From: Lancelot SIX <lancelot.six@amd.com>
To: Simon Marchi <simon.marchi@efficios.com>
Cc: <gdb-patches@sourceware.org>
Subject: Re: [PATCH v2 4/6] gdb/amd-dbgapi: pass amd_dbgapi_inferior_info to process_one_event
Date: Mon, 16 Jun 2025 11:32:59 +0100 [thread overview]
Message-ID: <rk2xk5bifjeusg2fmhonbl4axpoebm6r5bo3f6fleque5wubka@2s3hf2lbgffu> (raw)
In-Reply-To: <20250609161004.97787-4-simon.marchi@efficios.com>
On Mon, Jun 09, 2025 at 12:09:00PM -0400, Simon Marchi wrote:
> New in v2: this patch is new
>
> Pass the amd_dbgapi_inferior_info object from process_event_queue to
> process_one_event. Since process_event_queue pulls events for one
> specific inferior, we know for which inferior the event is. This
> removes the need for process_one_event to do two dbgapi calls to get the
> relevant pid. If also removes one inferior lookup.
>
Hi Simon,
Minor comments below.
> Change-Id: I22927e4b6251513eb3be95785082058aa3d09954
> ---
> gdb/amd-dbgapi-target.c | 58 +++++++++++++++--------------------------
> 1 file changed, 21 insertions(+), 37 deletions(-)
>
> diff --git a/gdb/amd-dbgapi-target.c b/gdb/amd-dbgapi-target.c
> index 888552e00d3e..80111359c8cd 100644
> --- a/gdb/amd-dbgapi-target.c
> +++ b/gdb/amd-dbgapi-target.c
> @@ -1143,32 +1143,14 @@ add_gpu_thread (inferior *inf, ptid_t wave_ptid)
> /* Process an event that was just pulled out of the amd-dbgapi library. */
>
> static void
> -process_one_event (amd_dbgapi_event_id_t event_id,
> +process_one_event (amd_dbgapi_inferior_info &info,
> + amd_dbgapi_event_id_t event_id,
> amd_dbgapi_event_kind_t event_kind)
> {
> /* Automatically mark this event processed when going out of scope. */
> scoped_amd_dbgapi_event_processed mark_event_processed (event_id);
>
> - amd_dbgapi_process_id_t process_id;
> - amd_dbgapi_status_t status
> - = amd_dbgapi_event_get_info (event_id, AMD_DBGAPI_EVENT_INFO_PROCESS,
> - sizeof (process_id), &process_id);
> - if (status != AMD_DBGAPI_STATUS_SUCCESS)
> - error (_("event_get_info for event_%ld failed (%s)"), event_id.handle,
> - get_status_string (status));
> -
> - amd_dbgapi_os_process_id_t pid;
> - status = amd_dbgapi_process_get_info (process_id,
> - AMD_DBGAPI_PROCESS_INFO_OS_ID,
> - sizeof (pid), &pid);
> - if (status != AMD_DBGAPI_STATUS_SUCCESS)
> - error (_("process_get_info for process_%ld failed (%s)"),
> - process_id.handle, get_status_string (status));
> -
> - auto *proc_target = current_inferior ()->process_target ();
> - inferior *inf = find_inferior_pid (proc_target, pid);
> - gdb_assert (inf != nullptr);
> - amd_dbgapi_inferior_info *info = get_amd_dbgapi_inferior_info (inf);
> + gdb_assert (info.inf != nullptr);
>
> switch (event_kind)
> {
> @@ -1176,14 +1158,14 @@ process_one_event (amd_dbgapi_event_id_t event_id,
> case AMD_DBGAPI_EVENT_KIND_WAVE_STOP:
> {
> amd_dbgapi_wave_id_t wave_id;
> - status
> + amd_dbgapi_status_t status
> = amd_dbgapi_event_get_info (event_id, AMD_DBGAPI_EVENT_INFO_WAVE,
> sizeof (wave_id), &wave_id);
> if (status != AMD_DBGAPI_STATUS_SUCCESS)
> error (_("event_get_info for event_%ld failed (%s)"),
> event_id.handle, get_status_string (status));
>
> - ptid_t event_ptid = make_gpu_ptid (pid, wave_id);
> + ptid_t event_ptid = make_gpu_ptid (info.inf->pid, wave_id);
> target_waitstatus ws;
>
> amd_dbgapi_wave_stop_reasons_t stop_reason;
> @@ -1224,9 +1206,10 @@ process_one_event (amd_dbgapi_event_id_t event_id,
> else
> ws.set_stopped (GDB_SIGNAL_0);
>
> - thread_info *thread = proc_target->find_thread (event_ptid);
> - if (thread == nullptr)
> - thread = add_gpu_thread (inf, event_ptid);
> + thread_info *thread =
> + info.inf->process_target()->find_thread(event_ptid);
> + if (thread == nullptr)
The indentation is off here. When applying the patch, GIT warns about
indenting with spaces:
.../binutils-gdb/rebase-apply/patch:69: indent with spaces.
thread_info *thread =
.../binutils-gdb/rebase-apply/patch:70: indent with spaces.
info.inf->process_target()->find_thread(event_ptid);
.../binutils-gdb/rebase-apply/patch:71: indent with spaces.
if (thread == nullptr)
warning: 3 lines add whitespace errors.
