[syzbot] possible deadlock in blkcg_deactivate_policy

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit: b9b4c79e5830 Merge tag 'sound-5.18-rc3' of git://git.kerne..
git tree: upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=118c96ccf00000
kernel config: https://syzkaller.appspot.com/x/.config?x=eb177500e563582f
dashboard link: https://syzkaller.appspot.com/bug?extid=b9e4c31d0a1efb34cb03
compiler: gcc (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.2
syz repro: https://syzkaller.appspot.com/x/repro.syz?x=16ab7a70f00000
C reproducer: https://syzkaller.appspot.com/x/repro.c?x=16e54770f00000

The issue was bisected to:

commit 0a9a25ca78437b39e691bcc3dc8240455b803d8d
Author: Ming Lei <ming...@redhat.com>
Date: Fri Mar 18 13:01:43 2022 +0000

block: let blkcg_gq grab request queue's refcnt

bisection log: https://syzkaller.appspot.com/x/bisect.txt?x=128129e0f00000
final oops: https://syzkaller.appspot.com/x/report.txt?x=118129e0f00000
console output: https://syzkaller.appspot.com/x/log.txt?x=168129e0f00000

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+b9e4c3...@syzkaller.appspotmail.com
Fixes: 0a9a25ca7843 ("block: let blkcg_gq grab request queue's refcnt")

RBP: 00007ffcdacd3400 R08: 0000000000000002 R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000004
R13: 431bde82d7b634db R14: 0000000000000000 R15: 0000000000000000
</TASK>
======================================================
WARNING: possible circular locking dependency detected
5.18.0-rc2-syzkaller-00122-gb9b4c79e5830 #0 Not tainted
------------------------------------------------------
syz-executor173/3613 is trying to acquire lock:
ffff8880168900a8 ((&sq->pending_timer)){+.-.}-{0:0}, at: del_timer_sync+0x33/0x1b0 kernel/time/timer.c:1363

but task is already holding lock:
ffffffff9070b710 (&blkcg->lock){....}-{2:2}, at: spin_lock include/linux/spinlock.h:349 [inline]
ffffffff9070b710 (&blkcg->lock){....}-{2:2}, at: blkcg_deactivate_policy block/blk-cgroup.c:1441 [inline]
ffffffff9070b710 (&blkcg->lock){....}-{2:2}, at: blkcg_deactivate_policy+0x1e7/0x4e0 block/blk-cgroup.c:1423

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #2 (&blkcg->lock){....}-{2:2}:
__raw_spin_lock include/linux/spinlock_api_smp.h:133 [inline]
_raw_spin_lock+0x2a/0x40 kernel/locking/spinlock.c:154
spin_lock include/linux/spinlock.h:349 [inline]
blkg_create+0x47b/0x1030 block/blk-cgroup.c:301
blkcg_init_queue+0xfe/0x810 block/blk-cgroup.c:1206
__alloc_disk_node+0x260/0x610 block/genhd.c:1381
__blk_alloc_disk+0x35/0x70 block/genhd.c:1421
brd_alloc.part.0+0x27f/0x7a0 drivers/block/brd.c:391
brd_alloc drivers/block/brd.c:374 [inline]
brd_init+0x1b8/0x24b drivers/block/brd.c:477
do_one_initcall+0x103/0x650 init/main.c:1298
do_initcall_level init/main.c:1371 [inline]
do_initcalls init/main.c:1387 [inline]
do_basic_setup init/main.c:1406 [inline]
kernel_init_freeable+0x6b1/0x73a init/main.c:1613
kernel_init+0x1a/0x1d0 init/main.c:1502
ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:298

-> #1 (&q->queue_lock){..-.}-{2:2}:
__raw_spin_lock_irq include/linux/spinlock_api_smp.h:119 [inline]
_raw_spin_lock_irq+0x32/0x50 kernel/locking/spinlock.c:170
spin_lock_irq include/linux/spinlock.h:374 [inline]
throtl_pending_timer_fn+0xf7/0x1690 block/blk-throttle.c:1152
call_timer_fn+0x1a5/0x6b0 kernel/time/timer.c:1421
expire_timers kernel/time/timer.c:1466 [inline]
__run_timers.part.0+0x67c/0xa30 kernel/time/timer.c:1734
__run_timers kernel/time/timer.c:1715 [inline]
run_timer_softirq+0xb3/0x1d0 kernel/time/timer.c:1747
__do_softirq+0x29b/0x9c2 kernel/softirq.c:558
invoke_softirq kernel/softirq.c:432 [inline]
__irq_exit_rcu+0x123/0x180 kernel/softirq.c:637
irq_exit_rcu+0x5/0x20 kernel/softirq.c:649
sysvec_apic_timer_interrupt+0x93/0xc0 arch/x86/kernel/apic/apic.c:1097
asm_sysvec_apic_timer_interrupt+0x12/0x20 arch/x86/include/asm/idtentry.h:645
native_safe_halt arch/x86/include/asm/irqflags.h:51 [inline]
arch_safe_halt arch/x86/include/asm/irqflags.h:89 [inline]
acpi_safe_halt drivers/acpi/processor_idle.c:115 [inline]
acpi_idle_do_entry+0x1c6/0x250 drivers/acpi/processor_idle.c:556
acpi_idle_enter+0x361/0x500 drivers/acpi/processor_idle.c:691
cpuidle_enter_state+0x1b1/0xc80 drivers/cpuidle/cpuidle.c:237
cpuidle_enter+0x4a/0xa0 drivers/cpuidle/cpuidle.c:351
call_cpuidle kernel/sched/idle.c:155 [inline]
cpuidle_idle_call kernel/sched/idle.c:236 [inline]
do_idle+0x3e8/0x590 kernel/sched/idle.c:303
cpu_startup_entry+0x14/0x20 kernel/sched/idle.c:400
start_secondary+0x265/0x340 arch/x86/kernel/smpboot.c:266
secondary_startup_64_no_verify+0xc3/0xcb

