All of lore.kernel.org
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Mika Kahola" <mika.kahola@intel.com>
Cc: intel-gfx@lists.freedesktop.org, intel-xe@lists.freedesktop.org
Subject: Re: [PATCH v2 6/6] drm/i915/display: Allow forcing the joiner on eDP
Date: Mon, 12 Oct 2026 12:52:13 +0000	[thread overview]
Message-ID: <sashiko-outbox-168129@kernel.org> (raw)
In-Reply-To: <20261012124145.1960050-7-mika.kahola@intel.com>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] drm/i915: data race on connector->force_joined_pipes via debugfs
--

commit 45be7b5e2e6fcbea23658dd271134587c4797f89
Author: Mika Kahola <mika.kahola@intel.com>

drm/i915/display: Allow forcing the joiner on eDP

intel_dp_has_joiner() rejects eDP unless the VBT enables the pipe
joiner, and the debugfs file is only created when that returns true, so
i915_joiner_force_enable can never turn it on.

Always create the file for DP and eDP, and let a forced pipe count
above one override the VBT.

> diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
> index d8c0d9381471d..ec0b4885e33ac 100644
> --- a/drivers/gpu/drm/i915/display/intel_dp.c
> +++ b/drivers/gpu/drm/i915/display/intel_dp.c

[ ... ]

> @@ -473,9 +473,15 @@ bool intel_dp_has_joiner(struct intel_dp *intel_dp)
>  	if (intel_dp->mso_link_count)
>  		return false;
>  
> -	if (intel_dp_is_edp(intel_dp) &&
> -	    !connector->panel.vbt.edp.pipe_joiner_enable)
> -		return false;
> +	if (intel_dp_is_edp(intel_dp)) {
> +		bool pipe_joiner_enable = connector->panel.vbt.edp.pipe_joiner_enable;
> +
> +		if (connector->force_joined_pipes > 1)

[Severity: Low]
Does this code introduce a data race with debugfs?

The debugfs handler i915_joiner_write() locklessly modifies
connector->force_joined_pipes:

drivers/gpu/drm/i915/display/intel_display_debugfs.c:i915_joiner_write() {
    ...
    connector->force_joined_pipes = force_joined_pipes;
    ...
}

Meanwhile, this read in intel_dp_has_joiner() during an atomic modeset
happens without holding any lock shared with the debugfs write path.

This problem wasn't introduced by this patch, but this commit expands a
similar pattern found in intel_dp_joiner_candidate_valid(), where a
double-read can cause a TOCTOU flaw if the compiler reloads the value
between reads:

drivers/gpu/drm/i915/display/intel_dp.c:intel_dp_joiner_candidate_valid() {
    ...
    if (connector->force_joined_pipes && connector->force_joined_pipes != num_joined_pipes)
        return false;
    ...
}

Could a lock or READ_ONCE() and WRITE_ONCE() annotations be added to
safely serialize or annotate these accesses?

> +			pipe_joiner_enable = true;
> +
> +		if (!pipe_joiner_enable)
> +			return false;
> +	}
>  
>  	return DISPLAY_VER(display) >= 12 ||
>  		(DISPLAY_VER(display) == 11 &&

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261012124145.1960050-1-mika.kahola@intel.com?part=6

  reply	other threads:[~2026-10-12 12:52 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-12 12:41 [PATCH v2 0/6] drm/i915/psr: Panel Replay and PSR2 with the joiner Mika Kahola
2026-10-12 12:41 ` [PATCH v2 1/6] drm/i915/psr: Track all joined pipes for frontbuffer tracking Mika Kahola
2026-10-12 12:41 ` [PATCH v2 2/6] drm/i915/psr: Exclude the joiner secondary from active_non_psr_pipes Mika Kahola
2026-10-12 12:41 ` [PATCH v2 3/6] drm/i915/psr: Take the selective fetch area from pipe_mode Mika Kahola
2026-10-12 12:41 ` [PATCH v2 4/6] drm/i915/psr: Merge the selective fetch area across joined pipes Mika Kahola
2026-10-12 12:56   ` sashiko-bot
2026-10-13 11:07   ` [PATCH v3 " Mika Kahola
2026-10-13 11:16     ` Mika Kahola
2026-10-12 12:41 ` [PATCH v2 5/6] drm/i915/psr: Allow Panel Replay and PSR2 with joiner Mika Kahola
2026-10-12 12:41 ` [PATCH v2 6/6] drm/i915/display: Allow forcing the joiner on eDP Mika Kahola
2026-10-12 12:52   ` sashiko-bot [this message]
2026-10-12 12:50 ` ✓ CI.KUnit: success for drm/i915/psr: Panel Replay and PSR2 with the joiner Patchwork
2026-10-12 14:28 ` ✓ Xe.CI.BAT: " Patchwork
2026-10-12 14:35 ` ✓ i915.CI.BAT: " Patchwork
2026-10-12 15:16 ` ✗ Xe.CI.FULL: failure " Patchwork
2026-10-12 17:25 ` ✗ i915.CI.Full: " Patchwork
2026-10-13 11:53 ` ✓ CI.KUnit: success for drm/i915/psr: Panel Replay and PSR2 with the joiner (rev3) Patchwork
2026-10-13 12:08 ` ✓ i915.CI.BAT: " Patchwork
2026-10-13 14:09 ` ✗ i915.CI.Full: failure " Patchwork
2026-10-13 16:21 ` ✓ Xe.CI.BAT: success " Patchwork

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-168129@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=mika.kahola@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 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.