Ksawlii/kernel_samsung_a53x
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
kernel_samsung_a53x/kernel
History
Haifeng Xu c52e3af387 perf/core: Fix missing wakeup when waiting for context reference 
[ Upstream commit 74751ef5c1912ebd3e65c3b65f45587e05ce5d36 ]

In our production environment, we found many hung tasks which are
blocked for more than 18 hours. Their call traces are like this:

[346278.191038] __schedule+0x2d8/0x890
[346278.191046] schedule+0x4e/0xb0
[346278.191049] perf_event_free_task+0x220/0x270
[346278.191056] ? init_wait_var_entry+0x50/0x50
[346278.191060] copy_process+0x663/0x18d0
[346278.191068] kernel_clone+0x9d/0x3d0
[346278.191072] __do_sys_clone+0x5d/0x80
[346278.191076] __x64_sys_clone+0x25/0x30
[346278.191079] do_syscall_64+0x5c/0xc0
[346278.191083] ? syscall_exit_to_user_mode+0x27/0x50
[346278.191086] ? do_syscall_64+0x69/0xc0
[346278.191088] ? irqentry_exit_to_user_mode+0x9/0x20
[346278.191092] ? irqentry_exit+0x19/0x30
[346278.191095] ? exc_page_fault+0x89/0x160
[346278.191097] ? asm_exc_page_fault+0x8/0x30
[346278.191102] entry_SYSCALL_64_after_hwframe+0x44/0xae

The task was waiting for the refcount become to 1, but from the vmcore,
we found the refcount has already been 1. It seems that the task didn't
get woken up by perf_event_release_kernel() and got stuck forever. The
below scenario may cause the problem.

Thread A					Thread B
...						...
perf_event_free_task				perf_event_release_kernel
						   ...
						   acquire event->child_mutex
						   ...
						   get_ctx
   ...						   release event->child_mutex
   acquire ctx->mutex
   ...
   perf_free_event (acquire/release event->child_mutex)
   ...
   release ctx->mutex
   wait_var_event
						   acquire ctx->mutex
						   acquire event->child_mutex
						   # move existing events to free_list
						   release event->child_mutex
						   release ctx->mutex
						   put_ctx
...						...

In this case, all events of the ctx have been freed, so we couldn't
find the ctx in free_list and Thread A will miss the wakeup. It's thus
necessary to add a wakeup after dropping the reference.

