The PERF_RECORD_THROTTLE records are dumped for all throttled events.
It's not necessary for group events, which are throttled altogether.
Optimize it by only dump the throttle log for the leader.
The sample right after the THROTTLE record must be generated by the
actual target event. It is good enough for the perf tool to locate the
actual target event.
Suggested-by: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20250520181644.2673067-3-kan.liang@linux.intel.com
The current throttle logic doesn't work well with a group, e.g., the
following sampling-read case.
$ perf record -e "{cycles,cycles}:S" ...
$ perf report -D | grep THROTTLE | tail -2
THROTTLE events: 426 ( 9.0%)
UNTHROTTLE events: 425 ( 9.0%)
$ perf report -D | grep PERF_RECORD_SAMPLE -a4 | tail -n 5
0 1020120874009167 0x74970 [0x68]: PERF_RECORD_SAMPLE(IP, 0x1):
... sample_read:
.... group nr 2
..... id 0000000000000327, value 000000000cbb993a, lost 0
..... id 0000000000000328, value 00000002211c26df, lost 0
The second cycles event has a much larger value than the first cycles
event in the same group.
The current throttle logic in the generic code only logs the THROTTLE
event. It relies on the specific driver implementation to disable
events. For all ARCHs, the implementation is similar. Only the event is
disabled, rather than the group.
The logic to disable the group should be generic for all ARCHs. Add the
logic in the generic code. The following patch will remove the buggy
driver-specific implementation.
The throttle only happens when an event is overflowed. Stop the entire
group when any event in the group triggers the throttle.
The MAX_INTERRUPTS is set to all throttle events.
The unthrottled could happen in 3 places.
- event/group sched. All events in the group are scheduled one by one.
All of them will be unthrottled eventually. Nothing needs to be
changed.
- The perf_adjust_freq_unthr_events for each tick. Needs to restart the
group altogether.
- The __perf_event_period(). The whole group needs to be restarted
altogether as well.
With the fix,
$ sudo perf report -D | grep PERF_RECORD_SAMPLE -a4 | tail -n 5
0 3573470770332 0x12f5f8 [0x70]: PERF_RECORD_SAMPLE(IP, 0x2):
... sample_read:
.... group nr 2
..... id 0000000000000a28, value 00000004fd3dfd8f, lost 0
..... id 0000000000000a29, value 00000004fd3dfd8f, lost 0
Suggested-by: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20250520181644.2673067-2-kan.liang@linux.intel.com
perf always allocates contiguous AUX pages based on aux_watermark.
However, this contiguous allocation doesn't benefit all PMUs. For
instance, ARM SPE and TRBE operate with virtual pages, and Coresight
ETR allocates a separate buffer. For these PMUs, allocating contiguous
AUX pages unnecessarily exacerbates memory fragmentation. This
fragmentation can prevent their use on long-running devices.
This patch modifies the perf driver to be memory-friendly by default,
by allocating non-contiguous AUX pages. For PMUs requiring contiguous
pages (Intel BTS and some Intel PT), the existing
PERF_PMU_CAP_AUX_NO_SG capability can be used. For PMUs that don't
require but can benefit from contiguous pages (some Intel PT), a new
capability, PERF_PMU_CAP_AUX_PREFER_LARGE, is added to maintain their
existing behavior.
Signed-off-by: Yabin Cui <yabinc@google.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: James Clark <james.clark@linaro.org>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20250508232642.148767-1-yabinc@google.com
While an event tears down all links to it as an aux, the iteration
happens on the event's group leader instead of the group itself.
If the event is a group leader, it has no effect because the event is
also its own group leader. But otherwise there would be a risk to detach
all the siblings events from the wrong group leader.
It just happens to work because each sibling's aux link is tested
against the right event before proceeding. Also the ctx lock is the same
for the events and their group leader so the iteration is safe.
Yet the iteration is confusing. Clarify the actual intent.
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20250424161128.29176-5-frederic@kernel.org
The CPU hotplug handlers are called twice: at prepare and online stage.
Their role is to:
1) Enable/disable a CPU context. This is irrelevant and even buggy at
the prepare stage because the CPU is still offline. On early
secondary CPU up, creating an event attached to that CPU might
silently fail because the CPU context is observed as online but the
context installation's IPI failure is ignored.
2) Update the scope cpumasks and re-migrate the events accordingly in
the CPU down case. This is irrelevant at the prepare stage.
3) Remove the events attached to the context of the offlining CPU. It
even uses an (unnecessary) IPI for it. This is also irrelevant at the
prepare stage.
Also none of the *_PREPARE and *_STARTING architecture perf related CPU
hotplug callbacks rely on CPUHP_PERF_PREPARE.
CPUHP_AP_PERF_ONLINE is enough and the right place to perform the work.
