *abfd)
> {
> return gdb_bfd_section_range (abfd->sections);
> }
>
> +static inline gdb_bfd_section_range
> +gdb_bfd_sections (const gdb_bfd_ref_ptr &abfd)
> +{
> + return gdb_bfd_section_range (abfd->sections);
> +};
> +
Just to confirm, the point of this is to avoid having to do `.get ()` everywhere?
(it LGTM)
Simon
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH 01/13] Add a new overload of gdb_bfd_sections
2020-08-28 16:34 ` Simon Marchi
@ 2020-08-28 19:25 ` Tom Tromey
0 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 19:25 UTC (permalink / raw)
To: Simon Marchi; +Cc: Tom Tromey, gdb-patches
Simon> Ah thanks for fixing the formatting. I'm confusing different styles for
Simon> different code bases I'm working with at the moment :).
It's definitely a pain when you work on more than one code base.
That happened to me when I was at Moz. Actually the Moz tree itself had
two different styles, which as you can imagine is impossible to get right.
>> +static inline gdb_bfd_section_range
>> +gdb_bfd_sections (const gdb_bfd_ref_ptr &abfd)
>> +{
>> + return gdb_bfd_section_range (abfd->sections);
>> +};
>> +
Simon> Just to confirm, the point of this is to avoid having to do `.get
Simon> ()` everywhere?
Yeah, it is just a convenience for exactly that reason.
Tom
^ permalink raw reply [flat|nested] 25+ messages in thread
* [PATCH 02/13] Use gdb_bfd_sections in core_target_open
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
2020-08-28 16:23 ` [PATCH 01/13] Add a new overload of gdb_bfd_sections Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 16:23 ` [PATCH 03/13] Use gdb_bfd_sections in gdb_bfd_close_or_warn Tom Tromey
` (13 subsequent siblings)
15 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes core_target_open to avoid bfd_map_over_sections, in favor
of iteration.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* corelow.c (add_to_thread_list): Change parameters.
(core_target_open): Use foreach.
---
gdb/ChangeLog | 5 +++++
gdb/corelow.c | 10 ++++------
2 files changed, 9 insertions(+), 6 deletions(-)
diff --git a/gdb/corelow.c b/gdb/corelow.c
index b6ee219f57c..1a07fcd0333 100644
--- a/gdb/corelow.c
+++ b/gdb/corelow.c
@@ -270,8 +270,6 @@ core_target::build_file_mappings ()
});
}
-static void add_to_thread_list (bfd *, asection *, void *);
-
/* An arbitrary identifier for the core inferior. */
#define CORELOW_PID 1
@@ -302,11 +300,10 @@ core_target::close ()
extract the list of threads in a core file. */
static void
-add_to_thread_list (bfd *abfd, asection *asect, void *reg_sect_arg)
+add_to_thread_list (asection *asect, asection *reg_sect)
{
int core_tid;
int pid, lwpid;
- asection *reg_sect = (asection *) reg_sect_arg;
bool fake_pid_p = false;
struct inferior *inf;
@@ -473,8 +470,9 @@ core_target_open (const char *arg, int from_tty)
/* Build up thread list from BFD sections, and possibly set the
current thread to the .reg/NN section matching the .reg
section. */
- bfd_map_over_sections (core_bfd, add_to_thread_list,
- bfd_get_section_by_name (core_bfd, ".reg"));
+ asection *reg_sect = bfd_get_section_by_name (core_bfd, ".reg");
+ for (asection *sect : gdb_bfd_sections (core_bfd))
+ add_to_thread_list (sect, reg_sect);
if (inferior_ptid == null_ptid)
{
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* [PATCH 03/13] Use gdb_bfd_sections in gdb_bfd_close_or_warn
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
2020-08-28 16:23 ` [PATCH 01/13] Add a new overload of gdb_bfd_sections Tom Tromey
2020-08-28 16:23 ` [PATCH 02/13] Use gdb_bfd_sections in core_target_open Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 16:23 ` [PATCH 04/13] Use gdb_bfd_sections in get_stap_base_address Tom Tromey
` (12 subsequent siblings)
15 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes gdb_bfd_close_or_warn to avoid bfd_map_over_sections, in
favor of iteration.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* gdb_bfd.c (free_one_bfd_section): Remove 'abfd' and 'ignore'
parameters.
(gdb_bfd_close_or_warn): Use foreach.
---
gdb/ChangeLog | 6 ++++++
gdb/gdb_bfd.c | 5 +++--
2 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/gdb/gdb_bfd.c b/gdb/gdb_bfd.c
index 25e0178a8b8..43b576f12a0 100644
--- a/gdb/gdb_bfd.c
+++ b/gdb/gdb_bfd.c
@@ -484,7 +484,7 @@ gdb_bfd_open (const char *name, const char *target, int fd,
BFD. */
static void
-free_one_bfd_section (bfd *abfd, asection *sectp, void *ignore)
+free_one_bfd_section (asection *sectp)
{
struct gdb_bfd_section_data *sect
= (struct gdb_bfd_section_data *) bfd_section_userdata (sectp);
@@ -513,7 +513,8 @@ gdb_bfd_close_or_warn (struct bfd *abfd)
int ret;
const char *name = bfd_get_filename (abfd);
- bfd_map_over_sections (abfd, free_one_bfd_section, NULL);
+ for (asection *sect : gdb_bfd_sections (abfd))
+ free_one_bfd_section (sect);
ret = bfd_close (abfd);
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* [PATCH 04/13] Use gdb_bfd_sections in get_stap_base_address
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (2 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 03/13] Use gdb_bfd_sections in gdb_bfd_close_or_warn Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 17:06 ` Simon Marchi
2020-08-28 16:23 ` [PATCH 05/13] Use gdb_bfd_sections in build_objfile_section_table Tom Tromey
` (11 subsequent siblings)
15 siblings, 1 reply; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes get_stap_base_address to avoid bfd_map_over_sections, in
favor of iteration.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* stap-probe.c (get_stap_base_address_1): Remove.
(get_stap_base_address): Use foreach.
---
gdb/ChangeLog | 5 +++++
gdb/stap-probe.c | 20 ++++++--------------
2 files changed, 11 insertions(+), 14 deletions(-)
diff --git a/gdb/stap-probe.c b/gdb/stap-probe.c
index 73596446cce..fb6010d14ff 100644
--- a/gdb/stap-probe.c
+++ b/gdb/stap-probe.c
@@ -1578,19 +1578,6 @@ handle_stap_probe (struct objfile *objfile, struct sdt_note *el,
probesp->emplace_back (ret);
}
-/* Helper function which tries to find the base address of the SystemTap
- base section named STAP_BASE_SECTION_NAME. */
-
-static void
-get_stap_base_address_1 (bfd *abfd, asection *sect, void *obj)
-{
- asection **ret = (asection **) obj;
-
- if ((sect->flags & (SEC_DATA | SEC_ALLOC | SEC_HAS_CONTENTS))
- && sect->name && !strcmp (sect->name, STAP_BASE_SECTION_NAME))
- *ret = sect;
-}
-
/* Helper function which iterates over every section in the BFD file,
trying to find the base address of the SystemTap base section.
