linux.git/kernel/sched/core.c, branch v6.2-rc5

sched/core: Use kfree_rcu() in do_set_cpus_allowed()

2023-01-09T10:43:23+00:00

Commit 851a723e45d1 ("sched: Always clear user_cpus_ptr in
do_set_cpus_allowed()") may call kfree() if user_cpus_ptr was previously
set. Unfortunately, some of the callers of do_set_cpus_allowed()
may have pi_lock held when calling it. So the following splats may be
printed especially when running with a PREEMPT_RT kernel:

   WARNING: possible circular locking dependency detected
   BUG: sleeping function called from invalid context

To avoid these problems, kfree_rcu() is used instead. An internal
cpumask_rcuhead union is created for the sole purpose of facilitating
the use of kfree_rcu() to free the cpumask.

Since user_cpus_ptr is not being used in non-SMP configs, the newly
introduced alloc_user_cpus_ptr() helper will return NULL in this case
and sched_setaffinity() is modified to handle this special case.

Fixes: 851a723e45d1 ("sched: Always clear user_cpus_ptr in do_set_cpus_allowed()")
Suggested-by: Peter Zijlstra 
Signed-off-by: Waiman Long 
Signed-off-by: Ingo Molnar 
Reviewed-by: Peter Zijlstra 
Link: https://lore.kernel.org/r/20221231041120.440785-3-longman@redhat.com

sched/core: Fix use-after-free bug in dup_user_cpus_ptr()

2023-01-09T10:43:07+00:00

Since commit 07ec77a1d4e8 ("sched: Allow task CPU affinity to be
restricted on asymmetric systems"), the setting and clearing of
user_cpus_ptr are done under pi_lock for arm64 architecture. However,
dup_user_cpus_ptr() accesses user_cpus_ptr without any lock
protection. Since sched_setaffinity() can be invoked from another
process, the process being modified may be undergoing fork() at
the same time.  When racing with the clearing of user_cpus_ptr in
__set_cpus_allowed_ptr_locked(), it can lead to user-after-free and
possibly double-free in arm64 kernel.

Commit 8f9ea86fdf99 ("sched: Always preserve the user requested
cpumask") fixes this problem as user_cpus_ptr, once set, will never
be cleared in a task's lifetime. However, this bug was re-introduced
in commit 851a723e45d1 ("sched: Always clear user_cpus_ptr in
do_set_cpus_allowed()") which allows the clearing of user_cpus_ptr in
do_set_cpus_allowed(). This time, it will affect all arches.

Fix this bug by always clearing the user_cpus_ptr of the newly
cloned/forked task before the copying process starts and check the
user_cpus_ptr state of the source task under pi_lock.

Note to stable, this patch won't be applicable to stable releases.
Just copy the new dup_user_cpus_ptr() function over.

Fixes: 07ec77a1d4e8 ("sched: Allow task CPU affinity to be restricted on asymmetric systems")
Fixes: 851a723e45d1 ("sched: Always clear user_cpus_ptr in do_set_cpus_allowed()")
Reported-by: David Wang 王标 
Signed-off-by: Waiman Long 
Signed-off-by: Ingo Molnar 
Reviewed-by: Peter Zijlstra 
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20221231041120.440785-2-longman@redhat.com

sched/core: Fix arch_scale_freq_tick() on tickless systems

2023-01-07T11:25:50+00:00

In order for the scheduler to be frequency invariant we measure the
ratio between the maximum CPU frequency and the actual CPU frequency.

During long tickless periods of time the calculations that keep track
of that might overflow, in the function scale_freq_tick():

  if (check_shl_overflow(acnt, 2*SCHED_CAPACITY_SHIFT, &acnt))
          goto error;

eventually forcing the kernel to disable the feature for all CPUs,
and show the warning message:

   "Scheduler frequency invariance went wobbly, disabling!".

Let's avoid that by limiting the frequency invariant calculations
to CPUs with regular tick.

Fixes: e2b0d619b400 ("x86, sched: check for counters overflow in frequency invariant accounting")
Suggested-by: "Peter Zijlstra (Intel)" 
Signed-off-by: Yair Podemsky 
Signed-off-by: Ingo Molnar 
Reviewed-by: Valentin Schneider 
Acked-by: Giovanni Gherdovich 
Link: https://lore.kernel.org/r/20221130125121.34407-1-ypodemsk@redhat.com

Merge tag 'hardening-v6.2-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux

2022-12-14T20:20:00+00:00

Pull kernel hardening updates from Kees Cook:

 - Convert flexible array members, fix -Wstringop-overflow warnings, and
   fix KCFI function type mismatches that went ignored by maintainers
   (Gustavo A. R. Silva, Nathan Chancellor, Kees Cook)

