Steven Rostedt (Google)
cf221a7880
ring-buffer: Update "shortest_full" in polling ...
commit 1e0cb399c7653462d9dadf8ab9425337c355d358 upstream.
It was discovered that the ring buffer polling was incorrectly stating
that read would not block, but that's because polling did not take into
account that reads will block if the "buffer-percent" was set. Instead,
the ring buffer polling would say reads would not block if there was any
data in the ring buffer. This was incorrect behavior from a user space
point of view. This was fixed by commit 42fb0a1e84ff by having the polling
code check if the ring buffer had more data than what the user specified
"buffer percent" had.
The problem now is that the polling code did not register itself to the
writer that it wanted to wait for a specific "full" value of the ring
buffer. The result was that the writer would wake the polling waiter
whenever there was a new event. The polling waiter would then wake up, see
that there's not enough data in the ring buffer to notify user space and
then go back to sleep. The next event would wake it up again.
Before the polling fix was added, the code would wake up around 100 times
for a hackbench 30 benchmark. After the "fix", due to the constant waking
of the writer, it would wake up over 11,0000 times! It would never leave
the kernel, so the user space behavior was still "correct", but this
definitely is not the desired effect.
To fix this, have the polling code add what it's waiting for to the
"shortest_full" variable, to tell the writer not to wake it up if the
buffer is not as full as it expects to be.
Note, after this fix, it appears that the waiter is now woken up around 2x
the times it was before (~200). This is a tremendous improvement from the
11,000 times, but I will need to spend some time to see why polling is
more aggressive in its wakeups than the read blocking code.
Link: https://lore.kernel.org/linux-trace-kernel/20230929180113.01c2cae3@rorschach.local.home
Cc: stable@vger.kernel.org
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Fixes: 42fb0a1e84ff ("tracing/ring-buffer: Have polling block on watermark")
Reported-by: Julia Lawall <julia.lawall@inria.fr>
Tested-by: Julia Lawall <julia.lawall@inria.fr>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2023-10-10 21:46:41 +02:00
bpf: Avoid deadlock when using queue and stack maps from NMI
2023-10-10 21:46:36 +02:00
cgroup: Do not corrupt task iteration when rebinding subsystem
2023-06-28 10:18:36 +02:00
treewide: Replace DECLARE_TASKLET() with DECLARE_TASKLET_OLD()
2023-04-20 12:07:32 +02:00
treewide: Remove uninitialized_var() usage
2023-06-09 10:29:01 +02:00
perf: Fix function pointer case
2023-08-11 11:53:54 +02:00
gcov: add support for checksum field
2023-01-18 11:41:42 +01:00
genirq: Add IRQF_NO_AUTOEN for request_irq/nmi()
2023-05-17 11:35:46 +02:00
livepatch: fix race between fork and KLP transition
2022-10-26 13:22:18 +02:00
treewide: Remove uninitialized_var() usage
2023-06-09 10:29:01 +02:00
PM: hibernate: Fix mistake in kerneldoc comment
2023-01-18 11:40:53 +01:00
printk: fix return value of printk.devkmsg __setup handler
2022-04-15 14:18:08 +02:00
rcu: Suppress smp_processor_id() complaint in synchronize_rcu_expedited_wait()
2023-03-11 16:43:54 +01:00
sched/fair: Don't balance task to its current running CPU
2023-07-27 08:37:42 +02:00
posix-timers: Ensure timer ID search-loop limit is valid
2023-07-27 08:37:42 +02:00
ring-buffer: Update "shortest_full" in polling
2023-10-10 21:46:41 +02:00
kbuild: update config_data.gz only when the content of .config is changed
2021-05-11 14:04:16 +02:00
acct: fix potential integer overflow in encode_comp_t()
2023-01-18 11:41:34 +01:00
treewide: Remove uninitialized_var() usage
2023-06-09 10:29:01 +02:00
treewide: Remove uninitialized_var() usage
2023-06-09 10:29:01 +02:00
audit: log AUDIT_TIME_* records only from rules
2022-04-15 14:18:04 +02:00
audit: fix potential double free on error path from fsnotify_add_inode_mark
2022-09-05 10:27:38 +02:00
audit: move put_tree() to avoid trim_trees refcount underflow and UAF
2021-09-03 10:08:16 +02:00
