commit dd29c017aed628076e915fe4cdfb5392fd4c5cab upstream.
RT migration is done aggressively. When a CPU schedules out a high
priority RT task for a lower priority task, it will look to see if there's
any RT tasks that are waiting to run on another CPU that is of higher
priority than the task this CPU is about to run. If it finds one, it will
pull that task over to the CPU and allow it to run there instead.
Normally, this pulling is done by looking at the RT overloaded mask (rto)
which contains all the CPUs in the scheduler domain with RT tasks that are
waiting to run due to a higher priority RT task currently running on their
CPU. The CPU that is about to schedule a lower priority task will grab the
rq lock of the overloaded CPU and move the RT task from that CPU's runqueue
to the local one and schedule the higher priority RT task.
This caused issues when a lot of CPUs would schedule a lower priority task
at the same time. They would all try to grab the same runqueue lock of
the CPU with the overloaded RT tasks. Only the first CPU that got in will
get that task. All the others would wait until they got the runqueue lock
and see there's nothing to pull and do nothing. On systems with lots of
CPUs, this caused a large latency (up to 500us) which is beyond what
PREEMPT_RT is to allow.
The solution to that was to create an RT_PUSH_IPI logic. When any CPU
wanted to pull a task, instead of grabbing the runqueue lock of the
overloaded CPU, it would start by sending an IPI to the overloaded CPU,
and that IPI handler would have the CPU with the waiting RT task do a push
instead. Then that handler would send an IPI to the next CPU with
overloaded RT tasks, and so on. Note, after the first CPU starts this
process, if another CPU wanted to do a pull, it would see that the process
has already begun and would only increment a counter to have the IPIs
continue again.
The RT_PUSH_IPI solved the latency problem with PREEMPT_RT but could cause
a new issue with non PREEMPT_RT. Namely, softirqs run in a threaded
context on PREEMPT_RT but they can run in an interrupt context in non-RT.
If an IPI lands on a CPU that has just woken up multiple RT tasks and the
current CPU is running a non RT or a low priority RT task, instead of
doing a push, it would simply do a schedule on that CPU. But if a softirq
was also executing on this CPU, the schedule would need to wait until the
softirq finished. Until then, the CPU would still be considered overloaded
as there are RT tasks still waiting to run on it.
A live lock occurred on a workload that was doing heavy networking traffic
on a large machine where the softirqs would run 500us out of 750us. And it
would also be waking up RT tasks, causing the RT pull logic to be
constantly executed.
When a softirq triggered on a CPU with RT tasks queued but not running
yet, and the other CPUs would see this CPU as being overloaded, they would
send an IPI over to it. The CPU would notice that the waiting RT tasks are
of higher priority than the currently running task and simply schedule
that CPU instead. But because the softirq was executing, before it could
schedule, it would receive another IPI to do the same. The amount of IPIs
would slow down the currently running softirq so much that before it could
return back to task context, it would execute another softirq never
allowing the CPU to schedule. This live locked that CPU.
As RT_PUSH_IPI was created to help PREEMPT_RT, make it default off if
PREEMPT_RT is not enabled.
Fixes: b6366f048e0c ("sched/rt: Use IPI to trigger RT task push migration instead of pulling")
Closes: https://lore.kernel.org/all/20260506235716.2530720-1-tj@kernel.org/
Reported-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20260515103740.25ccbed8@gandalf.local.home
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit de3ab9bd3133899efb92e4cd05ba4203e58fc0a3 upstream.
In mm_cid_fixup_cpus_to_tasks(), when rq->curr has the target mm and
mm_cid.active is set, the CID is checked with cid_in_transit() before
setting the transition bit. In per-CPU mode a newly forked or exec'd
task can be running with mm_cid.cid == MM_CID_UNSET because CIDs are
assigned lazily on schedule-in. With cid_in_transit() the guard passes
for MM_CID_UNSET (no transit bit), converts it to MM_CID_UNSET |
MM_CID_TRANSIT and stores it back; later mm_cid_schedout() feeds this
to clear_bit() with MM_CID_UNSET as the bit number, triggering an
out-of-bounds write.
Symptoms: this is genuine memory corruption, but a bounded out-of-bounds
write, not an arbitrary one. MM_CID_UNSET is the fixed sentinel BIT(31),
so once the bad value reaches mm_cid_schedout() the cid_from_transit_cid()
strip leaves MM_CID_UNSET, which fails the "cid < max_cids" convergence
test and falls into mm_drop_cid() -> clear_bit(MM_CID_UNSET,
mm_cidmask(mm)). The cid bitmap is embedded in the mm_struct slab object
(after cpu_bitmap and mm_cpus_allowed) and is only num_possible_cpus()
bits wide, so clearing bit 31 is a deterministic OOB bit-clear at a
fixed offset of 2^31 / 8 == 256 MiB past the bitmap base. The address is
not attacker-influenced (fixed sentinel -> fixed offset) and the op only
clears a single bit; what sits 256 MiB further along the direct map is
whatever kernel object happens to live there, so this corrupts one bit of
unpredictable kernel memory -- it is not an arbitrary-address or
arbitrary-value write.