Also, the `=` symbol should be after the newline, not just before it.
With those style issues fixed, this looks good to me, thanks.
Approved-by: Lancelot Six <lancelot.six@amd.com> (amdgpu)
Best,
Lancelot.
> + thread = add_gpu_thread (info.inf, event_ptid);
>
> /* If the wave is stopped because of a software breakpoint, the
> program counter needs to be adjusted so that it points to the
> @@ -1248,7 +1231,7 @@ process_one_event (amd_dbgapi_event_id_t event_id,
> error (_("wave_get_info for wave_%ld failed (%s)"),
> wave_id.handle, get_status_string (status));
>
> - info->wave_events.emplace_back (event_ptid, ws);
> + info.wave_events.emplace_back (event_ptid, ws);
> break;
> }
>
> @@ -1266,7 +1249,7 @@ process_one_event (amd_dbgapi_event_id_t event_id,
> When amd_dbgapi_target_breakpoint::check_status is called, the current
> inferior is the inferior that hit the breakpoint, which should still be
> the case now. */
> - gdb_assert (inf == current_inferior ());
> + gdb_assert (info.inf == current_inferior ());
> handle_solib_event ();
> break;
>
> @@ -1280,22 +1263,23 @@ process_one_event (amd_dbgapi_event_id_t event_id,
> {
> amd_dbgapi_runtime_state_t runtime_state;
>
> - status = amd_dbgapi_event_get_info (event_id,
> - AMD_DBGAPI_EVENT_INFO_RUNTIME_STATE,
> - sizeof (runtime_state),
> - &runtime_state);
> + amd_dbgapi_status_t status
> + = amd_dbgapi_event_get_info (event_id,
> + AMD_DBGAPI_EVENT_INFO_RUNTIME_STATE,
> + sizeof (runtime_state),
> + &runtime_state);
> if (status != AMD_DBGAPI_STATUS_SUCCESS)
> error (_("event_get_info for event_%ld failed (%s)"),
> event_id.handle, get_status_string (status));
>
> gdb_assert (runtime_state == AMD_DBGAPI_RUNTIME_STATE_UNLOADED);
> gdb_assert
> - (info->runtime_state == AMD_DBGAPI_RUNTIME_STATE_LOADED_SUCCESS);
> + (info.runtime_state == AMD_DBGAPI_RUNTIME_STATE_LOADED_SUCCESS);
>
> - info->runtime_state = runtime_state;
> + info.runtime_state = runtime_state;
>
> - gdb_assert (inf->target_is_pushed (&the_amd_dbgapi_target));
> - inf->unpush_target (&the_amd_dbgapi_target);
> + gdb_assert (info.inf->target_is_pushed (&the_amd_dbgapi_target));
> + info.inf->unpush_target (&the_amd_dbgapi_target);
> }
> break;
>
> @@ -1367,7 +1351,7 @@ process_event_queue (amd_dbgapi_inferior_info &info,
> if (event_id == AMD_DBGAPI_EVENT_NONE || event_kind == until_event_kind)
> return event_id;
>
> - process_one_event (event_id, event_kind);
> + process_one_event (info, event_id, event_kind);
> }
> }
>
> --
> 2.49.0
next prev parent reply other threads:[~2025-06-16 10:33 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-06-09 16:08 [PATCH v2 1/6] gdb/amd-dbgapi: remove unnecessary AMD_DBGAPI_EVENT_KIND_NONE argument Simon Marchi
2025-06-09 16:08 ` [PATCH v2 2/6] gdb/amd-dbgapi: add assert in require_forward_progress Simon Marchi
2025-06-09 16:08 ` [PATCH v2 3/6] gdb/amd-dbgapi: pass amd_dbgapi_inferior_info to process_event_queue Simon Marchi
2025-06-16 10:29 ` Lancelot SIX
2025-06-09 16:09 ` [PATCH v2 4/6] gdb/amd-dbgapi: pass amd_dbgapi_inferior_info to process_one_event Simon Marchi
2025-06-16 10:32 ` Lancelot SIX [this message]
2025-06-16 14:24 ` Simon Marchi
2025-06-09 16:09 ` [PATCH v2 5/6] gdb/amd-dbgapi: factor out require_forward_progress overload to target one inferior Simon Marchi
2025-06-09 16:09 ` [PATCH v2 6/6] gdb/amd-dbgapi: disable forward progress requirement in amd_dbgapi_target_breakpoint::check_status Simon Marchi
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=rk2xk5bifjeusg2fmhonbl4axpoebm6r5bo3f6fleque5wubka@2s3hf2lbgffu \
--to=lancelot.six@amd.com \
--cc=gdb-patches@sourceware.org \
--cc=simon.marchi@efficios.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