Returns 1 if found (setting BASE to the proper value), zero otherwise. */
@@ -1600,7 +1587,12 @@ get_stap_base_address (bfd *obfd, bfd_vma *base)
{
asection *ret = NULL;
- bfd_map_over_sections (obfd, get_stap_base_address_1, (void *) &ret);
+ for (asection *sect : gdb_bfd_sections (obfd))
+ {
+ if ((sect->flags & (SEC_DATA | SEC_ALLOC | SEC_HAS_CONTENTS))
+ && sect->name && !strcmp (sect->name, STAP_BASE_SECTION_NAME))
+ ret = sect;
+ }
if (ret == NULL)
{
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH 04/13] Use gdb_bfd_sections in get_stap_base_address
2020-08-28 16:23 ` [PATCH 04/13] Use gdb_bfd_sections in get_stap_base_address Tom Tromey
@ 2020-08-28 17:06 ` Simon Marchi
0 siblings, 0 replies; 25+ messages in thread
From: Simon Marchi @ 2020-08-28 17:06 UTC (permalink / raw)
To: Tom Tromey, gdb-patches
On 2020-08-28 12:23 p.m., Tom Tromey wrote:
> This changes get_stap_base_address to avoid bfd_map_over_sections, in
> favor of iteration.
>
> gdb/ChangeLog
> 2020-08-28 Tom Tromey <tom@tromey.com>
>
> * stap-probe.c (get_stap_base_address_1): Remove.
> (get_stap_base_address): Use foreach.
> ---
> gdb/ChangeLog | 5 +++++
> gdb/stap-probe.c | 20 ++++++--------------
> 2 files changed, 11 insertions(+), 14 deletions(-)
>
> diff --git a/gdb/stap-probe.c b/gdb/stap-probe.c
> index 73596446cce..fb6010d14ff 100644
> --- a/gdb/stap-probe.c
> +++ b/gdb/stap-probe.c
> @@ -1578,19 +1578,6 @@ handle_stap_probe (struct objfile *objfile, struct sdt_note *el,
> probesp->emplace_back (ret);
> }
>
> -/* Helper function which tries to find the base address of the SystemTap
> - base section named STAP_BASE_SECTION_NAME. */
> -
> -static void
> -get_stap_base_address_1 (bfd *abfd, asection *sect, void *obj)
> -{
> - asection **ret = (asection **) obj;
> -
> - if ((sect->flags & (SEC_DATA | SEC_ALLOC | SEC_HAS_CONTENTS))
> - && sect->name && !strcmp (sect->name, STAP_BASE_SECTION_NAME))
> - *ret = sect;
> -}
> -
> /* Helper function which iterates over every section in the BFD file,
> trying to find the base address of the SystemTap base section.
> Returns 1 if found (setting BASE to the proper value), zero otherwise. */
> @@ -1600,7 +1587,12 @@ get_stap_base_address (bfd *obfd, bfd_vma *base)
> {
> asection *ret = NULL;
>
> - bfd_map_over_sections (obfd, get_stap_base_address_1, (void *) &ret);
> + for (asection *sect : gdb_bfd_sections (obfd))
> + {
> + if ((sect->flags & (SEC_DATA | SEC_ALLOC | SEC_HAS_CONTENTS))
> + && sect->name && !strcmp (sect->name, STAP_BASE_SECTION_NAME))
> + ret = sect;
> + }
This is where I meant to say:
FYI: I've been taught by Pedro [1] that we don't actually need curly braces in this case,
and I've been misleading people about this for a while . For future code I write like
this, I won't use the curly braces.
[1] https://sourceware.org/pipermail/gdb-patches/2020-August/171532.html
Simon
^ permalink raw reply [flat|nested] 25+ messages in thread
* [PATCH 05/13] Use gdb_bfd_sections in build_objfile_section_table
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (3 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 04/13] Use gdb_bfd_sections in get_stap_base_address Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 16:23 ` [PATCH 06/13] Use gdb_bfd_sections in symfile.c Tom Tromey
` (10 subsequent siblings)
15 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes build_objfile_section_table to avoid
bfd_map_over_sections, in favor of iteration.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* objfiles.c (add_to_objfile_sections): Rename from
add_to_objfile_sections_full.
(add_to_objfile_sections): Remove.
(build_objfile_section_table): Use foreach.
---
gdb/ChangeLog | 7 +++++++
gdb/objfiles.c | 29 +++++++++++------------------
2 files changed, 18 insertions(+), 18 deletions(-)
diff --git a/gdb/objfiles.c b/gdb/objfiles.c
index 3aa7973e0d5..471a5c4c393 100644
--- a/gdb/objfiles.c
+++ b/gdb/objfiles.c
@@ -257,13 +257,13 @@ objfile_lookup_static_link (struct objfile *objfile,
\f
-/* Called via bfd_map_over_sections to build up the section table that
- the objfile references. The objfile contains pointers to the start
- of the table (objfile->sections) and to the first location after
- the end of the table (objfile->sections_end). */
+/* Build up the section table that the objfile references. The
+ objfile contains pointers to the start of the table
+ (objfile->sections) and to the first location after the end of the
+ table (objfile->sections_end). */
static void
-add_to_objfile_sections_full (struct bfd *abfd, struct bfd_section *asect,
+add_to_objfile_sections (struct bfd *abfd, struct bfd_section *asect,
struct objfile *objfile, int force)
{
struct obj_section *section;
@@ -283,13 +283,6 @@ add_to_objfile_sections_full (struct bfd *abfd, struct bfd_section *asect,
section->ovly_mapped = 0;
}
-static void
-add_to_objfile_sections (struct bfd *abfd, struct bfd_section *asect,
- void *objfilep)
-{
- add_to_objfile_sections_full (abfd, asect, (struct objfile *) objfilep, 0);
-}
-
/* Builds a section table for OBJFILE.
Note that the OFFSET and OVLY_MAPPED in each table entry are
@@ -304,14 +297,14 @@ build_objfile_section_table (struct objfile *objfile)
count,
struct obj_section);
objfile->sections_end = (objfile->sections + count);
- bfd_map_over_sections (objfile->obfd,
- add_to_objfile_sections, (void *) objfile);
+ for (asection *sect : gdb_bfd_sections (objfile->obfd))
+ add_to_objfile_sections (objfile->obfd, sect, objfile, 0);
/* See gdb_bfd_section_index. */
- add_to_objfile_sections_full (objfile->obfd, bfd_com_section_ptr, objfile, 1);
- add_to_objfile_sections_full (objfile->obfd, bfd_und_section_ptr, objfile, 1);
- add_to_objfile_sections_full (objfile->obfd, bfd_abs_section_ptr, objfile, 1);
- add_to_objfile_sections_full (objfile->obfd, bfd_ind_section_ptr, objfile, 1);
+ add_to_objfile_sections (objfile->obfd, bfd_com_section_ptr, objfile, 1);
+ add_to_objfile_sections (objfile->obfd, bfd_und_section_ptr, objfile, 1);
+ add_to_objfile_sections (objfile->obfd, bfd_abs_section_ptr, objfile, 1);
+ add_to_objfile_sections (objfile->obfd, bfd_ind_section_ptr, objfile, 1);
}
/* Given a pointer to an initialized bfd (ABFD) and some flag bits,
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* [PATCH 06/13] Use gdb_bfd_sections in symfile.c
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (4 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 05/13] Use gdb_bfd_sections in build_objfile_section_table Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 17:00 ` Simon Marchi
2020-08-28 16:23 ` [PATCH 07/13] Use gdb_bfd_sections in dwarf2/read.c Tom Tromey
` (9 subsequent siblings)
15 siblings, 1 reply; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes some functions in symfile.c to avoid
bfd_map_over_sections, in favor of iteration. Some helper types can
also be removed due to this change.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* symfile.h: (find_lowest_section): Don't declare.
* symfile.c (find_lowest_section): Now static. Change
parameters.
(struct place_section_arg): Remove.
