- Kernel - mailweb.openeuler.org

[PATCH] memcg: support ksm merge any mode per cgroup
by Nanyong Sun 22 Dec '23

22 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OIQR ---------------------------------------------------------------------- Add control file "memory.ksm" to enable ksm per cgroup. Echo to 1 will set all tasks currently in the cgroup to ksm merge any mode, which means ksm gets enabled for all vma's of a process. Meanwhile echo to 0 will disable ksm for them and unmerge the merged pages. Cat the file will show the above state and ksm related profits of this cgroup. Signed-off-by: Nanyong Sun <sunnanyong(a)huawei.com> --- .../admin-guide/cgroup-v1/memory.rst | 1 + mm/memcontrol.c | 110 +++++++++++++++++- 2 files changed, 109 insertions(+), 2 deletions(-) diff --git a/Documentation/admin-guide/cgroup-v1/memory.rst b/Documentation/admin-guide/cgroup-v1/memory.rst index ff456871bf4b..3fdb48435e8e 100644 --- a/Documentation/admin-guide/cgroup-v1/memory.rst +++ b/Documentation/admin-guide/cgroup-v1/memory.rst @@ -109,6 +109,7 @@ Brief summary of control files. memory.kmem.tcp.failcnt show the number of tcp buf memory usage hits limits memory.kmem.tcp.max_usage_in_bytes show max tcp buf memory usage recorded + memory.ksm set/show ksm merge any mode ==================================== ========================================== 1. History diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 2d9a873e5522..30cafc2f22c1 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -73,6 +73,7 @@ #include <linux/uaccess.h> #include <trace/events/vmscan.h> +#include <linux/ksm.h> struct cgroup_subsys memory_cgrp_subsys __read_mostly; EXPORT_SYMBOL(memory_cgrp_subsys); @@ -242,10 +243,15 @@ enum res_type { iter != NULL; \ iter = mem_cgroup_iter(NULL, iter, NULL)) +static inline bool __task_is_dying(struct task_struct *task) +{ + return tsk_is_oom_victim(task) || fatal_signal_pending(task) || + (task->flags & PF_EXITING); +} + static inline bool task_is_dying(void) { - return tsk_is_oom_victim(current) || fatal_signal_pending(current) || - (current->flags & PF_EXITING); + return __task_is_dying(current); } /* Some nice accessors for the vmpressure. */ @@ -5100,6 +5106,98 @@ static int mem_cgroup_slab_show(struct seq_file *m, void *p) } #endif +#ifdef CONFIG_KSM +static int memcg_set_ksm_for_tasks(struct mem_cgroup *memcg, bool enable) +{ + struct task_struct *task; + struct mm_struct *mm; + struct css_task_iter it; + int ret = 0; + + css_task_iter_start(&memcg->css, CSS_TASK_ITER_PROCS, &it); + while (!ret && (task = css_task_iter_next(&it))) { + if (__task_is_dying(task)) + continue; + + mm = get_task_mm(task); + if (!mm) + continue; + + if (mmap_write_lock_killable(mm)) { + mmput(mm); + continue; + } + + if (enable) + ret = ksm_enable_merge_any(mm); + else + ret = ksm_disable_merge_any(mm); + + mmap_write_unlock(mm); + mmput(mm); + } + css_task_iter_end(&it); + + return ret; +} + +static int memory_ksm_show(struct seq_file *m, void *v) +{ + unsigned long ksm_merging_pages = 0; + unsigned long ksm_rmap_items = 0; + long ksm_process_profits = 0; + unsigned int tasks = 0; + struct task_struct *task; + struct mm_struct *mm; + struct css_task_iter it; + struct mem_cgroup *memcg = mem_cgroup_from_seq(m); + + css_task_iter_start(&memcg->css, CSS_TASK_ITER_PROCS, &it); + while ((task = css_task_iter_next(&it))) { + mm = get_task_mm(task); + if (!mm) + continue; + + if (test_bit(MMF_VM_MERGE_ANY, &mm->flags)) + tasks++; + + ksm_rmap_items += mm->ksm_rmap_items; + ksm_merging_pages += mm->ksm_merging_pages; + ksm_process_profits += ksm_process_profit(mm); + mmput(mm); + } + css_task_iter_end(&it); + + seq_printf(m, "merge any tasks: %u\n", tasks); + seq_printf(m, "ksm_rmap_items %lu\n", ksm_rmap_items); + seq_printf(m, "ksm_merging_pages %lu\n", ksm_merging_pages); + seq_printf(m, "ksm_process_profits %ld\n", ksm_process_profits); + return 0; +} + +static ssize_t memory_ksm_write(struct kernfs_open_file *of, char *buf, + size_t nbytes, loff_t off) +{ + bool enable; + int err; + struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); + + buf = strstrip(buf); + if (!buf) + return -EINVAL; + + err = kstrtobool(buf, &enable); + if (err) + return err; + + err = memcg_set_ksm_for_tasks(memcg, enable); + if (err) + return err; + + return nbytes; +} +#endif /* CONFIG_KSM */ + static int memory_stat_show(struct seq_file *m, void *v); #ifdef CONFIG_MEMCG_V1_RECLAIM @@ -5337,6 +5435,14 @@ static struct cftype mem_cgroup_legacy_files[] = { .name = "reclaim", .write = memory_reclaim, }, +#endif +#ifdef CONFIG_KSM + { + .name = "ksm", + .flags = CFTYPE_NOT_ON_ROOT, + .write = memory_ksm_write, + .seq_show = memory_ksm_show, + }, #endif { }, /* terminate */ }; -- 2.25.1

1 0

[PATCH OLK-6.6 v4 0/7] arm64: Add support to turn an IPI as NMI
by Liao Chen 22 Dec '23

22 Dec '23

With pseudo NMIs support available its possible to configure SGIs to be triggered as pseudo NMIs running in NMI context. And kernel features such as: - NMI backtrace can leverage IPI turned as NMI to get a backtrace of CPU stuck in hard lockup using magic SYSRQ. - kgdb relies on NMI support to round up CPUs which are stuck in hard lockup state with interrupts disabled. This patch-set adds framework to turn an IPI as NMI which can be triggered as a pseudo NMI which in turn invokes registered NMI handlers. After this patch-set we should be able to get a backtrace for a CPU stuck in HARDLOCKUP. Sumit Garg (7): arm64: Add framework to turn IPI as NMI irqchip/gic-v3: Enable support for SGIs to act as NMIs arm64: smp: Assign and setup an IPI as NMI nmi: backtrace: Allow runtime arch specific override arm64: ipi_nmi: Add support for NMI backtrace kgdb: Expose default CPUs roundup fallback mechanism arm64: kgdb: Roundup cpus using IPI as NMI arch/arm/include/asm/irq.h | 2 +- arch/arm/kernel/smp.c | 3 +- arch/arm64/include/asm/irq.h | 6 +++ arch/arm64/include/asm/nmi.h | 17 +++++++ arch/arm64/kernel/Makefile | 2 +- arch/arm64/kernel/ipi_nmi.c | 84 ++++++++++++++++++++++++++++++++ arch/arm64/kernel/kgdb.c | 18 +++++++ arch/arm64/kernel/smp.c | 8 +++ arch/mips/include/asm/irq.h | 2 +- arch/mips/kernel/process.c | 3 +- arch/powerpc/include/asm/irq.h | 2 +- arch/powerpc/kernel/stacktrace.c | 3 +- arch/sparc/include/asm/irq_64.h | 2 +- arch/sparc/kernel/process_64.c | 4 +- arch/x86/include/asm/irq.h | 2 +- arch/x86/kernel/apic/hw_nmi.c | 3 +- drivers/irqchip/irq-gic-v3.c | 29 ++++++++--- include/linux/kgdb.h | 12 +++++ include/linux/nmi.h | 12 ++--- kernel/debug/debug_core.c | 8 ++- 20 files changed, 194 insertions(+), 28 deletions(-) create mode 100644 arch/arm64/include/asm/nmi.h create mode 100644 arch/arm64/kernel/ipi_nmi.c -- 2.34.1

2 8

[PATCH OLK-6.6 v4 0/7] arm64: Add support to turn an IPI as NMI
by Liao Chen 22 Dec '23

22 Dec '23

With pseudo NMIs support available its possible to configure SGIs to be triggered as pseudo NMIs running in NMI context. And kernel features such as: - NMI backtrace can leverage IPI turned as NMI to get a backtrace of CPU stuck in hard lockup using magic SYSRQ. - kgdb relies on NMI support to round up CPUs which are stuck in hard lockup state with interrupts disabled. This patch-set adds framework to turn an IPI as NMI which can be triggered as a pseudo NMI which in turn invokes registered NMI handlers. After this patch-set we should be able to get a backtrace for a CPU stuck in HARDLOCKUP. Sumit Garg (7): arm64: Add framework to turn IPI as NMI irqchip/gic-v3: Enable support for SGIs to act as NMIs arm64: smp: Assign and setup an IPI as NMI nmi: backtrace: Allow runtime arch specific override arm64: ipi_nmi: Add support for NMI backtrace kgdb: Expose default CPUs roundup fallback mechanism arm64: kgdb: Roundup cpus using IPI as NMI arch/arm/include/asm/irq.h | 2 +- arch/arm/kernel/smp.c | 3 +- arch/arm64/include/asm/irq.h | 6 +++ arch/arm64/include/asm/nmi.h | 17 +++++++ arch/arm64/kernel/Makefile | 2 +- arch/arm64/kernel/ipi_nmi.c | 84 ++++++++++++++++++++++++++++++++ arch/arm64/kernel/kgdb.c | 18 +++++++ arch/arm64/kernel/smp.c | 8 +++ arch/mips/include/asm/irq.h | 2 +- arch/mips/kernel/process.c | 3 +- arch/powerpc/include/asm/irq.h | 2 +- arch/powerpc/kernel/stacktrace.c | 3 +- arch/sparc/include/asm/irq_64.h | 2 +- arch/sparc/kernel/process_64.c | 4 +- arch/x86/include/asm/irq.h | 2 +- arch/x86/kernel/apic/hw_nmi.c | 3 +- drivers/irqchip/irq-gic-v3.c | 29 ++++++++--- include/linux/kgdb.h | 12 +++++ include/linux/nmi.h | 12 ++--- kernel/debug/debug_core.c | 8 ++- 20 files changed, 194 insertions(+), 28 deletions(-) create mode 100644 arch/arm64/include/asm/nmi.h create mode 100644 arch/arm64/kernel/ipi_nmi.c -- 2.34.1

1 1

[PATCH OLK-6.6 v4 1/7] arm64: Add framework to turn IPI as NMI
by Liao Chen 22 Dec '23

22 Dec '23

From: Sumit Garg <sumit.garg(a)linaro.org> maillist inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I7R4EN CVE: NA Reference: https://lore.kernel.org/all/1604317487-14543-1-git-send-email-sumit.garg@li… ------------------------------------------------- Introduce framework to turn an IPI as NMI using pseudo NMIs. The main motivation for this feature is to have an IPI that can be leveraged to invoke NMI functions on other CPUs. And current prospective users are NMI backtrace and KGDB CPUs round-up whose support is added via future patches. Signed-off-by: Sumit Garg <sumit.garg(a)linaro.org> Signed-off-by: Wei Li <liwei391(a)huawei.com> Reviewed-by: Xie XiuQi <xiexiuqi(a)huawei.com> Signed-off-by: Zheng Zengkai <zhengzengkai(a)huawei.com> Signed-off-by: Ruan Jinjie <ruanjinjie(a)huawei.com> Signed-off-by: Liao Chen <liaochen4(a)huawei.com> --- arch/arm64/include/asm/nmi.h | 17 ++++++++++ arch/arm64/kernel/ipi_nmi.c | 65 ++++++++++++++++++++++++++++++++++++ 2 files changed, 82 insertions(+) create mode 100644 arch/arm64/include/asm/nmi.h create mode 100644 arch/arm64/kernel/ipi_nmi.c diff --git a/arch/arm64/include/asm/nmi.h b/arch/arm64/include/asm/nmi.h new file mode 100644 index 000000000000..4cd14b6af88b --- /dev/null +++ b/arch/arm64/include/asm/nmi.h @@ -0,0 +1,17 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef __ASM_NMI_H +#define __ASM_NMI_H + +#ifndef __ASSEMBLER__ + +#include <linux/cpumask.h> + +extern bool arm64_supports_nmi(void); +extern void arm64_send_nmi(cpumask_t *mask); + +void set_smp_dynamic_ipi(int ipi); +void dynamic_ipi_setup(int cpu); +void dynamic_ipi_teardown(int cpu); + +#endif /* !__ASSEMBLER__ */ +#endif diff --git a/arch/arm64/kernel/ipi_nmi.c b/arch/arm64/kernel/ipi_nmi.c new file mode 100644 index 000000000000..a945dcf8015f --- /dev/null +++ b/arch/arm64/kernel/ipi_nmi.c @@ -0,0 +1,65 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * NMI support for IPIs + * + * Copyright (C) 2020 Linaro Limited + * Author: Sumit Garg <sumit.garg(a)linaro.org> + */ + +#include <linux/interrupt.h> +#include <linux/irq.h> +#include <linux/smp.h> + +#include <asm/nmi.h> + +static struct irq_desc *ipi_nmi_desc __read_mostly; +static int ipi_nmi_id __read_mostly; + +bool arm64_supports_nmi(void) +{ + if (ipi_nmi_desc) + return true; + + return false; +} + +void arm64_send_nmi(cpumask_t *mask) +{ + if (WARN_ON_ONCE(!ipi_nmi_desc)) + return; + + __ipi_send_mask(ipi_nmi_desc, mask); +} + +static irqreturn_t ipi_nmi_handler(int irq, void *data) +{ + /* nop, NMI handlers for special features can be added here. */ + + return IRQ_NONE; +} + +void dynamic_ipi_setup(int cpu) +{ + if (!ipi_nmi_desc) + return; + + if (!prepare_percpu_nmi(ipi_nmi_id)) + enable_percpu_nmi(ipi_nmi_id, IRQ_TYPE_NONE); +} + +void dynamic_ipi_teardown(int cpu) +{ + if (!ipi_nmi_desc) + return; + + disable_percpu_nmi(ipi_nmi_id); + teardown_percpu_nmi(ipi_nmi_id); +} + +void __init set_smp_dynamic_ipi(int ipi) +{ + if (!request_percpu_nmi(ipi, ipi_nmi_handler, "IPI", &cpu_number)) { + ipi_nmi_desc = irq_to_desc(ipi); + ipi_nmi_id = ipi; + } +} -- 2.34.1

2 7

[PATCH OLK-6.6] memcg: support ksm merge any mode per cgroup
by Nanyong Sun 22 Dec '23

22 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OIQR ---------------------------------------------------------------------- Add control file "memory.ksm" to enable ksm per cgroup. Echo to 1 will set all tasks currently in the cgroup to ksm merge any mode, which means ksm gets enabled for all vma's of a process. Meanwhile echo to 0 will disable ksm for them and unmerge the merged pages. Cat the file will show the above state and ksm related profits of this cgroup. Signed-off-by: Nanyong Sun <sunnanyong(a)huawei.com> --- .../admin-guide/cgroup-v1/memory.rst | 1 + mm/memcontrol.c | 110 +++++++++++++++++- 2 files changed, 109 insertions(+), 2 deletions(-) diff --git a/Documentation/admin-guide/cgroup-v1/memory.rst b/Documentation/admin-guide/cgroup-v1/memory.rst index ff456871bf4b..3fdb48435e8e 100644 --- a/Documentation/admin-guide/cgroup-v1/memory.rst +++ b/Documentation/admin-guide/cgroup-v1/memory.rst @@ -109,6 +109,7 @@ Brief summary of control files. memory.kmem.tcp.failcnt show the number of tcp buf memory usage hits limits memory.kmem.tcp.max_usage_in_bytes show max tcp buf memory usage recorded + memory.ksm set/show ksm merge any mode ==================================== ========================================== 1. History diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 2d9a873e5522..30cafc2f22c1 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -73,6 +73,7 @@ #include <linux/uaccess.h> #include <trace/events/vmscan.h> +#include <linux/ksm.h> struct cgroup_subsys memory_cgrp_subsys __read_mostly; EXPORT_SYMBOL(memory_cgrp_subsys); @@ -242,10 +243,15 @@ enum res_type { iter != NULL; \ iter = mem_cgroup_iter(NULL, iter, NULL)) +static inline bool __task_is_dying(struct task_struct *task) +{ + return tsk_is_oom_victim(task) || fatal_signal_pending(task) || + (task->flags & PF_EXITING); +} + static inline bool task_is_dying(void) { - return tsk_is_oom_victim(current) || fatal_signal_pending(current) || - (current->flags & PF_EXITING); + return __task_is_dying(current); } /* Some nice accessors for the vmpressure. */ @@ -5100,6 +5106,98 @@ static int mem_cgroup_slab_show(struct seq_file *m, void *p) } #endif +#ifdef CONFIG_KSM +static int memcg_set_ksm_for_tasks(struct mem_cgroup *memcg, bool enable) +{ + struct task_struct *task; + struct mm_struct *mm; + struct css_task_iter it; + int ret = 0; + + css_task_iter_start(&memcg->css, CSS_TASK_ITER_PROCS, &it); + while (!ret && (task = css_task_iter_next(&it))) { + if (__task_is_dying(task)) + continue; + + mm = get_task_mm(task); + if (!mm) + continue; + + if (mmap_write_lock_killable(mm)) { + mmput(mm); + continue; + } + + if (enable) + ret = ksm_enable_merge_any(mm); + else + ret = ksm_disable_merge_any(mm); + + mmap_write_unlock(mm); + mmput(mm); + } + css_task_iter_end(&it); + + return ret; +} + +static int memory_ksm_show(struct seq_file *m, void *v) +{ + unsigned long ksm_merging_pages = 0; + unsigned long ksm_rmap_items = 0; + long ksm_process_profits = 0; + unsigned int tasks = 0; + struct task_struct *task; + struct mm_struct *mm; + struct css_task_iter it; + struct mem_cgroup *memcg = mem_cgroup_from_seq(m); + + css_task_iter_start(&memcg->css, CSS_TASK_ITER_PROCS, &it); + while ((task = css_task_iter_next(&it))) { + mm = get_task_mm(task); + if (!mm) + continue; + + if (test_bit(MMF_VM_MERGE_ANY, &mm->flags)) + tasks++; + + ksm_rmap_items += mm->ksm_rmap_items; + ksm_merging_pages += mm->ksm_merging_pages; + ksm_process_profits += ksm_process_profit(mm); + mmput(mm); + } + css_task_iter_end(&it); + + seq_printf(m, "merge any tasks: %u\n", tasks); + seq_printf(m, "ksm_rmap_items %lu\n", ksm_rmap_items); + seq_printf(m, "ksm_merging_pages %lu\n", ksm_merging_pages); + seq_printf(m, "ksm_process_profits %ld\n", ksm_process_profits); + return 0; +} + +static ssize_t memory_ksm_write(struct kernfs_open_file *of, char *buf, + size_t nbytes, loff_t off) +{ + bool enable; + int err; + struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); + + buf = strstrip(buf); + if (!buf) + return -EINVAL; + + err = kstrtobool(buf, &enable); + if (err) + return err; + + err = memcg_set_ksm_for_tasks(memcg, enable); + if (err) + return err; + + return nbytes; +} +#endif /* CONFIG_KSM */ + static int memory_stat_show(struct seq_file *m, void *v); #ifdef CONFIG_MEMCG_V1_RECLAIM @@ -5337,6 +5435,14 @@ static struct cftype mem_cgroup_legacy_files[] = { .name = "reclaim", .write = memory_reclaim, }, +#endif +#ifdef CONFIG_KSM + { + .name = "ksm", + .flags = CFTYPE_NOT_ON_ROOT, + .write = memory_ksm_write, + .seq_show = memory_ksm_show, + }, #endif { }, /* terminate */ }; -- 2.25.1

2 1

[PATCH OLK-6.6 0/3] Print rootfs and tmpfs files charged by memcg
by Jinjiang Tu 22 Dec '23

22 Dec '23

Support to print rootfs files and tmpfs files that having pages charged in given memory cgroup. The files infomations can be printed through interface "memory.memfs_files_info" or printed when OOM is triggered. Jinjiang Tu (1): fs: move {lock, unlock}_mount_hash to fs/mount.h Liu Shixin (2): mm/memcg_memfs_info: show files that having pages charged in mem_cgroup config: enable CONFIG_MEMCG_MEMFS_INFO by default Documentation/vm/memcg_memfs_info.rst | 40 ++++ arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/mount.h | 10 + fs/namespace.c | 10 - include/linux/memcg_memfs_info.h | 23 ++ init/Kconfig | 10 + mm/Makefile | 1 + mm/memcg_memfs_info.c | 318 +++++++++++++++++++++++++ mm/memcontrol.c | 12 + 10 files changed, 416 insertions(+), 10 deletions(-) create mode 100644 Documentation/vm/memcg_memfs_info.rst create mode 100644 include/linux/memcg_memfs_info.h create mode 100644 mm/memcg_memfs_info.c -- 2.25.1

2 4

[PATCH OLK-6.6 v4 0/7] arm64: Add support to turn an IPI as NMI
by Liao Chen 22 Dec '23

22 Dec '23

With pseudo NMIs support available its possible to configure SGIs to be triggered as pseudo NMIs running in NMI context. And kernel features such as: - NMI backtrace can leverage IPI turned as NMI to get a backtrace of CPU stuck in hard lockup using magic SYSRQ. - kgdb relies on NMI support to round up CPUs which are stuck in hard lockup state with interrupts disabled. This patch-set adds framework to turn an IPI as NMI which can be triggered as a pseudo NMI which in turn invokes registered NMI handlers. After this patch-set we should be able to get a backtrace for a CPU stuck in HARDLOCKUP. Sumit Garg (7): arm64: Add framework to turn IPI as NMI irqchip/gic-v3: Enable support for SGIs to act as NMIs arm64: smp: Assign and setup an IPI as NMI nmi: backtrace: Allow runtime arch specific override arm64: ipi_nmi: Add support for NMI backtrace kgdb: Expose default CPUs roundup fallback mechanism arm64: kgdb: Roundup cpus using IPI as NMI arch/arm/include/asm/irq.h | 2 +- arch/arm/kernel/smp.c | 3 +- arch/arm64/include/asm/irq.h | 6 +++ arch/arm64/include/asm/nmi.h | 17 +++++++ arch/arm64/kernel/Makefile | 2 +- arch/arm64/kernel/ipi_nmi.c | 84 ++++++++++++++++++++++++++++++++ arch/arm64/kernel/kgdb.c | 18 +++++++ arch/arm64/kernel/smp.c | 8 +++ arch/mips/include/asm/irq.h | 2 +- arch/mips/kernel/process.c | 3 +- arch/powerpc/include/asm/irq.h | 2 +- arch/powerpc/kernel/stacktrace.c | 3 +- arch/sparc/include/asm/irq_64.h | 2 +- arch/sparc/kernel/process_64.c | 4 +- arch/x86/include/asm/irq.h | 2 +- arch/x86/kernel/apic/hw_nmi.c | 3 +- drivers/irqchip/irq-gic-v3.c | 29 ++++++++--- include/linux/kgdb.h | 12 +++++ include/linux/nmi.h | 12 ++--- kernel/debug/debug_core.c | 8 ++- 20 files changed, 194 insertions(+), 28 deletions(-) create mode 100644 arch/arm64/include/asm/nmi.h create mode 100644 arch/arm64/kernel/ipi_nmi.c -- 2.34.1

2 8

[Cancel] openEuler Kernel SIG双周例会
by openEuler conference 22 Dec '23

22 Dec '23

Sorry! The WeLink meeting will be held at 2023-12-22 14:00 scheduled by openEuler Kernel SIG has been cancelled.

