Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Paul Gevers]

Hi,

On 06-04-2022 21:30, Paul Gevers wrote:

On 06-04-2022 21:26, Paul Gevers wrote:

but ffmpeg succeeds.


But this is interesting in its output:
[libx264 @ 0x1a181a0] using cpu capabilities: ARMv6 NEON


https://www.baeldung.com/linux/proc-cpuinfo-flags#3-neon says we 
shouldn't look for "neon", but for asimd, which indeed is in the 
features set of this CPU (we run armhf lxc on amd64 host, so the CPU is 
actually 64 bits).


Paul


OpenPGP_signature
Description: OpenPGP digital signature

Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Paul Gevers]

Hi

On 06-04-2022 21:26, Paul Gevers wrote:

but ffmpeg succeeds.


But this is interesting in its output:
[libx264 @ 0x1a181a0] using cpu capabilities: ARMv6 NEON

Paul


OpenPGP_signature
Description: OpenPGP digital signature

Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Paul Gevers]

Hi,

On 06-04-2022 08:48, Ying-Chun Liu (PaulLiu) wrote:

I'm not sure if you have extra time.
But I think you are correct. We might need to report the bug to ffmpeg.

I'm not sure if we can simplify the test.
Like just run "ffplay test_raw.avi" or "ffmpeg -i test_raw.avi out.mp4" 
to see if it has the bus error as well.


ffplay fails with:
error: XDG_RUNTIME_DIR not set in the environment.
Could not initialize SDL - No available video device
(Did you set the DISPLAY variable?)

but ffmpeg succeeds. But then, I also had to install quite some packages 
to install ffmpeg. Maybe something is needed?


  libavdevice58 libcdio-cdda2 libcdio-paranoia2 libcdio19 libdc1394-25 
libdecor-0-0
  libdrm-amdgpu1 libdrm-nouveau2 libdrm-radeon1 libgbm1 libgl1 
libgl1-mesa-dri
  libglapi-mesa libglvnd0 libglx-mesa0 libglx0 libllvm13 libopenal-data 
libopenal1
  libsdl2-2.0-0 libsensors-config libsensors5 libsndio7.0 libusb-1.0-0 
libvulkan1
  libwayland-client0 libwayland-cursor0 libwayland-egl1 
libwayland-server0 libxcb-dri2-0
  libxcb-dri3-0 libxcb-glx0 libxcb-present0 libxcb-shape0 libxcb-sync1 
libxcb-xfixes0
  libxcursor1 libxi6 libxinerama1 libxkbcommon0 libxrandr2 
libxshmfence1 libxss1 libxv1

  libz3-4 x11-common xkb-data

Paul


OpenPGP_signature
Description: OpenPGP digital signature

Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Ying-Chun Liu (PaulLiu)]

Hi Paul,

I'm not sure if you have extra time.
But I think you are correct. We might need to report the bug to ffmpeg.

I'm not sure if we can simplify the test.
Like just run "ffplay test_raw.avi" or "ffmpeg -i test_raw.avi out.mp4" 
to see if it has the bus error as well.


Yours,
Paul

Paul Gevers 於 2022/4/6 02:52 寫道:

Hi,

On 04-04-2022 23:11, Ying-Chun Liu (PaulLiu) wrote:
It is because your environment supports "neon" and ffmpeg will use 
that for acceleration.


I wonder.


But abel and qemu-user-static doesn't support neon.
I think if you "cat /proc/cpuinfo" and neon will be in the Features.


root@ci-worker-armel-01:~# cat /proc/cpuinfo | grep neon
root@ci-worker-armel-01:~#

Nope.

Features    : fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics 
fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp ssbs


I now notice that the list of flags is also in the gdb output. If 
anything, it more looks like ffmpeg *thinks* it can use neon, but 
can't. Would that be a bug in ffmpeg?


I'll see if I can use cpuflags in ffmpeg to disable neon on my test. 
So that it can pass.


