Debian Bug report logs - #860805
beignet-opencl-icd: OpenCL fails with: drm_intel_gem_bo_context_exec() failed: Device or resource busy

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: "Michael Gernoth" <michael@zerfleddert.de>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: beignet-opencl-icd: OpenCL fails with: drm_intel_gem_bo_context_exec() failed: Device or resource busy

Date: Thu, 20 Apr 2017 12:34:35 +0200

Package: beignet-opencl-icd
Version: 1.3.0-2+b1
Severity: normal

Dear Maintainer,

running any opencl-using program (such as clinfo or osmocom_fft -F)
does not work but produces:

drm_intel_gem_bo_context_exec() failed: Device or resource busy
Beignet: "Exec event 0x55bc572926b0 error, type is 4592, error staus is -5"

This is on a Lenovo Thinkpad T440s with an Intel HD Graphics 4400.

Rebuilding beignet-opencl-icd with the following upstream-patch fixes
the problem and leads to working OpenCL:

https://cgit.freedesktop.org/beignet/commit/?id=8b04f0be372da8eabdc93d6ae1b81a3c83cba284

Cheers,
  Michael



-- Package-specific info:
Graphics hardware:
Providers: number : 1
Provider 0: id: 0x49 cap: 0xb, Source Output, Sink Output, Sink Offload crtcs: 4 outputs: 6 associated providers: 0 name:Intel
server glx vendor string: SGI
client glx vendor string: Mesa Project and SGI
Extended renderer info (GLX_MESA_query_renderer):
OpenGL vendor string: Intel Open Source Technology Center
OpenGL renderer string: Mesa DRI Intel(R) Haswell Mobile 
00:02.0 VGA compatible controller [0300]: Intel Corporation Haswell-ULT Integrated Graphics Controller [8086:0a16] (rev 0b)
04:00.0 VGA compatible controller [0300]: NVIDIA Corporation GK208M [GeForce GT 730M] [10de:1290] (rev a1)

Processor:
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                4
On-line CPU(s) list:   0-3
Thread(s) per core:    2
Core(s) per socket:    2
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 69
Model name:            Intel(R) Core(TM) i5-4300U CPU @ 1.90GHz
Stepping:              1
CPU MHz:               2606.964
CPU max MHz:           2900.0000
CPU min MHz:           800.0000
BogoMIPS:              4988.37
Virtualization:        VT-x
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              3072K
NUMA node0 CPU(s):     0-3
Flags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm epb tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid xsaveopt dtherm ida arat pln pts

OpenCL library:
rc 	amd-libopencl1:amd64	1:15.9-4~deb8u1
un 	libopencl-1.1-1	
un 	libopencl-1.2-1	
un 	libopencl-2.0-1	
un 	libopencl-2.1-1	
un 	libopencl1	
rc 	nvidia-libopencl1:amd64	340.96-1
un 	nvidia-libopencl1-dev	
un 	nvidia-libopencl1-ia32	
ii 	ocl-icd-libopencl1:amd64	2.2.11-1
ii 	ocl-icd-libopencl1:i386	2.2.11-1
ii 	amd-opencl-icd:amd64	1:15.9-4~deb8u2
un 	amd-opencl-icd-ia32	
un 	amd-opencl-icd-virtual	
ii 	beignet-opencl-icd:amd64	1.3.0-2+b1
rc 	mesa-opencl-icd	10.3.2-1+deb8u1
un 	nvidia-opencl-icd	
un 	opencl-icd	
==> /etc/OpenCL/vendors/amdocl64.icd <==
libamdocl64.so

==> /etc/OpenCL/vendors/intel-beignet-x86_64-linux-gnu.icd <==
/usr/lib/x86_64-linux-gnu/beignet//libcl.so

