Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Tom de Vries via Gdb-patches <gdb-patches@sourceware.org>
To: Pedro Alves <pedro@palves.net>, gdb-patches@sourceware.org
Subject: Re: [PATCH][gdb] Fix warning in foreach_arch selftests
Date: Wed, 1 Jun 2022 18:36:40 +0200	[thread overview]
Message-ID: <fa268877-17bd-490f-3231-d609d72993d8@suse.de> (raw)
In-Reply-To: <45511000-6883-53f4-cec6-395cd0c5b216@palves.net>

[-- Attachment #1: Type: text/plain, Size: 4326 bytes --]

On 6/1/22 16:32, Pedro Alves wrote:
> On 2022-06-01 11:41, Tom de Vries via Gdb-patches wrote:
>> Hi,
>>
>> 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.
>>
>> Tested on x86_64-linux.
>>
>> Any comments?
>>
>> Thanks,
>> - Tom
>>
>> [gdb] Fix warning in foreach_arch selftests
>>
>> ---
>>   gdb/osabi.c         | 13 +++++++++++++
>>   gdb/osabi.h         |  6 ++++++
>>   gdb/selftest-arch.c | 15 ++++++++++++++-
>>   3 files changed, 33 insertions(+), 1 deletion(-)
>>
>> diff --git a/gdb/osabi.c b/gdb/osabi.c
>> index bbd7635532f..c3f221df969 100644
>> --- a/gdb/osabi.c
>> +++ b/gdb/osabi.c
>> @@ -633,6 +633,19 @@ set_osabi (const char *args, int from_tty, struct cmd_list_element *c)
>>       internal_error (__FILE__, __LINE__, _("Updating OS ABI failed."));
>>   }
>>   
>> +void
>> +set_osabi (const char *arg)
>> +{
>> +  set_osabi_string = arg;
>> +  set_osabi (NULL, 0, NULL);
>> +}
>> +
>> +const char *
>> +get_osabi ()
>> +{
>> +  return set_osabi_string;
>> +}
> 
> Missing usual "see foo.h" comments.
> 

Hi,

thanks for the review.

Done.

>> +
>>   static void
>>   show_osabi (struct ui_file *file, int from_tty, struct cmd_list_element *c,
>>   	    const char *value)
>> diff --git a/gdb/osabi.h b/gdb/osabi.h
>> index be016732cbc..eb5d88699e7 100644
>> --- a/gdb/osabi.h
>> +++ b/gdb/osabi.h
>> @@ -89,4 +89,10 @@ const char *osabi_triplet_regexp (enum gdb_osabi osabi);
>>   void generic_elf_osabi_sniff_abi_tag_sections (bfd *, asection *,
>>   					       enum gdb_osabi *);
>>   
>> +/* Set osabi to ARG.  */
>> +extern void set_osabi (const char *arg);
>> +
>> +/* Return current osabi setting.  */
>> +extern const char *get_osabi ();
>> +
>>   #endif /* OSABI_H */
>> diff --git a/gdb/selftest-arch.c b/gdb/selftest-arch.c
>> index f434da718d5..a631f52e31e 100644
>> --- a/gdb/selftest-arch.c
>> +++ b/gdb/selftest-arch.c
>> @@ -66,12 +66,25 @@ foreach_arch_test_generator (const std::string &name,
>>         auto test_fn
>>   	= ([=] ()
>>   	   {
>> +	     /* Prevent warnings when setting architecture with current osabi
>> +		settings, like:
>> +		  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.  */
>> +	     const char *save_osabi = get_osabi ();
>> +	     set_osabi ("none");
>> +
> 
> A bit of an odd API to have to pass the string name in.  I'd think an API that
> takes an enum gdb_osabi would be better, as that way you don't have to worry
> about either passing the wrong string, or even worry about whether it's the
> string contents that counts (whether the function internally uses strcmp),
> or the string's address (the function internally compares pointers, assuming
> the only strings that will be passed down are the ones in the command's enum
> strings array).
> 

Done.


