- Kernel - mailweb.openeuler.org

[PATCH OLK-6.6 V1] sched: Support NUMA parallel scheduling for multiple processes
by Cheng Yu 28 Jul '25

28 Jul '25

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/ICBBNL -------------------------------- For architectures with multiple NUMA node levels and large distances between nodes, a better approach is to support processes running in parallel on each NUMA node. The usage is restricted to the following scenarios: 1. No CPU binding for user-space processes; 2. It is applicable to distributed applications, such as business architectures with one master and multiple slaves running in parallel; 3. The existing "qos dynamic affinity" and "qos smart grid" features must not be used simultaneously. Signed-off-by: Cheng Yu <serein.chengyu(a)huawei.com> --- arch/arm64/Kconfig | 1 + arch/arm64/configs/openeuler_defconfig | 1 + arch/arm64/include/asm/prefer_numa.h | 13 +++++ arch/arm64/kernel/Makefile | 1 + arch/arm64/kernel/prefer_numa.c | 72 ++++++++++++++++++++++++++ fs/proc/array.c | 3 -- include/linux/perf_event.h | 2 + include/linux/sched.h | 6 +++ init/Kconfig | 22 ++++++++ kernel/cgroup/cpuset.c | 6 ++- kernel/events/core.c | 13 +++++ kernel/fork.c | 11 ++-- kernel/sched/debug.c | 43 ++++++++++++++- kernel/sched/fair.c | 39 +++++++++++--- kernel/sched/features.h | 4 ++ 15 files changed, 216 insertions(+), 21 deletions(-) create mode 100644 arch/arm64/include/asm/prefer_numa.h create mode 100644 arch/arm64/kernel/prefer_numa.c diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 5422d1502fd6..b1f550c8c82a 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -105,6 +105,7 @@ config ARM64 select ARCH_SUPPORTS_ATOMIC_RMW select ARCH_SUPPORTS_INT128 if CC_HAS_INT128 select ARCH_SUPPORTS_NUMA_BALANCING + select ARCH_SUPPORTS_SCHED_PARAL select ARCH_SUPPORTS_PAGE_TABLE_CHECK select ARCH_SUPPORTS_PER_VMA_LOCK select ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 3cfff0701479..3d352fb1ae57 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -209,6 +209,7 @@ CONFIG_USER_NS=y CONFIG_PID_NS=y CONFIG_NET_NS=y CONFIG_SCHED_STEAL=y +CONFIG_SCHED_PARAL=y CONFIG_CHECKPOINT_RESTORE=y CONFIG_SCHED_AUTOGROUP=y CONFIG_RELAY=y diff --git a/arch/arm64/include/asm/prefer_numa.h b/arch/arm64/include/asm/prefer_numa.h new file mode 100644 index 000000000000..6c8e2b2142b9 --- /dev/null +++ b/arch/arm64/include/asm/prefer_numa.h @@ -0,0 +1,13 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +#ifndef __ASM_PREFER_NUMA_H +#define __ASM_PREFER_NUMA_H + +#include <linux/sched.h> + +#define PROBE_NUMA_PMU_NAME "hisi_sccl3_hha0" +#define PROBE_NUMA_PMU_EVENT 0x02 + +void set_task_paral_node(struct task_struct *p); +int probe_pmu_numa_event(void); + +#endif /* __ASM_PREFER_NUMA_H */ diff --git a/arch/arm64/kernel/Makefile b/arch/arm64/kernel/Makefile index 3d404a2cc961..b936be9d8baa 100644 --- a/arch/arm64/kernel/Makefile +++ b/arch/arm64/kernel/Makefile @@ -84,6 +84,7 @@ obj-$(CONFIG_IPI_AS_NMI) += ipi_nmi.o obj-$(CONFIG_HISI_VIRTCCA_GUEST) += virtcca_cvm_guest.o virtcca_cvm_tsi.o obj-$(CONFIG_HISI_VIRTCCA_HOST) += virtcca_cvm_host.o CFLAGS_patch-scs.o += -mbranch-protection=none +obj-$(CONFIG_SCHED_PARAL) += prefer_numa.o # Force dependency (vdso*-wrap.S includes vdso.so through incbin) $(obj)/vdso-wrap.o: $(obj)/vdso/vdso.so diff --git a/arch/arm64/kernel/prefer_numa.c b/arch/arm64/kernel/prefer_numa.c new file mode 100644 index 000000000000..394dd4098c8f --- /dev/null +++ b/arch/arm64/kernel/prefer_numa.c @@ -0,0 +1,72 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * choose a prefer numa node + * + * Copyright (C) 2025 Huawei Limited. + */ +#include <linux/perf_event.h> +#include <asm/prefer_numa.h> + +static atomic_t paral_nid_last = ATOMIC_INIT(-1); + +int probe_pmu_numa_event(void) +{ + int cpu = -1; + struct perf_event *event; + struct perf_event_attr attr = {}; + int type = perf_pmu_type_of_name(PROBE_NUMA_PMU_NAME); + + if (type == -1) + return -EINVAL; + + preempt_disable(); + cpu = smp_processor_id(); + preempt_enable(); + + attr.type = type; + attr.config = PROBE_NUMA_PMU_EVENT; + attr.size = sizeof(struct perf_event_attr); + attr.pinned = 1; + attr.disabled = 1; + attr.sample_period = 0; + + event = perf_event_create_kernel_counter(&attr, cpu, NULL, NULL, NULL); + if (IS_ERR(event)) + return PTR_ERR(event); + + perf_event_release_kernel(event); + + return 0; +} + +static inline unsigned int update_sched_paral_nid(void) +{ + return (unsigned int)atomic_inc_return(&paral_nid_last); +} + +void set_task_paral_node(struct task_struct *p) +{ + int nid; + int i = 0; + const cpumask_t *cpus_mask; + + if (is_global_init(current)) + return; + + if (p->flags & PF_KTHREAD || p->tgid != p->pid) + return; + + while (i < nr_node_ids) { + nid = update_sched_paral_nid() % nr_node_ids; + cpus_mask = cpumask_of_node(nid); + + if (cpumask_empty(cpus_mask) || + !cpumask_subset(cpus_mask, p->cpus_ptr)) { + i++; + continue; + } + + cpumask_copy(p->prefer_cpus, cpus_mask); + break; + } +} diff --git a/fs/proc/array.c b/fs/proc/array.c index a933a878df3c..6a4b0a850dce 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -439,9 +439,6 @@ __weak void arch_proc_pid_thread_features(struct seq_file *m, #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY static void task_cpus_preferred(struct seq_file *m, struct task_struct *task) { - if (!dynamic_affinity_enabled()) - return; - seq_printf(m, "Cpus_preferred:\t%*pb\n", cpumask_pr_args(task->prefer_cpus)); seq_printf(m, "Cpus_preferred_list:\t%*pbl\n", diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 826fb16906fe..14ed13f4b408 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -1778,6 +1778,7 @@ extern void perf_event_task_tick(void); extern int perf_event_account_interrupt(struct perf_event *event); extern int perf_event_period(struct perf_event *event, u64 value); extern u64 perf_event_pause(struct perf_event *event, bool reset); +extern int perf_pmu_type_of_name(const char *name); #else /* !CONFIG_PERF_EVENTS: */ static inline void * perf_aux_output_begin(struct perf_output_handle *handle, @@ -1864,6 +1865,7 @@ static inline u64 perf_event_pause(struct perf_event *event, bool reset) { return 0; } +static inline int perf_pmu_type_of_name(const char *name) { return -1; } #endif #if defined(CONFIG_PERF_EVENTS) && defined(CONFIG_CPU_SUP_INTEL) diff --git a/include/linux/sched.h b/include/linux/sched.h index 3979c34e9b83..ee10780715f1 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2627,6 +2627,12 @@ static inline bool dynamic_affinity_enabled(void) { return static_branch_unlikely(&__dynamic_affinity_switch); } + +#ifdef CONFIG_SCHED_PARAL +bool sched_paral_used(void); +#else +static inline bool sched_paral_used(void) { return false; } +#endif #endif #ifdef CONFIG_QOS_SCHED_SMART_GRID diff --git a/init/Kconfig b/init/Kconfig index c8bd58347a87..925e8517a7e8 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -1484,6 +1484,28 @@ config SCHED_STEAL If unsure, say N here. +# +# For architectures that want to enable the support for SCHED_PARAL +# +config ARCH_SUPPORTS_SCHED_PARAL + bool + +config SCHED_PARAL + bool "Parallelly schedule processes on different NUMA nodes" + depends on ARCH_SUPPORTS_SCHED_PARAL + depends on QOS_SCHED_DYNAMIC_AFFINITY + default n + help + By enabling this feature, processes can be scheduled in parallel + on various NUMA nodes to better utilize the cache in NUMA