From mboxrd@z Thu Jan 1 00:00:00 1970
From: bugzilla-daemon@freedesktop.org
Subject: [Bug 91278] Tonga GPU lock/reset fail with Unigine Valley
Date: Sun, 11 Oct 2015 16:14:11 +0000
Message-ID:
References:
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="===============1713792135=="
Return-path:
Received: from culpepper.freedesktop.org (unknown [131.252.210.165])
by gabe.freedesktop.org (Postfix) with ESMTP id 0FF8C6E1F5
for ; Sun, 11 Oct 2015 09:14:12 -0700 (PDT)
In-Reply-To:
List-Unsubscribe: ,
List-Archive:
List-Post:
List-Help:
List-Subscribe: ,
Errors-To: dri-devel-bounces@lists.freedesktop.org
Sender: "dri-devel"
To: dri-devel@lists.freedesktop.org
List-Id: dri-devel@lists.freedesktop.org
--===============1713792135==
Content-Type: multipart/alternative; boundary="1444580051.dbcbFb1a1.4150"; charset="UTF-8"
--1444580051.dbcbFb1a1.4150
Date: Sun, 11 Oct 2015 16:14:11 +0000
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
https://bugs.freedesktop.org/show_bug.cgi?id=3D91278
Grazvydas Ignotas changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |notasas@gmail.com
--- Comment #34 from Grazvydas Ignotas ---
Created attachment 118824
--> https://bugs.freedesktop.org/attachment.cgi?id=3D118824&action=3Dedit
test kernel patch
(In reply to Michel D=C3=A4nzer from comment #29)
> That is interesting, though; the radeonsi driver seems to think there sho=
uld
> be something mapped at the faulting address. This indicates that either t=
he
> kernel driver fails to handle the mapping properly, or maybe there's a
> problem with communicating the buffer mapping information from userspace =
to
> the kernel driver.
Judging by the symptoms it feels like some caching/buffering problem somewh=
ere.=20
If I understand the code right, most of things are mapped write-combine, wh=
ich
means the CPU is allowed to write data it any order it likes. Looking at
amdgpu/radeon code, there is surprising lack of barriers, basically it's ju=
st
amdgpu_ring_commit()/radeon_ring_commit() and that's it. But mb() doesn't
guarantee that the writes will arrive in program order, it just ensures that
all the writes are finished after that mb() statement.
So the question is, is it ok for the hardware if in something like
amdgpu_ib_schedule() the writes to the ring arrive before the writes to IB?=
I
do admit I don't understand how the hardware works, like what triggers the
hardware to start processing the ring contents, perhaps the write to the la=
st
word in the ring? If so you clearly need a wmb() before the write which
triggers the hardware so that everything is ready before the GPU kicks in.
Attached is a debug kernel patch to test if my guess is correct. It's way
overkill and will trash performance, but it should show if this is a problem
related to CPU caching/buffering. I don't have the hardware to test this
myself.
--=20
You are receiving this mail because:
You are the assignee for the bug.
--1444580051.dbcbFb1a1.4150
Date: Sun, 11 Oct 2015 16:14:11 +0000
MIME-Version: 1.0
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Grazvydas Ignotas
changed
bug 91278<=
/a>
What |
Removed |
Added |
CC |
|
notasas@gmail.com
|
Commen=
t # 34
on bug 91278<=
/a>
from Grazvydas Ignotas
Created attachment 118824 =
[details] [review]
test kernel patch
(In reply to Michel D=C3=A4nzer from comment #29)
> That is interesting, though; the radeonsi driver=
seems to think there should
> be something mapped at the faulting address. This indicates that eithe=
r the
> kernel driver fails to handle the mapping properly, or maybe there's a
> problem with communicating the buffer mapping information from userspa=
ce to
> the kernel driver.
Judging by the symptoms it feels like some caching/buffering problem somewh=
ere.=20
If I understand the code right, most of things are mapped write-combine, wh=
ich
means the CPU is allowed to write data it any order it likes. Looking at
amdgpu/radeon code, there is surprising lack of barriers, basically it's ju=
st
amdgpu_ring_commit()/radeon_ring_commit() and that's it. But mb() doesn't
guarantee that the writes will arrive in program order, it just ensures that
all the writes are finished after that mb() statement.
So the question is, is it ok for the hardware if in something like
amdgpu_ib_schedule() the writes to the ring arrive before the writes to IB?=
I
do admit I don't understand how the hardware works, like what triggers the
hardware to start processing the ring contents, perhaps the write to the la=
st
word in the ring? If so you clearly need a wmb() before the write which
triggers the hardware so that everything is ready before the GPU kicks in.
Attached is a debug kernel patch to test if my guess is correct. It's way
overkill and will trash performance, but it should show if this is a problem
related to CPU caching/buffering. I don't have the hardware to test this
myself.
You are receiving this mail because:
=20=20=20=20=20=20
- You are the assignee for the bug.
--1444580051.dbcbFb1a1.4150--
--===============1713792135==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: base64
Content-Disposition: inline
X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KZHJpLWRldmVs
IG1haWxpbmcgbGlzdApkcmktZGV2ZWxAbGlzdHMuZnJlZWRlc2t0b3Aub3JnCmh0dHA6Ly9saXN0
cy5mcmVlZGVza3RvcC5vcmcvbWFpbG1hbi9saXN0aW5mby9kcmktZGV2ZWwK
--===============1713792135==--