 - Remove the remaining side-effect users of ksize() by converting
   dma-buf, btrfs, and coredump to using kmalloc_size_roundup(), add
   more __alloc_size attributes, and introduce full testing of all
   allocator functions. Finally remove the ksize() side-effect so that
   each allocation-aware checker can finally behave without exceptions

 - Introduce oops_limit (default 10,000) and warn_limit (default off) to
   provide greater granularity of control for panic_on_oops and
   panic_on_warn (Jann Horn, Kees Cook)

 - Introduce overflows_type() and castable_to_type() helpers for cleaner
   overflow checking

 - Improve code generation for strscpy() and update str*() kern-doc

 - Convert strscpy and sigphash tests to KUnit, and expand memcpy tests

 - Always use a non-NULL argument for prepare_kernel_cred()

 - Disable structleak plugin in FORTIFY KUnit test (Anders Roxell)

 - Adjust orphan linker section checking to respect CONFIG_WERROR (Xin
   Li)

 - Make sure siginfo is cleared for forced SIGKILL (haifeng.xu)

 - Fix um vs FORTIFY warnings for always-NULL arguments

* tag 'hardening-v6.2-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux: (31 commits)
  ksmbd: replace one-element arrays with flexible-array members
  hpet: Replace one-element array with flexible-array member
  um: virt-pci: Avoid GCC non-NULL warning
  signal: Initialize the info in ksignal
  lib: fortify_kunit: build without structleak plugin
  panic: Expose "warn_count" to sysfs
  panic: Introduce warn_limit
  panic: Consolidate open-coded panic_on_warn checks
  exit: Allow oops_limit to be disabled
  exit: Expose "oops_count" to sysfs
  exit: Put an upper limit on how often we can oops
  panic: Separate sysctl logic from CONFIG_SMP
  mm/pgtable: Fix multiple -Wstringop-overflow warnings
  mm: Make ksize() a reporting-only function
  kunit/fortify: Validate __alloc_size attribute results
  drm/sti: Fix return type of sti_{dvo,hda,hdmi}_connector_mode_valid()
  drm/fsl-dcu: Fix return type of fsl_dcu_drm_connector_mode_valid()
  driver core: Add __alloc_size hint to devm allocators
  overflow: Introduce overflows_type() and castable_to_type()
  coredump: Proactively round up to kmalloc bucket size
  ...

Merge tag 'sysctl-6.2-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/mcgrof/linux

2022-12-13T22:16:44+00:00

Pull sysctl updates from Luis Chamberlain:
 "Only a small step forward on the sysctl cleanups for this cycle"

* tag 'sysctl-6.2-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/mcgrof/linux:
  sched: Move numa_balancing sysctls to its own file

Merge tag 'sched-core-2022-12-12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2022-12-12T23:33:42+00:00

Pull scheduler updates from Ingo Molnar:

 - Implement persistent user-requested affinity: introduce
   affinity_context::user_mask and unconditionally preserve the
   user-requested CPU affinity masks, for long-lived tasks to better
   interact with cpusets & CPU hotplug events over longer timespans,
   without destroying the original affinity intent if the underlying
   topology changes.

 - Uclamp updates: fix relationship between uclamp and fits_capacity()

 - PSI fixes

 - Misc fixes & updates

* tag 'sched-core-2022-12-12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched: Clear ttwu_pending after enqueue_task()
  sched/psi: Use task->psi_flags to clear in CPU migration
  sched/psi: Stop relying on timer_pending() for poll_work rescheduling
  sched/psi: Fix avgs_work re-arm in psi_avgs_work()
  sched/psi: Fix possible missing or delayed pending event
  sched: Always clear user_cpus_ptr in do_set_cpus_allowed()
  sched: Enforce user requested affinity
  sched: Always preserve the user requested cpumask
  sched: Introduce affinity_context
  sched: Add __releases annotations to affine_move_task()
  sched/fair: Check if prev_cpu has highest spare cap in feec()
  sched/fair: Consider capacity inversion in util_fits_cpu()
  sched/fair: Detect capacity inversion
  sched/uclamp: Cater for uclamp in find_energy_efficient_cpu()'s early exit condition
  sched/uclamp: Make cpu_overutilized() use util_fits_cpu()
  sched/uclamp: Make asym_fits_capacity() use util_fits_cpu()
  sched/uclamp: Make select_idle_capacity() use util_fits_cpu()
  sched/uclamp: Fix fits_capacity() check in feec()
  sched/uclamp: Make task_fits_capacity() use util_fits_cpu()
  sched/uclamp: Fix relationship between uclamp and migration margin

panic: Consolidate open-coded panic_on_warn checks

2022-12-02T21:04:44+00:00

Several run-time checkers (KASAN, UBSAN, KFENCE, KCSAN, sched) roll
their own warnings, and each check "panic_on_warn". Consolidate this
into a single function so that future instrumentation can be added in
a single location.