(place_section): Change parameters.
(addr_info_make_relative): Use foreach.
(symfile_dummy_outputs): Remove.
(default_symfile_relocate): Use foreach.
---
gdb/ChangeLog | 11 +++++++++++
gdb/symfile.c | 53 +++++++++++++++++----------------------------------
gdb/symfile.h | 2 --
3 files changed, 28 insertions(+), 38 deletions(-)
diff --git a/gdb/symfile.c b/gdb/symfile.c
index 2c38ce4431a..96d8b3f0b52 100644
--- a/gdb/symfile.c
+++ b/gdb/symfile.c
@@ -187,7 +187,6 @@ increment_reading_symtab (void)
}
/* Remember the lowest-addressed loadable section we've seen.
- This function is called via bfd_map_over_sections.
In case of equal vmas, the section with the largest size becomes the
lowest-addressed loadable section.
@@ -195,11 +194,9 @@ increment_reading_symtab (void)
If the vmas and sizes are equal, the last section is considered the
lowest-addressed loadable section. */
-void
-find_lowest_section (bfd *abfd, asection *sect, void *obj)
+static void
+find_lowest_section (asection *sect, asection **lowest)
{
- asection **lowest = (asection **) obj;
-
if (0 == (bfd_section_flags (sect) & (SEC_ALLOC | SEC_LOAD)))
return;
if (!*lowest)
@@ -335,22 +332,13 @@ init_objfile_sect_indices (struct objfile *objfile)
}
}
-/* The arguments to place_section. */
-
-struct place_section_arg
-{
- section_offsets *offsets;
- CORE_ADDR lowest;
-};
-
/* Find a unique offset to use for loadable section SECT if
the user did not provide an offset. */
static void
-place_section (bfd *abfd, asection *sect, void *obj)
+place_section (bfd *abfd, asection *sect, section_offsets &offsets,
+ CORE_ADDR &lowest)
{
- struct place_section_arg *arg = (struct place_section_arg *) obj;
- section_offsets &offsets = *arg->offsets;
CORE_ADDR start_addr;
int done;
ULONGEST align = ((ULONGEST) 1) << bfd_section_alignment (sect);
@@ -364,7 +352,7 @@ place_section (bfd *abfd, asection *sect, void *obj)
return;
/* Otherwise, let's try to find a place for the section. */
- start_addr = (arg->lowest + align - 1) & -align;
+ start_addr = (lowest + align - 1) & -align;
do {
asection *cur_sec;
@@ -405,7 +393,7 @@ place_section (bfd *abfd, asection *sect, void *obj)
while (!done);
offsets[gdb_bfd_section_index (abfd, sect)] = start_addr;
- arg->lowest = start_addr + bfd_section_size (sect);
+ lowest = start_addr + bfd_section_size (sect);
}
/* Store section_addr_info as prepared (made relative and with SECTINDEX
@@ -500,7 +488,8 @@ addr_info_make_relative (section_addr_info *addrs, bfd *abfd)
/* Find lowest loadable section to be used as starting point for
contiguous sections. */
lower_sect = NULL;
- bfd_map_over_sections (abfd, find_lowest_section, &lower_sect);
+ for (asection *iter : gdb_bfd_sections (abfd))
+ find_lowest_section (iter, &lower_sect);
if (lower_sect == NULL)
{
warning (_("no loadable sections found in added symbol-file %s"),
@@ -645,7 +634,6 @@ default_symfile_offsets (struct objfile *objfile,
small. */
if ((bfd_get_file_flags (objfile->obfd) & (EXEC_P | DYNAMIC)) == 0)
{
- struct place_section_arg arg;
bfd *abfd = objfile->obfd;
asection *cur_sec;
@@ -661,9 +649,10 @@ default_symfile_offsets (struct objfile *objfile,
/* Pick non-overlapping offsets for sections the user did not
place explicitly. */
- arg.offsets = &objfile->section_offsets;
- arg.lowest = 0;
- bfd_map_over_sections (objfile->obfd, place_section, &arg);
+ CORE_ADDR lowest = 0;
+ for (asection *sect : gdb_bfd_sections (objfile->obfd))
+ place_section (objfile->obfd, sect, objfile->section_offsets,
+ lowest);
/* Correctly filling in the section offsets is not quite
enough. Relocatable files have two properties that
@@ -3570,18 +3559,6 @@ simple_overlay_update (struct obj_section *osect)
}
}
-/* Set the output sections and output offsets for section SECTP in
- ABFD. The relocation code in BFD will read these offsets, so we
- need to be sure they're initialized. We map each section to itself,
- with no offset; this means that SECTP->vma will be honored. */
-
-static void
-symfile_dummy_outputs (bfd *abfd, asection *sectp, void *dummy)
-{
- sectp->output_section = sectp;
- sectp->output_offset = 0;
-}
-
/* Default implementation for sym_relocate. */
bfd_byte *
@@ -3599,7 +3576,11 @@ default_symfile_relocate (struct objfile *objfile, asection *sectp,
/* We will handle section offsets properly elsewhere, so relocate as if
all sections begin at 0. */
- bfd_map_over_sections (abfd, symfile_dummy_outputs, NULL);
+ for (asection *sect : gdb_bfd_sections (abfd))
+ {
+ sect->output_section = sect;
+ sect->output_offset = 0;
+ }
return bfd_simple_get_relocated_section_contents (abfd, sectp, buf, NULL);
}
diff --git a/gdb/symfile.h b/gdb/symfile.h
index fe79f79a048..203620bc284 100644
--- a/gdb/symfile.h
+++ b/gdb/symfile.h
@@ -474,8 +474,6 @@ extern bool auto_solib_add;
extern void set_initial_language (void);
-extern void find_lowest_section (bfd *, asection *, void *);
-
extern gdb_bfd_ref_ptr symfile_bfd_open (const char *);
extern int get_section_index (struct objfile *, const char *);
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH 06/13] Use gdb_bfd_sections in symfile.c
2020-08-28 16:23 ` [PATCH 06/13] Use gdb_bfd_sections in symfile.c Tom Tromey
@ 2020-08-28 17:00 ` Simon Marchi
2020-08-28 17:06 ` Simon Marchi
0 siblings, 1 reply; 25+ messages in thread
From: Simon Marchi @ 2020-08-28 17:00 UTC (permalink / raw)
To: Tom Tromey, gdb-patches
On 2020-08-28 12:23 p.m., Tom Tromey wrote:
> @@ -500,7 +488,8 @@ addr_info_make_relative (section_addr_info *addrs, bfd *abfd)
> /* Find lowest loadable section to be used as starting point for
> contiguous sections. */
> lower_sect = NULL;
> - bfd_map_over_sections (abfd, find_lowest_section, &lower_sect);
> + for (asection *iter : gdb_bfd_sections (abfd))
> + find_lowest_section (iter, &lower_sect);
> if (lower_sect == NULL)
> {
> warning (_("no loadable sections found in added symbol-file %s"),
FYI: I've been taught by Pedro [1] that we don't actually need curly braces in this case,
and I've been misleading people about this for a while :(. For future code I write like
this, I won't use the curly braces.
