KCSAN: data-race in rb_move_tail / ring_buffer_unlock_commit

Status: moderation: reported on 2026/11/20 00:09
Subsystems: trace
Labels: race:benign prio:low
[Documentation on labels]
Reported-by: syzbot+70e7520e4a7d94c17113@syzkaller.appspotmail.com
First crash: 2d22h, last: 2d06h
✨ AI-generated root cause summary:

In kernel/trace/ring_buffer.c, ring_buffer_unlock_commit() calls rb_wakeups() after rb_commit() has decremented cpu_buffer->committing to 0, with interrupts still enabled on the local CPU. Inside rb_wakeups(), cpu_buffer->commit_page is read without READ_ONCE() to compare against cpu_buffer->reader_page. An interrupt on the same CPU (e.g., perf hrtimer running a BPF program calling bpf_trace_printk()) can re-enter the ring buffer via a different interrupt context bit in trace_recursive_lock(), acquire committing == 1, and advance cpu_buffer->commit_page via rb_move_tail() -> rb_end_commit() -> rb_set_commit_to_write() -> rb_inc_page() without WRITE_ONCE(), causing a KCSAN data race.

▶ ▼ ✨ Full AI root cause analysis
Based on kernel commit 666539eecdf4:

Root Cause Analysis

The KCSAN data race is on struct ring_buffer_per_cpu::commit_page (at offset 0x60 in struct ring_buffer_per_cpu) in kernel/trace/ring_buffer.c, between a task-context commit finishing in ring_buffer_unlock_commit() and a hardirq-context reservation advancing the commit page on the same CPU.

Involved Functions and Fields

  • Struct / Field: struct ring_buffer_per_cpu::commit_page (struct buffer_page *, offset 768 bits / 96 bytes), cpu_buffer->committing, cpu_buffer->current_context, and cpu_buffer->reader_page.
  • Reader path (kernel/trace/ring_buffer.c):
    • ring_buffer_unlock_commit() (line 4573) calls rb_commit(cpu_buffer) (line 4580) and then rb_wakeups(buffer, cpu_buffer) (line 4582) before trace_recursive_unlock() and preempt_enable_notrace().
    • rb_commit() calls rb_end_commit(cpu_buffer) (line 4358), which decrements cpu_buffer->committing to 0 (local_dec(&cpu_buffer->committing) at line 4325).
    • rb_wakeups() then performs a plain (non-READ_ONCE) load of cpu_buffer->commit_page at line 4396:
      if (cpu_buffer->reader_page == cpu_buffer->commit_page)
              return;
      
  • Writer path (kernel/trace/ring_buffer.c):
    • While task 9170 is in rb_wakeups() on CPU 0 with interrupts enabled, an hrtimer interrupt fires (perf_swevent_hrtimer -> bpf_overflow_handler -> bpf_trace_printk).
    • ring_buffer_lock_reserve() (line 5043) calls trace_recursive_lock(cpu_buffer) (line 4483), which allows re-entry on the same CPU because the interrupt runs at a different context level (RB_CTX_IRQ) than the interrupted task (RB_CTX_NORMAL) in cpu_buffer->current_context.
    • rb_reserve_next_event() (line 4947) calls rb_start_commit(cpu_buffer) (line 4967), incrementing cpu_buffer->committing from 0 to 1.
    • When the current sub-buffer fills up, __rb_reserve_next() (line 4856) calls rb_move_tail() (line 3923), which advances tail_page and calls rb_end_commit(cpu_buffer) (line 4011).
    • Because cpu_buffer->committing == 1, rb_end_commit() calls rb_set_commit_to_write(cpu_buffer) (line 4323), which walks cpu_buffer->commit_page forward to tail_page via rb_inc_page(&cpu_buffer->commit_page) (line 4278):
      static inline void rb_inc_page(struct buffer_page **bpage)
      {
              struct list_head *p = rb_list_head((*bpage)->list.next);
      
              *bpage = list_entry(p, struct buffer_page, list);
      }
      
      This writes to cpu_buffer->commit_page at line 1471 without WRITE_ONCE().

Mechanism

Although ring_buffer_unlock_commit() still has preemption disabled and holds the RB_CTX_NORMAL recursion bit, cpu_buffer->committing has already dropped to 0 in rb_commit() and interrupts are enabled. An interrupt on the same CPU can therefore legitimately become the outermost active committer (committing == 1) and update cpu_buffer->commit_page concurrently with rb_wakeups() reading cpu_buffer->commit_page. Because neither the write in rb_inc_page() (or callers updating commit_page) nor the read in rb_wakeups() uses READ_ONCE()/WRITE_ONCE() (or data_race()), KCSAN reports a data race on the 8-byte commit_page pointer.

