user/sven/linux.git/kernel/sched, branch v6.6.111

sched: Fix sched_numa_find_nth_cpu() if mask offline

2025-09-09T16:56:26Z

commit 5ebf512f335053a42482ebff91e46c6dc156bf8c upstream. sched_numa_find_nth_cpu() uses a bsearch to look for the 'closest' CPU in sched_domains_numa_masks and given cpus mask. However they might not intersect if all CPUs in the cpus mask are offline. bsearch will return NULL in that case, bail out instead of dereferencing a bogus pointer. The previous behaviour lead to this bug when using maxcpus=4 on an rk3399 (LLLLbb) (i.e. booting with all big CPUs offline): [ 1.422922] Unable to handle kernel paging request at virtual address ffffff8000000000 [ 1.423635] Mem abort info: [ 1.423889] ESR = 0x0000000096000006 [ 1.424227] EC = 0x25: DABT (current EL), IL = 32 bits [ 1.424715] SET = 0, FnV = 0 [ 1.424995] EA = 0, S1PTW = 0 [ 1.425279] FSC = 0x06: level 2 translation fault [ 1.425735] Data abort info: [ 1.425998] ISV = 0, ISS = 0x00000006, ISS2 = 0x00000000 [ 1.426499] CM = 0, WnR = 0, TnD = 0, TagAccess = 0 [ 1.426952] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0 [ 1.427428] swapper pgtable: 4k pages, 39-bit VAs, pgdp=0000000004a9f000 [ 1.428038] [ffffff8000000000] pgd=18000000f7fff403, p4d=18000000f7fff403, pud=18000000f7fff403, pmd=0000000000000000 [ 1.429014] Internal error: Oops: 0000000096000006 [#1] SMP [ 1.429525] Modules linked in: [ 1.429813] CPU: 3 UID: 0 PID: 1 Comm: swapper/0 Not tainted 6.17.0-rc4-dirty #343 PREEMPT [ 1.430559] Hardware name: Pine64 RockPro64 v2.1 (DT) [ 1.431012] pstate: 60000005 (nZCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--) [ 1.431634] pc : sched_numa_find_nth_cpu+0x2a0/0x488 [ 1.432094] lr : sched_numa_find_nth_cpu+0x284/0x488 [ 1.432543] sp : ffffffc084e1b960 [ 1.432843] x29: ffffffc084e1b960 x28: ffffff80078a8800 x27: ffffffc0846eb1d0 [ 1.433495] x26: 0000000000000000 x25: 0000000000000000 x24: 0000000000000000 [ 1.434144] x23: 0000000000000000 x22: fffffffffff7f093 x21: ffffffc081de6378 [ 1.434792] x20: 0000000000000000 x19: 0000000ffff7f093 x18: 00000000ffffffff [ 1.435441] x17: 3030303866666666 x16: 66663d736b73616d x15: ffffffc104e1b5b7 [ 1.436091] x14: 0000000000000000 x13: ffffffc084712860 x12: 0000000000000372 [ 1.436739] x11: 0000000000000126 x10: ffffffc08476a860 x9 : ffffffc084712860 [ 1.437389] x8 : 00000000ffffefff x7 : ffffffc08476a860 x6 : 0000000000000000 [ 1.438036] x5 : 000000000000bff4 x4 : 0000000000000000 x3 : 0000000000000000 [ 1.438683] x2 : 0000000000000000 x1 : ffffffc0846eb000 x0 : ffffff8000407b68 [ 1.439332] Call trace: [ 1.439559] sched_numa_find_nth_cpu+0x2a0/0x488 (P) [ 1.440016] smp_call_function_any+0xc8/0xd0 [ 1.440416] armv8_pmu_init+0x58/0x27c [ 1.440770] armv8_cortex_a72_pmu_init+0x20/0x2c [ 1.441199] arm_pmu_device_probe+0x1e4/0x5e8 [ 1.441603] armv8_pmu_device_probe+0x1c/0x28 [ 1.442007] platform_probe+0x5c/0xac [ 1.442347] really_probe+0xbc/0x298 [ 1.442683] __driver_probe_device+0x78/0x12c [ 1.443087] driver_probe_device+0xdc/0x160 [ 1.443475] __driver_attach+0x94/0x19c [ 1.443833] bus_for_each_dev+0x74/0xd4 [ 1.444190] driver_attach+0x24/0x30 [ 1.444525] bus_add_driver+0xe4/0x208 [ 1.444874] driver_register+0x60/0x128 [ 1.445233] __platform_driver_register+0x24/0x30 [ 1.445662] armv8_pmu_driver_init+0x28/0x4c [ 1.446059] do_one_initcall+0x44/0x25c [ 1.446416] kernel_init_freeable+0x1dc/0x3bc [ 1.446820] kernel_init+0x20/0x1d8 [ 1.447151] ret_from_fork+0x10/0x20 [ 1.447493] Code: 90022e21 f000e5f5 910de2b5 2a1703e2 (f8767803) [ 1.448040] ---[ end trace 0000000000000000 ]--- [ 1.448483] note: swapper/0[1] exited with preempt_count 1 [ 1.449047] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b [ 1.449741] SMP: stopping secondary CPUs [ 1.450105] Kernel Offset: disabled [ 1.450419] CPU features: 0x000000,00080000,20002001,0400421b [ 1.450935] Memory Limit: none [ 1.451217] ---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b ]--- Yury: with the fix, the function returns cpu == nr_cpu_ids, and later in smp_call_function_any -> smp_call_function_single -> generic_exec_single we test the cpu for '>= nr_cpu_ids' and return -ENXIO. So everything is handled correctly. Fixes: cd7f55359c90 ("sched: add sched_numa_find_nth_cpu()") Cc: stable@vger.kernel.org Signed-off-by: Christian Loehle Signed-off-by: Yury Norov (NVIDIA) Signed-off-by: Greg Kroah-Hartman

sched/fair: Fix frequency selection for non-invariant case

2025-08-28T14:28:43Z

commit e37617c8e53a1f7fcba6d5e1041f4fd8a2425c27 upstream. Linus reported a ~50% performance regression on single-threaded workloads on his AMD Ryzen system, and bisected it to: 9c0b4bb7f630 ("sched/cpufreq: Rework schedutil governor performance estimation") When frequency invariance is not enabled, get_capacity_ref_freq(policy) is supposed to return the current frequency and the performance margin applied by map_util_perf(), enabling the utilization to go above the maximum compute capacity and to select a higher frequency than the current one. After the changes in 9c0b4bb7f630, the performance margin was applied earlier in the path to take into account utilization clampings and we couldn't get a utilization higher than the maximum compute capacity, and the CPU remained 'stuck' at lower frequencies. To fix this, we must use a frequency above the current frequency to get a chance to select a higher OPP when the current one becomes fully used. Apply the same margin and return a frequency 25% higher than the current one in order to switch to the next OPP before we fully use the CPU at the current one. [ mingo: Clarified the changelog. ] Fixes: 9c0b4bb7f630 ("sched/cpufreq: Rework schedutil governor performance estimation") Reported-by: Linus Torvalds Bisected-by: Linus Torvalds Reported-by: Wyes Karny Signed-off-by: Vincent Guittot Signed-off-by: Ingo Molnar Tested-by: Wyes Karny Link: https://lore.kernel.org/r/20240114183600.135316-1-vincent.guittot@linaro.org Signed-off-by: Wentao Guan Signed-off-by: Greg Kroah-Hartman

cpufreq/schedutil: Use a fixed reference frequency

2025-08-28T14:28:42Z

commit b3edde44e5d4504c23a176819865cd603fd16d6c upstream. cpuinfo.max_freq can change at runtime because of boost as an example. This implies that the value could be different than the one that has been used when computing the capacity of a CPU. The new arch_scale_freq_ref() returns a fixed and coherent reference frequency that can be used when computing a frequency based on utilization. Use this arch_scale_freq_ref() when available and fallback to policy otherwise. Signed-off-by: Vincent Guittot Signed-off-by: Ingo Molnar Tested-by: Lukasz Luba Reviewed-by: Lukasz Luba Reviewed-by: Dietmar Eggemann Acked-by: Rafael J. Wysocki Acked-by: Viresh Kumar Link: https://lore.kernel.org/r/20231211104855.558096-4-vincent.guittot@linaro.org Stable-dep-of: e37617c8e53a ("sched/fair: Fix frequency selection for non-invariant case") Signed-off-by: Wentao Guan Signed-off-by: Greg Kroah-Hartman