Ack, but I wonder if the bug needs to be cloned an reassigned to 
ffmpeg then, for detecting neon when not available.


Paul


OpenPGP_signature
Description: OpenPGP digital signature

Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Paul Gevers]

Hi,

On 04-04-2022 23:11, Ying-Chun Liu (PaulLiu) wrote:
It is because your environment supports "neon" and ffmpeg will use that 
for acceleration.


I wonder.


But abel and qemu-user-static doesn't support neon.
I think if you "cat /proc/cpuinfo" and neon will be in the Features.


root@ci-worker-armel-01:~# cat /proc/cpuinfo | grep neon
root@ci-worker-armel-01:~#

Nope.

Features: fp asimd evtstrm aes pmull sha1 sha2 crc32 atomics 
fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp ssbs


I now notice that the list of flags is also in the gdb output. If 
anything, it more looks like ffmpeg *thinks* it can use neon, but can't. 
Would that be a bug in ffmpeg?


I'll see if I can use cpuflags in ffmpeg to disable neon on my test. So 
that it can pass.


Ack, but I wonder if the bug needs to be cloned an reassigned to ffmpeg 
then, for detecting neon when not available.


Paul


OpenPGP_signature
Description: OpenPGP digital signature

Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Ying-Chun Liu (PaulLiu)]

Hi Paul,

Thanks for the report. Now I know why abel.debian.org or 
qemu-user-static doesn't reproduce the bug.
It is because your environment supports "neon" and ffmpeg will use that 
for acceleration.

But abel and qemu-user-static doesn't support neon.
I think if you "cat /proc/cpuinfo" and neon will be in the Features.

I'll see if I can use cpuflags in ffmpeg to disable neon on my test. So 
that it can pass.


Yours,
Paul

Paul Gevers 於 2022/4/5 03:16 寫道:

Hi

Bus error
autopkgtest [19:03:51]: test decoding-test-mp4-raw: 
---]
autopkgtest [19:03:51]: test decoding-test-mp4-raw:  - - - - - - - - - 
- results - - - - - - - - - -

decoding-test-mp4-raw FAIL non-zero exit status 135
autopkgtest [19:03:51]:  - - - - - - - - - - running shell - - - - - - 
- - - -
root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src# cd 
debian/tests/decoding-test-data/
root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 
./gentestavi.sh

Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Redistribute latency...
Pipeline is PREROLLED ...
Setting pipeline to PLAYING ...
Redistribute latency...
New clock: GstSystemClock
Got EOS from element "pipeline0".
Execution ended after 0:00:00.215248380
Setting pipeline to NULL ...
Freeing pipeline ...
Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Pipeline is PREROLLED ...
Setting pipeline to PLAYING ...
Redistribute latency...
New clock: GstSystemClock
Got EOS from element "pipeline0".
Execution ended after 0:00:00.219812907
Setting pipeline to NULL ...
Freeing pipeline ...
root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 

root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 

root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 
make

g++ -Wall -g -o test1 test1.cc \
    -laviplay -I/usr/include/avifile-0.7 \
    -laubio  \
    -lm
root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debianroot@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 
apt install gdb libavcodec58-dbgsym


< installing gdb and libavcodec58-dbgsym >

root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 
gdb ./test1

gdb: warning: Couldn't determine a path for the index cache directory.
GNU gdb (Debian 10.1-2+b1) 10.1.90.20210103-git
Copyright (C) 2021 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later 


This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "arm-linux-gnueabihf".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
.
Find the GDB manual and other documentation resources online at:
    .

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from ./test1...
(gdb) set args test_raw.avi
(gdb) run
Starting program: 
/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data/test1 
test_raw.avi

[Thread debugging using libthread_db enabled]
Using host libthread_db library 
"/lib/arm-linux-gnueabihf/libthread_db.so.1".

