<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix">Am 07.09.2017 um 12:14 schrieb Marek
Olšák:<br>
</div>
<blockquote type="cite"
cite="mid:CAAxE2A7NoHzmO=0B1JO-Yis_kNZFabWWQbCmibTe5mwgRoOxGQ@mail.gmail.com">
<div dir="auto">
<div><br>
<div class="gmail_extra"><br>
<div class="gmail_quote">On Sep 7, 2017 12:08 PM, "Christian
König" <<a href="mailto:deathsimple@vodafone.de"
moz-do-not-send="true">deathsimple@vodafone.de</a>>
wrote:<br type="attribution">
<blockquote class="quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="elided-text">Am 07.09.2017 um 11:23 schrieb
Michel Dänzer:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
On 01/09/17 07:40 PM, Christian König wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
Am 01.09.2017 um 12:28 schrieb Michel Dänzer:<br>
<blockquote class="gmail_quote" style="margin:0 0
0 .8ex;border-left:1px #ccc
solid;padding-left:1ex">
On 01/09/17 07:23 PM, Nicolai Hähnle wrote:<br>
<blockquote class="gmail_quote" style="margin:0
0 0 .8ex;border-left:1px #ccc
solid;padding-left:1ex">
On 01.09.2017 11:58, Michel Dänzer wrote:<br>
<blockquote class="gmail_quote"
style="margin:0 0 0 .8ex;border-left:1px
#ccc solid;padding-left:1ex">
On 29/08/17 11:47 PM, Christian König wrote:<br>
<blockquote class="gmail_quote"
style="margin:0 0 0 .8ex;border-left:1px
#ccc solid;padding-left:1ex">
From: Marek Olšák <<a
href="mailto:marek.olsak@amd.com"
target="_blank" moz-do-not-send="true">marek.olsak@amd.com</a>><br>
<br>
For lower overhead in the CS ioctl.<br>
Winsys allocators are not used with
interprocess-sharable resources.<br>
<br>
v2: It shouldn't crash anymore, but the
kernel will reject the new<br>
flag.<br>
v3 (christian): Rename the flag, avoid
sending those buffers in the<br>
BO list.<br>
v4 (christian): Remove setting the kernel
flag for now<br>
</blockquote>
This change seems to have caused a GPU hang
when running piglit on my<br>
Kaveri with the radeon kernel driver.<br>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
I think we can remove "seems to have". I'm still
reliably getting the<br>
GPUVM fault and hang with current master, but not if
I revert this<br>
commit (and the one after it).<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
<blockquote class="gmail_quote" style="margin:0 0
0 .8ex;border-left:1px #ccc
solid;padding-left:1ex">
<blockquote class="gmail_quote" style="margin:0
0 0 .8ex;border-left:1px #ccc
solid;padding-left:1ex">
<blockquote class="gmail_quote"
style="margin:0 0 0 .8ex;border-left:1px
#ccc solid;padding-left:1ex">
Haven't been able to isolate it to a
specific test, seems to only<br>
happen when running multiple tests
concurrently.<br>
</blockquote>
</blockquote>
</blockquote>
</blockquote>
I reproduced the problem with piglit process
separation enabled as well,<br>
and all four tests running when it hung were
textureGather tests.<br>
Before, reproducing the problem twice with piglit
process separation<br>
disabled, three textureGather tests were running
when it hung both times<br>
as well. I've been unable to reproduce the problem
by manually running<br>
the same textureGather tests in parallel though.<br>
<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
<blockquote class="gmail_quote" style="margin:0 0
0 .8ex;border-left:1px #ccc
solid;padding-left:1ex">
<blockquote class="gmail_quote" style="margin:0
0 0 .8ex;border-left:1px #ccc
solid;padding-left:1ex">
<blockquote class="gmail_quote"
style="margin:0 0 0 .8ex;border-left:1px
#ccc solid;padding-left:1ex">
There's a GPUVM fault before the hang, I
suspect it's related:<br>
<br>
radeon 0000:00:01.0: GPU fault detected: 146
0x0ae6760c<br>
radeon 0000:00:01.0:
VM_CONTEXT1_PROTECTION_FAULT_<wbr>ADDR
0x000001D7<br>
radeon 0000:00:01.0:
VM_CONTEXT1_PROTECTION_FAULT_<wbr>STATUS
0x0607600C<br>
VM fault (0x0c, vmid 3) at page 471, read
from 'CPF' (0x43504600) (118)<br>
<br>
<br>
Any ideas?<br>
</blockquote>
</blockquote>
</blockquote>
Not the slightest, but I'm still investigating
problems with that on<br>
amdgpu.<br>
<br>
If we can't find the root cause till Monday it
might be a good idea to<br>
revert the patches for now.<br>
</blockquote>
What's the status on that?<br>
</blockquote>
<br>
<br>
</div>
I've found and fixed the remaining kernel bugs over the
last weekend/beginning of this week.<br>
<br>
Still need to commit the fix for UVD/VCE, but that one
shouldn't affect GFX at all.</blockquote>
</div>
</div>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">Michel is seeing hangs on the radeon KMD, which
should be unaffected by you kernel work I think.</div>
<div dir="auto"><br>
</div>
<div dir="auto">We could revert this to unbreak Michel's Kaveri,
but I think it shouldn't be so difficult to find the culprit
in this patch if there is one.</div>
</div>
</blockquote>
<br>
The only crux is that the userspace patch shouldn't affect radeon at
all. So the real question is what the heck is going on here?<br>
<br>
Christian.<br>
<br>
<blockquote type="cite"
cite="mid:CAAxE2A7NoHzmO=0B1JO-Yis_kNZFabWWQbCmibTe5mwgRoOxGQ@mail.gmail.com">
<div dir="auto">
<div dir="auto"><br>
</div>
<div dir="auto">Marek</div>
<div dir="auto"><br>
</div>
<div dir="auto">
<div class="gmail_extra">
<div class="gmail_quote">
<blockquote class="quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex"><font
color="#888888"><br>
<br>
Christian.<br>
<br>
</font></blockquote>
</div>
<br>
</div>
</div>
</div>
</blockquote>
<p><br>
</p>
</body>
</html>