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20250424161128.29176-4-frederic@kernel.org
The following commit:
da916e96e2 ("perf: Make perf_pmu_unregister() useable")
has introduced two significant event's parent lifecycle changes:
1) An event that has exited now has EVENT_TOMBSTONE as a parent.
This can result in a situation where the delayed wakeup irq_work can
accidentally dereference EVENT_TOMBSTONE on:
CPU 0 CPU 1
----- -----
__schedule()
local_irq_disable()
rq_lock()
<NMI>
perf_event_overflow()
irq_work_queue(&child->pending_irq)
</NMI>
perf_event_task_sched_out()
raw_spin_lock(&ctx->lock)
ctx_sched_out()
ctx->is_active = 0
event_sched_out(child)
raw_spin_unlock(&ctx->lock)
perf_event_release_kernel(parent)
perf_remove_from_context(child)
raw_spin_lock_irq(&ctx->lock)
// Sees !ctx->is_active
// Removes from context inline
__perf_remove_from_context(child)
perf_child_detach(child)
event->parent = EVENT_TOMBSTONE
raw_spin_rq_unlock_irq(rq);
<IRQ>
perf_pending_irq()
perf_event_wakeup(child)
ring_buffer_wakeup(child)
rcu_dereference(child->parent->rb) <--- CRASH
This also concerns the call to kill_fasync() on parent->fasync.
2) The final parent reference count decrement can now happen before the
the final child reference count decrement. ie: the parent can now
be freed before its child. On PREEMPT_RT, this can result in a
situation where the delayed wakeup irq_work can accidentally
dereference a freed parent:
CPU 0 CPU 1 CPU 2
----- ----- ------
perf_pmu_unregister()
pmu_detach_events()
pmu_get_event()
atomic_long_inc_not_zero(&child->refcount)
<NMI>
perf_event_overflow()
irq_work_queue(&child->pending_irq);
</NMI>
<IRQ>
irq_work_run()
wake_irq_workd()
</IRQ>
preempt_schedule_irq()
=========> SWITCH to workd
irq_work_run_list()
perf_pending_irq()
perf_event_wakeup(child)
ring_buffer_wakeup(child)
event = child->parent
perf_event_release_kernel(parent)
// Not last ref, PMU holds it
put_event(child)
// Last ref
put_event(parent)
free_event()
call_rcu(...)
rcu_core()
free_event_rcu()
rcu_dereference(event->rb) <--- CRASH
This also concerns the call to kill_fasync() on parent->fasync.
The "easy" solution to 1) is to check that event->parent is not
EVENT_TOMBSTONE on perf_event_wakeup() (including both ring buffer
and fasync uses).
The "easy" solution to 2) is to turn perf_event_wakeup() to wholefully
run under rcu_read_lock().
However because of 2), sanity would prescribe to make event::parent
an __rcu pointer and annotate each and every users to prove they are
reliable.
Propose an alternate solution and restore the stable pointer to the
parent until all its children have called _free_event() themselves to
avoid any further accident. Also revert the EVENT_TOMBSTONE design
that is mostly here to determine which caller of perf_event_exit_event()
must perform the refcount decrement on a child event matching the
increment in inherit_event().
Arrange instead for checking the attach state of an event prior to its
removal and decrement the refcount of the child accordingly.
Fixes: da916e96e2 ("perf: Make perf_pmu_unregister() useable")
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
When inherit_event() fails after the child allocation but before the
parent refcount has been incremented, calling put_event() wrongly
decrements the reference to the parent, risking to free it too early.
Also pmu_get_event() can't be holding a reference to the child
concurrently at this point since it is under pmus_srcu critical section.
Fix it with restoring the deleted free_event() function and call it on
the failing child in order to free it directly under the verified
assumption that its refcount is only 1. The refcount to the parent is
then voluntarily omitted.
Fixes: da916e96e2 ("perf: Make perf_pmu_unregister() useable")
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20250424161128.29176-2-frederic@kernel.org
According to the throttling mechanism, the pmu interrupts number can not
exceed the max_samples_per_tick in one tick. But this mechanism is
ineffective when max_samples_per_tick=1, because the throttling check is
skipped during the first interrupt and only performed when the second
interrupt arrives.
Perhaps this bug may cause little influence in one tick, but if in a
larger time scale, the problem can not be underestimated.
When max_samples_per_tick = 1:
Allowed-interrupts-per-second max-samples-per-second default-HZ ARCH
200 100 100 X86
500 250 250 ARM64
...
Obviously, the pmu interrupt number far exceed the user's expect.
Fixes: e050e3f0a7 ("perf: Fix broken interrupt rate throttling")
Signed-off-by: Qing Wang <wangqing7171@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20250405141635.243786-3-wangqing7171@gmail.com
Pull sched_ext fixes from Tejun Heo:
- Use kvzalloc() so that large exit_dump buffer allocations don't fail
easily
- Remove cpu.weight / cpu.idle unimplemented warnings which are more
annoying than helpful.
