Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Julia Filipchuk" <julia.filipchuk@intel.com>,
	intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH] drm/i915/guc: Test GuC v70.83.0 for DG2, MTL
Date: Mon, 02 Nov 2026 22:22:16 +0000	[thread overview]
Message-ID: <sashiko-outbox-194275@kernel.org> (raw)
In-Reply-To: <20261102221518.546010-2-julia.filipchuk@intel.com>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] drm/i915/guc: guc_mmp() replacement breaks major-versioned firmware on MTL/DG2
--

commit 8a0f11765b445043e0614bba8e7df77aba661014
Author: Julia Filipchuk <julia.filipchuk@intel.com>

drm/i915/guc: Test GuC v70.83.0 for DG2, MTL

This commit updates the Meteor Lake and DG2 GuC firmware definitions in
intel_uc_fw.c from v70.53.0 using guc_maj() to v70.83.0 using guc_mmp() for
testing.

> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
> index d7c731fe972bf..c89a1b97108c0 100644
> --- a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
> +++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
> @@ -89,8 +89,8 @@ void intel_uc_fw_change_status(struct intel_uc_fw *uc_fw,
>   * security fixes, etc. to be enabled.
>   */
>  #define INTEL_GUC_FIRMWARE_DEFS(fw_def, guc_maj, guc_mmp) \
> -	fw_def(METEORLAKE,   0, guc_maj(mtl,  70, 53, 0)) \
> -	fw_def(DG2,          0, guc_maj(dg2,  70, 53, 0)) \
> +	fw_def(METEORLAKE,   0, guc_mmp(mtl,  70, 83, 0)) \
> +	fw_def(DG2,          0, guc_mmp(dg2,  70, 83, 0)) \

[Severity: High]
Does replacing guc_maj() with guc_mmp() here break GuC firmware loading on
Meteor Lake and DG2 systems that have standard linux-firmware installed?

With guc_mmp(), blobs_guc[] and the module firmware table only list
i915/mtl_guc_70.83.0.bin and i915/dg2_guc_70.83.0.bin, removing the
major-versioned i915/mtl_guc_70.bin and i915/dg2_guc_70.bin entries with no
fallback.

When intel_uc_fw_fetch() calls try_firmware_load() and those three-part
versioned test files are absent, the fallback loop in intel_uc_fw_fetch()
calls __uc_fw_auto_select():

drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c:intel_uc_fw_fetch() {
	...
	while (err == -ENOENT) {
		old_ver = true;

		__uc_fw_auto_select(i915, uc_fw);
		if (!uc_fw->file_selected.path) {
			...
}

Because no other blobs_guc[] entry exists for Meteor Lake or DG2,
__uc_fw_auto_select() leaves file_selected.path unset and firmware fetch
fails. Since GuC submission is enabled by default on both platforms,
__uc_init_hw() then aborts hardware initialization:

drivers/gpu/drm/i915/gt/uc/intel_uc.c:__uc_init_hw() {
	...
	if (!intel_uc_fw_is_loadable(&guc->fw)) {
		ret = __uc_check_hw(uc) ||
		      intel_uc_fw_is_overridden(&guc->fw) ||
		      intel_uc_wants_guc_submission(uc) ?
		      intel_uc_fw_status_to_error(guc->fw.status) : 0;
		goto err_out;
	}
	...
}

Should these entries remain guc_maj(mtl, 70, 83, 0) and
guc_maj(dg2, 70, 83, 0) (or retain the guc_maj() entries as fallbacks)
before this is merged beyond testing?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261102221518.546010-2-julia.filipchuk@intel.com?part=1

  reply	other threads:[~2026-11-02 22:22 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-02 22:15 [PATCH] drm/i915/guc: Test GuC v70.83.0 for DG2, MTL Julia Filipchuk
2026-11-02 22:22 ` sashiko-bot [this message]
2026-11-02 23:09 ` ✓ i915.CI.BAT: success for " Patchwork
2026-11-03  0:13 ` ✓ i915.CI.Full: " Patchwork
2026-11-03 19:37 ` ✓ i915.CI.BAT: success for drm/i915/guc: Test GuC v70.83.0 for DG2, MTL (rev2) Patchwork
2026-11-04 19:17   ` Julia Filipchuk
2026-11-03 23:43 ` ✗ i915.CI.Full: failure " Patchwork
2026-11-04 19:34   ` Julia Filipchuk

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=sashiko-outbox-194275@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=julia.filipchuk@intel.com \
    --cc=sashiko-reviews@lists.linux.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox