[RFC PATCH v2 0/5] ASoC: soc-pcm: fix trigger race conditions with shared BE

[Pierre-Louis Bossart]

> I did a quick test of your patches on my Tegra board and seeing issues
> with multiple streams. For instance, I ran it for a 2x1 mixer
> configuration and hitting below:
> 
> [  277.661886] BUG: scheduling while atomic: aplay/1306/0x00000100
> [  287.713193] BUG: spinlock cpu recursion on CPU#0, aplay/1307
> [  287.719138]  lock: 0xffff00008cc820f0, .magic: dead4ead, .owner:
> aplay/1306, .owner_cpu: 0
> [  287.727319] CPU: 0 PID: 1307 Comm: aplay Tainted: G W        
> 5.15.0-rc3-next-20210927-00026-gffdabce987b1 #12
> [  287.737783] Hardware name: NVIDIA Jetson AGX Xavier Developer Kit (DT)
> [  287.744228] Call trace:
> [  287.746656]  dump_backtrace+0x0/0x1c0
> [  287.750300]  show_stack+0x18/0x28
> [  287.753604]  dump_stack_lvl+0x7c/0xa8
> [  287.757236]  dump_stack+0x18/0x34
> [  287.760536]  spin_dump+0x70/0x90
> [  287.763732]  do_raw_spin_lock+0xd8/0x120
> [  287.767615]  _raw_spin_lock_irq+0x60/0x80
> [  287.771581]  snd_pcm_stream_lock_irq+0x20/0x48 [snd_pcm]
> [  287.776853]  snd_pcm_drain+0x1ec/0x348 [snd_pcm]
> [  287.781421]  snd_pcm_common_ioctl+0xacc/0x1938 [snd_pcm]
> [  287.786685]  snd_pcm_ioctl+0x2c/0x48 [snd_pcm]
> [  287.791101]  __arm64_sys_ioctl+0xb0/0xf0
> [  287.794982]  invoke_syscall+0x44/0x108
> [  287.798683]  el0_svc_common.constprop.3+0x74/0x100
> [  287.803416]  do_el0_svc+0x24/0x90
> [  287.806687]  el0_svc+0x20/0x60
> [  287.809705]  el0t_64_sync_handler+0x94/0xb8
> [  287.813839]  el0t_64_sync+0x180/0x184
> 
> 
> And in some case just below:
> 
> [ 1074.212276] BUG: scheduling while atomic: aplay/12327/0x00000100
> [ 1095.227509] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
> [ 1095.233443] rcu:     0-...0: (1 GPs behind)
> idle=4af/1/0x4000000000000004 softirq=19902/19902 fqs=2626
> [ 1095.242528] rcu:     2-...0: (1 GPs behind)
> idle=9d5/1/0x4000000000000000 softirq=22707/22707 fqs=262

Thanks Sameer for the overnight tests, much appreciated.

My patches don't change anything related to a spinlock or pcm stream
management, so not sure what could cause this rather spectacular
failure. That hints at a fundamental configuration difference, possibly
caused by your component chaining?

Since in your case you have a 1:1 mapping between FE and BE, would you
mind testing by backtracking, one patch at a time to see which one of
the three last patches could cause a problem on your board?

Thanks again for your time!
-Pierre