1 0

[Cancel] openEuler Kernel SIG双周例会
by openEuler conference 22 Dec '23

22 Dec '23

Sorry! The Zoom meeting will be held at 2022-11-04 14:00 scheduled by openEuler Kernel SIG has been cancelled.

1 0

openEuler Kernel SIG双周例会
by openEuler conference 22 Dec '23

22 Dec '23

您好！ Kernel SIG 邀请您参加 2023-12-22 14:00 召开的WeLink会议(自动录制) 会议主题：openEuler Kernel SIG双周例会会议内容： 1. 进展update 2. 议题征集中（新增议题可回复本邮件申报，也可直接填至会议看板）会议链接：https://bmeeting.huaweicloud.com:36443/#/j/966610402 会议纪要：https://etherpad.openeuler.org/p/Kernel-meetings 温馨提醒：建议接入会议后修改参会人的姓名，也可以使用您在gitee.com的ID 更多资讯尽在：https://openeuler.org/zh/ Hello! openEuler Kernel SIG invites you to attend the WeLink conference(auto recording) will be held at 2023-12-22 14:00, The subject of the conference is openEuler Kernel SIG双周例会, Summary: 1. 进展update 2. 议题征集中（新增议题可回复本邮件申报，也可直接填至会议看板） You can join the meeting at https://bmeeting.huaweicloud.com:36443/#/j/966610402. Add topics at https://etherpad.openeuler.org/p/Kernel-meetings. Note: You are advised to change the participant name after joining the conference or use your ID at gitee.com. More information: https://openeuler.org/en/

2 1

[PATCH OLK-6.6 v3 0/7] arm64: Add support to turn an IPI as NMI
by Liao Chen 22 Dec '23

22 Dec '23

With pseudo NMIs support available its possible to configure SGIs to be triggered as pseudo NMIs running in NMI context. And kernel features such as: - NMI backtrace can leverage IPI turned as NMI to get a backtrace of CPU stuck in hard lockup using magic SYSRQ. - kgdb relies on NMI support to round up CPUs which are stuck in hard lockup state with interrupts disabled. This patch-set adds framework to turn an IPI as NMI which can be triggered as a pseudo NMI which in turn invokes registered NMI handlers. After this patch-set we should be able to get a backtrace for a CPU stuck in HARDLOCKUP. Sumit Garg (7): arm64: Add framework to turn IPI as NMI irqchip/gic-v3: Enable support for SGIs to act as NMIs arm64: smp: Assign and setup an IPI as NMI nmi: backtrace: Allow runtime arch specific override arm64: ipi_nmi: Add support for NMI backtrace kgdb: Expose default CPUs roundup fallback mechanism arm64: kgdb: Roundup cpus using IPI as NMI arch/arm/include/asm/irq.h | 2 +- arch/arm/kernel/smp.c | 3 +- arch/arm64/include/asm/irq.h | 6 +++ arch/arm64/include/asm/nmi.h | 17 +++++++ arch/arm64/kernel/Makefile | 2 +- arch/arm64/kernel/ipi_nmi.c | 84 ++++++++++++++++++++++++++++++++ arch/arm64/kernel/kgdb.c | 18 +++++++ arch/arm64/kernel/smp.c | 8 +++ arch/mips/include/asm/irq.h | 2 +- arch/mips/kernel/process.c | 3 +- arch/powerpc/include/asm/irq.h | 2 +- arch/powerpc/kernel/stacktrace.c | 3 +- arch/sparc/include/asm/irq_64.h | 2 +- arch/sparc/kernel/process_64.c | 4 +- arch/x86/include/asm/irq.h | 2 +- arch/x86/kernel/apic/hw_nmi.c | 3 +- drivers/irqchip/irq-gic-v3.c | 29 ++++++++--- include/linux/kgdb.h | 12 +++++ include/linux/nmi.h | 12 ++--- kernel/debug/debug_core.c | 8 ++- 20 files changed, 194 insertions(+), 28 deletions(-) create mode 100644 arch/arm64/include/asm/nmi.h create mode 100644 arch/arm64/kernel/ipi_nmi.c -- 2.34.1

2 8

[PATCH v2 openEuler-1.0-LTS] umh: fix memory leak on execve failure
by Wenyu Huang 22 Dec '23

22 Dec '23

From: Vincent Minet <v.minet(a)criteo.com> mainline inclusion from mainline-v5.10-rc1 commit db803036ada7d61d096783726f9771b3fc540370 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8J9MA CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… ------------------------------------------------- If a UMH process created by fork_usermode_blob() fails to execute, a pair of struct file allocated by umh_pipe_setup() will leak. Under normal conditions, the caller (like bpfilter) needs to manage the lifetime of the UMH and its two pipes. But when fork_usermode_blob() fails, the caller doesn't really have a way to know what needs to be done. It seems better to do the cleanup ourselves in this case. Fixes: 449325b52b7a ("umh: introduce fork_usermode_blob() helper") Signed-off-by: Vincent Minet <v.minet(a)criteo.com> Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: Wenyu Huang <huangwenyu5(a)huawei.com> --- kernel/umh.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/kernel/umh.c b/kernel/umh.c index 53611efb10cb..715b4368f9ed 100644 --- a/kernel/umh.c +++ b/kernel/umh.c @@ -495,6 +495,12 @@ static void umh_clean_and_save_pid(struct subprocess_info *info) { struct umh_info *umh_info = info->data; + /* cleanup if umh_pipe_setup() was successful but exec failed */ + if (info->pid && info->retval) { + fput(umh_info->pipe_to_umh); + fput(umh_info->pipe_from_umh); + } + argv_free(info->argv); umh_info->pid = info->pid; } -- 2.34.1

2 1

[PATCH OLK-6.6 0/3] files cgroups feature
by chenridong 22 Dec '23

22 Dec '23

Binder Makin (1): cgroups: Resource controller for open files Hou Tao (1): cgroup/files: fix bugs as mentioned Lu Jialin (1): enable CONFIG_CGROUP_FILES in openeuler_defconfig for x86 and arm64 .../admin-guide/kernel-parameters.txt | 7 +- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Makefile | 1 + fs/file.c | 68 +++- fs/filescontrol.c | 312 ++++++++++++++++++ include/linux/cgroup-defs.h | 8 +- include/linux/cgroup_subsys.h | 4 + include/linux/fdtable.h | 1 + include/linux/filescontrol.h | 40 +++ init/Kconfig | 10 + 11 files changed, 445 insertions(+), 8 deletions(-) create mode 100644 fs/filescontrol.c create mode 100644 include/linux/filescontrol.h -- 2.34.1

2 4

[PATCH OLK-6.6 v2 0/7] arm64: Add support to turn an IPI as NMI
by Liao Chen 22 Dec '23

22 Dec '23

With pseudo NMIs support available its possible to configure SGIs to be triggered as pseudo NMIs running in NMI context. And kernel features such as: - NMI backtrace can leverage IPI turned as NMI to get a backtrace of CPU stuck in hard lockup using magic SYSRQ. - kgdb relies on NMI support to round up CPUs which are stuck in hard lockup state with interrupts disabled. This patch-set adds framework to turn an IPI as NMI which can be triggered as a pseudo NMI which in turn invokes registered NMI handlers. After this patch-set we should be able to get a backtrace for a CPU stuck in HARDLOCKUP. Sumit Garg (7): arm64: Add framework to turn IPI as NMI irqchip/gic-v3: Enable support for SGIs to act as NMIs arm64: smp: Assign and setup an IPI as NMI nmi: backtrace: Allow runtime arch specific override arm64: ipi_nmi: Add support for NMI backtrace kgdb: Expose default CPUs roundup fallback mechanism arm64: kgdb: Roundup cpus using IPI as NMI arch/arm/include/asm/irq.h | 2 +- arch/arm/kernel/smp.c | 3 +- arch/arm64/include/asm/irq.h | 6 +++ arch/arm64/include/asm/nmi.h | 17 +++++++ arch/arm64/kernel/Makefile | 2 +- arch/arm64/kernel/ipi_nmi.c | 84 ++++++++++++++++++++++++++++++++ arch/arm64/kernel/kgdb.c | 18 +++++++ arch/arm64/kernel/smp.c | 8 +++ arch/mips/include/asm/irq.h | 2 +- arch/mips/kernel/process.c | 3 +- arch/powerpc/include/asm/irq.h | 2 +- arch/powerpc/kernel/stacktrace.c | 3 +- arch/sparc/include/asm/irq_64.h | 2 +- arch/sparc/kernel/process_64.c | 4 +- arch/x86/include/asm/irq.h | 2 +- arch/x86/kernel/apic/hw_nmi.c | 3 +- drivers/irqchip/irq-gic-v3.c | 29 ++++++++--- include/linux/kgdb.h | 12 +++++ include/linux/nmi.h | 12 ++--- kernel/debug/debug_core.c | 8 ++- 20 files changed, 194 insertions(+), 28 deletions(-) create mode 100644 arch/arm64/include/asm/nmi.h create mode 100644 arch/arm64/kernel/ipi_nmi.c -- 2.34.1

2 8

[PATCH openEuler-22.03-LTS-SP2] tick/broadcast-hrtimer: Prevent the timer device on broadcast duty CPU from being disabled
by Yu Liao 22 Dec '23

22 Dec '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8PL17 CVE: NA ---------------------------------------- It was found that running the LTP hotplug stress test on a aarch64 system could produce rcu_sched stall warnings. The issue is the following: CPU1 (owns the broadcast hrtimer) CPU2 tick_broadcast_enter() //shut down local timer device ... tick_broadcast_exit() //exits with tick_broadcast_force_mask set, timer device remains disabled initiates offlining of CPU1 take_cpu_down() //CPU1 shuts down and does not send broadcast IPI anymore takedown_cpu() hotplug_cpu__broadcast_tick_pull() //move broadcast hrtimer to this CPU clockevents_program_event() bc_set_next() hrtimer_start() //does not call hrtimer_reprogram() to program timer device if expires equals dev->next_event, so the timer device remains disabled. CPU2 takes over the broadcast duty but local timer device is disabled, causing many CPUs to become stuck. Fix this by calling tick_program_event() to reprogram the local timer device in this scenario. Signed-off-by: Yu Liao <liaoyu15(a)huawei.com> --- kernel/time/tick-broadcast-hrtimer.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/kernel/time/tick-broadcast-hrtimer.c b/kernel/time/tick-broadcast-hrtimer.c index b5a65e212df2..b8e79f8012f3 100644 --- a/kernel/time/tick-broadcast-hrtimer.c +++ b/kernel/time/tick-broadcast-hrtimer.c @@ -42,6 +42,8 @@ static int bc_shutdown(struct clock_event_device *evt) */ static int bc_set_next(ktime_t expires, struct clock_event_device *bc) { + struct clock_event_device *dev = __this_cpu_read(tick_cpu_device.evtdev); + /* * This is called either from enter/exit idle code or from the * broadcast handler. In all cases tick_broadcast_lock is held. @@ -62,6 +64,18 @@ static int bc_set_next(ktime_t expires, struct clock_event_device *bc) * hrtimer_start() can call into tracing. */ RCU_NONIDLE( { + + /* + * This can be called from CPU offline operation to move broadcast + * assignment. If tick_broadcast_force_mask is set, the CPU local + * timer device may be disabled. And hrtimer_reprogram() will not + * called if the timer is not the first expiring timer. Reprogram + * the cpu local timer device to ensure we can take over the + * broadcast duty. + */ + if (tick_check_broadcast_expired() && expires >= dev->next_event) + clockevents_program_event(dev, dev->next_event, 1); + hrtimer_start(&bctimer, expires, HRTIMER_MODE_ABS_PINNED_HARD); /* * The core tick broadcast mode expects bc->bound_on to be set -- 2.33.0

2 1

[PATCH OLK-6.6 0/7] Support Features to Turn an IPI as NMI
by Liao Chen 22 Dec '23

22 Dec '23

With pseudo NMIs support available its possible to configure SGIs to be triggered as pseudo NMIs running in NMI context. And kernel features such as: - NMI backtrace can leverage IPI turned as NMI to get a backtrace of CPU stuck in hard lockup using magic SYSRQ. - kgdb relies on NMI support to round up CPUs which are stuck in hard lockup state with interrupts disabled. This patch-set adds framework to turn an IPI as NMI which can be triggered as a pseudo NMI which in turn invokes registered NMI handlers. After this patch-set we should be able to get a backtrace for a CPU stuck in HARDLOCKUP. Sumit Garg (7): arm64: Add framework to turn IPI as NMI irqchip/gic-v3: Enable support for SGIs to act as NMIs arm64: smp: Assign and setup an IPI as NMI nmi: backtrace: Allow runtime arch specific override arm64: ipi_nmi: Add support for NMI backtrace kgdb: Expose default CPUs roundup fallback mechanism arm64: kgdb: Roundup cpus using IPI as NMI arch/arm/include/asm/irq.h | 2 +- arch/arm/kernel/smp.c | 3 +- arch/arm64/include/asm/irq.h | 6 +++ arch/arm64/include/asm/nmi.h | 17 +++++++ arch/arm64/kernel/Makefile | 4 +- arch/arm64/kernel/ipi_nmi.c | 84 ++++++++++++++++++++++++++++++++ arch/arm64/kernel/kgdb.c | 18 +++++++ arch/arm64/kernel/smp.c | 8 +++ arch/mips/include/asm/irq.h | 2 +- arch/powerpc/kernel/stacktrace.c | 3 +- arch/sparc/include/asm/irq_64.h | 2 +- arch/sparc/kernel/process_64.c | 4 +- arch/x86/include/asm/irq.h | 2 +- arch/x86/kernel/apic/hw_nmi.c | 3 +- drivers/irqchip/irq-gic-v3.c | 29 ++++++++--- include/linux/kgdb.h | 12 +++++ include/linux/nmi.h | 12 ++--- kernel/debug/debug_core.c | 8 ++- 18 files changed, 192 insertions(+), 27 deletions(-) create mode 100644 arch/arm64/include/asm/nmi.h create mode 100644 arch/arm64/kernel/ipi_nmi.c -- 2.34.1

2 8

[PATCH OLK-6.6 v2 0/2] optimize inline
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Guo Xuenan (2): Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely" make OPTIMIZE_INLINING config editable arch/arm64/kvm/sys_regs.h | 5 +++++ arch/x86/configs/i386_defconfig | 1 + arch/x86/configs/x86_64_defconfig | 1 + drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h | 5 +++++ .../pci/hive_isp_css_include/print_support.h | 4 ++++ include/linux/compiler_types.h | 8 ++++++++ include/trace/trace_events.h | 2 +- kernel/configs/tiny.config | 1 + lib/Kconfig.debug | 13 +++++++++++++ 9 files changed, 39 insertions(+), 1 deletion(-) -- 2.34.1

2 3

[PATCH OLK-6.6 v2 0/2] optimize inline
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Guo Xuenan (2): Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely" make OPTIMIZE_INLINING config editable arch/arm64/kvm/sys_regs.h | 5 +++++ arch/x86/configs/i386_defconfig | 1 + arch/x86/configs/x86_64_defconfig | 1 + drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h | 5 +++++ .../pci/hive_isp_css_include/print_support.h | 4 ++++ include/linux/compiler_types.h | 8 ++++++++ include/trace/trace_events.h | 2 +- kernel/configs/tiny.config | 1 + lib/Kconfig.debug | 13 +++++++++++++ 9 files changed, 39 insertions(+), 1 deletion(-) -- 2.34.1

2 3

[PATCH OLK-5.10] tick/broadcast-hrtimer: Prevent the timer device on broadcast duty CPU from being disabled
by Yu Liao 21 Dec '23

21 Dec '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8PL17 CVE: NA ---------------------------------------- It was found that running the LTP hotplug stress test on a aarch64 system could produce rcu_sched stall warnings. The issue is the following: CPU1 (owns the broadcast hrtimer) CPU2 tick_broadcast_enter() //shut down local timer device ... tick_broadcast_exit() //exits with tick_broadcast_force_mask set, timer device remains disabled initiates offlining of CPU1 take_cpu_down() //CPU1 shuts down and does not send broadcast IPI anymore takedown_cpu() hotplug_cpu__broadcast_tick_pull() //move broadcast hrtimer to this CPU clockevents_program_event() bc_set_next() hrtimer_start() //does not call hrtimer_reprogram() to program timer device if expires equals dev->next_event, so the timer device remains disabled. CPU2 takes over the broadcast duty but local timer device is disabled, causing many CPUs to become stuck. Fix this by calling tick_program_event() to reprogram the local timer device in this scenario. Signed-off-by: Yu Liao <liaoyu15(a)huawei.com> --- kernel/time/tick-broadcast-hrtimer.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/kernel/time/tick-broadcast-hrtimer.c b/kernel/time/tick-broadcast-hrtimer.c index b5a65e212df2..b8e79f8012f3 100644 --- a/kernel/time/tick-broadcast-hrtimer.c +++ b/kernel/time/tick-broadcast-hrtimer.c @@ -42,6 +42,8 @@ static int bc_shutdown(struct clock_event_device *evt) */ static int bc_set_next(ktime_t expires, struct clock_event_device *bc) { + struct clock_event_device *dev = __this_cpu_read(tick_cpu_device.evtdev); + /* * This is called either from enter/exit idle code or from the * broadcast handler. In all cases tick_broadcast_lock is held. @@ -62,6 +64,18 @@ static int bc_set_next(ktime_t expires, struct clock_event_device *bc) * hrtimer_start() can call into tracing. */ RCU_NONIDLE( { + + /* + * This can be called from CPU offline operation to move broadcast + * assignment. If tick_broadcast_force_mask is set, the CPU local + * timer device may be disabled. And hrtimer_reprogram() will not + * called if the timer is not the first expiring timer. Reprogram + * the cpu local timer device to ensure we can take over the + * broadcast duty. + */ + if (tick_check_broadcast_expired() && expires >= dev->next_event) + clockevents_program_event(dev, dev->next_event, 1); + hrtimer_start(&bctimer, expires, HRTIMER_MODE_ABS_PINNED_HARD); /* * The core tick broadcast mode expects bc->bound_on to be set -- 2.33.0

2 1

[PATCH OLK-6.6] timekeeping: Avoiding false sharing in field access of tk_core
by Yu Liao 21 Dec '23

21 Dec '23

From: Wang ShaoBo <bobo.shaobowang(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I47W8L CVE: NA --------------------------- We detect a performance deterioration when using Unixbench, we use the dichotomy to locate the patch 7e66740ad725 ("MPAM / ACPI: Refactoring MPAM init process and set MPAM ACPI as entrance"), In comparing two commit df5defd901ff ("KVM: X86: MMU: Use the correct inherited permissions to get shadow page") and ac4dbb7554ef ("ACPI 6.x: Add definitions for MPAM table") we get following testing result: CMD: ./Run -c xx context1 RESULT: +-------------UnixBench context1-----------+ +---------+--------------+-----------------+ + + ac4dbb7554ef + df5defd901ff + +---------+--------------+---------+-------+ + Cores + Score + Score + +---------+--------------+-----------------+ + 1 + 522.8 + 535.7 + +---------+--------------+-----------------+ + 24 + 11231.5 + 12111.2 + +---------+--------------+-----------------+ + 48 + 8535.1 + 8745.1 + +---------+--------------+-----------------+ + 72 + 10821.9 + 10343.8 + +---------+--------------+-----------------+ + 96 + 15238.5 + 42947.8 + +---------+--------------+-----------------+ We found a irrefutable difference in latency sampling when using the perf tool: HEAD:ac4dbb7554ef HEAD:df5defd901ff 45.18% [kernel] [k] ktime_get_coarse_real_ts64 -> 1.78% [kernel] [k] ktime_get_coarse_real_ts64 ... 65.87 │ dmb ishld //smp_rmb() Through ftrace we get the calltrace and and detected the number of visits of ktime_get_coarse_real_ts64, which frequently visits tk_core->seq and tk_core->timekeeper->tkr_mono: - 48.86% [kernel] [k] ktime_get_coarse_real_ts64 - 5.76% ktime_get_coarse_real_ts64 #about 111437657 times per 10 seconds - 14.70% __audit_syscall_entry syscall_trace_enter el0_svc_common el0_svc_handler + el0_svc - 2.85% current_time So this may be performance degradation caused by interference when happened different fields access, We compare .bss and .data section of this two version: HEAD:ac4dbb7554ef `-> ffff00000962e680 l O .bss 0000000000000110 tk_core ffff000009355680 l O .data 0000000000000078 tk_fast_mono ffff0000093557a0 l O .data 0000000000000090 dummy_clock ffff000009355700 l O .data 0000000000000078 tk_fast_raw ffff000009355778 l O .data 0000000000000028 timekeeping_syscore_ops ffff00000962e640 l O .bss 0000000000000008 cycles_at_suspend HEAD:df5defd901ff `-> ffff00000957dbc0 l O .bss 0000000000000110 tk_core ffff0000092b4e80 l O .data 0000000000000078 tk_fast_mono ffff0000092b4fa0 l O .data 0000000000000090 dummy_clock ffff0000092b4f00 l O .data 0000000000000078 tk_fast_raw ffff0000092b4f78 l O .data 0000000000000028 timekeeping_syscore_ops ffff00000957db80 l O .bss 0000000000000008 cycles_at_suspend By comparing this two version tk_core's address: ffff00000962e680 is 128Byte aligned but latter df5defd901ff is 64Byte aligned, the memory storage layout of tk_core has undergone subtle changes: HEAD:ac4dbb7554ef `-> |<--------formmer 64Bytes---------->|<------------latter 64Byte------------->| 0xffff00000957dbc0_>|<-seq 8Bytes->|<-tkr_mono 56Bytes->|<-thr_raw 56Bytes->|<-xtime_sec 8Bytes->| 0xffff00000957dc00_>... HEAD:df5defd901ff `-> |<------formmer 64Bytes---->|<------------latter 64Byte-------->| 0xffff00000962e680_>|<-Other variables 64Bytes->|<-seq 8Bytes->|<-tkr_mono 56Bytes->| 0xffff00000962e6c0_>.. We testified thr_raw,xtime_sec fields interfere strongly with seq,tkr_mono field because of frequent load/store operation, this will cause as known false sharing. We add a 64Bytes padding field in tk_core for reservation of any after usefull usage and keep tk_core 128Byte aligned, this can avoid changes in the way tk_core's layout is stored, In this solution, layout of tk_core always like this: crash> struct -o tk_core_t struct tk_core_t { [0] u64 padding[8]; [64] seqcount_t seq; [72] struct timekeeper timekeeper; } SIZE: 336 crash> struct -o timekeeper struct timekeeper { [0] struct tk_read_base tkr_mono; [56] struct tk_read_base tkr_raw; [112] u64 xtime_sec; [120] unsigned long ktime_sec; ... } SIZE: 264 After appling our own solution: +---------+--------------+ + + Our solution + +---------+--------------+ + Cores + Score + +---------+--------------+ + 1 + 548.9 + +---------+--------------+ + 24 + 11018.3 + +---------+--------------+ + 48 + 8938.2 + +---------+--------------+ + 72 + 14610.7 + +---------+--------------+ + 96 + 40811.7 + +---------+--------------+ Signed-off-by: Wang ShaoBo <bobo.shaobowang(a)huawei.com> Signed-off-by: Yu Liao <liaoyu15(a)huawei.com> --- arch/arm64/Kconfig | 9 +++++++++ arch/arm64/include/asm/cache.h | 6 ++++++ kernel/time/timekeeping.c | 7 +++++++ 3 files changed, 22 insertions(+) diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 2aca373a7038..f5559e38b243 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -1493,6 +1493,15 @@ config HW_PERF_EVENTS def_bool y depends on ARM_PMU +config ARCH_LLC_128_LINE_SIZE + bool "Force 128 bytes alignment for fitting LLC cacheline" + depends on ARM64 + default y + help + As specific machine's LLC cacheline size may be up to + 128 bytes, gaining performance improvement from fitting + 128 Bytes LLC cache aligned. + # Supported by clang >= 7.0 or GCC >= 12.0.0 config CC_HAVE_SHADOW_CALL_STACK def_bool $(cc-option, -fsanitize=shadow-call-stack -ffixed-x18) diff --git a/arch/arm64/include/asm/cache.h b/arch/arm64/include/asm/cache.h index 1613779be63a..8455df351ef8 100644 --- a/arch/arm64/include/asm/cache.h +++ b/arch/arm64/include/asm/cache.h @@ -8,6 +8,12 @@ #define L1_CACHE_SHIFT (6) #define L1_CACHE_BYTES (1 << L1_CACHE_SHIFT) +#ifdef CONFIG_ARCH_LLC_128_LINE_SIZE +#ifndef ____cacheline_aligned_128 +#define ____cacheline_aligned_128 __attribute__((__aligned__(128))) +#endif +#endif + #define CLIDR_LOUU_SHIFT 27 #define CLIDR_LOC_SHIFT 24 #define CLIDR_LOUIS_SHIFT 21 diff --git a/kernel/time/timekeeping.c b/kernel/time/timekeeping.c index 266d02809dbb..35b366addf07 100644 --- a/kernel/time/timekeeping.c +++ b/kernel/time/timekeeping.c @@ -48,9 +48,16 @@ DEFINE_RAW_SPINLOCK(timekeeper_lock); * cache line. */ static struct { +#ifdef CONFIG_ARCH_LLC_128_LINE_SIZE + u64 padding[8]; +#endif seqcount_raw_spinlock_t seq; struct timekeeper timekeeper; +#ifdef CONFIG_ARCH_LLC_128_LINE_SIZE +} tk_core ____cacheline_aligned_128 = { +#else } tk_core ____cacheline_aligned = { +#endif .seq = SEQCNT_RAW_SPINLOCK_ZERO(tk_core.seq, &timekeeper_lock), }; -- 2.33.0

