Repository navigation
[BUG][WHL]Arecord hung after trigger successfully. #1692
Description
Describe the bug
Arecord will hang after trigger it.
To Reproduce
- Run command "arecord -Dhw:0,8 -M -N -r 16000 -c2 -f s24_le -vvv 1.wav" or with s16_le format
- Trigger it.
Reproduction Rate
80%
Expected behavior
Arecord successfully after trigger arecord.
Impact
Arecord hung after trigger it
Environment
Firmware: 442cfb2 https://github.com/thesofproject/sof/commits/cml-007-drop-stable
Kernel: f267c3d https://github.com/thesofproject/linux/commits/topic/keyword_detect
Topology:file: sof-cml-rt5682.tplg link from sof-whl-rt5682-kwd.tplg
platform: whl-i2s
logger
0 2 VOLUME 66817883.854167 4.427083 udio/volume/volume.c:405 volume_ctrl_set_cmd(), SOF_CTRL_CMD_VOLUME, channel = 0, value = 65536
0 2 VOLUME 66817888.854167 5.000000 udio/volume/volume.c:405 volume_ctrl_set_cmd(), SOF_CTRL_CMD_VOLUME, channel = 1, value = 65536
0 2 IPC 66818155.364583 266.510406 src/ipc/handler.c:628 ipc: pm -> restore
0 2 IPC 66819229.010417 1073.645874 src/ipc/handler.c:221 ipc: comp 34 -> params
0 2 PIPE 8.41 66819262.239583 33.229168 src/audio/pipeline.c:315 pipeline_params()
0 2 HOST 66819267.760417 5.520833 src/audio/host.c:502 host_params()
0 2 DMA 66819277.552083 9.791667 intel/cavs/hda-dma.c:462 hda-dmac: 4 channel 1 -> get
0 2 DMA 66819283.437500 5.885417 intel/cavs/hda-dma.c:687 hda-dmac: 4 channel 1 -> config
0 2 VOLUME 66819292.343750 8.906250 udio/volume/volume.c:247 volume_params()
0 2 DAI 8.40 66819297.968750 5.625000 src/audio/dai.c:356 dai_params()
0 2 PIPE 8.41 66819306.927083 8.958333 src/audio/pipeline.c:379 pipeline_prepare()
0 2 HOST 66819312.031250 5.104167 src/audio/host.c:632 host_prepare()
0 2 KPB 66819329.166667 17.135416 src/audio/kpb.c:338 kpb_prepare()
0 2 KPB 66819334.635417 5.468750 src/audio/kpb.c:178 kpb_allocate_history_buffer()
0 1 MEM 66820020.729167 686.093750 src/lib/alloc.c:255 error: 171 blocks needed for allocation but only 170 blocks are remaining
0 2 KPB 66820041.093750 20.364584 src/audio/kpb.c:204 kpb new memory block: 65280
0 2 KPB 66820101.250000 60.156250 src/audio/kpb.c:204 kpb new memory block: 203520
0 2 KPB 66820105.520833 4.270833 src/audio/kpb.c:251 kpb_allocate_history_buffer(): allocated 268800 bytes
0 2 KPB 66820110.260417 4.739583 src/audio/kpb.c:1137 kpb_clear_history_buffer()
0 2 VOLUME 66820692.083333 581.822937 udio/volume/volume.c:590 volume_prepare()
0 2 DAI 8.40 66820716.354167 24.270834 src/audio/dai.c:420 dai_prepare()
0 2 IPC 66821272.604167 556.250000 src/ipc/handler.c:221 ipc: comp 42 -> params
0 2 PIPE 9.47 66821295.833333 23.229166 src/audio/pipeline.c:315 pipeline_params()
0 2 SELECTOR 66821302.656250 6.822917 /selector/selector.c:164 selector_params()
0 2 PIPE 9.47 66821308.229167 5.572917 src/audio/pipeline.c:379 pipeline_prepare()
0 2 KEYWORD 66821313.072917 4.843750 /audio/detect_test.c:704 test_keyword_prepare()
0 2 SELECTOR 66821328.333333 15.260417 /selector/selector.c:350 selector_prepare()
0 2 SELECTOR 66821333.385417 5.052083 /selector/selector.c:379 selector_prepare(): source->params.channels = 2
Activity
mrajwa commentedon Jul 31, 2019
@YvonneYang2, logs look preatty fine. If that issue happens, you presumably terminate arecord and start again - now what happens? Are you able to trigger again and record with no problem or it no longer works.
PS: I can not reproduce this issue on neither CML nor WHL
YvonneYang2 commentedon Aug 1, 2019
Retest it again, when issue happened, terminate arecord and start again cannot recover it, will still reproduce the issue after trigger it.(logs please see again_xxx.log) Only reboot can recover it.
Logs as below
dmesg.log
logger.log
logger_error.log
again_dmesg.log
again_logger.log
again_logger_error.log
mrajwa commentedon Aug 1, 2019
Logs, again are clear. Can you try, before test run
and send logs again if they reveal sth new.
mrajwa commentedon Aug 27, 2019
@YvonneYang2 @keqiaozhang I think we can close this one?
keqiaozhang commentedon Aug 28, 2019
Confirmed that this issue is caused by a known issue with Host DMA(more like a hardware issue) and can be fixed by resetting or disabling the LI support("L1 Support Enable(L1SEN)" bit). So this is not a sof issue, let's close it.