[1] https://sourceware.org/pipermail/gdb-patches/2020-August/171532.html
Simon
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH 06/13] Use gdb_bfd_sections in symfile.c
2020-08-28 17:00 ` Simon Marchi
@ 2020-08-28 17:06 ` Simon Marchi
0 siblings, 0 replies; 25+ messages in thread
From: Simon Marchi @ 2020-08-28 17:06 UTC (permalink / raw)
To: Tom Tromey, gdb-patches
On 2020-08-28 1:00 p.m., Simon Marchi wrote:
> On 2020-08-28 12:23 p.m., Tom Tromey wrote:
>> @@ -500,7 +488,8 @@ addr_info_make_relative (section_addr_info *addrs, bfd *abfd)
>> /* Find lowest loadable section to be used as starting point for
>> contiguous sections. */
>> lower_sect = NULL;
>> - bfd_map_over_sections (abfd, find_lowest_section, &lower_sect);
>> + for (asection *iter : gdb_bfd_sections (abfd))
>> + find_lowest_section (iter, &lower_sect);
>> if (lower_sect == NULL)
>> {
>> warning (_("no loadable sections found in added symbol-file %s"),
>
> FYI: I've been taught by Pedro [1] that we don't actually need curly braces in this case,
> and I've been misleading people about this for a while :(. For future code I write like
> this, I won't use the curly braces.
>
> [1] https://sourceware.org/pipermail/gdb-patches/2020-August/171532.html
>
> Simon
>
Sorry, I replied this at the totally wrong place, in the wrong patch.
Simon
^ permalink raw reply [flat|nested] 25+ messages in thread
* [PATCH 07/13] Use gdb_bfd_sections in dwarf2/read.c
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (5 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 06/13] Use gdb_bfd_sections in symfile.c Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 16:23 ` [PATCH 08/13] Use gdb_bfd_sections in ELF osabi tag sniffing Tom Tromey
` (8 subsequent siblings)
15 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes some functions in dwarf2/read.c to avoid
bfd_map_over_sections, in favor of iteration.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* dwarf2/read.c (locate_dwz_sections): Change parameters.
(dwarf2_get_dwz_file): Use foreach.
(dwarf2_locate_dwo_sections): Change parameters.
(open_and_init_dwo_file): Use foreach.
(dwarf2_locate_common_dwp_sections): Change parameters.
(open_and_init_dwp_file): Use foreach.
---
gdb/ChangeLog | 9 +++++++++
gdb/dwarf2/read.c | 50 +++++++++++++++++++++++------------------------
2 files changed, 33 insertions(+), 26 deletions(-)
diff --git a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c
index 0ac8533263a..1da71c48084 100644
--- a/gdb/dwarf2/read.c
+++ b/gdb/dwarf2/read.c
@@ -2120,10 +2120,8 @@ dwarf2_get_section_info (struct objfile *objfile,
/* A helper function to find the sections for a .dwz file. */
static void
-locate_dwz_sections (bfd *abfd, asection *sectp, void *arg)
+locate_dwz_sections (bfd *abfd, asection *sectp, dwz_file *dwz_file)
{
- struct dwz_file *dwz_file = (struct dwz_file *) arg;
-
/* Note that we only support the standard ELF names, because .dwz
is ELF-only (at the time of writing). */
if (section_is_p (sectp->name, &dwarf2_elf_names.abbrev))
@@ -2246,8 +2244,8 @@ dwarf2_get_dwz_file (dwarf2_per_bfd *per_bfd)
std::unique_ptr<struct dwz_file> result
(new struct dwz_file (std::move (dwz_bfd)));
- bfd_map_over_sections (result->dwz_bfd.get (), locate_dwz_sections,
- result.get ());
+ for (asection *sec : gdb_bfd_sections (result->dwz_bfd))
+ locate_dwz_sections (result->dwz_bfd.get (), sec, result.get ());
gdb_bfd_record_inclusion (per_bfd->obfd, result->dwz_bfd.get ());
per_bfd->dwz_file = std::move (result);
@@ -11923,9 +11921,9 @@ create_dwp_hash_table (dwarf2_per_objfile *per_objfile,
/* Update SECTIONS with the data from SECTP.
- This function is like the other "locate" section routines that are
- passed to bfd_map_over_sections, but in this context the sections to
- read comes from the DWP V1 hash table, not the full ELF section table.
+ This function is like the other "locate" section routines, but in
+ this context the sections to read comes from the DWP V1 hash table,
+ not the full ELF section table.
The result is non-zero for success, or zero if an error was found. */
@@ -12738,9 +12736,9 @@ open_dwo_file (dwarf2_per_objfile *per_objfile,
size of each of the DWO debugging sections we are interested in. */
static void
-dwarf2_locate_dwo_sections (bfd *abfd, asection *sectp, void *dwo_sections_ptr)
+dwarf2_locate_dwo_sections (bfd *abfd, asection *sectp,
+ dwo_sections *dwo_sections)
{
- struct dwo_sections *dwo_sections = (struct dwo_sections *) dwo_sections_ptr;
const struct dwop_section_names *names = &dwop_section_names;
if (section_is_p (sectp->name, &names->abbrev_dwo))
@@ -12827,8 +12825,9 @@ open_and_init_dwo_file (dwarf2_cu *cu, const char *dwo_name,
dwo_file->comp_dir = comp_dir;
dwo_file->dbfd = std::move (dbfd);
- bfd_map_over_sections (dwo_file->dbfd.get (), dwarf2_locate_dwo_sections,
- &dwo_file->sections);
+ for (asection *sec : gdb_bfd_sections (dwo_file->dbfd))
+ dwarf2_locate_dwo_sections (dwo_file->dbfd.get (), sec,
+ &dwo_file->sections);
create_cus_hash_table (per_objfile, cu, *dwo_file, dwo_file->sections.info,
dwo_file->cus);
@@ -12857,9 +12856,8 @@ open_and_init_dwo_file (dwarf2_cu *cu, const char *dwo_name,
static void
dwarf2_locate_common_dwp_sections (bfd *abfd, asection *sectp,
- void *dwp_file_ptr)
+ dwp_file *dwp_file)
{
- struct dwp_file *dwp_file = (struct dwp_file *) dwp_file_ptr;
const struct dwop_section_names *names = &dwop_section_names;
unsigned int elf_section_nr = elf_section_data (sectp)->this_idx;
@@ -13123,9 +13121,9 @@ open_and_init_dwp_file (dwarf2_per_objfile *per_objfile)
OBSTACK_CALLOC (&per_objfile->per_bfd->obstack,
dwp_file->num_sections, asection *);
- bfd_map_over_sections (dwp_file->dbfd.get (),
- dwarf2_locate_common_dwp_sections,
- dwp_file.get ());
+ for (asection *sec : gdb_bfd_sections (dwp_file->dbfd))
+ dwarf2_locate_common_dwp_sections (dwp_file->dbfd.get (), sec,
+ dwp_file.get ());
dwp_file->cus = create_dwp_hash_table (per_objfile, dwp_file.get (), 0);
@@ -13151,15 +13149,15 @@ open_and_init_dwp_file (dwarf2_per_objfile *per_objfile)
else
dwp_file->version = 2;
- if (dwp_file->version == 2)
- bfd_map_over_sections (dwp_file->dbfd.get (),
- dwarf2_locate_v2_dwp_sections,
- dwp_file.get ());
- else if (dwp_file->version == 5)
- bfd_map_over_sections (dwp_file->dbfd.get (),
- dwarf2_locate_v5_dwp_sections,
- dwp_file.get ());
-
+ for (asection *sec : gdb_bfd_sections (dwp_file->dbfd))
+ {
+ if (dwp_file->version == 2)
+ dwarf2_locate_v2_dwp_sections (dwp_file->dbfd.get (), sec,
+ dwp_file.get ());
+ else
+ dwarf2_locate_v5_dwp_sections (dwp_file->dbfd.get (), sec,
+ dwp_file.get ());
+ }
dwp_file->loaded_cus = allocate_dwp_loaded_cutus_table ();
dwp_file->loaded_tus = allocate_dwp_loaded_cutus_table ();
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* [PATCH 08/13] Use gdb_bfd_sections in ELF osabi tag sniffing
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (6 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 07/13] Use gdb_bfd_sections in dwarf2/read.c Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 16:23 ` [PATCH 09/13] Use gdb_bfd_sections in gcore_memory_sections Tom Tromey
` (7 subsequent siblings)
15 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes some ELF osabi tag-sniffing functions to avoid
bfd_map_over_sections, in favor of iteration. I could only readily
test the generic one.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* osabi.h (generic_elf_osabi_sniff_abi_tag_sections): Update.