2 1

[PATCH OLK-6.6 0/2] support CLOCKSOURCE_VALIDATE_LAST_CYCLE on
by Yu Liao 21 Dec '23

21 Dec '23

Yu Liao (2): timekeeping: Make CLOCKSOURCE_VALIDATE_LAST_CYCLE configurable config: make CLOCKSOURCE_VALIDATE_LAST_CYCLE not set by default arch/arm64/configs/openeuler_defconfig | 1 + kernel/time/Kconfig | 13 ++++++++----- 2 files changed, 9 insertions(+), 5 deletions(-) -- 2.33.0

2 3

[PATCH OLK-6.6 v4 0/1] stop_machine: mask pseudo nmi before running the callback
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Wei Li (1): stop_machine: mask pseudo nmi before running the callback arch/arm64/include/asm/arch_gicv3.h | 12 ++++++++++++ kernel/stop_machine.c | 3 +++ 2 files changed, 15 insertions(+) -- 2.34.1

2 2

[PATCH OLK-6.6 0/3] Introduce CPU inspect feature
by Yu Liao 21 Dec '23

21 Dec '23

This patches series introduce CPU-inspect feature. CPU-inspect is designed to provide a framework for early detection of SDC by proactively executing CPU inspection test cases. Silent Data Corruption (SDC), sometimes referred to as Silent Data Error (SDE), is an industry-wide issue impacting not only long-protected memory, storage, and networking, but also computer CPUs. As with software issues, hardware-induced SDC can contribute to data loss and corruption. An SDC occurs when an impacted CPU inadvertently causes errors in the data it processes. For example, an impacted CPU might miscalculate data (i.e., 1+1=3). There may be no indication of these computational errors unless the software systematically checks for errors [1]. SDC issues have been around for many years, but as chips have become more advanced and compact in size, the transistors and lines have become so tiny that small electrical fluctuations can cause errors. Most of these errors are caused by defects during manufacturing and are screened out by the vendors; others are caught by hardware error detection or correction. However, some errors go undetected by hardware; therefore only detection software can protect against such errors [1]. [1] https://support.google.com/cloud/answer/10759085 Yu Liao (3): cpuinspect: add CPU-inspect infrastructure cpuinspect: add ATF inspector openeuler_defconfig: enable CPU inspect for arm64 by default arch/arm64/configs/openeuler_defconfig | 7 + drivers/Kconfig | 2 + drivers/Makefile | 1 + drivers/cpuinspect/Kconfig | 24 +++ drivers/cpuinspect/Makefile | 7 + drivers/cpuinspect/cpuinspect.c | 170 ++++++++++++++++ drivers/cpuinspect/cpuinspect.h | 46 +++++ drivers/cpuinspect/inspector-atf.c | 81 ++++++++ drivers/cpuinspect/inspector.c | 124 ++++++++++++ drivers/cpuinspect/sysfs.c | 258 +++++++++++++++++++++++++ include/linux/cpuinspect.h | 40 ++++ 11 files changed, 760 insertions(+) create mode 100644 drivers/cpuinspect/Kconfig create mode 100644 drivers/cpuinspect/Makefile create mode 100644 drivers/cpuinspect/cpuinspect.c create mode 100644 drivers/cpuinspect/cpuinspect.h create mode 100644 drivers/cpuinspect/inspector-atf.c create mode 100644 drivers/cpuinspect/inspector.c create mode 100644 drivers/cpuinspect/sysfs.c create mode 100644 include/linux/cpuinspect.h -- 2.33.0

2 4

[PATCH OLK-6.6 0/2] Introduce clear freelist feature
by Yu Liao 21 Dec '23

21 Dec '23

Yu Liao (2): mm: Add sysctl to clear free list pages config: make CONFIG_CLEAR_FREELIST_PAGE not set by default .../admin-guide/kernel-parameters.txt | 3 + Documentation/admin-guide/sysctl/vm.rst | 13 ++ arch/arm64/configs/openeuler_defconfig | 1 + mm/Kconfig | 12 ++ mm/Makefile | 2 + mm/clear_freelist_page.c | 178 ++++++++++++++++++ 6 files changed, 209 insertions(+) create mode 100644 mm/clear_freelist_page.c -- 2.33.0

2 3

[PATCH OLK-6.6 v3 0/1] stop_machine: mask pseudo nmi before running the callback
by Yuntao Liu 21 Dec '23

21 Dec '23

From: Jinjie Ruan <ruanjinjie(a)huawei.com> *** BLURB HERE *** Wei Li (1): stop_machine: mask pseudo nmi before running the callback arch/arm64/include/asm/arch_gicv3.h | 12 ++++++++++++ kernel/stop_machine.c | 3 +++ 2 files changed, 15 insertions(+) -- 2.34.1

2 2

[PATCH OLK-6.6 0/2] Introduce clear freelist feature
by Yu Liao 21 Dec '23

21 Dec '23

Yu Liao (2): mm: Add sysctl to clear free list pages config: make CONFIG_CLEAR_FREELIST_PAGE not set by default .../admin-guide/kernel-parameters.txt | 3 + Documentation/admin-guide/sysctl/vm.rst | 13 ++ arch/arm64/configs/openeuler_defconfig | 1 + mm/Kconfig | 13 ++ mm/Makefile | 2 + mm/clear_freelist_page.c | 178 ++++++++++++++++++ 6 files changed, 210 insertions(+) create mode 100644 mm/clear_freelist_page.c -- 2.33.0

2 3

[PATCH OLK-6.6 v2 0/1] stop_machine: mask pseudo nmi before running the callback
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Wei Li (1): stop_machine: mask pseudo nmi before running the callback arch/arm64/include/asm/arch_gicv3.h | 12 ++++++++++++ kernel/stop_machine.c | 3 +++ 2 files changed, 15 insertions(+) -- 2.34.1

2 2

[PATCH OLK-6.6 0/3] cgroup v1 writeback
by chenridong 21 Dec '23

21 Dec '23

Lu Jialin (2): cgroup: Factor out __cgroup_get_from_id() for cgroup v1 openeuler_defconfig: enable CONFIG_CGROUP_V1_WRITEBACK in openeuler_defconfig for x86 and arm64 chenridong (1): cgroup: support cgroup writeback on cgroupv1 arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + block/blk-cgroup.c | 3 + block/blk-cgroup.h | 3 + include/linux/backing-dev.h | 29 ++++++- include/linux/cgroup-defs.h | 2 + include/linux/cgroup.h | 1 + include/linux/memcontrol.h | 5 ++ init/Kconfig | 5 ++ kernel/cgroup/cgroup.c | 31 +++++-- mm/backing-dev.c | 116 ++++++++++++++++++++++++- mm/memcontrol.c | 83 +++++++++++++++++- 12 files changed, 267 insertions(+), 13 deletions(-) -- 2.34.1

2 4

[PATCH OLK-6.6 0/3] Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely"
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Guo Xuenan (3): Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely" make OPTIMIZE_INLINING config editable disable OPTIMIZE_INLINING by default arch/arm64/kvm/sys_regs.h | 5 +++++ arch/x86/configs/i386_defconfig | 1 + arch/x86/configs/x86_64_defconfig | 1 + drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h | 5 +++++ .../pci/hive_isp_css_include/print_support.h | 4 ++++ include/linux/compiler_types.h | 8 ++++++++ include/trace/trace_events.h | 2 +- kernel/configs/tiny.config | 1 + lib/Kconfig.debug | 13 +++++++++++++ 9 files changed, 39 insertions(+), 1 deletion(-) -- 2.34.1

2 4

[PATCH OLK-6.6 0/1] mbigen: fix mbigen driver defer probe
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Yang Yingliang (1): mbigen: fix mbigen driver defer probe drivers/irqchip/irq-mbigen.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) -- 2.34.1

2 2

[PATCH OLK-5.10] roh/hns3: Fix the processing flow of ROH CMDq during the reset process.
by Ke Chen 21 Dec '23

21 Dec '23

driver inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8PS3D ----------------------------------------------------------------------- Fix the sequence issue in the ROH CMDq processing flow during the FLR reset process, which may occasionally cause calltrace. In this patch, we added a flag in the reset process to indicate the availability of the ROH CMDq, in order to fix this issue. Fixes:1cb9fff33035("roh/hns3: Add ROH cmdq interface support") Signed-off-by: Ke Chen <chenke54(a)huawei.com> Reviewed-by: Gang Zhang <gang.zhang(a)huawei.com> --- drivers/roh/hw/hns3/hns3_cmdq.c | 5 +++++ drivers/roh/hw/hns3/hns3_device.h | 3 ++- drivers/roh/hw/hns3/hns3_main.c | 3 +++ 3 files changed, 10 insertions(+), 1 deletion(-) diff --git a/drivers/roh/hw/hns3/hns3_cmdq.c b/drivers/roh/hw/hns3/hns3_cmdq.c index 75fa2e852152..96c0286779eb 100644 --- a/drivers/roh/hw/hns3/hns3_cmdq.c +++ b/drivers/roh/hw/hns3/hns3_cmdq.c @@ -141,6 +141,8 @@ int hns3_roh_cmdq_init(struct hns3_roh_device *hroh_dev) /* Init CRQ REG */ hns3_roh_cmdq_init_regs(hroh_dev, HNS3_ROH_CMDQ_CRQ); + clear_bit(HNS3_ROH_STATE_CMD_DISABLE, &priv->handle->rohinfo.reset_state); + return 0; err_crq: @@ -340,6 +342,9 @@ int hns3_roh_cmdq_send(struct hns3_roh_device *hroh_dev, struct hns3_roh_desc *d int ntc = 0; int ret = 0; + if (test_bit(HNS3_ROH_STATE_CMD_DISABLE, &priv->handle->rohinfo.reset_state)) + return -EIO; + spin_lock_bh(&csq->lock); ret = hns3_roh_cmdq_build(hroh_dev, desc, num, &ntc); if (ret) { diff --git a/drivers/roh/hw/hns3/hns3_device.h b/drivers/roh/hw/hns3/hns3_device.h index 07e46879a4b9..4cc5f6aeffb8 100644 --- a/drivers/roh/hw/hns3/hns3_device.h +++ b/drivers/roh/hw/hns3/hns3_device.h @@ -8,7 +8,8 @@ enum hns3_roh_state { HNS3_ROH_STATE_RESETTING, HNS3_ROH_STATE_INIT, HNS3_ROH_STATE_INITED, - HNS3_ROH_STATE_DOWN + HNS3_ROH_STATE_DOWN, + HNS3_ROH_STATE_CMD_DISABLE, }; enum { HNS3_ROH_RST_DIRECT_RETURN = 0 }; diff --git a/drivers/roh/hw/hns3/hns3_main.c b/drivers/roh/hw/hns3/hns3_main.c index 64f9a13e5d83..5d5972f1af7f 100644 --- a/drivers/roh/hw/hns3/hns3_main.c +++ b/drivers/roh/hw/hns3/hns3_main.c @@ -388,6 +388,9 @@ static int hns3_roh_reset_notify(struct hnae3_handle *handle, case HNAE3_UNINIT_CLIENT: ret = hns3_roh_reset_notify_uninit(handle); break; + case HNAE3_DOWN_CLIENT: + set_bit(HNS3_ROH_STATE_CMD_DISABLE, &handle->rohinfo.reset_state); + break; default: break; } -- 2.30.0

2 1

[PATCH OLK-5.10] roh/hns3: Fix the processing flow of ROH CMDq during the reset process.
by Ke Chen 21 Dec '23

21 Dec '23

driver inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8PS3D ----------------------------------------------------------------------- Fix the sequence issue in the ROH CMDq processing flow during the FLR reset process, which may occasionally cause calltrace. In this patch, we added a flag in the reset process to indicate the availability of the ROH CMDq, in order to fix this issue. Fixes:1cb9fff33035("roh/hns3: Add ROH cmdq interface support") Signed-off-by: Ke Chen <chenke54(a)huawei.com> Reviewed-by: Gang Zhang <gang.zhang(a)huawei.com> --- drivers/roh/hw/hns3/hns3_cmdq.c | 5 +++++ drivers/roh/hw/hns3/hns3_device.h | 3 ++- drivers/roh/hw/hns3/hns3_main.c | 2 ++ 3 files changed, 9 insertions(+), 1 deletion(-) diff --git a/drivers/roh/hw/hns3/hns3_cmdq.c b/drivers/roh/hw/hns3/hns3_cmdq.c index 75fa2e852152..96c0286779eb 100644 --- a/drivers/roh/hw/hns3/hns3_cmdq.c +++ b/drivers/roh/hw/hns3/hns3_cmdq.c @@ -141,6 +141,8 @@ int hns3_roh_cmdq_init(struct hns3_roh_device *hroh_dev) /* Init CRQ REG */ hns3_roh_cmdq_init_regs(hroh_dev, HNS3_ROH_CMDQ_CRQ); + clear_bit(HNS3_ROH_STATE_CMD_DISABLE, &priv->handle->rohinfo.reset_state); + return 0; err_crq: @@ -340,6 +342,9 @@ int hns3_roh_cmdq_send(struct hns3_roh_device *hroh_dev, struct hns3_roh_desc *d int ntc = 0; int ret = 0; + if (test_bit(HNS3_ROH_STATE_CMD_DISABLE, &priv->handle->rohinfo.reset_state)) + return -EIO; + spin_lock_bh(&csq->lock); ret = hns3_roh_cmdq_build(hroh_dev, desc, num, &ntc); if (ret) { diff --git a/drivers/roh/hw/hns3/hns3_device.h b/drivers/roh/hw/hns3/hns3_device.h index 07e46879a4b9..4cc5f6aeffb8 100644 --- a/drivers/roh/hw/hns3/hns3_device.h +++ b/drivers/roh/hw/hns3/hns3_device.h @@ -8,7 +8,8 @@ enum hns3_roh_state { HNS3_ROH_STATE_RESETTING, HNS3_ROH_STATE_INIT, HNS3_ROH_STATE_INITED, - HNS3_ROH_STATE_DOWN + HNS3_ROH_STATE_DOWN, + HNS3_ROH_STATE_CMD_DISABLE, }; enum { HNS3_ROH_RST_DIRECT_RETURN = 0 }; diff --git a/drivers/roh/hw/hns3/hns3_main.c b/drivers/roh/hw/hns3/hns3_main.c index 64f9a13e5d83..78e57a3d8a95 100644 --- a/drivers/roh/hw/hns3/hns3_main.c +++ b/drivers/roh/hw/hns3/hns3_main.c @@ -388,6 +388,8 @@ static int hns3_roh_reset_notify(struct hnae3_handle *handle, case HNAE3_UNINIT_CLIENT: ret = hns3_roh_reset_notify_uninit(handle); break; + case HNAE3_DOWN_CLIENT: + set_bit(HNS3_ROH_STATE_CMD_DISABLE, &handle->rohinfo.reset_state); default: break; } -- 2.30.0

2 3

[PATCH OLK-6.6 v6 0/3] Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely"
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Guo Xuenan (3): Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely" make OPTIMIZE_INLINING config editable disable OPTIMIZE_INLINING by default hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8PGC4 arch/arm64/kvm/sys_regs.h | 5 +++++ arch/x86/configs/i386_defconfig | 1 + arch/x86/configs/x86_64_defconfig | 1 + drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h | 5 +++++ .../pci/hive_isp_css_include/print_support.h | 4 ++++ include/linux/compiler_types.h | 8 ++++++++ include/trace/trace_events.h | 2 +- kernel/configs/tiny.config | 1 + lib/Kconfig.debug | 13 +++++++++++++ 9 files changed, 39 insertions(+), 1 deletion(-) -- 2.34.1

2 4

[PATCH OLK-6.6 0/2] stop_machine: mask pseudo nmi before running the callback
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Wei Li (2): stop_machine: mask pseudo nmi before running the callback arm64: Fix wrong logic in gic_arch_restore_irqs() arch/arm64/include/asm/arch_gicv3.h | 11 +++++++++++ kernel/stop_machine.c | 3 +++ 2 files changed, 14 insertions(+) -- 2.34.1

2 3

[PATCH OLK-6.6 v5 0/3] Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely"
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Guo Xuenan (3): Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely" make OPTIMIZE_INLINING config editable disable OPTIMIZE_INLINING by default hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8PGC4 arch/arm64/kvm/sys_regs.h | 5 +++++ arch/x86/configs/i386_defconfig | 1 + arch/x86/configs/x86_64_defconfig | 1 + drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h | 5 +++++ .../pci/hive_isp_css_include/print_support.h | 4 ++++ include/linux/compiler_types.h | 8 ++++++++ include/trace/trace_events.h | 2 +- kernel/configs/tiny.config | 1 + lib/Kconfig.debug | 13 +++++++++++++ 9 files changed, 39 insertions(+), 1 deletion(-) -- 2.34.1

2 4

[PATCH OLK-6.6 0/1] ARM: LPAE: Use phys_addr_t instead of unsigned long in outercache hooks
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Zhen Lei (1): ARM: LPAE: Use phys_addr_t instead of unsigned long in outercache hooks arch/arm/include/asm/outercache.h | 6 ++-- arch/arm/mm/cache-feroceon-l2.c | 15 ++++++++-- arch/arm/mm/cache-l2x0.c | 50 ++++++++++++++++++++++--------- arch/arm/mm/cache-tauros2.c | 15 ++++++++-- arch/arm/mm/cache-uniphier.c | 6 ++-- arch/arm/mm/cache-xsc3l2.c | 12 ++++++-- 6 files changed, 75 insertions(+), 29 deletions(-) -- 2.34.1

2 2

[PATCH OLK-6.6 v4 0/3] Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely"
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Guo Xuenan (3): Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely" make OPTIMIZE_INLINING config editable disable OPTIMIZE_INLINING by default arch/arm64/kvm/sys_regs.h | 5 +++++ arch/x86/configs/i386_defconfig | 1 + arch/x86/configs/x86_64_defconfig | 1 + drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h | 2 +- .../pci/hive_isp_css_include/print_support.h | 4 ++++ include/linux/compiler_types.h | 8 ++++++++ include/trace/trace_events.h | 2 +- kernel/configs/tiny.config | 1 + lib/Kconfig.debug | 13 +++++++++++++ 9 files changed, 35 insertions(+), 2 deletions(-) -- 2.34.1

2 4

[PATCH OLK-6.6 v2 0/6] kabi: net: reserve space for net
by Zhengchao Shao 21 Dec '23

21 Dec '23

2 7

[PATCH OLK-6.6 0/1] fuse: reserve space for future expansion
by yangyun 21 Dec '23

21 Dec '23

fuse: reserve space for futer expansion Wenchao Hao (1): kabi:fuse: reserve space for future expansion fs/fuse/fuse_i.h | 10 ++++++++++ 1 file changed, 10 insertions(+) -- 2.33.0

2 2

[PATCH OLK-5.10] roh/hns3: Fix the processing flow of ROH CMDq during the reset process.
by Ke Chen 21 Dec '23

21 Dec '23

driver inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8PS3D ----------------------------------------------------------------------- Fix the sequence issue in the ROH CMDq processing flow during the FLR reset process, which may occasionally cause calltrace. In this patch, we added a flag in the reset process to indicate the availability of the ROH CMDq, in order to fix this issue. Fixes:1cb9fff33035("roh/hns3: Add ROH cmdq interface support") Signed-off-by: Ke Chen <chenke54(a)huawei.com> Reviewed-by: Gang Zhang <gang.zhang(a)huawei.com> --- drivers/roh/hw/hns3/hns3_cmdq.c | 5 +++++ drivers/roh/hw/hns3/hns3_device.h | 3 ++- drivers/roh/hw/hns3/hns3_main.c | 2 ++ 3 files changed, 9 insertions(+), 1 deletion(-) diff --git a/drivers/roh/hw/hns3/hns3_cmdq.c b/drivers/roh/hw/hns3/hns3_cmdq.c index 75fa2e852152..96c0286779eb 100644 --- a/drivers/roh/hw/hns3/hns3_cmdq.c +++ b/drivers/roh/hw/hns3/hns3_cmdq.c @@ -141,6 +141,8 @@ int hns3_roh_cmdq_init(struct hns3_roh_device *hroh_dev) /* Init CRQ REG */ hns3_roh_cmdq_init_regs(hroh_dev, HNS3_ROH_CMDQ_CRQ); + clear_bit(HNS3_ROH_STATE_CMD_DISABLE, &priv->handle->rohinfo.reset_state); + return 0; err_crq: @@ -340,6 +342,9 @@ int hns3_roh_cmdq_send(struct hns3_roh_device *hroh_dev, struct hns3_roh_desc *d int ntc = 0; int ret = 0; + if (test_bit(HNS3_ROH_STATE_CMD_DISABLE, &priv->handle->rohinfo.reset_state)) + return -EIO; + spin_lock_bh(&csq->lock); ret = hns3_roh_cmdq_build(hroh_dev, desc, num, &ntc); if (ret) { diff --git a/drivers/roh/hw/hns3/hns3_device.h b/drivers/roh/hw/hns3/hns3_device.h index 07e46879a4b9..4cc5f6aeffb8 100644 --- a/drivers/roh/hw/hns3/hns3_device.h +++ b/drivers/roh/hw/hns3/hns3_device.h @@ -8,7 +8,8 @@ enum hns3_roh_state { HNS3_ROH_STATE_RESETTING, HNS3_ROH_STATE_INIT, HNS3_ROH_STATE_INITED, - HNS3_ROH_STATE_DOWN + HNS3_ROH_STATE_DOWN, + HNS3_ROH_STATE_CMD_DISABLE, }; enum { HNS3_ROH_RST_DIRECT_RETURN = 0 }; diff --git a/drivers/roh/hw/hns3/hns3_main.c b/drivers/roh/hw/hns3/hns3_main.c index 64f9a13e5d83..78e57a3d8a95 100644 --- a/drivers/roh/hw/hns3/hns3_main.c +++ b/drivers/roh/hw/hns3/hns3_main.c @@ -388,6 +388,8 @@ static int hns3_roh_reset_notify(struct hnae3_handle *handle, case HNAE3_UNINIT_CLIENT: ret = hns3_roh_reset_notify_uninit(handle); break; + case HNAE3_DOWN_CLIENT: + set_bit(HNS3_ROH_STATE_CMD_DISABLE, &handle->rohinfo.reset_state); default: break; } -- 2.30.0

