Skip to contentSkip to conversation

bpf: Convert ringbuf.c to rqspinlock

PATCH Applied Kumar Kartikeya Dwivedi 3 participants 6 messages Apr 11, 2025

Subsystems (2)
lore
Share & tools
lore

Kumar Kartikeya Dwivedi

author maintainer
#0 lore raw

PATCH 1/1 bpf: Convert ringbuf.c to rqspinlock

Convert the raw spinlock used by BPF ringbuf to rqspinlock. Currently, we have an open syzbot report of a potential deadlock. In addition, the ringbuf can fail to reserve spuriously under contention from NMI context.

It is potentially attractive to enable unconstrained usage (incl. NMIs) while ensuring no deadlocks manifest at runtime, perform the conversion to rqspinlock to achieve this.

This change was benchmarked for BPF ringbuf's multi-producer contention case on an Intel Sapphire Rapids server, with hyperthreading disabled and performance governor turned on. 5 warm up runs were done for each case before obtaining the results.

Before (raw_spinlock_t):

Ringbuf, multi-producer contention
==================================

rb-libbpf nr_prod 1  11.440 ± 0.019M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 2  2.706 ± 0.010M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 3  3.130 ± 0.004M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 4  2.472 ± 0.003M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 8  2.352 ± 0.001M/s (drops 0.000 ± 0.000M/s)

rb-libbpf nr_prod 12 2.813 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 16 1.988 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 20 2.245 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 24 2.148 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 28 2.190 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 32 2.490 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 36 2.180 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 40 2.201 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 44 2.226 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 48 2.164 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 52 1.874 ± 0.001M/s (drops 0.000 ± 0.000M/s)

After (rqspinlock_t):

Ringbuf, multi-producer contention
==================================

rb-libbpf nr_prod 1  11.078 ± 0.019M/s (drops 0.000 ± 0.000M/s) (-3.16%)
rb-libbpf nr_prod 2  2.801 ± 0.014M/s (drops 0.000 ± 0.000M/s) (3.51%)
rb-libbpf nr_prod 3  3.454 ± 0.005M/s (drops 0.000 ± 0.000M/s) (10.35%)
rb-libbpf nr_prod 4  2.567 ± 0.002M/s (drops 0.000 ± 0.000M/s) (3.84%)
rb-libbpf nr_prod 8  2.468 ± 0.001M/s (drops 0.000 ± 0.000M/s) (4.93%)

rb-libbpf nr_prod 12 2.510 ± 0.001M/s (drops 0.000 ± 0.000M/s) (-10.77%) rb-libbpf nr_prod 16 2.075 ± 0.001M/s (drops 0.000 ± 0.000M/s) (4.38%) rb-libbpf nr_prod 20 2.640 ± 0.001M/s (drops 0.000 ± 0.000M/s) (17.59%) rb-libbpf nr_prod 24 2.092 ± 0.001M/s (drops 0.000 ± 0.000M/s) (-2.61%) rb-libbpf nr_prod 28 2.426 ± 0.005M/s (drops 0.000 ± 0.000M/s) (10.78%) rb-libbpf nr_prod 32 2.331 ± 0.004M/s (drops 0.000 ± 0.000M/s) (-6.39%) rb-libbpf nr_prod 36 2.306 ± 0.003M/s (drops 0.000 ± 0.000M/s) (5.78%) rb-libbpf nr_prod 40 2.178 ± 0.002M/s (drops 0.000 ± 0.000M/s) (-1.04%) rb-libbpf nr_prod 44 2.293 ± 0.001M/s (drops 0.000 ± 0.000M/s) (3.01%) rb-libbpf nr_prod 48 2.022 ± 0.001M/s (drops 0.000 ± 0.000M/s) (-6.56%) rb-libbpf nr_prod 52 1.809 ± 0.001M/s (drops 0.000 ± 0.000M/s) (-3.47%)

There's a fair amount of noise in the benchmark, with numbers on reruns going up and down by 10%, so all changes are in the range of this disturbance, and we see no major regressions.