>>   	     struct gdbarch_info info;
>>   	     info.bfd_arch_info = bfd_scan_arch (arch);
>>   	     struct gdbarch *gdbarch = gdbarch_find_by_info (info);
>>   	     SELF_CHECK (gdbarch != NULL);
>> +
>>   	     function (gdbarch);
>> -	     reset ();
>> +
>> +	     SCOPE_EXIT {
>> +	       reset ();
>> +	       set_osabi (save_osabi);
>> +	     };
> 
> Please format as:
> 
> 	     SCOPE_EXIT
>                 {
> 	         reset ();
> 	         set_osabi (save_osabi);
> 	       };
> 
> And you'll also need to move this to right after the set_osabi("none") call.  Doing it
> here is too late.  SCOPE_EXIT creates a RAII object on the stack, which needs to exist
> before any of the code that may throw right after set_osabi("none").
> 

Done, thanks for catching that.

How does this look?

Thanks,
- Tom


[-- Attachment #2: 0002-gdb-Fix-warning-in-foreach_arch-selftests.patch --]
[-- Type: text/x-patch, Size: 5403 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.

Tested on x86_64-linux.

---
 gdb/osabi.c         | 50 ++++++++++++++++++++++++++++++++++++++------------
 gdb/osabi.h         | 13 +++++++++++++
 gdb/selftest-arch.c | 18 +++++++++++++++++-
 3 files changed, 68 insertions(+), 13 deletions(-)

diff --git a/gdb/osabi.c b/gdb/osabi.c
index bbd7635532f..776e62069e5 100644
--- a/gdb/osabi.c
+++ b/gdb/osabi.c
@@ -32,7 +32,7 @@
 #endif
 
 /* State for the "set osabi" command.  */
-static enum { osabi_auto, osabi_default, osabi_user } user_osabi_state;
+static enum gdb_osabi_mode user_osabi_state;
 static enum gdb_osabi user_selected_osabi;
 static const char *gdb_osabi_available_names[GDB_OSABI_INVALID + 3] = {
   "auto",
@@ -595,15 +595,48 @@ generic_elf_osabi_sniffer (bfd *abfd)
   return osabi;
 }
 \f
+/* See osabi.h.  */
+
+void
+set_osabi (enum gdb_osabi_mode mode, enum gdb_osabi osabi)
+{
+  if (mode == osabi_auto)
+    user_osabi_state = osabi_auto;
+  else if (mode == osabi_default)
+    {
+      user_selected_osabi = GDB_OSABI_DEFAULT;
+      user_osabi_state = osabi_user;
+    }
+  else
+    {
+      user_selected_osabi = osabi;
+      user_osabi_state = osabi_user;
+    }
+
+  /* NOTE: At some point (true multiple architectures) we'll need to be more
+     graceful here.  */
+  gdbarch_info info;
+  if (! gdbarch_update_p (info))
+    internal_error (__FILE__, __LINE__, _("Updating OS ABI failed."));
+}
+
+/* See osabi.h.  */
+
+void
+get_osabi (enum gdb_osabi_mode &mode, enum gdb_osabi &osabi)
+{
+  mode = user_osabi_state;
+  osabi = user_selected_osabi;
+}
+
 static void
 set_osabi (const char *args, int from_tty, struct cmd_list_element *c)
 {
   if (strcmp (set_osabi_string, "auto") == 0)
-    user_osabi_state = osabi_auto;
+    set_osabi (osabi_auto, GDB_OSABI_INVALID);
   else if (strcmp (set_osabi_string, "default") == 0)
     {
-      user_selected_osabi = GDB_OSABI_DEFAULT;
-      user_osabi_state = osabi_user;
+      set_osabi (osabi_default, GDB_OSABI_INVALID);
     }
   else
     {
@@ -615,8 +648,7 @@ set_osabi (const char *args, int from_tty, struct cmd_list_element *c)
 
