| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core
Pull driver core fixes from Greg KH:
"Here are some small driver core and debugfs fixes for 6.0-rc5.
Included in here are:
- multiple attempts to get the arch_topology code to work properly on
non-cluster SMT systems. First attempt caused build breakages in
linux-next and 0-day, second try worked.
- debugfs fixes for a long-suffering memory leak. The pattern of
debugfs_remove(debugfs_lookup(...)) turns out to leak dentries, so
add debugfs_lookup_and_remove() to fix this problem. Also fix up
the scheduler debug code that highlighted this problem. Fixes for
other subsystems will be trickling in over the next few months for
this same issue once the debugfs function is merged.
All of these have been in linux-next since Wednesday with no reported
problems"
* tag 'driver-core-6.0-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
arch_topology: Make cluster topology span at least SMT CPUs
sched/debug: fix dentry leak in update_sched_domain_debugfs
debugfs: add debugfs_lookup_and_remove()
driver core: fix driver_set_override() issue with empty strings
Revert "arch_topology: Make cluster topology span at least SMT CPUs"
arch_topology: Make cluster topology span at least SMT CPUs
|
|
Kuyo reports that the pattern of using debugfs_remove(debugfs_lookup())
leaks a dentry and with a hotplug stress test, the machine eventually
runs out of memory.
Fix this up by using the newly created debugfs_lookup_and_remove() call
instead which properly handles the dentry reference counting logic.
Cc: Major Chen <major.chen@samsung.com>
Cc: stable <stable@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Ben Segall <bsegall@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Daniel Bristot de Oliveira <bristot@redhat.com>
Cc: Valentin Schneider <vschneid@redhat.com>
Cc: Matthias Brugger <matthias.bgg@gmail.com>
Reported-by: Kuyo Chang <kuyo.chang@mediatek.com>
Tested-by: Kuyo Chang <kuyo.chang@mediatek.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20220902123107.109274-2-gregkh@linuxfoundation.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
There are several places in the kernel where wait_on_bit is not followed
by a memory barrier (for example, in drivers/md/dm-bufio.c:new_read).
On architectures with weak memory ordering, it may happen that memory
accesses that follow wait_on_bit are reordered before wait_on_bit and
they may return invalid data.
Fix this class of bugs by introducing a new function "test_bit_acquire"
that works like test_bit, but has acquire memory ordering semantics.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Acked-by: Will Deacon <will@kernel.org>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
psi_trigger_create()'s 'nbytes' parameter is not used, so we can remove it.
Signed-off-by: Hao Jia <jiahao.os@bytedance.com>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
|
|
After commit 5f69a6577bc3 ("psi: dont alloc memory for psi by default"),
the memory used by struct psi_group is no longer allocated and zeroed
in cgroup_create().
Since the memory of struct psi_group is not zeroed, the data in this
memory is random, which will lead to inaccurate psi statistics when
creating a new cgroup.
So we use kzlloc() to allocate and zero the struct psi_group and
remove the redundant zeroing in group_init().
Steps to reproduce:
1. Use cgroup v2 and enable CONFIG_PSI
2. Create a new cgroup, and query psi statistics
mkdir /sys/fs/cgroup/test
cat /sys/fs/cgroup/test/cpu.pressure
some avg10=0.00 avg60=0.00 avg300=47927752200.00 total=12884901
full avg10=561815124.00 avg60=125835394188.00 avg300=1077090462000.00 total=10273561772
cat /sys/fs/cgroup/test/io.pressure
some avg10=1040093132823.95 avg60=1203770351379.21 avg300=3862252669559.46 total=4294967296
full avg10=921884564601.39 avg60=0.00 avg300=1984507298.35 total=442381631
cat /sys/fs/cgroup/test/memory.pressure
some avg10=232476085778.11 avg60=0.00 avg300=0.00 total=0
full avg10=0.00 avg60=0.00 avg300=2585658472280.57 total=12884901
Fixes: commit 5f69a6577bc3 ("psi: dont alloc memory for psi by default")
Signed-off-by: Hao Jia <jiahao.os@bytedance.com>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler fixes from Ingo Molnar:
"Various fixes: a deadline scheduler fix, a migration fix, a Sparse fix
and a comment fix"
* tag 'sched-urgent-2022-08-06' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/core: Do not requeue task on CPU excluded from cpus_mask
sched/rt: Fix Sparse warnings due to undefined rt.c declarations
exit: Fix typo in comment: s/sub-theads/sub-threads
sched, cpuset: Fix dl_cpu_busy() panic due to empty cs->cpus_allowed
|
|
The following warning was triggered on a large machine early in boot on
a distribution kernel but the same problem should also affect mainline.
WARNING: CPU: 439 PID: 10 at ../kernel/workqueue.c:2231 process_one_work+0x4d/0x440
Call Trace:
<TASK>
rescuer_thread+0x1f6/0x360
kthread+0x156/0x180
ret_from_fork+0x22/0x30
</TASK>
Commit c6e7bd7afaeb ("sched/core: Optimize ttwu() spinning on p->on_cpu")
optimises ttwu by queueing a task that is descheduling on the wakelist,
but does not check if the task descheduling is still allowed to run on that CPU.
In this warning, the problematic task is a workqueue rescue thread which
checks if the rescue is for a per-cpu workqueue and running on the wrong CPU.
While this is early in boot and it should be possible to create workers,
the rescue thread may still used if the MAYDAY_INITIAL_TIMEOUT is reached
or MAYDAY_INTERVAL and on a sufficiently large machine, the rescue
thread is being used frequently.
Tracing confirmed that the task should have migrated properly using the
stopper thread to handle the migration. However, a parallel wakeup from udev
running on another CPU that does not share CPU cache observes p->on_cpu and
uses task_cpu(p), queues the task on the old CPU and triggers the warning.
Check that the wakee task that is descheduling is still allowed to run
on its current CPU and if not, wait for the descheduling to complete
and select an allowed CPU.
Fixes: c6e7bd7afaeb ("sched/core: Optimize ttwu() spinning on p->on_cpu")
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20220804092119.20137-1-mgorman@techsingularity.net
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup
Pull cgroup updates from Tejun Heo:
"Several core optimizations:
- threadgroup_rwsem write locking is skipped when configuring
controllers in empty subtrees.
Combined with CLONE_INTO_CGROUP, this allows the common static
usage pattern to not grab threadgroup_rwsem at all (glibc still
doesn't seem ready for CLONE_INTO_CGROUP unfortunately).
- threadgroup_rwsem used to be put into non-percpu mode by default
due to latency concerns in specific use cases. There's no reason
for everyone else to pay for it. Make the behavior optional.
- psi no longer allocates memory when disabled.
... along with some code cleanups"
* tag 'cgroup-for-5.20' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup: Skip subtree root in cgroup_update_dfl_csses()
cgroup: remove "no" prefixed mount options
cgroup: Make !percpu threadgroup_rwsem operations optional
cgroup: Add "no" prefixed mount options
cgroup: Elide write-locking threadgroup_rwsem when updating csses on an empty subtree
cgroup.c: remove redundant check for mixable cgroup in cgroup_migrate_vet_dst
cgroup.c: add helper __cset_cgroup_from_root to cleanup duplicated codes
psi: dont alloc memory for psi by default
|
|
There are several symbols defined in kernel/sched/sched.h but get wrapped
in CONFIG_CGROUP_SCHED, even though dummy versions get built in rt.c and
therefore trigger Sparse warnings:
kernel/sched/rt.c:309:6: warning: symbol 'unregister_rt_sched_group' was not declared. Should it be static?
kernel/sched/rt.c:311:6: warning: symbol 'free_rt_sched_group' was not declared. Should it be static?
kernel/sched/rt.c:313:5: warning: symbol 'alloc_rt_sched_group' was not declared. Should it be static?
Fix this by moving them outside the CONFIG_CGROUP_SCHED block.
[ mingo: Refreshed to the latest scheduler tree, tweaked changelog. ]
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20220721145155.358366-1-ben-linux@fluff.org
|
|
With cgroup v2, the cpuset's cpus_allowed mask can be empty indicating
that the cpuset will just use the effective CPUs of its parent. So
cpuset_can_attach() can call task_can_attach() with an empty mask.
This can lead to cpumask_any_and() returns nr_cpu_ids causing the call
to dl_bw_of() to crash due to percpu value access of an out of bound
CPU value. For example:
[80468.182258] BUG: unable to handle page fault for address: ffffffff8b6648b0
:
[80468.191019] RIP: 0010:dl_cpu_busy+0x30/0x2b0
:
[80468.207946] Call Trace:
[80468.208947] cpuset_can_attach+0xa0/0x140
[80468.209953] cgroup_migrate_execute+0x8c/0x490
[80468.210931] cgroup_update_dfl_csses+0x254/0x270
[80468.211898] cgroup_subtree_control_write+0x322/0x400
[80468.212854] kernfs_fop_write_iter+0x11c/0x1b0
[80468.213777] new_sync_write+0x11f/0x1b0
[80468.214689] vfs_write+0x1eb/0x280
[80468.215592] ksys_write+0x5f/0xe0
[80468.216463] do_syscall_64+0x5c/0x80
[80468.224287] entry_SYSCALL_64_after_hwframe+0x44/0xae
Fix that by using effective_cpus instead. For cgroup v1, effective_cpus
is the same as cpus_allowed. For v2, effective_cpus is the real cpumask
to be used by tasks within the cpuset anyway.
Also update task_can_attach()'s 2nd argument name to cs_effective_cpus to
reflect the change. In addition, a check is added to task_can_attach()
to guard against the possibility that cpumask_any_and() may return a
value >= nr_cpu_ids.
Fixes: 7f51412a415d ("sched/deadline: Fix bandwidth check/update when migrating tasks between exclusive cpusets")
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Juri Lelli <juri.lelli@redhat.com>
Link: https://lore.kernel.org/r/20220803015451.2219567-1-longman@redhat.com
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu
Pull RCU updates from Paul McKenney:
- Documentation updates
- Miscellaneous fixes
- Callback-offload updates, perhaps most notably a new
RCU_NOCB_CPU_DEFAULT_ALL Kconfig option that causes all CPUs to be
offloaded at boot time, regardless of kernel boot parameters.
This is useful to battery-powered systems such as ChromeOS and
Android. In addition, a new RCU_NOCB_CPU_CB_BOOST kernel boot
parameter prevents offloaded callbacks from interfering with
real-time workloads and with energy-efficiency mechanisms
- Polled grace-period updates, perhaps most notably making these APIs
account for both normal and expedited grace periods
- Tasks RCU updates, perhaps most notably reducing the CPU overhead of
RCU tasks trace grace periods by more than a factor of two on a
system with 15,000 tasks.
The reduction is expected to increase with the number of tasks, so it
seems reasonable to hypothesize that a system with 150,000 tasks
might see a 20-fold reduction in CPU overhead
- Torture-test updates
- Updates that merge RCU's dyntick-idle tracking into context tracking,
thus reducing the overhead of transitioning to kernel mode from
either idle or nohz_full userspace execution for kernels that track
context independently of RCU.