==> /etc/OpenCL/vendors/mesa.icd <==
libMesaOpenCL.so

==> /etc/OpenCL/vendors/nvidia.icd <==
libnvidia-opencl.so.1

-- System Information:
Debian Release: 9.0
  APT prefers testing
  APT policy: (500, 'testing')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.9.0-2-amd64 (SMP w/4 CPU cores)
Locale: LANG=C, LC_CTYPE=de_DE@euro (charmap=ISO-8859-15)
Shell: /bin/sh linked to /bin/dash
Init: sysvinit (via /sbin/init)

Versions of packages beignet-opencl-icd depends on:
ii  libc6          2.24-9
ii  libdrm-intel1  2.4.74-1
ii  libdrm2        2.4.74-1
ii  libgcc1        1:6.3.0-12
ii  libstdc++6     6.3.0-12
ii  libtinfo5      6.0+20161126-1
ii  libx11-6       2:1.6.4-3
ii  libxext6       2:1.3.3-1+b2
ii  libxfixes3     1:5.0.3-1
ii  zlib1g         1:1.2.8.dfsg-5

beignet-opencl-icd recommends no packages.

beignet-opencl-icd suggests no packages.

-- no debconf information

Message #10 received at 860805@bugs.debian.org (full text, mbox, reply):

From: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

To: Michael Gernoth <michael@zerfleddert.de>, 860805@bugs.debian.org

Subject: Re: Bug#860805: beignet-opencl-icd: OpenCL fails with: drm_intel_gem_bo_context_exec() failed: Device or resource busy

Date: Thu, 20 Apr 2017 23:17:14 +0100

Control: severity -1 serious

I have been working on this upstream:
https://bugs.freedesktop.org/show_bug.cgi?id=100639

It appears to totally break beignet on Ivybridge/Haswell on recent 
(including sid/stretch) Linux.  (I didn't notice it before because I 
test in a chroot, i.e. with jessie's too-old-to-be-affected Linux.)

It's not yet entirely clear what to do about it:

The above patch works for me, but apparently not for everyone.

The other known partial fix is to disable softpin, but that also isn't a 
complete fix, and also disables OpenCL 2.0.

Message #17 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Andreas Tille <andreas@an3as.eu>

To: 860805@bugs.debian.org, debian-release@lists.debian.org

Subject: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Fri, 28 Apr 2017 08:33:38 +0200

Hi release team,

this morning 16 "marked for autoremoval from testing" mails hit my
mailbox for partly quite important Debian Med packages.  I'm sure other
packages will be affected as well so this package is somehow cruxial.

I wonder whether you might consider this bug stretch-is-blocker to make
sure we will not loose a larger set of packages.

Kind regards

        Andreas.

-- 
http://fam-tille.de

Message #24 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Niels Thykier <niels@thykier.net>

To: Andreas Tille <andreas@an3as.eu>, 860805@bugs.debian.org, debian-release@lists.debian.org

Cc: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

Subject: Re: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Fri, 28 Apr 2017 07:21:00 +0000

Andreas Tille:
> Hi release team,
> 
> this morning 16 "marked for autoremoval from testing" mails hit my
> mailbox for partly quite important Debian Med packages.  I'm sure other
> packages will be affected as well so this package is somehow cruxial.
> 
> I wonder whether you might consider this bug stretch-is-blocker to make
> sure we will not loose a larger set of packages.
> 
> Kind regards
> 
>         Andreas.
> 

Hi Andreas,

Sorry to hear that you are affected by an auto-removal.

With that and the removal date being in June, we should plenty of time
to see where this bug is going and possibly device a fix for it.   There
is some recent progress on the upstream bug with a patch, which might
bring us closer to a fix  (@Rebecca: Kudos for the timely response to
the bug when it was filed).
  Also, please keep in mind that regular updates to the bug will reset
the timer.  Even a simple """I tried this patch from upstream, it didn't
work because of X.  Notify upstream of the issue""" will reset the
automatic timer.

@Andreas: Assuming worst case, what are the options for dropping beignet
support in your packages?  (I appreciate that they are probably not the
best, but ...)

