Re: [syzbot] [bpf?] possible deadlock in kvfree_call_rcu

From: syzbot
Date: Wed Mar 27 2024 - 19:27:35 EST

Next message: Caleb Connolly: "Re: [PATCH] soc: qcom: cmd-db: map shared memory as WT, not WB"
Previous message: Paul Moore: "Re: [PATCH v2 2/2] selinux: improve symtab string hashing"
In reply to: Alexei Starovoitov: "Re: false positive deadlock? Was: [syzbot] [bpf?] possible deadlock in kvfree_call_rcu"
Next in thread: Hillf Danton: "Re: [syzbot] [bpf?] possible deadlock in kvfree_call_rcu"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

syzbot has found a reproducer for the following issue on:

HEAD commit: 443574b03387 riscv, bpf: Fix kfunc parameters incompatibil..
git tree: bpf
console+strace: https://syzkaller.appspot.com/x/log.txt?x=146398f9180000
kernel config: https://syzkaller.appspot.com/x/.config?x=6fb1be60a193d440
dashboard link: https://syzkaller.appspot.com/bug?extid=1fa663a2100308ab6eab
compiler: Debian clang version 15.0.6, GNU ld (GNU Binutils for Debian) 2.40
syz repro: https://syzkaller.appspot.com/x/repro.syz?x=12055cc6180000
C reproducer: https://syzkaller.appspot.com/x/repro.c?x=16ca53c9180000

Downloadable assets:
disk image: https://storage.googleapis.com/syzbot-assets/3f355021a085/disk-443574b0.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/44cf4de7472a/vmlinux-443574b0.xz
kernel image: https://storage.googleapis.com/syzbot-assets/a99a36c7ad65/bzImage-443574b0.xz

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+1fa663a2100308ab6eab@xxxxxxxxxxxxxxxxxxxxxxxxx

======================================================
WARNING: possible circular locking dependency detected
6.8.0-syzkaller-05236-g443574b03387 #0 Not tainted
------------------------------------------------------
syz-executor271/5074 is trying to acquire lock:
ffff8880b9529470 (krc.lock){....}-{2:2}, at: krc_this_cpu_lock kernel/rcu/tree.c:2960 [inline]
ffff8880b9529470 (krc.lock){....}-{2:2}, at: add_ptr_to_bulk_krc_lock kernel/rcu/tree.c:3359 [inline]
ffff8880b9529470 (krc.lock){....}-{2:2}, at: kvfree_call_rcu+0x18a/0x790 kernel/rcu/tree.c:3444

but task is already holding lock:
ffff888029e171f8 (&trie->lock){-...}-{2:2}, at: trie_update_elem+0xcb/0xc10 kernel/bpf/lpm_trie.c:324

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #2 (&trie->lock){-...}-{2:2}:
lock_acquire+0x1e4/0x530 kernel/locking/lockdep.c:5754
__raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
_raw_spin_lock_irqsave+0xd5/0x120 kernel/locking/spinlock.c:162
trie_delete_elem+0x96/0x6a0 kernel/bpf/lpm_trie.c:451
bpf_prog_2c29ac5cdc6b1842+0x42/0x46
bpf_dispatcher_nop_func include/linux/bpf.h:1234 [inline]
__bpf_prog_run include/linux/filter.h:657 [inline]
bpf_prog_run include/linux/filter.h:664 [inline]
__bpf_trace_run kernel/trace/bpf_trace.c:2381 [inline]
bpf_trace_run2+0x204/0x420 kernel/trace/bpf_trace.c:2420
trace_hrtimer_start include/trace/events/timer.h:222 [inline]
debug_activate kernel/time/hrtimer.c:479 [inline]
enqueue_hrtimer+0x335/0x3a0 kernel/time/hrtimer.c:1090
__hrtimer_start_range_ns kernel/time/hrtimer.c:1265 [inline]
hrtimer_start_range_ns+0xaa0/0xc60 kernel/time/hrtimer.c:1305
hrtimer_start_expires include/linux/hrtimer.h:289 [inline]
hrtimer_sleeper_start_expires kernel/time/hrtimer.c:1972 [inline]
schedule_hrtimeout_range_clock+0x277/0x480 kernel/time/hrtimer.c:2309
poll_schedule_timeout fs/select.c:244 [inline]
do_poll fs/select.c:965 [inline]
do_sys_poll+0xe40/0x1330 fs/select.c:1016
__do_sys_ppoll fs/select.c:1122 [inline]
__se_sys_ppoll+0x2a0/0x330 fs/select.c:1102
do_syscall_64+0xfb/0x240
entry_SYSCALL_64_after_hwframe+0x6d/0x75