This is expected to be helpful primarily for kernels built with
CONFIG_NO_HZ_FULL=y
* tag 'rcu.2022.07.26a' of git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu: (98 commits)
rcu: Add irqs-disabled indicator to expedited RCU CPU stall warnings
rcu: Diagnose extended sync_rcu_do_polled_gp() loops
rcu: Put panic_on_rcu_stall() after expedited RCU CPU stall warnings
rcutorture: Test polled expedited grace-period primitives
rcu: Add polled expedited grace-period primitives
rcutorture: Verify that polled GP API sees synchronous grace periods
rcu: Make Tiny RCU grace periods visible to polled APIs
rcu: Make polled grace-period API account for expedited grace periods
rcu: Switch polled grace-period APIs to ->gp_seq_polled
rcu/nocb: Avoid polling when my_rdp->nocb_head_rdp list is empty
rcu/nocb: Add option to opt rcuo kthreads out of RT priority
rcu: Add nocb_cb_kthread check to rcu_is_callbacks_kthread()
rcu/nocb: Add an option to offload all CPUs on boot
rcu/nocb: Fix NOCB kthreads spawn failure with rcu_nocb_rdp_deoffload() direct call
rcu/nocb: Invert rcu_state.barrier_mutex VS hotplug lock locking order
rcu/nocb: Add/del rdp to iterate from rcuog itself
rcu/tree: Add comment to describe GP-done condition in fqs loop
rcu: Initialize first_gp_fqs at declaration in rcu_gp_fqs()
rcu/kvfree: Remove useless monitor_todo flag
rcu: Cleanup RCU urgency state for offline CPU
...
|
|
Pull io_uring updates from Jens Axboe:
- As per (valid) complaint in the last merge window, fs/io_uring.c has
grown quite large these days. io_uring isn't really tied to fs
either, as it supports a wide variety of functionality outside of
that.
Move the code to io_uring/ and split it into files that either
implement a specific request type, and split some code into helpers
as well. The code is organized a lot better like this, and io_uring.c
is now < 4K LOC (me).
- Deprecate the epoll_ctl opcode. It'll still work, just trigger a
warning once if used. If we don't get any complaints on this, and I
don't expect any, then we can fully remove it in a future release
(me).
- Improve the cancel hash locking (Hao)
- kbuf cleanups (Hao)
- Efficiency improvements to the task_work handling (Dylan, Pavel)
- Provided buffer improvements (Dylan)
- Add support for recv/recvmsg multishot support. This is similar to
the accept (or poll) support for have for multishot, where a single
SQE can trigger everytime data is received. For applications that
expect to do more than a few receives on an instantiated socket, this
greatly improves efficiency (Dylan).
- Efficiency improvements for poll handling (Pavel)
- Poll cancelation improvements (Pavel)
- Allow specifiying a range for direct descriptor allocations (Pavel)
- Cleanup the cqe32 handling (Pavel)
- Move io_uring types to greatly cleanup the tracing (Pavel)
- Tons of great code cleanups and improvements (Pavel)
- Add a way to do sync cancelations rather than through the sqe -> cqe
interface, as that's a lot easier to use for some use cases (me).
- Add support to IORING_OP_MSG_RING for sending direct descriptors to a
different ring. This avoids the usually problematic SCM case, as we
disallow those. (me)
- Make the per-command alloc cache we use for apoll generic, place
limits on it, and use it for netmsg as well (me).
- Various cleanups (me, Michal, Gustavo, Uros)
* tag 'for-5.20/io_uring-2022-07-29' of git://git.kernel.dk/linux-block: (172 commits)
io_uring: ensure REQ_F_ISREG is set async offload
net: fix compat pointer in get_compat_msghdr()
io_uring: Don't require reinitable percpu_ref
io_uring: fix types in io_recvmsg_multishot_overflow
io_uring: Use atomic_long_try_cmpxchg in __io_account_mem
io_uring: support multishot in recvmsg
net: copy from user before calling __get_compat_msghdr
net: copy from user before calling __copy_msghdr
io_uring: support 0 length iov in buffer select in compat
io_uring: fix multishot ending when not polled
io_uring: add netmsg cache
io_uring: impose max limit on apoll cache
io_uring: add abstraction around apoll cache
io_uring: move apoll cache to poll.c
io_uring: consolidate hash_locked io-wq handling
io_uring: clear REQ_F_HASH_LOCKED on hash removal
io_uring: don't race double poll setting REQ_F_ASYNC_DATA
io_uring: don't miss setting REQ_F_DOUBLE_POLL
io_uring: disable multishot recvmsg
io_uring: only trace one of complete or overflow
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler updates from Ingo Molnar:
"Load-balancing improvements:
- Improve NUMA balancing on AMD Zen systems for affine workloads.
- Improve the handling of reduced-capacity CPUs in load-balancing.
- Energy Model improvements: fix & refine all the energy fairness
metrics (PELT), and remove the conservative threshold requiring 6%
energy savings to migrate a task. Doing this improves power
efficiency for most workloads, and also increases the reliability
of energy-efficiency scheduling.
- Optimize/tweak select_idle_cpu() to spend (much) less time
searching for an idle CPU on overloaded systems. There's reports of
several milliseconds spent there on large systems with large
workloads ...
[ Since the search logic changed, there might be behavioral side
effects. ]
- Improve NUMA imbalance behavior. On certain systems with spare
capacity, initial placement of tasks is non-deterministic, and such
an artificial placement imbalance can persist for a long time,
hurting (and sometimes helping) performance.
The fix is to make fork-time task placement consistent with runtime
NUMA balancing placement.
Note that some performance regressions were reported against this,
caused by workloads that are not memory bandwith limited, which
benefit from the artificial locality of the placement bug(s). Mel
Gorman's conclusion, with which we concur, was that consistency is
better than random workload benefits from non-deterministic bugs:
"Given there is no crystal ball and it's a tradeoff, I think
it's better to be consistent and use similar logic at both fork
time and runtime even if it doesn't have universal benefit."
- Improve core scheduling by fixing a bug in
sched_core_update_cookie() that caused unnecessary forced idling.
- Improve wakeup-balancing by allowing same-LLC wakeup of idle CPUs
for newly woken tasks.
- Fix a newidle balancing bug that introduced unnecessary wakeup
latencies.
ABI improvements/fixes:
- Do not check capabilities and do not issue capability check denial
messages when a scheduler syscall doesn't require privileges. (Such
as increasing niceness.)
- Add forced-idle accounting to cgroups too.
- Fix/improve the RSEQ ABI to not just silently accept unknown flags.
(No existing tooling is known to have learned to rely on the
previous behavior.)
- Depreciate the (unused) RSEQ_CS_FLAG_NO_RESTART_ON_* flags.
Optimizations:
- Optimize & simplify leaf_cfs_rq_list()
- Micro-optimize set_nr_{and_not,if}_polling() via try_cmpxchg().
Misc fixes & cleanups:
- Fix the RSEQ self-tests on RISC-V and Glibc 2.35 systems.
- Fix a full-NOHZ bug that can in some cases result in the tick not
being re-enabled when the last SCHED_RT task is gone from a
runqueue but there's still SCHED_OTHER tasks around.
- Various PREEMPT_RT related fixes.
- Misc cleanups & smaller fixes"
* tag 'sched-core-2022-08-01' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (32 commits)
rseq: Kill process when unknown flags are encountered in ABI structures
rseq: Deprecate RSEQ_CS_FLAG_NO_RESTART_ON_* flags
sched/core: Fix the bug that task won't enqueue into core tree when update cookie
nohz/full, sched/rt: Fix missed tick-reenabling bug in dequeue_task_rt()
sched/core: Always flush pending blk_plug
sched/fair: fix case with reduced capacity CPU
sched/core: Use try_cmpxchg in set_nr_{and_not,if}_polling
sched/core: add forced idle accounting for cgroups
sched/fair: Remove the energy margin in feec()
sched/fair: Remove task_util from effective utilization in feec()
sched/fair: Use the same cpumask per-PD throughout find_energy_efficient_cpu()
sched/fair: Rename select_idle_mask to select_rq_mask
sched, drivers: Remove max param from effective_cpu_util()/sched_cpu_util()
sched/fair: Decay task PELT values during wakeup migration
sched/fair: Provide u64 read for 32-bits arch helper
sched/fair: Introduce SIS_UTIL to search idle CPU based on sum of util_avg
sched: only perform capability check on privileged operation
sched: Remove unused function group_first_cpu()
sched/fair: Remove redundant word " *"
selftests/rseq: check if libc rseq support is registered
...
|
|
In preparation for splitting io_uring up a bit, move it into its own
top level directory. It didn't really belong in fs/ anyway, as it's
not a file system only API.
This adds io_uring/ and moves the core files in there, and updates the
MAINTAINERS file for the new location.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
ctxt.2022.07.05a: Linux-kernel memory model development branch.
|
|
cookie
In function sched_core_update_cookie(), a task will enqueue into the
core tree only when it enqueued before, that is, if an uncookied task
is cookied, it will not enqueue into the core tree until it enqueue
again, which will result in unnecessary force idle.
Here follows the scenario:
CPU x and CPU y are a pair of SMT siblings.
1. Start task a running on CPU x without sleeping, and task b and
task c running on CPU y without sleeping.
2. We create a cookie and share it to task a and task b, and then
we create another cookie and share it to task c.
3. Simpling core_forceidle_sum of task a and b from /proc/PID/sched
And we will find out that core_forceidle_sum of task a takes 30%
time of the sampling period, which shouldn't happen as task a and b
have the same cookie.
Then we migrate task a to CPU x', migrate task b and c to CPU y', where
CPU x' and CPU y' are a pair of SMT siblings, and sampling again, we
will found out that core_forceidle_sum of task a and b are almost zero.
To solve this problem, we enqueue the task into the core tree if it's
on rq.
Fixes: 6e33cad0af49("sched: Trivial core scheduling cookie management")
Signed-off-by: Cruz Zhao <CruzZhao@linux.alibaba.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/1656403045-100840-2-git-send-email-CruzZhao@linux.alibaba.com
|
|
dequeue_task_rt() only decrements 'rt_rq->rt_nr_running' after having
called sched_update_tick_dependency() preventing it from re-enabling the
tick on systems that no longer have pending SCHED_RT tasks but have
multiple runnable SCHED_OTHER tasks:
dequeue_task_rt()
dequeue_rt_entity()
dequeue_rt_stack()
dequeue_top_rt_rq()
sub_nr_running() // decrements rq->nr_running
sched_update_tick_dependency()
sched_can_stop_tick() // checks rq->rt.rt_nr_running,
...
__dequeue_rt_entity()
dec_rt_tasks() // decrements rq->rt.rt_nr_running
...