audit: CONFIG_CHANGE don't log internal bookkeeping as an event
2020-10-01 13:17:32 +02:00
audit: fix a net reference leak in audit_list_rules_send()
2020-06-22 09:30:59 +02:00
audit: fix possible soft lockup in __audit_inode_child()
2023-09-23 10:59:46 +02:00
treewide: Replace DECLARE_TASKLET() with DECLARE_TASKLET_OLD()
2023-04-20 12:07:32 +02:00
sched_getaffinity: don't assume 'cpumask_size()' is fully initialized
2023-04-05 11:16:42 +02:00
random: clear fast pool, crng, and batches in cpuhp bring up
2022-06-22 14:11:12 +02:00
kernel/cpu_pm: Fix uninitted local in cpu_pm
2020-06-22 09:31:22 +02:00
treewide: Remove uninitialized_var() usage
2023-06-09 10:29:01 +02:00
kernel/extable.c: use address-of operator on section symbols
2023-06-09 10:29:01 +02:00
kernel/fail_function: fix memory leak with using debugfs_lookup()
2023-03-11 16:44:15 +01:00
kernel/fork: beware of __put_task_struct() calling context
2023-09-23 11:00:03 +02:00
treewide: Remove uninitialized_var() usage
2023-06-09 10:29:01 +02:00
kbuild: add variables for compression tools
2020-09-03 11:27:10 +02:00
kallsyms: Refactor kallsyms_show_value() to take cred
2020-07-16 08:16:44 +02:00
exec: Transform exec_update_mutex into a rw_semaphore
2021-01-09 13:44:55 +01:00
kexec: fix a memory leak in crash_shrink_memory()
2023-07-27 08:37:10 +02:00
kexec: support purgatories with .text.hot sections
2023-06-21 15:44:10 +02:00
kheaders: Use array declaration instead of char
2023-05-17 11:35:33 +02:00
kmod: make request_module() return an error when autoloading is disabled
2020-04-17 10:50:22 +02:00
x86/kprobes: Fix arch_check_optimized_kprobe check within optimized_kprobe range
2023-03-11 16:44:02 +01:00
kthread: Fix PF_KTHREAD vs to_kthread() race
2021-09-12 08:56:39 +02:00
kbuild: update config_data.gz only when the content of .config is changed
2021-05-11 14:04:16 +02:00
modules: only allow symbol_get of EXPORT_SYMBOL_GPL modules
2023-09-23 10:59:36 +02:00
module: harden ELF info handling
2021-04-07 14:47:38 +02:00
module: harden ELF info handling
2021-04-07 14:47:38 +02:00
kernel/notifier.c: intercept duplicate registrations to avoid infinite loops
2020-10-01 13:17:23 +02:00
padata: add separate cpuhp node for CPUHP_PADATA_DEAD
2020-06-17 16:40:22 +02:00
exit: Use READ_ONCE() for all oops/warn limit reads
2023-02-06 07:52:50 +01:00
memcg: enable accounting for pids in nested pid namespaces
2021-09-22 12:26:37 +02:00
profiling: fix shift too large makes kernel panic
2022-08-25 11:18:02 +02:00
ptrace: Reimplement PTRACE_KILL by always sending SIGKILL
2022-06-14 18:11:24 +02:00
reboot: fix overflow parsing reboot cpu number
2020-11-18 19:20:30 +01:00
relayfs: fix out-of-bounds access in relay_file_read
2023-05-17 11:35:58 +02:00
/dev/mem: Revoke mappings when a driver claims the region
2020-06-24 17:50:35 +02:00
seccomp: Invalidate seccomp mode to catch death failures
2022-02-16 12:52:53 +01:00
signal handling: don't use BUG_ON() for debugging
2022-07-21 20:59:27 +02:00
smp: Fix offline cpu check in flush_smp_call_function_queue()
2022-04-20 09:19:39 +02:00
kthread: Extract KTHREAD_IS_PER_CPU
2021-02-07 15:35:49 +01:00
prlimit: do_prlimit needs to have a speculation check
2023-01-24 07:17:59 +01:00
kernel/sys_ni: add compat entry for fadvise64_64
2022-09-05 10:27:38 +02:00
kernel/sysctl-test: Add null pointer test for sysctl.c:proc_dointvec()
2020-10-01 13:17:10 +02:00
mm: allow a controlled amount of unfairness in the page lock
2023-08-30 16:27:26 +02:00
tracepoint: Add tracepoint_probe_register_may_exist() for BPF tracing
2021-07-14 16:53:08 +02:00
taskstats: Cleanup the use of task->exit_code
2022-02-23 11:59:57 +01:00
usermodehelper: reset umask to default before executing user process
2020-10-14 10:32:58 +02:00
smp: Fix smp_call_function_single_async prototype
2021-05-14 09:44:33 +02:00
watchdog: export lockup_detector_reconfigure
2022-08-25 11:18:37 +02:00
watchdog/perf: more properly prevent false positives with turbo modes
2023-07-27 08:37:10 +02:00
workqueue: clean up WORK_* constant types, clarify masking
2023-07-27 08:37:29 +02:00
cf221a7880