2 1

[PATCH OLK-6.6 v3 0/3] Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely"
by Yuntao Liu 21 Dec '23

21 Dec '23

*** BLURB HERE *** Guo Xuenan (3): Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely" make OPTIMIZE_INLINING config editable disable OPTIMIZE_INLINING by default arch/arm64/kvm/sys_regs.h | 5 +++++ arch/x86/configs/i386_defconfig | 1 + arch/x86/configs/x86_64_defconfig | 1 + drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h | 2 +- .../pci/hive_isp_css_include/print_support.h | 4 ++++ include/linux/compiler_types.h | 8 ++++++++ include/trace/trace_events.h | 2 +- kernel/configs/tiny.config | 1 + lib/Kconfig.debug | 13 +++++++++++++ 9 files changed, 35 insertions(+), 2 deletions(-) -- 2.34.1

2 4

[PATCH OLK-6.6 0/3] xfs: fix hung and warning
by Long Li 21 Dec '23

21 Dec '23

This patch set fix some hung task and warning bug: Long Li (1): xfs: fix hung when transaction commit fail in xfs_inactive_ifree Ye Bin (2): xfs: fix dead loop when do mount with IO fault injection xfs: fix warning in xfs_vm_writepages() fs/xfs/xfs_buf_item.c | 20 ++++++++++++++++++++ fs/xfs/xfs_icache.c | 6 ++++++ fs/xfs/xfs_inode.c | 3 ++- fs/xfs/xfs_inode_item.c | 3 ++- fs/xfs/xfs_log.c | 4 +++- 5 files changed, 33 insertions(+), 3 deletions(-) -- 2.31.1

2 4

[PATCH OLK-6.6 v2 0/3] Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely"
by Yuntao Liu 20 Dec '23

20 Dec '23

*** BLURB HERE *** Guo Xuenan (3): Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely" make OPTIMIZE_INLINING config editable disable OPTIMIZE_INLINING by default arch/arm64/kvm/sys_regs.h | 5 +++++ arch/x86/configs/i386_defconfig | 1 + arch/x86/configs/x86_64_defconfig | 1 + drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h | 2 +- .../pci/hive_isp_css_include/print_support.h | 4 ++++ include/linux/compiler_types.h | 8 ++++++++ include/trace/trace_events.h | 2 +- kernel/configs/tiny.config | 1 + lib/Kconfig.debug | 13 +++++++++++++ 9 files changed, 35 insertions(+), 2 deletions(-) -- 2.34.1

2 4

[PATCH OLK-6.6 0/6] kabi: net: reserve space for net
by Zhengchao Shao 20 Dec '23

20 Dec '23

2 7

[PATCH OLK-5.10 0/2] fix CVE-2023-6817 and CVE-2023-4244
by Dong Chenchen 20 Dec '23

20 Dec '23

fix CVE-2023-6817 and CVE-2023-4244 Florian Westphal (2): netfilter: nf_tables: don't skip expired elements during walk netfilter: nft_set_pipapo: skip inactive elements during set walk net/netfilter/nf_tables_api.c | 4 ++++ net/netfilter/nft_set_hash.c | 2 -- net/netfilter/nft_set_pipapo.c | 19 ++++++++++++++----- net/netfilter/nft_set_rbtree.c | 2 -- 4 files changed, 18 insertions(+), 9 deletions(-) -- 2.25.1

2 3

[PATCH OLK-6.6 0/8] adapt kabi reserve patch to OLK6.6
by Lin Yujun 20 Dec '23

20 Dec '23

*** BLURB HERE *** Chen Jiahao (2): KABI: reserve space for several i2c structures KABI: reserve space for struct input_dev Lin Ruizhe (6): bootparam: Add kabi_reserve in bootparam interrupt: Add kabi_reserve in interrupt.h irq: Add kabi_reserve in irq irq_desc: Add kabi_reserve in irq_desc irqdomain: Add kabi_reserve in irqdomain msi: Add kabi_reserve in msi.h arch/x86/include/uapi/asm/bootparam.h | 2 ++ include/linux/i2c.h | 11 +++++++++++ include/linux/input.h | 4 ++++ include/linux/interrupt.h | 4 ++++ include/linux/irq.h | 6 ++++++ include/linux/irqdesc.h | 3 ++- include/linux/irqdomain.h | 2 ++ include/linux/msi.h | 3 +++ 8 files changed, 34 insertions(+), 1 deletion(-) -- 2.34.1

2 9

[PATCH OLK-6.6 0/5] ubi: Enhanced fault injection capability for the UBI driver
by ZhaoLong Wang 20 Dec '23

20 Dec '23

CONFIG Dependency CONFIG_MTD=y CONFIG_MTD_NAND_NANDSIM=m CONFIG_MTD_UBI=m CONFIG_FAULT_INJECTION_DEBUG_FS=y CONFIG_MTD_PARTITIONED_MASTER The existing fault injection capability of UBI is too simple. It uses hard-coded fault probability values and lacks other configurable options. As a result, these interfaces are difficult to use when digging defects in the abnormal path of code and reproducing some problems. The kernel provides a powerful fault injection framework, which provides rich configurable fault injection attributes during runtime. So it can be used to improve the fault injection capability of the UBI driver. This series of patches refactor the existing fault injection interface and add some fault injection types to help testers and developers find potential problems in the code. This series of patches enhance the existing fault injection interface and retain the old debugfs interface, and add some fault injection types to help testers and developers Look for potential problems in the code. ZhaoLong Wang (5): ubi: Use the fault injection framework to enhance the fault injection capability ubi: Split io_failures into write_failure and erase_failure ubi: Add six fault injection type for testing ubi: Reserve sufficient buffer length for the input mask mtd: Add several functions to the fail_function list drivers/mtd/mtdcore.c | 5 + drivers/mtd/ubi/Kconfig | 9 ++ drivers/mtd/ubi/debug.c | 107 ++++++++++++-- drivers/mtd/ubi/debug.h | 304 +++++++++++++++++++++++++++++++++++++--- drivers/mtd/ubi/io.c | 86 +++++++++++- drivers/mtd/ubi/ubi.h | 45 +++--- 6 files changed, 497 insertions(+), 59 deletions(-) -- 2.39.2

2 6

[PATCH OLK-6.6 v5] ascend: export interfaces required by ascend drivers
by Yuan Can 20 Dec '23

20 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OX3Q CVE: NA -------------------------------- Export oom_type_notifier_call and __get_vm_area_caller. Export pm_autosleep_set_state and __vmalloc_node_range. Export alloc_workqueue_attrs, free_workqueue_attrs and apply_workqueue_attrs. Export find_get_task_by_vpid. Signed-off-by: Yuan Can <yuancan(a)huawei.com> --- kernel/pid.c | 1 + kernel/power/autosleep.c | 1 + kernel/workqueue.c | 3 +++ mm/oom_kill.c | 1 + mm/vmalloc.c | 2 ++ 5 files changed, 8 insertions(+) diff --git a/kernel/pid.c b/kernel/pid.c index 6500ef956f2f..383abde0c208 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -437,6 +437,7 @@ struct task_struct *find_get_task_by_vpid(pid_t nr) return task; } +EXPORT_SYMBOL_GPL(find_get_task_by_vpid); struct pid *get_task_pid(struct task_struct *task, enum pid_type type) { diff --git a/kernel/power/autosleep.c b/kernel/power/autosleep.c index b29c8aca7486..80ba474daa40 100644 --- a/kernel/power/autosleep.c +++ b/kernel/power/autosleep.c @@ -113,6 +113,7 @@ int pm_autosleep_set_state(suspend_state_t state) mutex_unlock(&autosleep_lock); return 0; } +EXPORT_SYMBOL_GPL(pm_autosleep_set_state); int __init pm_autosleep_init(void) { diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 0f682da96e1c..3eabf97c4e9a 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -3713,6 +3713,7 @@ void free_workqueue_attrs(struct workqueue_attrs *attrs) kfree(attrs); } } +EXPORT_SYMBOL_GPL(free_workqueue_attrs); /** * alloc_workqueue_attrs - allocate a workqueue_attrs @@ -3741,6 +3742,7 @@ struct workqueue_attrs *alloc_workqueue_attrs(void) free_workqueue_attrs(attrs); return NULL; } +EXPORT_SYMBOL_GPL(alloc_workqueue_attrs); static void copy_workqueue_attrs(struct workqueue_attrs *to, const struct workqueue_attrs *from) @@ -4482,6 +4484,7 @@ int apply_workqueue_attrs(struct workqueue_struct *wq, return ret; } +EXPORT_SYMBOL_GPL(apply_workqueue_attrs); /** * wq_update_pod - update pod affinity of a wq for CPU hot[un]plug diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 601ee56cc7d7..81b79f82f601 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -1141,6 +1141,7 @@ int oom_type_notifier_call(unsigned int type, struct oom_control *oc) return blocking_notifier_call_chain(&oom_type_notify_list, type, NULL); } +EXPORT_SYMBOL_GPL(oom_type_notifier_call); #endif /** diff --git a/mm/vmalloc.c b/mm/vmalloc.c index 30665fb33589..31761425d7a8 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -2622,6 +2622,7 @@ struct vm_struct *__get_vm_area_caller(unsigned long size, unsigned long flags, return __get_vm_area_node(size, 1, PAGE_SHIFT, flags, start, end, NUMA_NO_NODE, GFP_KERNEL, caller); } +EXPORT_SYMBOL(__get_vm_area_caller); /** * get_vm_area - reserve a contiguous kernel virtual area @@ -3362,6 +3363,7 @@ void *__vmalloc_node_range(unsigned long size, unsigned long align, return NULL; } +EXPORT_SYMBOL(__vmalloc_node_range); /** * __vmalloc_node - allocate virtually contiguous memory -- 2.17.1

2 1

[PATCH OLK-6.6 v4] ascend: export interfaces required by ascend drivers
by Yuan Can 20 Dec '23

20 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OX3Q CVE: NA -------------------------------- Export oom_type_notifier_call and __get_vm_area_caller. Export pm_autosleep_set_state and __vmalloc_node_range. Export alloc_workqueue_attrs, free_workqueue_attrs and apply_workqueue_attrs. Export find_get_task_by_vpid. Signed-off-by: Yuan Can <yuancan(a)huawei.com> --- kernel/pid.c | 1 + kernel/power/autosleep.c | 1 + kernel/workqueue.c | 3 +++ mm/oom_kill.c | 1 + mm/vmalloc.c | 2 ++ 5 files changed, 8 insertions(+) diff --git a/kernel/pid.c b/kernel/pid.c index 6500ef956f2f..383abde0c208 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -437,6 +437,7 @@ struct task_struct *find_get_task_by_vpid(pid_t nr) return task; } +EXPORT_SYMBOL_GPL(find_get_task_by_vpid); struct pid *get_task_pid(struct task_struct *task, enum pid_type type) { diff --git a/kernel/power/autosleep.c b/kernel/power/autosleep.c index b29c8aca7486..80ba474daa40 100644 --- a/kernel/power/autosleep.c +++ b/kernel/power/autosleep.c @@ -113,6 +113,7 @@ int pm_autosleep_set_state(suspend_state_t state) mutex_unlock(&autosleep_lock); return 0; } +EXPORT_SYMBOL_GPL(pm_autosleep_set_state); int __init pm_autosleep_init(void) { diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 0f682da96e1c..3eabf97c4e9a 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -3713,6 +3713,7 @@ void free_workqueue_attrs(struct workqueue_attrs *attrs) kfree(attrs); } } +EXPORT_SYMBOL_GPL(free_workqueue_attrs); /** * alloc_workqueue_attrs - allocate a workqueue_attrs @@ -3741,6 +3742,7 @@ struct workqueue_attrs *alloc_workqueue_attrs(void) free_workqueue_attrs(attrs); return NULL; } +EXPORT_SYMBOL_GPL(alloc_workqueue_attrs); static void copy_workqueue_attrs(struct workqueue_attrs *to, const struct workqueue_attrs *from) @@ -4482,6 +4484,7 @@ int apply_workqueue_attrs(struct workqueue_struct *wq, return ret; } +EXPORT_SYMBOL_GPL(apply_workqueue_attrs); /** * wq_update_pod - update pod affinity of a wq for CPU hot[un]plug diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 601ee56cc7d7..81b79f82f601 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -1141,6 +1141,7 @@ int oom_type_notifier_call(unsigned int type, struct oom_control *oc) return blocking_notifier_call_chain(&oom_type_notify_list, type, NULL); } +EXPORT_SYMBOL_GPL(oom_type_notifier_call); #endif /** diff --git a/mm/vmalloc.c b/mm/vmalloc.c index 30665fb33589..719539b32488 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -2622,6 +2622,7 @@ struct vm_struct *__get_vm_area_caller(unsigned long size, unsigned long flags, return __get_vm_area_node(size, 1, PAGE_SHIFT, flags, start, end, NUMA_NO_NODE, GFP_KERNEL, caller); } +EXPORT_SYMBOL(__get_vm_area_caller); /** * get_vm_area - reserve a contiguous kernel virtual area @@ -3362,6 +3363,7 @@ void *__vmalloc_node_range(unsigned long size, unsigned long align, return NULL; } +EXPORT_SYMBOL_GPL(__vmalloc_node_range); /** * __vmalloc_node - allocate virtually contiguous memory -- 2.17.1

2 1

[PATCH OLK-6.6 0/1] Support priority load balance for qos scheduler
by Song Zhang 20 Dec '23

20 Dec '23

Song Zhang (1): sched: Introduce priority load balance for qos scheduler arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + init/Kconfig | 9 +++ kernel/sched/core.c | 3 + kernel/sched/fair.c | 90 +++++++++++++++++++++++++- kernel/sched/sched.h | 3 + 6 files changed, 106 insertions(+), 1 deletion(-) -- 2.33.0

2 2

[PATCH V3 OLK-6.6 0/3] sync smmu patches for olk-6.6
by Zhang Zekun 19 Dec '23

19 Dec '23

Make the rache depot scale better. The origin depot size is fixed, may cause problem in some condition. Add another patch to fix smmu problem. v2->v3: - Add a smmu bugfix patch v1->v2: - Change the commit messge Robin Murphy (2): iommu/iova: Make the rcache depot scale better iommu/iova: Manage the depot list size Wenkai Lin (1): iommu/arm-smmu-v3: disable stall for quiet_cd drivers/iommu/arm/arm-smmu-v3/arm-smmu-v3.c | 3 + drivers/iommu/iova.c | 95 ++++++++++++++------- 2 files changed, 68 insertions(+), 30 deletions(-) -- 2.17.1

2 4

[PATCH OLK-6.6 0/8] adapt kabi reserve patch to OLK6.6
by Lin Yujun 19 Dec '23

19 Dec '23

*** BLURB HERE *** Chen Jiahao (2): KABI: reserve space for several i2c structures KABI: reserve space for struct input_dev Lin Ruizhe (6): bootparam: Add kabi_reserve in bootparam interrupt: Add kabi_reserve in interrupt.h irq: Add kabi_reserve in irq irq_desc: Add kabi_reserve in irq_desc irqdomain: Add kabi_reserve in irqdomain msi: Add kabi_reserve in msi.h arch/x86/include/uapi/asm/bootparam.h | 2 ++ include/linux/i2c.h | 11 +++++++++++ include/linux/input.h | 4 ++++ include/linux/interrupt.h | 4 ++++ include/linux/irq.h | 6 ++++++ include/linux/irqdesc.h | 3 ++- include/linux/irqdomain.h | 2 ++ include/linux/msi.h | 3 +++ 8 files changed, 34 insertions(+), 1 deletion(-) -- 2.34.1

2 9

[PATCH OLK-6.6 v2 0/5] Terrace Service Acceleration
by Zhengchao Shao 19 Dec '23

19 Dec '23

Terrace Service Acceleration. Liu Jian (3): bpf: Add bpf_get_sockops_uid_gid helper function bpf: Add new bpf helper to get SO_ORIGINAL_DST/REPLY_SRC bpf: Access bpf_sock's src_ip4 and sorc_port in BPF_CGROUP_INET_SOCK_RELEASE hook Lu Wei (1): net: core: Add a GID field to struct sock. Wang Yufen (1): bpf, sockmap: Add sk_rmem_alloc check for sockmap include/net/netfilter/nf_conntrack.h | 6 ++ include/net/sock.h | 18 ++++++ include/uapi/linux/bpf.h | 13 +++++ include/uapi/linux/netfilter_ipv4.h | 2 + net/Kconfig | 7 +++ net/core/filter.c | 84 ++++++++++++++++++++++++++++ net/core/skmsg.c | 10 +++- net/core/sock.c | 6 ++ net/ipv4/tcp_bpf.c | 10 ++++ net/netfilter/nf_conntrack_proto.c | 71 +++++++++++++++++++++++ net/socket.c | 8 ++- tools/include/uapi/linux/bpf.h | 13 +++++ 12 files changed, 245 insertions(+), 3 deletions(-) -- 2.34.1

2 6

[PATCH openEuler-1.0-LTS v2] net: Remove acked SYN flag from packet in the transmit queue correctly
by Dong Chenchen 19 Dec '23

19 Dec '23

mainline inclusion from mainline-v6.7-rc4 commit f99cd56230f56c8b6b33713c5be4da5d6766be1f category: bugfix bugzilla: 187254, https://gitee.com/openeuler/kernel/issues/I8PRTS Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… --------------------------- syzkaller report: kernel BUG at net/core/skbuff.c:3452! invalid opcode: 0000 [#1] PREEMPT SMP KASAN PTI CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.7.0-rc4-00009-gbee0e7762ad2-dirty #135 RIP: 0010:skb_copy_and_csum_bits (net/core/skbuff.c:3452) Call Trace: icmp_glue_bits (net/ipv4/icmp.c:357) __ip_append_data.isra.0 (net/ipv4/ip_output.c:1165) ip_append_data (net/ipv4/ip_output.c:1362 net/ipv4/ip_output.c:1341) icmp_push_reply (net/ipv4/icmp.c:370) __icmp_send (./include/net/route.h:252 net/ipv4/icmp.c:772) ip_fragment.constprop.0 (./include/linux/skbuff.h:1234 net/ipv4/ip_output.c:592 net/ipv4/ip_output.c:577) __ip_finish_output (net/ipv4/ip_output.c:311 net/ipv4/ip_output.c:295) ip_output (net/ipv4/ip_output.c:427) __ip_queue_xmit (net/ipv4/ip_output.c:535) __tcp_transmit_skb (net/ipv4/tcp_output.c:1462) __tcp_retransmit_skb (net/ipv4/tcp_output.c:3387) tcp_retransmit_skb (net/ipv4/tcp_output.c:3404) tcp_retransmit_timer (net/ipv4/tcp_timer.c:604) tcp_write_timer (./include/linux/spinlock.h:391 net/ipv4/tcp_timer.c:716) The panic issue was trigered by tcp simultaneous initiation. The initiation process is as follows: TCP A TCP B 1. CLOSED CLOSED 2. SYN-SENT --> <SEQ=100><CTL=SYN> ... 3. SYN-RECEIVED <-- <SEQ=300><CTL=SYN> <-- SYN-SENT 4. ... <SEQ=100><CTL=SYN> --> SYN-RECEIVED 5. SYN-RECEIVED --> <SEQ=100><ACK=301><CTL=SYN,ACK> ... // TCP B: not send challenge ack for ack limit or packet loss // TCP A: close tcp_close tcp_send_fin if (!tskb && tcp_under_memory_pressure(sk)) tskb = skb_rb_last(&sk->tcp_rtx_queue); //pick SYN_ACK packet TCP_SKB_CB(tskb)->tcp_flags |= TCPHDR_FIN; // set FIN flag 6. FIN_WAIT_1 --> <SEQ=100><ACK=301><END_SEQ=102><CTL=SYN,FIN,ACK> ... // TCP B: send challenge ack to SYN_FIN_ACK 7. ... <SEQ=301><ACK=101><CTL=ACK> <-- SYN-RECEIVED //challenge ack // TCP A: <SND.UNA=101> 8. FIN_WAIT_1 --> <SEQ=101><ACK=301><END_SEQ=102><CTL=SYN,FIN,ACK> ... // retransmit panic __tcp_retransmit_skb //skb->len=0 tcp_trim_head len = tp->snd_una - TCP_SKB_CB(skb)->seq // len=101-100 __pskb_trim_head skb->data_len -= len // skb->len=-1, wrap around ... ... ip_fragment icmp_glue_bits //BUG_ON If we use tcp_trim_head() to remove acked SYN from packet that contains data or other flags, skb->len will be incorrectly decremented. We can remove SYN flag that has been acked from rtx_queue earlier than tcp_trim_head(), which can fix the problem mentioned above. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Co-developed-by: Eric Dumazet <edumazet(a)google.com> Signed-off-by: Eric Dumazet <edumazet(a)google.com> Signed-off-by: Dong Chenchen <dongchenchen2(a)huawei.com> Link: https://lore.kernel.org/r/20231210020200.1539875-1-dongchenchen2@huawei.com Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: Dong Chenchen <dongchenchen2(a)huawei.com> --- net/ipv4/tcp_output.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 8cd673d75f68..90e0041ae9f7 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -2930,7 +2930,13 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb, int segs) if (skb_still_in_host_queue(sk, skb)) return -EBUSY; +start: if (before(TCP_SKB_CB(skb)->seq, tp->snd_una)) { + if (unlikely(TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)) { + TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_SYN; + TCP_SKB_CB(skb)->seq++; + goto start; + } if (unlikely(before(TCP_SKB_CB(skb)->end_seq, tp->snd_una))) { WARN_ON_ONCE(1); return -EINVAL; -- 2.25.1

2 1

[PATCH v3 OLK-6.6 0/2] Make the rcache depot scale better
by Zhang Zekun 19 Dec '23

19 Dec '23

Make the rache depot scale better. The origin depot size is fixed, may cause problem in some condition. Robin Murphy (2): [Backport] iommu/iova: Make the rcache depot scale better [Backport] iommu/iova: Manage the depot list size drivers/iommu/iova.c | 95 ++++++++++++++++++++++++++++++-------------- 1 file changed, 65 insertions(+), 30 deletions(-) -- 2.17.1