It triggers only in per-CPU CID mode, when a CPU is running an active
task of the target mm whose cid is still MM_CID_UNSET -- the
fork()/execve() window before that task's next schedule-in assigns it a
real CID -- and a per-CPU -> per-task fixup walks over it (the mode
fallback driven by a thread exit, sched_mm_cid_exit(), or by the deferred
max_cids recompute in mm_cid_work_fn()).
In practice syzkaller surfaced it as a KASAN use-after-free reported in
__schedule -> mm_cid_switch_to, where the offending clear_bit() is inlined
via mm_cid_schedout() -> mm_drop_cid().
Guard the transition-bit assignment against MM_CID_UNSET, in addition to
the existing cid_in_transit() check, so the bit is only set on a genuine
task-owned CID. A CPU-owned (MM_CID_ONCPU) CID of a running active task
is handled by the cid_on_cpu(pcp->cid) branch above and never reaches
this path, so excluding MM_CID_UNSET (and the already-transitioning case)
is sufficient.
Fixes: fbd0e71dc370 ("sched/mmcid: Provide CID ownership mode fixup functions")
Signed-off-by: Rik van Riel <riel@surriel.com>
Signed-off-by: Thomas Gleixner <tglx@kernel.org>
Assisted-by: Claude:claude-opus-4-8 syzkaller
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20260616203818.1516263-1-riel@surriel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit fad156c2af227f42ca796cbb20ddc354a6dd9932 upstream.
blk_time_get_ns() caches ktime_get_ns() in current->plug->cur_ktime
and marks the task with PF_BLOCK_TS. That cache is only valid while the
task keeps running; if the task is switched out, wall-clock time
advances and the cached value must not be reused when the task runs again.
The existing invalidation covers explicit plug flushes through
__blk_flush_plug(), and the schedule() / rtmutex paths through
sched_update_worker(). It does not cover in-kernel preemption paths such
as preempt_schedule(), preempt_schedule_notrace(), and
preempt_schedule_irq(), which enter __schedule(SM_PREEMPT) directly and
return without calling sched_update_worker().
As a result, a task preempted while holding a plug with PF_BLOCK_TS set
can reuse a stale plug->cur_ktime after it is scheduled back in. blk-iocost
then consumes that stale timestamp through ioc_now(), producing stale vnow
values for throttle decisions, and through ioc_rqos_done(), inflating
on-queue time and feeding false missed-QoS samples into vrate
adjustment.
Move the schedule-side invalidation to finish_task_switch(), which runs
for the scheduled-in task after every actual context switch regardless
of which schedule entry point was used. Keep __blk_flush_plug() as the
explicit flush/finish-plug invalidation path, and remove only the
PF_BLOCK_TS handling from sched_update_worker().
Fixes: 06b23f92af87 ("block: update cached timestamp post schedule/preemption")
Cc: stable@vger.kernel.org
Signed-off-by: Usama Arif <usama.arif@linux.dev>
Link: https://patch.msgid.link/20260616141604.328820-3-usama.arif@linux.dev
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Pull sched_ext fixes from Tejun Heo:
"Two low-risk fixes:
- Drop a spurious warning that can fire during cgroup migration while
a sched_ext scheduler is loaded
- Fix a drgn-based debug script that broke after scheduler state
moved into a per-scheduler struct"
* tag 'sched_ext-for-7.1-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext:
sched_ext: Don't warn on NULL cgrp_moving_from in scx_cgroup_move_task()
tools/sched_ext: Fix scx_show_state per-scheduler state reads
A WARN fires when systemd's user manager writes "+cpu +memory +pids" to
its own subtree_control while a sched_ext scheduler is loaded:
WARNING: at kernel/sched/ext.c:3227 scx_cgroup_move_task+0xa8/0xb0
scx_cgroup_move_task+0xa8/0xb0
sched_move_task+0x134/0x290
cpu_cgroup_attach+0x39/0x70
cgroup_migrate_execute+0x37d/0x450
cgroup_update_dfl_csses+0x1e3/0x270
cgroup_subtree_control_write+0x3e7/0x440
scx_cgroup_can_attach() arms cgrp_moving_from only when a task's cpu
cgroup changes. It can still be NULL when scx_cgroup_move_task() runs,
through this sequence:
Step Result
--------------------------------- ----------------------------------
1. cpu enabled on cgroup G cpu css = A
2. cpu toggled off then on for G A killed, B created (same cgroup)
3. an exiting task keeps A alive migration skips it, A now stale
4. +memory migrates G stale A vs current B pulls cpu in
5. cpu attach runs for all tasks hits a live, cpu-unchanged task
6. scx_cgroup_move_task() on it cgrp_moving_from NULL -> WARN
The mismatch is that scx_cgroup_can_attach() keys on cgroup identity
while migration drives the move on css identity, so a NULL cgrp_moving_from
here is a legitimate css-only migration, not a missing prep.