---
 kernel/bpf/ringbuf.c | 17 +++++++----------
 1 file changed, 7 insertions(+), 10 deletions(-)
kernel/bpf/ringbuf.c +7 -10
index 1499d8caa9a3..719d73299397 100644
@@ -11,6 +11,7 @@
 #include <linux/kmemleak.h>
 #include <uapi/linux/btf.h>
 #include <linux/btf_ids.h>
+#include <asm/rqspinlock.h>
 
 #define RINGBUF_CREATE_FLAG_MASK (BPF_F_NUMA_NODE)
 
@@ -29,7 +30,7 @@ struct bpf_ringbuf {
 	u64 mask;
 	struct page **pages;
 	int nr_pages;
-	raw_spinlock_t spinlock ____cacheline_aligned_in_smp;
+	rqspinlock_t spinlock ____cacheline_aligned_in_smp;
 	/* For user-space producer ring buffers, an atomic_t busy bit is used
 	 * to synchronize access to the ring buffers in the kernel, rather than
 	 * the spinlock that is used for kernel-producer ring buffers. This is
@@ -173,7 +174,7 @@ static struct bpf_ringbuf *bpf_ringbuf_alloc(size_t data_sz, int numa_node)
 	if (!rb)
 		return NULL;
 
-	raw_spin_lock_init(&rb->spinlock);
+	raw_res_spin_lock_init(&rb->spinlock);
 	atomic_set(&rb->busy, 0);
 	init_waitqueue_head(&rb->waitq);
 	init_irq_work(&rb->work, bpf_ringbuf_notify);
@@ -416,12 +417,8 @@ static void *__bpf_ringbuf_reserve(struct bpf_ringbuf *rb, u64 size)
 
 	cons_pos = smp_load_acquire(&rb->consumer_pos);
 
-	if (in_nmi()) {
-		if (!raw_spin_trylock_irqsave(&rb->spinlock, flags))
-			return NULL;
-	} else {
-		raw_spin_lock_irqsave(&rb->spinlock, flags);
-	}
+	if (raw_res_spin_lock_irqsave(&rb->spinlock, flags))
+		return NULL;
 
 	pend_pos = rb->pending_pos;
 	prod_pos = rb->producer_pos;
@@ -446,7 +443,7 @@ static void *__bpf_ringbuf_reserve(struct bpf_ringbuf *rb, u64 size)
 	 */
 	if (new_prod_pos - cons_pos > rb->mask ||
 	    new_prod_pos - pend_pos > rb->mask) {
-		raw_spin_unlock_irqrestore(&rb->spinlock, flags);
+		raw_res_spin_unlock_irqrestore(&rb->spinlock, flags);
 		return NULL;
 	}
 
@@ -458,7 +455,7 @@ static void *__bpf_ringbuf_reserve(struct bpf_ringbuf *rb, u64 size)
 	/* pairs with consumer's smp_load_acquire() */
 	smp_store_release(&rb->producer_pos, new_prod_pos);
 
-	raw_spin_unlock_irqrestore(&rb->spinlock, flags);
+	raw_res_spin_unlock_irqrestore(&rb->spinlock, flags);
 
 	return (void *)hdr + BPF_RINGBUF_HDR_SZ;
 }

--
2.47.1

author maintainer
#1 lore raw
Quoted 57 lines · Kumar Kartikeya Dwivedi

Convert the raw spinlock used by BPF ringbuf to rqspinlock. Currently, we have an open syzbot report of a potential deadlock. In addition, the ringbuf can fail to reserve spuriously under contention from NMI context.

It is potentially attractive to enable unconstrained usage (incl. NMIs) while ensuring no deadlocks manifest at runtime, perform the conversion to rqspinlock to achieve this.

This change was benchmarked for BPF ringbuf's multi-producer contention case on an Intel Sapphire Rapids server, with hyperthreading disabled and performance governor turned on. 5 warm up runs were done for each case before obtaining the results.