Every other scheduler class performs the operation in the opposite
order, and sched_update_tick_dependency() expects the values to be
updated as such. So avoid the misbehaviour by inverting the order in
which the above operations are performed in the RT scheduler.
Fixes: 76d92ac305f2 ("sched: Migrate sched to use new tick dependency mask model")
Signed-off-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Reviewed-by: Phil Auld <pauld@redhat.com>
Link: https://lore.kernel.org/r/20220628092259.330171-1-nsaenzju@redhat.com
|
|
Tasks the are being deboosted from SCHED_DEADLINE might enter
enqueue_task_dl() one last time and hit an erroneous BUG_ON condition:
since they are not boosted anymore, the if (is_dl_boosted()) branch is
not taken, but the else if (!dl_prio) is and inside this one we
BUG_ON(!is_dl_boosted), which is of course false (BUG_ON triggered)
otherwise we had entered the if branch above. Long story short, the
current condition doesn't make sense and always leads to triggering of a
BUG.
Fix this by only checking enqueue flags, properly: ENQUEUE_REPLENISH has
to be present, but additional flags are not a problem.
Fixes: 64be6f1f5f71 ("sched/deadline: Don't replenish from a !SCHED_DEADLINE entity")
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20220714151908.533052-1-juri.lelli@redhat.com
|
|
With CONFIG_PREEMPT_RT, it is possible to hit a deadlock between two
normal priority tasks (SCHED_OTHER, nice level zero):
INFO: task kworker/u8:0:8 blocked for more than 491 seconds.
Not tainted 5.15.49-rt46 #1
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
task:kworker/u8:0 state:D stack: 0 pid: 8 ppid: 2 flags:0x00000000
Workqueue: writeback wb_workfn (flush-7:0)
[<c08a3a10>] (__schedule) from [<c08a3d84>] (schedule+0xdc/0x134)
[<c08a3d84>] (schedule) from [<c08a65a0>] (rt_mutex_slowlock_block.constprop.0+0xb8/0x174)
[<c08a65a0>] (rt_mutex_slowlock_block.constprop.0) from [<c08a6708>]
+(rt_mutex_slowlock.constprop.0+0xac/0x174)
[<c08a6708>] (rt_mutex_slowlock.constprop.0) from [<c0374d60>] (fat_write_inode+0x34/0x54)
[<c0374d60>] (fat_write_inode) from [<c0297304>] (__writeback_single_inode+0x354/0x3ec)
[<c0297304>] (__writeback_single_inode) from [<c0297998>] (writeback_sb_inodes+0x250/0x45c)
[<c0297998>] (writeback_sb_inodes) from [<c0297c20>] (__writeback_inodes_wb+0x7c/0xb8)
[<c0297c20>] (__writeback_inodes_wb) from [<c0297f24>] (wb_writeback+0x2c8/0x2e4)
[<c0297f24>] (wb_writeback) from [<c0298c40>] (wb_workfn+0x1a4/0x3e4)
[<c0298c40>] (wb_workfn) from [<c0138ab8>] (process_one_work+0x1fc/0x32c)
[<c0138ab8>] (process_one_work) from [<c0139120>] (worker_thread+0x22c/0x2d8)
[<c0139120>] (worker_thread) from [<c013e6e0>] (kthread+0x16c/0x178)
[<c013e6e0>] (kthread) from [<c01000fc>] (ret_from_fork+0x14/0x38)
Exception stack(0xc10e3fb0 to 0xc10e3ff8)
3fa0: 00000000 00000000 00000000 00000000
3fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
3fe0: 00000000 00000000 00000000 00000000 00000013 00000000
INFO: task tar:2083 blocked for more than 491 seconds.
Not tainted 5.15.49-rt46 #1
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
task:tar state:D stack: 0 pid: 2083 ppid: 2082 flags:0x00000000
[<c08a3a10>] (__schedule) from [<c08a3d84>] (schedule+0xdc/0x134)
[<c08a3d84>] (schedule) from [<c08a41b0>] (io_schedule+0x14/0x24)
[<c08a41b0>] (io_schedule) from [<c08a455c>] (bit_wait_io+0xc/0x30)
[<c08a455c>] (bit_wait_io) from [<c08a441c>] (__wait_on_bit_lock+0x54/0xa8)
[<c08a441c>] (__wait_on_bit_lock) from [<c08a44f4>] (out_of_line_wait_on_bit_lock+0x84/0xb0)
[<c08a44f4>] (out_of_line_wait_on_bit_lock) from [<c0371fb0>] (fat_mirror_bhs+0xa0/0x144)
[<c0371fb0>] (fat_mirror_bhs) from [<c0372a68>] (fat_alloc_clusters+0x138/0x2a4)
[<c0372a68>] (fat_alloc_clusters) from [<c0370b14>] (fat_alloc_new_dir+0x34/0x250)
[<c0370b14>] (fat_alloc_new_dir) from [<c03787c0>] (vfat_mkdir+0x58/0x148)
[<c03787c0>] (vfat_mkdir) from [<c0277b60>] (vfs_mkdir+0x68/0x98)
[<c0277b60>] (vfs_mkdir) from [<c027b484>] (do_mkdirat+0xb0/0xec)
[<c027b484>] (do_mkdirat) from [<c0100060>] (ret_fast_syscall+0x0/0x1c)
Exception stack(0xc2e1bfa8 to 0xc2e1bff0)
bfa0: 01ee42f0 01ee4208 01ee42f0 000041ed 00000000 00004000
bfc0: 01ee42f0 01ee4208 00000000 00000027 01ee4302 00000004 000dcb00 01ee4190
bfe0: 000dc368 bed11924 0006d4b0 b6ebddfc
Here the kworker is waiting on msdos_sb_info::s_lock which is held by
tar which is in turn waiting for a buffer which is locked waiting to be
flushed, but this operation is plugged in the kworker.
The lock is a normal struct mutex, so tsk_is_pi_blocked() will always
return false on !RT and thus the behaviour changes for RT.
It seems that the intent here is to skip blk_flush_plug() in the case
where a non-preemptible lock (such as a spinlock) has been converted to
a rtmutex on RT, which is the case covered by the SM_RTLOCK_WAIT
schedule flag. But sched_submit_work() is only called from schedule()
which is never called in this scenario, so the check can simply be
deleted.
Looking at the history of the -rt patchset, in fact this change was
present from v5.9.1-rt20 until being dropped in v5.13-rt1 as it was part
of a larger patch [1] most of which was replaced by commit b4bfa3fcfe3b
("sched/core: Rework the __schedule() preempt argument").
As described in [1]:
The schedule process must distinguish between blocking on a regular
sleeping lock (rwsem and mutex) and a RT-only sleeping lock (spinlock
and rwlock):
- rwsem and mutex must flush block requests (blk_schedule_flush_plug())
even if blocked on a lock. This can not deadlock because this also
happens for non-RT.
There should be a warning if the scheduling point is within a RCU read
section.
- spinlock and rwlock must not flush block requests. This will deadlock
if the callback attempts to acquire a lock which is already acquired.
Similarly to being preempted, there should be no warning if the
scheduling point is within a RCU read section.
and with the tsk_is_pi_blocked() in the scheduler path, we hit the first
issue.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git/tree/patches/0022-locking-rtmutex-Use-custom-scheduling-function-for-s.patch?h=linux-5.10.y-rt-patches
Signed-off-by: John Keeping <john@metanate.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Link: https://lkml.kernel.org/r/20220708162702.1758865-1-john@metanate.com
|
|
The capacity of the CPU available for CFS tasks can be reduced because of
other activities running on the latter. In such case, it's worth trying to
move CFS tasks on a CPU with more available capacity.
The rework of the load balance has filtered the case when the CPU is
classified to be fully busy but its capacity is reduced.
Check if CPU's capacity is reduced while gathering load balance statistic
and classify it group_misfit_task instead of group_fully_busy so we can
try to move the load on another CPU.
Reported-by: David Chen <david.chen@nutanix.com>
Reported-by: Zhang Qiao <zhangqiao22@huawei.com>
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: David Chen <david.chen@nutanix.com>
Tested-by: Zhang Qiao <zhangqiao22@huawei.com>
Link: https://lkml.kernel.org/r/20220708154401.21411-1-vincent.guittot@linaro.org
|
|
The RCU dynticks counter is going to be merged into the context tracking
subsystem. Start with moving the idle extended quiescent states
entrypoints to context tracking. For now those are dumb redirections to
existing RCU calls.
[ paulmck: Apply kernel test robot feedback. ]
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Cc: Uladzislau Rezki <uladzislau.rezki@sony.com>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Nicolas Saenz Julienne <nsaenz@kernel.org>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Xiongfeng Wang <wangxiongfeng2@huawei.com>
Cc: Yu Liao <liaoyu15@huawei.com>
Cc: Phil Auld <pauld@redhat.com>
Cc: Paul Gortmaker<paul.gortmaker@windriver.com>
Cc: Alex Belits <abelits@marvell.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Tested-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
|
|
Use try_cmpxchg instead of cmpxchg (*ptr, old, new) != old in
set_nr_{and_not,if}_polling. x86 cmpxchg returns success in ZF flag,
so this change saves a compare after cmpxchg.
The definition of cmpxchg based fetch_or was changed in the
same way as atomic_fetch_##op definitions were changed
in e6790e4b5d5e97dc287f3496dd2cf2dbabdfdb35.
Also declare these two functions as inline to ensure inlining. In the
case of set_nr_and_not_polling, the compiler (gcc) tries to outsmart
itself by constructing the boolean return value with logic operations
on the fetched value, and these extra operations enlarge the function
over the inlining threshold value.
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220629151552.6015-1-ubizjak@gmail.com
|
|
4feee7d1260 previously added per-task forced idle accounting. This patch
extends this to also include cgroups.
rstat is used for cgroup accounting, except for the root, which uses
kcpustat in order to bypass the need for doing an rstat flush when
reading root stats.
Only cgroup v2 is supported. Similar to the task accounting, the cgroup
accounting requires that schedstats is enabled.
Signed-off-by: Josh Don <joshdon@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Tejun Heo <tj@kernel.org>
Link: https://lkml.kernel.org/r/20220629211426.3329954-1-joshdon@google.com
|
|
Context tracking is going to be used not only to track user transitions
but also idle/IRQs/NMIs. The user tracking part will then become a
separate feature. Prepare Kconfig for that.
[ frederic: Apply Max Filippov feedback. ]
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Cc: Uladzislau Rezki <uladzislau.rezki@sony.com>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Nicolas Saenz Julienne <nsaenz@kernel.org>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Xiongfeng Wang <wangxiongfeng2@huawei.com>
Cc: Yu Liao <liaoyu15@huawei.com>
Cc: Phil Auld <pauld@redhat.com>
Cc: Paul Gortmaker<paul.gortmaker@windriver.com>
Cc: Alex Belits <abelits@marvell.com>
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Tested-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
|
|
find_energy_efficient_cpu() integrates a margin to protect tasks from
bouncing back and forth from a CPU to another. This margin is set as being
6% of the total current energy estimated on the system. This however does
not work for two reasons:
1. The energy estimation is not a good absolute value:
compute_energy() used in feec() is a good estimation for task placement as
it allows to compare the energy with and without a task. The computed
delta will give a good overview of the cost for a certain task placement.