This makes SCX_OPS_HAS_CGROUP_WEIGHT unnecessary. Mark it for
deprecation
* tag 'sched_ext-for-6.15-rc3-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext:
sched_ext: Mark SCX_OPS_HAS_CGROUP_WEIGHT for deprecation
sched_ext: Remove cpu.weight / cpu.idle unimplemented warnings
sched_ext: Use kvzalloc for large exit_dump allocation
Pull cgroup fixes from Tejun Heo:
- Fix compilation in CONFIG_LOCKDEP && !CONFIG_PROVE_RCU configurations
- Allow "cpuset_v2_mode" mount option for "cpuset" filesystem type to
make life easier for android
* tag 'cgroup-for-6.15-rc3-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup/cpuset-v1: Add missing support for cpuset_v2_mode
cgroup: Fix compilation issue due to cgroup_mutex not being exported
Pull vfs fixes from Christian Brauner:
- Revert the hfs{plus} deprecation warning that's also included in this
pull request. The commit introducing the deprecation warning resides
rather early in this branch. So simply dropping it would've rebased
all other commits which I decided to avoid. Hence the revert in the
same branch
[ Background - the deprecation warning discussion resulted in people
stepping up, and so hfs{plus} will have a maintainer taking care of
it after all.. - Linus ]
- Switch CONFIG_SYSFS_SYCALL default to n and decouple from
CONFIG_EXPERT
- Fix an audit bug caused by changes to our kernel path lookup helpers
this cycle. Audit needs the parent path even if the dentry it tried
to look up is negative
- Ensure that the kernel path lookup helpers leave the passed in path
argument clean when they return an error. This is consistent with all
our other helpers
- Ensure that vfs_getattr_nosec() calls bdev_statx() so the relevant
information is available to kernel consumers as well
- Don't set a timer and call schedule() if the timer will expire
immediately in epoll
- Make netfs lookup tables with __nonstring
* tag 'vfs-6.15-rc3.fixes.2' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
Revert "hfs{plus}: add deprecation warning"
fs: move the bdex_statx call to vfs_getattr_nosec
netfs: Mark __nonstring lookup tables
eventpoll: Set epoll timeout if it's in the future
fs: ensure that *path_locked*() helpers leave passed path pristine
fs: add kern_path_locked_negative()
hfs{plus}: add deprecation warning
Kconfig: switch CONFIG_SYSFS_SYCALL default to n
Pull tracing fixes from Steven Rostedt:
- Initialize hash variables in ftrace subops logic
The fix that simplified the ftrace subops logic opened a path where
some variables could be used without being initialized, and done
subtly where the compiler did not catch it. Initialize those
variables to the EMPTY_HASH, which is the default hash.
- Reinitialize the hash pointers after they are freed
Some of the hash pointers in the subop logic were freed but may still
be referenced later. To prevent use-after-free bugs, initialize them
back to the EMPTY_HASH.
- Free the ftrace hashes when they are replaced
The fix that simplified the subops logic updated some hash pointers,
but left the original hash that they were pointing to where they are
no longer used. This caused a memory leak. Free the hashes that are
pointed to by the pointers when they are replaced.
- Fix size initialization of ftrace direct function hash
The ftrace direct function hash used by BPF initialized the hash size
incorrectly. It checked the size of items to a hard coded 32, which
made the hash bit size of 5. The hash size is supposed to be limited
by the bit size of the hash, as the bitmask is allowed to be greater
than 5. Rework the size check to first pass the number of elements to
fls() and then compare that to FTRACE_HASH_MAX_BITS before allocating
the hash.
- Fix format output of ftrace_graph_ent_entry event
The field depth of the ftrace_graph_ent_entry event is of size 4 but
the output showed it as unsigned long and use "%lu". Change it to
unsigned int and use "%u" in the print format that is displayed to
user space.
- Fix the trace event filter on strings
Events can be filtered on numbers or string values. The return value
checked from strncpy_from_kernel_nofault() and
strncpy_from_user_nofault() was used to determine if reading the
strings would fault or not. It would return fault if the value was
non zero, which is basically meant that it was always considering the
read as a fault.
- Add selftest to test trace event string filtering
In order to catch the breakage of the string filtering, add a self
test to make sure that it continues to work.