Before (raw_spinlock_t):

Ringbuf, multi-producer contention
==================================

rb-libbpf nr_prod 1  11.440 ± 0.019M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 2  2.706 ± 0.010M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 3  3.130 ± 0.004M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 4  2.472 ± 0.003M/s (drops 0.000 ± 0.000M/s)
rb-libbpf nr_prod 8  2.352 ± 0.001M/s (drops 0.000 ± 0.000M/s)

rb-libbpf nr_prod 12 2.813 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 16 1.988 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 20 2.245 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 24 2.148 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 28 2.190 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 32 2.490 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 36 2.180 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 40 2.201 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 44 2.226 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 48 2.164 ± 0.001M/s (drops 0.000 ± 0.000M/s) rb-libbpf nr_prod 52 1.874 ± 0.001M/s (drops 0.000 ± 0.000M/s)

After (rqspinlock_t):

Ringbuf, multi-producer contention
==================================

rb-libbpf nr_prod 1  11.078 ± 0.019M/s (drops 0.000 ± 0.000M/s) (-3.16%)
rb-libbpf nr_prod 2  2.801 ± 0.014M/s (drops 0.000 ± 0.000M/s) (3.51%)
rb-libbpf nr_prod 3  3.454 ± 0.005M/s (drops 0.000 ± 0.000M/s) (10.35%)
rb-libbpf nr_prod 4  2.567 ± 0.002M/s (drops 0.000 ± 0.000M/s) (3.84%)
rb-libbpf nr_prod 8  2.468 ± 0.001M/s (drops 0.000 ± 0.000M/s) (4.93%)

rb-libbpf nr_prod 12 2.510 ± 0.001M/s (drops 0.000 ± 0.000M/s) (-10.77%) rb-libbpf nr_prod 16 2.075 ± 0.001M/s (drops 0.000 ± 0.000M/s) (4.38%) rb-libbpf nr_prod 20 2.640 ± 0.001M/s (drops 0.000 ± 0.000M/s) (17.59%) rb-libbpf nr_prod 24 2.092 ± 0.001M/s (drops 0.000 ± 0.000M/s) (-2.61%) rb-libbpf nr_prod 28 2.426 ± 0.005M/s (drops 0.000 ± 0.000M/s) (10.78%) rb-libbpf nr_prod 32 2.331 ± 0.004M/s (drops 0.000 ± 0.000M/s) (-6.39%) rb-libbpf nr_prod 36 2.306 ± 0.003M/s (drops 0.000 ± 0.000M/s) (5.78%) rb-libbpf nr_prod 40 2.178 ± 0.002M/s (drops 0.000 ± 0.000M/s) (-1.04%) rb-libbpf nr_prod 44 2.293 ± 0.001M/s (drops 0.000 ± 0.000M/s) (3.01%) rb-libbpf nr_prod 48 2.022 ± 0.001M/s (drops 0.000 ± 0.000M/s) (-6.56%) rb-libbpf nr_prod 52 1.809 ± 0.001M/s (drops 0.000 ± 0.000M/s) (-3.47%)

Quoted 8 lines

There's a fair amount of noise in the benchmark, with numbers on reruns going up and down by 10%, so all changes are in the range of this disturbance, and we see no major regressions.

---

#syz test

syzbot
#2 lore raw

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

syzbot has tested the proposed patch but the reproducer is still triggering an issue: possible deadlock in __bpf_ringbuf_reserve

Hello,

syzbot has tested the proposed patch but the reproducer is still triggering an issue:
possible deadlock in __bpf_ringbuf_reserve

============================================
WARNING: possible recursive locking detected
6.15.0-rc1-syzkaller-ge618ee89561b #0 Not tainted
--------------------------------------------
kworker/2:3/6044 is trying to acquire lock:
ffffc90006f360d8 (&rb->spinlock){-.-.}-{2:2}, at: __bpf_ringbuf_reserve+0x36e/0x4b0 kernel/bpf/ringbuf.c:423

but task is already holding lock:
ffffc900070410d8 (&rb->spinlock){-.-.}-{2:2}, at: __bpf_ringbuf_reserve+0x36e/0x4b0 kernel/bpf/ringbuf.c:423

other info that might help us debug this:

 Possible unsafe locking scenario:
       CPU0
       ----

lock(&rb->spinlock);
lock(&rb->spinlock);