sched/fair: Bump sd->max_newidle_lb_cost when newidle balance fails

2025-08-28T14:28:21Z

[ Upstream commit 155213a2aed42c85361bf4f5c817f5cb68951c3b ] schbench (https://github.com/masoncl/schbench.git) is showing a regression from previous production kernels that bisected down to: sched/fair: Remove sysctl_sched_migration_cost condition (c5b0a7eefc) The schbench command line was: schbench -L -m 4 -M auto -t 256 -n 0 -r 0 -s 0 This creates 4 message threads pinned to CPUs 0-3, and 256x4 worker threads spread across the rest of the CPUs. Neither the worker threads or the message threads do any work, they just wake each other up and go back to sleep as soon as possible. The end result is the first 4 CPUs are pegged waking up those 1024 workers, and the rest of the CPUs are constantly banging in and out of idle. If I take a v6.9 Linus kernel and revert that one commit, performance goes from 3.4M RPS to 5.4M RPS. schedstat shows there are ~100x more new idle balance operations, and profiling shows the worker threads are spending ~20% of their CPU time on new idle balance. schedstats also shows that almost all of these new idle balance attemps are failing to find busy groups. The fix used here is to crank up the cost of the newidle balance whenever it fails. Since we don't want sd->max_newidle_lb_cost to grow out of control, this also changes update_newidle_cost() to use sysctl_sched_migration_cost as the upper limit on max_newidle_lb_cost. Signed-off-by: Chris Mason Signed-off-by: Peter Zijlstra (Intel) Acked-by: Vincent Guittot Link: https://lkml.kernel.org/r/20250626144017.1510594-2-clm@fb.com Signed-off-by: Sasha Levin

freezer,sched: Use saved_state to reduce some spurious wakeups

2025-08-15T10:09:07Z

commit 8f0eed4a78a81668bc78923ea09f51a7a663c2b0 upstream. After commit f5d39b020809 ("freezer,sched: Rewrite core freezer logic"), tasks that transition directly from TASK_FREEZABLE to TASK_FROZEN are always woken up on the thaw path. Prior to that commit, tasks could ask freezer to consider them "frozen enough" via freezer_do_not_count(). The commit replaced freezer_do_not_count() with a TASK_FREEZABLE state which allows freezer to immediately mark the task as TASK_FROZEN without waking up the task. This is efficient for the suspend path, but on the thaw path, the task is always woken up even if the task didn't need to wake up and goes back to its TASK_(UN)INTERRUPTIBLE state. Although these tasks are capable of handling of the wakeup, we can observe a power/perf impact from the extra wakeup. We observed on Android many tasks wait in the TASK_FREEZABLE state (particularly due to many of them being binder clients). We observed nearly 4x the number of tasks and a corresponding linear increase in latency and power consumption when thawing the system. The latency increased from ~15ms to ~50ms. Avoid the spurious wakeups by saving the state of TASK_FREEZABLE tasks. If the task was running before entering TASK_FROZEN state (__refrigerator()) or if the task received a wake up for the saved state, then the task is woken on thaw. saved_state from PREEMPT_RT locks can be re-used because freezer would not stomp on the rtlock wait flow: TASK_RTLOCK_WAIT isn't considered freezable. Reported-by: Prakash Viswalingam Signed-off-by: Elliot Berman Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Ingo Molnar Signed-off-by: Chen Ridong Signed-off-by: Greg Kroah-Hartman

sched/core: Remove ifdeffery for saved_state

2025-08-15T10:09:07Z

commit fbaa6a181a4b1886cbf4214abdf9a2df68471510 upstream. In preparation for freezer to also use saved_state, remove the CONFIG_PREEMPT_RT compilation guard around saved_state. On the arm64 platform I tested which did not have CONFIG_PREEMPT_RT, there was no statistically significant deviation by applying this patch. Test methodology: perf bench sched message -g 40 -l 40 Signed-off-by: Elliot Berman Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Ingo Molnar Signed-off-by: Chen Ridong Signed-off-by: Greg Kroah-Hartman

sched/psi: Fix psi_seq initialization

2025-08-15T10:09:00Z

[ Upstream commit 99b773d720aeea1ef2170dce5fcfa80649e26b78 ] With the seqcount moved out of the group into a global psi_seq, re-initializing the seqcount on group creation is causing seqcount corruption. Fixes: 570c8efd5eb7 ("sched/psi: Optimize psi_group_change() cpu_clock() usage") Reported-by: Chris Mason Suggested-by: Beata Michalska Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Linus Torvalds Signed-off-by: Sasha Levin

