Re: [PATCH] nd_btt: Make BTT lanes preemptible

[Ira Weiny]

Tomas Glozar wrote:
> čt 14. 9. 2023 v 22:18 odesílatel Ira Weiny <ira.weiny@xxxxxxxxx> napsal:
> > Is the bug in 1 of 2 places?
> >
> > 1) When btt_write_pg()->lock_map() (when the number of lanes is < number
> >    of cpus) and the lane is acquired is called?
> >
> > *or*
> >
> > 2) When nd_region_acquire_lane() internally trys to take it's lock?
> >
> > A copy/paste of the BUG observed would have been more clear I think.
> >
> 
> The BUG was observed on btt_write_pg()->lock_map(), but I assume the
> BUG will also happen on the lock in nd_region_acquire_lane, since that
> is also a spin lock, i.e. a sleeping lock on RT.
> 
> BUG observed in dmesg when running ndctl tests on RT kernel without the patch:

Thanks for clarifying.  Could you respin the patch with the text below?
That would have saved me a lot of time digging to see what the code path
was.

...
	BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48
	in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 4903, name: libndctl
	preempt_count: 1, expected: 0
	RCU nest depth: 0, expected: 0
	1 lock held by libndctl/4903:
	 #0: ffff8c184a270060 (&arena->map_locks[i].lock){+.+.}-{2:2}, at: btt_write_pg+0x2d7/0x500 [nd_btt]
	Preemption disabled at:
	[<ffffffffc1313db5>] nd_region_acquire_lane+0x15/0x90 [libnvdimm]
	Call Trace:
	 <TASK>
	 dump_stack_lvl+0x8e/0xb0
	 __might_resched+0x19b/0x250
	 rt_spin_lock+0x4c/0x100
	 ? btt_write_pg+0x2d7/0x500 [nd_btt]
	 btt_write_pg+0x2d7/0x500 [nd_btt]
	 ? local_clock_noinstr+0x9/0xc0
	 btt_submit_bio+0x16d/0x270 [nd_btt]
	 __submit_bio+0x48/0x80
	 __submit_bio_noacct+0x7e/0x1e0
	 submit_bio_wait+0x58/0xb0
	 __blkdev_direct_IO_simple+0x107/0x240
	 ? inode_set_ctime_current+0x51/0x110
	 ? __pfx_submit_bio_wait_endio+0x10/0x10
	 blkdev_write_iter+0x1d8/0x290
	 vfs_write+0x237/0x330
...

With a respin including this trace:

Reviewed-by: Ira Weiny <ira.weiny@xxxxxxxxx>