It, however, doesn't work as an absolute estimation for the total energy
of the system. First it adds the contribution to idle CPUs into the
energy, second it mixes util_avg with util_est values. util_avg contains
the near history for a CPU usage, it doesn't tell at all what the current
utilization is. A system that has been quite busy in the near past will
hold a very high energy and then a high margin preventing any task
migration to a lower capacity CPU, wasting energy. It even creates a
negative feedback loop: by holding the tasks on a less efficient CPU, the
margin contributes in keeping the energy high.
2. The margin handicaps small tasks:
On a system where the workload is composed mostly of small tasks (which is
often the case on Android), the overall energy will be high enough to
create a margin none of those tasks can cross. On a Pixel4, a small
utilization of 5% on all the CPUs creates a global estimated energy of 140
joules, as per the Energy Model declaration of that same device. This
means, after applying the 6% margin that any migration must save more than
8 joules to happen. No task with a utilization lower than 40 would then be
able to migrate away from the biggest CPU of the system.
The 6% of the overall system energy was brought by the following patch:
(eb92692b2544 sched/fair: Speed-up energy-aware wake-ups)
It was previously 6% of the prev_cpu energy. Also, the following one
made this margin value conditional on the clusters where the task fits:
(8d4c97c105ca sched/fair: Only compute base_energy_pd if necessary)
We could simply revert that margin change to what it was, but the original
version didn't have strong grounds neither and as demonstrated in (1.) the
estimated energy isn't a good absolute value. Instead, removing it
completely. It is indeed, made possible by recent changes that improved
energy estimation comparison fairness (sched/fair: Remove task_util from
effective utilization in feec()) (PM: EM: Increase energy calculation
precision) and task utilization stabilization (sched/fair: Decay task
util_avg during migration)
Without a margin, we could have feared bouncing between CPUs. But running
LISA's eas_behaviour test coverage on three different platforms (Hikey960,
RB-5 and DB-845) showed no issue.
Removing the energy margin enables more energy-optimized placements for a
more energy efficient system.
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Vincent Donnefort <vdonnefort@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://lkml.kernel.org/r/20220621090414.433602-8-vdonnefort@google.com
|
|
The energy estimation in find_energy_efficient_cpu() (feec()) relies on
the computation of the effective utilization for each CPU of a perf domain
(PD). This effective utilization is then used as an estimation of the busy
time for this pd. The function effective_cpu_util() which gives this value,
scales the utilization relative to IRQ pressure on the CPU to take into
account that the IRQ time is hidden from the task clock. The IRQ scaling is
as follow:
effective_cpu_util = irq + (cpu_cap - irq)/cpu_cap * util
Where util is the sum of CFS/RT/DL utilization, cpu_cap the capacity of
the CPU and irq the IRQ avg time.
If now we take as an example a task placement which doesn't raise the OPP
on the candidate CPU, we can write the energy delta as:
delta = OPPcost/cpu_cap * (effective_cpu_util(cpu_util + task_util) -
effective_cpu_util(cpu_util))
= OPPcost/cpu_cap * (cpu_cap - irq)/cpu_cap * task_util
We end-up with an energy delta depending on the IRQ avg time, which is a
problem: first the time spent on IRQs by a CPU has no effect on the
additional energy that would be consumed by a task. Second, we don't want
to favour a CPU with a higher IRQ avg time value.
Nonetheless, we need to take the IRQ avg time into account. If a task
placement raises the PD's frequency, it will increase the energy cost for
the entire time where the CPU is busy. A solution is to only use
effective_cpu_util() with the CPU contribution part. The task contribution
is added separately and scaled according to prev_cpu's IRQ time.
No change for the FREQUENCY_UTIL component of the energy estimation. We
still want to get the actual frequency that would be selected after the
task placement.
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Vincent Donnefort <vdonnefort@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://lkml.kernel.org/r/20220621090414.433602-7-vdonnefort@google.com
|
|
The Perf Domain (PD) cpumask (struct em_perf_domain.cpus) stays
invariant after Energy Model creation, i.e. it is not updated after
CPU hotplug operations.
That's why the PD mask is used in conjunction with the cpu_online_mask
(or Sched Domain cpumask). Thereby the cpu_online_mask is fetched
multiple times (in compute_energy()) during a run-queue selection
for a task.
cpu_online_mask may change during this time which can lead to wrong
energy calculations.
To be able to avoid this, use the select_rq_mask per-cpu cpumask to
create a cpumask out of PD cpumask and cpu_online_mask and pass it
through the function calls of the EAS run-queue selection path.
The PD cpumask for max_spare_cap_cpu/compute_prev_delta selection
(find_energy_efficient_cpu()) is now ANDed not only with the SD mask
but also with the cpu_online_mask. This is fine since this cpumask
has to be in syc with the one used for energy computation
(compute_energy()).
An exclusive cpuset setup with at least one asymmetric CPU capacity
island (hence the additional AND with the SD cpumask) is the obvious
exception here.
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Tested-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://lkml.kernel.org/r/20220621090414.433602-6-vdonnefort@google.com
|
|
On 21/06/2022 11:04, Vincent Donnefort wrote:
> From: Dietmar Eggemann <dietmar.eggemann@arm.com>
https://lkml.kernel.org/r/202206221253.ZVyGQvPX-lkp@intel.com discovered
that this patch doesn't build anymore (on tip sched/core or linux-next)
because of commit f5b2eeb499910 ("sched/fair: Consider CPU affinity when
allowing NUMA imbalance in find_idlest_group()").
New version of [PATCH v11 4/7] sched/fair: Rename select_idle_mask to
select_rq_mask below.
-- >8 --
Decouple the name of the per-cpu cpumask select_idle_mask from its usage
in select_idle_[cpu/capacity]() of the CFS run-queue selection
(select_task_rq_fair()).
This is to support the reuse of this cpumask in the Energy Aware
Scheduling (EAS) path (find_energy_efficient_cpu()) of the CFS run-queue
selection.
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Tested-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://lkml.kernel.org/r/250691c7-0e2b-05ab-bedf-b245c11d9400@arm.com
|
|
effective_cpu_util() already has a `int cpu' parameter which allows to
retrieve the CPU capacity scale factor (or maximum CPU capacity) inside
this function via an arch_scale_cpu_capacity(cpu).
A lot of code calling effective_cpu_util() (or the shim
sched_cpu_util()) needs the maximum CPU capacity, i.e. it will call
arch_scale_cpu_capacity() already.
But not having to pass it into effective_cpu_util() will make the EAS
wake-up code easier, especially when the maximum CPU capacity reduced
by the thermal pressure is passed through the EAS wake-up functions.
Due to the asymmetric CPU capacity support of arm/arm64 architectures,
arch_scale_cpu_capacity(int cpu) is a per-CPU variable read access via
per_cpu(cpu_scale, cpu) on such a system.
On all other architectures it is a a compile-time constant
(SCHED_CAPACITY_SCALE).
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Vincent Guittot <vincent.guittot@linaro.org>
Tested-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://lkml.kernel.org/r/20220621090414.433602-4-vdonnefort@google.com
|
|
Before being migrated to a new CPU, a task sees its PELT values
synchronized with rq last_update_time. Once done, that same task will also
have its sched_avg last_update_time reset. This means the time between
the migration and the last clock update will not be accounted for in
util_avg and a discontinuity will appear. This issue is amplified by the
PELT clock scaling. It takes currently one tick after the CPU being idle
to let clock_pelt catching up clock_task.
This is especially problematic for asymmetric CPU capacity systems which
need stable util_avg signals for task placement and energy estimation.
Ideally, this problem would be solved by updating the runqueue clocks
before the migration. But that would require taking the runqueue lock
which is quite expensive [1]. Instead estimate the missing time and update
the task util_avg with that value.
To that end, we need sched_clock_cpu() but it is a costly function. Limit
the usage to the case where the source CPU is idle as we know this is when
the clock is having the biggest risk of being outdated.
See comment in migrate_se_pelt_lag() for more details about how the PELT
value is estimated. Notice though this estimation doesn't take into account
IRQ and Paravirt time.
[1] https://lkml.kernel.org/r/20190709115759.10451-1-chris.redpath@arm.com
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Vincent Donnefort <vdonnefort@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://lkml.kernel.org/r/20220621090414.433602-3-vdonnefort@google.com
|
|
Introducing macro helpers u64_u32_{store,load}() to factorize lockless
accesses to u64 variables for 32-bits architectures.
Users are for now cfs_rq.min_vruntime and sched_avg.last_update_time. To
accommodate the later where the copy lies outside of the structure
(cfs_rq.last_udpate_time_copy instead of sched_avg.last_update_time_copy),
use the _copy() version of those helpers.
Those new helpers encapsulate smp_rmb() and smp_wmb() synchronization and
therefore, have a small penalty for 32-bits machines in set_task_rq_fair()
and init_cfs_rq().
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Vincent Donnefort <vdonnefort@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://lkml.kernel.org/r/20220621090414.433602-2-vdonnefort@google.com
|
|
[Problem Statement]
select_idle_cpu() might spend too much time searching for an idle CPU,
when the system is overloaded.
The following histogram is the time spent in select_idle_cpu(),
when running 224 instances of netperf on a system with 112 CPUs
per LLC domain:
@usecs:
[0] 533 | |
[1] 5495 | |
[2, 4) 12008 | |
[4, 8) 239252 | |
[8, 16) 4041924 |@@@@@@@@@@@@@@ |
[16, 32) 12357398 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ |
[32, 64) 14820255 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@|
[64, 128) 13047682 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ |
[128, 256) 8235013 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@ |
[256, 512) 4507667 |@@@@@@@@@@@@@@@ |
[512, 1K) 2600472 |@@@@@@@@@ |
[1K, 2K) 927912 |@@@ |
[2K, 4K) 218720 | |
[4K, 8K) 98161 | |
[8K, 16K) 37722 | |
[16K, 32K) 6715 | |
[32K, 64K) 477 | |
[64K, 128K) 7 | |
netperf latency usecs:
=======
case load Lat_99th std%
TCP_RR thread-224 257.39 ( 0.21)
The time spent in select_idle_cpu() is visible to netperf and might have a negative
impact.
[Symptom analysis]
The patch [1] from Mel Gorman has been applied to track the efficiency
of select_idle_sibling. Copy the indicators here:
SIS Search Efficiency(se_eff%):
A ratio expressed as a percentage of runqueues scanned versus
idle CPUs found. A 100% efficiency indicates that the target,
prev or recent CPU of a task was idle at wakeup. The lower the
efficiency, the more runqueues were scanned before an idle CPU
was found.