node. + The usage is restricted to the following scenarios: + 1. No CPU binding is performed for user-space processes; + 2. It is applicable to distributed applications, such as business + architectures with one master and multiple slaves running in + parallel; + 3. The existing "qos dynamic affinity" and "qos smart grid" + features must not be used simultaneously. + config CHECKPOINT_RESTORE bool "Checkpoint/restore support" depends on PROC_FS diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index 417827f2c043..01a9b18d80ce 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -3488,7 +3488,8 @@ static void cpuset_attach_task(struct cpuset *cs, struct task_struct *task) WARN_ON_ONCE(set_cpus_allowed_ptr(task, cpus_attach)); #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY cpumask_copy(prefer_cpus_attach, cs->prefer_cpus); - set_prefer_cpus_ptr(task, prefer_cpus_attach); + if (!sched_paral_used() || !cpumask_empty(prefer_cpus_attach)) + set_prefer_cpus_ptr(task, prefer_cpus_attach); #endif cpuset_change_task_nodemask(task, &cpuset_attach_nodemask_to); @@ -4348,7 +4349,8 @@ static void cpuset_fork(struct task_struct *task) set_cpus_allowed_ptr(task, current->cpus_ptr); #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY - set_prefer_cpus_ptr(task, current->prefer_cpus); + if (!sched_paral_used() || !cpumask_empty(cs->prefer_cpus)) + set_prefer_cpus_ptr(task, current->prefer_cpus); #endif task->mems_allowed = current->mems_allowed; return; diff --git a/kernel/events/core.c b/kernel/events/core.c index f042d6101932..99f46f6ea198 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -13855,6 +13855,19 @@ static int __init perf_event_sysfs_init(void) } device_initcall(perf_event_sysfs_init); +int perf_pmu_type_of_name(const char *name) +{ + unsigned int i; + struct pmu *pmu; + + idr_for_each_entry(&pmu_idr, pmu, i) { + if (!strcmp(pmu->name, name)) + return pmu->type; + } + + return -1; +} + #ifdef CONFIG_CGROUP_PERF static struct cgroup_subsys_state * perf_cgroup_css_alloc(struct cgroup_subsys_state *parent_css) diff --git a/kernel/fork.c b/kernel/fork.c index 96c6a9e446ac..8b2ff47de685 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -631,8 +631,7 @@ void free_task(struct task_struct *tsk) free_kthread_struct(tsk); bpf_task_storage_free(tsk); #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY - if (dynamic_affinity_enabled()) - sched_prefer_cpus_free(tsk); + sched_prefer_cpus_free(tsk); #endif #ifdef CONFIG_QOS_SCHED_SMART_GRID if (smart_grid_enabled()) @@ -2451,11 +2450,9 @@ __latent_entropy struct task_struct *copy_process( #endif #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY - if (dynamic_affinity_enabled()) { - retval = sched_prefer_cpus_fork(p, current->prefer_cpus); - if (retval) - goto bad_fork_free; - } + retval = sched_prefer_cpus_fork(p, current->prefer_cpus); + if (retval) + goto bad_fork_free; #endif lockdep_assert_irqs_enabled(); diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index 7a9e6896c699..793019869da9 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -7,6 +7,10 @@ * Copyright(C) 2007, Red Hat, Inc., Ingo Molnar */ +#ifdef CONFIG_SCHED_PARAL +#include <asm/prefer_numa.h> +#endif + /* * This allows printing both to /proc/sched_debug and * to the console @@ -95,6 +99,39 @@ static void sched_feat_disable(int i) { }; static void sched_feat_enable(int i) { }; #endif /* CONFIG_JUMP_LABEL */ +#ifdef CONFIG_SCHED_PARAL +static void sched_feat_disable_paral(char *cmp) +{ + struct task_struct *tsk, *t; + + if (strncmp(cmp, "PARAL", 5) == 0) { + read_lock(&tasklist_lock); + for_each_process(tsk) { + if (tsk->flags & PF_KTHREAD || is_global_init(tsk)) + continue; + + for_each_thread(tsk, t) + cpumask_clear(t->prefer_cpus); + } + read_unlock(&tasklist_lock); + } +} + +static bool sched_feat_enable_paral(char *cmp) +{ + if (strncmp(cmp, "PARAL", 5) != 0) + return true; + + if (probe_pmu_numa_event() != 0) + return false; + + return true; +} +#else +static void sched_feat_disable_paral(char *cmp) {}; +static bool sched_feat_enable_paral(char *cmp) { return true; }; +#endif /* CONFIG_SCHED_PARAL */ + static int sched_feat_set(char *cmp) { int i; @@ -111,8 +148,12 @@ static int sched_feat_set(char *cmp) if (neg) { sysctl_sched_features &= ~(1UL << i); + sched_feat_disable_paral(cmp); sched_feat_disable(i); } else { + if (!sched_feat_enable_paral(cmp)) + return -EPERM; + sysctl_sched_features |= (1UL << i); sched_feat_enable(i); } @@ -1045,7 +1086,7 @@ void proc_sched_show_task(struct task_struct *p, struct pid_namespace *ns, P_SCHEDSTAT(nr_wakeups_passive); P_SCHEDSTAT(nr_wakeups_idle); #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY - if (dynamic_affinity_enabled()) { + if (dynamic_affinity_enabled() || sched_paral_used()) { P_SCHEDSTAT(nr_wakeups_preferred_cpus); P_SCHEDSTAT(nr_wakeups_force_preferred_cpus); } diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 71661d6c5b54..8a32d0ac4a8b 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -75,6 +75,10 @@ #endif #include <linux/sched/grid_qos.h> +#ifdef CONFIG_SCHED_PARAL +#include <asm/prefer_numa.h> +#endif + /* * The initial- and re-scaling of tunables is configurable * @@ -9057,6 +9061,12 @@ static int find_energy_efficient_cpu(struct task_struct *p, int prev_cpu) } #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY +#ifdef CONFIG_SCHED_PARAL +bool sched_paral_used(void) +{ + return sched_feat(PARAL); +} +#endif DEFINE_STATIC_KEY_FALSE(__dynamic_affinity_switch); @@ -9084,16 +9094,15 @@ __setup("dynamic_affinity=", dynamic_affinity_switch_setup); static inline bool prefer_cpus_valid(struct task_struct *p) { - struct cpumask *prefer_cpus; + struct cpumask *prefer_cpus = task_prefer_cpus(p); - if (!dynamic_affinity_enabled()) - return false; - - prefer_cpus = task_prefer_cpus(p); + if (dynamic_affinity_enabled() || sched_paral_used()) { + return !cpumask_empty(prefer_cpus) && + !cpumask_equal(prefer_cpus, p->cpus_ptr) && + cpumask_subset(prefer_cpus, p->cpus_ptr); + } - return !cpumask_empty(prefer_cpus) && - !cpumask_equal(prefer_cpus, p->cpus_ptr) && - cpumask_subset(prefer_cpus, p->cpus_ptr); + return false; } static inline unsigned long taskgroup_cpu_util(struct task_group *tg, @@ -9193,6 +9202,14 @@ static void set_task_select_cpus(struct task_struct *p, int *idlest_cpu, } rcu_read_unlock(); + /* In extreme cases, it may cause uneven system load. */ + if (sched_paral_used() && sysctl_sched_util_low_pct == 100 && nr_cpus_valid > 0) { + p->select_cpus = p->prefer_cpus; + if (sd_flag & SD_BALANCE_WAKE) + schedstat_inc(p->stats.nr_wakeups_preferred_cpus); + return; + } + /* * Follow cases should select cpus_ptr, checking by condition of * tg_capacity > nr_cpus_valid: @@ -14679,6 +14696,12 @@ static void task_fork_fair(struct task_struct *p) if (curr) update_curr(cfs_rq); place_entity(cfs_rq, se, ENQUEUE_INITIAL); + +#ifdef CONFIG_SCHED_PARAL + if (sched_paral_used()) + set_task_paral_node(p); +#endif + rq_unlock(rq, &rf); } diff --git a/kernel/sched/features.h b/kernel/sched/features.h index ea7ba74810e3..67939d04542f 100644 --- a/kernel/sched/features.h +++ b/kernel/sched/features.h @@ -61,6 +61,10 @@ SCHED_FEAT(SIS_UTIL, true) SCHED_FEAT(STEAL, false) #endif +#ifdef CONFIG_SCHED_PARAL +SCHED_FEAT(PARAL, false) +#endif + /* * Issue a WARN when we do multiple update_rq_clock() calls * in a single rq->lock section. Default disabled because the -- 2.25.1