* osabi.c (generic_elf_osabi_sniff_abi_tag_sections): Change
parameters.
(generic_elf_osabi_sniffer): Use foreach.
* mips-sde-tdep.c (mips_sde_elf_osabi_sniffer): Use foreach.
* arm-tdep.c (arm_elf_osabi_sniffer): Use foreach.
---
gdb/ChangeLog | 9 +++++++++
gdb/arm-tdep.c | 7 ++++---
gdb/mips-sde-tdep.c | 5 ++---
gdb/osabi.c | 10 +++++-----
gdb/osabi.h | 6 +++---
5 files changed, 23 insertions(+), 14 deletions(-)
diff --git a/gdb/arm-tdep.c b/gdb/arm-tdep.c
index 074eedb4800..962ad1a40bc 100644
--- a/gdb/arm-tdep.c
+++ b/gdb/arm-tdep.c
@@ -8754,9 +8754,10 @@ arm_elf_osabi_sniffer (bfd *abfd)
if (elfosabi == ELFOSABI_ARM)
/* GNU tools use this value. Check note sections in this case,
as well. */
- bfd_map_over_sections (abfd,
- generic_elf_osabi_sniff_abi_tag_sections,
- &osabi);
+ {
+ for (asection *sect : gdb_bfd_sections (abfd))
+ generic_elf_osabi_sniff_abi_tag_sections (abfd, sect, &osabi);
+ }
/* Anything else will be handled by the generic ELF sniffer. */
return osabi;
diff --git a/gdb/mips-sde-tdep.c b/gdb/mips-sde-tdep.c
index 685c3c7263c..1d282c219e5 100644
--- a/gdb/mips-sde-tdep.c
+++ b/gdb/mips-sde-tdep.c
@@ -226,9 +226,8 @@ mips_sde_elf_osabi_sniffer (bfd *abfd)
/* If the generic sniffer gets a hit, return and let other sniffers
get a crack at it. */
- bfd_map_over_sections (abfd,
- generic_elf_osabi_sniff_abi_tag_sections,
- &osabi);
+ for (asection *sect : gdb_bfd_sections (abfd))
+ generic_elf_osabi_sniff_abi_tag_sections (abfd, sect, &osabi);
if (osabi != GDB_OSABI_UNKNOWN)
return GDB_OSABI_UNKNOWN;
diff --git a/gdb/osabi.c b/gdb/osabi.c
index 627b9d98151..e8a813b62f2 100644
--- a/gdb/osabi.c
+++ b/gdb/osabi.c
@@ -23,6 +23,7 @@
#include "arch-utils.h"
#include "gdbcmd.h"
#include "command.h"
+#include "gdb_bfd.h"
#include "elf-bfd.h"
@@ -439,9 +440,9 @@ check_note (bfd *abfd, asection *sect, char *note, unsigned int *sectsize,
/* Generic sniffer for ELF flavoured files. */
void
-generic_elf_osabi_sniff_abi_tag_sections (bfd *abfd, asection *sect, void *obj)
+generic_elf_osabi_sniff_abi_tag_sections (bfd *abfd, asection *sect,
+ enum gdb_osabi *osabi)
{
- enum gdb_osabi *osabi = (enum gdb_osabi *) obj;
const char *name;
unsigned int sectsize;
char *note;
@@ -561,9 +562,8 @@ generic_elf_osabi_sniffer (bfd *abfd)
/* And likewise ELFOSABI_HPUX. For some reason the default
value for the EI_OSABI field is ELFOSABI_HPUX for all PA-RISC
targets (with the exception of GNU/Linux). */
- bfd_map_over_sections (abfd,
- generic_elf_osabi_sniff_abi_tag_sections,
- &osabi);
+ for (asection *sect : gdb_bfd_sections (abfd))
+ generic_elf_osabi_sniff_abi_tag_sections (abfd, sect, &osabi);
break;
case ELFOSABI_FREEBSD:
diff --git a/gdb/osabi.h b/gdb/osabi.h
index a7e6a10d019..9b7ab1818c0 100644
--- a/gdb/osabi.h
+++ b/gdb/osabi.h
@@ -86,8 +86,8 @@ const char *gdbarch_osabi_name (enum gdb_osabi);
const char *osabi_triplet_regexp (enum gdb_osabi osabi);
/* Helper routine for ELF file sniffers. This looks at ABI tag note
- sections to determine the OS ABI from the note. It should be called
- via bfd_map_over_sections. */
-void generic_elf_osabi_sniff_abi_tag_sections (bfd *, asection *, void *);
+ sections to determine the OS ABI from the note. */
+void generic_elf_osabi_sniff_abi_tag_sections (bfd *, asection *,
+ enum gdb_osabi *);
#endif /* OSABI_H */
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* [PATCH 09/13] Use gdb_bfd_sections in gcore_memory_sections
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (7 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 08/13] Use gdb_bfd_sections in ELF osabi tag sniffing Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 16:23 ` [PATCH 10/13] Use gdb_bfd_sections in restore_command Tom Tromey
` (6 subsequent siblings)
15 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes gcore_memory_sections to avoid bfd_map_over_sections, in
favor of iteration.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* gcore.c (make_output_phdrs): Remove 'ignored' parameter.
(gcore_copy_callback): Likewise.
(gcore_memory_sections): Use foreach.
---
gdb/ChangeLog | 6 ++++++
gdb/gcore.c | 10 ++++++----
2 files changed, 12 insertions(+), 4 deletions(-)
diff --git a/gdb/gcore.c b/gdb/gcore.c
index d0e36b1a708..46ca2da5003 100644
--- a/gdb/gcore.c
+++ b/gdb/gcore.c
@@ -364,7 +364,7 @@ derive_heap_segment (bfd *abfd, bfd_vma *bottom, bfd_vma *top)
}
static void
-make_output_phdrs (bfd *obfd, asection *osec, void *ignored)
+make_output_phdrs (bfd *obfd, asection *osec)
{
int p_flags = 0;
int p_type = 0;
@@ -531,7 +531,7 @@ objfile_find_memory_regions (struct target_ops *self,
}
static void
-gcore_copy_callback (bfd *obfd, asection *osec, void *ignored)
+gcore_copy_callback (bfd *obfd, asection *osec)
{
bfd_size_type size, total_size = bfd_section_size (osec);
file_ptr offset = 0;
@@ -587,10 +587,12 @@ gcore_memory_sections (bfd *obfd)
}
/* Record phdrs for section-to-segment mapping. */
- bfd_map_over_sections (obfd, make_output_phdrs, NULL);
+ for (asection *sect : gdb_bfd_sections (obfd))
+ make_output_phdrs (obfd, sect);
/* Copy memory region contents. */
- bfd_map_over_sections (obfd, gcore_copy_callback, NULL);
+ for (asection *sect : gdb_bfd_sections (obfd))
+ gcore_copy_callback (obfd, sect);
return 1;
}
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* [PATCH 10/13] Use gdb_bfd_sections in restore_command
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (8 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 09/13] Use gdb_bfd_sections in gcore_memory_sections Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 17:12 ` Simon Marchi
2020-08-28 16:23 ` [PATCH 11/13] Use gdb_bfd_sections in elf_symfile_read Tom Tromey
` (5 subsequent siblings)
15 siblings, 1 reply; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes restore_command to avoid bfd_map_over_sections, in favor
of iteration. A helper data structure can also be removed by this
patch.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* cli/cli-dump.c (struct callback_data): Remove.