SIS Domain Search Efficiency(dom_eff%):
Similar, except only for the slower SIS
patch.
SIS Fast Success Rate(fast_rate%):
Percentage of SIS that used target, prev or
recent CPUs.
SIS Success rate(success_rate%):
Percentage of scans that found an idle CPU.
The test is based on Aubrey's schedtests tool, including netperf, hackbench,
schbench and tbench.
Test on vanilla kernel:
schedstat_parse.py -f netperf_vanilla.log
case load se_eff% dom_eff% fast_rate% success_rate%
TCP_RR 28 threads 99.978 18.535 99.995 100.000
TCP_RR 56 threads 99.397 5.671 99.964 100.000
TCP_RR 84 threads 21.721 6.818 73.632 100.000
TCP_RR 112 threads 12.500 5.533 59.000 100.000
TCP_RR 140 threads 8.524 4.535 49.020 100.000
TCP_RR 168 threads 6.438 3.945 40.309 99.999
TCP_RR 196 threads 5.397 3.718 32.320 99.982
TCP_RR 224 threads 4.874 3.661 25.775 99.767
UDP_RR 28 threads 99.988 17.704 99.997 100.000
UDP_RR 56 threads 99.528 5.977 99.970 100.000
UDP_RR 84 threads 24.219 6.992 76.479 100.000
UDP_RR 112 threads 13.907 5.706 62.538 100.000
UDP_RR 140 threads 9.408 4.699 52.519 100.000
UDP_RR 168 threads 7.095 4.077 44.352 100.000
UDP_RR 196 threads 5.757 3.775 35.764 99.991
UDP_RR 224 threads 5.124 3.704 28.748 99.860
schedstat_parse.py -f schbench_vanilla.log
(each group has 28 tasks)
case load se_eff% dom_eff% fast_rate% success_rate%
normal 1 mthread 99.152 6.400 99.941 100.000
normal 2 mthreads 97.844 4.003 99.908 100.000
normal 3 mthreads 96.395 2.118 99.917 99.998
normal 4 mthreads 55.288 1.451 98.615 99.804
normal 5 mthreads 7.004 1.870 45.597 61.036
normal 6 mthreads 3.354 1.346 20.777 34.230
normal 7 mthreads 2.183 1.028 11.257 21.055
normal 8 mthreads 1.653 0.825 7.849 15.549
schedstat_parse.py -f hackbench_vanilla.log
(each group has 28 tasks)
case load se_eff% dom_eff% fast_rate% success_rate%
process-pipe 1 group 99.991 7.692 99.999 100.000
process-pipe 2 groups 99.934 4.615 99.997 100.000
process-pipe 3 groups 99.597 3.198 99.987 100.000
process-pipe 4 groups 98.378 2.464 99.958 100.000
process-pipe 5 groups 27.474 3.653 89.811 99.800
process-pipe 6 groups 20.201 4.098 82.763 99.570
process-pipe 7 groups 16.423 4.156 77.398 99.316
process-pipe 8 groups 13.165 3.920 72.232 98.828
process-sockets 1 group 99.977 5.882 99.999 100.000
process-sockets 2 groups 99.927 5.505 99.996 100.000
process-sockets 3 groups 99.397 3.250 99.980 100.000
process-sockets 4 groups 79.680 4.258 98.864 99.998
process-sockets 5 groups 7.673 2.503 63.659 92.115
process-sockets 6 groups 4.642 1.584 58.946 88.048
process-sockets 7 groups 3.493 1.379 49.816 81.164
process-sockets 8 groups 3.015 1.407 40.845 75.500
threads-pipe 1 group 99.997 0.000 100.000 100.000
threads-pipe 2 groups 99.894 2.932 99.997 100.000
threads-pipe 3 groups 99.611 4.117 99.983 100.000
threads-pipe 4 groups 97.703 2.624 99.937 100.000
threads-pipe 5 groups 22.919 3.623 87.150 99.764
threads-pipe 6 groups 18.016 4.038 80.491 99.557
threads-pipe 7 groups 14.663 3.991 75.239 99.247
threads-pipe 8 groups 12.242 3.808 70.651 98.644
threads-sockets 1 group 99.990 6.667 99.999 100.000
threads-sockets 2 groups 99.940 5.114 99.997 100.000
threads-sockets 3 groups 99.469 4.115 99.977 100.000
threads-sockets 4 groups 87.528 4.038 99.400 100.000
threads-sockets 5 groups 6.942 2.398 59.244 88.337
threads-sockets 6 groups 4.359 1.954 49.448 87.860
threads-sockets 7 groups 2.845 1.345 41.198 77.102
threads-sockets 8 groups 2.871 1.404 38.512 74.312
schedstat_parse.py -f tbench_vanilla.log
case load se_eff% dom_eff% fast_rate% success_rate%
loopback 28 threads 99.976 18.369 99.995 100.000
loopback 56 threads 99.222 7.799 99.934 100.000
loopback 84 threads 19.723 6.819 70.215 100.000
loopback 112 threads 11.283 5.371 55.371 99.999
loopback 140 threads 0.000 0.000 0.000 0.000
loopback 168 threads 0.000 0.000 0.000 0.000
loopback 196 threads 0.000 0.000 0.000 0.000
loopback 224 threads 0.000 0.000 0.000 0.000
According to the test above, if the system becomes busy, the
SIS Search Efficiency(se_eff%) drops significantly. Although some
benchmarks would finally find an idle CPU(success_rate% = 100%), it is
doubtful whether it is worth it to search the whole LLC domain.
[Proposal]
It would be ideal to have a crystal ball to answer this question:
How many CPUs must a wakeup path walk down, before it can find an idle
CPU? Many potential metrics could be used to predict the number.
One candidate is the sum of util_avg in this LLC domain. The benefit
of choosing util_avg is that it is a metric of accumulated historic
activity, which seems to be smoother than instantaneous metrics
(such as rq->nr_running). Besides, choosing the sum of util_avg
would help predict the load of the LLC domain more precisely, because
SIS_PROP uses one CPU's idle time to estimate the total LLC domain idle
time.
In summary, the lower the util_avg is, the more select_idle_cpu()
should scan for idle CPU, and vice versa. When the sum of util_avg
in this LLC domain hits 85% or above, the scan stops. The reason to
choose 85% as the threshold is that this is the imbalance_pct(117)
when a LLC sched group is overloaded.
Introduce the quadratic function:
y = SCHED_CAPACITY_SCALE - p * x^2
and y'= y / SCHED_CAPACITY_SCALE
x is the ratio of sum_util compared to the CPU capacity:
x = sum_util / (llc_weight * SCHED_CAPACITY_SCALE)
y' is the ratio of CPUs to be scanned in the LLC domain,
and the number of CPUs to scan is calculated by:
nr_scan = llc_weight * y'
Choosing quadratic function is because:
[1] Compared to the linear function, it scans more aggressively when the
sum_util is low.
[2] Compared to the exponential function, it is easier to calculate.
[3] It seems that there is no accurate mapping between the sum of util_avg
and the number of CPUs to be scanned. Use heuristic scan for now.
For a platform with 112 CPUs per LLC, the number of CPUs to scan is:
sum_util% 0 5 15 25 35 45 55 65 75 85 86 ...
scan_nr 112 111 108 102 93 81 65 47 25 1 0 ...
For a platform with 16 CPUs per LLC, the number of CPUs to scan is:
sum_util% 0 5 15 25 35 45 55 65 75 85 86 ...
scan_nr 16 15 15 14 13 11 9 6 3 0 0 ...
Furthermore, to minimize the overhead of calculating the metrics in
select_idle_cpu(), borrow the statistics from periodic load balance.
As mentioned by Abel, on a platform with 112 CPUs per LLC, the
sum_util calculated by periodic load balance after 112 ms would
decay to about 0.5 * 0.5 * 0.5 * 0.7 = 8.75%, thus bringing a delay
in reflecting the latest utilization. But it is a trade-off.
Checking the util_avg in newidle load balance would be more frequent,
but it brings overhead - multiple CPUs write/read the per-LLC shared
variable and introduces cache contention. Tim also mentioned that,
it is allowed to be non-optimal in terms of scheduling for the
short-term variations, but if there is a long-term trend in the load
behavior, the scheduler can adjust for that.
When SIS_UTIL is enabled, the select_idle_cpu() uses the nr_scan
calculated by SIS_UTIL instead of the one from SIS_PROP. As Peter and
Mel suggested, SIS_UTIL should be enabled by default.
This patch is based on the util_avg, which is very sensitive to the
CPU frequency invariance. There is an issue that, when the max frequency
has been clamp, the util_avg would decay insanely fast when
the CPU is idle. Commit addca285120b ("cpufreq: intel_pstate: Handle no_turbo
in frequency invariance") could be used to mitigate this symptom, by adjusting
the arch_max_freq_ratio when turbo is disabled. But this issue is still
not thoroughly fixed, because the current code is unaware of the user-specified
max CPU frequency.
[Test result]
netperf and tbench were launched with 25% 50% 75% 100% 125% 150%
175% 200% of CPU number respectively. Hackbench and schbench were launched
by 1, 2 ,4, 8 groups. Each test lasts for 100 seconds and repeats 3 times.
The following is the benchmark result comparison between
baseline:vanilla v5.19-rc1 and compare:patched kernel. Positive compare%
indicates better performance.
Each netperf test is a:
netperf -4 -H 127.0.1 -t TCP/UDP_RR -c -C -l 100
netperf.throughput
=======
case load baseline(std%) compare%( std%)
TCP_RR 28 threads 1.00 ( 0.34) -0.16 ( 0.40)
TCP_RR 56 threads 1.00 ( 0.19) -0.02 ( 0.20)
TCP_RR 84 threads 1.00 ( 0.39) -0.47 ( 0.40)
TCP_RR 112 threads 1.00 ( 0.21) -0.66 ( 0.22)
TCP_RR 140 threads 1.00 ( 0.19) -0.69 ( 0.19)
TCP_RR 168 threads 1.00 ( 0.18) -0.48 ( 0.18)
TCP_RR 196 threads 1.00 ( 0.16) +194.70 ( 16.43)
TCP_RR 224 threads 1.00 ( 0.16) +197.30 ( 7.85)
UDP_RR 28 threads 1.00 ( 0.37) +0.35 ( 0.33)
UDP_RR 56 threads 1.00 ( 11.18) -0.32 ( 0.21)
UDP_RR 84 threads 1.00 ( 1.46) -0.98 ( 0.32)
UDP_RR 112 threads 1.00 ( 28.85) -2.48 ( 19.61)
UDP_RR 140 threads 1.00 ( 0.70) -0.71 ( 14.04)
UDP_RR 168 threads 1.00 ( 14.33) -0.26 ( 11.16)
UDP_RR 196 threads 1.00 ( 12.92) +186.92 ( 20.93)
UDP_RR 224 threads 1.00 ( 11.74) +196.79 ( 18.62)
Take the 224 threads as an example, the SIS search metrics changes are
illustrated below:
vanilla patched
4544492 +237.5% 15338634 sched_debug.cpu.sis_domain_search.avg
38539 +39686.8% 15333634 sched_debug.cpu.sis_failed.avg
128300000 -87.9% 15551326 sched_debug.cpu.sis_scanned.avg
5842896 +162.7% 15347978 sched_debug.cpu.sis_search.avg
There is -87.9% less CPU scans after patched, which indicates lower overhead.