2 1

[PATCH -next v7 0/7] arm64: entry: Convert to generic irq entry
by Jinjie Ruan 28 Jul '25

28 Jul '25

Currently, x86, Riscv, Loongarch use the generic entry. Also convert arm64 to use the generic entry infrastructure from kernel/entry/*. The generic entry makes maintainers' work easier and codes more elegant, which will make PREEMPT_DYNAMIC and PREEMPT_LAZY use the generic entry common code and remove a lot of duplicate code. Since commit a70e9f647f50 ("entry: Split generic entry into generic exception and syscall entry") split the generic entry into generic irq entry and generic syscall entry, it is time to convert arm64 to use the generic irq entry. And ARM64 will be completely converted to generic entry in the upcoming patch series. The main convert steps are as follows: - Split generic entry into generic irq entry and generic syscall to make the single patch more concentrated in switching to one thing. - Make arm64 easier to use irqentry_enter/exit(). - Make arm64 closer to the PREEMPT_DYNAMIC code of generic entry. - Switch to generic irq entry. It was tested ok with following test cases on QEMU virt platform: - Perf tests. - Different `dynamic preempt` mode switch. - Pseudo NMI tests. - Stress-ng CPU stress test. - MTE test case in Documentation/arch/arm64/memory-tagging-extension.rst and all test cases in tools/testing/selftests/arm64/mte/*. The test QEMU configuration is as follows: qemu-system-aarch64 \ -M virt,gic-version=3,virtualization=on,mte=on \ -cpu max,pauth-impdef=on \ -kernel Image \ -smp 8,sockets=1,cores=4,threads=2 \ -m 512m \ -nographic \ -no-reboot \ -device virtio-rng-pci \ -append "root=/dev/vda rw console=ttyAMA0 kgdboc=ttyAMA0,115200 \ earlycon preempt=voluntary irqchip.gicv3_pseudo_nmi=1" \ -drive if=none,file=images/rootfs.ext4,format=raw,id=hd0 \ -device virtio-blk-device,drive=hd0 \ Changes in v7: - Rebased on v6.16-rc7 and remove the merged first patch. - Update the commit message. Changes in v6: - Rebased on 6.14 rc2 next. - Put the syscall bits aside and split it out. - Have the split patch before the arm64 changes. - Merge some tightly coupled patches. - Adjust the order of some patches to make them more reasonable. - Define regs_irqs_disabled() by inline function. - Define interrupts_enabled() in terms of regs_irqs_disabled(). - Delete the fast_interrupts_enabled() macro. - irqentry_state_t -> arm64_irqentry_state_t. - Remove arch_exit_to_user_mode_prepare() and pull local_daif_mask() later in the arm64 exit sequence - Update the commit message. Changes in v5: - Not change arm32 and keep inerrupts_enabled() macro for gicv3 driver. - Move irqentry_state definition into arch/arm64/kernel/entry-common.c. - Avoid removing the __enter_from_*() and __exit_to_*() wrappers. - Update "irqentry_state_t ret/irq_state" to "state" to keep it consistently. - Use generic irq entry header for PREEMPT_DYNAMIC after split the generic entry. - Also refactor the ARM64 syscall code. - Introduce arch_ptrace_report_syscall_entry/exit(), instead of arch_pre/post_report_syscall_entry/exit() to simplify code. - Make the syscall patches clear separation. - Update the commit message. Changes in v4: - Rework/cleanup split into a few patches as Mark suggested. - Replace interrupts_enabled() macro with regs_irqs_disabled(), instead of left it here. - Remove rcu and lockdep state in pt_regs by using temporary irqentry_state_t as Mark suggested. - Remove some unnecessary intermediate functions to make it clear. - Rework preempt irq and PREEMPT_DYNAMIC code to make the switch more clear. - arch_prepare_*_entry/exit() -> arch_pre_*_entry/exit(). - Expand the arch functions comment. - Make arch functions closer to its caller. - Declare saved_reg in for block. - Remove arch_exit_to_kernel_mode_prepare(), arch_enter_from_kernel_mode(). - Adjust "Add few arch functions to use generic entry" patch to be the penultimate. - Update the commit message. - Add suggested-by. Changes in v3: - Test the MTE test cases. - Handle forget_syscall() in arch_post_report_syscall_entry() - Make the arch funcs not use __weak as Thomas suggested, so move the arch funcs to entry-common.h, and make arch_forget_syscall() folded in arch_post_report_syscall_entry() as suggested. - Move report_single_step() to thread_info.h for arm64 - Change __always_inline() to inline, add inline for the other arch funcs. - Remove unused signal.h for entry-common.h. - Add Suggested-by. - Update the commit message. Changes in v2: - Add tested-by. - Fix a bug that not call arch_post_report_syscall_entry() in syscall_trace_enter() if ptrace_report_syscall_entry() return not zero. - Refactor report_syscall(). - Add comment for arch_prepare_report_syscall_exit(). - Adjust entry-common.h header file inclusion to alphabetical order. - Update the commit message. Jinjie Ruan (7): arm64: ptrace: Replace interrupts_enabled() with regs_irqs_disabled() arm64: entry: Refactor the entry and exit for exceptions from EL1 arm64: entry: Rework arm64_preempt_schedule_irq() arm64: entry: Use preempt_count() and need_resched() helper arm64: entry: Refactor preempt_schedule_irq() check code arm64: entry: Move arm64_preempt_schedule_irq() into __exit_to_kernel_mode() arm64: entry: Switch to generic IRQ entry arch/arm64/Kconfig | 1 + arch/arm64/include/asm/daifflags.h | 2 +- arch/arm64/include/asm/entry-common.h | 56 ++++ arch/arm64/include/asm/preempt.h | 2 - arch/arm64/include/asm/ptrace.h | 13 +- arch/arm64/include/asm/xen/events.h | 2 +- arch/arm64/kernel/acpi.c | 2 +- arch/arm64/kernel/debug-monitors.c | 2 +- arch/arm64/kernel/entry-common.c | 411 +++++++++----------------- arch/arm64/kernel/sdei.c | 2 +- arch/arm64/kernel/signal.c | 3 +- kernel/entry/common.c | 16 +- 12 files changed, 217 insertions(+), 295 deletions(-) create mode 100644 arch/arm64/include/asm/entry-common.h -- 2.34.1