* tag 'trace-v6.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
tracing: selftests: Add testing a user string to filters
tracing: Fix filter string testing
ftrace: Fix type of ftrace_graph_ent_entry.depth
ftrace: fix incorrect hash size in register_ftrace_direct()
ftrace: Free ftrace hashes after they are replaced in the subops code
ftrace: Reinitialize hash to EMPTY_HASH after freeing
ftrace: Initialize variables for ftrace_startup/shutdown_subops()
Android has mounted the v1 cpuset controller using filesystem type
"cpuset" (not "cgroup") since 2015 [1], and depends on the resulting
behavior where the controller name is not added as a prefix for cgroupfs
files. [2]
Later, a problem was discovered where cpu hotplug onlining did not
affect the cpuset/cpus files, which Android carried an out-of-tree patch
to address for a while. An attempt was made to upstream this patch, but
the recommendation was to use the "cpuset_v2_mode" mount option
instead. [3]
An effort was made to do so, but this fails with "cgroup: Unknown
parameter 'cpuset_v2_mode'" because commit e1cba4b85d ("cgroup: Add
mount flag to enable cpuset to use v2 behavior in v1 cgroup") did not
update the special cased cpuset_mount(), and only the cgroup (v1)
filesystem type was updated.
Add parameter parsing to the cpuset filesystem type so that
cpuset_v2_mode works like the cgroup filesystem type:
$ mkdir /dev/cpuset
$ mount -t cpuset -ocpuset_v2_mode none /dev/cpuset
$ mount|grep cpuset
none on /dev/cpuset type cgroup (rw,relatime,cpuset,noprefix,cpuset_v2_mode,release_agent=/sbin/cpuset_release_agent)
[1] b769c8d24f
[2] https://cs.android.com/android/platform/superproject/main/+/main:system/core/libprocessgroup/setup/cgroup_map_write.cpp;drc=2dac5d89a0f024a2d0cc46a80ba4ee13472f1681;l=192
[3] https://lore.kernel.org/lkml/f795f8be-a184-408a-0b5a-553d26061385@redhat.com/T/
Fixes: e1cba4b85d ("cgroup: Add mount flag to enable cpuset to use v2 behavior in v1 cgroup")
Signed-off-by: T.J. Mercier <tjmercier@google.com>
Acked-by: Waiman Long <longman@redhat.com>
Reviewed-by: Kamalesh Babulal <kamalesh.babulal@oracle.com>
Acked-by: Michal Koutný <mkoutny@suse.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
When adding folio_memcg function call in the zram module for
Android16-6.12, the following error occurs during compilation:
ERROR: modpost: "cgroup_mutex" [../soc-repo/zram.ko] undefined!
This error is caused by the indirect call to lockdep_is_held(&cgroup_mutex)
within folio_memcg. The export setting for cgroup_mutex is controlled by
the CONFIG_PROVE_RCU macro. If CONFIG_LOCKDEP is enabled while
CONFIG_PROVE_RCU is not, this compilation error will occur.
To resolve this issue, add a parallel macro CONFIG_LOCKDEP control to
ensure cgroup_mutex is properly exported when needed.
Signed-off-by: gao xu <gaoxu2@honor.com>
Acked-by: Michal Koutný <mkoutny@suse.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Notice that ignore_dl_rate_limit() need not piggy back on the
limits_changed handling to achieve its goal (which is to enforce a
frequency update before its due time).
Namely, if sugov_should_update_freq() is updated to check
sg_policy->need_freq_update and return 'true' if it is set when
sg_policy->limits_changed is not set, ignore_dl_rate_limit() may
set the former directly instead of setting the latter, so it can
avoid hitting the memory barrier in sugov_should_update_freq().
Update the code accordingly.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Christian Loehle <christian.loehle@arm.com>
Link: https://patch.msgid.link/10666429.nUPlyArG6x@rjwysocki.net
The handling of the limits_changed flag in struct sugov_policy needs to
be explicitly synchronized to ensure that cpufreq policy limits updates
will not be missed in some cases.
Without that synchronization it is theoretically possible that
the limits_changed update in sugov_should_update_freq() will be
reordered with respect to the reads of the policy limits in
cpufreq_driver_resolve_freq() and in that case, if the limits_changed
update in sugov_limits() clobbers the one in sugov_should_update_freq(),
the new policy limits may not take effect for a long time.
Likewise, the limits_changed update in sugov_limits() may theoretically
get reordered with respect to the updates of the policy limits in
cpufreq_set_policy() and if sugov_should_update_freq() runs between
them, the policy limits change may be missed.
To ensure that the above situations will not take place, add memory
barriers preventing the reordering in question from taking place and
add READ_ONCE() and WRITE_ONCE() annotations around all of the
limits_changed flag updates to prevent the compiler from messing up
with that code.
Fixes: 600f5badb7 ("cpufreq: schedutil: Don't skip freq update when limits change")
Cc: 5.3+ <stable@vger.kernel.org> # 5.3+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Christian Loehle <christian.loehle@arm.com>
Link: https://patch.msgid.link/3376719.44csPzL39Z@rjwysocki.net
Commit 8e461a1cb4 ("cpufreq: schedutil: Fix superfluous updates caused
by need_freq_update") modified sugov_should_update_freq() to set the
need_freq_update flag only for drivers with CPUFREQ_NEED_UPDATE_LIMITS
set, but that flag generally needs to be set when the policy limits
change because the driver callback may need to be invoked for the new
limits to take effect.