2 3

[PATCH OLK-6.6 0/5] Terrace Service Acceleration
by Zhengchao Shao 19 Dec '23

19 Dec '23

Terrace Service Acceleration. Liu Jian (3): bpf: Add bpf_get_sockops_uid_gid helper function bpf: Add new bpf helper to get SO_ORIGINAL_DST/REPLY_SRC bpf: Access bpf_sock's src_ip4 and sorc_port in BPF_CGROUP_INET_SOCK_RELEASE hook Lu Wei (1): net: core: Add a GID field to struct sock. Wang Yufen (1): bpf, sockmap: Add sk_rmem_alloc check for sockmap include/net/netfilter/nf_conntrack.h | 6 ++ include/net/sock.h | 18 ++++++ include/uapi/linux/bpf.h | 13 +++++ include/uapi/linux/netfilter_ipv4.h | 2 + net/Kconfig | 7 +++ net/core/filter.c | 84 ++++++++++++++++++++++++++++ net/core/skmsg.c | 10 +++- net/core/sock.c | 6 ++ net/ipv4/tcp_bpf.c | 10 ++++ net/netfilter/nf_conntrack_proto.c | 71 +++++++++++++++++++++++ net/socket.c | 8 ++- tools/include/uapi/linux/bpf.h | 13 +++++ 12 files changed, 245 insertions(+), 3 deletions(-) -- 2.34.1

2 6

[PATCH OLK-6.6 v4] cache: Workaround HiSilicon Linxicore DC CVAU
by Yuan Can 19 Dec '23

19 Dec '23

From: Weilong Chen <chenweilong(a)huawei.com> ascend inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OX47 CVE: NA ------------------------------------- Linxicore's L1/L2 cache is inclusive, and the data is consistent. Any change of L1 does not require DC operation to brush CL in L1 to L2. It's safe that don't clean data cache by address to point of unification. Some of the Linxicore report wrong value to kernel. Without reporting this IDC featrue, the kernel needs to flush icache as well as dcache, causes performance degradation. Signed-off-by: Weilong Chen <chenweilong(a)huawei.com> --- Documentation/arch/arm64/silicon-errata.rst | 2 ++ arch/arm64/Kconfig | 10 +++++++ arch/arm64/include/asm/cache.h | 9 ++++++ arch/arm64/kernel/cpu_errata.c | 33 +++++++++++++++++++++ arch/arm64/tools/cpucaps | 1 + 5 files changed, 55 insertions(+) diff --git a/Documentation/arch/arm64/silicon-errata.rst b/Documentation/arch/arm64/silicon-errata.rst index bb62e0dbd546..30d3091452f9 100644 --- a/Documentation/arch/arm64/silicon-errata.rst +++ b/Documentation/arch/arm64/silicon-errata.rst @@ -205,6 +205,8 @@ stable kernels. +----------------+-----------------+-----------------+-----------------------------+ | Hisilicon | LINXICORE9100 | #162100125 | HISILICON_ERRATUM_162100125 | +----------------+-----------------+-----------------+-----------------------------+ +| Hisilicon | LINXICORE9100 | #1980005 | HISILICON_ERRATUM_1980005 | ++----------------+-----------------+-----------------+-----------------------------+ +----------------+-----------------+-----------------+-----------------------------+ | Qualcomm Tech. | Kryo/Falkor v1 | E1003 | QCOM_FALKOR_ERRATUM_1003 | +----------------+-----------------+-----------------+-----------------------------+ diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 73333648a98c..2aca373a7038 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -1170,6 +1170,16 @@ config HISILICON_ERRATUM_162100125 If unsure, say Y. +config HISILICON_ERRATUM_1980005 + bool "Hisilicon erratum 1980005: IDC support" + default n + help + Some of the Hisilicon linxicore support idc but report wrong value to + the kernel, which result to the unnecessary cache flush, leading to + performance degradation. + + If unsure, say N. + config QCOM_FALKOR_ERRATUM_1003 bool "Falkor E1003: Incorrect translation due to ASID change" default y diff --git a/arch/arm64/include/asm/cache.h b/arch/arm64/include/asm/cache.h index ceb368d33bf4..1613779be63a 100644 --- a/arch/arm64/include/asm/cache.h +++ b/arch/arm64/include/asm/cache.h @@ -112,6 +112,15 @@ int cache_line_size(void); static inline u32 __attribute_const__ read_cpuid_effective_cachetype(void) { u32 ctr = read_cpuid_cachetype(); +#ifdef CONFIG_HISILICON_ERRATUM_1980005 + static const struct midr_range idc_support_list[] = { + MIDR_ALL_VERSIONS(MIDR_HISI_TSV110), + MIDR_REV(MIDR_HISI_LINXICORE9100, 1, 0), + { /* sentinel */ } + }; + if (is_midr_in_range_list(read_cpuid_id(), idc_support_list)) + ctr |= BIT(CTR_EL0_IDC_SHIFT); +#endif if (!(ctr & BIT(CTR_EL0_IDC_SHIFT))) { u64 clidr = read_sysreg(clidr_el1); diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c index e40b76ac4711..a686a96d966a 100644 --- a/arch/arm64/kernel/cpu_errata.c +++ b/arch/arm64/kernel/cpu_errata.c @@ -60,6 +60,30 @@ is_kryo_midr(const struct arm64_cpu_capabilities *entry, int scope) return model == entry->midr_range.model; } +#ifdef CONFIG_HISILICON_ERRATUM_1980005 +static bool +hisilicon_1980005_match(const struct arm64_cpu_capabilities *entry, + int scope) +{ + static const struct midr_range idc_support_list[] = { + MIDR_ALL_VERSIONS(MIDR_HISI_TSV110), + MIDR_REV(MIDR_HISI_LINXICORE9100, 1, 0), + { /* sentinel */ } + }; + + return is_midr_in_range_list(read_cpuid_id(), idc_support_list); +} + +static void +hisilicon_1980005_enable(const struct arm64_cpu_capabilities *__unused) +{ + __set_bit(ARM64_HAS_CACHE_IDC, system_cpucaps); + arm64_ftr_reg_ctrel0.sys_val |= BIT(CTR_EL0_IDC_SHIFT); + arm64_ftr_reg_ctrel0.strict_mask &= ~BIT(CTR_EL0_IDC_SHIFT); + sysreg_clear_set(sctlr_el1, SCTLR_EL1_UCT, 0); +} +#endif + static bool has_mismatched_cache_type(const struct arm64_cpu_capabilities *entry, int scope) @@ -567,6 +591,15 @@ const struct arm64_cpu_capabilities arm64_errata[] = { ERRATA_MIDR_RANGE_LIST(hisilicon_erratum_162100125_cpus), }, #endif +#ifdef CONFIG_HISILICON_ERRATUM_1980005 + { + .desc = "Hisilicon erratum 1980005 (IDC)", + .capability = ARM64_WORKAROUND_HISILICON_1980005, + .matches = hisilicon_1980005_match, + .type = ARM64_CPUCAP_SYSTEM_FEATURE, + .cpu_enable = hisilicon_1980005_enable, + }, +#endif #ifdef CONFIG_QCOM_FALKOR_ERRATUM_1003 { .desc = "Qualcomm Technologies Falkor/Kryo erratum 1003", diff --git a/arch/arm64/tools/cpucaps b/arch/arm64/tools/cpucaps index cb4996edc09f..0b62edb91876 100644 --- a/arch/arm64/tools/cpucaps +++ b/arch/arm64/tools/cpucaps @@ -102,3 +102,4 @@ WORKAROUND_REPEAT_TLBI WORKAROUND_SPECULATIVE_AT WORKAROUND_HISILICON_ERRATUM_162100125 WORKAROUND_HISI_HIP08_RU_PREFETCH +WORKAROUND_HISILICON_1980005 -- 2.17.1

2 1

[PATCH OLK-6.6 0/3] memcg: support OOM priority for memcg
by Jinjiang Tu 19 Dec '23

19 Dec '23

support OOM priority for memcg. Jing Xiangfeng (2): memcg: support priority for oom memcg: Add sysctl memcg_qos_enable Jinjiang Tu (1): memcg: enable CONFIG_MEMCG_OOM_PROIRITY by default arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + include/linux/memcontrol.h | 26 +++ init/Kconfig | 14 ++ mm/memcontrol.c | 230 +++++++++++++++++++++++++ mm/oom_kill.c | 59 ++++++- 6 files changed, 326 insertions(+), 5 deletions(-) -- 2.25.1

2 4

[PATCH OLK-6.6 0/3] memcg: support OOM priority for memcg
by Jinjiang Tu 19 Dec '23

19 Dec '23

Support OOM priority for memcg. Jing Xiangfeng (2): memcg: support priority for oom memcg: Add sysctl memcg_qos_enable memcg: enable CONFIG_MEMCG_OOM_PROIRITY by default arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + include/linux/memcontrol.h | 26 +++ init/Kconfig | 14 ++ mm/memcontrol.c | 230 +++++++++++++++++++++++++ mm/oom_kill.c | 59 ++++++- 6 files changed, 326 insertions(+), 5 deletions(-) -- 2.25.1

2 4

[PATCH OLK-6.6 0/2] Make the rcache depot scale better
by Zhang Zekun 19 Dec '23

19 Dec '23

Make the rache depot scale better. The origin depot size is fixed, may cause problem in some condition. Robin Murphy (2): [Backport] iommu/iova: Make the rcache depot scale better [Backport] iommu/iova: Manage the depot list size drivers/iommu/iova.c | 95 ++++++++++++++++++++++++++++++-------------- 1 file changed, 65 insertions(+), 30 deletions(-) -- 2.17.1

2 3

[PATCH OLK-6.6 0/9] Support dynamic affinity scheduler
by Hui Tang 19 Dec '23

19 Dec '23

patch 1->7: Introduce dynamic affinity for cfs scheduler. patch 8->9: Tiny bugfix. Hui Tang (7): sched: Introduce dynamic affinity for cfs scheduler cpuset: Introduce new interface for scheduler dynamic affinity sched: Adjust wakeup cpu range according CPU util dynamicly sched: Adjust cpu allowed in load balance dynamicly sched: Add statistics for scheduler dynamic affinity sched: Add cmdline for dynamic affinity config: enable CONFIG_QOS_SCHED_DYNAMIC_AFFINITY by default zhangwei123171 (2): sched/fair: Remove invalid cpu selection logic in dynamic affinity sched/fair: Modify idle cpu judgment in dynamic affinity arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/proc/base.c | 73 +++++++++++ include/linux/sched.h | 22 ++++ include/linux/sched/sysctl.h | 3 + init/Kconfig | 10 ++ init/init_task.c | 3 + kernel/cgroup/cpuset.c | 153 +++++++++++++++++++++- kernel/fork.c | 16 +++ kernel/sched/core.c | 98 ++++++++++++++ kernel/sched/debug.c | 6 + kernel/sched/fair.c | 170 +++++++++++++++++++++++++ kernel/sched/features.h | 7 + kernel/sysctl.c | 11 ++ 14 files changed, 573 insertions(+), 1 deletion(-) -- 2.34.1

2 10

[PATCH 1/2] [Backport] iommu/iova: Make the rcache depot scale better
by Zhang Zekun 19 Dec '23

19 Dec '23

From: Robin Murphy <robin.murphy(a)arm.com> mainline inclusion from mainline-v6.6-rc4 commit 911aa1245da83ff5e76d33bb612d8b5a3f2ec4a5 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8PJ8L Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------------- The algorithm in the original paper specifies the storage of full magazines in the depot as an unbounded list rather than a fixed-size array. It turns out to be pretty straightforward to do this in our implementation with no significant loss of efficiency. This allows the depot to scale up to the working set sizes of larger systems, while also potentially saving some memory on smaller ones too. Since this involves touching struct iova_magazine with the requisite care, we may as well reinforce the comment with a proper assertion too. Reviewed-by: John Garry <john.g.garry(a)oracle.com> Reviewed-by: Jerry Snitselaar <jsnitsel(a)redhat.com> Signed-off-by: Robin Murphy <robin.murphy(a)arm.com> Link: https://lore.kernel.org/r/f597aa72fc3e1d315bc4574af0ce0ebe5c31cd22.16945355… Signed-off-by: Joerg Roedel <jroedel(a)suse.de> Signed-off-by: Zhang Zekun <zhangzekun11(a)huawei.com> --- drivers/iommu/iova.c | 65 ++++++++++++++++++++++++-------------------- 1 file changed, 35 insertions(+), 30 deletions(-) diff --git a/drivers/iommu/iova.c b/drivers/iommu/iova.c index 10b964600948..dd2309e9a6c5 100644 --- a/drivers/iommu/iova.c +++ b/drivers/iommu/iova.c @@ -622,15 +622,19 @@ EXPORT_SYMBOL_GPL(reserve_iova); /* * As kmalloc's buffer size is fixed to power of 2, 127 is chosen to * assure size of 'iova_magazine' to be 1024 bytes, so that no memory - * will be wasted. + * will be wasted. Since only full magazines are inserted into the depot, + * we don't need to waste PFN capacity on a separate list head either. */ #define IOVA_MAG_SIZE 127 -#define MAX_GLOBAL_MAGS 32 /* magazines per bin */ struct iova_magazine { - unsigned long size; + union { + unsigned long size; + struct iova_magazine *next; + }; unsigned long pfns[IOVA_MAG_SIZE]; }; +static_assert(!(sizeof(struct iova_magazine) & (sizeof(struct iova_magazine) - 1))); struct iova_cpu_rcache { spinlock_t lock; @@ -640,8 +644,7 @@ struct iova_cpu_rcache { struct iova_rcache { spinlock_t lock; - unsigned long depot_size; - struct iova_magazine *depot[MAX_GLOBAL_MAGS]; + struct iova_magazine *depot; struct iova_cpu_rcache __percpu *cpu_rcaches; }; @@ -717,6 +720,21 @@ static void iova_magazine_push(struct iova_magazine *mag, unsigned long pfn) mag->pfns[mag->size++] = pfn; } +static struct iova_magazine *iova_depot_pop(struct iova_rcache *rcache) +{ + struct iova_magazine *mag = rcache->depot; + + rcache->depot = mag->next; + mag->size = IOVA_MAG_SIZE; + return mag; +} + +static void iova_depot_push(struct iova_rcache *rcache, struct iova_magazine *mag) +{ + mag->next = rcache->depot; + rcache->depot = mag; +} + int iova_domain_init_rcaches(struct iova_domain *iovad) { unsigned int cpu; @@ -734,7 +752,6 @@ int iova_domain_init_rcaches(struct iova_domain *iovad) rcache = &iovad->rcaches[i]; spin_lock_init(&rcache->lock); - rcache->depot_size = 0; rcache->cpu_rcaches = __alloc_percpu(sizeof(*cpu_rcache), cache_line_size()); if (!rcache->cpu_rcaches) { @@ -776,7 +793,6 @@ static bool __iova_rcache_insert(struct iova_domain *iovad, struct iova_rcache *rcache, unsigned long iova_pfn) { - struct iova_magazine *mag_to_free = NULL; struct iova_cpu_rcache *cpu_rcache; bool can_insert = false; unsigned long flags; @@ -794,12 +810,7 @@ static bool __iova_rcache_insert(struct iova_domain *iovad, if (new_mag) { spin_lock(&rcache->lock); - if (rcache->depot_size < MAX_GLOBAL_MAGS) { - rcache->depot[rcache->depot_size++] = - cpu_rcache->loaded; - } else { - mag_to_free = cpu_rcache->loaded; - } + iova_depot_push(rcache, cpu_rcache->loaded); spin_unlock(&rcache->lock); cpu_rcache->loaded = new_mag; @@ -812,11 +823,6 @@ static bool __iova_rcache_insert(struct iova_domain *iovad, spin_unlock_irqrestore(&cpu_rcache->lock, flags); - if (mag_to_free) { - iova_magazine_free_pfns(mag_to_free, iovad); - iova_magazine_free(mag_to_free); - } - return can_insert; } @@ -854,9 +860,9 @@ static unsigned long __iova_rcache_get(struct iova_rcache *rcache, has_pfn = true; } else { spin_lock(&rcache->lock); - if (rcache->depot_size > 0) { + if (rcache->depot) { iova_magazine_free(cpu_rcache->loaded); - cpu_rcache->loaded = rcache->depot[--rcache->depot_size]; + cpu_rcache->loaded = iova_depot_pop(rcache); has_pfn = true; } spin_unlock(&rcache->lock); @@ -895,9 +901,8 @@ static void free_iova_rcaches(struct iova_domain *iovad) struct iova_rcache *rcache; struct iova_cpu_rcache *cpu_rcache; unsigned int cpu; - int i, j; - for (i = 0; i < IOVA_RANGE_CACHE_MAX_SIZE; ++i) { + for (int i = 0; i < IOVA_RANGE_CACHE_MAX_SIZE; ++i) { rcache = &iovad->rcaches[i]; if (!rcache->cpu_rcaches) break; @@ -907,8 +912,8 @@ static void free_iova_rcaches(struct iova_domain *iovad) iova_magazine_free(cpu_rcache->prev); } free_percpu(rcache->cpu_rcaches); - for (j = 0; j < rcache->depot_size; ++j) - iova_magazine_free(rcache->depot[j]); + while (rcache->depot) + iova_magazine_free(iova_depot_pop(rcache)); } kfree(iovad->rcaches); @@ -942,16 +947,16 @@ static void free_global_cached_iovas(struct iova_domain *iovad) { struct iova_rcache *rcache; unsigned long flags; - int i, j; - for (i = 0; i < IOVA_RANGE_CACHE_MAX_SIZE; ++i) { + for (int i = 0; i < IOVA_RANGE_CACHE_MAX_SIZE; ++i) { rcache = &iovad->rcaches[i]; spin_lock_irqsave(&rcache->lock, flags); - for (j = 0; j < rcache->depot_size; ++j) { - iova_magazine_free_pfns(rcache->depot[j], iovad); - iova_magazine_free(rcache->depot[j]); + while (rcache->depot) { + struct iova_magazine *mag = iova_depot_pop(rcache); + + iova_magazine_free_pfns(mag, iovad); + iova_magazine_free(mag); } - rcache->depot_size = 0; spin_unlock_irqrestore(&rcache->lock, flags); } } -- 2.17.1

1 1

[PATCH OLK-6.6 v3 0/8] kaslr: ppc64: Introduce KASLR for PPC64
by GUO Zihua 19 Dec '23

19 Dec '23

This patchset introduces KASLR for PowerPC64 chips. v3: Fixed one more commit message. v2: Replaced bugzilla in commit messages. GUO Zihua (2): powerpc/fsl_booke/kaslr: Provide correct r5 value for relocated kernel powerpc/fsl_booke/kaslr: Fix preserved memory size for int-vectors issue Jason Yan (6): powerpc/fsl_booke/kaslr: refactor kaslr_legal_offset() and kaslr_early_init() powerpc/fsl_booke/64: introduce reloc_kernel_entry() helper powerpc/fsl_booke/64: implement KASLR for fsl_booke64 powerpc/fsl_booke/64: do not clear the BSS for the second pass powerpc/fsl_booke/64: clear the original kernel if randomized powerpc/fsl_booke/kaslr: rename kaslr-booke32.rst to kaslr-booke.rst and add 64bit part Documentation/powerpc/index.rst | 2 +- .../{kaslr-booke32.rst => kaslr-booke.rst} | 35 ++++++- arch/powerpc/Kconfig | 5 +- arch/powerpc/kernel/exceptions-64e.S | 27 ++++++ arch/powerpc/kernel/head_64.S | 22 +++++ arch/powerpc/kernel/prom.c | 8 +- arch/powerpc/kernel/setup_64.c | 3 + arch/powerpc/mm/mmu_decl.h | 25 ++--- arch/powerpc/mm/nohash/kaslr_booke.c | 91 +++++++++++++------ 9 files changed, 169 insertions(+), 49 deletions(-) rename Documentation/powerpc/{kaslr-booke32.rst => kaslr-booke.rst} (59%) -- 2.34.1

2 9

[PATCH OLK-6.6 00/19] Support basic functions of kernel livepatch
by Zheng Yejian 19 Dec '23

19 Dec '23

Port patches from v5.10 to support basic functions of kernel livepatch on arch x86_64/arm64/arm32/ppc32. Zheng Yejian (19): livepatch/core: Allow implementation without ftrace livepatch/core: Reuse common codes in the solution without ftrace Revert "x86/insn: Make insn_complete() static" livepatch/x86: Support livepatch without ftrace livepatch/core: Disable support for replacing livepatch/core: Restrict livepatch patched/unpatched when plant kprobe livepatch/core: Support load and unload hooks livepatch: samples: Adapt livepatch-sample for solution without ftrace livepatch/core: Support jump_label livepatch: Fix crash when access the global variable in hook livepatch: Fix patching functions which have static_call livepatch/core: Avoid conflict with static {call,key} livepatch/arm64: Support livepatch without ftrace livepatch/core: Revert module_enable_ro and module_disable_ro livepatch: Enable livepatch configs in openeuler_defconfig arm/module: Use plt section indices for relocations livepatch/core: Add support for arm for klp relocation livepatch/arm: Support livepatch without ftrace livepatch/ppc32: Support livepatch without ftrace Documentation/filesystems/proc.rst | 2 +- arch/arm/Kconfig | 3 + arch/arm/include/asm/livepatch.h | 51 + arch/arm/include/asm/module.h | 4 +- arch/arm/kernel/Makefile | 1 + arch/arm/kernel/ftrace.c | 4 +- arch/arm/kernel/livepatch.c | 256 ++++ arch/arm/kernel/module-plts.c | 22 +- arch/arm/kernel/module.c | 4 +- arch/arm64/Kconfig | 3 + arch/arm64/configs/openeuler_defconfig | 12 + arch/arm64/include/asm/livepatch.h | 48 + arch/arm64/kernel/Makefile | 1 + arch/arm64/kernel/livepatch.c | 225 ++++ arch/powerpc/Kconfig | 5 +- arch/powerpc/include/asm/livepatch.h | 24 + arch/powerpc/kernel/Makefile | 1 + arch/powerpc/kernel/livepatch.c | 282 +++++ arch/powerpc/kernel/livepatch_32.c | 124 ++ arch/s390/Kconfig | 2 +- arch/s390/configs/debug_defconfig | 1 + arch/s390/configs/defconfig | 1 + arch/x86/Kconfig | 3 +- arch/x86/configs/openeuler_defconfig | 12 +- arch/x86/include/asm/insn.h | 7 + arch/x86/include/asm/livepatch.h | 33 + arch/x86/kernel/Makefile | 1 + arch/x86/kernel/livepatch.c | 289 +++++ arch/x86/kernel/module.c | 2 +- arch/x86/lib/insn.c | 7 - include/linux/jump_label.h | 10 + include/linux/livepatch.h | 99 +- include/linux/livepatch_sched.h | 6 +- include/linux/module.h | 33 + include/linux/moduleloader.h | 4 +- include/linux/static_call.h | 6 + kernel/jump_label.c | 22 + kernel/livepatch/Kconfig | 78 +- kernel/livepatch/Makefile | 3 +- kernel/livepatch/core.c | 1242 +++++++++++++++++++- kernel/livepatch/core.h | 4 + kernel/module/main.c | 10 +- kernel/module/strict_rwx.c | 17 + kernel/static_call_inline.c | 20 + lib/Kconfig.debug | 2 +- samples/livepatch/Makefile | 2 + samples/livepatch/livepatch-sample.c | 36 + tools/arch/x86/include/asm/insn.h | 7 + tools/arch/x86/lib/insn.c | 7 - tools/testing/selftests/bpf/config.aarch64 | 1 + tools/testing/selftests/bpf/config.s390x | 1 + tools/testing/selftests/livepatch/README | 1 + tools/testing/selftests/livepatch/config | 1 + 53 files changed, 2985 insertions(+), 57 deletions(-) create mode 100644 arch/arm/include/asm/livepatch.h create mode 100644 arch/arm/kernel/livepatch.c create mode 100644 arch/arm64/include/asm/livepatch.h create mode 100644 arch/arm64/kernel/livepatch.c create mode 100644 arch/powerpc/kernel/livepatch.c create mode 100644 arch/powerpc/kernel/livepatch_32.c create mode 100644 arch/x86/include/asm/livepatch.h create mode 100644 arch/x86/kernel/livepatch.c -- 2.25.1