1 7

[PATCH OLK-6.6] mm/dpool: do not use uninitialized variable
by Wupeng Ma 28 Jul '25

28 Jul '25

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8S9BY -------------------------------- Variable ret is not initialized if dpool_create() failed during dpool_init(). Do not dump it to fix this problem. Fixes: cf98ffbc116b ("mm/dynamic_pool: fill dpool with pagelist") Signed-off-by: Wupeng Ma <mawupeng1(a)huawei.com> --- mm/dynamic_pool.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/dynamic_pool.c b/mm/dynamic_pool.c index 063fb6092401..1b79fb099cc9 100644 --- a/mm/dynamic_pool.c +++ b/mm/dynamic_pool.c @@ -1701,7 +1701,7 @@ int dpool_init(struct dpool_info *arg) dpool = dpool_create(arg->memcg, &pagelist_dpool_ops); if (!dpool) { - pr_err("init failed, create failed. ret: %d\n", ret); + pr_err("init failed, create failed.\n"); ret = -ENOMEM; goto unlock; } -- 2.43.0

2 1

[PATCH OLK-5.10] hugeltb: Fix null ptr exception during drain movable pcp
by Wupeng Ma 28 Jul '25

28 Jul '25

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICLL1L -------------------------------- During our test, with movable_node enabled, null ptr exception will happen with the following command: echo 10 > /proc/sys/vm/nr_hugepages Which will lead to kernel panic with the following log: Unable to handle kernel paging request at virtual address 00000000000013f0 pc : set_max_huge_pages+0x52c/0x5fc lr : set_max_huge_pages+0x514/0x5fc Call trace: set_max_huge_pages+0x52c/0x5fc hugetlb_sysctl_handler+0xe4/0x10c proc_sys_call_handler+0x13c/0x260 proc_sys_write+0x18/0x3c new_sync_write+0xec/0x1a0 vfs_write+0x208/0x2b0 ksys_write+0x70/0x110 __arm64_sys_write+0x20/0x40 el0_sync+0x168/0x180 To fix this, stop drian pcp if nid is not valid. Fixes: 539370f5231b ("hugetlb: drain pcp for movable zone before alloc") Signed-off-by: Wupeng Ma <mawupeng1(a)huawei.com> --- mm/hugetlb.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/mm/hugetlb.c b/mm/hugetlb.c index eff60dc4135d9..5307e9c2b0538 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -3113,7 +3113,8 @@ static int set_max_huge_pages(struct hstate *h, unsigned long count, int nid, * drain pcp for movable zone to increase the success rate for * hugetlb memory allocation if movable_node enabled */ - if (movable_node_is_enabled() && count > persistent_huge_pages(h)) + if ((nid != NUMA_NO_NODE) && movable_node_is_enabled() && + count > persistent_huge_pages(h)) hugetlb_drain_movable_pcp(h, nid); while (count > persistent_huge_pages(h)) { -- 2.43.0

2 1

[PATCH OLK-6.6 v6 0/3] arm64: Add hardware xcall framework support
by Jinjie Ruan 28 Jul '25

28 Jul '25

Add hardware xcall framework support. Changes in v6: - Use system_uses_xcall_xint() instead of static key for hardware xcall. - xcall_scno_enabled -> xcall_scno_count for each task_struct. - Use xcall_scno_enabled instead of static_branch_unlikely() for performance-insensitive code. - Some cleanup. Changes in v5: - Not adjust the software xcall and hardware xcall hwcap detect order. - Not udapte xcall 1.0 enrty code. - Some cleanup. Changes in v4: - Merge the percpu and csel patch. - Use static key to improve the software and hardware xcall code. - Some cleanup. - Fix a build problem. Changes in v3: - Optimize the use of hardware xcall and not use them at the same time. - Use syscall table instead of xcall enable bitmap in hardware xcall. - Use percpu syscall table cache to simplfy code. - Use csel to improve the jump instruction. Changes in v2: - Split software xcall and hardware xcall code and use same bitmap to check xcall enable. Jinjie Ruan (3): arm64: Refactor the xcall proc code arm64: Reserve a kabi in task_struct exclusively for xcall arm64: Add hardware xcall framework support arch/arm64/Kconfig.turbo | 1 + arch/arm64/include/asm/cpufeature.h | 2 - arch/arm64/include/asm/exception.h | 5 + arch/arm64/include/asm/mmu_context.h | 7 + arch/arm64/include/asm/xcall.h | 86 ++++++++++++ arch/arm64/kernel/Makefile | 1 + arch/arm64/kernel/asm-offsets.c | 2 +- arch/arm64/kernel/cpufeature.c | 41 +++--- arch/arm64/kernel/entry-common.c | 2 +- arch/arm64/kernel/entry.S | 6 +- arch/arm64/kernel/xcall/Makefile | 2 + arch/arm64/kernel/xcall/entry.S | 195 +++++++++++++++++++++++++++ arch/arm64/kernel/xcall/xcall.c | 75 +++++++++++ fs/proc/Makefile | 1 + fs/proc/base.c | 109 --------------- fs/proc/internal.h | 4 + fs/proc/proc_xcall.c | 178 ++++++++++++++++++++++++ include/linux/sched.h | 2 +- kernel/fork.c | 15 +-- 19 files changed, 593 insertions(+), 141 deletions(-) create mode 100644 arch/arm64/include/asm/xcall.h create mode 100644 arch/arm64/kernel/xcall/Makefile create mode 100644 arch/arm64/kernel/xcall/entry.S create mode 100644 arch/arm64/kernel/xcall/xcall.c create mode 100644 fs/proc/proc_xcall.c -- 2.34.1

