From: sashiko-bot@kernel.org
To: "Dave Airlie" <airlied@redhat.com>
Cc: amd-gfx@lists.freedesktop.org, dri-devel@lists.freedesktop.org
Subject: Re: [PATCH 09/19] drm/amd/display/amdgpu_dm: Use DC accessors in debugfs paths
Date: Thu, 12 Nov 2026 01:53:22 +0000 [thread overview]
Message-ID: <sashiko-outbox-210110@kernel.org> (raw)
In-Reply-To: <20261112-amdgpu-dc-cleanup-v1-9-0fae41718ad2@redhat.com>
Thank you for your contribution! Sashiko AI review found no new issues
in this patch, but noticed 1 related pre-existing issue it does not
introduce:
- [Critical] drm/amd/display: use-after-free in dp_dsc_clock_en_write()
[1]
--
commit 8d9d786f748dd7764b2335ea8779ba852b1d97b2
Author: Dave Airlie <airlied@redhat.com>
drm/amd/display/amdgpu_dm: Use DC accessors in debugfs paths
This commit replaces direct struct dc member dereferences in
amdgpu_dm_debugfs.c with DC accessor functions and adds a setter for the
eDP hotplug detection mode.
> diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
> index 68356cb8e4dec..fddb47a32fb9e 100644
> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
> +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
[ ... ]
> @@ -1696,7 +1696,7 @@ static ssize_t dp_dsc_clock_en_write(struct file *f, const char __user *buf,
> }
>
> for (i = 0; i < MAX_PIPES; i++) {
> - pipe_ctx = &aconnector->dc_link->dc->current_state->res_ctx.pipe_ctx[i];
> + pipe_ctx = &dc_get_current_state(aconnector->dc_link->dc)->res_ctx.pipe_ctx[i];
> if (pipe_ctx->stream &&
> pipe_ctx->stream->link == aconnector->dc_link &&
> pipe_ctx->stream->sink &&
[Severity: Critical]
This problem wasn't introduced by this patch, but can a concurrent atomic
commit free current_state while dp_dsc_clock_en_write() iterates over
res_ctx.pipe_ctx[] [1]?
dp_dsc_clock_en_write() walks dc_get_current_state()->res_ctx.pipe_ctx[]
before taking dev->mode_config.mutex and dev->mode_config.connection_mutex:
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c:dp_dsc_clock_en_write() {
...
for (i = 0; i < MAX_PIPES; i++) {
pipe_ctx = &dc_get_current_state(aconnector->dc_link->dc)->res_ctx.pipe_ctx[i];
if (pipe_ctx->stream &&
...
break;
}
if (!pipe_ctx->stream)
goto done;
// Get CRTC state
mutex_lock(&dev->mode_config.mutex);
drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
...
}
CPU0 CPU1
dp_dsc_clock_en_write()
state = dc_get_current_state(dc)
amdgpu_dm_atomic_commit_tail()
...
dc_state_release(old_state)
pipe_ctx = &state->res_ctx.pipe_ctx[i]
if (pipe_ctx->stream && ...) <-- use-after-free
If an atomic modeset commit swaps out and frees current_state during the
unlocked loop, dp_dsc_clock_en_write() dereferences freed memory.
[1] https://sashiko.dev/#/bug/linux-3a852cda-9b25-49c2-b119-47f2a71a94c3
--
Sashiko AI review · https://sashiko.dev/#/patchset/20261112-amdgpu-dc-cleanup-v1-0-0fae41718ad2@redhat.com?part=9
next prev parent reply other threads:[~2026-11-12 1:53 UTC|newest]
Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-12 1:43 [PATCH 00/19] drm/amd/display: refactor DC interfaces to abstract it better Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:43 ` [PATCH 01/19] drm/amd/display: Drop the dead vmid module Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:43 ` [PATCH 02/19] drm/amd/display: Add DC accessors as the first step to opaque struct dc Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:43 ` [PATCH 03/19] drm/amd/display: Take struct dc in public DSC compute helpers Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:43 ` [PATCH 04/19] drm/amd/display/amdgpu_dm: Use DC accessors in audio/color/backlight paths Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:50 ` sashiko-bot
2026-11-12 1:43 ` [PATCH 05/19] drm/amd/display/amdgpu_dm: Use DC accessors in psr/replay/dmub/hdcp/irq paths Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 2:07 ` sashiko-bot
2026-11-12 1:43 ` [PATCH 06/19] drm/amd/display/amdgpu_dm: Use DC accessors in plane/crtc/helpers paths Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:58 ` sashiko-bot
2026-11-12 1:43 ` [PATCH 07/19] drm/amd/display/amdgpu_dm: Use DC accessors in connector paths Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:43 ` [PATCH 08/19] drm/amd/display: Use DC accessors in DAL modules Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:43 ` [PATCH 09/19] drm/amd/display/amdgpu_dm: Use DC accessors in debugfs paths Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:53 ` sashiko-bot [this message]
2026-11-12 1:43 ` [PATCH 10/19] drm/amd/display/amdgpu_dm: Use DC accessors in the DM core paths Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:43 ` [PATCH 11/19] drm/amd/display/amdgpu_dm: Finish DC accessor migration for non-test code Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:53 ` sashiko-bot
2026-11-12 1:43 ` [PATCH 12/19] drm/amd/display: Migrate amdgpu_dm KUnit tests off struct dc internals Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:59 ` sashiko-bot
2026-11-12 1:43 ` [PATCH 13/19] drm/amd/display: Move struct dc out of the public dc.h Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:43 ` [PATCH 14/19] drm/amd/display: Restrict the dc directory root to DC-internal builds Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 2:08 ` sashiko-bot
2026-11-12 1:43 ` [PATCH 15/19] drm/amd/display: Add dc_force_max_dramclk for the DP compliance workaround Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 2:02 ` sashiko-bot
2026-11-12 1:43 ` [PATCH 16/19] drm/amd/display: Add DC primitives for ABM hardware programming Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 2:01 ` sashiko-bot
2026-11-12 1:43 ` [PATCH 17/19] drm/amd/display: Replace the dc_get_res_pool accessor with purpose APIs Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 1:43 ` [PATCH 18/19] drm/amd/display: Replace dc_get_config with per-option config reads Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 20:31 ` Harry Wentland
2026-11-12 1:43 ` [PATCH 19/19] drm/amd/display: Replace dc_get_caps with purpose-built capability queries Dave Airlie
2026-11-12 1:43 ` Dave Airlie via B4 Relay
2026-11-12 2:14 ` sashiko-bot
2026-11-12 19:41 ` Harry Wentland
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-210110@kernel.org \
--to=sashiko-bot@kernel.org \
--cc=airlied@redhat.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=dri-devel@lists.freedesktop.org \
--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.