From: Tom de Vries <tdevries@suse.de>
To: Simon Marchi <simark@simark.ca>, Pedro Alves <pedro@palves.net>,
gdb-patches@sourceware.org
Subject: Re: [PATCH][gdb] Fix warning in foreach_arch selftests
Date: Thu, 2 Jun 2022 20:35:03 +0200 [thread overview]
Message-ID: <79403061-0d96-654b-03dd-032ede7fbdb3@suse.de> (raw)
In-Reply-To: <ac81790c-d64f-16b2-8787-e108b2af75ed@simark.ca>
[-- Attachment #1: Type: text/plain, Size: 1646 bytes --]
On 6/2/22 19:44, Simon Marchi wrote:
> On 6/2/22 11:44, Tom de Vries via Gdb-patches wrote:
>> On 6/1/22 21:08, Simon Marchi wrote:
>>> Hi Tom,
>>>
>>> This introduces some failures for me:
>>>
>>> Running selftest print_one_insn::A6.^M
>>> Self test failed: Cannot access memory at address 0x0^M
>>
>> Hi Simon,
>>
>> sorry for not catching this, I forgot to test with --enable-targets=all, I'll try to fix the default in my build scripts.
>>
>> I've done a new version of the existing patch (so it applies once the old one is reverted).
>>
>> Mainly because I realized that there's an easier way to handle the osabi none setting:
>> ...
>> @@ -68,6 +68,7 @@ foreach_arch_test_generator (const std::string &name,
>> {
>> struct gdbarch_info info;
>> info.bfd_arch_info = bfd_scan_arch (arch);
>> + info.osabi = GDB_OSABI_NONE;
>> struct gdbarch *gdbarch = gdbarch_find_by_info (info);
>> SELF_CHECK (gdbarch != NULL);
>> function (gdbarch);
>> ...
>>
>> Currently regression testing, but at least fixes the ARC regression.
>
> It fixes the gdb.gdb/unittest.exp failure, but introduces:
>
> FAIL: gdb.arch/amd64-osabi.exp: arch=i386:x86-64: print sizeof (void*)
> FAIL: gdb.arch/amd64-osabi.exp: arch=i386:x86-64: print sizeof (long double)
> FAIL: gdb.arch/amd64-osabi.exp: arch=i386:x64-32: print sizeof (long double)
Ack, I also encountered that during testing, that was a thinko in the
refactoring of gdbarch_init_osabi, now fixed and fully tested.
Thanks,
- Tom
[-- Attachment #2: 0001-gdb-Fix-warning-in-foreach_arch-selftests.patch --]
[-- Type: text/x-patch, Size: 6228 bytes --]
[gdb] Fix warning in foreach_arch selftests
When running the selftests, I run into:
...
$ gdb -q -batch -ex "maint selftest"
...
Running selftest execute_cfa_program::aarch64:ilp32.
warning: A handler for the OS ABI "GNU/Linux" is not built into this
configuration of GDB. Attempting to continue with the default aarch64:ilp32
settings.
...
and likewise for execute_cfa_program::i8086 and
execute_cfa_program::ia64-elf32.
The warning can easily be reproduced outside the selftests by doing:
...
$ gdb -q -batch -ex "set arch aarch64:ilp32"
...
and can be prevented by first doing "set osabi none".
Fix the warning by setting osabi to none while doing selftests that iterate
over all architectures.
This causes a regression in the print_one_insn selftests for the ARC
archictecture.
The problem is pre-existing, and can be demonstrated (already without this
patch) using:
...
$ gdb -q -batch -ex "set osabi none" -ex "maint selftest print_one_insn::A6"
Running selftest print_one_insn::A6.
Self test failed: Cannot access memory at address 0x0
Ran 1 unit tests, 1 failed
$
...
For ARC, we use the generic case in print_one_insn_test, containing this code:
...
int kind = gdbarch_breakpoint_kind_from_pc (gdbarch, &pc);
...
insn = gdbarch_sw_breakpoint_from_kind (gdbarch, kind, &bplen);
...
The problem is that with osabi linux we trigger:
...
static int
arc_linux_breakpoint_kind_from_pc (struct gdbarch *gdbarch, CORE_ADDR *pcptr)
{
return trap_size;
}
...
but with osabi none:
...
arc_breakpoint_kind_from_pc (struct gdbarch *gdbarch, CORE_ADDR *pcptr)
{
size_t length_with_limm = gdb_insn_length (gdbarch, *pcptr);
...
which needs access to memory, and will consequently fail.
Fix this in print_one_insn_test, in the default case, by iterating over
supported osabi's to makes sure we trigger arc_linux_breakpoint_kind_from_pc
which will give us a usable instruction to disassemble.
Tested on x86_64-linux.
