Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.32 as permitted sender) client-ip=23.128.96.32;
Date:   Fri, 17 Nov 2023 16:20:12 +0000
To:     =?utf-8?Q?Andr=C3=A9_Almeida?= <andrealmeid@igalia.com>
From:   Simon Ser <contact@emersion.fr>
Cc:     dri-devel@lists.freedesktop.org, amd-gfx@lists.freedesktop.org,
        linux-kernel@vger.kernel.org, kernel-dev@igalia.com,
        alexander.deucher@amd.com, christian.koenig@amd.com,
        pierre-eric.pelloux-prayer@amd.com,
        Rob Clark <robdclark@gmail.com>,
        Pekka Paalanen <ppaalanen@gmail.com>,
        Daniel Vetter <daniel@ffwll.ch>,
        Daniel Stone <daniel@fooishbar.org>,
        =?utf-8?Q?=27Marek_Ol=C5=A1=C3=A1k=27?= <maraeo@gmail.com>,
        Dave Airlie <airlied@gmail.com>,
        =?utf-8?Q?Michel_D=C3=A4nzer?= <michel.daenzer@mailbox.org>,
        Randy Dunlap <rdunlap@infradead.org>
Subject: Re: [PATCH v8 0/6] drm: Add support for atomic async page-flip
Message-ID: <Xvnc1dvx8ij7QQjMrEG5AE_wpnGxYwEwk0PxRir17B34AM8y9HZTorD18qGzm-mMrgq9svucDnTS2d-8VyPT44jHuOLpdTavQV5Rgjz-dYc=@emersion.fr>
In-Reply-To: <20231025005318.293690-1-andrealmeid@igalia.com>
References: <20231025005318.293690-1-andrealmeid@igalia.com>
Feedback-ID: 1358184:user:proton
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

It seems like commits were re-ordered at some point. I think we need "drm:
introduce drm_mode_config.atomic_async_page_flip_not_supported" to come bef=
ore
"drm: allow DRM_MODE_PAGE_FLIP_ASYNC for atomic commits" because the latter
uses atomic_async_page_flip_not_supported. Similarly, "drm: Refuse to async=
 flip
with atomic prop changes" should probably come before that same patch becau=
se
we don't want to have an intermediary state where we allow user-space to ch=
ange
arbitrary properties. In general, commits should be constructed so that the=
re
is no "broken state" in-between: each commit needs to build without errors =
and
not have transient bugs.

While reading this again, I think that now that we only allow primary FB_ID=
 to
change, we don't need atomic_async_page_flip_not_supported anymore? In othe=
r
words, allowing async atomic commits on all drivers doesn't require anythin=
g
more than they support already, because the atomic codepath can't do anythi=
ng
more than the legacy codepath.

With that in mind, I also wonder if the new cap is helpful. Since async ato=
mic
commits can fail for pretty much any reason, user-space can just try and
fallback to something else. The cap could be useful to indicate whether asy=
nc
atomic commits would always fail, but not sure how useful that is? I don't =
have
strong opinions either way.