-> #0 ((&sq->pending_timer)){+.-.}-{0:0}:
check_prev_add kernel/locking/lockdep.c:3065 [inline]
check_prevs_add kernel/locking/lockdep.c:3188 [inline]
validate_chain kernel/locking/lockdep.c:3803 [inline]
__lock_acquire+0x2ac6/0x56c0 kernel/locking/lockdep.c:5029
lock_acquire kernel/locking/lockdep.c:5641 [inline]
lock_acquire+0x1ab/0x510 kernel/locking/lockdep.c:5606
del_timer_sync+0x5b/0x1b0 kernel/time/timer.c:1364
throtl_pd_free+0x15/0x40 block/blk-throttle.c:500
blkcg_deactivate_policy block/blk-cgroup.c:1445 [inline]
blkcg_deactivate_policy+0x2d2/0x4e0 block/blk-cgroup.c:1423
blk_throtl_exit+0x8a/0x1a0 block/blk-throttle.c:2335
blkcg_init_queue+0x225/0x810 block/blk-cgroup.c:1226
__alloc_disk_node+0x260/0x610 block/genhd.c:1381
__blk_mq_alloc_disk+0x133/0x1c0 block/blk-mq.c:3930
loop_add+0x33d/0x910 drivers/block/loop.c:2012
loop_control_ioctl+0x130/0x4d0 drivers/block/loop.c:2186
vfs_ioctl fs/ioctl.c:51 [inline]
__do_sys_ioctl fs/ioctl.c:870 [inline]
__se_sys_ioctl fs/ioctl.c:856 [inline]
__x64_sys_ioctl+0x193/0x200 fs/ioctl.c:856
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x44/0xae

other info that might help us debug this:

Chain exists of:
(&sq->pending_timer) --> &q->queue_lock --> &blkcg->lock

Possible unsafe locking scenario:

CPU0 CPU1
---- ----
lock(&blkcg->lock);
lock(&q->queue_lock);
lock(&blkcg->lock);
lock((&sq->pending_timer));

*** DEADLOCK ***

2 locks held by syz-executor173/3613:
#0: ffff88801dd168e0 (&q->queue_lock){..-.}-{2:2}, at: spin_lock_irq include/linux/spinlock.h:374 [inline]
#0: ffff88801dd168e0 (&q->queue_lock){..-.}-{2:2}, at: blkcg_deactivate_policy block/blk-cgroup.c:1434 [inline]
#0: ffff88801dd168e0 (&q->queue_lock){..-.}-{2:2}, at: blkcg_deactivate_policy+0xfe/0x4e0 block/blk-cgroup.c:1423
#1: ffffffff9070b710 (&blkcg->lock){....}-{2:2}, at: spin_lock include/linux/spinlock.h:349 [inline]
#1: ffffffff9070b710 (&blkcg->lock){....}-{2:2}, at: blkcg_deactivate_policy block/blk-cgroup.c:1441 [inline]
#1: ffffffff9070b710 (&blkcg->lock){....}-{2:2}, at: blkcg_deactivate_policy+0x1e7/0x4e0 block/blk-cgroup.c:1423

