Re: Problems running DAMON kunit tests with spinlock debugging enabled

[SeongJae Park <sj@kernel.org>]

Hi Guenter,

On Tue, 3 Sep 2024 12:23:50 -0700 Guenter Roeck <linux@roeck-us.net> wrote:

> Hi,
> 
> when trying to run DAMON kunit tests with spinlock debugging enabled,
> I get the following error messages.
> 
>     BUG: spinlock bad magic on CPU#0, kunit_try_catch/209
>      lock: mm.13+0x40/0x840, .magic: 00000000, .owner: <none>/-1, .owner_cpu: 0
>     CPU: 0 UID: 0 PID: 209 Comm: kunit_try_catch Tainted: G                 N 6.11.0-rc6-00148-g178297ec52d6 #1
>     Tainted: [N]=TEST
>     Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014
>     Call Trace:
>      <TASK>
>      dump_stack_lvl+0x9e/0xe0
>      do_raw_spin_lock+0x63/0xb0
>      damon_test_three_regions_in_vmas+0x121/0x450S
>      ...
> 
> This happens because damon_test_three_regions_in_vmas() calls
> mt_init_flags() with MM_MT_FLAGS, which sets MT_FLAGS_LOCK_EXTERN.
> Because of this, the spinlock is not initialized, which then results
> in the error message when mas_lock() is called during the test.
> 
> Unfortunately, replacing MM_MT_FLAGS with MT_FLAGS_ALLOC_RANGE |
> MT_FLAGS_USE_RCU or just with MT_FLAGS_ALLOC_RANGE doesn't help;
> it results in various "suspicious RCU usage" messages.

Thank you very much for this report with the grateful detailed investigation!
I was able to make and send a fix[1], thanks to the investigation.  Seems it
may need some more revisions[2], though.  I'll continue discussion on the
thread.

[1] https://lore.kernel.org/20240904004534.1189-1-sj@kernel.org
[2] https://lore.kernel.org/jy6263g6em4jsdhp6tknmh2cljpuvq652kvcet4ko3z2xt7pym@ltc5h5twsszu/


Thanks,
SJ

> 
> Thanks,
> Guenter