The below commit introduces a dummy lockdep map, but didn't get
the initialization quite right (it should mimic the initialization
of the real ww_mutex lockdep maps). It also introduced a separate
locking api selftest failure. Fix these.
Closes: https://lore.kernel.org/lkml/Zw19sMtnKdyOVQoh@boqun-archlinux/
Fixes: 823a566221 ("locking/ww_mutex: Adjust to lockdep nest_lock requirements")
Reported-by: Boqun Feng <boqun.feng@gmail.com>
Suggested-by: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20241127085430.3045-1-thomas.hellstrom@linux.intel.com
Currently REACHABLE is weird for being on the instruction after the
instruction it modifies.
Since all REACHABLE annotations have an explicit instruction, flip
them around.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
Link: https://lore.kernel.org/r/20241128094312.494176035@infradead.org
There are no users of annotate_reachable() left.
And the annotate_unreachable() usage in unreachable() is plain wrong;
it will hide dangerous fall-through code-gen.
Remove both.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
Link: https://lore.kernel.org/r/20241128094312.235637588@infradead.org
annotate_reachable() is unreliable since the compiler is free to place
random code inbetween two consecutive asm() statements.
This removes the last and only annotate_reachable() user.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
Link: https://lore.kernel.org/r/20241128094312.133437051@infradead.org
Avoid unreachable() as it can (and will in the absence of UBSAN)
generate fallthrough code. Use BUG() so we get a UD2 trap (with
unreachable annotation).
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
Link: https://lore.kernel.org/r/20241128094312.028316261@infradead.org
Since barrier_before_unreachable() is empty for !GCC it is trivial to
unify the two definitions. Less is more.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
Link: https://lore.kernel.org/r/20241128094311.924381359@infradead.org
Reduce read_annotate() runs by collapsing subsequent runs into a
single call.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
Link: https://lore.kernel.org/r/20241128094311.688871544@infradead.org
Avoid endless .discard.foo sections for each annotation, create a
single .discard.annotate_insn section that takes an annotation type along
with the instruction.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
Link: https://lore.kernel.org/r/20241128094310.932794537@infradead.org
At the point where find_active_uprobe_rcu() is used we know that VMA in
question has triggered software breakpoint, so we don't need to validate
vma->vm_flags. Keep only vma->vm_file NULL check.
Suggested-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Link: https://lkml.kernel.org/r/20241122035922.3321100-2-andrii@kernel.org
Add helper functions to speculatively perform operations without
read-locking mmap_lock, expecting that mmap_lock will not be
write-locked and mm is not modified from under us.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Liam R. Howlett <Liam.Howlett@Oracle.com>
Link: https://lkml.kernel.org/r/20241122174416.1367052-3-surenb@google.com
Convert mm_lock_seq to be seqcount_t and change all mmap_write_lock
variants to increment it, in-line with the usual seqcount usage pattern.
This lets us check whether the mmap_lock is write-locked by checking
mm_lock_seq.sequence counter (odd=locked, even=unlocked). This will be
used when implementing mmap_lock speculation functions.
As a result vm_lock_seq is also change to be unsigned to match the type
of mm_lock_seq.sequence.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Liam R. Howlett <Liam.Howlett@Oracle.com>
Link: https://lkml.kernel.org/r/20241122174416.1367052-2-surenb@google.com
David pointed out that gup_fast() does exactly what the new
raw_seqcount_try_begin() does -- use it.
Suggested-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: David Hildenbrand <david@redhat.com>
Add raw_seqcount_try_begin() to opens a read critical section of the given
seqcount_t if the counter is even. This enables eliding the critical
section entirely if the counter is odd, instead of doing the speculation
knowing it will fail.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Liam R. Howlett <Liam.Howlett@Oracle.com>
Link: https://lkml.kernel.org/r/20241122174416.1367052-1-surenb@google.com
Add a new "power_core" PMU and "energy-core" event for monitoring
energy consumption by each individual core. The existing energy-cores
event aggregates the energy consumption of CPU cores at the package level.
This new event aligns with the AMD's per-core energy counters.
Tested the package level and core level PMU counters with workloads
pinned to different CPUs.