Thanks,
~Niels

Message #29 received at 860805@bugs.debian.org (full text, mbox, reply):

From: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

To: Andreas Tille <andreas@an3as.eu>, 860805@bugs.debian.org, michael@zerfleddert.de, debian-release@lists.debian.org

Subject: Re: [Pkg-opencl-devel] Bug#860805: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Fri, 28 Apr 2017 08:20:34 +0100

Upstream status update:

- The patch is probably better than disabling softpin, but neither is 
perfect; in particular, there is now a report of *silently wrong 
results* (it's now yet clear with which), which means a partial fix 
*may* be worse than nothing.

- We now have a test case for the error remaining after the patch, but 
it doesn't fail on my system: any of you with Haswell hardware and a 
real (not chrooted) sid/stretch, please apply the patch then try
https://bugs.freedesktop.org/show_bug.cgi?id=100639

> this morning 16 "marked for autoremoval from testing" mails hit my
> mailbox for partly quite important Debian Med packages.  I'm sure other
> packages will be affected as well so this package is somehow cruxial.

I'm not sure why, given that this package is one of several OpenCL 
implementations.

Message #34 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Andreas Tille <andreas@an3as.eu>

To: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

Cc: 860805@bugs.debian.org, michael@zerfleddert.de, debian-release@lists.debian.org

Subject: Re: [Pkg-opencl-devel] Bug#860805: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Fri, 28 Apr 2017 09:35:53 +0200

Hi Rebecca,

On Fri, Apr 28, 2017 at 08:20:34AM +0100, Rebecca N. Palmer wrote:
> - We now have a test case for the error remaining after the patch, but it
> doesn't fail on my system: any of you with Haswell hardware and a real (not
> chrooted) sid/stretch, please apply the patch then try
> https://bugs.freedesktop.org/show_bug.cgi?id=100639

Please excuse my hardware ignorance but how can I tell whether my hardware
has Haswell?  Any one-liner to read out the relevant part from the system
information?
 
> >this morning 16 "marked for autoremoval from testing" mails hit my
> >mailbox for partly quite important Debian Med packages.  I'm sure other
> >packages will be affected as well so this package is somehow cruxial.
> 
> I'm not sure why, given that this package is one of several OpenCL
> implementations.

I guess is based on libhmsbeagle1v5.  It is

   Depends: beignet-opencl-icd|mesa-opencl-icd|opencl-icd

but I've got the "marked for autoremoval from testing" despite the fact
that there are alternatives.  May be it is caused by a failure in the
testing migration script?

Kind regards

        Andreas.

-- 
http://fam-tille.de

Message #39 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Andreas Tille <andreas@an3as.eu>

To: Niels Thykier <niels@thykier.net>

Cc: 860805@bugs.debian.org, debian-release@lists.debian.org, "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

Subject: Re: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Fri, 28 Apr 2017 10:50:46 +0200

Hi Niels,

On Fri, Apr 28, 2017 at 07:21:00AM +0000, Niels Thykier wrote:
> @Andreas: Assuming worst case, what are the options for dropping beignet
> support in your packages?  (I appreciate that they are probably not the
> best, but ...)

As I wrote in my other mail I wonder if the testing removal script is
even correct since the dependencies are

$ apt-cache show libhmsbeagle1v5 | grep ^Depends
Depends: libc6 (>= 2.14), libgcc1 (>= 1:3.0), libstdc++6 (>= 5.2), ocl-icd-libopencl1 | libopencl1, ocl-icd-libopencl1 (>= 1.0) | libopencl-1.1-1, beignet-opencl-icd | mesa-opencl-icd | opencl-icd

so providing alternatives that could fullfill the Dependencies.

Kind regards

       Andreas.