However, if the return value of cpufreq_driver_resolve_freq() after
applying the new limits is still equal to the previously selected
frequency, the driver callback needs to be invoked only in the case
when CPUFREQ_NEED_UPDATE_LIMITS is set (which means that the driver
specifically wants its callback to be invoked every time the policy
limits change).
Update the code accordingly to avoid missing policy limits changes for
drivers without CPUFREQ_NEED_UPDATE_LIMITS.
Fixes: 8e461a1cb4 ("cpufreq: schedutil: Fix superfluous updates caused by need_freq_update")
Closes: https://lore.kernel.org/lkml/Z_Tlc6Qs-tYpxWYb@linaro.org/
Reported-by: Stephan Gerhold <stephan.gerhold@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Christian Loehle <christian.loehle@arm.com>
Link: https://patch.msgid.link/3010358.e9J7NaK4W3@rjwysocki.net
So there are three situations:
* If perf_event_free_task() has removed all the children from the parent list
before perf_event_release_kernel() got a chance to even iterate them, then
it's all good as there is no get_ctx() pending.
* If perf_event_release_kernel() iterates a child event, but it gets freed
meanwhile by perf_event_free_task() while the mutexes are temporarily
unlocked, it's all good because while locking again the ctx mutex,
perf_event_release_kernel() observes TASK_TOMBSTONE.
* But if perf_event_release_kernel() frees the child event before
perf_event_free_task() got a chance, we may face this scenario:
perf_event_release_kernel() perf_event_free_task()
-------------------------- ------------------------
mutex_lock(&event->child_mutex)
get_ctx(child->ctx)
mutex_unlock(&event->child_mutex)
mutex_lock(ctx->mutex)
mutex_lock(&event->child_mutex)
perf_remove_from_context(child)
mutex_unlock(&event->child_mutex)
mutex_unlock(ctx->mutex)
// This lock acquires ctx->refcount == 2
// visibility
mutex_lock(ctx->mutex)
ctx->task = TASK_TOMBSTONE
mutex_unlock(ctx->mutex)
wait_var_event()
// enters prepare_to_wait() since
// ctx->refcount == 2
// is guaranteed to be seen
set_current_state(TASK_INTERRUPTIBLE)
smp_mb()
if (ctx->refcount != 1)
schedule()
put_ctx()
// NOT fully ordered! Only RELEASE semantics
refcount_dec_and_test()
atomic_fetch_sub_release()
// So TASK_TOMBSTONE is not guaranteed to be seen
if (ctx->task == TASK_TOMBSTONE)
wake_up_var()
Basically it's a broken store buffer:
perf_event_release_kernel() perf_event_free_task()
-------------------------- ------------------------
ctx->task = TASK_TOMBSTONE smp_store_release(&ctx->refcount, ctx->refcount - 1)
smp_mb()
READ_ONCE(ctx->refcount) READ_ONCE(ctx->task)
So we need a smp_mb__after_atomic() before looking at ctx->task.
Fixes: 59f3aa4a3e ("perf: Simplify perf_event_free_task() wait")
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lkml.kernel.org/r/Z_ZvmEhjkAhplCBE@localhost.localdomain
The audit code relies on the fact that kern_path_locked() returned a
path even for a negative dentry. If it doesn't find a valid dentry it
immediately calls:
audit_find_parent(d_backing_inode(parent_path.dentry));
which assumes that parent_path.dentry is still valid. But it isn't since
kern_path_locked() has been changed to path_put() also for a negative
dentry.
Fix this by adding a helper that implements the required audit semantics
and allows us to fix the immediate bleeding. We can find a unified
solution for this afterwards.
Link: https://lore.kernel.org/20250414-rennt-wimmeln-f186c3a780f1@brauner
Fixes: 1c3cb50b58 ("VFS: change kern_path_locked() and user_path_locked_at() to never return negative dentry")
Reported-and-tested-by: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Pull tracing fixes from Steven Rostedt:
- Hide get_vm_area() from MMUless builds
The function get_vm_area() is not defined when CONFIG_MMU is not
defined. Hide that function within #ifdef CONFIG_MMU.
- Fix output of synthetic events when they have dynamic strings
The print fmt of the synthetic event's format file use to have "%.*s"
for dynamic size strings even though the user space exported
arguments had only __get_str() macro that provided just a nul
terminated string. This was fixed so that user space could parse this
properly.
But the reason that it had "%.*s" was because internally it provided
the maximum size of the string as one of the arguments. The fix that
replaced "%.*s" with "%s" caused the trace output (when the kernel
reads the event) to write "(efault)" as it would now read the length
of the string as "%s".
As the string provided is always nul terminated, there's no reason
for the internal code to use "%.*s" anyway. Just remove the length
argument to match the "%s" that is now in the format.