2 4

[PATCH OLK-5.10] ovl: fix general protection fault in security_inode_getattr
by Zizhi Wo 28 Jul '25

28 Jul '25

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICP68T -------------------------------- When mounting overlay with "metacopy=on", if lowerdir contains a file A: 1) Thread p1 executes ln A B (creates a hard link to file A as B); 2) Thread p2 simultaneously opens file B; then a UAF issue can occur due to concurrency. p1 p2 sys_open ovl_lookup B ovl_alloc_entry oe->numlower = 0 ln A B ovl_link ovl_open ovl_maybe_copy_up ovl_open_need_copy_up ovl_already_copied_up ovl_dentry_needs_data_copy_up ovl_has_upperdata // false as metacopy=on ovl_copy_up_flags ovl_copy_up_one ovl_path_lower path = {} vfs_getattr(path, xxx) security_inode_getattr(path) ...path->dentry... ----------null-ptr-deref!!! To resolve this issue, add a check in ovl_copy_up_one(). If lowerpath does not exist, return an error immediately. Fixes: 0c2888749363 ("ovl: A new xattr OVL_XATTR_METACOPY for file on upper") Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- fs/overlayfs/copy_up.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/fs/overlayfs/copy_up.c b/fs/overlayfs/copy_up.c index a1ec45fc77d8..4324d84ac1d7 100644 --- a/fs/overlayfs/copy_up.c +++ b/fs/overlayfs/copy_up.c @@ -879,6 +879,12 @@ static int ovl_copy_up_one(struct dentry *parent, struct dentry *dentry, return -EROFS; ovl_path_lower(dentry, &ctx.lowerpath); + + if (unlikely(!ctx.lowerpath.dentry)) { + pr_err("prevention GPF in security_inode_getattr()\n"); + return -EIO; + } + err = vfs_getattr(&ctx.lowerpath, &ctx.stat, STATX_BASIC_STATS, AT_STATX_SYNC_AS_STAT); if (err) -- 2.46.1

2 1

[PATCH openEuler-1.0-LTS] ovl: fix general protection fault in security_inode_getattr
by Zizhi Wo 28 Jul '25

28 Jul '25

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/ICP68T -------------------------------- When mounting overlay with "metacopy=on", if lowerdir contains a file A: 1) Thread p1 executes ln A B (creates a hard link to file A as B); 2) Thread p2 simultaneously opens file B; then a UAF issue can occur due to concurrency. p1 p2 sys_open ovl_lookup B ovl_alloc_entry oe->numlower = 0 ln A B ovl_link ovl_open ovl_maybe_copy_up ovl_open_need_copy_up ovl_already_copied_up ovl_dentry_needs_data_copy_up ovl_has_upperdata // false as metacopy=on ovl_copy_up_flags ovl_copy_up_one ovl_path_lower path = {} vfs_getattr(path, xxx) security_inode_getattr(path) ...path->dentry... ----------null-ptr-deref!!! To resolve this issue, add a check in ovl_copy_up_one(). If lowerpath does not exist, return an error immediately. Fixes: 0c2888749363 ("ovl: A new xattr OVL_XATTR_METACOPY for file on upper") Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- fs/overlayfs/copy_up.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/fs/overlayfs/copy_up.c b/fs/overlayfs/copy_up.c index 96749f0af9b7..bdf8aee5983b 100644 --- a/fs/overlayfs/copy_up.c +++ b/fs/overlayfs/copy_up.c @@ -773,6 +773,12 @@ static int ovl_copy_up_one(struct dentry *parent, struct dentry *dentry, return -EROFS; ovl_path_lower(dentry, &ctx.lowerpath); + + if (unlikely(!ctx.lowerpath.dentry)) { + pr_err("prevention GPF in security_inode_getattr()\n"); + return -EIO; + } + err = vfs_getattr(&ctx.lowerpath, &ctx.stat, STATX_BASIC_STATS, AT_STATX_SYNC_AS_STAT); if (err) -- 2.39.2

2 1

[PATCH openEuler-1.0-LTS] usb: typec: altmodes/displayport: do not index invalid pin_assignments
by Tengda Wu 28 Jul '25

28 Jul '25

From: RD Babiera <rdbabiera(a)google.com> mainline inclusion from mainline-v6.16-rc5 commit af4db5a35a4ef7a68046883bfd12468007db38f1 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/ICOXG8 CVE: CVE-2025-38391 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- A poorly implemented DisplayPort Alt Mode port partner can indicate that its pin assignment capabilities are greater than the maximum value, DP_PIN_ASSIGN_F. In this case, calls to pin_assignment_show will cause a BRK exception due to an out of bounds array access. Prevent for loop in pin_assignment_show from accessing invalid values in pin_assignments by adding DP_PIN_ASSIGN_MAX value in typec_dp.h and using i < DP_PIN_ASSIGN_MAX as a loop condition. Fixes: 0e3bb7d6894d ("usb: typec: Add driver for DisplayPort alternate mode") Cc: stable <stable(a)kernel.org> Signed-off-by: RD Babiera <rdbabiera(a)google.com> Reviewed-by: Badhri Jagan Sridharan <badhri(a)google.com> Reviewed-by: Heikki Krogerus <heikki.krogerus(a)linux.intel.com> Link: https://lore.kernel.org/r/20250618224943.3263103-2-rdbabiera@google.com Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Conflicts: drivers/usb/typec/altmodes/displayport.c [Context conflict introduced by a refactoring patch] Signed-off-by: Tengda Wu <wutengda2(a)huawei.com> --- drivers/usb/typec/altmodes/displayport.c | 2 +- include/linux/usb/typec_dp.h | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/usb/typec/altmodes/displayport.c b/drivers/usb/typec/altmodes/displayport.c index 3f06e94771a7..90bb0baed624 100644 --- a/drivers/usb/typec/altmodes/displayport.c +++ b/drivers/usb/typec/altmodes/displayport.c @@ -482,7 +482,7 @@ static ssize_t pin_assignment_show(struct device *dev, else assignments = DP_CAP_DFP_D_PIN_ASSIGN(dp->alt->vdo); - for (i = 0; assignments; assignments >>= 1, i++) { + for (i = 0; assignments && i < DP_PIN_ASSIGN_MAX; assignments >>= 1, i++) { if (assignments & 1) { if (i == cur) len += sprintf(buf + len, "[%s] ", diff --git a/include/linux/usb/typec_dp.h b/include/linux/usb/typec_dp.h index 55ae781d60a9..28f71228ae92 100644 --- a/include/linux/usb/typec_dp.h +++ b/include/linux/usb/typec_dp.h @@ -51,6 +51,7 @@ enum { DP_PIN_ASSIGN_D, DP_PIN_ASSIGN_E, DP_PIN_ASSIGN_F, /* Not supported after v1.0b */ + DP_PIN_ASSIGN_MAX, }; /* DisplayPort alt mode specific commands */ -- 2.34.1