2 20

[PATCH OLK-6.6 v2 0/8] kaslr: ppc64: Introduce KASLR for PPC64
by GUO Zihua 19 Dec '23

19 Dec '23

This patchset introduces KASLR for PowerPC64 chips. v2: Replaced bugzilla in commit messages. GUO Zihua (2): powerpc/fsl_booke/kaslr: Provide correct r5 value for relocated kernel powerpc/fsl_booke/kaslr: Fix preserved memory size for int-vectors issue Jason Yan (6): powerpc/fsl_booke/kaslr: refactor kaslr_legal_offset() and kaslr_early_init() powerpc/fsl_booke/64: introduce reloc_kernel_entry() helper powerpc/fsl_booke/64: implement KASLR for fsl_booke64 powerpc/fsl_booke/64: do not clear the BSS for the second pass powerpc/fsl_booke/64: clear the original kernel if randomized powerpc/fsl_booke/kaslr: rename kaslr-booke32.rst to kaslr-booke.rst and add 64bit part Documentation/powerpc/index.rst | 2 +- .../{kaslr-booke32.rst => kaslr-booke.rst} | 35 ++++++- arch/powerpc/Kconfig | 5 +- arch/powerpc/kernel/exceptions-64e.S | 27 ++++++ arch/powerpc/kernel/head_64.S | 22 +++++ arch/powerpc/kernel/prom.c | 8 +- arch/powerpc/kernel/setup_64.c | 3 + arch/powerpc/mm/mmu_decl.h | 25 ++--- arch/powerpc/mm/nohash/kaslr_booke.c | 91 +++++++++++++------ 9 files changed, 169 insertions(+), 49 deletions(-) rename Documentation/powerpc/{kaslr-booke32.rst => kaslr-booke.rst} (59%) -- 2.34.1

2 9

[PATCH OLK-6.6 v6 0/3] add redis sockmap sample code
by Zhengchao Shao 19 Dec '23

19 Dec '23

add redis sockmap sample code. Liu Jian (3): net: let sockops can use bpf_get_current_comm() net: add local_skb parameter to identify local tcp connection tools: add sample sockmap code for redis include/linux/filter.h | 3 + include/linux/skbuff.h | 5 + include/uapi/linux/bpf.h | 1 + net/Kconfig | 6 + net/core/filter.c | 18 ++ net/ipv4/tcp_input.c | 6 +- net/ipv4/tcp_output.c | 9 + tools/include/uapi/linux/bpf.h | 1 + tools/netacc/Makefile | 24 +++ tools/netacc/bpf_sockmap.h | 167 +++++++++++++++++++ tools/netacc/net-acc | 45 +++++ tools/netacc/netacc.c | 296 +++++++++++++++++++++++++++++++++ tools/netacc/netaccsockmap.c | 160 ++++++++++++++++++ 13 files changed, 740 insertions(+), 1 deletion(-) create mode 100644 tools/netacc/Makefile create mode 100644 tools/netacc/bpf_sockmap.h create mode 100755 tools/netacc/net-acc create mode 100644 tools/netacc/netacc.c create mode 100644 tools/netacc/netaccsockmap.c -- 2.34.1

2 4

[PATCH OLK-5.10] tick/broadcast-hrtimer: Prevent the timer device on broadcast duty CPU from being disabled
by Yu Liao 19 Dec '23

19 Dec '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8PL17 CVE: NA ---------------------------------------- It was found that running the LTP hotplug stress test on a aarch64 system could produce rcu_sched stall warnings. The issue is the following: CPU1 (owns the broadcast hrtimer) CPU2 tick_broadcast_enter() //shut down local timer device ... tick_broadcast_exit() //exits with tick_broadcast_force_mask set, timer device remains disabled initiates offlining of CPU1 take_cpu_down() //CPU1 shuts down and does not send broadcast IPI anymore takedown_cpu() hotplug_cpu__broadcast_tick_pull() //move broadcast hrtimer to this CPU clockevents_program_event() bc_set_next() hrtimer_start() //does not call hrtimer_reprogram() to program timer device if expires equals dev->next_event, so the timer device remains disabled. CPU2 takes over the broadcast duty but local timer device is disabled, causing many CPUs to become stuck. Fix this by calling tick_program_event() to reprogram the local timer device in this scenario. Signed-off-by: Yu Liao <liaoyu15(a)huawei.com> --- kernel/time/tick-broadcast-hrtimer.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/kernel/time/tick-broadcast-hrtimer.c b/kernel/time/tick-broadcast-hrtimer.c index b5a65e212df2..9cc9e6703694 100644 --- a/kernel/time/tick-broadcast-hrtimer.c +++ b/kernel/time/tick-broadcast-hrtimer.c @@ -42,6 +42,8 @@ static int bc_shutdown(struct clock_event_device *evt) */ static int bc_set_next(ktime_t expires, struct clock_event_device *bc) { + ktime_t next_event; + /* * This is called either from enter/exit idle code or from the * broadcast handler. In all cases tick_broadcast_lock is held. @@ -62,6 +64,19 @@ static int bc_set_next(ktime_t expires, struct clock_event_device *bc) * hrtimer_start() can call into tracing. */ RCU_NONIDLE( { + next_event = this_cpu_ptr(&tick_cpu_device)->evtdev->next_event; + + /* + * This can be called from CPU offline operation to move broadcast + * assignment. If tick_broadcast_force_mask is set, the CPU local + * timer device may be disabled. And hrtimer_reprogram() will not + * called if the timer is not the first expiring timer. Reprogram + * the cpu local timer device to ensure we can take over the + * broadcast duty. + */ + if (tick_check_broadcast_expired() && expires >= next_event) + tick_program_event(next_event, 1); + hrtimer_start(&bctimer, expires, HRTIMER_MODE_ABS_PINNED_HARD); /* * The core tick broadcast mode expects bc->bound_on to be set -- 2.33.0

2 1

[PATCH OLK-6.6 0/1] Support priority load balance for qos scheduler
by Song Zhang 19 Dec '23

19 Dec '23

Song Zhang (1): sched: Introduce priority load balance for qos scheduler arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + include/linux/sched/sysctl.h | 4 ++ init/Kconfig | 9 +++ kernel/sched/core.c | 3 + kernel/sched/fair.c | 90 +++++++++++++++++++++++++- kernel/sched/sched.h | 3 + 7 files changed, 110 insertions(+), 1 deletion(-) -- 2.33.0

2 2

[PATCH OLK-6.6 0/1] Support priority load balance for qos scheduler
by Song Zhang 19 Dec '23

19 Dec '23

Song Zhang (1): sched: Introduce priority load balance for qos scheduler arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + include/linux/sched/sysctl.h | 4 ++ init/Kconfig | 9 +++ kernel/sched/core.c | 3 + kernel/sched/fair.c | 90 +++++++++++++++++++++++++- kernel/sched/sched.h | 3 + 7 files changed, 110 insertions(+), 1 deletion(-) -- 2.33.0

2 2

[PATCH OLK-6.6 0/3] Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely"
by Yuntao Liu 19 Dec '23

19 Dec '23

*** BLURB HERE *** Guo Xuenan (3): Revert "compiler: remove CONFIG_OPTIMIZE_INLINING entirely" disable OPTIMIZE_INLINING by default make OPTIMIZE_INLINING config editable arch/arm64/kvm/sys_regs.h | 5 +++++ arch/x86/configs/i386_defconfig | 1 + arch/x86/configs/x86_64_defconfig | 1 + .../pci/hive_isp_css_include/print_support.h | 4 ++++ include/linux/compiler_types.h | 9 ++++++++- kernel/configs/tiny.config | 1 + lib/Kconfig.debug | 13 +++++++++++++ 7 files changed, 33 insertions(+), 1 deletion(-) -- 2.34.1

2 4

[PATCH OLK-5.10] livepatch/core: Disable support for replacing
by Zheng Yejian 19 Dec '23

19 Dec '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8ORHJ -------------------------------- Replacing is an optimization for livepatching based on ftrace, the wo_ftrace solution do not support it. Signed-off-by: Zheng Yejian <zhengyejian1(a)huawei.com> --- kernel/livepatch/core.c | 10 ++++++++++ kernel/livepatch/patch.c | 2 ++ 2 files changed, 12 insertions(+) diff --git a/kernel/livepatch/core.c b/kernel/livepatch/core.c index 8768ec1bddf3..7288456906b6 100644 --- a/kernel/livepatch/core.c +++ b/kernel/livepatch/core.c @@ -1156,6 +1156,9 @@ static void klp_init_func_early(struct klp_object *obj, kobject_init(&func->kobj, &klp_ktype_func); list_add_tail(&func->node, &obj->func_list); func->func_node = NULL; +#ifdef CONFIG_LIVEPATCH_STOP_MACHINE_CONSISTENCY + func->nop = false; +#endif } static void klp_init_object_early(struct klp_patch *patch, @@ -1166,6 +1169,7 @@ static void klp_init_object_early(struct klp_patch *patch, list_add_tail(&obj->node, &patch->obj_list); #ifdef CONFIG_LIVEPATCH_STOP_MACHINE_CONSISTENCY obj->mod = NULL; + obj->dynamic = false; #endif } @@ -1245,6 +1249,12 @@ static int klp_init_patch(struct klp_patch *patch) if (ret) return ret; +#ifdef CONFIG_LIVEPATCH_STOP_MACHINE_CONSISTENCY + if (patch->replace) { + pr_err("Replacing is not supported\n"); + return -EINVAL; + } +#endif if (patch->replace) { ret = klp_add_nops(patch); if (ret) diff --git a/kernel/livepatch/patch.c b/kernel/livepatch/patch.c index bea6c5d0af94..df0c487a04fd 100644 --- a/kernel/livepatch/patch.c +++ b/kernel/livepatch/patch.c @@ -315,7 +315,9 @@ int klp_patch_object(struct klp_object *obj, bool rollback) if (obj->patched) return 0; + WARN_ON(obj->dynamic); klp_for_each_func(obj, func) { + WARN_ON(func->nop); ret = klp_patch_func(func); if (ret && klp_need_rollback(ret, rollback)) { klp_unpatch_object(obj); -- 2.25.1

2 1

[PATCH OLK-6.6 0/1] Support priority load balance for qos scheduler
by Song Zhang 19 Dec '23

19 Dec '23

Song Zhang (1): sched: Introduce priority load balance for qos scheduler arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + include/linux/sched/sysctl.h | 4 ++ init/Kconfig | 9 +++ kernel/sched/core.c | 3 + kernel/sched/fair.c | 90 +++++++++++++++++++++++++- kernel/sched/sched.h | 3 + 7 files changed, 110 insertions(+), 1 deletion(-) -- 2.33.0

2 2

[PATCH OLK-6.6 0/1] Support priority load balance for qos scheduler
by Song Zhang 19 Dec '23

19 Dec '23

Song Zhang (1): sched: Introduce priority load balance for qos scheduler arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + include/linux/sched/sysctl.h | 4 ++ init/Kconfig | 9 +++ kernel/sched/core.c | 3 + kernel/sched/fair.c | 90 +++++++++++++++++++++++++- kernel/sched/sched.h | 3 + 7 files changed, 110 insertions(+), 1 deletion(-) -- 2.33.0

2 2

[PATCH OLK-5.10 0/2] Support node_reclaim_distance adjust
by Hui Tang 19 Dec '23

19 Dec '23

Allow arch to adjust node_reclaim_distance. Hui Tang (2): arm64/numa: Support node_reclaim_distance adjust for arch config: enable COBFIG_ARCH_CUSTOM_NUMA_DISTANCE for arm64 arch/arm64/Kconfig | 13 +++++ arch/arm64/configs/openeuler_defconfig | 2 + arch/arm64/mm/numa.c | 79 ++++++++++++++++++++++++++ 3 files changed, 94 insertions(+) -- 2.34.1

2 3

[PATCH OLK-5.10 0/2] Support node_reclaim_distance adjust
by Hui Tang 19 Dec '23

19 Dec '23

Allow arch to adjust node_reclaim_distance. Hui Tang (2): arm64/numa: Support node_reclaim_distance adjust for arch config: enable COBFIG_ARCH_CUSTOM_NUMA_DISTANCE for arm64 arch/arm64/Kconfig | 13 +++++ arch/arm64/configs/openeuler_defconfig | 2 + arch/arm64/mm/numa.c | 80 ++++++++++++++++++++++++++ 3 files changed, 95 insertions(+) -- 2.34.1

2 3

[PATCH v3 OLK-5.10 0/1] Revert "mm/mempolicy: don't handle MPOL_LOCA
by Wupeng Ma 18 Dec '23

18 Dec '23

From: Ma Wupeng <mawupeng1(a)huawei.com> Revert to bring MPOL_LOCAL back. changelog since v2: - introduce mpol_rebind_preferred changelog since v1: - bugfix in __mpol_equal Ma Wupeng (1): Revert "mm/mempolicy: don't handle MPOL_LOCAL like a fake MPOL_PREFERRED policy" include/uapi/linux/mempolicy.h | 1 + mm/mempolicy.c | 142 ++++++++++++++++++++------------- 2 files changed, 87 insertions(+), 56 deletions(-) -- 2.25.1

2 2

[PATCH v2 OLK-5.10 0/1] Revert "mm/mempolicy: don't handle MPOL_LOCA
by Wupeng Ma 18 Dec '23

18 Dec '23

From: Ma Wupeng <mawupeng1(a)huawei.com> Revert to bring MPOL_LOCAL back. changelog since v1: - bugfix in __mpol_equal Ma Wupeng (1): Revert "mm/mempolicy: don't handle MPOL_LOCAL like a fake MPOL_PREFERRED policy" include/uapi/linux/mempolicy.h | 1 + mm/mempolicy.c | 138 +++++++++++++++++++-------------- 2 files changed, 81 insertions(+), 58 deletions(-) -- 2.25.1

2 2

[PATCH v2 OLK-5.10 0/1] Revert "mm/mempolicy: don't handle MPOL_LOCA
by Wupeng Ma 18 Dec '23

18 Dec '23

From: Ma Wupeng <mawupeng1(a)huawei.com> Revert to bring MPOL_LOCAL back. changelog since v1: - bugfix in __mpol_equal Ma Wupeng (1): Revert "mm/mempolicy: don't handle MPOL_LOCAL like a fake MPOL_PREFERRED policy" include/uapi/linux/mempolicy.h | 1 + mm/mempolicy.c | 138 +++++++++++++++++++-------------- 2 files changed, 81 insertions(+), 58 deletions(-) -- 2.25.1

2 4

[PATCH OLK-5.10 0/1] Revert "mm/mempolicy: don't handle MPOL_LOCA
by Wupeng Ma 18 Dec '23

18 Dec '23

From: Ma Wupeng <mawupeng1(a)huawei.com> Revert to bring MPOL_LOCAL back. Ma Wupeng (1): Revert "mm/mempolicy: don't handle MPOL_LOCAL like a fake MPOL_PREFERRED policy" include/uapi/linux/mempolicy.h | 1 + mm/mempolicy.c | 134 +++++++++++++++++++-------------- 2 files changed, 80 insertions(+), 55 deletions(-) -- 2.25.1

2 2

[PATCH OLK-6.6 0/3] xfs: fix attr inactive problems
by Long Li 18 Dec '23

18 Dec '23

This patch set fix two attr inactive problems: Guo Xuenan (1): xfs: force shutdown xfs when xfs_attr_inactive fails Zhang Yi (2): xfs: factor out __xfs_da3_node_read() xfs: atomic drop extent entries when inactiving attr fs/xfs/libxfs/xfs_da_btree.c | 5 +-- fs/xfs/libxfs/xfs_da_btree.h | 15 ++++++-- fs/xfs/xfs_attr_inactive.c | 66 ++++++++++++++++++++++++++---------- 3 files changed, 65 insertions(+), 21 deletions(-) -- 2.31.1

2 4

[PATCH OLK-5.10] sched: Support custom numa distance for schedluler
by Hui Tang 18 Dec '23

18 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8PG0C CVE: NA -------------------------------- Support arch redefine far numa distance, which affects the initialization of sched domain, such as SD_WAKE_AFFINE. Signed-off-by: Hui Tang <tanghui20(a)huawei.com> --- arch/arm64/Kconfig | 5 +++++ arch/arm64/include/asm/numa.h | 1 + arch/arm64/mm/numa.c | 40 +++++++++++++++++++++++++++++++++++ kernel/sched/topology.c | 9 +++++++- 4 files changed, 54 insertions(+), 1 deletion(-) diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index c5eca65c27fc..88d9862aa139 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -2102,6 +2102,11 @@ config STACKPROTECTOR_PER_TASK def_bool y depends on STACKPROTECTOR && CC_HAVE_STACKPROTECTOR_SYSREG +config ARCH_CUSTOM_NUMA_DISTANCE + bool "Support custom far numa distance" + depends on ARCH_HISI + default n + menuconfig ASCEND_FEATURES bool "Support Ascend Features" depends on ARM64 diff --git a/arch/arm64/include/asm/numa.h b/arch/arm64/include/asm/numa.h index 43bfff72a32f..3af06fa3e969 100644 --- a/arch/arm64/include/asm/numa.h +++ b/arch/arm64/include/asm/numa.h @@ -43,6 +43,7 @@ void __init early_map_cpu_to_node(unsigned int cpu, int nid); void numa_store_cpu_info(unsigned int cpu); void numa_add_cpu(unsigned int cpu); void numa_remove_cpu(unsigned int cpu); +bool arch_numa_distance_is_far(int distance, int numa_levels); #else /* CONFIG_NUMA */ diff --git a/arch/arm64/mm/numa.c b/arch/arm64/mm/numa.c index 480fcfb1b917..871394038ae0 100644 --- a/arch/arm64/mm/numa.c +++ b/arch/arm64/mm/numa.c @@ -619,6 +619,46 @@ static int __init dummy_numa_init(void) return 0; } +#ifdef CONFIG_ARCH_CUSTOM_NUMA_DISTANCE +static int arch_custom_numa_distance_flag = 1; +static int __init arch_custom_numa_distance_flag_setup(char *str) +{ + int val; + + if (kstrtoint(str, 0, &val)) + pr_warn("Unable to set arch_custom_numa_distance\n"); + + if (val != 0 || val != 1) + return 1; + + arch_custom_numa_distance_flag = val; + return 1; +} +early_param("arch_custom_numa_distance_flag=", arch_custom_numa_distance_flag_setup); + +bool arch_numa_distance_is_far(int distance, int numa_levels) +{ + unsigned int model = read_cpuid_id() & MIDR_CPU_MODEL_MASK; + int far_numa_distance = node_reclaim_distance; + + if (arch_custom_numa_distance_flag == 0) + return distance > node_reclaim_distance; + + switch (model) { + case MIDR_HISI_TSV200: + if (nr_node_ids != 4 || numa_levels <= 3) + break; + + far_numa_distance = 40; + break; + default: + break; + } + + return distance > far_numa_distance; +} +#endif + /** * arm64_numa_init() - Initialize NUMA * diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c index e7413d6dd75b..e0fd4e68616c 100644 --- a/kernel/sched/topology.c +++ b/kernel/sched/topology.c @@ -1529,6 +1529,12 @@ int sched_max_numa_distance; static int *sched_domains_numa_distance; static struct cpumask ***sched_domains_numa_masks; int __read_mostly node_reclaim_distance = RECLAIM_DISTANCE; + +#ifndef CONFIG_ARCH_CUSTOM_NUMA_DISTANCE +bool arch_numa_distance_is_far(int distance, int numa_levels) { + return distance > node_reclaim_distance; +} +#endif #endif /* @@ -1640,7 +1646,8 @@ sd_init(struct sched_domain_topology_level *tl, sd->flags &= ~SD_PREFER_SIBLING; sd->flags |= SD_SERIALIZE; - if (sched_domains_numa_distance[tl->numa_level] > node_reclaim_distance) { + if (arch_numa_distance_is_far(tl->numa_level, + sched_domains_numa_levels)) { sd->flags &= ~(SD_BALANCE_EXEC | SD_BALANCE_FORK | SD_WAKE_AFFINE); -- 2.34.1

2 1

[PATCH] sched: Support custom numa distance for schedluler
by Hui Tang 18 Dec '23

18 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8PG0C CVE: NA -------------------------------- Support arch redefine far numa distance, which affects the initialization of sched domain, such as SD_WAKE_AFFINE. Signed-off-by: Hui Tang <tanghui20(a)huawei.com> --- arch/arm64/Kconfig | 5 +++++ arch/arm64/include/asm/numa.h | 1 + arch/arm64/mm/numa.c | 40 +++++++++++++++++++++++++++++++++++ kernel/sched/topology.c | 9 +++++++- 4 files changed, 54 insertions(+), 1 deletion(-) diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index c5eca65c27fc..88d9862aa139 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -2102,6 +2102,11 @@ config STACKPROTECTOR_PER_TASK def_bool y depends on STACKPROTECTOR && CC_HAVE_STACKPROTECTOR_SYSREG +config ARCH_CUSTOM_NUMA_DISTANCE + bool "Support custom far numa distance" + depends on ARCH_HISI + default n + menuconfig ASCEND_FEATURES bool "Support Ascend Features" depends on ARM64 diff --git a/arch/arm64/include/asm/numa.h b/arch/arm64/include/asm/numa.h index 43bfff72a32f..3af06fa3e969 100644 --- a/arch/arm64/include/asm/numa.h +++ b/arch/arm64/include/asm/numa.h @@ -43,6 +43,7 @@ void __init early_map_cpu_to_node(unsigned int cpu, int nid); void numa_store_cpu_info(unsigned int cpu); void numa_add_cpu(unsigned int cpu); void numa_remove_cpu(unsigned int cpu); +bool arch_numa_distance_is_far(int distance, int numa_levels); #else /* CONFIG_NUMA */ diff --git a/arch/arm64/mm/numa.c b/arch/arm64/mm/numa.c index 480fcfb1b917..871394038ae0 100644 --- a/arch/arm64/mm/numa.c +++ b/arch/arm64/mm/numa.c @@ -619,6 +619,46 @@ static int __init dummy_numa_init(void) return 0; } +#ifdef CONFIG_ARCH_CUSTOM_NUMA_DISTANCE +static int arch_custom_numa_distance_flag = 1; +static int __init arch_custom_numa_distance_flag_setup(char *str) +{ + int val; + + if (kstrtoint(str, 0, &val)) + pr_warn("Unable to set arch_custom_numa_distance\n"); + + if (val != 0 || val != 1) + return 1; + + arch_custom_numa_distance_flag = val; + return 1; +} +early_param("arch_custom_numa_distance_flag=", arch_custom_numa_distance_flag_setup); + +bool arch_numa_distance_is_far(int distance, int numa_levels) +{ + unsigned int model = read_cpuid_id() & MIDR_CPU_MODEL_MASK; + int far_numa_distance = node_reclaim_distance; + + if (arch_custom_numa_distance_flag == 0) + return distance > node_reclaim_distance; + + switch (model) { + case MIDR_HISI_TSV200: + if (nr_node_ids != 4 || numa_levels <= 3) + break; + + far_numa_distance = 40; + break; + default: + break; + } + + return distance > far_numa_distance; +} +#endif + /** * arm64_numa_init() - Initialize NUMA * diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c index e7413d6dd75b..e0fd4e68616c 100644 --- a/kernel/sched/topology.c +++ b/kernel/sched/topology.c @@ -1529,6 +1529,12 @@ int sched_max_numa_distance; static int *sched_domains_numa_distance; static struct cpumask ***sched_domains_numa_masks; int __read_mostly node_reclaim_distance = RECLAIM_DISTANCE; + +#ifndef CONFIG_ARCH_CUSTOM_NUMA_DISTANCE +bool arch_numa_distance_is_far(int distance, int numa_levels) { + return distance > node_reclaim_distance; +} +#endif #endif /* @@ -1640,7 +1646,8 @@ sd_init(struct sched_domain_topology_level *tl, sd->flags &= ~SD_PREFER_SIBLING; sd->flags |= SD_SERIALIZE; - if (sched_domains_numa_distance[tl->numa_level] > node_reclaim_distance) { + if (arch_numa_distance_is_far(tl->numa_level, + sched_domains_numa_levels)) { sd->flags &= ~(SD_BALANCE_EXEC | SD_BALANCE_FORK | SD_WAKE_AFFINE); -- 2.34.1

