Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: "Six, Lancelot" <Lancelot.Six@amd.com>
To: "Aktemur, Baris" <TankutBaris.Aktemur@amd.com>,
	"gdb-patches@sourceware.org" <gdb-patches@sourceware.org>,
	"tom@tromey.com" <tom@tromey.com>
Subject: RE: [PATCH] gdb, amdgpu: add config guards for AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE
Date: Thu, 15 Oct 2026 17:13:23 +0000	[thread overview]
Message-ID: <PH0PR12MB5607A3A17F9E971A24198D6E839C2@PH0PR12MB5607.namprd12.prod.outlook.com> (raw)
In-Reply-To: <20261015170450.2520815-1-tankutbaris.aktemur@amd.com>

AMD General

> -----Original Message-----
> From: Aktemur, Baris <TankutBaris.Aktemur@amd.com>
> Sent: 15 October 2026 18:05
> To: gdb-patches@sourceware.org; Six, Lancelot <Lancelot.Six@amd.com>;
> tom@tromey.com
> Subject: [PATCH] gdb, amdgpu: add config guards for
> AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE
>
> Commit e52a8aca2baf6481c4e8fea45530283759096be5 "gdb, amdgpu:
> handle
> UNAVAILABLE error when accessing memory" used an enum value from
> amd-dbgapi named AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE
> that may
> not be available in all installed packages.  This causes a build
> failure as reported by Tom Tromey at
>
>   https://sourceware.org/pipermail/gdb-patches/2026-October/231115.html
>
> Detect at configuration time if the enum value exists and guard the
> newly added code accordingly.
> ---
>  gdb/amd-dbgapi-target.c |  4 ++++
>  gdb/config.in           |  3 +++
>  gdb/configure           | 36 ++++++++++++++++++++++++++++++++++++
>  gdb/configure.ac        | 21 +++++++++++++++++++++
>  4 files changed, 64 insertions(+)
>
> diff --git a/gdb/amd-dbgapi-target.c b/gdb/amd-dbgapi-target.c
> index 2fd175de6f7..1446187c75f 100644
> --- a/gdb/amd-dbgapi-target.c
> +++ b/gdb/amd-dbgapi-target.c
> @@ -786,11 +786,13 @@ amd_dbgapi_target::xfer_partial (enum
> target_object object, const char *annex,
>                                     AMD_DBGAPI_ADDRESS_SPACE_GLOBAL,
>                                     offset, &len, writebuf);
>
> +#if AMD_DBGAPI_HAS_STATUS_ERROR_MEMORY_UNAVAILABLE
>    if (status == AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE)
>      {
>        *xfered_len = requested_len;
>        return TARGET_XFER_UNAVAILABLE;
>      }
> +#endif
>
>    if (status != AMD_DBGAPI_STATUS_SUCCESS)
>      return TARGET_XFER_E_IO;
> @@ -2576,8 +2578,10 @@ amd_dbgapi_xfer_global_memory_callback
>
>    if (status == TARGET_XFER_EOF)
>      return AMD_DBGAPI_STATUS_ERROR_PROCESS_EXITED;
> +#if AMD_DBGAPI_HAS_STATUS_ERROR_MEMORY_UNAVAILABLE
>    else if (status == TARGET_XFER_UNAVAILABLE)
>      return AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE;
> +#endif
>    else if (status != TARGET_XFER_OK)
>      return AMD_DBGAPI_STATUS_ERROR_MEMORY_ACCESS;
>
> diff --git a/gdb/config.in b/gdb/config.in
> index 7ea0e9cfce8..93060f573c5 100644
> --- a/gdb/config.in
> +++ b/gdb/config.in
> @@ -6,6 +6,9 @@
>  /* Additional directories to look for separate debug info. */
>  #undef ADDITIONAL_DEBUG_DIRS
>
> +/* Define if AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE is
> available. */
> +#undef AMD_DBGAPI_HAS_STATUS_ERROR_MEMORY_UNAVAILABLE
> +
>  /* Directories from which to load auto-loaded scripts. */
>  #undef AUTO_LOAD_DIR
>
> diff --git a/gdb/configure b/gdb/configure
> index 0ed49887480..eaa8588fe69 100755
> --- a/gdb/configure
> +++ b/gdb/configure
> @@ -25105,6 +25105,42 @@ $as_echo "#define HAVE_AMD_DBGAPI 1"
> >>confdefs.h
>       as_fn_error $? "amd-dbgapi not supported for host ${gdb_host}"
> "$LINENO" 5
>       ;;
>      esac
> +
> +    # Detect if AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE can be
> used.
> +    save_CPPFLAGS="$CPPFLAGS"
> +    CPPFLAGS="$CPPFLAGS $AMD_DBGAPI_CFLAGS"
> +    { $as_echo "$as_me:${as_lineno-$LINENO}: checking whether
> AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE is available" >&5
> +$as_echo_n "checking whether
> AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE is available... " >&6; }
> +    cat confdefs.h - <<_ACEOF >conftest.$ac_ext
> +/* end confdefs.h.  */
> +#include <amd-dbgapi/amd-dbgapi.h>
> +int
> +main ()
> +{
> +amd_dbgapi_status_t s =
> AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE;
> +
> +  ;
> +  return 0;
> +}
> +_ACEOF
> +if ac_fn_c_try_compile "$LINENO"; then :
> +  amd_dbgapi_has_status_error_memory_unavailable=yes
> +       { $as_echo "$as_me:${as_lineno-$LINENO}: result: yes" >&5
> +$as_echo "yes" >&6; }
> +else
> +  amd_dbgapi_has_status_error_memory_unavailable=no
> +       { $as_echo "$as_me:${as_lineno-$LINENO}: result: no" >&5
> +$as_echo "no" >&6; }
> +fi
> +rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
> +    CPPFLAGS="$save_CPPFLAGS"
> +
> +    if test "x${amd_dbgapi_has_status_error_memory_unavailable}" = "xyes";
> then
> +
> +$as_echo "#define
> AMD_DBGAPI_HAS_STATUS_ERROR_MEMORY_UNAVAILABLE 1"
> >>confdefs.h
> +
> +    fi
> +
>    elif test "$gdb_require_amd_dbgapi" = true -o "$with_amd_dbgapi" = yes;
> then
>      # amd-dbgapi was not found and...
>      #
> diff --git a/gdb/configure.ac b/gdb/configure.ac
> index e0117ede292..274d248a64d 100644
> --- a/gdb/configure.ac
> +++ b/gdb/configure.ac
> @@ -368,6 +368,27 @@ if test "$gdb_require_amd_dbgapi" = true \
>       AC_MSG_ERROR([amd-dbgapi not supported for host ${gdb_host}])
>       ;;
>      esac
> +
> +    # Detect if AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE can be
> used.
> +    save_CPPFLAGS="$CPPFLAGS"
> +    CPPFLAGS="$CPPFLAGS $AMD_DBGAPI_CFLAGS"
> +    AC_MSG_CHECKING([whether
> AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE is available])
> +    AC_COMPILE_IFELSE(
> +      [AC_LANG_PROGRAM(
> +       [[#include <amd-dbgapi/amd-dbgapi.h>]],
> +       [[amd_dbgapi_status_t s =
> AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE;]]
> +       )],

