From: "Timur Kristóf" <timur.kristof@gmail.com>
To: amd-gfx@lists.freedesktop.org, alexander.deucher@amd.com,
Bing.Ma@amd.com, James Zhu <jamesz@amd.com>
Cc: "James Zhu" <James.Zhu@amd.com>,
"Natalie Vock" <nat@pixelcluster.dev>,
"Christian König" <christian.koenig@amd.com>
Subject: Re: [PATCH 01/12] drm/amdgpu: add profiler UAPI definitions
Date: Tue, 22 Sep 2026 17:34:13 +0200 [thread overview]
Message-ID: <-2dFiumdQ7an8e02CFlfgw@gmail.com> (raw)
In-Reply-To: <3553c2d2-ca84-4333-a3cf-e3c745eaad1f@amd.com>
On Tuesday, September 22, 2026 5:32:12 PM Central European Summer Time James
Zhu wrote:
> On 2026-09-22 11:22, Timur Kristóf wrote:
> > On Tuesday, September 22, 2026 5:20:41 PM Central European Summer Time
> > James>
> > Zhu wrote:
> >> On 2026-09-22 11:09, Timur Kristóf wrote:
> >>> On Monday, September 21, 2026 10:03:02 PM Central European Summer Time
> >>> James>
> >>>
> >>> Zhu wrote:
> >>>> Add the DRM_AMDGPU_PROFILER ioctl number and its argument structures,
> >>>> the
> >>>> UAPI for GPU performance monitoring features such as SPM (Stream
> >>>> Performance Monitor).
> >>>>
> >>>> struct drm_amdgpu_profiler_args carries an operation selector and a
> >>>> union
> >>>> of per-operation parameters, currently AMDGPU_PROFILER_OP_GET_VER and
> >>>> AMDGPU_PROFILER_OP_SPM. The SPM operations and the layout of the
> >>>> destination buffer are described by struct drm_amdgpu_spm_args and
> >>>> struct drm_amdgpu_spm_buffer_header.
> >>>>
> >>>> The ioctl number and struct layout are frozen here. The implementation
> >>>> lands in the rest of this series and the ioctl is not registered in
> >>>> amdgpu_ioctls_kms[] until then, so this header change alone does not
> >>>> make
> >>>> the interface reachable.
> >>>>
> >>>> Signed-off-by: James Zhu<James.Zhu@amd.com>
> >>>> Reviewed-by: Bing Ma<bing.ma@amd.com>
> >>>
> >>> Hi,
> >>>
> >>> Can you guys please clarify why this is necessary and what the use case
> >>> is?
> >>>
> >>> Are we sure we want a uAPI for this? RADV has been using SPM from
> >>> userspace
> >>> for many years now.
> >>
> >> [JZ] current RADV SPM is for graphic engine. These patches are for
> >> compute engine SPM.
> >>
> >> Thanks!
> >
> > Is there any reason why it can't be done in userspace for compute too?
>
> [JZ] for compute engine, user space is using ROCM stack , not RADV.
Is there any reason why SPM can't be handled from userspace?
(Regardless of whether that userspace is RADV or ROCm)
next prev parent reply other threads:[~2026-09-22 15:34 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-21 20:03 [PATCH 01/12] drm/amdgpu: add profiler UAPI definitions James Zhu
2026-09-21 20:03 ` [PATCH 02/12] drm/amdgpu: Add profiler UAPI infrastructure James Zhu
2026-09-21 20:03 ` [PATCH 03/12] drm/amdgpu: add generic RLC SPM interface layer James Zhu
2026-09-21 20:03 ` [PATCH 04/12] drm/amdgpu: add amdgpu_xcp_get_xcc_mask helper James Zhu
2026-09-22 5:22 ` Lazar, Lijo
2026-09-22 13:35 ` James Zhu
2026-09-22 13:45 ` Lazar, Lijo
2026-09-22 14:04 ` James Zhu
2026-09-22 14:11 ` James Zhu
2026-09-22 14:26 ` Lazar, Lijo
2026-09-22 15:10 ` James Zhu
2026-09-23 3:39 ` Lazar, Lijo
2026-09-23 12:36 ` James Zhu
2026-09-23 14:55 ` Lazar, Lijo
2026-09-23 15:08 ` James Zhu
2026-09-21 20:03 ` [PATCH 05/12] drm/amdgpu: add profiler/spm operation AMDGPU_SPM_OP_ACQUIRE James Zhu
2026-09-21 20:03 ` [PATCH 06/12] drm/amdgpu: add profiler/spm operation AMDGPU_SPM_OP_SET_DEST_BUF James Zhu
2026-09-21 20:03 ` [PATCH 07/12] drm/amdgpu: add profiler/spm operation AMDGPU_SPM_OP_RELEASE James Zhu
2026-09-21 20:03 ` [PATCH 08/12] drm/amdgpu: implement data dump from spm ring buffer James Zhu
2026-09-21 20:03 ` [PATCH 09/12] drm/amdgpu: workaround for spm ring buffer overflow James Zhu
2026-09-21 20:03 ` [PATCH 10/12] drm/amdgpu: add SPM ring-buffer polling thread for pre-GFX12 James Zhu
2026-09-21 20:03 ` [PATCH 11/12] drm/amdgpu: add profiler/spm support for gfx12_0 James Zhu
2026-09-21 20:03 ` [PATCH 12/12] drm/amdgpu: bump KMS driver version to 3.65 for profiler UAPI James Zhu
2026-09-22 15:09 ` [PATCH 01/12] drm/amdgpu: add profiler UAPI definitions Timur Kristóf
2026-09-22 15:20 ` James Zhu
2026-09-22 15:22 ` Timur Kristóf
2026-09-22 15:32 ` James Zhu
2026-09-22 15:34 ` Timur Kristóf [this message]
2026-09-22 15:40 ` Alex Deucher
2026-09-22 15:43 ` Christian König
2026-09-23 1:20 ` Ma, Bing
2026-09-23 7:47 ` Christian König
2026-09-24 12:33 ` Ma, Bing
2026-09-22 15:13 ` Christian König
2026-09-22 15:29 ` James Zhu
-- strict thread matches above, loose matches on Subject: below --
2026-09-21 20:00 James Zhu
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=-2dFiumdQ7an8e02CFlfgw@gmail.com \
--to=timur.kristof@gmail.com \
--cc=Bing.Ma@amd.com \
--cc=James.Zhu@amd.com \
--cc=alexander.deucher@amd.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=christian.koenig@amd.com \
--cc=jamesz@amd.com \
--cc=nat@pixelcluster.dev \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.