-- 
http://fam-tille.de

Message #44 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Holger Levsen <holger@layer-acht.org>

To: Andreas Tille <andreas@an3as.eu>

Cc: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>, 860805@bugs.debian.org, michael@zerfleddert.de

Subject: Re: [Pkg-opencl-devel] Bug#860805: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Fri, 28 Apr 2017 10:03:31 +0000

[Message part 1 (text/plain, inline)]

On Fri, Apr 28, 2017 at 09:35:53AM +0200, Andreas Tille wrote:
> Please excuse my hardware ignorance but how can I tell whether my hardware
> has Haswell?  Any one-liner to read out the relevant part from the system
> information?
 
compare /proc/cpuinfo with 
https://en.wikipedia.org/wiki/Haswell_%28microarchitecture%29#List_of_Haswell_processors

probably hwinfo from that package will also detected it, dunno.


-- 
cheers,
	Holger

[signature.asc (application/pgp-signature, inline)]

Message #49 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Andreas Tille <andreas@fam-tille.de>

To: Holger Levsen <holger@layer-acht.org>

Cc: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>, 860805@bugs.debian.org, michael@zerfleddert.de

Subject: Re: [Pkg-opencl-devel] Bug#860805: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Sat, 29 Apr 2017 07:15:23 +0200

On Fri, Apr 28, 2017 at 10:03:31AM +0000, Holger Levsen wrote:
> On Fri, Apr 28, 2017 at 09:35:53AM +0200, Andreas Tille wrote:
> > Please excuse my hardware ignorance but how can I tell whether my hardware
> > has Haswell?  Any one-liner to read out the relevant part from the system
> > information?
>  
> compare /proc/cpuinfo with 

 $ grep '^model name'  /proc/cpuinfo | head -n1
model name      : Intel(R) Core(TM) i5-4200U CPU @ 1.60GHz

> https://en.wikipedia.org/wiki/Haswell_%28microarchitecture%29#List_of_Haswell_processors

... 4200U is listed so I guess I can do the testing, right?
 
Kind regards

       Andreas.

-- 
http://fam-tille.de

Message #54 received at 860805@bugs.debian.org (full text, mbox, reply):

From: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

To: Andreas Tille <andreas@fam-tille.de>, 860805@bugs.debian.org, michael@zerfleddert.de

Subject: Re: Bug#860805: beignet-opencl-icd: OpenCL fails with: drm_intel_gem_bo_context_exec() failed: Device or resource busy

Date: Sat, 29 Apr 2017 12:43:19 +0100

The above patch is now in Alioth, but please do *not* upload it yet.

>  $ grep '^model name'  /proc/cpuinfo | head -n1
> model name      : Intel(R) Core(TM) i5-4200U CPU @ 1.60GHz
> ... 4200U is [Haswell] so I guess I can do the testing, right?

If your integrated GPU is enabled (it might not be if you also have a 
discrete GPU - xrandr --listproviders), yes.  Please try, with both 
current beignet-opencl-icd and the above fix:

- the test suite (/usr/lib/x86_64-linux-gnu/beignet/utest_run from 
beignet-dev): many failures before this fix, should be none after.

- the clFFT test from https://bugs.freedesktop.org/show_bug.cgi?id=100639
This needs a small change to work in sid (its reporter was using 
jessie-backports):
-DCMAKE_C_FLAGS:STRING="-O3 -DCL_USE_DEPRECATED_OPENCL_1_2_APIS=1"
as linking to an __inline symbol (clfftInitSetupData) without 
optimization fails in recent gcc.

> Depends: beignet-opencl-icd|mesa-opencl-icd|opencl-icd

I'd suggest just Depends: opencl-icd, as which one will actually work is 
hardware dependent.

Message #59 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Andreas Beckmann <anbe@debian.org>

To: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>, 860805@bugs.debian.org, Andreas Tille <andreas@fam-tille.de>, michael@zerfleddert.de

Subject: Re: [Pkg-opencl-devel] Bug#860805: beignet-opencl-icd: OpenCL fails with: drm_intel_gem_bo_context_exec() failed: Device or resource busy

Date: Sat, 29 Apr 2017 23:25:39 +0200

On 2017-04-29 13:43, Rebecca N. Palmer wrote:
>> Depends: beignet-opencl-icd|mesa-opencl-icd|opencl-icd
> 
> I'd suggest just Depends: opencl-icd, as which one will actually work is
> hardware dependent.