The call is already gated on cgrp_moving_from, so just drop the warning.
ops.cgroup_prep_move() and ops.cgroup_move() stay paired.
Fixes: 819513666966 ("sched_ext: Add cgroup support")
Cc: stable@vger.kernel.org # v6.12+
Reported-by: Matt Fleming <mfleming@cloudflare.com>
Closes: https://lore.kernel.org/all/20260601124156.2205704-1-mfleming@cloudflare.com/
Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Andrea Righi <arighi@nvidia.com>
Pull sched_ext fixes from Tejun Heo: - Spurious WARN in ops_dequeue() racing with concurrent dispatch - Self-deadlock between scheduler disable and a concurrent sub-sched enable * tag 'sched_ext-for-7.1-rc4-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext: sched_ext: Fix spurious WARN on stale ops_state in ops_dequeue() sched_ext: Fix deadlock between scx_root_disable() and concurrent forks
ops_dequeue() can race with finish_dispatch() and spuriously trigger the
"queued task must be in BPF scheduler's custody" warning.
ops_dequeue() snapshots p->scx.ops_state via atomic_long_read_acquire()
and then, in the SCX_OPSS_QUEUED arm, asserts that SCX_TASK_IN_CUSTODY
is set. The two reads are not atomic w.r.t. a concurrent
finish_dispatch() running on another CPU:
CPU 1 CPU 2
===== =====
dequeue_task_scx()
ops_dequeue()
opss = read_acquire(ops_state)
= SCX_OPSS_QUEUED
finish_dispatch()
cmpxchg ops_state:
SCX_OPSS_QUEUED -> SCX_OPSS_DISPATCHING [succeeds]
dispatch_enqueue(SCX_DSQ_GLOBAL,
SCX_ENQ_CLEAR_OPSS)
call_task_dequeue()
p->scx.flags &= ~SCX_TASK_IN_CUSTODY
WARN_ON_ONCE(!(p->scx.flags &
SCX_TASK_IN_CUSTODY))
/* opss is stale: QUEUED,
* but task already claimed */
set_release(ops_state, SCX_OPSS_NONE)
The race has been observed via two distinct call chains: the most common
goes through sched_setaffinity(), a rarer variant through
sched_change_begin().
For SCX_DSQ_GLOBAL / SCX_DSQ_BYPASS, dispatch_enqueue() clears
SCX_TASK_IN_CUSTODY before clearing ops_state to SCX_OPSS_NONE
(intentional, to avoid concurrent non-atomic RMW of p->scx.flags against
ops_dequeue()). The window between those two writes is exactly what
ops_dequeue() observes as "QUEUED without custody".
The observed state is not actually inconsistent, it just means CPU 1 has
already claimed the task and the QUEUED value held by CPU 2 is stale.
Re-read ops_state in that case; the next read is guaranteed to return
SCX_OPSS_DISPATCHING or SCX_OPSS_NONE, both of which exit the switch
cleanly. The retry is bounded: once IN_CUSTODY is cleared, ops_state has
already advanced past QUEUED for this dispatch cycle, and a fresh QUEUED
would require re-enqueue under p's rq lock, which CPU 2 holds.
Changes in v2:
- Use READ_ONCE() for p->scx.flags to ensure fresh reads and prevent
compiler reordering in the lockless path
- Add cpu_relax() to reduce power consumption and improve performance
during the spin-wait
- Use unlikely() to optimize branch prediction for the common case
- Expand the in-code comment to document the race condition and
bounded retry guarantee
Fixes: ebf1ccff79c4 ("sched_ext: Fix ops.dequeue() semantics")
Suggested-by: Andrea Righi <arighi@nvidia.com>
Signed-off-by: Samuele Mariotti <smariotti@disroot.org>
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: Tejun Heo <tj@kernel.org>
scx_root_disable() enters SCX_DISABLING before it grabs scx_enable_mutex to
clear __scx_switched_all and scx_switching_all. task_should_scx() short-circuits on DISABLING,
so forks in that window land on fair while next_active_class() still skips
fair - the new tasks stall.