- Fix the ftrace subops hash logic of the manager ops hash
The function_graph uses the ftrace subops code. The subops code is a
way to have a single ftrace_ops registered with ftrace to determine
what functions will call the ftrace_ops callback. More than one user
of function graph can register a ftrace_ops with it. The function
graph infrastructure will then add this ftrace_ops as a subops with
the main ftrace_ops it registers with ftrace. This is because the
functions will always call the function graph callback which in turn
calls the subops ftrace_ops callbacks.
The main ftrace_ops must add a callback to all the functions that the
subops want a callback from. When a subops is registered, it will
update the main ftrace_ops hash to include the functions it wants.
This is the logic that was broken.
The ftrace_ops hash has a "filter_hash" and a "notrace_hash" where
all the functions in the filter_hash but not in the notrace_hash are
attached by ftrace. The original logic would have the main ftrace_ops
filter_hash be a union of all the subops filter_hashes and the main
notrace_hash would be a intersect of all the subops filter hashes.
But this was incorrect because the notrace hash depends on the
filter_hash it is associated to and not the union of all
filter_hashes.
Instead, when a subops is added, just include all the functions of
the subops hash that are in its filter_hash but not in its
notrace_hash. The main subops hash should not use its notrace hash,
unless all of its subops hashes have an empty filter_hash (which
means to attach to all functions), and then, and only then, the main
ftrace_ops notrace hash can be the intersect of all the subops
hashes.
This not only fixes the bug, but also simplifies the code.
- Add a selftest to better test the subops filtering
Add a selftest that would catch the bug fixed by the above change.
- Fix extra newline printed in function tracing with retval
The function parameter code changed the output logic slightly and
called print_graph_retval() and also printed a newline. The
print_graph_retval() also prints a newline which caused blank lines
to be printed in the function graph tracer when retval was added.
This caused one of the selftests to fail if retvals were enabled.
Instead remove the new line output from print_graph_retval() and have
the callers always print the new line so that it doesn't have to do
special logic if it calls print_graph_retval() or not.
- Fix out-of-bound memory access in the runtime verifier
When rv_is_container_monitor() is called on the last entry on the
link list it references the next entry, which is the list head and
causes an out-of-bound memory access.
* tag 'trace-v6.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
rv: Fix out-of-bound memory access in rv_is_container_monitor()
ftrace: Do not have print_graph_retval() add a newline
tracing/selftest: Add test to better test subops filtering of function graph
ftrace: Fix accounting of subop hashes
ftrace: Properly merge notrace hashes
tracing: Do not add length to print format in synthetic events
tracing: Hide get_vm_area() from MMUless builds
Pull bpf fixes from Alexei Starovoitov:
- Followup fixes for resilient spinlock (Kumar Kartikeya Dwivedi):
- Make res_spin_lock test less verbose, since it was spamming BPF
CI on failure, and make the check for AA deadlock stronger
- Fix rebasing mistake and use architecture provided
res_smp_cond_load_acquire
- Convert BPF maps (queue_stack and ringbuf) to resilient spinlock
to address long standing syzbot reports
- Make sure that classic BPF load instruction from SKF_[NET|LL]_OFF
offsets works when skb is fragmeneted (Willem de Bruijn)
* tag 'bpf-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
bpf: Convert ringbuf map to rqspinlock
bpf: Convert queue_stack map to rqspinlock
bpf: Use architecture provided res_smp_cond_load_acquire
selftests/bpf: Make res_spin_lock AA test condition stronger
selftests/net: test sk_filter support for SKF_NET_OFF on frags
bpf: support SKF_NET_OFF and SKF_LL_OFF on skb frags
selftests/bpf: Make res_spin_lock test less verbose
When rv_is_container_monitor() is called on the last monitor in
rv_monitors_list, KASAN yells:
BUG: KASAN: global-out-of-bounds in rv_is_container_monitor+0x101/0x110
Read of size 8 at addr ffffffff97c7c798 by task setup/221
The buggy address belongs to the variable:
rv_monitors_list+0x18/0x40
This is due to list_next_entry() is called on the last entry in the list.
It wraps around to the first list_head, and the first list_head is not
embedded in struct rv_monitor_def.
Fix it by checking if the monitor is last in the list.
Cc: stable@vger.kernel.org
Cc: Gabriele Monaco <gmonaco@redhat.com>
Fixes: cb85c660fc ("rv: Add option for nested monitors and include sched")
Link: https://lore.kernel.org/e85b5eeb7228bfc23b8d7d4ab5411472c54ae91b.1744355018.git.namcao@linutronix.de
Signed-off-by: Nam Cao <namcao@linutronix.de>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
The retval and retaddr options for function_graph tracer will add a
comment at the end of a function for both leaf and non leaf functions that
looks like:
__wake_up_common(); /* ret=0x1 */
} /* pick_next_task_fair ret=0x0 */
The function print_graph_retval() adds a newline after the "*/". But if
that's not called, the caller function needs to make sure there's a
newline added.
