linux-kernel - Re: [PATCH 0/8] drm: writeback: clean up writeback connector initialization

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <pwnrvebmf77p77z3pn53kzivy5t5qglh3ngdmj33f7i7v5r4jg@sseokeuqindn>
Date: Tue, 5 Aug 2025 07:58:04 +0300
From: Dmitry Baryshkov <dmitry.baryshkov@....qualcomm.com>
To: Liviu Dudau <liviu.dudau@....com>
Cc: Jani Nikula <jani.nikula@...ux.intel.com>,
        "Kandpal, Suraj" <suraj.kandpal@...el.com>,
        Harry Wentland <harry.wentland@....com>, Leo Li <sunpeng.li@....com>,
        Rodrigo Siqueira <siqueira@...lia.com>,
        Alex Deucher <alexander.deucher@....com>,
        Christian König <christian.koenig@....com>,
        David Airlie <airlied@...il.com>, Simona Vetter <simona@...ll.ch>,
        Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
        Maxime Ripard <mripard@...nel.org>,
        Thomas Zimmermann <tzimmermann@...e.de>,
        Rob Clark <robin.clark@....qualcomm.com>,
        Dmitry Baryshkov <lumag@...nel.org>,
        Abhinav Kumar <abhinav.kumar@...ux.dev>,
        Jessica Zhang <jessica.zhang@....qualcomm.com>,
        Sean Paul <sean@...rly.run>,
        Marijn Suijten <marijn.suijten@...ainline.org>,
        Laurent Pinchart <laurent.pinchart+renesas@...asonboard.com>,
        Tomi Valkeinen <tomi.valkeinen+renesas@...asonboard.com>,
        Kieran Bingham <kieran.bingham+renesas@...asonboard.com>,
        Geert Uytterhoeven <geert+renesas@...der.be>,
        Magnus Damm <magnus.damm@...il.com>,
        Dave Stevenson <dave.stevenson@...pberrypi.com>,
        Maíra Canal <mcanal@...lia.com>,
        Raspberry Pi Kernel Maintenance <kernel-list@...pberrypi.com>,
        amd-gfx@...ts.freedesktop.org, dri-devel@...ts.freedesktop.org,
        linux-kernel@...r.kernel.org, linux-arm-msm@...r.kernel.org,
        freedreno@...ts.freedesktop.org, linux-renesas-soc@...r.kernel.org
Subject: Re: [PATCH 0/8] drm: writeback: clean up writeback connector
 initialization

On Mon, Aug 04, 2025 at 03:43:12PM +0100, Liviu Dudau wrote:
> Hi,
> 
> On Fri, Aug 01, 2025 at 04:51:08PM +0300, Dmitry Baryshkov wrote:
> > Drivers using drm_writeback_connector_init() / _with_encoder() don't
> > perform cleanup in a manner similar to drmm_writeback_connector_init()
> > (see drm_writeback_connector_cleanup()). Migrate all existing drivers
> > to use drmm_writeback_connector_init(), drop
> > drm_writeback_connector_init() and drm_writeback_connector::encoder
> > (it's unused afterwards).
> > 
> > This series leaves former drm_writeback_connector_init_with_encoder()
> > (renamed to drm_writeback_connector_init as a non-managed counterpart
> > for drmm_writeback_connector_init()). It is supposed to be used by
> > drivers which can not use drmm functions (like Intel). However I think
> > it would be better to drop it completely.
> 
> The intent of _init_with_encoder() was to be a special case for drivers
> that use their own specific encoder and the rest use the generic function
> that creates the virtual encoder inside the call. The API for
> _init_with_encoder() was actually introduced 4 years after the original
> patch, so that should give a hint.
> 
> drmm_writeback_connector_init() is more like _init_with_encoder() and
> I don't remember reviewing it, so I'm not sure why that was considered
> to be the better behaviour for the managed version. Now you're moving
> all the drivers to the managed version and you have to duplicate code
> in each driver to create the ENCODER_VIRTUAL encoder.

This follows e.g. the process of deprecating drm_simple_* /
drm_simple_encoder. The drivers are expected to open code empty encoder
handling on their own.

> I'm not against the changes being made in the series, I just want to
> see a better justification on why _init_with_encoder() behaviour is
> better than the previous default that you're removing.

This was triggered by the discussion of Intel writeback patchset, see
the threads for first three patches of [1]. We have an optional
non-pointer field inside drm_writeback_connector, which can be left
uninitialized (or zero-filled). I have checked and the drivers are not
actually using the embedded connector for anything after linking it to
the drm_connector. So, by removing the encoder from the
drm_writeback_connector structure we are tying a loose end.

[1] https://lore.kernel.org/dri-devel/20250725050409.2687242-1-suraj.kandpal@intel.com/

-- 
With best wishes
Dmitry