2 1

[PATCH OLK-6.6 V1] sched: Support NUMA parallel scheduling for multiple processes
by Cheng Yu 28 Jul '25

28 Jul '25

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/ICBBNL -------------------------------- For architectures with multiple NUMA node levels and large distances between nodes, a better approach is to support processes running in parallel on each NUMA node. The usage is restricted to the following scenarios: 1. No CPU binding for user-space processes; 2. It is applicable to distributed applications, such as business architectures with one master and multiple slaves running in parallel; 3. The existing "qos dynamic affinity" and "qos smart grid" features must not be used simultaneously. Signed-off-by: Cheng Yu <serein.chengyu(a)huawei.com> --- arch/arm64/Kconfig | 1 + arch/arm64/configs/openeuler_defconfig | 1 + arch/arm64/include/asm/prefer_numa.h | 13 +++++ arch/arm64/kernel/Makefile | 1 + arch/arm64/kernel/prefer_numa.c | 68 ++++++++++++++++++++++++++ fs/proc/array.c | 3 -- include/linux/perf_event.h | 2 + include/linux/sched.h | 6 +++ init/Kconfig | 22 +++++++++ kernel/cgroup/cpuset.c | 6 ++- kernel/events/core.c | 13 +++++ kernel/fork.c | 11 ++--- kernel/sched/debug.c | 43 +++++++++++++++- kernel/sched/fair.c | 39 ++++++++++++--- kernel/sched/features.h | 4 ++ 15 files changed, 212 insertions(+), 21 deletions(-) create mode 100644 arch/arm64/include/asm/prefer_numa.h create mode 100644 arch/arm64/kernel/prefer_numa.c diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 5422d1502fd6..b1f550c8c82a 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -105,6 +105,7 @@ config ARM64 select ARCH_SUPPORTS_ATOMIC_RMW select ARCH_SUPPORTS_INT128 if CC_HAS_INT128 select ARCH_SUPPORTS_NUMA_BALANCING + select ARCH_SUPPORTS_SCHED_PARAL select ARCH_SUPPORTS_PAGE_TABLE_CHECK select ARCH_SUPPORTS_PER_VMA_LOCK select ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 3cfff0701479..3d352fb1ae57 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -209,6 +209,7 @@ CONFIG_USER_NS=y CONFIG_PID_NS=y CONFIG_NET_NS=y CONFIG_SCHED_STEAL=y +CONFIG_SCHED_PARAL=y CONFIG_CHECKPOINT_RESTORE=y CONFIG_SCHED_AUTOGROUP=y CONFIG_RELAY=y diff --git a/arch/arm64/include/asm/prefer_numa.h b/arch/arm64/include/asm/prefer_numa.h new file mode 100644 index 000000000000..6c8e2b2142b9 --- /dev/null +++ b/arch/arm64/include/asm/prefer_numa.h @@ -0,0 +1,13 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +#ifndef __ASM_PREFER_NUMA_H +#define __ASM_PREFER_NUMA_H + +#include <linux/sched.h> + +#define PROBE_NUMA_PMU_NAME "hisi_sccl3_hha0" +#define PROBE_NUMA_PMU_EVENT 0x02 + +void set_task_paral_node(struct task_struct *p); +int probe_pmu_numa_event(void); + +#endif /* __ASM_PREFER_NUMA_H */ diff --git a/arch/arm64/kernel/Makefile b/arch/arm64/kernel/Makefile index 3d404a2cc961..b936be9d8baa 100644 --- a/arch/arm64/kernel/Makefile +++ b/arch/arm64/kernel/Makefile @@ -84,6 +84,7 @@ obj-$(CONFIG_IPI_AS_NMI) += ipi_nmi.o obj-$(CONFIG_HISI_VIRTCCA_GUEST) += virtcca_cvm_guest.o virtcca_cvm_tsi.o obj-$(CONFIG_HISI_VIRTCCA_HOST) += virtcca_cvm_host.o CFLAGS_patch-scs.o += -mbranch-protection=none +obj-$(CONFIG_SCHED_PARAL) += prefer_numa.o # Force dependency (vdso*-wrap.S includes vdso.so through incbin) $(obj)/vdso-wrap.o: $(obj)/vdso/vdso.so diff --git a/arch/arm64/kernel/prefer_numa.c b/arch/arm64/kernel/prefer_numa.c new file mode 100644 index 000000000000..e6f3f43fb97a --- /dev/null +++ b/arch/arm64/kernel/prefer_numa.c @@ -0,0 +1,68 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * choose a prefer numa node + * + * Copyright (C) 2025 Huawei Limited. + */ +#include <linux/perf_event.h> +#include <asm/prefer_numa.h> + +static atomic_t paral_nid_last = ATOMIC_INIT(-1); + +int probe_pmu_numa_event(void) +{ + struct perf_event *event; + struct perf_event_attr attr = {}; + int type = perf_pmu_type_of_name(PROBE_NUMA_PMU_NAME); + + if (type == -1) + return -EINVAL; + + attr.type = type; + attr.config = PROBE_NUMA_PMU_EVENT; + attr.size = sizeof(struct perf_event_attr); + attr.pinned = 1; + attr.disabled = 1; + attr.sample_period = 0; + + event = perf_event_create_kernel_counter(&attr, smp_processor_id(), + NULL, NULL, NULL); + if (IS_ERR(event)) + return PTR_ERR(event); + + perf_event_release_kernel(event); + + return 0; +} + +static inline unsigned int update_sched_paral_nid(void) +{ + return (unsigned int)atomic_inc_return(&paral_nid_last); +} + +void set_task_paral_node(struct task_struct *p) +{ + int nid; + int i = 0; + const cpumask_t *cpus_mask; + + if (is_global_init(current)) + return; + + if (p->flags & PF_KTHREAD || p->tgid != p->pid) + return; + + while (i < nr_node_ids) { + nid = update_sched_paral_nid() % nr_node_ids; + cpus_mask = cpumask_of_node(nid); + + if (cpumask_empty(cpus_mask) || + !cpumask_subset(cpus_mask, p->cpus_ptr)) { + i++; + continue; + } + + cpumask_copy(p->prefer_cpus, cpus_mask); + break; + } +} diff --git a/fs/proc/array.c b/fs/proc/array.c index a933a878df3c..6a4b0a850dce 100644 --- a/fs/proc/array.c +++ b/fs/proc/array.c @@ -439,9 +439,6 @@ __weak void arch_proc_pid_thread_features(struct seq_file *m, #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY static void task_cpus_preferred(struct seq_file *m, struct task_struct *task) { - if (!dynamic_affinity_enabled()) - return; - seq_printf(m, "Cpus_preferred:\t%*pb\n", cpumask_pr_args(task->prefer_cpus)); seq_printf(m, "Cpus_preferred_list:\t%*pbl\n", diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index 826fb16906fe..14ed13f4b408 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -1778,6 +1778,7 @@ extern void perf_event_task_tick(void); extern int perf_event_account_interrupt(struct perf_event *event); extern int perf_event_period(struct perf_event *event, u64 value); extern u64 perf_event_pause(struct perf_event *event, bool reset); +extern int perf_pmu_type_of_name(const char *name); #else /* !CONFIG_PERF_EVENTS: */ static inline void * perf_aux_output_begin(struct perf_output_handle *handle, @@ -1864,6 +1865,7 @@ static inline u64 perf_event_pause(struct perf_event *event, bool reset) { return 0; } +static inline int perf_pmu_type_of_name(const char *name) { return -1; } #endif #if defined(CONFIG_PERF_EVENTS) && defined(CONFIG_CPU_SUP_INTEL) diff --git a/include/linux/sched.h b/include/linux/sched.h index 3979c34e9b83..ee10780715f1 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2627,6 +2627,12 @@ static inline bool dynamic_affinity_enabled(void) { return static_branch_unlikely(&__dynamic_affinity_switch); } + +#ifdef CONFIG_SCHED_PARAL +bool sched_paral_used(void); +#else +static inline bool sched_paral_used(void) { return false; } +#endif #endif #ifdef CONFIG_QOS_SCHED_SMART_GRID diff --git a/init/Kconfig b/init/Kconfig index c8bd58347a87..925e8517a7e8 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -1484,6 +1484,28 @@ config SCHED_STEAL If unsure, say N here. +# +# For architectures that want to enable the support for SCHED_PARAL +# +config ARCH_SUPPORTS_SCHED_PARAL + bool + +config SCHED_PARAL + bool "Parallelly schedule processes on different NUMA nodes" + depends on ARCH_SUPPORTS_SCHED_PARAL + depends on QOS_SCHED_DYNAMIC_AFFINITY + default n + help + By enabling this feature, processes can be scheduled in parallel + on various NUMA nodes to better utilize the cache in NUMA node. + The usage is restricted to the following