 : Avifile RELEASE-0.7.48-220318-20:07-../src/configure
 : Available CPU flags: fp asimd evtstrm aes pmull sha1 sha2 
crc32 atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp ssbs

This binary was compiled for Avifile ver. 1840, the library is ver. 1840
 : Checking: test_raw.avi
 : MainHeader: MicroSecPerFrame=4 MaxBytesPerSec=6112800
 PaddingGranularity=0 Flags=[ HAS_INDEX IS_INTERLEAVED ] TotalFrames=250
 InitialFrames=0 Streams=2 SuggestedBufferSize=0 WxH=320x240
 Scale=0 Rate=0 Start=0 Length=0
 : StreamHeader: Type=vids Handler=DIVX Flags=[ ]
 InitialFrames=0 Scale=1 Rate=25 Start=0 Length=250
 SuggestedBufferSize=12413 Quality=0 SampleSize=0 Rect l,r,t,b=0,0,0,0
 : StreamHeader: Type=auds Handler=0x1 Flags=[ ]
 InitialFrames=0 Scale=1 Rate=44100 Start=0 Length=450559
 SuggestedBufferSize=8192 Quality=0 SampleSize=8 Rect l,r,t,b=0,0,0,0
 : Reading index from offset: 5581514
 : Stream 0 vids : DIVX (0x4944) 250 chunks (0.98KB)
 : WARNING: stream header has incorrect dwLength (450559 
-> 450560)

 : Stream 1 auds : Microsoft PCM (0x1) 440 chunks (3.44KB)
[New Thread 0xefdfea20 (LWP 6059)]
 : Initialized video stream (chunk tblsz: 250, fmtsz: 40)
 : Found 7 plugins 
(/usr/lib/arm-linux-gnueabihf/avifile-0.7,A:28,V:36)

 : looking for mpeg4  1482049860
 : Created video decoder: FF OpenDivX
 : Initialized audio stream (chunk tblsz: 450560, fmtsz: 18)
 : PCM audio decoder created
bh.biHeight = -240, bh.biWidth = 320
bh.biHeight < 0, 

Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Paul Gevers]

Hi

Bus error
autopkgtest [19:03:51]: test decoding-test-mp4-raw: ---]
autopkgtest [19:03:51]: test decoding-test-mp4-raw:  - - - - - - - - - - 
results - - - - - - - - - -

decoding-test-mp4-raw FAIL non-zero exit status 135
autopkgtest [19:03:51]:  - - - - - - - - - - running shell - - - - - - - 
- - -
root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src# cd 
debian/tests/decoding-test-data/
root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 
./gentestavi.sh

Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Redistribute latency...
Pipeline is PREROLLED ...
Setting pipeline to PLAYING ...
Redistribute latency...
New clock: GstSystemClock
Got EOS from element "pipeline0".
Execution ended after 0:00:00.215248380
Setting pipeline to NULL ...
Freeing pipeline ...
Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Pipeline is PREROLLED ...
Setting pipeline to PLAYING ...
Redistribute latency...
New clock: GstSystemClock
Got EOS from element "pipeline0".
Execution ended after 0:00:00.219812907
Setting pipeline to NULL ...
Freeing pipeline ...
root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 

root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 

root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 
make

g++ -Wall -g -o test1 test1.cc \
-laviplay -I/usr/include/avifile-0.7 \
-laubio  \
-lm
root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debianroot@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 
apt install gdb libavcodec58-dbgsym


< installing gdb and libavcodec58-dbgsym >

root@elbrus:/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data# 
gdb ./test1

gdb: warning: Couldn't determine a path for the index cache directory.
GNU gdb (Debian 10.1-2+b1) 10.1.90.20210103-git
Copyright (C) 2021 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later 


This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Type "show copying" and "show warranty" for details.
This GDB was configured as "arm-linux-gnueabihf".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
.
Find the GDB manual and other documentation resources online at:
.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from ./test1...
(gdb) set args test_raw.avi
(gdb) run
Starting program: 
/tmp/autopkgtest-lxc.ze1_2t2a/downtmp/build.e5m/src/debian/tests/decoding-test-data/test1 
test_raw.avi

[Thread debugging using libthread_db enabled]
Using host libthread_db library 
"/lib/arm-linux-gnueabihf/libthread_db.so.1".