Besides, with this patch applied, there is -13% less rq lock contention
in perf-profile.calltrace.cycles-pp._raw_spin_lock.raw_spin_rq_lock_nested
.try_to_wake_up.default_wake_function.woken_wake_function.
This might help explain the performance improvement - Because this patch allows
the waking task to remain on the previous CPU, rather than grabbing other CPUs'
lock.
Each hackbench test is a:
hackbench -g $job --process/threads --pipe/sockets -l 1000000 -s 100
hackbench.throughput
=========
case load baseline(std%) compare%( std%)
process-pipe 1 group 1.00 ( 1.29) +0.57 ( 0.47)
process-pipe 2 groups 1.00 ( 0.27) +0.77 ( 0.81)
process-pipe 4 groups 1.00 ( 0.26) +1.17 ( 0.02)
process-pipe 8 groups 1.00 ( 0.15) -4.79 ( 0.02)
process-sockets 1 group 1.00 ( 0.63) -0.92 ( 0.13)
process-sockets 2 groups 1.00 ( 0.03) -0.83 ( 0.14)
process-sockets 4 groups 1.00 ( 0.40) +5.20 ( 0.26)
process-sockets 8 groups 1.00 ( 0.04) +3.52 ( 0.03)
threads-pipe 1 group 1.00 ( 1.28) +0.07 ( 0.14)
threads-pipe 2 groups 1.00 ( 0.22) -0.49 ( 0.74)
threads-pipe 4 groups 1.00 ( 0.05) +1.88 ( 0.13)
threads-pipe 8 groups 1.00 ( 0.09) -4.90 ( 0.06)
threads-sockets 1 group 1.00 ( 0.25) -0.70 ( 0.53)
threads-sockets 2 groups 1.00 ( 0.10) -0.63 ( 0.26)
threads-sockets 4 groups 1.00 ( 0.19) +11.92 ( 0.24)
threads-sockets 8 groups 1.00 ( 0.08) +4.31 ( 0.11)
Each tbench test is a:
tbench -t 100 $job 127.0.0.1
tbench.throughput
======
case load baseline(std%) compare%( std%)
loopback 28 threads 1.00 ( 0.06) -0.14 ( 0.09)
loopback 56 threads 1.00 ( 0.03) -0.04 ( 0.17)
loopback 84 threads 1.00 ( 0.05) +0.36 ( 0.13)
loopback 112 threads 1.00 ( 0.03) +0.51 ( 0.03)
loopback 140 threads 1.00 ( 0.02) -1.67 ( 0.19)
loopback 168 threads 1.00 ( 0.38) +1.27 ( 0.27)
loopback 196 threads 1.00 ( 0.11) +1.34 ( 0.17)
loopback 224 threads 1.00 ( 0.11) +1.67 ( 0.22)
Each schbench test is a:
schbench -m $job -t 28 -r 100 -s 30000 -c 30000
schbench.latency_90%_us
========
case load baseline(std%) compare%( std%)
normal 1 mthread 1.00 ( 31.22) -7.36 ( 20.25)*
normal 2 mthreads 1.00 ( 2.45) -0.48 ( 1.79)
normal 4 mthreads 1.00 ( 1.69) +0.45 ( 0.64)
normal 8 mthreads 1.00 ( 5.47) +9.81 ( 14.28)
*Consider the Standard Deviation, this -7.36% regression might not be valid.
Also, a OLTP workload with a commercial RDBMS has been tested, and there
is no significant change.
There were concerns that unbalanced tasks among CPUs would cause problems.
For example, suppose the LLC domain is composed of 8 CPUs, and 7 tasks are
bound to CPU0~CPU6, while CPU7 is idle:
CPU0 CPU1 CPU2 CPU3 CPU4 CPU5 CPU6 CPU7
util_avg 1024 1024 1024 1024 1024 1024 1024 0
Since the util_avg ratio is 87.5%( = 7/8 ), which is higher than 85%,
select_idle_cpu() will not scan, thus CPU7 is undetected during scan.
But according to Mel, it is unlikely the CPU7 will be idle all the time
because CPU7 could pull some tasks via CPU_NEWLY_IDLE.
lkp(kernel test robot) has reported a regression on stress-ng.sock on a
very busy system. According to the sched_debug statistics, it might be caused
by SIS_UTIL terminates the scan and chooses a previous CPU earlier, and this
might introduce more context switch, especially involuntary preemption, which
impacts a busy stress-ng. This regression has shown that, not all benchmarks
in every scenario benefit from idle CPU scan limit, and it needs further
investigation.
Besides, there is slight regression in hackbench's 16 groups case when the
LLC domain has 16 CPUs. Prateek mentioned that we should scan aggressively
in an LLC domain with 16 CPUs. Because the cost to search for an idle one
among 16 CPUs is negligible. The current patch aims to propose a generic
solution and only considers the util_avg. Something like the below could
be applied on top of the current patch to fulfill the requirement:
if (llc_weight <= 16)
nr_scan = nr_scan * 32 / llc_weight;
For LLC domain with 16 CPUs, the nr_scan will be expanded to 2 times large.
The smaller the CPU number this LLC domain has, the larger nr_scan will be
expanded. This needs further investigation.
There is also ongoing work[2] from Abel to filter out the busy CPUs during
wakeup, to further speed up the idle CPU scan. And it could be a following-up
optimization on top of this change.
Suggested-by: Tim Chen <tim.c.chen@intel.com>
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Yicong Yang <yangyicong@hisilicon.com>
Tested-by: Mohini Narkhede <mohini.narkhede@intel.com>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20220612163428.849378-1-yu.c.chen@intel.com
|
|
sched_setattr(2) issues via kernel/sched/core.c:__sched_setscheduler()
a CAP_SYS_NICE audit event unconditionally, even when the requested
operation does not require that capability / is unprivileged, i.e. for
reducing niceness.
This is relevant in connection with SELinux, where a capability check
results in a policy decision and by default a denial message on
insufficient permission is issued.
It can lead to three undesired cases:
1. A denial message is generated, even in case the operation was an
unprivileged one and thus the syscall succeeded, creating noise.
2. To avoid the noise from 1. the policy writer adds a rule to ignore
those denial messages, hiding future syscalls, where the task
performs an actual privileged operation, leading to hidden limited
functionality of that task.
3. To avoid the noise from 1. the policy writer adds a rule to allow
the task the capability CAP_SYS_NICE, while it does not need it,
violating the principle of least privilege.
Conduct privilged/unprivileged categorization first and perform a
capable test (and at most once) only if needed.
Signed-off-by: Christian Göttsche <cgzones@googlemail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220615152505.310488-1-cgzones@googlemail.com
|
|
As of commit afe06efdf07c ("sched: Extend scheduler's asym packing")
group_first_cpu() became an unused function, remove it.
Signed-off-by: Zhang Qiao <zhangqiao22@huawei.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lore.kernel.org/r/20220617181151.29980-3-zhangqiao22@huawei.com
|
|
" *" is redundant. so remove it.
Signed-off-by: Zhang Qiao <zhangqiao22@huawei.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20220617181151.29980-2-zhangqiao22@huawei.com
|
|
Currently, the RCU Tasks Trace grace-period kthread IPIs each online CPU
using smp_call_function_single() in order to track any tasks currently in
RCU Tasks Trace read-side critical sections during which the corresponding
task has neither blocked nor been preempted. These IPIs are annoying
and are also not strictly necessary because any task that blocks or is
preempted within its current RCU Tasks Trace read-side critical section
will be tracked on one of the per-CPU rcu_tasks_percpu structure's
->rtp_blkd_tasks list. So the only time that this is a problem is if
one of the CPUs runs through a long-duration RCU Tasks Trace read-side
critical section without a context switch.
Note that the task_call_func() function cannot help here because there is
no safe way to identify the target task. Of course, the task_call_func()
function will be very useful later, when processing the list of tasks,
but it needs to know the task.
This commit therefore creates a cpu_curr_snapshot() function that returns
a pointer the task_struct structure of some task that happened to be
running on the specified CPU more or less during the time that the
cpu_curr_snapshot() function was executing. If there was no context
switch during this time, this function will return a pointer to the
task_struct structure of the task that was running throughout. If there
was a context switch, then the outgoing task will be taken care of by
RCU's context-switch hook, and the incoming task was either already taken
care during some previous context switch, or it is not currently within an
RCU Tasks Trace read-side critical section. And in this latter case, the
grace period already started, so there is no need to wait on this task.
This new cpu_curr_snapshot() function is invoked on each CPU early in
the RCU Tasks Trace grace-period processing, and the resulting tasks
are queued for later quiescent-state inspection.
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Neeraj Upadhyay <quic_neeraju@quicinc.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: Martin KaFai Lau <kafai@fb.com>
Cc: KP Singh <kpsingh@kernel.org>
|
|
Wakelist can help avoid cache bouncing and offload the overhead of waker
cpu. So far, using wakelist within the same llc only happens on
WF_ON_CPU, and this limitation could be removed to further improve
wakeup performance.
The commit 518cd6234178 ("sched: Only queue remote wakeups when
crossing cache boundaries") disabled queuing tasks on wakelist when
the cpus share llc. This is because, at that time, the scheduler must
send IPIs to do ttwu_queue_wakelist. Nowadays, ttwu_queue_wakelist also
supports TIF_POLLING, so this is not a problem now when the wakee cpu is
in idle polling.
Benefits:
Queuing the task on idle cpu can help improving performance on waker cpu
and utilization on wakee cpu, and further improve locality because
the wakee cpu can handle its own rq. This patch helps improving rt on
our real java workloads where wakeup happens frequently.
Consider the normal condition (CPU0 and CPU1 share same llc)
Before this patch:
CPU0 CPU1
select_task_rq() idle
rq_lock(CPU1->rq)
enqueue_task(CPU1->rq)
notify CPU1 (by sending IPI or CPU1 polling)
resched()
After this patch:
CPU0 CPU1
select_task_rq() idle
add to wakelist of CPU1
notify CPU1 (by sending IPI or CPU1 polling)
rq_lock(CPU1->rq)
enqueue_task(CPU1->rq)
resched()
We see CPU0 can finish its work earlier. It only needs to put task to
wakelist and return.
While CPU1 is idle, so let itself handle its own runqueue data.
This patch brings no difference about IPI.
This patch only takes effect when the wakee cpu is:
1) idle polling
2) idle not polling
For 1), there will be no IPI with or without this patch.
For 2), there will always be an IPI before or after this patch.
