2412306_en-US

[#1]

Bug: iMX8QM: Audio over HDMI Creates a Kernel Panic When Audio Stops

Hello,

This is a follow up from the thread Max created back then.

It seems that the patch he provided was at the end not applied as it's not present in lf-6.18.y.

Currently, there is a kernel Panic when Audio stops.

root@apalis-imx8-15304854:~# aplay -D sysdefault:CARD=imxaudiohdmitx sound/Gong.wav          
Playing WAVE 'sound/Gong.wav' : [  100.599725] [drm] pclkc 533250 is not supported!
Signed 16 bit Little Endian, Rate 44100 Hz, Stereo
^CAborted by signal Interrupt...
[  102.674593] SError Interrupt on CPU3, code 0x00000000bf000002 -- SError
[  102.674642] CPU: 3 UID: 0 PID: 630 Comm: aplay Tainted: G   M                6.18.39+ #115
 PREEMPT 
[  102.674664] Tainted: [M]=MACHINE_CHECK
[  102.674671] Hardware name: Toradex Apalis iMX8QM V1.1 on Apalis Evaluation Board V1.2 (DT)
[  102.674680] pstate: 20000005 (nzCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[  102.674694] pc : cdns_mhdp_bus_write+0x0/0xe4 [cdns_mhdp_drmcore]
[  102.674744] lr : cdns_mhdp_audio_stop+0x40/0x180 [cdns_mhdp_drmcore]
[  102.674769] sp : ffff80009290bc70
[  102.674774] x29: ffff80009290bc70 x28: ffff00080e893240 x27: 0000000000000000
[  102.674801] x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000
[  102.674818] x23: 0000000000000000 x22: ffff0008090f1880 x21: ffff000801114a00
[  102.674837] x20: ffff000809a34b78 x19: ffff000809a34080 x18: 0000000000000000
[  102.674856] x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000
[  102.674874] x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000
[  102.674892] x11: 0000000000000000 x10: 0000000000000000 x9 : 0000000000000000
[  102.674910] x8 : ffff80009290bd78 x7 : 0000000000000000 x6 : 0000000000000000
[  102.674927] x5 : ffff00080905a880 x4 : ffff00080905a918 x3 : ffff000801150e80
[  102.674946] x2 : 000000000003004c x1 : ffff000809a34080 x0 : 0000000000000000
[  102.674968] Kernel panic - not syncing: Asynchronous SError Interrupt
[  102.674978] CPU: 3 UID: 0 PID: 630 Comm: aplay Tainted: G   M                6.18.39+ #115
 PREEMPT 
[  102.674995] Tainted: [M]=MACHINE_CHECK
[  102.675000] Hardware name: Toradex Apalis iMX8QM V1.1 on Apalis Evaluation Board V1.2 (DT)
[  102.675009] Call trace:
[  102.675016]  show_stack+0x28/0x7c (C)
[  102.675046]  dump_stack_lvl+0x60/0x80
[  102.675070]  dump_stack+0x18/0x24
[  102.675088]  vpanic+0xf8/0x2b4
[  102.675106]  abort+0x0/0x4
[  102.675120]  add_taint+0x0/0xbc
[  102.675145]  arm64_serror_panic+0x70/0x80
[  102.675166]  do_serror+0x34/0x74
[  102.675184]  el1h_64_error_handler+0x38/0x60
[  102.675206]  el1h_64_error+0x6c/0x70
[  102.675222]  cdns_mhdp_bus_write+0x0/0xe4 [cdns_mhdp_drmcore] (P)
[  102.675250]  audio_shutdown+0x20/0x40 [cdns_mhdp_drmcore]
[  102.675276]  hdmi_codec_shutdown+0x58/0x7c [snd_soc_hdmi_codec]
[  102.675302]  snd_soc_dai_shutdown+0x54/0x80
[  102.675327]  soc_pcm_clean.isra.0+0x5c/0x1c8
[  102.675345]  soc_pcm_close+0x34/0x58
[  102.675360]  snd_pcm_release_substream.part.0+0x3c/0xc0
[  102.675378]  snd_pcm_release+0x58/0xe0
[  102.675392]  __fput+0xc8/0x2cc
[  102.675408]  fput_close_sync+0x38/0x114
[  102.675423]  __arm64_sys_close+0x38/0x7c
[  102.675443]  invoke_syscall.constprop.0+0x48/0xc0
[  102.675464]  do_el0_svc+0x3c/0xc0
[  102.675480]  el0_svc+0x40/0x170
[  102.675495]  el0t_64_sync_handler+0xa0/0xe4
[  102.675511]  el0t_64_sync+0x198/0x19c
[  102.675529] SMP: stopping secondary CPUs
[  102.675547] Kernel Offset: 0x4f62f0a00000 from 0xffff800080000000
[  102.675555] PHYS_OFFSET: 0xfff1000080000000
[  102.675560] CPU features: 0x100000,00025800,40004001,0400420b
[  102.675571] Memory Limit: none
[  102.952968] ---[ end Kernel panic - not syncing: Asynchronous SError Interrupt ]---

This bug has been introduced with https://github.com/nxp-imx/linux-imx/commit/bbdf7c86aa79c9ead532a2779920a0ac68db4863

Can the fix be included to the next BSP release?

@Zhiming_Liu 

Kind regards

Franz Schnyder

Re: Bug: iMX8QM: Audio over HDMI Creates a Kernel Panic When Audio Stops

Hi,

Already follow up the internal case, wait for reply.

Best Regards,
Zhiming