This is confusing and when the function parameters code was added, it
added a newline even when calling print_graph_retval() as the fact that
the print_graph_retval() function prints a newline isn't obvious.
This caused an extra newline to be printed and that made it fail the
selftests when the retval option was set, as the selftests were not
expecting blank lines being injected into the trace.
Instead of having print_graph_retval() print a newline, just have the
caller always print the newline regardless if it calls print_graph_retval()
or not. This not only fixes this bug, but it also simplifies the code.
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://lore.kernel.org/20250411133015.015ca393@gandalf.local.home
Reported-by: Mark Brown <broonie@kernel.org>
Tested-by: Mark Brown <broonie@kernel.org>
Closes: https://lore.kernel.org/all/ccc40f2b-4b9e-4abd-8daf-d22fce2a86f0@sirena.org.uk/
Fixes: ff5c9c576e ("ftrace: Add support for function argument to graph tracer")
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
The function graph infrastructure uses ftrace to hook to functions. It has
a single ftrace_ops to manage all the users of function graph. Each
individual user (tracing, bpf, fprobes, etc) has its own ftrace_ops to
track the functions it will have its callback called from. These
ftrace_ops are "subops" to the main ftrace_ops of the function graph
infrastructure.
Each ftrace_ops has a filter_hash and a notrace_hash that is defined as:
Only trace functions that are in the filter_hash but not in the
notrace_hash.
If the filter_hash is empty, it means to trace all functions.
If the notrace_hash is empty, it means do not disable any function.
The function graph main ftrace_ops needs to be a superset containing all
the functions to be traced by all the subops it has. The algorithm to
perform this merge was incorrect.
When the first subops was added to the main ops, it simply made the main
ops a copy of the subops (same filter_hash and notrace_hash).
When a second ops was added, it joined the new subops filter_hash with the
main ops filter_hash as a union of the two sets. The intersect between the
new subops notrace_hash and the main ops notrace_hash was created as the
new notrace_hash of the main ops.
The issue here is that it would then start tracing functions than no
subops were tracing. For example if you had two subops that had:
subops 1:
filter_hash = '*sched*' # trace all functions with "sched" in it
notrace_hash = '*time*' # except do not trace functions with "time"
subops 2:
filter_hash = '*lock*' # trace all functions with "lock" in it
notrace_hash = '*clock*' # except do not trace functions with "clock"
The intersect of '*time*' functions with '*clock*' functions could be the
empty set. That means the main ops will be tracing all functions with
'*time*' and all "*clock*" in it!
Instead, modify the algorithm to be a bit simpler and correct.
First, when adding a new subops, even if it's the first one, do not add
the notrace_hash if the filter_hash is not empty. Instead, just add the
functions that are in the filter_hash of the subops but not in the
notrace_hash of the subops into the main ops filter_hash. There's no
reason to add anything to the main ops notrace_hash.
The notrace_hash of the main ops should only be non empty iff all subops
filter_hashes are empty (meaning to trace all functions) and all subops
notrace_hashes include the same functions.
That is, the main ops notrace_hash is empty if any subops filter_hash is
non empty.
The main ops notrace_hash only has content in it if all subops
filter_hashes are empty, and the content are only functions that intersect
all the subops notrace_hashes. If any subops notrace_hash is empty, then
so is the main ops notrace_hash.
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Shuah Khan <skhan@linuxfoundation.org>
Cc: Andy Chiu <andybnac@gmail.com>
Link: https://lore.kernel.org/20250409152720.216356767@goodmis.org
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Pull misc timer fixes from Ingo Molnar:
- Fix missing ACCESS_PRIVATE() that triggered a Sparse warning
- Fix lockdep false positive in tick_freeze() on CONFIG_PREEMPT_RT=y
- Avoid <vdso/unaligned.h> macro's variable shadowing to address build
warning that triggers under W=2 builds
* tag 'timers-urgent-2025-04-10' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
vdso: Address variable shadowing in macros
timekeeping: Add a lockdep override in tick_freeze()
hrtimer: Add missing ACCESS_PRIVATE() for hrtimer::function
Pull misc perf events fixes from Ingo Molnar:
- Fix __free_event() corner case splat
- Fix false-positive uprobes related lockdep splat on
CONFIG_PREEMPT_RT=y kernels
- Fix a complicated perf sigtrap race that may result in hangs
* tag 'perf-urgent-2025-04-10' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf: Fix hang while freeing sigtrap event
uprobes: Avoid false-positive lockdep splat on CONFIG_PREEMPT_RT=y in the ri_timer() uprobe timer callback, use raw_write_seqcount_*()
perf/core: Fix WARN_ON(!ctx) in __free_event() for partial init
The following causes a vsnprintf fault:
