linux-kernel - Re: [PATCH v4 27/27] drm/panel: panel-simple: Prepare/unprepare are refcounted, not forced

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 for Android: free password hash cracker in your pocket

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

Date:   Fri, 23 Apr 2021 09:46:57 -0700
From:   Doug Anderson <dianders@...omium.org>
To:     Andrzej Hajda <a.hajda@...sung.com>,
        Neil Armstrong <narmstrong@...libre.com>,
        Laurent Pinchart <Laurent.pinchart@...asonboard.com>,
        Jonas Karlman <jonas@...boo.se>,
        Jernej Skrabec <jernej.skrabec@...l.net>,
        Sam Ravnborg <sam@...nborg.org>, Wolfram Sang <wsa@...nel.org>
Cc:     Stephen Boyd <swboyd@...omium.org>,
        Rob Clark <robdclark@...omium.org>,
        Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
        Stanislav Lisovskiy <stanislav.lisovskiy@...el.com>,
        Steev Klimaszewski <steev@...i.org>,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        linux-arm-msm <linux-arm-msm@...r.kernel.org>,
        Linus W <linus.walleij@...aro.org>,
        Daniel Vetter <daniel@...ll.ch>,
        David Airlie <airlied@...ux.ie>,
        Thierry Reding <thierry.reding@...il.com>,
        dri-devel <dri-devel@...ts.freedesktop.org>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v4 27/27] drm/panel: panel-simple: Prepare/unprepare are
 refcounted, not forced

Hi,

On Fri, Apr 16, 2021 at 3:41 PM Douglas Anderson <dianders@...omium.org> wrote:
>
> Historically simple-panel had code to make sure that if prepare() was
> called on an already-prepared panel that it was a no-op. Similarly if
> unprepare() was called on an already-unprepared panel it was also a
> no-op. Essentially it means that these functions always "forced" the
> value to be whatever the caller wanted it to be. You can see that the
> forcing behavior dates back at least as far as 2014 by looking at
> commit 613a633e7a56 ("drm/panel: simple: Add proper definition for
> prepare and unprepare").
>
> Apparently the code supporting the historical behavior may not be
> needed [1] and prepare() / unprepare() are supposed to be
> balanced. Let's try removing it and see if anyone breaks! If they do
> then we can have a debate about whether we should change that code or
> revert this patch. :-) If nobody breaks then we've nicely saved a few
> lines of code and some complexity.
>
> [1] https://lore.kernel.org/r/YHePsQgqOau1V5lD@pendragon.ideasonboard.com
>
> Suggested-by: Laurent Pinchart <laurent.pinchart@...asonboard.com>
> Signed-off-by: Douglas Anderson <dianders@...omium.org>
> ---
>
> (no changes since v1)
>
>  drivers/gpu/drm/panel/panel-simple.c | 14 --------------
>  1 file changed, 14 deletions(-)

So it turns out that when looking at panel_simple_remove() I already
found a case that's not necessarily refcounting. There I see
drm_panel_unprepare() just simply hardcoded, but (as I understand it)
there's no reason to believe that the panel has been prepared at
remove() time. Yes, I could fix panel_simple_remove() but instead, for
now, I think I'm going to drop this patch from the series. If someone
wants to pick it up then I certainly won't object, but for now keeping
the way things are seems the best way to keep from getting shouted at.

-Doug