Nope, opencl-icd should *never* occur as the first alternative, since it
could pull in a non-free ICD (and corresponding non-free driver etc.).

I vaguely remember we had discussions about providing some
meta-packages, but I don't think anything was ever implemented.
We should pick this up again for buster :-)


Andreas

Message #64 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Julien Cristau <jcristau@debian.org>

To: Andreas Tille <andreas@an3as.eu>

Cc: Niels Thykier <niels@thykier.net>, 860805@bugs.debian.org, debian-release@lists.debian.org, "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

Subject: Re: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Mon, 1 May 2017 12:56:32 +0200

On Fri, Apr 28, 2017 at 10:50:46 +0200, Andreas Tille wrote:

> Hi Niels,
> 
> On Fri, Apr 28, 2017 at 07:21:00AM +0000, Niels Thykier wrote:
> > @Andreas: Assuming worst case, what are the options for dropping beignet
> > support in your packages?  (I appreciate that they are probably not the
> > best, but ...)
> 
> As I wrote in my other mail I wonder if the testing removal script is
> even correct since the dependencies are
> 
> $ apt-cache show libhmsbeagle1v5 | grep ^Depends
> Depends: libc6 (>= 2.14), libgcc1 (>= 1:3.0), libstdc++6 (>= 5.2), ocl-icd-libopencl1 | libopencl1, ocl-icd-libopencl1 (>= 1.0) | libopencl-1.1-1, beignet-opencl-icd | mesa-opencl-icd | opencl-icd
> 
> so providing alternatives that could fullfill the Dependencies.
> 
It's very much correct.  If beignet-opencl-icd doesn't ship, then it
shouldn't be the first alternative in some other package's Depends.

Cheers,
Julien

Message #69 received at 860805@bugs.debian.org (full text, mbox, reply):

From: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

To: 860805@bugs.debian.org, Andreas Tille <andreas@an3as.eu>, Michael Gernoth <michael@zerfleddert.de>

Subject: Re: [Pkg-opencl-devel] Bug#860805: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Tue, 2 May 2017 07:40:43 +0100

Has anyone tried the clFFT test I requested above?

Nothing further from upstream.

Message #74 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Andreas Tille <andreas@an3as.eu>

To: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

Cc: 860805@bugs.debian.org, michael@zerfleddert.de

Subject: Re: Bug#860805: beignet-opencl-icd: OpenCL fails with: drm_intel_gem_bo_context_exec() failed: Device or resource busy

Date: Tue, 2 May 2017 08:57:17 +0200

Hi Rebecca,

On Sat, Apr 29, 2017 at 12:43:19PM +0100, Rebecca N. Palmer wrote:
> The above patch is now in Alioth, but please do *not* upload it yet.

I have tested beignet-dev_1.3.0-2 and compared it to what you get when
building commit 7d60bf92c6706f174305393172841e2078ecd508.


> > $ grep '^model name'  /proc/cpuinfo | head -n1
> >model name      : Intel(R) Core(TM) i5-4200U CPU @ 1.60GHz
> >... 4200U is [Haswell] so I guess I can do the testing, right?
> 
> If your integrated GPU is enabled (it might not be if you also have a
> discrete GPU - xrandr --listproviders), yes.  Please try, with both current
> beignet-opencl-icd and the above fix:
> 
> - the test suite (/usr/lib/x86_64-linux-gnu/beignet/utest_run from
> beignet-dev): many failures before this fix, should be none after.

I can confirm that beignet-dev_1.3.0-2 produces a lot of errors which
vanished when using said status from Git.  So the patch works perfectly
on my machine.
 
> - the clFFT test from https://bugs.freedesktop.org/show_bug.cgi?id=100639
> This needs a small change to work in sid (its reporter was using
> jessie-backports):
> -DCMAKE_C_FLAGS:STRING="-O3 -DCL_USE_DEPRECATED_OPENCL_1_2_APIS=1"
> as linking to an __inline symbol (clfftInitSetupData) without optimization
> fails in recent gcc.