 : Avifile RELEASE-0.7.48-220318-20:07-../src/configure
 : Available CPU flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 
atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp ssbs

This binary was compiled for Avifile ver. 1840, the library is ver. 1840
 : Checking: test_raw.avi
 : MainHeader: MicroSecPerFrame=4 MaxBytesPerSec=6112800
 PaddingGranularity=0 Flags=[ HAS_INDEX IS_INTERLEAVED ] TotalFrames=250
 InitialFrames=0 Streams=2 SuggestedBufferSize=0 WxH=320x240
 Scale=0 Rate=0 Start=0 Length=0
 : StreamHeader: Type=vids Handler=DIVX Flags=[ ]
 InitialFrames=0 Scale=1 Rate=25 Start=0 Length=250
 SuggestedBufferSize=12413 Quality=0 SampleSize=0 Rect l,r,t,b=0,0,0,0
 : StreamHeader: Type=auds Handler=0x1 Flags=[ ]
 InitialFrames=0 Scale=1 Rate=44100 Start=0 Length=450559
 SuggestedBufferSize=8192 Quality=0 SampleSize=8 Rect l,r,t,b=0,0,0,0
 : Reading index from offset: 5581514
 : Stream 0 vids : DIVX (0x4944) 250 chunks (0.98KB)
 : WARNING: stream header has incorrect dwLength (450559 -> 
450560)

 : Stream 1 auds : Microsoft PCM (0x1) 440 chunks (3.44KB)
[New Thread 0xefdfea20 (LWP 6059)]
 : Initialized video stream (chunk tblsz: 250, fmtsz: 40)
 : Found 7 plugins 
(/usr/lib/arm-linux-gnueabihf/avifile-0.7,A:28,V:36)

 : looking for mpeg4  1482049860
 : Created video decoder: FF OpenDivX
 : Initialized audio stream (chunk tblsz: 450560, fmtsz: 18)
 : PCM audio decoder created
bh.biHeight = -240, bh.biWidth = 320
bh.biHeight < 0, correct the value to 240
Movie size: 320 x 240  [YV12]
audio format: 1, Channels: 2, Samples/sec: 44100, Bits/Sample: 32, 
Bytes/sec: 352800

Video Length: 250
Video Pos: 0
Audio Length: 450560
Audio Pos: 0
Decoder YUV capabilities: 0x8080
CAPS is CAP_YV12
TIME 0 0
 : using DR1
[mpeg4 @ 0x455690] Requested frame threading with a custom get_buffer2() 
implementation which is not marked as thread safe. This is not supported 
anymore, make your callback thread-safe.

Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Ying-Chun Liu (PaulLiu)]

Hi Paul,

In debian/tests/decoding-test-data, please run "./gentestavi.sh". This 
command generates the test data (test_raw.avi).


And run "make". It will compile the test program (test1).
Then run "gdb ./test1".
In gdb prompt, execute "set args test_raw.avi".
And then "run".
When the bus error happened, execute "bt -full" and it will print out 
the backtrace.


You might want to install the debug symbols (-dbg) packages to get the 
symbol resolved.


Yours,
Paul

Ying-Chun Liu (PaulLiu) 於 2022/4/1 20:53 寫道:

Hi Paul,


I need some help. Actually I know this bug for a while but I just 
cannot fix it. I've tried to reproduce the bug on qemu-user-static 
armhf chroot envorionment and PorterBox (abel.debian.org), on both 
platform the autopkgtest just passed without problem.


I think the bug should be easy to locate if I can run gdb and see the 
backtrace when the bus error happened.
But I just cannot reproduce this bug on my side or on porter box. If 
possible I hope to get some core dump when bus error happened on the 
machine that runs the test.