(restore_one_section): Rename from restore_section_callback.
Change parameters.
(restore_binary_file): Change parameters.
(restore_command): Use foreach.
---
gdb/ChangeLog | 8 ++++
gdb/cli/cli-dump.c | 94 ++++++++++++++++++++++------------------------
2 files changed, 53 insertions(+), 49 deletions(-)
diff --git a/gdb/cli/cli-dump.c b/gdb/cli/cli-dump.c
index 567ef2eeded..849964ece6b 100644
--- a/gdb/cli/cli-dump.c
+++ b/gdb/cli/cli-dump.c
@@ -371,22 +371,14 @@ add_dump_command (const char *name,
c->doc = concat ("Append ", c->doc + 6, (char *)NULL);
}
-/* Opaque data for restore_section_callback. */
-struct callback_data {
- CORE_ADDR load_offset;
- CORE_ADDR load_start;
- CORE_ADDR load_end;
-};
-
-/* Function: restore_section_callback.
-
- Callback function for bfd_map_over_sections.
- Selectively loads the sections into memory. */
+/* Selectively loads the sections into memory. */
static void
-restore_section_callback (bfd *ibfd, asection *isec, void *args)
+restore_one_section (bfd *ibfd, asection *isec,
+ CORE_ADDR load_offset,
+ CORE_ADDR load_start,
+ CORE_ADDR load_end)
{
- struct callback_data *data = (struct callback_data *) args;
bfd_vma sec_start = bfd_section_vma (isec);
bfd_size_type size = bfd_section_size (isec);
bfd_vma sec_end = sec_start + size;
@@ -399,8 +391,8 @@ restore_section_callback (bfd *ibfd, asection *isec, void *args)
return;
/* Does the section overlap with the desired restore range? */
- if (sec_end <= data->load_start
- || (data->load_end > 0 && sec_start >= data->load_end))
+ if (sec_end <= load_start
+ || (load_end > 0 && sec_start >= load_end))
{
/* No, no useable data in this section. */
printf_filtered (_("skipping section %s...\n"),
@@ -411,12 +403,12 @@ restore_section_callback (bfd *ibfd, asection *isec, void *args)
/* Compare section address range with user-requested
address range (if any). Compute where the actual
transfer should start and end. */
- if (sec_start < data->load_start)
- sec_offset = data->load_start - sec_start;
+ if (sec_start < load_start)
+ sec_offset = load_start - sec_start;
/* Size of a partial transfer. */
sec_load_count -= sec_offset;
- if (data->load_end > 0 && sec_end > data->load_end)
- sec_load_count -= sec_end - data->load_end;
+ if (load_end > 0 && sec_end > load_end)
+ sec_load_count -= sec_end - load_end;
/* Get the data. */
gdb::byte_vector buf (size);
@@ -429,26 +421,28 @@ restore_section_callback (bfd *ibfd, asection *isec, void *args)
(unsigned long) sec_start,
(unsigned long) sec_end);
- if (data->load_offset != 0 || data->load_start != 0 || data->load_end != 0)
+ if (load_offset != 0 || load_start != 0 || load_end != 0)
printf_filtered (" into memory (%s to %s)\n",
paddress (target_gdbarch (),
(unsigned long) sec_start
- + sec_offset + data->load_offset),
+ + sec_offset + load_offset),
paddress (target_gdbarch (),
(unsigned long) sec_start + sec_offset
- + data->load_offset + sec_load_count));
+ + load_offset + sec_load_count));
else
puts_filtered ("\n");
/* Write the data. */
- ret = target_write_memory (sec_start + sec_offset + data->load_offset,
+ ret = target_write_memory (sec_start + sec_offset + load_offset,
&buf[sec_offset], sec_load_count);
if (ret != 0)
warning (_("restore: memory write failed (%s)."), safe_strerror (ret));
}
static void
-restore_binary_file (const char *filename, struct callback_data *data)
+restore_binary_file (const char *filename, CORE_ADDR load_offset,
+ CORE_ADDR load_start, CORE_ADDR load_end)
+
{
gdb_file_up file = gdb_fopen_cloexec (filename, FOPEN_RB);
long len;
@@ -466,25 +460,25 @@ restore_binary_file (const char *filename, struct callback_data *data)
else
perror_with_name (filename);
- if (len <= data->load_start)
+ if (len <= load_start)
error (_("Start address is greater than length of binary file %s."),
filename);
/* Chop off "len" if it exceeds the requested load_end addr. */
- if (data->load_end != 0 && data->load_end < len)
- len = data->load_end;
+ if (load_end != 0 && load_end < len)
+ len = load_end;
/* Chop off "len" if the requested load_start addr skips some bytes. */
- if (data->load_start > 0)
- len -= data->load_start;
+ if (load_start > 0)
+ len -= load_start;
printf_filtered
("Restoring binary file %s into memory (0x%lx to 0x%lx)\n",
filename,
- (unsigned long) (data->load_start + data->load_offset),
- (unsigned long) (data->load_start + data->load_offset + len));
+ (unsigned long) (load_start + load_offset),
+ (unsigned long) (load_start + load_offset + len));
/* Now set the file pos to the requested load start pos. */
- if (fseek (file.get (), data->load_start, SEEK_SET) != 0)
+ if (fseek (file.get (), load_start, SEEK_SET) != 0)
perror_with_name (filename);
/* Now allocate a buffer and read the file contents. */
@@ -493,8 +487,7 @@ restore_binary_file (const char *filename, struct callback_data *data)
perror_with_name (filename);
/* Now write the buffer into target memory. */
- len = target_write_memory (data->load_start + data->load_offset,
- buf.data (), len);
+ len = target_write_memory (load_start + load_offset, buf.data (), len);
if (len != 0)
warning (_("restore: memory write failed (%s)."), safe_strerror (len));
}
@@ -502,15 +495,14 @@ restore_binary_file (const char *filename, struct callback_data *data)
static void
restore_command (const char *args, int from_tty)
{
- struct callback_data data;
int binary_flag = 0;
if (!target_has_execution)
noprocess ();
- data.load_offset = 0;
- data.load_start = 0;
- data.load_end = 0;
+ CORE_ADDR load_offset = 0;
+ CORE_ADDR load_start = 0;
+ CORE_ADDR load_end = 0;
/* Parse the input arguments. First is filename (required). */
gdb::unique_xmalloc_ptr<char> filename = scan_filename (&args, NULL);
@@ -527,19 +519,20 @@ restore_command (const char *args, int from_tty)
}
/* Parse offset (optional). */
if (args != NULL && *args != '\0')
- data.load_offset = binary_flag ?
