public inbox for intel-xe@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Matt Roper <matthew.d.roper@intel.com>
To: Clint Taylor <clinton.a.taylor@intel.com>
Cc: <Intel-xe@lists.freedesktop.org>
Subject: Re: [PATCH] drm/xe: Apply WA_14026810691 to engine
Date: Thu, 9 Apr 2026 15:58:57 -0700	[thread overview]
Message-ID: <20260409225857.GA6301@mdroper-desk1.amr.corp.intel.com> (raw)
In-Reply-To: <20260409215705.1456888-1-clinton.a.taylor@intel.com>

On Thu, Apr 09, 2026 at 02:57:05PM -0700, Clint Taylor wrote:
> Apply WA_14026810691 to following IPs:
> Xe3p_LPG
> Xe3p_XPC
> 
> Signed-off-by: Clint Taylor <clinton.a.taylor@intel.com>
> ---
>  drivers/gpu/drm/xe/regs/xe_gt_regs.h | 3 +++
>  drivers/gpu/drm/xe/xe_wa.c           | 6 ++++++
>  2 files changed, 9 insertions(+)
> 
> diff --git a/drivers/gpu/drm/xe/regs/xe_gt_regs.h b/drivers/gpu/drm/xe/regs/xe_gt_regs.h
> index aa267c2f6162..6a16ac1aca96 100644
> --- a/drivers/gpu/drm/xe/regs/xe_gt_regs.h
> +++ b/drivers/gpu/drm/xe/regs/xe_gt_regs.h
> @@ -536,6 +536,9 @@
>  #define   SLM_WMTP_RESTORE			REG_BIT(11)
>  #define   RES_CHK_SPR_DIS			REG_BIT(6)
>  
> +#define TDL_TSL_CHICKEN2			XE_REG_MCR(0xe4cc, XE_REG_OPTION_MASKED)
> +#define   TILEY_LOCALID				REG_BIT(2)
> +
>  #define ROW_CHICKEN				XE_REG_MCR(0xe4f0, XE_REG_OPTION_MASKED)
>  #define   UGM_BACKUP_MODE			REG_BIT(13)
>  #define   MDQ_ARBITRATION_MODE			REG_BIT(12)
> diff --git a/drivers/gpu/drm/xe/xe_wa.c b/drivers/gpu/drm/xe/xe_wa.c
> index 2ec70be78bf9..30dd1687cc8d 100644
> --- a/drivers/gpu/drm/xe/xe_wa.c
> +++ b/drivers/gpu/drm/xe/xe_wa.c
> @@ -609,6 +609,12 @@ static const struct xe_rtp_entry_sr engine_was[] = {
>  		       FUNC(xe_rtp_match_first_render_or_compute)),
>  	  XE_RTP_ACTIONS(SET(ROW_CHICKEN3, DIS_EU_GRF_POISON_TO_LSC))
>  	},
> +	{ XE_RTP_NAME("14026810691"),

There doesn't appear to be any record for this in the workaround
database.  Looking closer, this appears to be a general (non-workaround)
software programming guidance, currently documented in a programming
note on bspec page 73720:

        "SW to set the value to TILEY_2X2 walk pattern"

and also the tuning guide on bspec page 72161:

        "Program bit 2 to 1 for all clients to have TileY generated as
        2x2 walk pattern"

Since this is a tuning setting rather than a workaround, we should move
the implementation over to the engine_tunings[] table of xe_tuning.c
(and give it a descriptive name like "Tuning: TileY 2x2 Walk" rather
than a number).  Also for now this appears to be an open-ended
recommendation that should carry forward to future IP versions as well,
so we can change the range to (3510, XE_RTP_END_VERSION_UNDEFINED) to
match that. If/when the hardware teams eventually change the hardware's
default setting on some future platform down the road, we'll come back
and add a specific upper bound.


Matt

> +	  XE_RTP_RULES(GRAPHICS_VERSION_RANGE(3510, 3511),
> +		       FUNC(xe_rtp_match_first_render_or_compute)),
> +	  XE_RTP_ACTIONS(SET(TDL_TSL_CHICKEN2, TILEY_LOCALID))
> +	},
> +
>  };
>  
>  static const struct xe_rtp_entry_sr lrc_was[] = {
> -- 
> 2.25.1
> 

-- 
Matt Roper
Graphics Software Engineer
Linux GPU Platform Enablement
Intel Corporation

  parent reply	other threads:[~2026-04-09 22:59 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-09 21:57 [PATCH] drm/xe: Apply WA_14026810691 to engine Clint Taylor
2026-04-09 22:06 ` ✓ CI.KUnit: success for " Patchwork
2026-04-09 22:48 ` ✓ Xe.CI.BAT: " Patchwork
2026-04-09 22:58 ` Matt Roper [this message]
2026-04-10  0:10 ` ✗ Xe.CI.FULL: failure " Patchwork
  -- strict thread matches above, loose matches on Subject: below --
2026-04-13 20:08 [PATCH] " Clint Taylor
2026-04-15 19:47 ` Matt Roper
2026-04-15 20:18   ` Taylor, Clinton A

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=20260409225857.GA6301@mdroper-desk1.amr.corp.intel.com \
    --to=matthew.d.roper@intel.com \
    --cc=Intel-xe@lists.freedesktop.org \
    --cc=clinton.a.taylor@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