Before this patch: waker cpu will enqueue task and check preempt. Since
"idle" will be sure to be preempted, waker cpu must send a resched IPI.
After this patch: waker cpu will put the task to the wakelist of wakee
cpu, and send an IPI.
Benchmark:
We've tested schbench, unixbench, and hachbench on both x86 and arm64.
On x86 (Intel Xeon Platinum 8269CY):
schbench -m 2 -t 8
Latency percentiles (usec) before after
50.0000th: 8 6
75.0000th: 10 7
90.0000th: 11 8
95.0000th: 12 8
*99.0000th: 13 10
99.5000th: 15 11
99.9000th: 18 14
Unixbench with full threads (104)
before after
Dhrystone 2 using register variables 3011862938 3009935994 -0.06%
Double-Precision Whetstone 617119.3 617298.5 0.03%
Execl Throughput 27667.3 27627.3 -0.14%
File Copy 1024 bufsize 2000 maxblocks 785871.4 784906.2 -0.12%
File Copy 256 bufsize 500 maxblocks 210113.6 212635.4 1.20%
File Copy 4096 bufsize 8000 maxblocks 2328862.2 2320529.1 -0.36%
Pipe Throughput 145535622.8 145323033.2 -0.15%
Pipe-based Context Switching 3221686.4 3583975.4 11.25%
Process Creation 101347.1 103345.4 1.97%
Shell Scripts (1 concurrent) 120193.5 123977.8 3.15%
Shell Scripts (8 concurrent) 17233.4 17138.4 -0.55%
System Call Overhead 5300604.8 5312213.6 0.22%
hackbench -g 1 -l 100000
before after
Time 3.246 2.251
On arm64 (Ampere Altra):
schbench -m 2 -t 8
Latency percentiles (usec) before after
50.0000th: 14 10
75.0000th: 19 14
90.0000th: 22 16
95.0000th: 23 16
*99.0000th: 24 17
99.5000th: 24 17
99.9000th: 28 25
Unixbench with full threads (80)
before after
Dhrystone 2 using register variables 3536194249 3537019613 0.02%
Double-Precision Whetstone 629383.6 629431.6 0.01%
Execl Throughput 65920.5 65846.2 -0.11%
File Copy 1024 bufsize 2000 maxblocks 1063722.8 1064026.8 0.03%
File Copy 256 bufsize 500 maxblocks 322684.5 318724.5 -1.23%
File Copy 4096 bufsize 8000 maxblocks 2348285.3 2328804.8 -0.83%
Pipe Throughput 133542875.3 131619389.8 -1.44%
Pipe-based Context Switching 3215356.1 3576945.1 11.25%
Process Creation 108520.5 120184.6 10.75%
Shell Scripts (1 concurrent) 122636.3 121888 -0.61%
Shell Scripts (8 concurrent) 17462.1 17381.4 -0.46%
System Call Overhead 4429998.9 4435006.7 0.11%
hackbench -g 1 -l 100000
before after
Time 4.217 2.916
Our patch has improvement on schbench, hackbench
and Pipe-based Context Switching of unixbench
when there exists idle cpus,
and no obvious regression on other tests of unixbench.
This can help improve rt in scenes where wakeup happens frequently.
Signed-off-by: Tianchen Ding <dtcccc@linux.alibaba.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lore.kernel.org/r/20220608233412.327341-3-dtcccc@linux.alibaba.com
|
|
The commit 2ebb17717550 ("sched/core: Offload wakee task activation if it
the wakee is descheduling") checked rq->nr_running <= 1 to avoid task
stacking when WF_ON_CPU.
Per the ordering of writes to p->on_rq and p->on_cpu, observing p->on_cpu
(WF_ON_CPU) in ttwu_queue_cond() implies !p->on_rq, IOW p has gone through
the deactivate_task() in __schedule(), thus p has been accounted out of
rq->nr_running. As such, the task being the only runnable task on the rq
implies reading rq->nr_running == 0 at that point.
The benchmark result is in [1].
[1] https://lore.kernel.org/all/e34de686-4e85-bde1-9f3c-9bbc86b38627@linux.alibaba.com/
Suggested-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Tianchen Ding <dtcccc@linux.alibaba.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lore.kernel.org/r/20220608233412.327341-2-dtcccc@linux.alibaba.com
|
|
While doing newidle load balancing, it is possible for new tasks to
arrive, such as with pending wakeups. newidle_balance() already accounts
for this by exiting the sched_domain load_balance() iteration if it
detects these cases. This is very important for minimizing wakeup
latency.
However, if we are already in load_balance(), we may stay there for a
while before returning back to newidle_balance(). This is most
exacerbated if we enter a 'goto redo' loop in the LBF_ALL_PINNED case. A
very straightforward workaround to this is to adjust should_we_balance()
to bail out if we're doing a CPU_NEWLY_IDLE balance and new tasks are
detected.
This was tested with the following reproduction:
- two threads that take turns sleeping and waking each other up are
affined to two cores
- a large number of threads with 100% utilization are pinned to all
other cores
Without this patch, wakeup latency was ~120us for the pair of threads,
almost entirely spent in load_balance(). With this patch, wakeup latency
is ~6us.
Signed-off-by: Josh Don <joshdon@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220609025515.2086253-1-joshdon@google.com
|
|
sysctl_sched_dl_period_max and sysctl_sched_dl_period_min are unsigned
integer, but proc_dointvec() wouldn't return error even if we set a
negative number.
Use proc_douintvec_minmax() instead of proc_dointvec(). Add extra1 for
sysctl_sched_dl_period_max and extra2 for sysctl_sched_dl_period_min.
It's just an optimization for match data and proc_handler in struct
ctl_table. The 'if (period < min || period > max)' in __checkparam_dl()
will work fine even if there hasn't this patch.
Signed-off-by: Yajun Deng <yajun.deng@linux.dev>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Daniel Bristot de Oliveira <bristot@kernel.org>
Link: https://lore.kernel.org/r/20220607101807.249965-1-yajun.deng@linux.dev
|
|
We notice the rq leaf_cfs_rq_list has two problems when do bugfix
backports and some test profiling.
1. cfs_rqs under throttled subtree could be added to the list, and
make their fully decayed ancestors on the list, even though not needed.
2. #1 also make the leaf_cfs_rq_list management complex and error prone,
this is the list of related bugfix so far:
commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
commit fe61468b2cbc ("sched/fair: Fix enqueue_task_fair warning")
commit b34cb07dde7c ("sched/fair: Fix enqueue_task_fair() warning some more")
commit 39f23ce07b93 ("sched/fair: Fix unthrottle_cfs_rq() for leaf_cfs_rq list")
commit 0258bdfaff5b ("sched/fair: Fix unfairness caused by missing load decay")
commit a7b359fc6a37 ("sched/fair: Correctly insert cfs_rq's to list on unthrottle")
commit fdaba61ef8a2 ("sched/fair: Ensure that the CFS parent is added after unthrottling")
commit 2630cde26711 ("sched/fair: Add ancestors of unthrottled undecayed cfs_rq")
commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
delete every cfs_rq under throttled subtree from rq->leaf_cfs_rq_list,
and delete the throttled_hierarchy() test in update_blocked_averages(),
which optimized update_blocked_averages().
But those later bugfix add cfs_rqs under throttled subtree back to
rq->leaf_cfs_rq_list again, with their fully decayed ancestors, for
the integrity of rq->leaf_cfs_rq_list.
This patch takes another method, skip all cfs_rqs under throttled
hierarchy when list_add_leaf_cfs_rq(), to completely make cfs_rqs
under throttled subtree off the leaf_cfs_rq_list.
So we don't need to consider throttled related things in
enqueue_entity(), unthrottle_cfs_rq() and enqueue_task_fair(),
which simplify the code a lot. Also optimize update_blocked_averages()
since cfs_rqs under throttled hierarchy and their ancestors
won't be on the leaf_cfs_rq_list.
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lore.kernel.org/r/20220601021848.76943-1-zhouchengming@bytedance.com
|
|
find_idlest_group()
In the case of systems containing multiple LLCs per socket, like
AMD Zen systems, users want to spread bandwidth hungry applications
across multiple LLCs. Stream is one such representative workload where
the best performance is obtained by limiting one stream thread per LLC.
To ensure this, users are known to pin the tasks to a specify a subset
of the CPUs consisting of one CPU per LLC while running such bandwidth
hungry tasks.
Suppose we kickstart a multi-threaded task like stream with 8 threads
using taskset or numactl to run on a subset of CPUs on a 2 socket Zen3
server where each socket contains 128 CPUs
(0-63,128-191 in one socket, 64-127,192-255 in another socket)
Eg: numactl -C 0,16,32,48,64,80,96,112 ./stream8
Here each CPU in the list is from a different LLC and 4 of those LLCs
are on one socket, while the other 4 are on another socket.
Ideally we would prefer that each stream thread runs on a different
CPU from the allowed list of CPUs. However, the current heuristics in
find_idlest_group() do not allow this during the initial placement.
Suppose the first socket (0-63,128-191) is our local group from which
we are kickstarting the stream tasks. The first four stream threads
will be placed in this socket. When it comes to placing the 5th
thread, all the allowed CPUs are from the local group (0,16,32,48)
would have been taken.
However, the current scheduler code simply checks if the number of
tasks in the local group is fewer than the allowed numa-imbalance
threshold. This threshold was previously 25% of the NUMA domain span
(in this case threshold = 32) but after the v6 of Mel's patchset
"Adjust NUMA imbalance for multiple LLCs", got merged in sched-tip,
Commit: e496132ebedd ("sched/fair: Adjust the allowed NUMA imbalance
when SD_NUMA spans multiple LLCs") it is now equal to number of LLCs
in the NUMA domain, for processors with multiple LLCs.
(in this case threshold = 8).
For this example, the number of tasks will always be within threshold
and thus all the 8 stream threads will be woken up on the first socket
thereby resulting in sub-optimal performance.
The following sched_wakeup_new tracepoint output shows the initial
placement of tasks in the current tip/sched/core on the Zen3 machine:
stream-5313 [016] d..2. 627.005036: sched_wakeup_new: comm=stream pid=5315 prio=120 target_cpu=032
stream-5313 [016] d..2. 627.005086: sched_wakeup_new: comm=stream pid=5316 prio=120 target_cpu=048
stream-5313 [016] d..2. 627.005141: sched_wakeup_new: comm=stream pid=5317 prio=120 target_cpu=000
stream-5313 [016] d..2. 627.005183: sched_wakeup_new: comm=stream pid=5318 prio=120 target_cpu=016
stream-5313 [016] d..2. 627.005218: sched_wakeup_new: comm=stream pid=5319 prio=120 target_cpu=016
stream-5313 [016] d..2. 627.005256: sched_wakeup_new: comm=stream pid=5320 prio=120 target_cpu=016
stream-5313 [016] d..2. 627.005295: sched_wakeup_new: comm=stream pid=5321 prio=120 target_cpu=016
Once the first four threads are distributed among the allowed CPUs of
socket one, the rest of the treads start piling on these same CPUs
when clearly there are CPUs on the second socket that can be used.