 	  if (strcmp (set_osabi_string, gdbarch_osabi_name (osabi)) == 0)
 	    {
-	      user_selected_osabi = osabi;
-	      user_osabi_state = osabi_user;
+	      set_osabi (osabi_user, osabi);
 	      break;
 	    }
 	}
@@ -625,12 +657,6 @@ set_osabi (const char *args, int from_tty, struct cmd_list_element *c)
 			_("Invalid OS ABI \"%s\" passed to command handler."),
 			set_osabi_string);
     }
-
-  /* NOTE: At some point (true multiple architectures) we'll need to be more
-     graceful here.  */
-  gdbarch_info info;
-  if (! gdbarch_update_p (info))
-    internal_error (__FILE__, __LINE__, _("Updating OS ABI failed."));
 }
 
 static void
diff --git a/gdb/osabi.h b/gdb/osabi.h
index be016732cbc..3737a77d50e 100644
--- a/gdb/osabi.h
+++ b/gdb/osabi.h
@@ -50,6 +50,13 @@ enum gdb_osabi
   GDB_OSABI_INVALID		/* keep this last */
 };
 
+enum gdb_osabi_mode
+{
+  osabi_auto,
+  osabi_default,
+  osabi_user
+};
+
 /* Register an OS ABI sniffer.  Each arch/flavour may have more than
    one sniffer.  This is used to e.g. differentiate one OS's a.out from
    another.  The first sniffer to return something other than
@@ -89,4 +96,10 @@ const char *osabi_triplet_regexp (enum gdb_osabi osabi);
 void generic_elf_osabi_sniff_abi_tag_sections (bfd *, asection *,
 					       enum gdb_osabi *);
 
+/* Set osabi to MODE/OSABI.  */
+extern void set_osabi (enum gdb_osabi_mode mode, enum gdb_osabi osabi);
+
+/* Return current osabi setting in MODE/OSABI.  */
+extern void get_osabi (enum gdb_osabi_mode &mode, enum gdb_osabi &osabi);
+
 #endif /* OSABI_H */
diff --git a/gdb/selftest-arch.c b/gdb/selftest-arch.c
index f434da718d5..a4d9de5a74d 100644
--- a/gdb/selftest-arch.c
+++ b/gdb/selftest-arch.c
@@ -66,12 +66,28 @@ foreach_arch_test_generator (const std::string &name,
       auto test_fn
 	= ([=] ()
 	   {
+	     /* Prevent warnings when setting architecture with current osabi
+		settings, like:
+		  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.  */
+	     enum gdb_osabi_mode mode;
+	     enum gdb_osabi osabi;
+	     get_osabi (mode, osabi);
+
+	     set_osabi (osabi_user, GDB_OSABI_NONE);
+	     SCOPE_EXIT
+	       {
+		 reset ();
+		 set_osabi (mode, osabi);
+	       };
+
 	     struct gdbarch_info info;
 	     info.bfd_arch_info = bfd_scan_arch (arch);
 	     struct gdbarch *gdbarch = gdbarch_find_by_info (info);
 	     SELF_CHECK (gdbarch != NULL);
+
 	     function (gdbarch);
-	     reset ();
 	   });
 
       tests.emplace_back (string_printf ("%s::%s", name.c_str (), arch),

  reply	other threads:[~2022-06-01 16:37 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-06-01 10:41 Tom de Vries via Gdb-patches
2022-06-01 14:32 ` Pedro Alves
2022-06-01 16:36   ` Tom de Vries via Gdb-patches [this message]
2022-06-01 17:10     ` Pedro Alves
2022-06-01 19:08     ` Simon Marchi via Gdb-patches
2022-06-02 15:44       ` Tom de Vries via Gdb-patches
2022-06-02 17:44         ` Simon Marchi via Gdb-patches
2022-06-02 18:35           ` Tom de Vries via Gdb-patches
2022-06-04  7:23             ` Tom de Vries via Gdb-patches

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=fa268877-17bd-490f-3231-d609d72993d8@suse.de \
    --to=gdb-patches@sourceware.org \
    --cc=pedro@palves.net \
    --cc=tdevries@suse.de \
    /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