Results with workload pinned to CPU 4 in core 4 on an AMD Zen4 Genoa
machine:
$ sudo perf stat --per-core -e power_core/energy-core/ -- taskset -c 4 stress-ng --matrix 1 --timeout 5s
stress-ng: info: [21250] setting to a 5 second run per stressor
stress-ng: info: [21250] dispatching hogs: 1 matrix
stress-ng: info: [21250] successful run completed in 5.00s
Performance counter stats for 'system wide':
S0-D0-C0 1 0.00 Joules power_core/energy-core/
S0-D0-C1 1 0.00 Joules power_core/energy-core/
S0-D0-C2 1 0.00 Joules power_core/energy-core/
S0-D0-C3 1 0.00 Joules power_core/energy-core/
S0-D0-C4 1 8.43 Joules power_core/energy-core/
S0-D0-C5 1 0.00 Joules power_core/energy-core/
S0-D0-C6 1 0.00 Joules power_core/energy-core/
S0-D0-C7 1 0.00 Joules power_core/energy-core/
S0-D1-C8 1 0.00 Joules power_core/energy-core/
S0-D1-C9 1 0.00 Joules power_core/energy-core/
S0-D1-C10 1 0.00 Joules power_core/energy-core/
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: "Gautham R. Shenoy" <gautham.shenoy@amd.com>
Link: https://lore.kernel.org/r/20241115060805.447565-11-Dhananjay.Ugwekar@amd.com
Prepare for the addition of RAPL core energy counter support.
Move cntr_mask to rapl_pmus struct instead of adding a new global
cntr_mask for the new RAPL power_core PMU. This will also ensure that
the second "core_cntr_mask" is only created if needed (i.e. in case of
AMD CPUs).
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: "Gautham R. Shenoy" <gautham.shenoy@amd.com>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20241115060805.447565-10-Dhananjay.Ugwekar@amd.com
Prepare for the addition of RAPL core energy counter support.
After making the rapl_model struct global, the rapl_msrs global
variable isn't needed, so remove it.
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: "Gautham R. Shenoy" <gautham.shenoy@amd.com>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20241115060805.447565-9-Dhananjay.Ugwekar@amd.com
Prepare for the addition of RAPL core energy counter support.
Replace the generic names with *_pkg*, to later on differentiate between
the scopes of the two different PMUs and their variables.
No functional change.
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: "Gautham R. Shenoy" <gautham.shenoy@amd.com>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20241115060805.447565-8-Dhananjay.Ugwekar@amd.com
Prepare for the addition of RAPL core energy counter support.
Add arguments to the init and cleanup functions, which will help in
initialization and cleaning up of two separate PMUs.
No functional change.
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: "Gautham R. Shenoy" <gautham.shenoy@amd.com>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20241115060805.447565-7-Dhananjay.Ugwekar@amd.com
Prepare for the addition of RAPL core energy counter support.
As there will always be just one rapl_model variable on a system, make it
global, to make it easier to access it from any function.
No functional change.
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Reviewed-by: "Gautham R. Shenoy" <gautham.shenoy@amd.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20241115060805.447565-6-Dhananjay.Ugwekar@amd.com
Rename struct rapl_pmu variables from "pmu" to "rapl_pmu", to
avoid any confusion between the variables of two different
structs pmu and rapl_pmu. As rapl_pmu also contains a pointer to
struct pmu, which leads to situations in code like pmu->pmu,
which is needlessly confusing. Above scenario is replaced with
much more readable rapl_pmu->pmu with this change.
Also rename "pmus" member in rapl_pmus struct, for same reason.
No functional change.
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: "Gautham R. Shenoy" <gautham.shenoy@amd.com>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20241115060805.447565-5-Dhananjay.Ugwekar@amd.com
Prepare for the addition of RAPL core energy counter support.
Post which, one CPU might be mapped to more than one rapl_pmu
(package/die one and a core one). So, remove the cpu_to_rapl_pmu()
function.
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20241115060805.447565-4-Dhananjay.Ugwekar@amd.com
On x86, topology_core_id() returns a unique core ID within the PKG
domain. Looking at match_smt() suggests that a core ID just needs to be
unique within a LLC domain. For use cases such as the core RAPL PMU,
there exists a need for a unique core ID across the entire system with
multiple PKG domains. Introduce topology_logical_core_id() to derive a
unique core ID across the system.
Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com>
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Reviewed-by: "Gautham R. Shenoy" <gautham.shenoy@amd.com>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Link: https://lore.kernel.org/r/20241115060805.447565-3-Dhananjay.Ugwekar@amd.com
commit 9e9af8bbb5 ("perf/x86/rapl: Clean up cpumask and hotplug")
removes the cpumask handling from rapl. Post that, we no longer need the
get_rapl_pmu_cpumask() function. So remove it.
Signed-off-by: Dhananjay Ugwekar <Dhananjay.Ugwekar@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20241115060805.447565-2-Dhananjay.Ugwekar@amd.com
The current code may iterate all the PEBS records in the DS area several
times. The first loop is to find all active events and calculate the
available records for each event. Then iterate the whole buffer again
and again to process available records until all active events are
processed.
The algorithm is inherited from the old generations. The old PEBS
hardware does not deal well with the situation when events happen near
each other. SW has to drop the error records. Multiple iterations are
required.
The hardware limit has been addressed on newer platforms with adaptive
PEBS. A simple one-iteration algorithm is introduced.
The samples are output by record order with the patch, rather than the
event order. It doesn't impact the post-processing. The perf tool always
sorts the records by time before presenting them to the end user.
In an NMI, the last record has to be specially handled. Add a last[]
variable to track the last unprocessed record of each event.
Test:
11 PEBS events are used in the perf test. Only the basic information is
collected.
perf record -e instructions:up,...,instructions:up -c 2000003 benchmark
The ftrace is used to record the duration of the
intel_pmu_drain_pebs_icl().
The average duration reduced from 62.04us to 57.94us.
A small improvement can be observed with the new algorithm.
Also, the implementation becomes simpler and more straightforward.
Suggested-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Link: https://lore.kernel.org/r/20241119135504.1463839-5-kan.liang@linux.intel.com
Factor out functions to process normal and the last PEBS records, which
can be shared with the later patch.
Move the event updating related codes (intel_pmu_save_and_restart())
to the end, where all samples have been processed.
For the current usage, it doesn't matter when perf updates event counts
and reset the counter. Because all counters are stopped when the PEBS
buffer is drained.
Drop the return of the !intel_pmu_save_and_restart(event) check. Because
it never happen. The intel_pmu_save_and_restart(event) only returns 0,
when !hwc->event_base or the period_left > 0.
- The !hwc->event_base is impossible for the PEBS event, since the PEBS
event is only available on GP and fixed counters, which always have
a valid hwc->event_base.
- The check only happens for the case of non-AUTO_RELOAD and single
PEBS, which implies that the event must be overflowed. The period_left
must be always <= 0 for an overflowed event after the
x86_pmu_update().
Co-developed-by: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Signed-off-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>
Link: https://lore.kernel.org/r/20241119135504.1463839-4-kan.liang@linux.intel.com
Modify the pebs_basic and pebs_meminfo structs to make the bitfields
more explicit to ease readability of the code.
Co-developed-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241119135504.1463839-3-kan.liang@linux.intel.com
The PEBS kernel warnings can still be observed with the below case.
when the below commands are running in parallel for a while.
while true;
do
perf record --no-buildid -a --intr-regs=AX \
-e cpu/event=0xd0,umask=0x81/pp \
-c 10003 -o /dev/null ./triad;
done &
while true;
do
perf record -e 'cpu/mem-loads,ldlat=3/uP' -W -d -- ./dtlb
done
The commit b752ea0c28 ("perf/x86/intel/ds: Flush PEBS DS when changing
PEBS_DATA_CFG") intends to flush the entire PEBS buffer before the
hardware is reprogrammed. However, it fails in the above case.
The first perf command utilizes the large PEBS, while the second perf
command only utilizes a single PEBS. When the second perf event is
added, only the n_pebs++. The intel_pmu_pebs_enable() is invoked after
intel_pmu_pebs_add(). So the cpuc->n_pebs == cpuc->n_large_pebs check in
the intel_pmu_drain_large_pebs() fails. The PEBS DS is not flushed.
The new PEBS event should not be taken into account when flushing the
existing PEBS DS.
The check is unnecessary here. Before the hardware is reprogrammed, all
the stale records must be drained unconditionally.
For single PEBS or PEBS-vi-pt, the DS must be empty. The drain_pebs()
can handle the empty case. There is no harm to unconditionally drain the
PEBS DS.