*** DEADLOCK ***

May be due to missing lock nesting notation

6 locks held by kworker/2:3/6044:
#0: ffff88801b48ad48 ((wq_completion)rcu_gp){+.+.}-{0:0}, at: process_one_work+0x12a2/0x1b70 kernel/workqueue.c:3213 #1: ffffc90004c1fd18 ((work_completion)(&(&ssp->srcu_sup->work)->work)){+.+.}-{0:0}, at: process_one_work+0x929/0x1b70 kernel/workqueue.c:3214 #2: ffff88801ea8f158 (&ssp->srcu_sup->srcu_gp_mutex){+.+.}-{4:4}, at: srcu_advance_state kernel/rcu/srcutree.c:1701 [inline] #2: ffff88801ea8f158 (&ssp->srcu_sup->srcu_gp_mutex){+.+.}-{4:4}, at: process_srcu+0x73/0x1920 kernel/rcu/srcutree.c:1861 #3: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: rcu_lock_acquire include/linux/rcupdate.h:331 [inline]
#3: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: rcu_read_lock include/linux/rcupdate.h:841 [inline]
#3: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: __bpf_trace_run kernel/trace/bpf_trace.c:2362 [inline]
#3: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: bpf_trace_run2+0x1b6/0x590 kernel/trace/bpf_trace.c:2404
#4: ffffc900070410d8 (&rb->spinlock){-.-.}-{2:2}, at: __bpf_ringbuf_reserve+0x36e/0x4b0 kernel/bpf/ringbuf.c:423
#5: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: rcu_lock_acquire include/linux/rcupdate.h:331 [inline]
#5: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: rcu_read_lock include/linux/rcupdate.h:841 [inline]
#5: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: __bpf_trace_run kernel/trace/bpf_trace.c:2362 [inline]
#5: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: bpf_trace_run2+0x1b6/0x590 kernel/trace/bpf_trace.c:2404

stack backtrace:
CPU: 2 UID: 0 PID: 6044 Comm: kworker/2:3 Not tainted 6.15.0-rc1-syzkaller-ge618ee89561b #0 PREEMPT(full) 
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014
Workqueue: rcu_gp process_srcu
Call Trace:
 <TASK>
 __dump_stack lib/dump_stack.c:94 [inline]
 dump_stack_lvl+0x116/0x1f0 lib/dump_stack.c:120
 print_deadlock_bug+0x1e9/0x240 kernel/locking/lockdep.c:3042
 check_deadlock kernel/locking/lockdep.c:3094 [inline]
 validate_chain kernel/locking/lockdep.c:3896 [inline]
 __lock_acquire+0xff7/0x1ba0 kernel/locking/lockdep.c:5235
 lock_acquire kernel/locking/lockdep.c:5866 [inline]
 lock_acquire+0x179/0x350 kernel/locking/lockdep.c:5823
 __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
 _raw_spin_lock_irqsave+0x3a/0x60 kernel/locking/spinlock.c:162
 __bpf_ringbuf_reserve+0x36e/0x4b0 kernel/bpf/ringbuf.c:423
 ____bpf_ringbuf_reserve kernel/bpf/ringbuf.c:474 [inline]
 bpf_ringbuf_reserve+0x57/0x90 kernel/bpf/ringbuf.c:466
 bpf_prog_385141c453c15099+0x36/0x5d
 bpf_dispatcher_nop_func include/linux/bpf.h:1316 [inline]
 __bpf_prog_run include/linux/filter.h:718 [inline]
 bpf_prog_run include/linux/filter.h:725 [inline]
 __bpf_trace_run kernel/trace/bpf_trace.c:2363 [inline]
 bpf_trace_run2+0x230/0x590 kernel/trace/bpf_trace.c:2404
 __bpf_trace_contention_begin+0xc9/0x110 include/trace/events/lock.h:95
 __traceiter_contention_begin+0x5a/0xa0 include/trace/events/lock.h:95
 __preempt_count_dec_and_test arch/x86/include/asm/preempt.h:95 [inline]
 class_preempt_notrace_destructor include/linux/preempt.h:482 [inline]
 __do_trace_contention_begin include/trace/events/lock.h:95 [inline]
 trace_contention_begin.constprop.0+0xde/0x160 include/trace/events/lock.h:95
 __pv_queued_spin_lock_slowpath+0x109/0xcf0 kernel/locking/qspinlock.c:219
 pv_queued_spin_lock_slowpath arch/x86/include/asm/paravirt.h:572 [inline]
 queued_spin_lock_slowpath arch/x86/include/asm/qspinlock.h:51 [inline]
 queued_spin_lock include/asm-generic/qspinlock.h:114 [inline]
 do_raw_spin_lock+0x20e/0x2b0 kernel/locking/spinlock_debug.c:116
 __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:111 [inline]
 _raw_spin_lock_irqsave+0x42/0x60 kernel/locking/spinlock.c:162
 __bpf_ringbuf_reserve+0x36e/0x4b0 kernel/bpf/ringbuf.c:423
 ____bpf_ringbuf_reserve kernel/bpf/ringbuf.c:474 [inline]
 bpf_ringbuf_reserve+0x57/0x90 kernel/bpf/ringbuf.c:466
 bpf_prog_385141c453c15099+0x36/0x5d
 bpf_dispatcher_nop_func include/linux/bpf.h:1316 [inline]
 __bpf_prog_run include/linux/filter.h:718 [inline]
 bpf_prog_run include/linux/filter.h:725 [inline]
 __bpf_trace_run kernel/trace/bpf_trace.c:2363 [inline]
 bpf_trace_run2+0x230/0x590 kernel/trace/bpf_trace.c:2404
 __bpf_trace_contention_begin+0xc9/0x110 include/trace/events/lock.h:95
 __traceiter_contention_begin+0x5a/0xa0 include/trace/events/lock.h:95
 __do_trace_contention_begin include/trace/events/lock.h:95 [inline]
 trace_contention_begin+0xc1/0x130 include/trace/events/lock.h:95
 __mutex_lock_common kernel/locking/mutex.c:603 [inline]
 __mutex_lock+0x1a6/0xb90 kernel/locking/mutex.c:746
 srcu_advance_state kernel/rcu/srcutree.c:1701 [inline]
 process_srcu+0x73/0x1920 kernel/rcu/srcutree.c:1861
 process_one_work+0x9cc/0x1b70 kernel/workqueue.c:3238
 process_scheduled_works kernel/workqueue.c:3319 [inline]
 worker_thread+0x6c8/0xf10 kernel/workqueue.c:3400
 kthread+0x3c2/0x780 kernel/kthread.c:464
 ret_from_fork+0x45/0x80 arch/x86/kernel/process.c:153
 ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:245
 </TASK>