Following the initial pile up on a small number of CPUs, though the
load-balancer eventually kicks in, it takes a while to get to {4}{4}
and even {4}{4} isn't stable as we observe a bunch of ping ponging
between {4}{4} to {5}{3} and back before a stable state is reached
much later (1 Stream thread per allowed CPU) and no more migration is
required.
We can detect this piling and avoid it by checking if the number of
allowed CPUs in the local group are fewer than the number of tasks
running in the local group and use this information to spread the
5th task out into the next socket (after all, the goal in this
slowpath is to find the idlest group and the idlest CPU during the
initial placement!).
The following sched_wakeup_new tracepoint output shows the initial
placement of tasks after adding this fix on the Zen3 machine:
stream-4485 [016] d..2. 230.784046: sched_wakeup_new: comm=stream pid=4487 prio=120 target_cpu=032
stream-4485 [016] d..2. 230.784123: sched_wakeup_new: comm=stream pid=4488 prio=120 target_cpu=048
stream-4485 [016] d..2. 230.784167: sched_wakeup_new: comm=stream pid=4489 prio=120 target_cpu=000
stream-4485 [016] d..2. 230.784222: sched_wakeup_new: comm=stream pid=4490 prio=120 target_cpu=112
stream-4485 [016] d..2. 230.784271: sched_wakeup_new: comm=stream pid=4491 prio=120 target_cpu=096
stream-4485 [016] d..2. 230.784322: sched_wakeup_new: comm=stream pid=4492 prio=120 target_cpu=080
stream-4485 [016] d..2. 230.784368: sched_wakeup_new: comm=stream pid=4493 prio=120 target_cpu=064
We see that threads are using all of the allowed CPUs and there is
no pileup.
No output is generated for tracepoint sched_migrate_task with this
patch due to a perfect initial placement which removes the need
for balancing later on - both across NUMA boundaries and within
NUMA boundaries for stream.
Following are the results from running 8 Stream threads with and
without pinning on a dual socket Zen3 Machine (2 x 64C/128T):
During the testing of this patch, the tip sched/core was at
commit: 089c02ae2771 "ftrace: Use preemption model accessors for trace
header printout"
Pinning is done using: numactl -C 0,16,32,48,64,80,96,112 ./stream8
5.18.0-rc1 5.18.0-rc1 5.18.0-rc1
tip sched/core tip sched/core tip sched/core
(no pinning) + pinning + this-patch
+ pinning
Copy: 109364.74 (0.00 pct) 94220.50 (-13.84 pct) 158301.28 (44.74 pct)
Scale: 109670.26 (0.00 pct) 90210.59 (-17.74 pct) 149525.64 (36.34 pct)
Add: 129029.01 (0.00 pct) 101906.00 (-21.02 pct) 186658.17 (44.66 pct)
Triad: 127260.05 (0.00 pct) 106051.36 (-16.66 pct) 184327.30 (44.84 pct)
Pinning currently hurts the performance compared to unbound case on
tip/sched/core. With the addition of this patch, we are able to
outperform tip/sched/core by a good margin with pinning.
Following are the results from running 16 Stream threads with and
without pinning on a dual socket IceLake Machine (2 x 32C/64T):
NUMA Topology of Intel Skylake machine:
Node 1: 0,2,4,6 ... 126 (Even numbers)
Node 2: 1,3,5,7 ... 127 (Odd numbers)
Pinning is done using: numactl -C 0-15 ./stream16
5.18.0-rc1 5.18.0-rc1 5.18.0-rc1
tip sched/core tip sched/core tip sched/core
(no pinning) +pinning + this-patch
+ pinning
Copy: 85815.31 (0.00 pct) 149819.21 (74.58 pct) 156807.48 (82.72 pct)
Scale: 64795.60 (0.00 pct) 97595.07 (50.61 pct) 99871.96 (54.13 pct)
Add: 71340.68 (0.00 pct) 111549.10 (56.36 pct) 114598.33 (60.63 pct)
Triad: 68890.97 (0.00 pct) 111635.16 (62.04 pct) 114589.24 (66.33 pct)
In case of Icelake machine, with single LLC per socket, pinning across
the two sockets reduces cache contention, thus showing great
improvement in pinned case which is further benefited by this patch.
Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Link: https://lkml.kernel.org/r/20220407111222.22649-1-kprateek.nayak@amd.com
|
|
For a single LLC per node, a NUMA imbalance is allowed up until 25%
of CPUs sharing a node could be active. One intent of the cut-off is
to avoid an imbalance of memory channels but there is no topological
information based on active memory channels. Furthermore, there can
be differences between nodes depending on the number of populated
DIMMs.
A cut-off of 25% was arbitrary but generally worked. It does have a severe
corner cases though when an parallel workload is using 25% of all available
CPUs over-saturates memory channels. This can happen due to the initial
forking of tasks that get pulled more to one node after early wakeups
(e.g. a barrier synchronisation) that is not quickly corrected by the
load balancer. The LB may fail to act quickly as the parallel tasks are
considered to be poor migrate candidates due to locality or cache hotness.
On a range of modern Intel CPUs, 12.5% appears to be a better cut-off
assuming all memory channels are populated and is used as the new cut-off
point. A minimum of 1 is specified to allow a communicating pair to
remain local even for CPUs with low numbers of cores. For modern AMDs,
there are multiple LLCs and are not affected.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20220520103519.1863-5-mgorman@techsingularity.net
|
|
The imbalance limitations are applied inconsistently at fork time
and at runtime. At fork, a new task can remain local until there are
too many running tasks even if the degree of imbalance is larger than
NUMA_IMBALANCE_MIN which is different to runtime. Secondly, the imbalance
figure used during load balancing is different to the one used at NUMA
placement. Load balancing uses the number of tasks that must move to
restore imbalance where as NUMA balancing uses the total imbalance.
In combination, it is possible for a parallel workload that uses a small
number of CPUs without applying scheduler policies to have very variable
run-to-run performance.
[lkp@intel.com: Fix build breakage for arc-allyesconfig]
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20220520103519.1863-4-mgorman@techsingularity.net
|
|
If a destination node has spare capacity but there is an imbalance then
two tasks are selected for swapping. If the tasks have no numa group
or are within the same NUMA group, it's simply shuffling tasks around
without having any impact on the compute imbalance. Instead, it's just
punishing one task to help another.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20220520103519.1863-3-mgorman@techsingularity.net
|
|
On clone, numa_migrate_retry is inherited from the parent which means
that the first NUMA placement of a task is non-deterministic. This
affects when load balancing recognises numa tasks and whether to
migrate "regular", "remote" or "all" tasks between NUMA scheduler
domains.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: K Prateek Nayak <kprateek.nayak@amd.com>
Link: https://lore.kernel.org/r/20220520103519.1863-2-mgorman@techsingularity.net
|
|
The purpose of balance_push() is to act as a filter on task selection
in the case of CPU hotplug, specifically when taking the CPU out.
It does this by (ab)using the balance callback infrastructure, with
the express purpose of keeping all the unlikely/odd cases in a single
place.
In order to serve its purpose, the balance_push_callback needs to be
(exclusively) on the callback list at all times (noting that the
callback always places itself back on the list the moment it runs,
also noting that when the CPU goes down, regular balancing concerns
are moot, so ignoring them is fine).
And here-in lies the problem, __sched_setscheduler()'s use of
splice_balance_callbacks() takes the callbacks off the list across a
lock-break, making it possible for, an interleaving, __schedule() to
see an empty list and not get filtered.
Fixes: ae7927023243 ("sched: Optimize finish_lock_switch()")
Reported-by: Jing-Ting Wu <jing-ting.wu@mediatek.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Jing-Ting Wu <jing-ting.wu@mediatek.com>
Link: https://lkml.kernel.org/r/20220519134706.GH2578@worktop.programming.kicks-ass.net
|
|
Memory about struct psi_group is allocated by default for
each cgroup even if psi_disabled is true, in this case, these
allocated memory is waste, so alloc memory for struct psi_group
only when psi_disabled is false.
Signed-off-by: Chen Wandun <chenwandun@huawei.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull scheduler fix from Thomas Gleixner:
"Fix the fallout of sysctl code move which placed the init function
wrong"
* tag 'sched-urgent-2022-06-05' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/autogroup: Fix sysctl move
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace
Pull ptrace_stop cleanups from Eric Biederman:
"While looking at the ptrace problems with PREEMPT_RT and the problems
Peter Zijlstra was encountering with ptrace in his freezer rewrite I
identified some cleanups to ptrace_stop that make sense on their own
and move make resolving the other problems much simpler.
The biggest issue is the habit of the ptrace code to change
task->__state from the tracer to suppress TASK_WAKEKILL from waking up
the tracee. No other code in the kernel does that and it is straight
forward to update signal_wake_up and friends to make that unnecessary.
Peter's task freezer sets frozen tasks to a new state TASK_FROZEN and
then it stores them by calling "wake_up_state(t, TASK_FROZEN)" relying
on the fact that all stopped states except the special stop states can
tolerate spurious wake up and recover their state.
The state of stopped and traced tasked is changed to be stored in
task->jobctl as well as in task->__state. This makes it possible for
the freezer to recover tasks in these special states, as well as
serving as a general cleanup. With a little more work in that
direction I believe TASK_STOPPED can learn to tolerate spurious wake
ups and become an ordinary stop state.
The TASK_TRACED state has to remain a special state as the registers
for a process are only reliably available when the process is stopped
in the scheduler. Fundamentally ptrace needs acess to the saved
register values of a task.
There are bunch of semi-random ptrace related cleanups that were found
while looking at these issues.
One cleanup that deserves to be called out is from commit 57b6de08b5f6
("ptrace: Admit ptrace_stop can generate spuriuos SIGTRAPs"). This
makes a change that is technically user space visible, in the handling
of what happens to a tracee when a tracer dies unexpectedly. According
to our testing and our understanding of userspace nothing cares that
spurious SIGTRAPs can be generated in that case"
* tag 'ptrace_stop-cleanup-for-v5.19' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace:
sched,signal,ptrace: Rework TASK_TRACED, TASK_STOPPED state
ptrace: Always take siglock in ptrace_resume
ptrace: Don't change __state
ptrace: Admit ptrace_stop can generate spuriuos SIGTRAPs
ptrace: Document that wait_task_inactive can't fail
ptrace: Reimplement PTRACE_KILL by always sending SIGKILL
signal: Use lockdep_assert_held instead of assert_spin_locked
ptrace: Remove arch_ptrace_attach
ptrace/xtensa: Replace PT_SINGLESTEP with TIF_SINGLESTEP
ptrace/um: Replace PT_DTRACE with TIF_SINGLESTEP
signal: Replace __group_send_sig_info with send_signal_locked
signal: Rename send_signal send_signal_locked
|