Yours,
Paul




Paul Gevers 於 2022/4/1 03:31 寫道:

Source: avifile
Version: 1:0.7.48~20090503.ds-23
Severity: serious
User: debian...@lists.debian.org
Usertags: regression

Dear maintainer(s),

With a recent upload of avifile the autopkgtest of avifile fails in 
testing on armhf when that autopkgtest is run with the binary 
packages of avifile from unstable. It passes when run with only 
packages from testing. In tabular form:


   pass    fail
avifile    from testing    1:0.7.48~20090503.ds-23
all others from testing    from testing

I copied some of the output at the bottom of this report.

Currently this regression is blocking the migration to testing [1]. 
Can you please investigate the situation and fix it?


More information about this bug and the reason for filing it can be 
found on

https://wiki.debian.org/ContinuousIntegration/RegressionEmailInformation

Paul

[1] https://qa.debian.org/excuses.php?package=avifile

https://ci.debian.net/data/autopkgtest/testing/armhf/a/avifile/20131436/log.gz 



Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Redistribute latency...
Pipeline is PREROLLED ...
Setting pipeline to PLAYING ...
Redistribute latency...
New clock: GstSystemClock
Got EOS from element "pipeline0".
Execution ended after 0:00:00.312851570
Setting pipeline to NULL ...
Freeing pipeline ...
Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Pipeline is PREROLLED ...
Setting pipeline to PLAYING ...
Redistribute latency...
New clock: GstSystemClock
Got EOS from element "pipeline0".
Execution ended after 0:00:00.217510393
Setting pipeline to NULL ...
Freeing pipeline ...
g++ -Wall -g -o test1 test1.cc \
-laviplay -I/usr/include/avifile-0.7 \
-laubio  \
-lm
 : Avifile RELEASE-0.7.48-220318-20:07-../src/configure
 : Available CPU flags: fp asimd evtstrm aes pmull sha1 sha2 
crc32 atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp ssbs

This binary was compiled for Avifile ver. 1840, the library is ver. 1840
 : Checking: ./test_raw.avi
 : MainHeader: MicroSecPerFrame=4 MaxBytesPerSec=6112800
 PaddingGranularity=0 Flags=[ HAS_INDEX IS_INTERLEAVED ] TotalFrames=250
 InitialFrames=0 Streams=2 SuggestedBufferSize=0 WxH=320x240
 Scale=0 Rate=0 Start=0 Length=0
 : StreamHeader: Type=vids Handler=DIVX Flags=[ ]
 InitialFrames=0 Scale=1 Rate=25 Start=0 Length=250
 SuggestedBufferSize=12413 Quality=0 SampleSize=0 Rect l,r,t,b=0,0,0,0
 : StreamHeader: Type=auds Handler=0x1 Flags=[ ]
 InitialFrames=0 Scale=1 Rate=44100 Start=0 Length=450559
 SuggestedBufferSize=8192 Quality=0 SampleSize=8 Rect l,r,t,b=0,0,0,0
 : Reading index from offset: 5581514
 : Stream 0 vids : DIVX (0x4944) 250 chunks (0.98KB)
 : WARNING: stream header has incorrect dwLength (450559 
-> 450560)

 : Stream 1 auds : Microsoft PCM (0x1) 440 chunks (3.44KB)
 : Initialized video stream (chunk tblsz: 250, fmtsz: 40)
 : Found 7 plugins 
(/usr/lib/arm-linux-gnueabihf/avifile-0.7,A:28,V:36)

 : creating dir: /home/debci/.avm
 : looking for mpeg4  1482049860
 : Created video decoder: FF OpenDivX
 : Initialized audio stream (chunk tblsz: 450560, fmtsz: 18)
 : PCM audio decoder created
bh.biHeight = -240, bh.biWidth = 320
bh.biHeight < 0, correct the value to 240
Movie size: 320 x 240  [YV12]
audio format: 1, Channels: 2, Samples/sec: 44100, Bits/Sample: 32, 
Bytes/sec: 352800

Video Length: 250
Video Pos: 0
Audio Length: 450560
Audio Pos: 0
Decoder YUV capabilities: 0x8080
CAPS is CAP_YV12
TIME 0 0
[mpeg4 @ 0x1ad7e40] Requested frame threading with a custom 
get_buffer2() implementation which is not marked as thread safe. This 
is not supported anymore, make your callback thread-safe.