# echo 's:wake_lat char[] wakee; u64 delta;' >> /sys/kernel/tracing/dynamic_events
# echo 'hist:keys=pid:ts=common_timestamp.usecs if !(common_flags & 0x18)' > /sys/kernel/tracing/events/sched/sched_waking/trigger
# echo 'hist:keys=next_pid:delta=common_timestamp.usecs-$ts:onmatch(sched.sched_waking).trace(wake_lat,next_comm,$delta)' > /sys/kernel/tracing/events/sched/sched_switch/trigger
Because the synthetic event's "wakee" field is created as a dynamic string
(even though the string copied is not). The print format to print the
dynamic string changed from "%*s" to "%s" because another location
(__set_synth_event_print_fmt()) exported this to user space, and user
space did not need that. But it is still used in print_synth_event(), and
the output looks like:
<idle>-0 [001] d..5. 193.428167: wake_lat: wakee=(efault)sshd-sessiondelta=155
sshd-session-879 [001] d..5. 193.811080: wake_lat: wakee=(efault)kworker/u34:5delta=58
<idle>-0 [002] d..5. 193.811198: wake_lat: wakee=(efault)bashdelta=91
bash-880 [002] d..5. 193.811371: wake_lat: wakee=(efault)kworker/u35:2delta=21
<idle>-0 [001] d..5. 193.811516: wake_lat: wakee=(efault)sshd-sessiondelta=129
sshd-session-879 [001] d..5. 193.967576: wake_lat: wakee=(efault)kworker/u34:5delta=50
The length isn't needed as the string is always nul terminated. Just print
the string and not add the length (which was hard coded to the max string
length anyway).
Cc: stable@vger.kernel.org
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Tom Zanussi <zanussi@kernel.org>
Cc: Douglas Raillard <douglas.raillard@arm.com>
Acked-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Link: https://lore.kernel.org/20250407154139.69955768@gandalf.local.home
Fixes: 4d38328eb4 ("tracing: Fix synth event printk format for str fields");
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Pull probes fixes from Masami Hiramatsu:
- fprobe: remove fprobe_hlist_node when module unloading
When a fprobe target module is removed, the fprobe_hlist_node should
be removed from the fprobe's hash table to prevent reusing
accidentally if another module is loaded at the same address.
- fprobe: lock module while registering fprobe
The module containing the function to be probeed is locked using a
reference counter until the fprobe registration is complete, which
prevents use after free.
- fprobe-events: fix possible UAF on modules
Basically as same as above, but in the fprobe-events layer we also
need to get module reference counter when we find the tracepoint in
the module.
* tag 'probes-fixes-v6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
tracing: fprobe: Cleanup fprobe hash when module unloading
tracing: fprobe events: Fix possible UAF on modules
tracing: fprobe: Fix to lock module while registering fprobe
Pull cgroup fixes from Tejun Heo:
- A number of cpuset remote partition related fixes and cleanups along
with selftest updates.
- A change from this merge window made cgroup_rstat_updated_list()
called outside cgroup_rstat_lock leading to list corruptions. Fix it
by relocating the call inside the lock.
* tag 'cgroup-for-6.15-rc1-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup/cpuset: Fix race between newly created partition and dying one
cgroup: rstat: call cgroup_rstat_updated_list with cgroup_rstat_lock
selftest/cgroup: Add a remote partition transition test to test_cpuset_prs.sh
selftest/cgroup: Clean up and restructure test_cpuset_prs.sh
selftest/cgroup: Update test_cpuset_prs.sh to use | as effective CPUs and state separator
cgroup/cpuset: Remove unneeded goto in sched_partition_write() and rename it
cgroup/cpuset: Code cleanup and comment update
cgroup/cpuset: Don't allow creation of local partition over a remote one
cgroup/cpuset: Remove remote_partition_check() & make update_cpumasks_hier() handle remote partition
cgroup/cpuset: Fix error handling in remote_partition_disable()
cgroup/cpuset: Fix incorrect isolated_cpus update in update_parent_effective_cpumask()
Previously it was only safe to call perf_pmu_unregister() if there
were no active events of that pmu around -- which was impossible to
guarantee since it races all sorts against perf_init_event().
Rework the whole thing by:
- keeping track of all events for a given pmu
- 'hiding' the pmu from perf_init_event()
- waiting for the appropriate (s)rcu grace periods such that all
prior references to the PMU will be completed
- detaching all still existing events of that pmu (see first point)
and moving them to a new REVOKED state.
- actually freeing the pmu data.
Where notably the new REVOKED state must inhibit all event actions
from reaching code that wants to use event->pmu.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lkml.kernel.org/r/20250307193723.525402029@infradead.org
The task passed to perf_event_exit_task() is not a child, it is
current. Fix this confusing naming, since much of the rest of the code
also relies on it being current.
Specifically, both exec() and exit() callers use it with current as
the argument.
Notably, task_ctx_sched_out() doesn't make much sense outside of
current.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lkml.kernel.org/r/20250307193305.486326750@infradead.org
