Re: [syzbot] [fuse?] possible deadlock in __folio_end_writeback

[Hillf Danton]

> Date: Tue, 07 Oct 2025 14:19:29 -0700	[thread overview]
> Hello,
> 
> syzbot found the following issue on:
> 
> HEAD commit:    cbf33b8e0b36 Merge tag 'bpf-fixes' of git://git.kernel.org..
> git tree:       upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=17a25ee2580000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=1b4263e12240e6e1
> dashboard link: https://syzkaller.appspot.com/bug?extid=27727256237e6bdd3649
> compiler:       Debian clang version 20.1.8 (++20250708063551+0c9f909b7976-1~exp1~20250708183702.136), Debian LLD 20.1.8
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=14eaea7c580000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=134c4304580000

#syz test

--- x/lib/flex_proportions.c
+++ y/lib/flex_proportions.c
@@ -64,12 +64,14 @@ void fprop_global_destroy(struct fprop_g
 bool fprop_new_period(struct fprop_global *p, int periods)
 {
 	s64 events = percpu_counter_sum(&p->events);
+	unsigned long flags;
 
 	/*
 	 * Don't do anything if there are no events.
 	 */
 	if (events <= 1)
 		return false;
+	local_irq_save(flags);
 	preempt_disable_nested();
 	write_seqcount_begin(&p->sequence);
 	if (periods < 64)
@@ -79,6 +81,7 @@ bool fprop_new_period(struct fprop_globa
 	p->period += periods;
 	write_seqcount_end(&p->sequence);
 	preempt_enable_nested();
+	local_irq_restore(flags);
 
 	return true;
 }
--