Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: John Harrison <john.c.harrison@intel.com>
To: Lucas De Marchi <lucas.demarchi@intel.com>,
	intel-xe <intel-xe@lists.freedesktop.org>
Cc: Matt Roper <matthew.d.roper@intel.com>
Subject: Re: [PATCH 0/2] drm/xe: Support different GuCs per GT
Date: Thu, 20 Mar 2025 13:09:36 -0700	[thread overview]
Message-ID: <cc30ee16-f0bd-45de-b709-de35b0a70ef7@intel.com> (raw)
In-Reply-To: <20250307-xe-per-gt-fw-v1-0-459574d76400@intel.com>

On 3/7/2025 10:13 AM, Lucas De Marchi wrote:
> This is an alternative approach to what John sent at
> https://lore.kernel.org/intel-xe/20250303195215.1046192-1-John.C.Harrison@Intel.com/
>
> Main thing is not making the name of the firmware magic, depending on
> part of starting with "g" or "m". So here I add the gt type in a
> pahole-friendly way and use it in the tables.
>
> Tested on BMG with the following table, just for tests:
>
> 	fw_def(PANTHERLAKE,     GT_TYPE_ANY,    mmp_ver(xe,     guc,    ptl,    70, 38, 1))     \
> 	fw_def(BATTLEMAGE,      GT_TYPE_MEDIA,  mmp_ver(xe,     guc,    bmg,    70, 40, 2))     \
> 	fw_def(BATTLEMAGE,      GT_TYPE_ANY,    major_ver(xe,   guc,    bmg,    70, 29, 2))     \
Longer lines, but otherwise seems like a better way of doing it.

PS: Sorry for taking so long on the review, lots going on lately!

> 	...
>
> And with that I get (pasting just the relevant parts:
>
> 	xe 0000:03:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version 70.36.0
> 	xe 0000:03:00.0: [drm:xe_guc_ct_enable [xe]] GT0: GuC CT communication channel enabled
> 	xe 0000:03:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.40.2.bin version 70.40.2
> 	xe 0000:03:00.0: [drm:xe_guc_ct_enable [xe]] GT1: GuC CT communication channel enabled
>
> Something that we could also discuss is getting rid of the macro
> indirections and just type the name of the file. Once upon a time we
> needed to use different dirs, but now this only matters per for the file
> name. The only repetition would be probably on the version. But I think
> it would make it clearer what files we expect to have without the macro
> indirections.
Not convinced about that plan. Having the filename auto-generated from 
specific fields ensures the format is correct. And duplicating the 
version number seems wasteful.

John.

>
> Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
> ---
> Lucas De Marchi (2):
>        drm/xe/uc: Remove static from loop variable
>        drm/xe/uc: Add support for different firmware files on each GT
>
>   drivers/gpu/drm/xe/xe_uc_fw.c | 102 ++++++++++++++++++++++++------------------
>   1 file changed, 59 insertions(+), 43 deletions(-)
> ---
> base-commit: 2c4fba1d065427efba4aaed5f80f4664c9f4a729
> change-id: 20250307-xe-per-gt-fw-4938dff9b339
>
> Best regards,


  parent reply	other threads:[~2025-03-20 20:09 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-03-07 18:13 [PATCH 0/2] drm/xe: Support different GuCs per GT Lucas De Marchi
2025-03-07 18:13 ` [PATCH 1/2] drm/xe/uc: Remove static from loop variable Lucas De Marchi
2025-03-20 20:08   ` John Harrison
2025-03-07 18:13 ` [PATCH 2/2] drm/xe/uc: Add support for different firmware files on each GT Lucas De Marchi
2025-03-20 20:09   ` John Harrison
2025-03-07 20:01 ` ✓ CI.Patch_applied: success for drm/xe: Support different GuCs per GT Patchwork
2025-03-07 20:01 ` ✗ CI.checkpatch: warning " Patchwork
2025-03-07 20:02 ` ✓ CI.KUnit: success " Patchwork
2025-03-07 20:19 ` ✓ CI.Build: " Patchwork
2025-03-07 20:21 ` ✓ CI.Hooks: " Patchwork
2025-03-07 20:23 ` ✓ CI.checksparse: " Patchwork
2025-03-07 20:45 ` ✓ Xe.CI.BAT: " Patchwork
2025-03-09  0:37 ` ✗ Xe.CI.Full: failure " Patchwork
2025-03-20 20:09 ` John Harrison [this message]
2025-03-21 18:42   ` [PATCH 0/2] " Lucas De 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=cc30ee16-f0bd-45de-b709-de35b0a70ef7@intel.com \
    --to=john.c.harrison@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=lucas.demarchi@intel.com \
    --cc=matthew.d.roper@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