In other words:  I would need to rebuild beignet-dev_1.3.0-2 with the
said change in d/rules as well as rebuild
7d60bf92c6706f174305393172841e2078ecd508 and compare again?  Could
you please give a direct link what program to run once I have rebuild?
 
Thanks a lot for your work on this

       Andreas. 

-- 
http://fam-tille.de

Message #79 received at 860805@bugs.debian.org (full text, mbox, reply):

From: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

To: Andreas Tille <andreas@an3as.eu>, 860805@bugs.debian.org

Cc: michael@zerfleddert.de

Subject: Re: [Pkg-opencl-devel] Bug#860805: beignet-opencl-icd: OpenCL fails with: drm_intel_gem_bo_context_exec() failed: Device or resource busy

Date: Tue, 2 May 2017 19:04:20 +0100

On 02/05/17 07:57, Andreas Tille wrote:
> I can confirm that beignet-dev_1.3.0-2 produces a lot of errors which
> vanished when using said status from Git.

Good so far.

>> - the clFFT test from https://bugs.freedesktop.org/show_bug.cgi?id=100639
>> This needs a small change to work in sid (its reporter was using
>> jessie-backports):
>> -DCMAKE_C_FLAGS:STRING="-O3 -DCL_USE_DEPRECATED_OPENCL_1_2_APIS=1"
>> as linking to an __inline symbol (clfftInitSetupData) without optimization
>> fails in recent gcc.
>
> In other words:  I would need to rebuild beignet-dev_1.3.0-2 with the
> said change in d/rules as well as rebuild
> 7d60bf92c6706f174305393172841e2078ecd508 and compare again?

No, that change is for the test program (which won't build in sid 
without it), not beignet itself.

Could
> you please give a direct link what program to run once I have rebuild?

https://bugs.freedesktop.org/show_bug.cgi?id=100639#c15
where patch_clfft_haswell.patch is 
https://bugs.freedesktop.org/attachment.cgi?id=131090
(Posting the actual binary would be a security hole, since I don't have 
a GPG key yet...)

Message #84 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Andreas Tille <tille@debian.org>

To: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

Cc: 860805@bugs.debian.org, michael@zerfleddert.de

Subject: Re: [Pkg-opencl-devel] Bug#860805: beignet-opencl-icd: OpenCL fails with: drm_intel_gem_bo_context_exec() failed: Device or resource busy

Date: Tue, 2 May 2017 21:43:22 +0200

Hi Rebecca,

On Tue, May 02, 2017 at 07:04:20PM +0100, Rebecca N. Palmer wrote:
> 
> Could
> >you please give a direct link what program to run once I have rebuild?
> 
> https://bugs.freedesktop.org/show_bug.cgi?id=100639#c15
> where patch_clfft_haswell.patch is
> https://bugs.freedesktop.org/attachment.cgi?id=131090
> (Posting the actual binary would be a security hole, since I don't have a
> GPG key yet...)

I downloaded

   https://bugs.freedesktop.org/attachment.cgi?id=131090

and tried to add it as quilt patch at the end of sequence but it does
not apply.  Would you mind creating an extra branch in Git applying what
should be applied and I just build this?

BTW, it would be really great if you would get your key signed soon. ;-)

Kind regards

       Andreas.

-- 
http://fam-tille.de

Message #89 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Michael Gernoth <michael@zerfleddert.de>

To: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

Cc: 860805@bugs.debian.org, Andreas Tille <andreas@an3as.eu>

Subject: Re: [Pkg-opencl-devel] Bug#860805: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Tue, 2 May 2017 22:42:17 +0200

Hi,

On Tue, May 02, 2017 at 07:40:43AM +0100, Rebecca N. Palmer wrote:
> Has anyone tried the clFFT test I requested above?

