On (23/04/04 09:37), Tetsuo Handa wrote: > syzbot is reporting circular locking dependency which involves > zonelist_update_seq seqlock [1], for this lock is checked by memory > allocation requests which do not need to be retried. > > We somehow need to prevent __alloc_pages_slowpath() from checking > this lock. Since Petr Mladek thinks that __build_all_zonelists() can > become a candidate for deferring printk() [2], let's make sure that > current CPU/thread won't reach __alloc_pages_slowpath() while this lock > is in use. > > Reported-by: syzbot <syzbot+223c7461c58c58a4cb10@xxxxxxxxxxxxxxxxxxxxxxxxx> > Link: https://syzkaller.appspot.com/bug?extid=223c7461c58c58a4cb10 [1] > Fixes: 3d36424b3b58 ("mm/page_alloc: fix race condition between build_all_zonelists and page allocation") > Link: https://lkml.kernel.org/r/ZCrs+1cDqPWTDFNM@alley [2] > Signed-off-by: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx> > Cc: Michal Hocko <mhocko@xxxxxxxx> > Cc: Petr Mladek <pmladek@xxxxxxxx> Yeah that looks like one of those printk_deferred() cases. FWIW Reviewed-by: Sergey Senozhatsky <senozhatsky@xxxxxxxxxxxx>