- parse_and_eval_address (scan_expression (&args, NULL).get ()) :
- parse_and_eval_long (scan_expression (&args, NULL).get ());
+ load_offset
+ = (binary_flag
+ ? parse_and_eval_address (scan_expression (&args, NULL).get ())
+ : parse_and_eval_long (scan_expression (&args, NULL).get ()));
if (args != NULL && *args != '\0')
{
/* Parse start address (optional). */
- data.load_start =
+ load_start =
parse_and_eval_long (scan_expression (&args, NULL).get ());
if (args != NULL && *args != '\0')
{
/* Parse end address (optional). */
- data.load_end = parse_and_eval_long (args);
- if (data.load_end <= data.load_start)
+ load_end = parse_and_eval_long (args);
+ if (load_end <= load_start)
error (_("Start must be less than end."));
}
}
@@ -547,13 +540,14 @@ restore_command (const char *args, int from_tty)
if (info_verbose)
printf_filtered ("Restore file %s offset 0x%lx start 0x%lx end 0x%lx\n",
- filename.get (), (unsigned long) data.load_offset,
- (unsigned long) data.load_start,
- (unsigned long) data.load_end);
+ filename.get (), (unsigned long) load_offset,
+ (unsigned long) load_start,
+ (unsigned long) load_end);
if (binary_flag)
{
- restore_binary_file (filename.get (), &data);
+ restore_binary_file (filename.get (), load_offset, load_start,
+ load_end);
}
else
{
@@ -561,7 +555,9 @@ restore_command (const char *args, int from_tty)
gdb_bfd_ref_ptr ibfd (bfd_openr_or_error (filename.get (), NULL));
/* Process the sections. */
- bfd_map_over_sections (ibfd.get (), restore_section_callback, &data);
+ for (asection *sect : gdb_bfd_sections (ibfd))
+ restore_one_section (ibfd.get (), sect, load_offset, load_start,
+ load_end);
}
}
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH 10/13] Use gdb_bfd_sections in restore_command
2020-08-28 16:23 ` [PATCH 10/13] Use gdb_bfd_sections in restore_command Tom Tromey
@ 2020-08-28 17:12 ` Simon Marchi
2020-08-29 15:47 ` Tom Tromey
0 siblings, 1 reply; 25+ messages in thread
From: Simon Marchi @ 2020-08-28 17:12 UTC (permalink / raw)
To: Tom Tromey, gdb-patches
On 2020-08-28 12:23 p.m., Tom Tromey wrote:
> This changes restore_command to avoid bfd_map_over_sections, in favor
> of iteration. A helper data structure can also be removed by this
> patch.
>
> gdb/ChangeLog
> 2020-08-28 Tom Tromey <tom@tromey.com>
>
> * cli/cli-dump.c (struct callback_data): Remove.
> (restore_one_section): Rename from restore_section_callback.
> Change parameters.
> (restore_binary_file): Change parameters.
> (restore_command): Use foreach.
Warnings I got while applying:
Applying: Use gdb_bfd_sections in restore_command
.git/rebase-apply/patch:44: trailing whitespace.
if (sec_end <= load_start
.git/rebase-apply/patch:76: trailing whitespace.
+ sec_offset + load_offset),
.git/rebase-apply/patch:86: trailing whitespace.
ret = target_write_memory (sec_start + sec_offset + load_offset,
.git/rebase-apply/patch:178: trailing whitespace.
load_start =
.git/rebase-apply/patch:198: trailing whitespace.
(unsigned long) load_start,
warning: 5 lines add whitespace errors.
Simon
^ permalink raw reply [flat|nested] 25+ messages in thread
* [PATCH 11/13] Use gdb_bfd_sections in elf_symfile_read
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (9 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 10/13] Use gdb_bfd_sections in restore_command Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 16:23 ` [PATCH 12/13] Use gdb_bfd_sections in build_section_table Tom Tromey
` (4 subsequent siblings)
15 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes elf_symfile_read to avoid bfd_map_over_sections, in favor
of iteration.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* elfread.c (elf_locate_sections): Change parameters.
(elf_symfile_read): Use foreach.
---
gdb/ChangeLog | 5 +++++
gdb/elfread.c | 10 +++++-----
2 files changed, 10 insertions(+), 5 deletions(-)
diff --git a/gdb/elfread.c b/gdb/elfread.c
index 75bdd75250b..5dad3ddd940 100644
--- a/gdb/elfread.c
+++ b/gdb/elfread.c
@@ -176,11 +176,8 @@ elf_symfile_segments (bfd *abfd)
-kingdon). */
static void
-elf_locate_sections (bfd *ignore_abfd, asection *sectp, void *eip)
+elf_locate_sections (asection *sectp, struct elfinfo *ei)
{
- struct elfinfo *ei;
-
- ei = (struct elfinfo *) eip;
if (strcmp (sectp->name, ".stab") == 0)
{
ei->stabsect = sectp;
@@ -1211,7 +1208,10 @@ elf_symfile_read (struct objfile *objfile, symfile_add_flags symfile_flags)
memset ((char *) &ei, 0, sizeof (ei));
if (!(objfile->flags & OBJF_READNEVER))
- bfd_map_over_sections (abfd, elf_locate_sections, (void *) & ei);
+ {
+ for (asection *sect : gdb_bfd_sections (abfd))
+ elf_locate_sections (sect, &ei);
+ }
elf_read_minimal_symbols (objfile, symfile_flags, &ei);
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* [PATCH 12/13] Use gdb_bfd_sections in build_section_table
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (10 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 11/13] Use gdb_bfd_sections in elf_symfile_read Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 16:23 ` [PATCH 13/13] Use gdb_bfd_sections in generic_load Tom Tromey
` (3 subsequent siblings)
15 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes build_section_table to avoid bfd_map_over_sections, in
favor of iteration. In this situation it seemed simple to just remove
the helper function entirely.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* exec.c (add_to_section_table): Remove.
(build_section_table): Use foreach.
---
gdb/ChangeLog | 5 +++++
gdb/exec.c | 50 ++++++++++++++++++++------------------------------
2 files changed, 25 insertions(+), 30 deletions(-)
diff --git a/gdb/exec.c b/gdb/exec.c
index e50f38899d6..14cc6af7de0 100644
--- a/gdb/exec.c
+++ b/gdb/exec.c
@@ -587,35 +587,6 @@ file_command (const char *arg, int from_tty)
}
\f
-/* Locate all mappable sections of a BFD file.
- table_pp_char is a char * to get it through bfd_map_over_sections;
- we cast it back to its proper type. */
-
-static void
-add_to_section_table (bfd *abfd, struct bfd_section *asect,
- void *table_pp_char)
-{
- struct target_section **table_pp = (struct target_section **) table_pp_char;
- flagword aflag;
-
- gdb_assert (abfd == asect->owner);
-
- /* Check the section flags, but do not discard zero-length sections, since
- some symbols may still be attached to this section. For instance, we
- encountered on sparc-solaris 2.10 a shared library with an empty .bss
- section to which a symbol named "_end" was attached. The address
- of this symbol still needs to be relocated. */
- aflag = bfd_section_flags (asect);
- if (!(aflag & SEC_ALLOC))
- return;
-
- (*table_pp)->owner = NULL;
- (*table_pp)->the_bfd_section = asect;
- (*table_pp)->addr = bfd_section_vma (asect);
- (*table_pp)->endaddr = (*table_pp)->addr + bfd_section_size (asect);
- (*table_pp)++;
-}
-
/* See exec.h. */
void
@@ -665,7 +636,26 @@ build_section_table (struct bfd *some_bfd, struct target_section **start,
xfree (*start);
*start = XNEWVEC (struct target_section, count);
*end = *start;
- bfd_map_over_sections (some_bfd, add_to_section_table, (char *) end);
+ for (asection *asect : gdb_bfd_sections (some_bfd))
+ {
+ flagword aflag;
+
+ /* Check the section flags, but do not discard zero-length
+ sections, since some symbols may still be attached to this
+ section. For instance, we encountered on sparc-solaris 2.10
+ a shared library with an empty .bss section to which a symbol
+ named "_end" was attached. The address of this symbol still
+ needs to be relocated. */
+ aflag = bfd_section_flags (asect);
+ if (!(aflag & SEC_ALLOC))
+ continue;
+
+ (*end)->owner = NULL;
+ (*end)->the_bfd_section = asect;
+ (*end)->addr = bfd_section_vma (asect);
+ (*end)->endaddr = (*end)->addr + bfd_section_size (asect);
+ (*end)++;
+ }
gdb_assert (*end <= *start + count);
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* [PATCH 13/13] Use gdb_bfd_sections in generic_load
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (11 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 12/13] Use gdb_bfd_sections in build_section_table Tom Tromey
@ 2020-08-28 16:23 ` Tom Tromey
2020-08-28 17:14 ` [PATCH 00/13] Use gdb_bfd_sections in more places Simon Marchi
` (2 subsequent siblings)
15 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-08-28 16:23 UTC (permalink / raw)
To: gdb-patches; +Cc: Tom Tromey
This changes generic_load to avoid bfd_map_over_sections, in favor of
iteration.