scenarios: + 1. No CPU binding is performed for user-space processes; + 2. It is applicable to distributed applications, such as business + architectures with one master and multiple slaves running in + parallel; + 3. The existing "qos dynamic affinity" and "qos smart grid" + features must not be used simultaneously. + config CHECKPOINT_RESTORE bool "Checkpoint/restore support" depends on PROC_FS diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index 417827f2c043..01a9b18d80ce 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -3488,7 +3488,8 @@ static void cpuset_attach_task(struct cpuset *cs, struct task_struct *task) WARN_ON_ONCE(set_cpus_allowed_ptr(task, cpus_attach)); #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY cpumask_copy(prefer_cpus_attach, cs->prefer_cpus); - set_prefer_cpus_ptr(task, prefer_cpus_attach); + if (!sched_paral_used() || !cpumask_empty(prefer_cpus_attach)) + set_prefer_cpus_ptr(task, prefer_cpus_attach); #endif cpuset_change_task_nodemask(task, &cpuset_attach_nodemask_to); @@ -4348,7 +4349,8 @@ static void cpuset_fork(struct task_struct *task) set_cpus_allowed_ptr(task, current->cpus_ptr); #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY - set_prefer_cpus_ptr(task, current->prefer_cpus); + if (!sched_paral_used() || !cpumask_empty(cs->prefer_cpus)) + set_prefer_cpus_ptr(task, current->prefer_cpus); #endif task->mems_allowed = current->mems_allowed; return; diff --git a/kernel/events/core.c b/kernel/events/core.c index f042d6101932..99f46f6ea198 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -13855,6 +13855,19 @@ static int __init perf_event_sysfs_init(void) } device_initcall(perf_event_sysfs_init); +int perf_pmu_type_of_name(const char *name) +{ + unsigned int i; + struct pmu *pmu; + + idr_for_each_entry(&pmu_idr, pmu, i) { + if (!strcmp(pmu->name, name)) + return pmu->type; + } + + return -1; +} + #ifdef CONFIG_CGROUP_PERF static struct cgroup_subsys_state * perf_cgroup_css_alloc(struct cgroup_subsys_state *parent_css) diff --git a/kernel/fork.c b/kernel/fork.c index 96c6a9e446ac..8b2ff47de685 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -631,8 +631,7 @@ void free_task(struct task_struct *tsk) free_kthread_struct(tsk); bpf_task_storage_free(tsk); #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY - if (dynamic_affinity_enabled()) - sched_prefer_cpus_free(tsk); + sched_prefer_cpus_free(tsk); #endif #ifdef CONFIG_QOS_SCHED_SMART_GRID if (smart_grid_enabled()) @@ -2451,11 +2450,9 @@ __latent_entropy struct task_struct *copy_process( #endif #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY - if (dynamic_affinity_enabled()) { - retval = sched_prefer_cpus_fork(p, current->prefer_cpus); - if (retval) - goto bad_fork_free; - } + retval = sched_prefer_cpus_fork(p, current->prefer_cpus); + if (retval) + goto bad_fork_free; #endif lockdep_assert_irqs_enabled(); diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c index 7a9e6896c699..793019869da9 100644 --- a/kernel/sched/debug.c +++ b/kernel/sched/debug.c @@ -7,6 +7,10 @@ * Copyright(C) 2007, Red Hat, Inc., Ingo Molnar */ +#ifdef CONFIG_SCHED_PARAL +#include <asm/prefer_numa.h> +#endif + /* * This allows printing both to /proc/sched_debug and * to the console @@ -95,6 +99,39 @@ static void sched_feat_disable(int i) { }; static void sched_feat_enable(int i) { }; #endif /* CONFIG_JUMP_LABEL */ +#ifdef CONFIG_SCHED_PARAL +static void sched_feat_disable_paral(char *cmp) +{ + struct task_struct *tsk, *t; + + if (strncmp(cmp, "PARAL", 5) == 0) { + read_lock(&tasklist_lock); + for_each_process(tsk) { + if (tsk->flags & PF_KTHREAD || is_global_init(tsk)) + continue; + + for_each_thread(tsk, t) + cpumask_clear(t->prefer_cpus); + } + read_unlock(&tasklist_lock); + } +} + +static bool sched_feat_enable_paral(char *cmp) +{ + if (strncmp(cmp, "PARAL", 5) != 0) + return true; + + if (probe_pmu_numa_event() != 0) + return false; + + return true; +} +#else +static void sched_feat_disable_paral(char *cmp) {}; +static bool sched_feat_enable_paral(char *cmp) { return true; }; +#endif /* CONFIG_SCHED_PARAL */ + static int sched_feat_set(char *cmp) { int i; @@ -111,8 +148,12 @@ static int sched_feat_set(char *cmp) if (neg) { sysctl_sched_features &= ~(1UL << i); + sched_feat_disable_paral(cmp); sched_feat_disable(i); } else { + if (!sched_feat_enable_paral(cmp)) + return -EPERM; + sysctl_sched_features |= (1UL << i); sched_feat_enable(i); } @@ -1045,7 +1086,7 @@ void proc_sched_show_task(struct task_struct *p, struct pid_namespace *ns, P_SCHEDSTAT(nr_wakeups_passive); P_SCHEDSTAT(nr_wakeups_idle); #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY - if (dynamic_affinity_enabled()) { + if (dynamic_affinity_enabled() || sched_paral_used()) { P_SCHEDSTAT(nr_wakeups_preferred_cpus); P_SCHEDSTAT(nr_wakeups_force_preferred_cpus); } diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 71661d6c5b54..8a32d0ac4a8b 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -75,6 +75,10 @@ #endif #include <linux/sched/grid_qos.h> +#ifdef CONFIG_SCHED_PARAL +#include <asm/prefer_numa.h> +#endif + /* * The initial- and re-scaling of tunables is configurable * @@ -9057,6 +9061,12 @@ static int find_energy_efficient_cpu(struct task_struct *p, int prev_cpu) } #ifdef CONFIG_QOS_SCHED_DYNAMIC_AFFINITY +#ifdef CONFIG_SCHED_PARAL +bool sched_paral_used(void) +{ + return sched_feat(PARAL); +} +#endif DEFINE_STATIC_KEY_FALSE(__dynamic_affinity_switch); @@ -9084,16 +9094,15 @@ __setup("dynamic_affinity=", dynamic_affinity_switch_setup); static inline bool prefer_cpus_valid(struct task_struct *p) { - struct cpumask *prefer_cpus; + struct cpumask *prefer_cpus = task_prefer_cpus(p); - if (!dynamic_affinity_enabled()) - return false; - - prefer_cpus = task_prefer_cpus(p); + if (dynamic_affinity_enabled() || sched_paral_used()) { + return !cpumask_empty(prefer_cpus) && + !cpumask_equal(prefer_cpus, p->cpus_ptr) && + cpumask_subset(prefer_cpus, p->cpus_ptr); + } - return !cpumask_empty(prefer_cpus) && - !cpumask_equal(prefer_cpus, p->cpus_ptr) && - cpumask_subset(prefer_cpus, p->cpus_ptr); + return false; } static inline unsigned long taskgroup_cpu_util(struct task_group *tg, @@ -9193,6 +9202,14 @@ static void set_task_select_cpus(struct task_struct *p, int *idlest_cpu, } rcu_read_unlock(); + /* In extreme cases, it may cause uneven system load. */ + if (sched_paral_used() && sysctl_sched_util_low_pct == 100 && nr_cpus_valid > 0) { + p->select_cpus = p->prefer_cpus; + if (sd_flag & SD_BALANCE_WAKE) + schedstat_inc(p->stats.nr_wakeups_preferred_cpus); + return; + } + /* * Follow cases should select cpus_ptr, checking by condition of * tg_capacity > nr_cpus_valid: @@ -14679,6 +14696,12 @@ static void task_fork_fair(struct task_struct *p) if (curr) update_curr(cfs_rq); place_entity(cfs_rq, se, ENQUEUE_INITIAL); + +#ifdef CONFIG_SCHED_PARAL + if (sched_paral_used()) + set_task_paral_node(p); +#endif + rq_unlock(rq, &rf); } diff --git a/kernel/sched/features.h b/kernel/sched/features.h index ea7ba74810e3..67939d04542f 100644 --- a/kernel/sched/features.h +++ b/kernel/sched/features.h @@ -61,6 +61,10 @@ SCHED_FEAT(SIS_UTIL, true) SCHED_FEAT(STEAL, false) #endif +#ifdef CONFIG_SCHED_PARAL +SCHED_FEAT(PARAL, false) +#endif + /* * Issue a WARN when we do multiple update_rq_clock() calls * in a single rq->lock section. Default disabled because the -- 2.25.1