 : using DR1
Audio read 192 samples, 1536 bytes.
Audio frequency 469.27 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.188 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.095 hz
Audio read 192 samples, 1536 

Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Paul Gevers]

Hi,

On 01-04-2022 14:53, Ying-Chun Liu (PaulLiu) wrote:
I need some help. Actually I know this bug for a while but I just cannot 
fix it. I've tried to reproduce the bug on qemu-user-static armhf chroot 
envorionment and PorterBox (abel.debian.org), on both platform the 
autopkgtest just passed without problem.


I think the bug should be easy to locate if I can run gdb and see the 
backtrace when the bus error happened.
But I just cannot reproduce this bug on my side or on porter box. If 
possible I hope to get some core dump when bus error happened on the 
machine that runs the test.


Glad to help, but can you please give me exact commands what you want me 
to run in the testbed? I'm not very familiar with gdb and last time I 
used it is years ago.


Paul


OpenPGP_signature
Description: OpenPGP digital signature

Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Ying-Chun Liu (PaulLiu)]

Hi Paul,


I need some help. Actually I know this bug for a while but I just cannot 
fix it. I've tried to reproduce the bug on qemu-user-static armhf chroot 
envorionment and PorterBox (abel.debian.org), on both platform the 
autopkgtest just passed without problem.


I think the bug should be easy to locate if I can run gdb and see the 
backtrace when the bus error happened.
But I just cannot reproduce this bug on my side or on porter box. If 
possible I hope to get some core dump when bus error happened on the 
machine that runs the test.


Yours,
Paul




Paul Gevers 於 2022/4/1 03:31 寫道:

Source: avifile
Version: 1:0.7.48~20090503.ds-23
Severity: serious
User: debian...@lists.debian.org
Usertags: regression

Dear maintainer(s),

With a recent upload of avifile the autopkgtest of avifile fails in 
testing on armhf when that autopkgtest is run with the binary packages 
of avifile from unstable. It passes when run with only packages from 
testing. In tabular form:


   pass    fail
avifile    from testing    1:0.7.48~20090503.ds-23
all others from testing    from testing

I copied some of the output at the bottom of this report.

Currently this regression is blocking the migration to testing [1]. 
Can you please investigate the situation and fix it?


More information about this bug and the reason for filing it can be 
found on

https://wiki.debian.org/ContinuousIntegration/RegressionEmailInformation

Paul

[1] https://qa.debian.org/excuses.php?package=avifile

https://ci.debian.net/data/autopkgtest/testing/armhf/a/avifile/20131436/log.gz 



Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Redistribute latency...
Pipeline is PREROLLED ...
Setting pipeline to PLAYING ...
Redistribute latency...
New clock: GstSystemClock
Got EOS from element "pipeline0".
Execution ended after 0:00:00.312851570
Setting pipeline to NULL ...
Freeing pipeline ...
Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Pipeline is PREROLLED ...
Setting pipeline to PLAYING ...
Redistribute latency...
New clock: GstSystemClock
Got EOS from element "pipeline0".
Execution ended after 0:00:00.217510393
Setting pipeline to NULL ...
Freeing pipeline ...
g++ -Wall -g -o test1 test1.cc \
-laviplay -I/usr/include/avifile-0.7 \
-laubio  \
-lm
 : Avifile RELEASE-0.7.48-220318-20:07-../src/configure
 : Available CPU flags: fp asimd evtstrm aes pmull sha1 sha2 
crc32 atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp ssbs

