Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Gustavo Sousa <gustavo.sousa@intel.com>
To: <intel-xe@lists.freedesktop.org>
Cc: Matt Atwood <matthew.s.atwood@intel.com>,
	Dnyaneshwar Bhadane <dnyaneshwar.bhadane@intel.com>,
	Matt Roper <matthew.d.roper@intel.com>,
	Shekhar Chauhan <shekhar.chauhan@intel.com>,
	Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Subject: Re: [PATCH v3 00/10] drm/xe: Clean up adjust_idledly() and implement Wa_16031259911
Date: Mon, 9 Nov 2026 20:50:30 -0300	[thread overview]
Message-ID: <871p8t36pl.fsf@intel.com> (raw)
In-Reply-To: <20261105-wa_16031259911-v3-0-da6ec790aa29@intel.com>

Gustavo Sousa <gustavo.sousa@intel.com> writes:

> There is a new workaround, namely Wa_16031259911, to be added to the
> xe driver that applies to graphics IP versions 35.10 and 35.11.  The
> workaround is similar to existing ones that apply adjustments to
> RING_IDLEDLY.
>
> The current implementation can be improved though: adjust_idledly()
> can be cleaned up and refactored to make the code easier to follow
> sequentially as well as to add new workarounds (such as
> Wa_16031259911); the GuC ADS regset accounting has a subtle issue
> where the addition of RING_IDLEDLY is not accounted for.
>
> This series addresses those issues first and then adds the
> implementation of Wa_16031259911 as the final step.
>
> Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>
> ---
> Changes in v3:
> - Rebased on top of drm-tip to fix a conflict and resending for fresh
>   CI results. 
> - Link to v2: https://patch.msgid.link/20261030-wa_16031259911-v2-0-919fa1c27070@intel.com

Added MattA's r-b from v2 to the equivalent patch in v3 that was missing
a r-b and applied to drm-xe-next.

Thanks MattA and Bala for the reviews!

--
Gustavo Sousa

>
> Changes in v2:
> - Change idledly_units_ps and maxcnt_units_ps to u64 in adjust_idledly()
>   to avoid 32-bit multiplication overflow. (Sashiko, Matt)
> - Replace the call-site comment for adjust_idledly() with a comment at
>   the function definition instead of just dropping it. (Matt)
> - Link to v1: https://patch.msgid.link/20261023-wa_16031259911-v1-0-b0453f79c4e1@intel.com
>
> ---
> Gustavo Sousa (10):
>       drm/xe: Bail from adjust_idledly() if in VF mode
>       drm/xe: Sync Wa_16023105232 in xe_was to match OOB rule
>       drm/xe: Use picosecond as common unit in adjust_idledly()
>       drm/xe: Factor out {min,max}_idledly_ps in adjust_idledly()
>       drm/xe: Bail from adjust_idledly() if no applicable workaround
>       drm/xe: Drop early and redundant workaround checks in adjust_idledly()
>       drm/xe: Replace MAXCNT comment with a comment on adjust_idledly()
>       drm/xe/guc: Increment regset size for Wa_1607983814 inside hwe loop
>       drm/xe/guc: Increment regset size for RING_IDLEDLY workarounds
>       drm/xe: Implement Wa_16031259911
>
>  drivers/gpu/drm/xe/xe_guc_ads.c    |  16 ++--
>  drivers/gpu/drm/xe/xe_hw_engine.c  | 158 ++++++++++++++++++-------------------
>  drivers/gpu/drm/xe/xe_wa.c         |   5 +-
>  drivers/gpu/drm/xe/xe_wa_oob.rules |   6 +-
>  4 files changed, 93 insertions(+), 92 deletions(-)
> ---
> base-commit: be7a4fe8d1795219f7d033830ec4cf6e8ca1ecdb
> change-id: 20261021-wa_16031259911-3a83c69d8570
>
> Best regards,
> --  
> Gustavo Sousa <gustavo.sousa@intel.com>

      parent reply	other threads:[~2026-11-09 23:50 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-05 23:31 [PATCH v3 00/10] drm/xe: Clean up adjust_idledly() and implement Wa_16031259911 Gustavo Sousa
2026-11-05 23:31 ` [PATCH v3 01/10] drm/xe: Bail from adjust_idledly() if in VF mode Gustavo Sousa
2026-11-05 23:31 ` [PATCH v3 02/10] drm/xe: Sync Wa_16023105232 in xe_was to match OOB rule Gustavo Sousa
2026-11-05 23:31 ` [PATCH v3 03/10] drm/xe: Use picosecond as common unit in adjust_idledly() Gustavo Sousa
2026-11-05 23:31 ` [PATCH v3 04/10] drm/xe: Factor out {min,max}_idledly_ps " Gustavo Sousa
2026-11-05 23:31 ` [PATCH v3 05/10] drm/xe: Bail from adjust_idledly() if no applicable workaround Gustavo Sousa
2026-11-05 23:31 ` [PATCH v3 06/10] drm/xe: Drop early and redundant workaround checks in adjust_idledly() Gustavo Sousa
2026-11-05 23:31 ` [PATCH v3 07/10] drm/xe: Replace MAXCNT comment with a comment on adjust_idledly() Gustavo Sousa
2026-11-05 23:31 ` [PATCH v3 08/10] drm/xe/guc: Increment regset size for Wa_1607983814 inside hwe loop Gustavo Sousa
2026-11-05 23:31 ` [PATCH v3 09/10] drm/xe/guc: Increment regset size for RING_IDLEDLY workarounds Gustavo Sousa
2026-11-05 23:31 ` [PATCH v3 10/10] drm/xe: Implement Wa_16031259911 Gustavo Sousa
2026-11-05 23:39 ` ✗ CI.checkpatch: warning for drm/xe: Clean up adjust_idledly() and implement Wa_16031259911 (rev3) Patchwork
2026-11-05 23:41 ` ✓ CI.KUnit: success " Patchwork
2026-11-06  0:23 ` ✓ Xe.CI.BAT: " Patchwork
2026-11-07  2:17 ` ✗ Xe.CI.FULL: failure " Patchwork
2026-11-09 23:11   ` Gustavo Sousa
2026-11-09 23:50 ` Gustavo Sousa [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=871p8t36pl.fsf@intel.com \
    --to=gustavo.sousa@intel.com \
    --cc=balasubramani.vivekanandan@intel.com \
    --cc=dnyaneshwar.bhadane@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=matthew.d.roper@intel.com \
    --cc=matthew.s.atwood@intel.com \
    --cc=shekhar.chauhan@intel.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