Tested on:

commit:         e618ee89 Merge tag 'spi-fix-v6.15-rc1' of git://git.ke..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=10461c04580000
kernel config:  https://syzkaller.appspot.com/x/.config?x=36c5de4d99134dda
dashboard link: https://syzkaller.appspot.com/bug?extid=850aaf14624dc0c6d366
compiler:       gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

Note: no patches were applied.

author maintainer
#3 lore raw

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

Quoted 109 lines · syzbot

Hello,

syzbot has tested the proposed patch but the reproducer is still triggering an issue:
possible deadlock in __bpf_ringbuf_reserve

============================================
WARNING: possible recursive locking detected
6.15.0-rc1-syzkaller-ge618ee89561b #0 Not tainted
--------------------------------------------
kworker/2:3/6044 is trying to acquire lock:
ffffc90006f360d8 (&rb->spinlock){-.-.}-{2:2}, at: __bpf_ringbuf_reserve+0x36e/0x4b0 kernel/bpf/ringbuf.c:423

but task is already holding lock:
ffffc900070410d8 (&rb->spinlock){-.-.}-{2:2}, at: __bpf_ringbuf_reserve+0x36e/0x4b0 kernel/bpf/ringbuf.c:423

other info that might help us debug this:

 Possible unsafe locking scenario:
       CPU0
       ----

lock(&rb->spinlock);
lock(&rb->spinlock);