This binary was compiled for Avifile ver. 1840, the library is ver. 1840
 : Checking: ./test_raw.avi
 : MainHeader: MicroSecPerFrame=4 MaxBytesPerSec=6112800
 PaddingGranularity=0 Flags=[ HAS_INDEX IS_INTERLEAVED ] TotalFrames=250
 InitialFrames=0 Streams=2 SuggestedBufferSize=0 WxH=320x240
 Scale=0 Rate=0 Start=0 Length=0
 : StreamHeader: Type=vids Handler=DIVX Flags=[ ]
 InitialFrames=0 Scale=1 Rate=25 Start=0 Length=250
 SuggestedBufferSize=12413 Quality=0 SampleSize=0 Rect l,r,t,b=0,0,0,0
 : StreamHeader: Type=auds Handler=0x1 Flags=[ ]
 InitialFrames=0 Scale=1 Rate=44100 Start=0 Length=450559
 SuggestedBufferSize=8192 Quality=0 SampleSize=8 Rect l,r,t,b=0,0,0,0
 : Reading index from offset: 5581514
 : Stream 0 vids : DIVX (0x4944) 250 chunks (0.98KB)
 : WARNING: stream header has incorrect dwLength (450559 
-> 450560)

 : Stream 1 auds : Microsoft PCM (0x1) 440 chunks (3.44KB)
 : Initialized video stream (chunk tblsz: 250, fmtsz: 40)
 : Found 7 plugins 
(/usr/lib/arm-linux-gnueabihf/avifile-0.7,A:28,V:36)

 : creating dir: /home/debci/.avm
 : looking for mpeg4  1482049860
 : Created video decoder: FF OpenDivX
 : Initialized audio stream (chunk tblsz: 450560, fmtsz: 18)
 : PCM audio decoder created
bh.biHeight = -240, bh.biWidth = 320
bh.biHeight < 0, correct the value to 240
Movie size: 320 x 240  [YV12]
audio format: 1, Channels: 2, Samples/sec: 44100, Bits/Sample: 32, 
Bytes/sec: 352800

Video Length: 250
Video Pos: 0
Audio Length: 450560
Audio Pos: 0
Decoder YUV capabilities: 0x8080
CAPS is CAP_YV12
TIME 0 0
[mpeg4 @ 0x1ad7e40] Requested frame threading with a custom 
get_buffer2() implementation which is not marked as thread safe. This 
is not supported anymore, make your callback thread-safe.

 : using DR1
Audio read 192 samples, 1536 bytes.
Audio frequency 469.27 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.188 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.095 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.086 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.17 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.262 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.267 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.18 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.091 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.09 hz
TIME 0.04 1
Bus error
autopkgtest [10:24:27]: test decoding-test-mp4-raw



OpenPGP_signature
Description: 

Bug#1008752: avifile: autopkgtest regression on armhf: Bus error

[Paul Gevers]

Source: avifile
Version: 1:0.7.48~20090503.ds-23
Severity: serious
User: debian...@lists.debian.org
Usertags: regression

Dear maintainer(s),

With a recent upload of avifile the autopkgtest of avifile fails in 
testing on armhf when that autopkgtest is run with the binary packages 
of avifile from unstable. It passes when run with only packages from 
testing. In tabular form:


   passfail
avifilefrom testing1:0.7.48~20090503.ds-23
all others from testingfrom testing

I copied some of the output at the bottom of this report.

Currently this regression is blocking the migration to testing [1]. Can 
you please investigate the situation and fix it?