I've now done a test on my system (Thinkpad T440s, i5-4300U):

1. Without the patch (stderr):

   drm_intel_gem_bo_context_exec() failed: Device or resource busy
   Beignet: "Exec event 0x55e6d5920fd0 error, type is 4592, error staus is -5"
   (err: 0)after clGetPlatformIDs
   (err: 0) Platform found: Intel Gen OCL Driver
   (err: 0)after clGetDeviceIDs
   (err: 0)Device found on the above platform: Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile
   (err: 0) after clCreateContext
   (err: 0)after clCreateCommandQueue
   (err: 0)after clfftInitSetupData
   (err: 0)after clfftSetup
   
   Performing fft on an one dimensional array of size N = 8192
   (err: 0)after clCreateBuffer
   (err: 0)after clfftCreateDefaultPlan
   (err: 0)after clfftSetPlanPrecision
   (err: 0)after clfftSetLayout
   (err: 0)after clfftSetResultLocation
   (err: 0)after clfftBakePlan
   (err: 0)after clEnqueueWriteBuffer
   drm_intel_gem_bo_context_exec() failed: Device or resource busy
   Beignet: "Exec event 0x55e6d61cf540 error, type is 4592, error staus is -5"
   clEnqueueNDRangeKernel failed second... -14 params: 1 1024 128 0
   ERROR executing the kernel -5
   (err: -14)after clfftEnqueueTransform
   (err: -14)after clfftEnqueueTransform = CL_EXEC_STATUS_ERROR_FOR_EVENTS_IN_WAIT_LIST
   (err: 0)after clFinish
   (err: 0)after clEnqueueReadBuffer
   Time took 1: 0.000406


ERROR: found same numbers !!! ... FFT was not executed !!!

2. With the patch (stderr):

   (err: 0)after clGetPlatformIDs
   (err: 0) Platform found: Intel Gen OCL Driver
   (err: 0)after clGetDeviceIDs
   (err: 0)Device found on the above platform: Intel(R) HD Graphics Haswell Ultrabook GT2 Mobile
   (err: 0) after clCreateContext
   (err: 0)after clCreateCommandQueue
   (err: 0)after clfftInitSetupData
   (err: 0)after clfftSetup
   
   Performing fft on an one dimensional array of size N = 8192
   (err: 0)after clCreateBuffer
   (err: 0)after clfftCreateDefaultPlan
   (err: 0)after clfftSetPlanPrecision
   (err: 0)after clfftSetLayout
   (err: 0)after clfftSetResultLocation
   (err: 0)after clfftBakePlan
   (err: 0)after clEnqueueWriteBuffer
   (err: 0)after clfftEnqueueTransform
   (err: 0)after clFinish
   (err: 0)after clEnqueueReadBuffer
   Time took 1: 0.000887

stdout contains values which could be the in- and output of an FFT in
the 2. case.

Cheers,
  Michael

Message #94 received at 860805@bugs.debian.org (full text, mbox, reply):

From: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

To: Michael Gernoth <michael@zerfleddert.de>, 860805@bugs.debian.org

Cc: Andreas Tille <andreas@an3as.eu>

Subject: Re: [Pkg-opencl-devel] Bug#860805: Bug#860805: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Tue, 2 May 2017 23:31:08 +0100

Control: tags -1 pending

The above looks like success - thanks.

Someone please upload beignet (as currently in Alioth); I'll file an 
unblock request.

Message #101 received at 860805@bugs.debian.org (full text, mbox, reply):

From: Andreas Beckmann <anbe@debian.org>

To: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>, 860805@bugs.debian.org, Michael Gernoth <michael@zerfleddert.de>

Cc: Andreas Tille <andreas@an3as.eu>

Subject: Re: Bug#860805: Could we set bug #860805 against beignet-opencl-icd to stretch-is-blocker?

Date: Wed, 3 May 2017 01:07:18 +0200

On 2017-05-03 00:31, Rebecca N. Palmer wrote:
> Someone please upload beignet (as currently in Alioth); I'll file an
> unblock request.