*** DEADLOCK ***

May be due to missing lock nesting notation

6 locks held by kworker/2:3/6044:
#0: ffff88801b48ad48 ((wq_completion)rcu_gp){+.+.}-{0:0}, at: process_one_work+0x12a2/0x1b70 kernel/workqueue.c:3213 #1: ffffc90004c1fd18 ((work_completion)(&(&ssp->srcu_sup->work)->work)){+.+.}-{0:0}, at: process_one_work+0x929/0x1b70 kernel/workqueue.c:3214 #2: ffff88801ea8f158 (&ssp->srcu_sup->srcu_gp_mutex){+.+.}-{4:4}, at: srcu_advance_state kernel/rcu/srcutree.c:1701 [inline] #2: ffff88801ea8f158 (&ssp->srcu_sup->srcu_gp_mutex){+.+.}-{4:4}, at: process_srcu+0x73/0x1920 kernel/rcu/srcutree.c:1861 #3: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: rcu_lock_acquire include/linux/rcupdate.h:331 [inline]
#3: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: rcu_read_lock include/linux/rcupdate.h:841 [inline]
#3: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: __bpf_trace_run kernel/trace/bpf_trace.c:2362 [inline]
#3: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: bpf_trace_run2+0x1b6/0x590 kernel/trace/bpf_trace.c:2404
#4: ffffc900070410d8 (&rb->spinlock){-.-.}-{2:2}, at: __bpf_ringbuf_reserve+0x36e/0x4b0 kernel/bpf/ringbuf.c:423
#5: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: rcu_lock_acquire include/linux/rcupdate.h:331 [inline]
#5: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: rcu_read_lock include/linux/rcupdate.h:841 [inline]
#5: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: __bpf_trace_run kernel/trace/bpf_trace.c:2362 [inline]
#5: ffffffff8e3c15c0 (rcu_read_lock){....}-{1:3}, at: bpf_trace_run2+0x1b6/0x590 kernel/trace/bpf_trace.c:2404

stack backtrace:
CPU: 2 UID: 0 PID: 6044 Comm: kworker/2:3 Not tainted 6.15.0-rc1-syzkaller-ge618ee89561b #0 PREEMPT(full)
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014
Workqueue: rcu_gp process_srcu
Call Trace:
 <TASK>
 __dump_stack lib/dump_stack.c:94 [inline]
 dump_stack_lvl+0x116/0x1f0 lib/dump_stack.c:120
 print_deadlock_bug+0x1e9/0x240 kernel/locking/lockdep.c:3042
 check_deadlock kernel/locking/lockdep.c:3094 [inline]
 validate_chain kernel/locking/lockdep.c:3896 [inline]
 __lock_acquire+0xff7/0x1ba0 kernel/locking/lockdep.c:5235
 lock_acquire kernel/locking/lockdep.c:5866 [inline]
 lock_acquire+0x179/0x350 kernel/locking/lockdep.c:5823
 __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
 _raw_spin_lock_irqsave+0x3a/0x60 kernel/locking/spinlock.c:162
 __bpf_ringbuf_reserve+0x36e/0x4b0 kernel/bpf/ringbuf.c:423
 ____bpf_ringbuf_reserve kernel/bpf/ringbuf.c:474 [inline]
 bpf_ringbuf_reserve+0x57/0x90 kernel/bpf/ringbuf.c:466
 bpf_prog_385141c453c15099+0x36/0x5d
 bpf_dispatcher_nop_func include/linux/bpf.h:1316 [inline]
 __bpf_prog_run include/linux/filter.h:718 [inline]
 bpf_prog_run include/linux/filter.h:725 [inline]
 __bpf_trace_run kernel/trace/bpf_trace.c:2363 [inline]
 bpf_trace_run2+0x230/0x590 kernel/trace/bpf_trace.c:2404
 __bpf_trace_contention_begin+0xc9/0x110 include/trace/events/lock.h:95
 __traceiter_contention_begin+0x5a/0xa0 include/trace/events/lock.h:95
 __preempt_count_dec_and_test arch/x86/include/asm/preempt.h:95 [inline]
 class_preempt_notrace_destructor include/linux/preempt.h:482 [inline]
 __do_trace_contention_begin include/trace/events/lock.h:95 [inline]
 trace_contention_begin.constprop.0+0xde/0x160 include/trace/events/lock.h:95
 __pv_queued_spin_lock_slowpath+0x109/0xcf0 kernel/locking/qspinlock.c:219
 pv_queued_spin_lock_slowpath arch/x86/include/asm/paravirt.h:572 [inline]
 queued_spin_lock_slowpath arch/x86/include/asm/qspinlock.h:51 [inline]
 queued_spin_lock include/asm-generic/qspinlock.h:114 [inline]
 do_raw_spin_lock+0x20e/0x2b0 kernel/locking/spinlock_debug.c:116
 __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:111 [inline]
 _raw_spin_lock_irqsave+0x42/0x60 kernel/locking/spinlock.c:162
 __bpf_ringbuf_reserve+0x36e/0x4b0 kernel/bpf/ringbuf.c:423
 ____bpf_ringbuf_reserve kernel/bpf/ringbuf.c:474 [inline]
 bpf_ringbuf_reserve+0x57/0x90 kernel/bpf/ringbuf.c:466
 bpf_prog_385141c453c15099+0x36/0x5d
 bpf_dispatcher_nop_func include/linux/bpf.h:1316 [inline]
 __bpf_prog_run include/linux/filter.h:718 [inline]
 bpf_prog_run include/linux/filter.h:725 [inline]
 __bpf_trace_run kernel/trace/bpf_trace.c:2363 [inline]
 bpf_trace_run2+0x230/0x590 kernel/trace/bpf_trace.c:2404
 __bpf_trace_contention_begin+0xc9/0x110 include/trace/events/lock.h:95
 __traceiter_contention_begin+0x5a/0xa0 include/trace/events/lock.h:95
 __do_trace_contention_begin include/trace/events/lock.h:95 [inline]
 trace_contention_begin+0xc1/0x130 include/trace/events/lock.h:95
 __mutex_lock_common kernel/locking/mutex.c:603 [inline]
 __mutex_lock+0x1a6/0xb90 kernel/locking/mutex.c:746
 srcu_advance_state kernel/rcu/srcutree.c:1701 [inline]
 process_srcu+0x73/0x1920 kernel/rcu/srcutree.c:1861
 process_one_work+0x9cc/0x1b70 kernel/workqueue.c:3238
 process_scheduled_works kernel/workqueue.c:3319 [inline]
 worker_thread+0x6c8/0xf10 kernel/workqueue.c:3400
 kthread+0x3c2/0x780 kernel/kthread.c:464
 ret_from_fork+0x45/0x80 arch/x86/kernel/process.c:153
 ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:245
 </TASK>