Fixes: b752ea0c28 ("perf/x86/intel/ds: Flush PEBS DS when changing PEBS_DATA_CFG")
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20241119135504.1463839-2-kan.liang@linux.intel.com
Andy reported that clang gets upset with CONFIG_CFS_BANDWIDTH=n:
kernel/sched/fair.c:6580:20: error: unused function 'cfs_bandwidth_used' [-Werror,-Wunused-function]
6580 | static inline bool cfs_bandwidth_used(void)
| ^~~~~~~~~~~~~~~~~~
Indeed, cfs_bandwidth_used() is only used within functions defined under
CONFIG_CFS_BANDWIDTH=y. Remove its CONFIG_CFS_BANDWIDTH=n declaration &
definition.
Reported-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Link: https://lore.kernel.org/r/20241127165501.160004-1-vschneid@redhat.com
After commit b58652db66 ("sched/deadline: Fix task_struct reference
leak"), I identified additional calls to hrtimer_try_to_cancel that
might also require a dl_server check. It remains unclear whether this
omission was intentional or accidental in those contexts.
This patch consolidates the timer cancellation logic into dedicated
functions, ensuring consistent behavior across all calls.
Additionally, it reduces code duplication and improves overall code
cleanliness.
Note the use of the __always_inline keyword. In some instances, we
have a task_struct pointer, dereference the dl member, and then use
the container_of macro to retrieve the task_struct pointer again. By
inlining the code, the compiler can potentially optimize out this
redundant round trip.
Signed-off-by: Wander Lairson Costa <wander@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Juri Lelli <juri.lelli@redhat.com>
Link: https://lore.kernel.org/r/20240724142253.27145-3-wander@redhat.com
Currently we check for bandwidth overflow potentially due to hotplug
operations at the end of sched_cpu_deactivate(), after the cpu going
offline has already been removed from scheduling, active_mask, etc.
This can create issues for DEADLINE tasks, as there is a substantial
race window between the start of sched_cpu_deactivate() and the moment
we possibly decide to roll-back the operation if dl_bw_deactivate()
returns failure in cpuset_cpu_inactive(). An example is a throttled
task that sees its replenishment timer firing while the cpu it was
previously running on is considered offline, but before
dl_bw_deactivate() had a chance to say no and roll-back happened.
Fix this by directly calling dl_bw_deactivate() first thing in
sched_cpu_deactivate() and do the required calculation in the former
function considering the cpu passed as an argument as offline already.
By doing so we also simplify sched_cpu_deactivate(), as there is no need
anymore for any kind of roll-back if we fail early.
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Phil Auld <pauld@redhat.com>
Tested-by: Waiman Long <longman@redhat.com>
Link: https://lore.kernel.org/r/Zzc1DfPhbvqDDIJR@jlelli-thinkpadt14gen4.remote.csb
For hotplug operations, DEADLINE needs to check that there is still enough
bandwidth left after removing the CPU that is going offline. We however
fail to do so currently.
Restore the correct behavior by restructuring dl_bw_manage() a bit, so
that overflow conditions (not enough bandwidth left) are properly
checked. Also account for dl_server bandwidth, i.e. discount such
bandwidth in the calculation since NORMAL tasks will be anyway moved
away from the CPU as a result of the hotplug operation.
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Phil Auld <pauld@redhat.com>
Tested-by: Waiman Long <longman@redhat.com>
Link: https://lore.kernel.org/r/20241114142810.794657-3-juri.lelli@redhat.com
When root domain non-destructive changes (e.g., only modifying one of
the existing root domains while the rest is not touched) happen we still
need to clear DEADLINE bandwidth accounting so that it's then properly
restored, taking into account DEADLINE tasks associated to each cpuset
(associated to each root domain). After the introduction of dl_servers,
we fail to restore such servers contribution after non-destructive
changes (as they are only considered on destructive changes when
runqueues are attached to the new domains).
Fix this by making sure we iterate over the dl_servers attached to
domains that have not been destroyed and add their bandwidth
contribution back correctly.
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Phil Auld <pauld@redhat.com>
Tested-by: Waiman Long <longman@redhat.com>
Link: https://lore.kernel.org/r/20241114142810.794657-2-juri.lelli@redhat.com
task_on_rq_queued read p->on_rq without READ_ONCE, though p->on_rq is
set with WRITE_ONCE in {activate|deactivate}_task and smp_store_release
in __block_task, and also read with READ_ONCE in task_on_rq_migrating.