2 1

[PATCH OLK-5.10] usb: typec: altmodes/displayport: do not index invalid pin_assignments
by Tengda Wu 28 Jul '25

28 Jul '25

From: RD Babiera <rdbabiera(a)google.com> stable inclusion from stable-5.10.240 commit 114a977e0f6bf278e05eade055e13fc271f69cf7 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/ICOXG8 CVE: CVE-2025-38391 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit af4db5a35a4ef7a68046883bfd12468007db38f1 upstream. A poorly implemented DisplayPort Alt Mode port partner can indicate that its pin assignment capabilities are greater than the maximum value, DP_PIN_ASSIGN_F. In this case, calls to pin_assignment_show will cause a BRK exception due to an out of bounds array access. Prevent for loop in pin_assignment_show from accessing invalid values in pin_assignments by adding DP_PIN_ASSIGN_MAX value in typec_dp.h and using i < DP_PIN_ASSIGN_MAX as a loop condition. Fixes: 0e3bb7d6894d ("usb: typec: Add driver for DisplayPort alternate mode") Cc: stable <stable(a)kernel.org> Signed-off-by: RD Babiera <rdbabiera(a)google.com> Reviewed-by: Badhri Jagan Sridharan <badhri(a)google.com> Reviewed-by: Heikki Krogerus <heikki.krogerus(a)linux.intel.com> Link: https://lore.kernel.org/r/20250618224943.3263103-2-rdbabiera@google.com Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Tengda Wu <wutengda2(a)huawei.com> --- drivers/usb/typec/altmodes/displayport.c | 2 +- include/linux/usb/typec_dp.h | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/usb/typec/altmodes/displayport.c b/drivers/usb/typec/altmodes/displayport.c index 344bd0de69b7..f063ac8d8101 100644 --- a/drivers/usb/typec/altmodes/displayport.c +++ b/drivers/usb/typec/altmodes/displayport.c @@ -490,7 +490,7 @@ static ssize_t pin_assignment_show(struct device *dev, assignments = get_current_pin_assignments(dp); - for (i = 0; assignments; assignments >>= 1, i++) { + for (i = 0; assignments && i < DP_PIN_ASSIGN_MAX; assignments >>= 1, i++) { if (assignments & 1) { if (i == cur) len += sprintf(buf + len, "[%s] ", diff --git a/include/linux/usb/typec_dp.h b/include/linux/usb/typec_dp.h index 296909ea04f2..afb73b3e0b80 100644 --- a/include/linux/usb/typec_dp.h +++ b/include/linux/usb/typec_dp.h @@ -56,6 +56,7 @@ enum { DP_PIN_ASSIGN_D, DP_PIN_ASSIGN_E, DP_PIN_ASSIGN_F, /* Not supported after v1.0b */ + DP_PIN_ASSIGN_MAX, }; /* DisplayPort alt mode specific commands */ -- 2.34.1

2 1