Tested on:

Quoted 8 lines
commit:         e618ee89 Merge tag 'spi-fix-v6.15-rc1' of git://git.ke..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=10461c04580000
kernel config:  https://syzkaller.appspot.com/x/.config?x=36c5de4d99134dda
dashboard link: https://syzkaller.appspot.com/bug?extid=850aaf14624dc0c6d366
compiler:       gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

Note: no patches were applied.

#syz test: git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf.git master

syzbot
#4 lore raw

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

syzbot has tested the proposed patch but the reproducer is still triggering an issue: unregister_netdevice: waiting for DEV to become free

Hello,

syzbot has tested the proposed patch but the reproducer is still triggering an issue:
unregister_netdevice: waiting for DEV to become free

unregister_netdevice: waiting for batadv0 to become free. Usage count = 3

Tested on:

commit:         a650d389 bpf: Convert ringbuf map to rqspinlock
git tree:       bpf
console output: https://syzkaller.appspot.com/x/log.txt?x=17928870580000
kernel config:  https://syzkaller.appspot.com/x/.config?x=ea2b297a0891c87e
dashboard link: https://syzkaller.appspot.com/bug?extid=850aaf14624dc0c6d366
compiler:       gcc (Debian 12.2.0-14) 12.2.0, GNU ld (GNU Binutils for Debian) 2.40

Note: no patches were applied.

#5 lore raw
In reply to Kumar Kartikeya Dwivedi
This patch was applied to bpf/bpf.git (master) by Alexei Starovoitov <[email protected]>:

Hello:

This patch was applied to bpf/bpf.git (master)
by Alexei Starovoitov <[email protected]>:

Quoted 10 lines · On Fri, 11 Apr 2025 03:17:59 -0700 you wrote:

Convert the raw spinlock used by BPF ringbuf to rqspinlock. Currently, we have an open syzbot report of a potential deadlock. In addition, the ringbuf can fail to reserve spuriously under contention from NMI context.

It is potentially attractive to enable unconstrained usage (incl. NMIs) while ensuring no deadlocks manifest at runtime, perform the conversion to rqspinlock to achieve this.

[...]

Here is the summary with links:
- [bpf,v1] bpf: Convert ringbuf.c to rqspinlock

    https://git.kernel.org/bpf/bpf/c/0b51f0ac3dc5

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html