Cc: Marco Elver 
Cc: Dmitry Vyukov 
Cc: Ingo Molnar 
Cc: Peter Zijlstra 
Cc: Juri Lelli 
Cc: Vincent Guittot 
Cc: Dietmar Eggemann 
Cc: Steven Rostedt 
Cc: Ben Segall 
Cc: Mel Gorman 
Cc: Daniel Bristot de Oliveira 
Cc: Valentin Schneider 
Cc: Andrey Ryabinin 
Cc: Alexander Potapenko 
Cc: Andrey Konovalov 
Cc: Vincenzo Frascino 
Cc: Andrew Morton 
Cc: David Gow 
Cc: tangmeng 
Cc: Jann Horn 
Cc: Shuah Khan 
Cc: Petr Mladek 
Cc: "Paul E. McKenney" 
Cc: Sebastian Andrzej Siewior 
Cc: "Guilherme G. Piccoli" 
Cc: Tiezhu Yang 
Cc: kasan-dev@googlegroups.com
Cc: linux-mm@kvack.org
Reviewed-by: Luis Chamberlain 
Signed-off-by: Kees Cook 
Reviewed-by: Marco Elver 
Reviewed-by: Andrey Konovalov 
Link: https://lore.kernel.org/r/20221117234328.594699-4-keescook@chromium.org

sched: Move numa_balancing sysctls to its own file

2022-11-21T04:55:26+00:00

The sysctl_numa_balancing_promote_rate_limit and sysctl_numa_balancing
are part of sched, move them to its own file.

Signed-off-by: Kefeng Wang 
Signed-off-by: Luis Chamberlain

sched: Clear ttwu_pending after enqueue_task()

2022-11-16T09:13:05+00:00

We found a long tail latency in schbench whem m*t is close to nr_cpus.
(e.g., "schbench -m 2 -t 16" on a machine with 32 cpus.)

This is because when the wakee cpu is idle, rq->ttwu_pending is cleared
too early, and idle_cpu() will return true until the wakee task enqueued.
This will mislead the waker when selecting idle cpu, and wake multiple
worker threads on the same wakee cpu. This situation is enlarged by
commit f3dd3f674555 ("sched: Remove the limitation of WF_ON_CPU on
wakelist if wakee cpu is idle") because it tends to use wakelist.

Here is the result of "schbench -m 2 -t 16" on a VM with 32vcpu
(Intel(R) Xeon(R) Platinum 8369B).

Latency percentiles (usec):
                base      base+revert_f3dd3f674555   base+this_patch
50.0000th:         9                            13                 9
75.0000th:        12                            19                12
90.0000th:        15                            22                15
95.0000th:        18                            24                17
*99.0000th:       27                            31                24
99.5000th:      3364                            33                27
99.9000th:     12560                            36                30

We also tested on unixbench and hackbench, and saw no performance
change.

Signed-off-by: Tianchen Ding 
Signed-off-by: Peter Zijlstra (Intel) 
Acked-by: Mel Gorman 
Link: https://lkml.kernel.org/r/20221104023601.12844-1-dtcccc@linux.alibaba.com

sched: Fix race in task_call_func()

2022-11-14T08:58:32+00:00

There is a very narrow race between schedule() and task_call_func().

  CPU0						CPU1

  __schedule()
    rq_lock();
    prev_state = READ_ONCE(prev->__state);
    if (... && prev_state) {
      deactivate_tasl(rq, prev, ...)
        prev->on_rq = 0;

						task_call_func()
						  raw_spin_lock_irqsave(p->pi_lock);
						  state = READ_ONCE(p->__state);
						  smp_rmb();
						  if (... || p->on_rq) // false!!!
						    rq = __task_rq_lock()

						  ret = func();

    next = pick_next_task();
    rq = context_switch(prev, next)
      prepare_lock_switch()
        spin_release(&__rq_lockp(rq)->dep_map...)

So while the task is on it's way out, it still holds rq->lock for a
little while, and right then task_call_func() comes in and figures it
doesn't need rq->lock anymore (because the task is already dequeued --
but still running there) and then the __set_task_frozen() thing observes
it's holding rq->lock and yells murder.

Avoid this by waiting for p->on_cpu to get cleared, which guarantees
the task is fully finished on the old CPU.

( While arguably the fixes tag is 'wrong' -- none of the previous
  task_call_func() users appears to care for this case. )

Fixes: f5d39b020809 ("freezer,sched: Rewrite core freezer logic")
Reported-by: Ville Syrjälä 
Signed-off-by: Peter Zijlstra (Intel) 
Tested-by: Ville Syrjälä 
Link: https://lkml.kernel.org/r/Y1kdRNNfUeAU+FNl@hirez.programming.kicks-ass.net