From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
To: "Christian König" <deathsimple@vodafone.de>,
"Jerome Glisse" <j.glisse@gmail.com>,
"Daniel Vetter" <daniel@ffwll.ch>
Cc: daniel.vetter@ffwll.ch, dri-devel@lists.freedesktop.org,
bskeggs@redhat.com
Subject: Re: Fence, timeline and android sync points
Date: Thu, 14 Aug 2014 14:37:15 +0200 [thread overview]
Message-ID: <53ECAD7B.8040304@canonical.com> (raw)
In-Reply-To: <53ECA34B.6030501@vodafone.de>
Op 14-08-14 om 13:53 schreef Christian König:
>> But because of driver differences I can't implement it as a straight wait queue. Some drivers may not have a reliable interrupt, so they need a custom wait function. (qxl)
>> Some may need to do extra flushing to get fences signaled (vmwgfx), others need some locking to protect against gpu lockup races (radeon, i915??). And nouveau
>> doesn't use wait queues, but rolls its own (nouveau).
> But when all those drivers need a special wait function how can you still justify the common callback when a fence is signaled?
>
> If I understood it right the use case for this was waiting for any fence of a list of fences from multiple drivers, but if each driver needs special handling how for it's wait how can that work reliable?
TTM doesn't rely on the callbacks. It will call .enable_signaling when .is_signaled is NULL, to make sure that fence_is_signaled returns true sooner.
QXL is completely awful, I've seen some patches to add dma-buf support but I'll make sure it never supports importing from/exporting to other devices. This should reduce insanity factor there.
If I understand QXL correctly, sometimes fences may never signal at all due to virt-hw bugs.
nouveau (pre nv84) has no interrupt for completed work, but it has a reliable is_signaled. So .enable_signaling only checks if fence is signaled here.
A custom waiting function makes sure things work correctly, and also signals all unsignaled fences for that context. I preserved the original wait from before the fence conversion.
Nouveau keeps a global list of unsignaled fences, so they will all signal eventually.
I may have to limit importing/exporting dma-buf's to other devices, or add delayed work that periodically checks all contexts for completed fences for this to work cross-device.
nv84+ has a sane interrupt, so I use it. :-)
Radeon with fence conversion has the delayed work for handling lockups that also checks.
~Maarten
next prev parent reply other threads:[~2014-08-14 12:37 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-08-12 22:13 Fence, timeline and android sync points Jerome Glisse
2014-08-13 1:23 ` Jerome Glisse
2014-08-13 7:59 ` Christian König
2014-08-13 13:41 ` Jerome Glisse
2014-08-13 14:08 ` Christian König
2014-08-13 15:56 ` Jerome Glisse
2014-08-13 8:28 ` Daniel Vetter
2014-08-13 13:36 ` Jerome Glisse
2014-08-13 15:54 ` Daniel Vetter
2014-08-13 17:07 ` Jerome Glisse
2014-08-14 9:08 ` Daniel Vetter
2014-08-14 14:23 ` Jerome Glisse
2014-08-14 15:55 ` Daniel Vetter
2014-08-14 18:18 ` Jerome Glisse
2014-08-14 18:47 ` Daniel Vetter
2014-08-14 19:15 ` Jerome Glisse
2014-08-14 19:40 ` Maarten Lankhorst
2014-08-14 19:56 ` Jerome Glisse
2014-08-14 21:20 ` Daniel Vetter
2014-08-14 21:23 ` Daniel Vetter
2014-08-14 23:03 ` Jerome Glisse
2014-08-15 8:07 ` Daniel Vetter
2014-08-15 14:53 ` Jerome Glisse
2014-08-14 21:30 ` Daniel Vetter
2014-08-15 6:54 ` Thomas Hellstrom
2014-08-15 14:52 ` Jerome Glisse
2014-08-16 7:01 ` Thomas Hellstrom
2014-08-16 15:30 ` Jerome Glisse
2014-08-14 9:15 ` Maarten Lankhorst
2014-08-14 11:53 ` Christian König
2014-08-14 12:37 ` Maarten Lankhorst [this message]
2014-08-14 14:31 ` Christian König
2014-08-14 14:09 ` Jerome Glisse
2014-08-14 13:16 ` Rob Clark
2014-08-14 14:12 ` Jerome Glisse
2014-08-14 15:58 ` Daniel Vetter
2014-08-14 18:26 ` Jerome Glisse
2014-08-14 19:16 ` Maarten Lankhorst
2014-08-14 22:11 ` Rob Clark
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=53ECAD7B.8040304@canonical.com \
--to=maarten.lankhorst@canonical.com \
--cc=bskeggs@redhat.com \
--cc=daniel.vetter@ffwll.ch \
--cc=daniel@ffwll.ch \
--cc=deathsimple@vodafone.de \
--cc=dri-devel@lists.freedesktop.org \
--cc=j.glisse@gmail.com \
/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.