Re: [PATCH v5 03/21] drm: bridge: dw_hdmi: Call poweron/poweroff from atomic enable/disable

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The commit message incorrectly assumes that the `drm mode_config mutex` is always held during `atomic_enable` and `atomic_disable`.
--

commit dbe86704cd583592fb999cc80ea4e4d22f4f6084
Author: Jonas Karlman <jonas@kwiboo.se>

drm: bridge: dw_hdmi: Call poweron/poweroff from atomic enable/disable

This commit moves the poweron and poweroff calls directly into the
atomic_enable and atomic_disable bridge callbacks. It removes the
bridge_is_on state tracking previously handled in hotplug IRQ paths.

> A benefit of this is that drm mode_config mutex is always held at
> poweron/off, something that may reduce the need for the dw-hdmi mutex.

Is the statement about the mode_config mutex being held accurate here?

In the DRM atomic framework, bridge atomic_enable and atomic_disable
callbacks are executed from commit_tail. For non-blocking atomic commits,
commit_tail runs asynchronously in a workqueue after the IOCTL handler has
already dropped all modeset locks, including the mode_config mutex.

Because of this, no global DRM modeset locks are guaranteed to be held
during these sequences. 

While this patch appears safe since it still explicitly acquires hdmi->mutex
in both callbacks, could relying on the mode_config mutex in future changes
introduce data races when interacting with concurrent IRQ handlers?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260510124111.1226584-1-jonas@kwiboo.se?part=3