Re: [syzbot] BUG: MAX_LOCK_DEPTH too low! (3)

On 2022/2/9 4:21 am, syzbot wrote:

The issue was bisected to:

commit 341adeec9adad0874f29a0a1af35638207352a39
Author: Wen Gu [off-list ref]
Date:   Wed Jan 26 15:33:04 2022 +0000

     net/smc: Forward wakeup to smc socket waitqueue after fallback

bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=11c2637c700000
final oops:     https://syzkaller.appspot.com/x/report.txt?x=13c2637c700000
console output: https://syzkaller.appspot.com/x/log.txt?x=15c2637c700000

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+4de3c0e8a263e1e499bc@syzkaller.appspotmail.com
Fixes: 341adeec9ada ("net/smc: Forward wakeup to smc socket waitqueue after fallback")

Thanks for all the details provided by syzbot.

I reproduced this issue in my environment. It is caused by repeated calls to
smc_switch_to_fallback().

In 341adeec9ada ("net/smc: Forward wakeup to smc socket waitqueue after fallback"),
smc_switch_to_fallback() saves the original callback function of clcsock in
smc->clcsk_error_report and set clcsk->sk_error_report as smc_fback_error_report().

If smc_switch_to_fallback() is called repeatedly, the smc->clcsk_error_report will be
reset as clcsk->sk_error_report, which now is smc_fback_error_report().

And the call trace will be:

clcsk->sk_error_report
   |- smc_fback_error_report() <----------------|
        |- smc_fback_forward_wakeup()           |
             |- clcsock_callback()              |
                  |- smc->clcsk_error_report() -|

Thus resulting in this issue.

I will send a patch to fix it.