---
gdb/disasm-selftests.c | 47 ++++++++++++++++++++++++++++++++++++++++++++---
gdb/osabi.c | 35 +++++++++++++++++++++++++++++------
gdb/osabi.h | 3 +++
gdb/selftest-arch.c | 1 +
4 files changed, 77 insertions(+), 9 deletions(-)
diff --git a/gdb/disasm-selftests.c b/gdb/disasm-selftests.c
index 928d26f7018..ccb60a0bc87 100644
--- a/gdb/disasm-selftests.c
+++ b/gdb/disasm-selftests.c
@@ -101,12 +101,53 @@ print_one_insn_test (struct gdbarch *gdbarch)
{
/* Test disassemble breakpoint instruction. */
CORE_ADDR pc = 0;
- int kind = gdbarch_breakpoint_kind_from_pc (gdbarch, &pc);
+ int kind;
int bplen;
- insn = gdbarch_sw_breakpoint_from_kind (gdbarch, kind, &bplen);
- len = bplen;
+ const char *arch = gdbarch_bfd_arch_info (gdbarch)->printable_name;
+
+ enum gdb_osabi it;
+ bool found = false;
+ for (it = GDB_OSABI_UNKNOWN; it != GDB_OSABI_INVALID;
+ it = static_cast<enum gdb_osabi>(static_cast<int>(it) + 1))
+ {
+ if (it == GDB_OSABI_UNKNOWN)
+ continue;
+
+ struct gdbarch_info info;
+ info.bfd_arch_info = bfd_scan_arch (arch);
+ info.osabi = it;
+
+ if (it != GDB_OSABI_NONE)
+ {
+ if (!has_gdb_osabi_handler (info))
+ /* Unsupported. Skip to prevent warnings like:
+ A handler for the OS ABI <x> is not built into this
+ configuration of GDB. Attempting to continue with the
+ default <y> settings. */
+ continue;
+ }
+
+ gdbarch = gdbarch_find_by_info (info);
+ SELF_CHECK (gdbarch != NULL);
+
+ try
+ {
+ kind = gdbarch_breakpoint_kind_from_pc (gdbarch, &pc);
+ insn = gdbarch_sw_breakpoint_from_kind (gdbarch, kind, &bplen);
+ }
+ catch (...)
+ {
+ continue;
+ }
+ found = true;
+ break;
+ }
+
+ /* Assert that we have found an instruction to disassemble. */
+ SELF_CHECK (found);
+ len = bplen;
break;
}
}
diff --git a/gdb/osabi.c b/gdb/osabi.c
index bbd7635532f..f1b7f227b87 100644
--- a/gdb/osabi.c
+++ b/gdb/osabi.c
@@ -332,9 +332,10 @@ can_run_code_for (const struct bfd_arch_info *a, const struct bfd_arch_info *b)
return (a == b || a->compatible (a, b) == a);
}
+/* Return OS ABI handler for INFO. */
-void
-gdbarch_init_osabi (struct gdbarch_info info, struct gdbarch *gdbarch)
+static struct gdb_osabi_handler *
+gdbarch_osabi_handler (struct gdbarch_info info)
{
struct gdb_osabi_handler *handler;
@@ -367,10 +368,32 @@ gdbarch_init_osabi (struct gdbarch_info info, struct gdbarch *gdbarch)
ISA"). BFD doesn't normally consider 32-bit and 64-bit
"compatible" so it doesn't succeed. */
if (can_run_code_for (info.bfd_arch_info, handler->arch_info))
- {
- (*handler->init_osabi) (info, gdbarch);
- return;
- }
+ return handler;
+ }
+
+ return nullptr;
+}
+
+/* See osabi.h. */
+
+bool
+has_gdb_osabi_handler (struct gdbarch_info info)
+{
+ return gdbarch_osabi_handler (info) != nullptr;
+}
+
+void
+gdbarch_init_osabi (struct gdbarch_info info, struct gdbarch *gdbarch)
+{
+ struct gdb_osabi_handler *handler;
+
+ gdb_assert (info.osabi != GDB_OSABI_UNKNOWN);
+ handler = gdbarch_osabi_handler (info);
+
+ if (handler != nullptr)
+ {
+ (*handler->init_osabi) (info, gdbarch);
+ return;
}
if (info.osabi == GDB_OSABI_NONE)
diff --git a/gdb/osabi.h b/gdb/osabi.h
index be016732cbc..478a418aac2 100644
--- a/gdb/osabi.h
+++ b/gdb/osabi.h
@@ -74,6 +74,9 @@ enum gdb_osabi gdbarch_lookup_osabi (bfd *);
string. */
enum gdb_osabi osabi_from_tdesc_string (const char *text);
+/* Return true if there's an OS ABI handler for INFO. */
+bool has_gdb_osabi_handler (struct gdbarch_info info);
+
/* Initialize the gdbarch for the specified OS ABI variant. */
void gdbarch_init_osabi (struct gdbarch_info, struct gdbarch *);
diff --git a/gdb/selftest-arch.c b/gdb/selftest-arch.c
index f434da718d5..1fe0b2d59b4 100644
--- a/gdb/selftest-arch.c
+++ b/gdb/selftest-arch.c
@@ -68,6 +68,7 @@ foreach_arch_test_generator (const std::string &name,
{
struct gdbarch_info info;
info.bfd_arch_info = bfd_scan_arch (arch);
+ info.osabi = GDB_OSABI_NONE;
struct gdbarch *gdbarch = gdbarch_find_by_info (info);
SELF_CHECK (gdbarch != NULL);
function (gdbarch);
next prev parent reply other threads:[~2022-06-02 18:35 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-06-01 10:41 Tom de Vries
2022-06-01 14:32 ` Pedro Alves
2022-06-01 16:36 ` Tom de Vries
2022-06-01 17:10 ` Pedro Alves
2022-06-01 19:08 ` Simon Marchi
2022-06-02 15:44 ` Tom de Vries
2022-06-02 17:44 ` Simon Marchi
2022-06-02 18:35 ` Tom de Vries [this message]
2022-06-04 7:23 ` Tom de Vries
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=79403061-0d96-654b-03dd-032ede7fbdb3@suse.de \
--to=tdevries@suse.de \
--cc=gdb-patches@sourceware.org \
--cc=pedro@palves.net \
--cc=simark@simark.ca \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).