-> #1 (hrtimer_bases.lock){-.-.}-{2:2}:
lock_acquire+0x1e4/0x530 kernel/locking/lockdep.c:5754
__raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
_raw_spin_lock_irqsave+0xd5/0x120 kernel/locking/spinlock.c:162
lock_hrtimer_base kernel/time/hrtimer.c:175 [inline]
hrtimer_start_range_ns+0xdf/0xc60 kernel/time/hrtimer.c:1303
hrtimer_start include/linux/hrtimer.h:275 [inline]
run_page_cache_worker kernel/rcu/tree.c:3341 [inline]
kvfree_call_rcu+0x5e6/0x790 kernel/rcu/tree.c:3446
rtnl_register_internal+0x482/0x590 net/core/rtnetlink.c:265
rtnl_register+0x36/0x80 net/core/rtnetlink.c:315
ip_rt_init+0x2f5/0x3a0 net/ipv4/route.c:3719
ip_init+0xe/0x20 net/ipv4/ip_output.c:1664
inet_init+0x3d8/0x580 net/ipv4/af_inet.c:2022
do_one_initcall+0x238/0x830 init/main.c:1241
do_initcall_level+0x157/0x210 init/main.c:1303
do_initcalls+0x3f/0x80 init/main.c:1319
kernel_init_freeable+0x435/0x5d0 init/main.c:1557
kernel_init+0x1d/0x2a0 init/main.c:1446
ret_from_fork+0x4b/0x80 arch/x86/kernel/process.c:147
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:243

-> #0 (krc.lock){....}-{2:2}:
check_prev_add kernel/locking/lockdep.c:3134 [inline]
check_prevs_add kernel/locking/lockdep.c:3253 [inline]
validate_chain+0x18cb/0x58e0 kernel/locking/lockdep.c:3869
__lock_acquire+0x1346/0x1fd0 kernel/locking/lockdep.c:5137
lock_acquire+0x1e4/0x530 kernel/locking/lockdep.c:5754
__raw_spin_lock include/linux/spinlock_api_smp.h:133 [inline]
_raw_spin_lock+0x2e/0x40 kernel/locking/spinlock.c:154
krc_this_cpu_lock kernel/rcu/tree.c:2960 [inline]
add_ptr_to_bulk_krc_lock kernel/rcu/tree.c:3359 [inline]
kvfree_call_rcu+0x18a/0x790 kernel/rcu/tree.c:3444
trie_update_elem+0x819/0xc10 kernel/bpf/lpm_trie.c:385
bpf_map_update_value+0x4d3/0x540 kernel/bpf/syscall.c:203
generic_map_update_batch+0x60d/0x900 kernel/bpf/syscall.c:1876
bpf_map_do_batch+0x3e0/0x690 kernel/bpf/syscall.c:5145
__sys_bpf+0x377/0x810
__do_sys_bpf kernel/bpf/syscall.c:5738 [inline]
__se_sys_bpf kernel/bpf/syscall.c:5736 [inline]
__x64_sys_bpf+0x7c/0x90 kernel/bpf/syscall.c:5736
do_syscall_64+0xfb/0x240
entry_SYSCALL_64_after_hwframe+0x6d/0x75

other info that might help us debug this:

Chain exists of:
krc.lock --> hrtimer_bases.lock --> &trie->lock

Possible unsafe locking scenario:

CPU0 CPU1
---- ----
lock(&trie->lock);
lock(hrtimer_bases.lock);
lock(&trie->lock);
lock(krc.lock);