Make all of these accesses pair together by adding READ_ONCE in the
task_on_rq_queued.
Signed-off-by: Harshit Agarwal <harshit@nutanix.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Phil Auld <pauld@redhat.com>
Link: https://lkml.kernel.org/r/20241114210812.1836587-1-jon@nutanix.com
When steal time exceeds the measured delta when updating clock_task, we
currently try to catch up the excess in future updates.
However, this results in inaccurate run times for the future things using
clock_task, in some situations, as they end up getting additional steal
time that did not actually happen.
This is because there is a window between reading the elapsed time in
update_rq_clock() and sampling the steal time in update_rq_clock_task().
If the VCPU gets preempted between those two points, any additional
steal time is accounted to the outgoing task even though the calculated
delta did not actually contain any of that "stolen" time.
When this race happens, we can end up with steal time that exceeds the
calculated delta, and the previous code would try to catch up that excess
steal time in future clock updates, which is given to the next,
incoming task, even though it did not actually have any time stolen.
This behavior is particularly bad when steal time can be very long,
which we've seen when trying to extend steal time to contain the duration
that the host was suspended [0]. When this happens, clock_task stays
frozen, during which the running task stays running for the whole
duration, since its run time doesn't increase.
However the race can happen even under normal operation.
Ideally we would read the elapsed cpu time and the steal time atomically,
to prevent this race from happening in the first place, but doing so
is non-trivial.
Since the time between those two points isn't otherwise accounted anywhere,
neither to the outgoing task nor the incoming task (because the "end of
outgoing task" and "start of incoming task" timestamps are the same),
I would argue that the right thing to do is to simply drop any excess steal
time, in order to prevent these issues.
[0] https://lore.kernel.org/kvm/20240820043543.837914-1-suleiman@google.com/
Signed-off-by: Suleiman Souhlal <suleiman@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20241118043745.1857272-1-suleiman@google.com
Anders had bisected a crash using PREEMPT_RT with linux-next and
isolated it down to commit 894d1b3db4 ("locking/mutex: Remove
wakeups from under mutex::wait_lock"), where it seemed the
wake_q structure was somehow getting corrupted causing a null
pointer traversal.
I was able to easily repoduce this with PREEMPT_RT and managed
to isolate down that through various call stacks we were
actually calling wake_up_q() twice on the same wake_q.
I found that in the problematic commit, I had added the
wake_up_q() call in task_blocks_on_rt_mutex() around
__ww_mutex_add_waiter(), following a similar pattern in
__mutex_lock_common().
However, its just wrong. We haven't dropped the lock->wait_lock,
so its contrary to the point of the original patch. And it
didn't match the __mutex_lock_common() logic of re-initializing
the wake_q after calling it midway in the stack.
Looking at it now, the wake_up_q() call is incorrect and should
just be removed. So drop the erronious logic I had added.
Fixes: 894d1b3db4 ("locking/mutex: Remove wakeups from under mutex::wait_lock")
Closes: https://lore.kernel.org/lkml/6afb936f-17c7-43fa-90e0-b9e780866097@app.fastmail.com/
Reported-by: Anders Roxell <anders.roxell@linaro.org>
Reported-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: John Stultz <jstultz@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Juri Lelli <juri.lelli@redhat.com>
Tested-by: Anders Roxell <anders.roxell@linaro.org>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20241114190051.552665-1-jstultz@google.com
When running the following command:
while true; do
stress-ng --cyclic 30 --timeout 30s --minimize --quiet
done
a warning is eventually triggered:
WARNING: CPU: 43 PID: 2848 at kernel/sched/deadline.c:794
setup_new_dl_entity+0x13e/0x180
...