Hi,

Silly question.  If CPPFLAGS contains "-Werror -Wall", wouldn't this risk to complain about s not being used?  Would fallback to the conservative solution to not use MEMORY_UNAVAILABLE, so nothing really dramatic.

Might be worth using "(void) AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE" (or anything to avoid the potential warning).

Best,
Lancelot.

> +       [amd_dbgapi_has_status_error_memory_unavailable=yes
> +       AC_MSG_RESULT([yes])],
> +       [amd_dbgapi_has_status_error_memory_unavailable=no
> +       AC_MSG_RESULT([no])])
> +    CPPFLAGS="$save_CPPFLAGS"
> +
> +    if test "x${amd_dbgapi_has_status_error_memory_unavailable}" = "xyes";
> then
> +
> AC_DEFINE(AMD_DBGAPI_HAS_STATUS_ERROR_MEMORY_UNAVAILABLE, 1,
> +             [Define if
> AMD_DBGAPI_STATUS_ERROR_MEMORY_UNAVAILABLE is available.])
> +    fi
> +
>    elif test "$gdb_require_amd_dbgapi" = true -o "$with_amd_dbgapi" = yes;
> then
>      # amd-dbgapi was not found and...
>      #
> --
> 2.53.0


      reply	other threads:[~2026-10-15 17:14 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-15 17:04 Tankut Baris Aktemur
2026-10-15 17:13 ` Six, Lancelot [this message]

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=PH0PR12MB5607A3A17F9E971A24198D6E839C2@PH0PR12MB5607.namprd12.prod.outlook.com \
    --to=lancelot.six@amd.com \
    --cc=TankutBaris.Aktemur@amd.com \
    --cc=gdb-patches@sourceware.org \
    --cc=tom@tromey.com \
    /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