▶ ▼ ✨ AI Jobs (2)
ID Workflow Result Correct Bug Created Started Finished Revision
8ffab941 assessment-kcsan Benign: ✅ ❓ KCSAN: data-race in rb_move_tail / ring_buffer_unlock_commit 2026/11/19 19:39 2026/11/19 19:39 2026/11/19 19:41 229865fb
df1d6592 assessment-security DenialOfService: ❌ Exploitable: ❌ FilesystemTrigger: ❌ FuzzingArtifact: ❌ MemoryCorruption: ❌ NetworkNeedsSetup: ❌ NetworkTrigger: ❌ PeripheralTrigger: ❌ RemoteTrigger: ❌ Unprivileged: ❌ UserNamespace: ❌ VMGuestTrigger: ❌ VMHostTrigger: ❌ ❓ KCSAN: data-race in rb_move_tail / ring_buffer_unlock_commit 2026/11/19 19:38 2026/11/19 19:38 2026/11/19 19:48 229865fb

Sample crash report:
==================================================================
BUG: KCSAN: data-race in rb_move_tail / ring_buffer_unlock_commit

read-write to 0xffff88810007a860 of 8 bytes by interrupt on cpu 0:
 rb_inc_page kernel/trace/ring_buffer.c:1471 [inline]
 rb_set_commit_to_write kernel/trace/ring_buffer.c:4278 [inline]
 rb_end_commit kernel/trace/ring_buffer.c:4323 [inline]
 rb_move_tail+0x620/0xdc0 kernel/trace/ring_buffer.c:4011
 __rb_reserve_next+0x51c/0x790 kernel/trace/ring_buffer.c:4856
 rb_reserve_next_event kernel/trace/ring_buffer.c:5012 [inline]
 ring_buffer_lock_reserve+0x7e4/0x9c0 kernel/trace/ring_buffer.c:5071
 __trace_buffer_lock_reserve kernel/trace/trace.h:1674 [inline]
 trace_event_buffer_lock_reserve+0x14a/0x300 kernel/trace/trace.c:1935
 trace_event_buffer_reserve+0x120/0x180 kernel/trace/trace_events.c:724
 do_trace_event_raw_event_bpf_trace_printk kernel/trace/bpf_trace.h:11 [inline]
 trace_event_raw_event_bpf_trace_printk+0x9d/0x190 kernel/trace/bpf_trace.h:11
 __do_trace_bpf_trace_printk kernel/trace/bpf_trace.h:11 [inline]
 trace_bpf_trace_printk kernel/trace/bpf_trace.h:11 [inline]
 ____bpf_trace_printk kernel/trace/bpf_trace.c:409 [inline]
 bpf_trace_printk+0x1b9/0x1e0 kernel/trace/bpf_trace.c:392
 bpf_prog_484c593dea3d67e5+0x3b/0x41
 bpf_dispatcher_nop_func include/linux/bpf.h:1547 [inline]
 __bpf_prog_run include/linux/filter.h:813 [inline]
 bpf_prog_run include/linux/filter.h:820 [inline]
 bpf_overflow_handler kernel/events/core.c:10822 [inline]
 __perf_event_overflow+0x331/0x950 kernel/events/core.c:10921
 perf_event_overflow kernel/events/core.c:11004 [inline]
 perf_swevent_hrtimer+0x293/0x320 kernel/events/core.c:12424
 __run_hrtimer kernel/time/hrtimer.c:2079 [inline]
 __hrtimer_run_queues+0x1f8/0x540 kernel/time/hrtimer.c:2136
 hrtimer_interrupt+0x27b/0x870 kernel/time/hrtimer.c:2255
 local_apic_timer_interrupt arch/x86/kernel/apic/apic.c:1051 [inline]
 __sysvec_apic_timer_interrupt+0x5f/0x1f0 arch/x86/kernel/apic/apic.c:1068
 instr_sysvec_apic_timer_interrupt arch/x86/kernel/apic/apic.c:1062 [inline]
 sysvec_apic_timer_interrupt+0x6f/0x80 arch/x86/kernel/apic/apic.c:1062
 asm_sysvec_apic_timer_interrupt+0x1a/0x20 arch/x86/include/asm/idtentry.h:674
 rdtsc_ordered arch/x86/include/asm/tsc.h:57 [inline]
 delay_tsc+0x5b/0xa0 arch/x86/lib/delay.c:70
 udelay include/asm-generic/delay.h:64 [inline]
 delay_access kernel/kcsan/core.c:332 [inline]
 kcsan_setup_watchpoint+0x255/0x3f0 kernel/kcsan/core.c:622
 rb_wakeups kernel/trace/ring_buffer.c:4396 [inline]
 ring_buffer_unlock_commit+0x191/0x3b0 kernel/trace/ring_buffer.c:4582
 __buffer_unlock_commit kernel/trace/trace.h:1695 [inline]
 trace_buffer_unlock_commit_regs+0x43/0x310 kernel/trace/trace.c:2060
 trace_event_buffer_commit+0x149/0x5b0 kernel/trace/trace.c:2035
 do_trace_event_raw_event_bpf_trace_printk kernel/trace/bpf_trace.h:11 [inline]
 trace_event_raw_event_bpf_trace_printk+0xee/0x190 kernel/trace/bpf_trace.h:11
 __do_trace_bpf_trace_printk kernel/trace/bpf_trace.h:11 [inline]
 trace_bpf_trace_printk kernel/trace/bpf_trace.h:11 [inline]
 ____bpf_trace_printk kernel/trace/bpf_trace.c:409 [inline]
 bpf_trace_printk+0x1b9/0x1e0 kernel/trace/bpf_trace.c:392
 bpf_prog_ac2694f0a96ddbc2+0x43/0x6a
 bpf_dispatcher_nop_func include/linux/bpf.h:1547 [inline]
 __bpf_prog_run include/linux/filter.h:813 [inline]
 bpf_prog_run include/linux/filter.h:820 [inline]
 bpf_test_run+0x29b/0x620 net/bpf/test_run.c:422
 bpf_prog_test_run_skb+0xaa1/0x11d0 net/bpf/test_run.c:1226
 bpf_prog_test_run+0x21b/0x360 kernel/bpf/syscall.c:4860
 __sys_bpf+0x914/0xc50 kernel/bpf/syscall.c:6478
 __do_sys_bpf kernel/bpf/syscall.c:6582 [inline]
 __se_sys_bpf kernel/bpf/syscall.c:6579 [inline]
 __x64_sys_bpf+0x69/0x80 kernel/bpf/syscall.c:6579
 x64_sys_call+0xf26/0x2550 arch/x86/include/generated/asm/syscalls_64.h:322
 do_syscall_x64 arch/x86/entry/syscall_64.c:61 [inline]
 do_syscall_64+0x11e/0x3a0 arch/x86/entry/syscall_64.c:89
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