Fixes: 1cf8dfe8a661 ("perf/core: Fix race between close() and fork()")
Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20240513103948.33570-1-haifeng.xu@shopee.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
2024-11-19 14:19:30 +01:00
..
bpf Revert "fs: add file and path permissions helpers" 2024-11-19 13:30:21 +01:00
cgroup sched/fair: Allow disabling sched_balance_newidle with sched_relax_domain_level 2024-11-19 12:27:00 +01:00
configs Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
debug kdb: Use format-specifiers rather than memset() for padding in kdb_read() 2024-11-19 12:27:19 +01:00
dma dma-mapping: clear dev->dma_mem to NULL after freeing it 2024-11-18 12:12:37 +01:00
entry Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
events perf/core: Fix missing wakeup when waiting for context reference 2024-11-19 14:19:30 +01:00
futex futex: Don't include process MM in futex key on no-MMU 2024-11-18 11:42:47 +01:00
gcov gcov: add support for GCC 14 2024-11-19 14:19:10 +01:00
irq genirq/cpuhotplug, x86/vector: Prevent vector leak during CPU offline 2024-11-19 12:27:11 +01:00
kcsan Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
livepatch kallsyms: refactor {,module_}kallsyms_on_each_symbol 2024-11-19 12:27:32 +01:00
locking lockdep: Fix block chain corruption 2024-11-18 12:10:56 +01:00
power PM: suspend: Set mem_sleep_current during kernel command line setup 2024-11-19 09:22:17 +01:00
printk printk: Update @console_may_schedule in console_trylock_spinning() 2024-11-19 09:22:40 +01:00
rcu rcutorture: Fix invalid context warning when enable srcu barrier testing 2024-11-19 14:19:06 +01:00
sched sched/fair: Allow disabling sched_balance_newidle with sched_relax_domain_level 2024-11-19 12:27:00 +01:00
time tick/nohz_full: Don't abuse smp_call_function_single() in tick_setup_device() 2024-11-19 14:19:05 +01:00
trace tracing: Add MODULE_DESCRIPTION() to preemptirq_delay_test 2024-11-19 14:19:29 +01:00
acct.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
async.c async: Introduce async_schedule_dev_nocall() 2024-11-18 12:12:56 +01:00
audit.c audit: Send netlink ACK before setting connection in auditd_set 2024-11-18 12:13:09 +01:00
audit.h Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
audit_fsnotify.c fsnotify: make allow_dups a property of the group 2024-11-19 12:27:55 +01:00
audit_tree.c fsnotify: pass flags argument to fsnotify_alloc_group() 2024-11-19 12:27:55 +01:00
audit_watch.c fsnotify: pass flags argument to fsnotify_alloc_group() 2024-11-19 12:27:55 +01:00
auditfilter.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
auditsc.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
backtracetest.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
bounds.c bounds: Use the right number of bits for power-of-two CONFIG_NR_CPUS 2024-11-19 11:32:40 +01:00
capability.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
cfi.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
compat.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
configs.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
context_tracking.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
cpu.c cpu: Re-enable CPU mitigations by default for !X86 architectures 2024-11-19 11:32:38 +01:00
cpu_pm.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
crash_core.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
crash_dump.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
cred.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
delayacct.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
dma.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
exec_domain.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
exit.c security: samsung: defex_lsm: nuke 2024-06-15 16:20:49 -03:00
extable.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
fail_function.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
fork.c exec: Simplify unshare_files 2024-11-19 12:27:27 +01:00
freezer.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
gen_kheaders.sh kheaders: explicitly define file modes for archived headers 2024-11-19 14:19:30 +01:00
groups.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
hung_task.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
iomem.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
irq_work.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
jump_label.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
kallsyms.c kallsyms: only build {,module_}kallsyms_on_each_symbol when required 2024-11-19 12:27:33 +01:00
kcmp.c kcmp: In get_file_raw_ptr use task_lookup_fd_rcu 2024-11-19 12:27:27 +01:00
Kconfig.freezer Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
Kconfig.hz Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
Kconfig.locks Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
Kconfig.preempt Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
kcov.c kcov: don't lose track of remote references during softirqs 2024-11-19 14:19:10 +01:00
kexec.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
kexec_core.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
kexec_elf.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
kexec_file.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
kexec_internal.h Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
kheaders.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
kmod.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
kprobes.c kprobes: Fix possible use-after-free issue on kprobe registration 2024-11-19 11:32:21 +01:00
ksysfs.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
kthread.c exit: Implement kthread_exit 2024-11-19 12:27:49 +01:00
latencytop.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
Makefile kernel: Use the stock defconfig for /proc/config.gz 2024-06-15 16:20:14 -03:00
module-internal.h Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
module.c NFSD: Remove svc_serv_ops::svo_module 2024-11-19 12:27:54 +01:00
module_signature.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
module_signing.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
notifier.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
nsproxy.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
padata.c padata: Disable BH when taking works lock on MT path 2024-11-19 14:19:06 +01:00
panic.c panic: Flush kernel log buffer at the end 2024-11-19 09:23:11 +01:00
params.c params: lift param_set_uint_minmax to common code 2024-11-19 12:27:09 +01:00
pid.c kernel/pid.c: implement additional checks upon pidfd_create() parameters 2024-11-19 12:27:43 +01:00
pid_namespace.c zap_pid_ns_processes: clear TIF_NOTIFY_SIGNAL along with TIF_SIGPENDING 2024-11-19 14:19:05 +01:00
profile.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
ptrace.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
range.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
reboot.c kernel/reboot: emergency_restart: Set correct system_state 2024-11-18 11:43:25 +01:00
regset.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
relay.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
resource.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
rseq.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
scftorture.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
scs.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
seccomp.c seccomp: Invalidate seccomp mode to catch death failures 2024-11-18 22:25:35 +01:00
signal.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
smp.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
smpboot.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
smpboot.h Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
softirq.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
stackleak.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
stacktrace.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
static_call.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
stop_machine.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
sys.c Revert "fs: add file and path permissions helpers" 2024-11-19 13:30:21 +01:00
sys_ni.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
sysctl-test.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
sysctl.c sysctl: introduce new proc handler proc_dobool 2024-11-19 12:27:44 +01:00
task_work.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
taskstats.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
test_kprobes.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
torture.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
tracepoint.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
tsacct.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
ucount.c fanotify: configurable limits via sysfs 2024-11-19 12:27:37 +01:00
uid16.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
uid16.h Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
umh.c security: samsung: defex_lsm: nuke 2024-06-15 16:20:49 -03:00
up.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
user-return-notifier.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
user.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
user_namespace.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
usermode_driver.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
utsname.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
utsname_sysctl.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
watch_queue.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
watchdog.c watchdog: move softlockup_panic back to early_param 2024-11-18 11:43:21 +01:00
watchdog_hld.c Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00
workqueue.c Revert "workqueue: remove unused cancel_work()" 2024-11-18 12:11:11 +01:00
workqueue_internal.h Import A536BXXU9EXDC 2024-06-15 16:02:09 -03:00