1 0

[PATCH OLK-5.10] sched: Support custom numa distance for schedluler
by Hui Tang 18 Dec '23

18 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8PG0C CVE: NA -------------------------------- Support arch redefine far numa distance, which affects the initialization of sched domain, such as SD_WAKE_AFFINE. Signed-off-by: Hui Tang <tanghui20(a)huawei.com> --- arch/arm64/Kconfig | 5 +++++ arch/arm64/include/asm/numa.h | 1 + arch/arm64/mm/numa.c | 40 +++++++++++++++++++++++++++++++++++ kernel/sched/topology.c | 9 +++++++- 4 files changed, 54 insertions(+), 1 deletion(-) diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index c5eca65c27fc..88d9862aa139 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -2102,6 +2102,11 @@ config STACKPROTECTOR_PER_TASK def_bool y depends on STACKPROTECTOR && CC_HAVE_STACKPROTECTOR_SYSREG +config ARCH_CUSTOM_NUMA_DISTANCE + bool "Support custom far numa distance" + depends on ARCH_HISI + default n + menuconfig ASCEND_FEATURES bool "Support Ascend Features" depends on ARM64 diff --git a/arch/arm64/include/asm/numa.h b/arch/arm64/include/asm/numa.h index 43bfff72a32f..3af06fa3e969 100644 --- a/arch/arm64/include/asm/numa.h +++ b/arch/arm64/include/asm/numa.h @@ -43,6 +43,7 @@ void __init early_map_cpu_to_node(unsigned int cpu, int nid); void numa_store_cpu_info(unsigned int cpu); void numa_add_cpu(unsigned int cpu); void numa_remove_cpu(unsigned int cpu); +bool arch_numa_distance_is_far(int distance, int numa_levels); #else /* CONFIG_NUMA */ diff --git a/arch/arm64/mm/numa.c b/arch/arm64/mm/numa.c index 480fcfb1b917..871394038ae0 100644 --- a/arch/arm64/mm/numa.c +++ b/arch/arm64/mm/numa.c @@ -619,6 +619,46 @@ static int __init dummy_numa_init(void) return 0; } +#ifdef CONFIG_ARCH_CUSTOM_NUMA_DISTANCE +static int arch_custom_numa_distance_flag = 1; +static int __init arch_custom_numa_distance_flag_setup(char *str) +{ + int val; + + if (kstrtoint(str, 0, &val)) + pr_warn("Unable to set arch_custom_numa_distance\n"); + + if (val != 0 || val != 1) + return 1; + + arch_custom_numa_distance_flag = val; + return 1; +} +early_param("arch_custom_numa_distance_flag=", arch_custom_numa_distance_flag_setup); + +bool arch_numa_distance_is_far(int distance, int numa_levels) +{ + unsigned int model = read_cpuid_id() & MIDR_CPU_MODEL_MASK; + int far_numa_distance = node_reclaim_distance; + + if (arch_custom_numa_distance_flag == 0) + return distance > node_reclaim_distance; + + switch (model) { + case MIDR_HISI_TSV200: + if (nr_node_ids != 4 || numa_levels <= 3) + break; + + far_numa_distance = 40; + break; + default: + break; + } + + return distance > far_numa_distance; +} +#endif + /** * arm64_numa_init() - Initialize NUMA * diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c index e7413d6dd75b..fb6f20956ee7 100644 --- a/kernel/sched/topology.c +++ b/kernel/sched/topology.c @@ -1529,6 +1529,12 @@ int sched_max_numa_distance; static int *sched_domains_numa_distance; static struct cpumask ***sched_domains_numa_masks; int __read_mostly node_reclaim_distance = RECLAIM_DISTANCE; + +#ifdef CONFIG_ARCH_CUSTOM_NUMA_DISTANCE +static inline bool arch_numa_distance_is_far(int distance, int numa_levels) { + return distance > node_reclaim_distance; +} +#endif #endif /* @@ -1640,7 +1646,8 @@ sd_init(struct sched_domain_topology_level *tl, sd->flags &= ~SD_PREFER_SIBLING; sd->flags |= SD_SERIALIZE; - if (sched_domains_numa_distance[tl->numa_level] > node_reclaim_distance) { + if (arch_numa_distance_is_far(tl->numa_level, + sched_domains_numa_levels)) { sd->flags &= ~(SD_BALANCE_EXEC | SD_BALANCE_FORK | SD_WAKE_AFFINE); -- 2.34.1

2 1

[PATCH] sched: Support custom numa distance for schedluler
by Hui Tang 18 Dec '23

18 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8PG0C CVE: NA -------------------------------- Support arch redefine far numa distance, which affects the initialization of sched domain, such as SD_WAKE_AFFINE. Signed-off-by: Hui Tang <tanghui20(a)huawei.com> --- arch/arm64/Kconfig | 5 +++++ arch/arm64/include/asm/numa.h | 1 + arch/arm64/mm/numa.c | 40 +++++++++++++++++++++++++++++++++++ kernel/sched/topology.c | 9 +++++++- 4 files changed, 54 insertions(+), 1 deletion(-) diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index c5eca65c27fc..88d9862aa139 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -2102,6 +2102,11 @@ config STACKPROTECTOR_PER_TASK def_bool y depends on STACKPROTECTOR && CC_HAVE_STACKPROTECTOR_SYSREG +config ARCH_CUSTOM_NUMA_DISTANCE + bool "Support custom far numa distance" + depends on ARCH_HISI + default n + menuconfig ASCEND_FEATURES bool "Support Ascend Features" depends on ARM64 diff --git a/arch/arm64/include/asm/numa.h b/arch/arm64/include/asm/numa.h index 43bfff72a32f..3af06fa3e969 100644 --- a/arch/arm64/include/asm/numa.h +++ b/arch/arm64/include/asm/numa.h @@ -43,6 +43,7 @@ void __init early_map_cpu_to_node(unsigned int cpu, int nid); void numa_store_cpu_info(unsigned int cpu); void numa_add_cpu(unsigned int cpu); void numa_remove_cpu(unsigned int cpu); +bool arch_numa_distance_is_far(int distance, int numa_levels); #else /* CONFIG_NUMA */ diff --git a/arch/arm64/mm/numa.c b/arch/arm64/mm/numa.c index 480fcfb1b917..871394038ae0 100644 --- a/arch/arm64/mm/numa.c +++ b/arch/arm64/mm/numa.c @@ -619,6 +619,46 @@ static int __init dummy_numa_init(void) return 0; } +#ifdef CONFIG_ARCH_CUSTOM_NUMA_DISTANCE +static int arch_custom_numa_distance_flag = 1; +static int __init arch_custom_numa_distance_flag_setup(char *str) +{ + int val; + + if (kstrtoint(str, 0, &val)) + pr_warn("Unable to set arch_custom_numa_distance\n"); + + if (val != 0 || val != 1) + return 1; + + arch_custom_numa_distance_flag = val; + return 1; +} +early_param("arch_custom_numa_distance_flag=", arch_custom_numa_distance_flag_setup); + +bool arch_numa_distance_is_far(int distance, int numa_levels) +{ + unsigned int model = read_cpuid_id() & MIDR_CPU_MODEL_MASK; + int far_numa_distance = node_reclaim_distance; + + if (arch_custom_numa_distance_flag == 0) + return distance > node_reclaim_distance; + + switch (model) { + case MIDR_HISI_TSV200: + if (nr_node_ids != 4 || numa_levels <= 3) + break; + + far_numa_distance = 40; + break; + default: + break; + } + + return distance > far_numa_distance; +} +#endif + /** * arm64_numa_init() - Initialize NUMA * diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c index e7413d6dd75b..fb6f20956ee7 100644 --- a/kernel/sched/topology.c +++ b/kernel/sched/topology.c @@ -1529,6 +1529,12 @@ int sched_max_numa_distance; static int *sched_domains_numa_distance; static struct cpumask ***sched_domains_numa_masks; int __read_mostly node_reclaim_distance = RECLAIM_DISTANCE; + +#ifdef CONFIG_ARCH_CUSTOM_NUMA_DISTANCE +static inline bool arch_numa_distance_is_far(int distance, int numa_levels) { + return distance > node_reclaim_distance; +} +#endif #endif /* @@ -1640,7 +1646,8 @@ sd_init(struct sched_domain_topology_level *tl, sd->flags &= ~SD_PREFER_SIBLING; sd->flags |= SD_SERIALIZE; - if (sched_domains_numa_distance[tl->numa_level] > node_reclaim_distance) { + if (arch_numa_distance_is_far(tl->numa_level, + sched_domains_numa_levels)) { sd->flags &= ~(SD_BALANCE_EXEC | SD_BALANCE_FORK | SD_WAKE_AFFINE); -- 2.34.1

1 0

[PATCH OLK-6.6 v5 0/4] add redis sockmap sample code
by Zhengchao Shao 18 Dec '23

18 Dec '23

add redis sockmap sample code. Liu Jian (4): cgroup: make cgroup_bpf_prog_attach work when cgroup2 is not mounted net: let sockops can use bpf_get_current_comm() net: add local_skb parameter to identify local tcp connection tools: add sample sockmap code for redis include/linux/filter.h | 3 + include/linux/skbuff.h | 5 + include/uapi/linux/bpf.h | 1 + kernel/bpf/cgroup.c | 16 ++ net/Kconfig | 5 + net/core/filter.c | 18 ++ net/ipv4/tcp_input.c | 6 +- net/ipv4/tcp_output.c | 9 + tools/include/uapi/linux/bpf.h | 1 + tools/netacc/Makefile | 24 +++ tools/netacc/bpf_sockmap.h | 167 +++++++++++++++++++ tools/netacc/net-acc | 35 ++++ tools/netacc/netacc.c | 296 +++++++++++++++++++++++++++++++++ tools/netacc/netaccsockmap.c | 160 ++++++++++++++++++ 14 files changed, 745 insertions(+), 1 deletion(-) create mode 100644 tools/netacc/Makefile create mode 100644 tools/netacc/bpf_sockmap.h create mode 100755 tools/netacc/net-acc create mode 100644 tools/netacc/netacc.c create mode 100644 tools/netacc/netaccsockmap.c -- 2.34.1

2 5

[PATCH OLK-5.10] livepatch/core: Disable support for replacing
by Zheng Yejian 18 Dec '23

18 Dec '23

Offering: HULK hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8ORHJ -------------------------------- Replacing is an optimization for livepatching based on ftrace, the wo_ftrace solution do not support it. Signed-off-by: Zheng Yejian <zhengyejian1(a)huawei.com> --- kernel/livepatch/core.c | 10 ++++++++++ kernel/livepatch/patch.c | 2 ++ 2 files changed, 12 insertions(+) diff --git a/kernel/livepatch/core.c b/kernel/livepatch/core.c index 8768ec1bddf3..7288456906b6 100644 --- a/kernel/livepatch/core.c +++ b/kernel/livepatch/core.c @@ -1156,6 +1156,9 @@ static void klp_init_func_early(struct klp_object *obj, kobject_init(&func->kobj, &klp_ktype_func); list_add_tail(&func->node, &obj->func_list); func->func_node = NULL; +#ifdef CONFIG_LIVEPATCH_STOP_MACHINE_CONSISTENCY + func->nop = false; +#endif } static void klp_init_object_early(struct klp_patch *patch, @@ -1166,6 +1169,7 @@ static void klp_init_object_early(struct klp_patch *patch, list_add_tail(&obj->node, &patch->obj_list); #ifdef CONFIG_LIVEPATCH_STOP_MACHINE_CONSISTENCY obj->mod = NULL; + obj->dynamic = false; #endif } @@ -1245,6 +1249,12 @@ static int klp_init_patch(struct klp_patch *patch) if (ret) return ret; +#ifdef CONFIG_LIVEPATCH_STOP_MACHINE_CONSISTENCY + if (patch->replace) { + pr_err("Replacing is not supported\n"); + return -EINVAL; + } +#endif if (patch->replace) { ret = klp_add_nops(patch); if (ret) diff --git a/kernel/livepatch/patch.c b/kernel/livepatch/patch.c index bea6c5d0af94..df0c487a04fd 100644 --- a/kernel/livepatch/patch.c +++ b/kernel/livepatch/patch.c @@ -315,7 +315,9 @@ int klp_patch_object(struct klp_object *obj, bool rollback) if (obj->patched) return 0; + WARN_ON(obj->dynamic); klp_for_each_func(obj, func) { + WARN_ON(func->nop); ret = klp_patch_func(func); if (ret && klp_need_rollback(ret, rollback)) { klp_unpatch_object(obj); -- 2.25.1

2 1

[PATCH openEuler-1.0-LTS] net: Remove acked SYN flag from packet in the transmit queue correctly
by Dong Chenchen 18 Dec '23

18 Dec '23

mainline inclusion from mainline-v6.7-rc4 commit f99cd56230f56c8b6b33713c5be4da5d6766be1f category: bugfix bugzilla: 187254 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… --------------------------- syzkaller report: kernel BUG at net/core/skbuff.c:3452! invalid opcode: 0000 [#1] PREEMPT SMP KASAN PTI CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.7.0-rc4-00009-gbee0e7762ad2-dirty #135 RIP: 0010:skb_copy_and_csum_bits (net/core/skbuff.c:3452) Call Trace: icmp_glue_bits (net/ipv4/icmp.c:357) __ip_append_data.isra.0 (net/ipv4/ip_output.c:1165) ip_append_data (net/ipv4/ip_output.c:1362 net/ipv4/ip_output.c:1341) icmp_push_reply (net/ipv4/icmp.c:370) __icmp_send (./include/net/route.h:252 net/ipv4/icmp.c:772) ip_fragment.constprop.0 (./include/linux/skbuff.h:1234 net/ipv4/ip_output.c:592 net/ipv4/ip_output.c:577) __ip_finish_output (net/ipv4/ip_output.c:311 net/ipv4/ip_output.c:295) ip_output (net/ipv4/ip_output.c:427) __ip_queue_xmit (net/ipv4/ip_output.c:535) __tcp_transmit_skb (net/ipv4/tcp_output.c:1462) __tcp_retransmit_skb (net/ipv4/tcp_output.c:3387) tcp_retransmit_skb (net/ipv4/tcp_output.c:3404) tcp_retransmit_timer (net/ipv4/tcp_timer.c:604) tcp_write_timer (./include/linux/spinlock.h:391 net/ipv4/tcp_timer.c:716) The panic issue was trigered by tcp simultaneous initiation. The initiation process is as follows: TCP A TCP B 1. CLOSED CLOSED 2. SYN-SENT --> <SEQ=100><CTL=SYN> ... 3. SYN-RECEIVED <-- <SEQ=300><CTL=SYN> <-- SYN-SENT 4. ... <SEQ=100><CTL=SYN> --> SYN-RECEIVED 5. SYN-RECEIVED --> <SEQ=100><ACK=301><CTL=SYN,ACK> ... // TCP B: not send challenge ack for ack limit or packet loss // TCP A: close tcp_close tcp_send_fin if (!tskb && tcp_under_memory_pressure(sk)) tskb = skb_rb_last(&sk->tcp_rtx_queue); //pick SYN_ACK packet TCP_SKB_CB(tskb)->tcp_flags |= TCPHDR_FIN; // set FIN flag 6. FIN_WAIT_1 --> <SEQ=100><ACK=301><END_SEQ=102><CTL=SYN,FIN,ACK> ... // TCP B: send challenge ack to SYN_FIN_ACK 7. ... <SEQ=301><ACK=101><CTL=ACK> <-- SYN-RECEIVED //challenge ack // TCP A: <SND.UNA=101> 8. FIN_WAIT_1 --> <SEQ=101><ACK=301><END_SEQ=102><CTL=SYN,FIN,ACK> ... // retransmit panic __tcp_retransmit_skb //skb->len=0 tcp_trim_head len = tp->snd_una - TCP_SKB_CB(skb)->seq // len=101-100 __pskb_trim_head skb->data_len -= len // skb->len=-1, wrap around ... ... ip_fragment icmp_glue_bits //BUG_ON If we use tcp_trim_head() to remove acked SYN from packet that contains data or other flags, skb->len will be incorrectly decremented. We can remove SYN flag that has been acked from rtx_queue earlier than tcp_trim_head(), which can fix the problem mentioned above. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Co-developed-by: Eric Dumazet <edumazet(a)google.com> Signed-off-by: Eric Dumazet <edumazet(a)google.com> Signed-off-by: Dong Chenchen <dongchenchen2(a)huawei.com> Link: https://lore.kernel.org/r/20231210020200.1539875-1-dongchenchen2@huawei.com Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: Dong Chenchen <dongchenchen2(a)huawei.com> --- net/ipv4/tcp_output.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 8cd673d75f68..90e0041ae9f7 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -2930,7 +2930,13 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb, int segs) if (skb_still_in_host_queue(sk, skb)) return -EBUSY; +start: if (before(TCP_SKB_CB(skb)->seq, tp->snd_una)) { + if (unlikely(TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)) { + TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_SYN; + TCP_SKB_CB(skb)->seq++; + goto start; + } if (unlikely(before(TCP_SKB_CB(skb)->end_seq, tp->snd_una))) { WARN_ON_ONCE(1); return -EINVAL; -- 2.25.1

2 1

[PATCH] sched: Introduce priority load balance for qos scheduler
by Song Zhang 18 Dec '23