read to 0xffff88810007a860 of 8 bytes by task 9170 on cpu 0:
 rb_wakeups kernel/trace/ring_buffer.c:4396 [inline]
 ring_buffer_unlock_commit+0x191/0x3b0 kernel/trace/ring_buffer.c:4582
 __buffer_unlock_commit kernel/trace/trace.h:1695 [inline]
 trace_buffer_unlock_commit_regs+0x43/0x310 kernel/trace/trace.c:2060
 trace_event_buffer_commit+0x149/0x5b0 kernel/trace/trace.c:2035
 do_trace_event_raw_event_bpf_trace_printk kernel/trace/bpf_trace.h:11 [inline]
 trace_event_raw_event_bpf_trace_printk+0xee/0x190 kernel/trace/bpf_trace.h:11
 __do_trace_bpf_trace_printk kernel/trace/bpf_trace.h:11 [inline]
 trace_bpf_trace_printk kernel/trace/bpf_trace.h:11 [inline]
 ____bpf_trace_printk kernel/trace/bpf_trace.c:409 [inline]
 bpf_trace_printk+0x1b9/0x1e0 kernel/trace/bpf_trace.c:392
 bpf_prog_ac2694f0a96ddbc2+0x43/0x6a
 bpf_dispatcher_nop_func include/linux/bpf.h:1547 [inline]
 __bpf_prog_run include/linux/filter.h:813 [inline]
 bpf_prog_run include/linux/filter.h:820 [inline]
 bpf_test_run+0x29b/0x620 net/bpf/test_run.c:422
 bpf_prog_test_run_skb+0xaa1/0x11d0 net/bpf/test_run.c:1226
 bpf_prog_test_run+0x21b/0x360 kernel/bpf/syscall.c:4860
 __sys_bpf+0x914/0xc50 kernel/bpf/syscall.c:6478
 __do_sys_bpf kernel/bpf/syscall.c:6582 [inline]
 __se_sys_bpf kernel/bpf/syscall.c:6579 [inline]
 __x64_sys_bpf+0x69/0x80 kernel/bpf/syscall.c:6579
 x64_sys_call+0xf26/0x2550 arch/x86/include/generated/asm/syscalls_64.h:322
 do_syscall_x64 arch/x86/entry/syscall_64.c:61 [inline]
 do_syscall_64+0x11e/0x3a0 arch/x86/entry/syscall_64.c:89
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

value changed: 0xffff8881206ca040 -> 0xffff8881206ca000

Reported by Kernel Concurrency Sanitizer on:
CPU: 0 UID: 0 PID: 9170 Comm: syz.1.1277 Not tainted syzkaller #0 PREEMPT(lazy) 
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 11/06/2026
==================================================================

Crashes (2):
Time Kernel Commit Syzkaller Config Log Report Syz repro C repro VM info Assets (help?) Manager Title
2026/11/17 18:36 upstream 666539eecdf4 122f94d5 .config console log report info [disk image] [vmlinux] [kernel image] ci2-upstream-kcsan-gce KCSAN: data-race in rb_move_tail / ring_buffer_unlock_commit
2026/11/17 03:04 upstream 666539eecdf4 a3e64cfc .config console log report info [disk image] [vmlinux] [kernel image] ci2-upstream-kcsan-gce KCSAN: data-race in rb_move_tail / ring_buffer_unlock_commit
* Struck through repros no longer work on HEAD.