gdb/ChangeLog
2020-08-28 Tom Tromey <tom@tromey.com>
* symfile.c (add_section_size_callback): Remove.
(load_one_section): Rename from load_section_callback. Change
parameters.
(generic_load): Use foreach.
---
gdb/ChangeLog | 7 +++++++
gdb/symfile.c | 23 +++++++----------------
2 files changed, 14 insertions(+), 16 deletions(-)
diff --git a/gdb/symfile.c b/gdb/symfile.c
index 96d8b3f0b52..1c22d8582c0 100644
--- a/gdb/symfile.c
+++ b/gdb/symfile.c
@@ -1842,16 +1842,6 @@ load_command (const char *arg, int from_tty)
static int validate_download = 0;
-/* Callback service function for generic_load (bfd_map_over_sections). */
-
-static void
-add_section_size_callback (bfd *abfd, asection *asec, void *data)
-{
- bfd_size_type *sum = (bfd_size_type *) data;
-
- *sum += bfd_section_size (asec);
-}
-
/* Opaque data for load_progress. */
struct load_progress_data
{
@@ -1966,12 +1956,12 @@ load_progress (ULONGEST bytes, void *untyped_arg)
totals->total_size);
}
-/* Callback service function for generic_load (bfd_map_over_sections). */
+/* Service function for generic_load. */
static void
-load_section_callback (bfd *abfd, asection *asec, void *data)
+load_one_section (bfd *abfd, asection *asec,
+ struct load_section_data *args)
{
- struct load_section_data *args = (struct load_section_data *) data;
bfd_size_type size = bfd_section_size (asec);
const char *sect_name = bfd_section_name (asec);
@@ -2040,10 +2030,11 @@ generic_load (const char *args, int from_tty)
bfd_errmsg (bfd_get_error ()));
}
- bfd_map_over_sections (loadfile_bfd.get (), add_section_size_callback,
- (void *) &total_progress.total_size);
+ for (asection *asec : gdb_bfd_sections (loadfile_bfd))
+ total_progress.total_size += bfd_section_size (asec);
- bfd_map_over_sections (loadfile_bfd.get (), load_section_callback, &cbdata);
+ for (asection *asec : gdb_bfd_sections (loadfile_bfd))
+ load_one_section (loadfile_bfd.get (), asec, &cbdata);
using namespace std::chrono;
--
2.17.2
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH 00/13] Use gdb_bfd_sections in more places
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (12 preceding siblings ...)
2020-08-28 16:23 ` [PATCH 13/13] Use gdb_bfd_sections in generic_load Tom Tromey
@ 2020-08-28 17:14 ` Simon Marchi
2020-08-31 18:19 ` John Baldwin
2020-09-19 18:07 ` Tom Tromey
15 siblings, 0 replies; 25+ messages in thread
From: Simon Marchi @ 2020-08-28 17:14 UTC (permalink / raw)
To: Tom Tromey, gdb-patches
On 2020-08-28 12:23 p.m., Tom Tromey wrote:
> As promised, here are some patches to use gdb_bfd_sections in more
> places.
>
> The main motivation for this patch is that ordinary iteration is
> easier to understand than callback functions. Often, it is also
> shorter -- some patches in this series remove helper functions or
> helper data structures. Finally, iteration is more type-safe (at
> least without additional wrappers to pass function-views through
> bfd_map_over_sections).
>
> I did not convert every call to bfd_map_over_sections. Some of the
> remaining calls are not easy to test. For the calls in compile/, I
> have another series that makes many changes in that code, and so I
> didn't want to introduce clashes at this time. (I don't recall
> offhand if that series fixes this, but I can revisit when it lands.)
>
> As usual with these cleanups, I don't plan to commit this until the
> release branch is made.
>
> Let me know what you think.
>
> Tom
>
>
Woohoo, thanks!
Apart from the small nits I sent, this all LGTM.
Simon
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH 00/13] Use gdb_bfd_sections in more places
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (13 preceding siblings ...)
2020-08-28 17:14 ` [PATCH 00/13] Use gdb_bfd_sections in more places Simon Marchi
@ 2020-08-31 18:19 ` John Baldwin
2020-09-19 18:07 ` Tom Tromey
15 siblings, 0 replies; 25+ messages in thread
From: John Baldwin @ 2020-08-31 18:19 UTC (permalink / raw)
To: Tom Tromey, gdb-patches
On 8/28/20 9:23 AM, Tom Tromey wrote:
> As promised, here are some patches to use gdb_bfd_sections in more
> places.
>
> The main motivation for this patch is that ordinary iteration is
> easier to understand than callback functions. Often, it is also
> shorter -- some patches in this series remove helper functions or
> helper data structures. Finally, iteration is more type-safe (at
> least without additional wrappers to pass function-views through
> bfd_map_over_sections).
>
> I did not convert every call to bfd_map_over_sections. Some of the
> remaining calls are not easy to test. For the calls in compile/, I
> have another series that makes many changes in that code, and so I
> didn't want to introduce clashes at this time. (I don't recall
> offhand if that series fixes this, but I can revisit when it lands.)
>
> As usual with these cleanups, I don't plan to commit this until the
> release branch is made.
>
> Let me know what you think.
I didn't look at all the patches in detail, but in general I agree that
this is more readable and really like this change. Ranged-for also
makes this very readable relative to what you would have to do in
C++03.
I see in some cases you also inlined the callback entirely when it was
short which I also think is a good idea.
--
John Baldwin
^ permalink raw reply [flat|nested] 25+ messages in thread* Re: [PATCH 00/13] Use gdb_bfd_sections in more places
2020-08-28 16:23 [PATCH 00/13] Use gdb_bfd_sections in more places Tom Tromey
` (14 preceding siblings ...)
2020-08-31 18:19 ` John Baldwin
@ 2020-09-19 18:07 ` Tom Tromey
15 siblings, 0 replies; 25+ messages in thread
From: Tom Tromey @ 2020-09-19 18:07 UTC (permalink / raw)
To: Tom Tromey; +Cc: gdb-patches
>>>>> "Tom" == Tom Tromey <tom@tromey.com> writes:
Tom> As promised, here are some patches to use gdb_bfd_sections in more
Tom> places.
I've fixed the review comments, which were very minor, and now I am
checking this in.
Tom
^ permalink raw reply [flat|nested] 25+ messages in thread