*** DEADLOCK ***

2 locks held by syz-executor271/5074:
#0: ffffffff8e131920 (rcu_read_lock){....}-{1:2}, at: rcu_lock_acquire include/linux/rcupdate.h:298 [inline]
#0: ffffffff8e131920 (rcu_read_lock){....}-{1:2}, at: rcu_read_lock include/linux/rcupdate.h:750 [inline]
#0: ffffffff8e131920 (rcu_read_lock){....}-{1:2}, at: bpf_map_update_value+0x3c4/0x540 kernel/bpf/syscall.c:202
#1: ffff888029e171f8 (&trie->lock){-...}-{2:2}, at: trie_update_elem+0xcb/0xc10 kernel/bpf/lpm_trie.c:324

stack backtrace:
CPU: 1 PID: 5074 Comm: syz-executor271 Not tainted 6.8.0-syzkaller-05236-g443574b03387 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 02/29/2024
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:88 [inline]
dump_stack_lvl+0x1e7/0x2e0 lib/dump_stack.c:106
check_noncircular+0x36a/0x4a0 kernel/locking/lockdep.c:2187
check_prev_add kernel/locking/lockdep.c:3134 [inline]
check_prevs_add kernel/locking/lockdep.c:3253 [inline]
validate_chain+0x18cb/0x58e0 kernel/locking/lockdep.c:3869
__lock_acquire+0x1346/0x1fd0 kernel/locking/lockdep.c:5137
lock_acquire+0x1e4/0x530 kernel/locking/lockdep.c:5754
__raw_spin_lock include/linux/spinlock_api_smp.h:133 [inline]
_raw_spin_lock+0x2e/0x40 kernel/locking/spinlock.c:154
krc_this_cpu_lock kernel/rcu/tree.c:2960 [inline]
add_ptr_to_bulk_krc_lock kernel/rcu/tree.c:3359 [inline]
kvfree_call_rcu+0x18a/0x790 kernel/rcu/tree.c:3444
trie_update_elem+0x819/0xc10 kernel/bpf/lpm_trie.c:385
bpf_map_update_value+0x4d3/0x540 kernel/bpf/syscall.c:203
generic_map_update_batch+0x60d/0x900 kernel/bpf/syscall.c:1876
bpf_map_do_batch+0x3e0/0x690 kernel/bpf/syscall.c:5145
__sys_bpf+0x377/0x810
__do_sys_bpf kernel/bpf/syscall.c:5738 [inline]
__se_sys_bpf kernel/bpf/syscall.c:5736 [inline]
__x64_sys_bpf+0x7c/0x90 kernel/bpf/syscall.c:5736
do_syscall_64+0xfb/0x240
entry_SYSCALL_64_after_hwframe+0x6d/0x75
RIP: 0033:0x7f485af3c8e9
Code: 48 83 c4 28 c3 e8 37 17 00 00 0f 1f 80 00 00 00 00 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 b8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007ffc5f419378 EFLAGS: 00000246 ORIG_RAX: 0000000000000141
RAX: ffffffffffffffda RBX: 00007ffc5f419548 RCX: 00007f485af3c8e9
RDX: 0000000000000038 RSI: 0000000020000000 RDI: 000000000000001a
RBP: 00007f485afaf610 R08: 00007ffc5f419548 R09: 00007ffc5f419548
R10: 00007ffc5f419548 R11: 0000000000000246 R12: 0000000000000001
R13: 00007ffc5f419538 R14: 0000000000000001 R15: 0000000000000001
</TASK>

---
If you want syzbot to run the reproducer, reply with:
#syz test: git://repo/address.git branch-or-commit-hash
If you attach or paste a git patch, syzbot will apply it before testing.

Next message: Caleb Connolly: "Re: [PATCH] soc: qcom: cmd-db: map shared memory as WT, not WB"
Previous message: Paul Moore: "Re: [PATCH v2 2/2] selinux: improve symtab string hashing"
In reply to: Alexei Starovoitov: "Re: false positive deadlock? Was: [syzbot] [bpf?] possible deadlock in kvfree_call_rcu"
Next in thread: Hillf Danton: "Re: [syzbot] [bpf?] possible deadlock in kvfree_call_rcu"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]