sched/psi: Optimize psi_group_change() cpu_clock() usage

2025-08-15T10:08:46Z

[ Upstream commit 570c8efd5eb79c3725ba439ce105ed1bedc5acd9 ] Dietmar reported that commit 3840cbe24cf0 ("sched: psi: fix bogus pressure spikes from aggregation race") caused a regression for him on a high context switch rate benchmark (schbench) due to the now repeating cpu_clock() calls. In particular the problem is that get_recent_times() will extrapolate the current state to 'now'. But if an update uses a timestamp from before the start of the update, it is possible to get two reads with inconsistent results. It is effectively back-dating an update. (note that this all hard-relies on the clock being synchronized across CPUs -- if this is not the case, all bets are off). Combine this problem with the fact that there are per-group-per-cpu seqcounts, the commit in question pushed the clock read into the group iteration, causing tree-depth cpu_clock() calls. On architectures where cpu_clock() has appreciable overhead, this hurts. Instead move to a per-cpu seqcount, which allows us to have a single clock read for all group updates, increasing internal consistency and lowering update overhead. This comes at the cost of a longer update side (proportional to the tree depth) which can cause the read side to retry more often. Fixes: 3840cbe24cf0 ("sched: psi: fix bogus pressure spikes from aggregation race") Reported-by: Dietmar Eggemann Signed-off-by: Peter Zijlstra (Intel) Acked-by: Johannes Weiner Tested-by: Dietmar Eggemann , Link: https://lkml.kernel.org/20250522084844.GC31726@noisy.programming.kicks-ass.net Signed-off-by: Sasha Levin

sched: Change nr_uninterruptible type to unsigned long

2025-07-24T06:53:20Z

commit 36569780b0d64de283f9d6c2195fd1a43e221ee8 upstream. The commit e6fe3f422be1 ("sched: Make multiple runqueue task counters 32-bit") changed nr_uninterruptible to an unsigned int. But the nr_uninterruptible values for each of the CPU runqueues can grow to large numbers, sometimes exceeding INT_MAX. This is valid, if, over time, a large number of tasks are migrated off of one CPU after going into an uninterruptible state. Only the sum of all nr_interruptible values across all CPUs yields the correct result, as explained in a comment in kernel/sched/loadavg.c. Change the type of nr_uninterruptible back to unsigned long to prevent overflows, and thus the miscalculation of load average. Fixes: e6fe3f422be1 ("sched: Make multiple runqueue task counters 32-bit") Signed-off-by: Aruna Ramakrishna Signed-off-by: Peter Zijlstra (Intel) Link: https://lkml.kernel.org/r/20250709173328.606794-1-aruna.ramakrishna@oracle.com Signed-off-by: Greg Kroah-Hartman

sched: Reduce the default slice to avoid tasks getting an extra tick

2025-06-04T12:42:09Z

[ Upstream commit 2ae891b826958b60919ea21c727f77bcd6ffcc2c ] The old default value for slice is 0.75 msec * (1 + ilog(ncpus)) which means that we have a default slice of: 0.75 for 1 cpu 1.50 up to 3 cpus 2.25 up to 7 cpus 3.00 for 8 cpus and above. For HZ=250 and HZ=100, because of the tick accuracy, the runtime of tasks is far higher than their slice. For HZ=1000 with 8 cpus or more, the accuracy of tick is already satisfactory, but there is still an issue that tasks will get an extra tick because the tick often arrives a little faster than expected. In this case, the task can only wait until the next tick to consider that it has reached its deadline, and will run 1ms longer. vruntime + sysctl_sched_base_slice = deadline |-----------|-----------|-----------|-----------| 1ms 1ms 1ms 1ms ^ ^ ^ ^ tick1 tick2 tick3 tick4(nearly 4ms) There are two reasons for tick error: clockevent precision and the CONFIG_IRQ_TIME_ACCOUNTING/CONFIG_PARAVIRT_TIME_ACCOUNTING. with CONFIG_IRQ_TIME_ACCOUNTING every tick will be less than 1ms, but even without it, because of clockevent precision, tick still often less than 1ms. In order to make scheduling more precise, we changed 0.75 to 0.70, Using 0.70 instead of 0.75 should not change much for other configs and would fix this issue: 0.70 for 1 cpu 1.40 up to 3 cpus 2.10 up to 7 cpus 2.8 for 8 cpus and above. This does not guarantee that tasks can run the slice time accurately every time, but occasionally running an extra tick has little impact. Signed-off-by: zihan zhou <15645113830zzh@gmail.com> Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Vincent Guittot Link: https://lkml.kernel.org/r/20250208075322.13139-1-15645113830zzh@gmail.com Signed-off-by: Sasha Levin