Call Trace:
<TASK>
? show_trace_log_lvl+0x1c4/0x2df
? enqueue_dl_entity+0x631/0x6e0
? setup_new_dl_entity+0x13e/0x180
? __warn+0x7e/0xd0
? report_bug+0x11a/0x1a0
? handle_bug+0x3c/0x70
? exc_invalid_op+0x14/0x70
? asm_exc_invalid_op+0x16/0x20
enqueue_dl_entity+0x631/0x6e0
enqueue_task_dl+0x7d/0x120
__do_set_cpus_allowed+0xe3/0x280
__set_cpus_allowed_ptr_locked+0x140/0x1d0
__set_cpus_allowed_ptr+0x54/0xa0
migrate_enable+0x7e/0x150
rt_spin_unlock+0x1c/0x90
group_send_sig_info+0xf7/0x1a0
? kill_pid_info+0x1f/0x1d0
kill_pid_info+0x78/0x1d0
kill_proc_info+0x5b/0x110
__x64_sys_kill+0x93/0xc0
do_syscall_64+0x5c/0xf0
entry_SYSCALL_64_after_hwframe+0x6e/0x76
RIP: 0033:0x7f0dab31f92b
This warning occurs because set_cpus_allowed dequeues and enqueues tasks
with the ENQUEUE_RESTORE flag set. If the task is boosted, the warning
is triggered. A boosted task already had its parameters set by
rt_mutex_setprio, and a new call to setup_new_dl_entity is unnecessary,
hence the WARN_ON call.
Check if we are requeueing a boosted task and avoid calling
setup_new_dl_entity if that's the case.
Fixes: 295d6d5e37 ("sched/deadline: Fix switching to -deadline")
Signed-off-by: Wander Lairson Costa <wander@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Juri Lelli <juri.lelli@redhat.com>
Link: https://lore.kernel.org/r/20240724142253.27145-2-wander@redhat.com
Update the documentation for the `preempt=' parameter which now also
accepts `lazy'.
Fixes: 7c70cb94d2 ("sched: Add Lazy preemption model")
Reported-by: Shrikanth Hegde <sshegde@linux.ibm.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Shrikanth Hegde <sshegde@linux.ibm.com>
Link: https://lore.kernel.org/r/20241122173557.MYOtT95Q@linutronix.de
Scheduler raises a SCHED_SOFTIRQ to trigger a load balancing event on
from the IPI handler on the idle CPU. If the SMP function is invoked
from an idle CPU via flush_smp_call_function_queue() then the HARD-IRQ
flag is not set and raise_softirq_irqoff() needlessly wakes ksoftirqd
because soft interrupts are handled before ksoftirqd get on the CPU.
Adding a trace_printk() in nohz_csd_func() at the spot of raising
SCHED_SOFTIRQ and enabling trace events for sched_switch, sched_wakeup,
and softirq_entry (for SCHED_SOFTIRQ vector alone) helps observing the
current behavior:
<idle>-0 [000] dN.1.: nohz_csd_func: Raising SCHED_SOFTIRQ from nohz_csd_func
<idle>-0 [000] dN.4.: sched_wakeup: comm=ksoftirqd/0 pid=16 prio=120 target_cpu=000
<idle>-0 [000] .Ns1.: softirq_entry: vec=7 [action=SCHED]
<idle>-0 [000] .Ns1.: softirq_exit: vec=7 [action=SCHED]
<idle>-0 [000] d..2.: sched_switch: prev_comm=swapper/0 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=ksoftirqd/0 next_pid=16 next_prio=120
ksoftirqd/0-16 [000] d..2.: sched_switch: prev_comm=ksoftirqd/0 prev_pid=16 prev_prio=120 prev_state=S ==> next_comm=swapper/0 next_pid=0 next_prio=120
...
Use __raise_softirq_irqoff() to raise the softirq. The SMP function call
is always invoked on the requested CPU in an interrupt handler. It is
guaranteed that soft interrupts are handled at the end.
Following are the observations with the changes when enabling the same
set of events:
<idle>-0 [000] dN.1.: nohz_csd_func: Raising SCHED_SOFTIRQ for nohz_idle_balance
<idle>-0 [000] dN.1.: softirq_raise: vec=7 [action=SCHED]
<idle>-0 [000] .Ns1.: softirq_entry: vec=7 [action=SCHED]
No unnecessary ksoftirqd wakeups are seen from idle task's context to
service the softirq.
Fixes: b2a02fc43a ("smp: Optimize send_call_function_single_ipi()")
Closes: https://lore.kernel.org/lkml/fcf823f-195e-6c9a-eac3-25f870cb35ac@inria.fr/ [1]
Reported-by: Julia Lawall <julia.lawall@inria.fr>
Suggested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: https://lore.kernel.org/r/20241119054432.6405-5-kprateek.nayak@amd.com