Upload is on its way. You can probably wait a few days with the unblock
request, the release-team will probably pick up the upload fixing an RC
bug on their own.

Andreas

Message #106 received at 860805-close@bugs.debian.org (full text, mbox, reply):

From: rebecca_palmer@zoho.com (Rebecca N. Palmer)

To: 860805-close@bugs.debian.org

Subject: Bug#860805: fixed in beignet 1.3.0-3

Date: Tue, 02 May 2017 23:18:32 +0000

Source: beignet
Source-Version: 1.3.0-3

We believe that the bug you reported is fixed in the latest version of
beignet, which is due to be installed in the Debian FTP archive.

A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 860805@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Rebecca N. Palmer <rebecca_palmer@zoho.com> (supplier of updated beignet package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@ftp-master.debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Format: 1.8
Date: Tue, 02 May 2017 23:23:11 +0100
Source: beignet
Binary: beignet-opencl-icd beignet-dev beignet
Architecture: source
Version: 1.3.0-3
Distribution: unstable
Urgency: medium
Maintainer: Debian OpenCL Maintainers <pkg-opencl-devel@lists.alioth.debian.org>
Changed-By: Rebecca N. Palmer <rebecca_palmer@zoho.com>
Description:
 beignet    - OpenCL library for Intel GPUs - transitional dummy package
 beignet-dev - OpenCL for Intel GPUs (development files and documentation)
 beignet-opencl-icd - OpenCL library for Intel GPUs
Closes: 860805
Changes:
 beignet (1.3.0-3) unstable; urgency=medium
 .
   * Fix "Exec...-5" error on older hardware.  (Closes: #860805)
   * Use LLVM 3.8 on x32 to fix FTBFS.
Checksums-Sha1:
 d40af8ab304bd4ef0082a00b311abd5a366ead59 2660 beignet_1.3.0-3.dsc
 0c839365da8219bb995439a0fdecac6403ad417b 25120 beignet_1.3.0-3.debian.tar.xz
 706762251bd3a49e60f4f7c6e3d21f70ad069189 7827 beignet_1.3.0-3_source.buildinfo
Checksums-Sha256:
 3aaefe938b2c16ea6e80761c62d9290dd9345fb27f8fc1d406094d14f3c224a3 2660 beignet_1.3.0-3.dsc
 3fbd09eef1e796773a0b31e1c07da9ef4de4fc28782bfeff9d1f828034048072 25120 beignet_1.3.0-3.debian.tar.xz
 b6844013f8ae31ba83e25b7502b3a7fcdf329762bbde0bf0e3b719685f24512e 7827 beignet_1.3.0-3_source.buildinfo
Files:
 a4dac95188292604efa365a480967ece 2660 libs extra beignet_1.3.0-3.dsc
 8900717a98abbf62d05f83aeace1d7e1 25120 libs extra beignet_1.3.0-3.debian.tar.xz
 897f9d754d5ed8d17f9cac27683208c8 7827 libs extra beignet_1.3.0-3_source.buildinfo

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=gaaQ
-----END PGP SIGNATURE-----

Message #111 received at 860805@bugs.debian.org (full text, mbox, reply):

From: "Rebecca N. Palmer" <rebecca_palmer@zoho.com>

To: 860805@bugs.debian.org

Subject: Re: Bug#860805: beignet-opencl-icd: OpenCL fails with: drm_intel_gem_bo_context_exec() failed: Device or resource busy

Date: Sat, 13 May 2017 10:46:24 +0100

The upstream report of "this isn't a complete fix" turned out to be 
mistaken: the user had both versions and had accidentally used the 
unfixed one.  https://bugs.freedesktop.org/show_bug.cgi?id=100639#c19

This bug also affects jessie-backports beignet when used with 
jessie-backports Linux (which is likely as, for unrelated reasons, 
OpenCL doesn't work properly on Haswell with jessie Linux).  Fixed 
backport in Alioth, ready for upload.

Send a report that this bug log contains spam.