This can deadlock the disable path itself: scx_alloc_and_add_sched() runs
under scx_enable_mutex and creates a helper kthread; if that new kthread is
one of the stalled fair tasks, the mutex holder waits forever and
scx_root_disable() can never make progress. Only sub-sched support exposes
this, since sub-sched enables are the only path where
scx_alloc_and_add_sched() can race the root's disable.
Move the DISABLING check after @scx_switching_all. @scx_switching_all
serves as a proxy for __scx_switched_all, so while it's set, forks keep
going to scx. Once cleared, DISABLING applies normally.
v2: Reword in-source comment and description. (Andrea)
Fixes: 337ec00b1d9c ("sched_ext: Implement cgroup sub-sched enabling and disabling")
Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Andrea Righi <arighi@nvidia.com>
Pull sched_ext fixes from Tejun Heo:
"The bulk of this is hardening of the new sub-scheduler infrastructure.
- UAFs and lifecycle bugs on the sub-sched attach/detach paths:
parent sub_kset freed under a racing child, list_del_rcu on an
uninitialized list head, ops->priv stomped by concurrent
attach/detach, and a UAF in the init-failure error path
- Task state-machine reorg closing concurrent enable-vs-dead races: a
task exiting during the unlocked init window could trip NULL ops
derefs or skip exit_task() cleanup
- A scx_link_sched() self-deadlock on scx_sched_lock
- isolcpus: stop dereferencing the now-RCU-protected HK_TYPE_DOMAIN
cpumask without RCU, and stop rejecting BPF schedulers when only
cpuset isolated partitions are active
- PREEMPT_RT: disable irq_work runs in hardirq context so dumps show
the failing task rather than the irq_work kthread
- Assorted !CONFIG_EXT_SUB_SCHED, randconfig, and selftest build
fixes"
* tag 'sched_ext-for-7.1-rc3-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext:
sched_ext: Use HK_TYPE_DOMAIN_BOOT to detect isolcpus= domain isolation
sched_ext: Defer sub_kset base put to scx_sched_free_rcu_work
sched_ext: INIT_LIST_HEAD() &sch->all in scx_alloc_and_add_sched()
sched_ext: Drop NONE early return in scx_disable_and_exit_task()
sched_ext: Avoid UAF in scx_root_enable_workfn() init failure path
sched_ext: Clear ops->priv on scx_alloc_and_add_sched() error paths
sched_ext: Fix ops->priv clobber on concurrent attach/detach
selftests/sched_ext: Fix build error in dequeue selftest
sched_ext: Handle SCX_TASK_NONE in disable/switched_from paths
sched_ext: Close sub-sched init race with post-init DEAD recheck
sched_ext: Close root-enable vs sched_ext_dead() race with SCX_TASK_INIT_BEGIN
sched_ext: Replace SCX_TASK_OFF_TASKS flag with SCX_TASK_DEAD state
sched_ext: Inline scx_init_task() and move RESET_RUNNABLE_AT into scx_set_task_state()
sched_ext: Cleanups in preparation for the SCX_TASK_INIT_BEGIN/DEAD work
sched_ext: Use IRQ_WORK_INIT_HARD() to initialize sch->disable_irq_work
sched_ext: Fix !CONFIG_EXT_SUB_SCHED build warnings
sched_ext: Drop unused scx_find_sub_sched() stub
sched_ext: Move scx_error() out of scx_link_sched()'s lock region
Pull cgroup fixes from Tejun Heo:
- cpuset fixes:
- Partition invalidation could return CPUs still in use by sibling
partitions, producing overlapping effective_cpus
- cpuset_can_attach() over-reserved DL bandwidth on moves that
stayed within the same root domain
- Pending DL migration state leaked into later attaches when a
later can_attach() check failed
- Reorder PF_EXITING and __GFP_HARDWALL checks so dying tasks can
allocate from any node and exit quickly
- dmem: propagate -ENOMEM instead of spinning forever when the fallback
pool allocation also fails
- selftests/cgroup: percpu test error-path leak, bogus numeric
comparison of cpuset strings, and a zero-length read() that silently
passed OOM-kill tests
* tag 'cgroup-for-7.1-rc3-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
cgroup/cpuset: Return only actually allocated CPUs during partition invalidation
selftests/cgroup: Fix error path leaks in test_percpu_basic
cgroup/cpuset: Reserve DL bandwidth only for root-domain moves
cgroup/cpuset: Reset DL migration state on can_attach() failure
selftests/cgroup: Fix string comparison in write_test
selftests/cgroup: Fix cg_read_strcmp() empty string comparison
cgroup/dmem: Return -ENOMEM on failed pool preallocation
cgroup/cpuset: move PF_EXITING check before __GFP_HARDWALL in cpuset_current_node_allowed()