stack backtrace:
CPU: 1 PID: 3613 Comm: syz-executor173 Not tainted 5.18.0-rc2-syzkaller-00122-gb9b4c79e5830 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:88 [inline]
dump_stack_lvl+0xcd/0x134 lib/dump_stack.c:106
check_noncircular+0x25f/0x2e0 kernel/locking/lockdep.c:2145
check_prev_add kernel/locking/lockdep.c:3065 [inline]
check_prevs_add kernel/locking/lockdep.c:3188 [inline]
validate_chain kernel/locking/lockdep.c:3803 [inline]
__lock_acquire+0x2ac6/0x56c0 kernel/locking/lockdep.c:5029
lock_acquire kernel/locking/lockdep.c:5641 [inline]
lock_acquire+0x1ab/0x510 kernel/locking/lockdep.c:5606
del_timer_sync+0x5b/0x1b0 kernel/time/timer.c:1364
throtl_pd_free+0x15/0x40 block/blk-throttle.c:500
blkcg_deactivate_policy block/blk-cgroup.c:1445 [inline]
blkcg_deactivate_policy+0x2d2/0x4e0 block/blk-cgroup.c:1423
blk_throtl_exit+0x8a/0x1a0 block/blk-throttle.c:2335
blkcg_init_queue+0x225/0x810 block/blk-cgroup.c:1226
__alloc_disk_node+0x260/0x610 block/genhd.c:1381
__blk_mq_alloc_disk+0x133/0x1c0 block/blk-mq.c:3930
loop_add+0x33d/0x910 drivers/block/loop.c:2012
loop_control_ioctl+0x130/0x4d0 drivers/block/loop.c:2186
vfs_ioctl fs/ioctl.c:51 [inline]
__do_sys_ioctl fs/ioctl.c:870 [inline]
__se_sys_ioctl fs/ioctl.c:856 [inline]
__x64_sys_ioctl+0x193/0x200 fs/ioctl.c:856
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x44/0xae
RIP: 0033:0x7fde66153079
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 b1 14 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007ffcdacd33e8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
RAX: ffffffffffffffda RBX: 0000000000000002 RCX: 00007fde66153079
RDX: 0000000000000000 RSI: 0000000000004c80 RDI: 0000000000000003
RBP: 00007ffcdacd3400 R08: 0000000000000002 R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000004
R13: 431bde82d7b634db R14: 0000000000000000 R15: 0000000000000000
</TASK>


---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzk...@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
For information about bisection process see: https://goo.gl/tpsmEJ#bisection
syzbot can test patches for this issue, for details see:
https://goo.gl/tpsmEJ#testing-patches

[Hillf Danton]

On Fri, 15 Apr 2022 11:29:33 -0700

To fix the deadlock, free throttle group in workqueue with no lock held.

#syz test https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/ b9b4c79e5830

diff -pur a/block/blk-throttle.c b/block/blk-throttle.c
--- a/block/blk-throttle.c 2022-04-16 06:56:37.857427700 +0800
+++ b/block/blk-throttle.c 2022-04-16 07:15:44.803445700 +0800
@@ -493,9 +493,9 @@ static void throtl_pd_offline(struct blk
throtl_upgrade_state(tg->td);
}

-static void throtl_pd_free(struct blkg_policy_data *pd)
+static void throtl_grp_free_workfn(struct work_struct *work)
{
- struct throtl_grp *tg = pd_to_tg(pd);
+ struct throtl_grp *tg = container_of(work, struct throtl_grp, free_work);

del_timer_sync(&tg->service_queue.pending_timer);
blkg_rwstat_exit(&tg->stat_bytes);
@@ -503,6 +503,14 @@ static void throtl_pd_free(struct blkg_p
kfree(tg);
}

+static void throtl_pd_free(struct blkg_policy_data *pd)
+{
+ struct throtl_grp *tg = pd_to_tg(pd);
+
+ INIT_WORK(&tg->free_work, throtl_grp_free_workfn);
+ queue_work(system_unbound_wq, &tg->free_work);
+}
+
static struct throtl_grp *
throtl_rb_first(struct throtl_service_queue *parent_sq)
{
diff -pur a/block/blk-throttle.h b/block/blk-throttle.h
--- a/block/blk-throttle.h 2022-04-16 06:56:24.885706800 +0800
+++ b/block/blk-throttle.h 2022-04-16 07:06:21.260907700 +0800
@@ -141,6 +141,8 @@ struct throtl_grp {

struct blkg_rwstat stat_bytes;
struct blkg_rwstat stat_ios;
+
+ struct work_struct free_work;
};

extern struct blkcg_policy blkcg_policy_throtl;
--

[syzbot]

Hello,

syzbot has tested the proposed patch and the reproducer did not trigger any issue:

Reported-and-tested-by: syzbot+b9e4c3...@syzkaller.appspotmail.com

Tested on:

commit: b9b4c79e Merge tag 'sound-5.18-rc3' of git://git.kerne..
git tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/

kernel config: https://syzkaller.appspot.com/x/.config?x=eb177500e563582f
dashboard link: https://syzkaller.appspot.com/bug?extid=b9e4c31d0a1efb34cb03
compiler: gcc (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.2

patch: https://syzkaller.appspot.com/x/patch.diff?x=11b965c0f00000

Note: testing is done by a robot and is best-effort only.

[syzbot]

Auto-closing this bug as obsolete.
No recent activity, existing reproducers are no longer triggering the issue.