More information about this bug and the reason for filing it can be found on
https://wiki.debian.org/ContinuousIntegration/RegressionEmailInformation

Paul

[1] https://qa.debian.org/excuses.php?package=avifile

https://ci.debian.net/data/autopkgtest/testing/armhf/a/avifile/20131436/log.gz

Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Redistribute latency...
Pipeline is PREROLLED ...
Setting pipeline to PLAYING ...
Redistribute latency...
New clock: GstSystemClock
Got EOS from element "pipeline0".
Execution ended after 0:00:00.312851570
Setting pipeline to NULL ...
Freeing pipeline ...
Setting pipeline to PAUSED ...
Pipeline is PREROLLING ...
Pipeline is PREROLLED ...
Setting pipeline to PLAYING ...
Redistribute latency...
New clock: GstSystemClock
Got EOS from element "pipeline0".
Execution ended after 0:00:00.217510393
Setting pipeline to NULL ...
Freeing pipeline ...
g++ -Wall -g -o test1 test1.cc \
-laviplay -I/usr/include/avifile-0.7 \
-laubio  \
-lm
 : Avifile RELEASE-0.7.48-220318-20:07-../src/configure
 : Available CPU flags: fp asimd evtstrm aes pmull sha1 sha2 crc32 
atomics fphp asimdhp cpuid asimdrdm lrcpc dcpop asimddp ssbs

This binary was compiled for Avifile ver. 1840, the library is ver. 1840
 : Checking: ./test_raw.avi
 : MainHeader: MicroSecPerFrame=4 MaxBytesPerSec=6112800
 PaddingGranularity=0 Flags=[ HAS_INDEX IS_INTERLEAVED ] TotalFrames=250
 InitialFrames=0 Streams=2 SuggestedBufferSize=0 WxH=320x240
 Scale=0 Rate=0 Start=0 Length=0
 : StreamHeader: Type=vids Handler=DIVX Flags=[ ]
 InitialFrames=0 Scale=1 Rate=25 Start=0 Length=250
 SuggestedBufferSize=12413 Quality=0 SampleSize=0 Rect l,r,t,b=0,0,0,0
 : StreamHeader: Type=auds Handler=0x1 Flags=[ ]
 InitialFrames=0 Scale=1 Rate=44100 Start=0 Length=450559
 SuggestedBufferSize=8192 Quality=0 SampleSize=8 Rect l,r,t,b=0,0,0,0
 : Reading index from offset: 5581514
 : Stream 0 vids : DIVX (0x4944) 250 chunks (0.98KB)
 : WARNING: stream header has incorrect dwLength (450559 -> 
450560)

 : Stream 1 auds : Microsoft PCM (0x1) 440 chunks (3.44KB)
 : Initialized video stream (chunk tblsz: 250, fmtsz: 40)
 : Found 7 plugins 
(/usr/lib/arm-linux-gnueabihf/avifile-0.7,A:28,V:36)

 : creating dir: /home/debci/.avm
 : looking for mpeg4  1482049860
 : Created video decoder: FF OpenDivX
 : Initialized audio stream (chunk tblsz: 450560, fmtsz: 18)
 : PCM audio decoder created
bh.biHeight = -240, bh.biWidth = 320
bh.biHeight < 0, correct the value to 240
Movie size: 320 x 240  [YV12]
audio format: 1, Channels: 2, Samples/sec: 44100, Bits/Sample: 32, 
Bytes/sec: 352800

Video Length: 250
Video Pos: 0
Audio Length: 450560
Audio Pos: 0
Decoder YUV capabilities: 0x8080
CAPS is CAP_YV12
TIME 0 0
[mpeg4 @ 0x1ad7e40] Requested frame threading with a custom 
get_buffer2() implementation which is not marked as thread safe. This is 
not supported anymore, make your callback thread-safe.

 : using DR1
Audio read 192 samples, 1536 bytes.
Audio frequency 469.27 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.188 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.095 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.086 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.17 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.262 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.267 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.18 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.091 hz
Audio read 192 samples, 1536 bytes.
Audio frequency 469.09 hz
TIME 0.04 1
Bus error
autopkgtest [10:24:27]: test decoding-test-mp4-raw



OpenPGP_signature
Description: OpenPGP digital signature