18 Dec '23

euleros inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8MWDD CVE: NA Add new sysctl interface: /proc/sys/kernel/sched_prio_load_balance_enabled 0: default behavior 1: enable priority load balance for qos scheduler For tasks co-location with qos scheduler, when CFS do load balance, it is reasonable to prefer migrating online(Latency Sensitive) tasks. So the CFS load balance can be changed to below: cfs_tasks list is owned by online tasks. Add new cfs_offline_tasks list which is owned by offline tasks. Prefer to migrate the online tasks of cfs_tasks list to dst rq. In the scenario of hyperthread interference, if the smt expeller feature enabled, CPU A and CPU B are two hyperthreads on a physical core, CPU A runs online tasks while CPU B only has offline tasks, The offline tasks on CPU B are expelled by the online tasks on CPU A and cannot be scheduled. However, when load balance is triggered, before CPU B can migrate some online tasks from CPU A, the load on the two cpus is already balanced. As a result, CPU B cannot run online tasks and online tasks cannot be evenly distributed among different cpus. Signed-off-by: Song Zhang <zhangsong34(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + include/linux/sched/sysctl.h | 4 ++ init/Kconfig | 9 +++ kernel/sched/core.c | 3 + kernel/sched/fair.c | 90 +++++++++++++++++++++++++- kernel/sched/sched.h | 3 + 7 files changed, 110 insertions(+), 1 deletion(-) diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 039f3496af..296421c0d5 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -162,6 +162,7 @@ CONFIG_BLK_CGROUP=y CONFIG_CGROUP_WRITEBACK=y CONFIG_CGROUP_SCHED=y CONFIG_QOS_SCHED=y +CONFIG_QOS_SCHED_PRIO_LB=y CONFIG_FAIR_GROUP_SCHED=y CONFIG_CFS_BANDWIDTH=y CONFIG_RT_GROUP_SCHED=y diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index f481c1c8b4..efb9a1efa7 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -184,6 +184,7 @@ CONFIG_BLK_CGROUP=y CONFIG_CGROUP_WRITEBACK=y CONFIG_CGROUP_SCHED=y CONFIG_QOS_SCHED=y +CONFIG_QOS_SCHED_PRIO_LB=y CONFIG_FAIR_GROUP_SCHED=y CONFIG_CFS_BANDWIDTH=y CONFIG_RT_GROUP_SCHED=y diff --git a/include/linux/sched/sysctl.h b/include/linux/sched/sysctl.h index 5a64582b08..efee7213a0 100644 --- a/include/linux/sched/sysctl.h +++ b/include/linux/sched/sysctl.h @@ -29,4 +29,8 @@ extern int sysctl_numa_balancing_mode; #define sysctl_numa_balancing_mode 0 #endif +#ifdef CONFIG_QOS_SCHED_PRIO_LB +extern unsigned int sysctl_sched_prio_load_balance_enabled; +#endif + #endif /* _LINUX_SCHED_SYSCTL_H */ diff --git a/init/Kconfig b/init/Kconfig index b6952df34e..c177b7591a 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -997,6 +997,15 @@ config QOS_SCHED If in doubt, say N. +config QOS_SCHED_PRIO_LB + bool "Priority load balance for Qos scheduler" + depends on QOS_SCHED + default n + help + This feature enable priority load balance + for Qos scheduler, which prefer migrating online tasks + and migrating offline tasks secondly between CPUs. + config FAIR_GROUP_SCHED bool "Group scheduling for SCHED_OTHER" depends on CGROUP_SCHED diff --git a/kernel/sched/core.c b/kernel/sched/core.c index a1c73dea1f..d68ddb4d28 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -10045,6 +10045,9 @@ void __init sched_init(void) rq->max_idle_balance_cost = sysctl_sched_migration_cost; INIT_LIST_HEAD(&rq->cfs_tasks); +#ifdef CONFIG_QOS_SCHED_PRIO_LB + INIT_LIST_HEAD(&rq->cfs_offline_tasks); +#endif rq_attach_root(rq, &def_root_domain); #ifdef CONFIG_NO_HZ_COMMON diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 8ae0d65713..3b35448448 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -147,6 +147,10 @@ static int hundred_thousand = 100000; static int unthrottle_qos_cfs_rqs(int cpu); #endif +#ifdef CONFIG_QOS_SCHED_PRIO_LB +unsigned int sysctl_sched_prio_load_balance_enabled; +#endif + #ifdef CONFIG_CFS_BANDWIDTH /* * Amount of runtime to allocate from global (tg) to local (per-cfs_rq) pool @@ -214,6 +218,17 @@ static struct ctl_table sched_fair_sysctls[] = { .extra1 = SYSCTL_ONE_HUNDRED, .extra2 = &one_thousand, }, +#endif +#ifdef CONFIG_QOS_SCHED_PRIO_LB + { + .procname = "sched_prio_load_balance_enabled", + .data = &sysctl_sched_prio_load_balance_enabled, + .maxlen = sizeof(unsigned int), + .mode = 0644, + .proc_handler = proc_dointvec_minmax, + .extra1 = SYSCTL_ZERO, + .extra2 = SYSCTL_ONE, + }, #endif {} }; @@ -3565,6 +3580,21 @@ static inline void update_scan_period(struct task_struct *p, int new_cpu) #endif /* CONFIG_NUMA_BALANCING */ +#ifdef CONFIG_QOS_SCHED_PRIO_LB +static void +adjust_rq_cfs_tasks(void (*list_op)(struct list_head *, struct list_head *), + struct rq *rq, + struct sched_entity *se) +{ + struct task_group *tg = task_group(task_of(se)); + + if (sysctl_sched_prio_load_balance_enabled && tg->qos_level == -1) + (*list_op)(&se->group_node, &rq->cfs_offline_tasks); + else + (*list_op)(&se->group_node, &rq->cfs_tasks); +} +#endif + static void account_entity_enqueue(struct cfs_rq *cfs_rq, struct sched_entity *se) { @@ -3574,7 +3604,11 @@ account_entity_enqueue(struct cfs_rq *cfs_rq, struct sched_entity *se) struct rq *rq = rq_of(cfs_rq); account_numa_enqueue(rq, task_of(se)); +#ifdef CONFIG_QOS_SCHED_PRIO_LB + adjust_rq_cfs_tasks(list_add, rq, se); +#else list_add(&se->group_node, &rq->cfs_tasks); +#endif } #endif cfs_rq->nr_running++; @@ -8611,7 +8645,11 @@ done: __maybe_unused; * the list, so our cfs_tasks list becomes MRU * one. */ +#ifdef CONFIG_QOS_SCHED_PRIO_LB + adjust_rq_cfs_tasks(list_move, rq, &p->se); +#else list_move(&p->se.group_node, &rq->cfs_tasks); +#endif #endif if (hrtick_enabled_fair(rq)) @@ -8959,6 +8997,14 @@ static int task_hot(struct task_struct *p, struct lb_env *env) (&p->se == cfs_rq_of(&p->se)->next)) return 1; +#ifdef CONFIG_QOS_SCHED_PRIO_LB + /* Preempt sched idle cpu do not consider migration cost */ + if (sysctl_sched_prio_load_balance_enabled && + cpus_share_cache(env->src_cpu, env->dst_cpu) && + sched_idle_cpu(env->dst_cpu)) + return 0; +#endif + if (sysctl_sched_migration_cost == -1) return 1; @@ -9149,11 +9195,18 @@ static void detach_task(struct task_struct *p, struct lb_env *env) static struct task_struct *detach_one_task(struct lb_env *env) { struct task_struct *p; + struct list_head *tasks = &env->src_rq->cfs_tasks; +#ifdef CONFIG_QOS_SCHED_PRIO_LB + int loop = 0; +#endif lockdep_assert_rq_held(env->src_rq); +#ifdef CONFIG_QOS_SCHED_PRIO_LB +again: +#endif list_for_each_entry_reverse(p, - &env->src_rq->cfs_tasks, se.group_node) { + tasks, se.group_node) { if (!can_migrate_task(p, env)) continue; @@ -9168,6 +9221,15 @@ static struct task_struct *detach_one_task(struct lb_env *env) schedstat_inc(env->sd->lb_gained[env->idle]); return p; } +#ifdef CONFIG_QOS_SCHED_PRIO_LB + if (sysctl_sched_prio_load_balance_enabled) { + loop++; + if (loop == 1) { + tasks = &env->src_rq->cfs_offline_tasks; + goto again; + } + } +#endif return NULL; } @@ -9183,6 +9245,9 @@ static int detach_tasks(struct lb_env *env) unsigned long util, load; struct task_struct *p; int detached = 0; +#ifdef CONFIG_QOS_SCHED_PRIO_LB + int loop = 0; +#endif lockdep_assert_rq_held(env->src_rq); @@ -9198,6 +9263,9 @@ static int detach_tasks(struct lb_env *env) if (env->imbalance <= 0) return 0; +#ifdef CONFIG_QOS_SCHED_PRIO_LB +again: +#endif while (!list_empty(tasks)) { /* * We don't want to steal all, otherwise we may be treated likewise, @@ -9303,6 +9371,22 @@ static int detach_tasks(struct lb_env *env) list_move(&p->se.group_node, tasks); } +#ifdef CONFIG_QOS_SCHED_PRIO_LB + if (sysctl_sched_prio_load_balance_enabled && env->imbalance > 0) { + /* + * Avoid offline tasks starve to death if env->loop exceed + * env->loop_max, so we should set env->loop to 0 and detach + * offline tasks again. + */ + if (env->loop > env->loop_max) + env->loop = 0; + loop++; + if (loop == 1) { + tasks = &env->src_rq->cfs_offline_tasks; + goto again; + } + } +#endif /* * Right now, this is one of only two places we collect this stat * so we can safely collect detach_one_task() stats here rather @@ -12920,7 +13004,11 @@ static void set_next_task_fair(struct rq *rq, struct task_struct *p, bool first) * Move the next running task to the front of the list, so our * cfs_tasks list becomes MRU one. */ +#ifdef CONFIG_QOS_SCHED_PRIO_LB + adjust_rq_cfs_tasks(list_move, rq, se); +#else list_move(&se->group_node, &rq->cfs_tasks); +#endif } #endif diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 3de84e95ba..76951226b6 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -1075,6 +1075,9 @@ struct rq { int online; struct list_head cfs_tasks; +#ifdef CONFIG_QOS_SCHED_PRIO_LB + struct list_head cfs_offline_tasks; +#endif struct sched_avg avg_rt; struct sched_avg avg_dl; -- 2.33.0

1 0

[PATCH OLK-6.6 0/2] Support iocost for cgroup v1
by Li Nan 18 Dec '23

18 Dec '23

Li Nan (2): cgroup: Support iocost for cgroup v1 iocost: add config option for iocost in cgroup v1 block/blk-iocost.c | 27 +++++++++++++++++++++++++++ block/Kconfig | 8 ++++++++ 2 files changed, 35 insertions(+) -- 2.39.2

2 3

[PATCH OLK-6.6 0/4] Backport nbd bugfix patch
by Li Nan 18 Dec '23

18 Dec '23

Li Nan (4): nbd: fold nbd config initialization into nbd_alloc_config() nbd: factor out a helper to get nbd_config without holding 'config_lock' nbd: fix null-ptr-dereference while accessing 'nbd->config' nbd: pass nbd_sock to nbd_read_reply() instead of index drivers/block/nbd.c | 117 ++++++++++++++++++++++++++++---------------- 1 file changed, 75 insertions(+), 42 deletions(-) -- 2.39.2

2 5

[PATCH openEuler-1.0-LTS v2 0/2] Fix data-races around
by Zhengchao Shao 18 Dec '23

18 Dec '23

Fix data-races around sysctl_ip_fwd_update_priority and sysctl_ip_no_pmtu_disc. Kuniyuki Iwashima (2): ip: Fix data-races around sysctl_ip_fwd_update_priority. ip: Fix data-races around sysctl_ip_no_pmtu_disc. drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 3 ++- net/ipv4/af_inet.c | 2 +- net/ipv4/icmp.c | 2 +- net/ipv4/ip_forward.c | 2 +- net/ipv6/af_inet6.c | 2 +- 5 files changed, 6 insertions(+), 5 deletions(-) -- 2.34.1

2 3

[PATCH OLK-6.6 v3] ascend: export interfaces required by ascend drivers
by Yuan Can 18 Dec '23

18 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OX3Q CVE: NA -------------------------------- Export oom_type_notifier_call and __get_vm_area_caller. Export pm_autosleep_set_state and __vmalloc_node_range. Export alloc_workqueue_attrs, free_workqueue_attrs and apply_workqueue_attrs. Export find_get_task_by_vpid. Signed-off-by: Yuan Can <yuancan(a)huawei.com> --- kernel/pid.c | 1 + kernel/power/autosleep.c | 1 + kernel/workqueue.c | 3 +++ mm/oom_kill.c | 1 + mm/vmalloc.c | 2 ++ 5 files changed, 8 insertions(+) diff --git a/kernel/pid.c b/kernel/pid.c index 6500ef956f2f..383abde0c208 100644 --- a/kernel/pid.c +++ b/kernel/pid.c @@ -437,6 +437,7 @@ struct task_struct *find_get_task_by_vpid(pid_t nr) return task; } +EXPORT_SYMBOL_GPL(find_get_task_by_vpid); struct pid *get_task_pid(struct task_struct *task, enum pid_type type) { diff --git a/kernel/power/autosleep.c b/kernel/power/autosleep.c index b29c8aca7486..80ba474daa40 100644 --- a/kernel/power/autosleep.c +++ b/kernel/power/autosleep.c @@ -113,6 +113,7 @@ int pm_autosleep_set_state(suspend_state_t state) mutex_unlock(&autosleep_lock); return 0; } +EXPORT_SYMBOL_GPL(pm_autosleep_set_state); int __init pm_autosleep_init(void) { diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 0f682da96e1c..3eabf97c4e9a 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -3713,6 +3713,7 @@ void free_workqueue_attrs(struct workqueue_attrs *attrs) kfree(attrs); } } +EXPORT_SYMBOL_GPL(free_workqueue_attrs); /** * alloc_workqueue_attrs - allocate a workqueue_attrs @@ -3741,6 +3742,7 @@ struct workqueue_attrs *alloc_workqueue_attrs(void) free_workqueue_attrs(attrs); return NULL; } +EXPORT_SYMBOL_GPL(alloc_workqueue_attrs); static void copy_workqueue_attrs(struct workqueue_attrs *to, const struct workqueue_attrs *from) @@ -4482,6 +4484,7 @@ int apply_workqueue_attrs(struct workqueue_struct *wq, return ret; } +EXPORT_SYMBOL_GPL(apply_workqueue_attrs); /** * wq_update_pod - update pod affinity of a wq for CPU hot[un]plug diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 601ee56cc7d7..81b79f82f601 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -1141,6 +1141,7 @@ int oom_type_notifier_call(unsigned int type, struct oom_control *oc) return blocking_notifier_call_chain(&oom_type_notify_list, type, NULL); } +EXPORT_SYMBOL_GPL(oom_type_notifier_call); #endif /** diff --git a/mm/vmalloc.c b/mm/vmalloc.c index 30665fb33589..719539b32488 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -2622,6 +2622,7 @@ struct vm_struct *__get_vm_area_caller(unsigned long size, unsigned long flags, return __get_vm_area_node(size, 1, PAGE_SHIFT, flags, start, end, NUMA_NO_NODE, GFP_KERNEL, caller); } +EXPORT_SYMBOL(__get_vm_area_caller); /** * get_vm_area - reserve a contiguous kernel virtual area @@ -3362,6 +3363,7 @@ void *__vmalloc_node_range(unsigned long size, unsigned long align, return NULL; } +EXPORT_SYMBOL_GPL(__vmalloc_node_range); /** * __vmalloc_node - allocate virtually contiguous memory -- 2.17.1

2 1

[PATCH OLK-6.6] ubi: block: Fix use-after-free in ubiblock_cleanup
by Li Nan 18 Dec '23

18 Dec '23

From: ZhaoLong Wang <wangzhaolong1(a)huawei.com> mainline inclusion from mainline-v6.7-rc1 commit d07cec9c238ae8fc6c1a9f3f5d30a2f8ec6cdc71 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8P68G Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- The following BUG is reported when a ubiblock is removed: ================================================================== BUG: KASAN: slab-use-after-free in ubiblock_cleanup+0x88/0xa0 [ubi] Read of size 4 at addr ffff88810c8f3804 by task ubiblock/1716 CPU: 5 PID: 1716 Comm: ubiblock Not tainted 6.6.0-rc2+ #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014 Call Trace: <TASK> dump_stack_lvl+0x37/0x50 print_report+0xd0/0x620 kasan_report+0xb6/0xf0 ubiblock_cleanup+0x88/0xa0 [ubi] ubiblock_remove+0x121/0x190 [ubi] vol_cdev_ioctl+0x355/0x630 [ubi] __x64_sys_ioctl+0xc7/0x100 do_syscall_64+0x3f/0x90 entry_SYSCALL_64_after_hwframe+0x6e/0xd8 RIP: 0033:0x7f08d7445577 Code: b3 66 90 48 8b 05 11 89 2c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d e1 8 RSP: 002b:00007ffde05a3018 EFLAGS: 00000206 ORIG_RAX: 0000000000000010 RAX: ffffffffffffffda RBX: 00000000ffffffff RCX: 00007f08d7445577 RDX: 0000000000000000 RSI: 0000000000004f08 RDI: 0000000000000003 RBP: 0000000000816010 R08: 00000000008163a7 R09: 0000000000000000 R10: 0000000000000003 R11: 0000000000000206 R12: 0000000000000003 R13: 00007ffde05a3130 R14: 0000000000000000 R15: 0000000000000000 </TASK> Allocated by task 1715: kasan_save_stack+0x22/0x50 kasan_set_track+0x25/0x30 __kasan_kmalloc+0x7f/0x90 __alloc_disk_node+0x40/0x2b0 __blk_mq_alloc_disk+0x3e/0xb0 ubiblock_create+0x2ba/0x620 [ubi] vol_cdev_ioctl+0x581/0x630 [ubi] __x64_sys_ioctl+0xc7/0x100 do_syscall_64+0x3f/0x90 entry_SYSCALL_64_after_hwframe+0x6e/0xd8 Freed by task 0: kasan_save_stack+0x22/0x50 kasan_set_track+0x25/0x30 kasan_save_free_info+0x2b/0x50 __kasan_slab_free+0x10e/0x190 __kmem_cache_free+0x96/0x220 bdev_free_inode+0xa4/0xf0 rcu_core+0x496/0xec0 __do_softirq+0xeb/0x384 The buggy address belongs to the object at ffff88810c8f3800 which belongs to the cache kmalloc-1k of size 1024 The buggy address is located 4 bytes inside of freed 1024-byte region [ffff88810c8f3800, ffff88810c8f3c00) The buggy address belongs to the physical page: page:00000000d03de848 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x10c8f0 head:00000000d03de848 order:3 entire_mapcount:0 nr_pages_mapped:0 pincount:0 flags: 0x200000000000840(slab|head|node=0|zone=2) page_type: 0xffffffff() raw: 0200000000000840 ffff888100042dc0 ffffea0004244400 dead000000000002 raw: 0000000000000000 0000000080100010 00000001ffffffff 0000000000000000 page dumped because: kasan: bad access detected Memory state around the buggy address: ffff88810c8f3700: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc ffff88810c8f3780: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc >ffff88810c8f3800: fa fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb ^ ffff88810c8f3880: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb ffff88810c8f3900: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb ================================================================== Fix it by using a local variable to record the gendisk ID. Fixes: 77567b25ab9f ("ubi: use blk_mq_alloc_disk and blk_cleanup_disk") Signed-off-by: ZhaoLong Wang <wangzhaolong1(a)huawei.com> Reviewed-by: Zhihao Cheng <chengzhihao1(a)huawei.com> Signed-off-by: Richard Weinberger <richard(a)nod.at> Signed-off-by: Li Nan <linan122(a)huawei.com> --- drivers/mtd/ubi/block.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/mtd/ubi/block.c b/drivers/mtd/ubi/block.c index 437c5b83ffe5..309a42aeaa4c 100644 --- a/drivers/mtd/ubi/block.c +++ b/drivers/mtd/ubi/block.c @@ -447,13 +447,15 @@ int ubiblock_create(struct ubi_volume_info *vi) static void ubiblock_cleanup(struct ubiblock *dev) { + int id = dev->gd->first_minor; + /* Stop new requests to arrive */ del_gendisk(dev->gd); /* Finally destroy the blk queue */ dev_info(disk_to_dev(dev->gd), "released"); put_disk(dev->gd); blk_mq_free_tag_set(&dev->tag_set); - idr_remove(&ubiblock_minor_idr, dev->gd->first_minor); + idr_remove(&ubiblock_minor_idr, id); } int ubiblock_remove(struct ubi_volume_info *vi) -- 2.39.2

2 1

[PATCH OLK-5.10] dm: add error handling support for add_disk()
by Li Lingfeng 18 Dec '23

18 Dec '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I81XCK CVE: NA --------------------------- NULL pointer dereference and uaf will be triggered in del_gendisk() if add_disk has not complete. Clear request_queue and gendisk early if add_disk fail. Signed-off-by: Li Lingfeng <lilingfeng3(a)huawei.com> --- drivers/md/dm.c | 17 ++++++++++------- 1 file changed, 10 insertions(+), 7 deletions(-) diff --git a/drivers/md/dm.c b/drivers/md/dm.c index e2df27c373f9..654aae1dd777 100644 --- a/drivers/md/dm.c +++ b/drivers/md/dm.c @@ -1853,13 +1853,11 @@ static struct mapped_device *alloc_dev(int minor) * established. If request-based table is loaded: blk-mq will * override accordingly. */ - md->queue = blk_alloc_queue(numa_node_id); - if (!md->queue) - goto bad; - - md->disk = alloc_disk_node(1, md->numa_node_id); + md->disk = blk_alloc_disk(md->numa_node_id); if (!md->disk) goto bad; + md->disk->minors = 1; + md->queue = md->disk->queue; init_waitqueue_head(&md->wait); INIT_WORK(&md->work, dm_wq_work); @@ -1873,11 +1871,16 @@ static struct mapped_device *alloc_dev(int minor) md->disk->major = _major; md->disk->first_minor = minor; md->disk->fops = &dm_blk_dops; - md->disk->queue = md->queue; md->disk->private_data = md; sprintf(md->disk->disk_name, "dm-%d", minor); - add_disk_no_queue_reg(md->disk); + r = add_disk_no_queue_reg_safe(md->disk); + if (r) { + blk_cleanup_disk(md->disk); + md->queue = NULL; + md->disk = NULL; + goto bad; + } if (IS_ENABLED(CONFIG_DAX_DRIVER)) { md->dax_dev = alloc_dax(md, md->disk->disk_name, -- 2.31.1

2 1

[PATCH OLK-6.6 v3] blk-throttle: enable hierarchical throttle in cgroup v1
by Yu Kuai 18 Dec '23

18 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I4R7RE CVE: NA ----------------------------------------------- blkio subsytem is not under default hierarchy in cgroup v1 by default, which means configurations will only be effective on current cgroup for io throttle. This patch introduces a new feature that enable default hierarchy for io throttle, which means configurations will be effective on child cgroups. Such feature is disabled by default, and can be enabled by: 1) set CONFIG_BLK_DEV_SUPPORT_LEGACY_GLOBAL_LIMIT, then 2) adding "blkcg_global_limit=1" or "blkcg_global_limit=Y" or "blkcg_global_limit=y" in boot cmd; Signed-off-by: Yu Kuai <yukuai3(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + block/Kconfig | 11 ++++++++++ block/blk-throttle.c | 28 +++++++++++++++++++++++++- 4 files changed, 40 insertions(+), 1 deletion(-) diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 039f3496af78..3e76da491416 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -855,6 +855,7 @@ CONFIG_BLK_DEV_INTEGRITY_T10=m CONFIG_BLK_DEV_ZONED=y CONFIG_BLK_DEV_THROTTLING=y # CONFIG_BLK_DEV_THROTTLING_LOW is not set +CONFIG_BLK_DEV_SUPPORT_LEGACY_GLOBAL_LIMIT=y CONFIG_BLK_WBT=y CONFIG_BLK_WBT_MQ=y # CONFIG_BLK_CGROUP_IOLATENCY is not set diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index f481c1c8b4d1..a4bd203e6db5 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -925,6 +925,7 @@ CONFIG_BLK_DEV_INTEGRITY_T10=m CONFIG_BLK_DEV_ZONED=y CONFIG_BLK_DEV_THROTTLING=y # CONFIG_BLK_DEV_THROTTLING_LOW is not set +CONFIG_BLK_DEV_SUPPORT_LEGACY_GLOBAL_LIMIT=y CONFIG_BLK_WBT=y CONFIG_BLK_WBT_MQ=y # CONFIG_BLK_CGROUP_IOLATENCY is not set diff --git a/block/Kconfig b/block/Kconfig index f1364d1c0d93..95378563a05a 100644 --- a/block/Kconfig +++ b/block/Kconfig @@ -111,6 +111,17 @@ config BLK_DEV_THROTTLING_LOW Note, this is an experimental interface and could be changed someday. +config BLK_DEV_SUPPORT_LEGACY_GLOBAL_LIMIT + bool "Block layer global limit in cgroup v1" + depends on BLK_DEV_THROTTLING + default n + help + Blkio subsytem is not under default hierarchy in cgroup v1 by default, + enabling this will support globlal limit in cgroup v1. + + Note, a cmdline "blkcg_global_limit=1" is still required to enabled this + feature. + config BLK_WBT bool "Enable support for block device writeback throttling" help diff --git a/block/blk-throttle.c b/block/blk-throttle.c index 13e4377a8b28..76a54f43135c 100644 --- a/block/blk-throttle.c +++ b/block/blk-throttle.c @@ -41,6 +41,31 @@ /* A workqueue to queue throttle related work */ static struct workqueue_struct *kthrotld_workqueue; +#ifdef CONFIG_BLK_DEV_SUPPORT_LEGACY_GLOBAL_LIMIT +/* True if global limit is enabled in cgroup v1 */ +static bool global_limit; + +static inline bool blkcg_global_limit_enabled(void) +{ + return global_limit; +} + +static int __init setup_global_limit(char *str) +{ + if (!strcmp(str, "1") || !strcmp(str, "Y") || !strcmp(str, "y")) + global_limit = true; + + return 1; +} + +__setup("blkcg_global_limit=", setup_global_limit); + +#else +static inline bool blkcg_global_limit_enabled(void) +{ + return false; +} +#endif #define rb_entry_tg(node) rb_entry((node), struct throtl_grp, rb_node) /* We measure latency for request size from <= 4k to >= 1M */ @@ -405,7 +430,8 @@ static void throtl_pd_init(struct blkg_policy_data *pd) * regardless of the position of the group in the hierarchy. */ sq->parent_sq = &td->service_queue; - if (cgroup_subsys_on_dfl(io_cgrp_subsys) && blkg->parent) + if ((cgroup_subsys_on_dfl(io_cgrp_subsys) || + blkcg_global_limit_enabled()) && blkg->parent) sq->parent_sq = &blkg_to_tg(blkg->parent)->service_queue; tg->td = td; } -- 2.39.2

2 1

[PATCH OLK-6.6] fs/dirty_pages: dump the number of dirty pages for each inode
by Zizhi Wo 18 Dec '23

18 Dec '23

From: yu kuai <yukuai3(a)huawei.com> hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8OHM4?from=project-issue CVE: NA --------------------------- In order to analyse the IO performance when using buffer IO, it's useful to obtain the number of dirty pages for each inode in the filesystem. This feature depends on 'CONFIG_DIRTY_PAGES'. It creates 3 interfaces by using procfs. /proc/dirty/buffer_size for buffer allocation and release, /proc/dirty/page_threshold to filter result and /proc/dirty/dirty_list to get dirty pages. Signed-off-by: yu kuai <yukuai3(a)huawei.com> Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Kconfig | 13 + fs/Makefile | 1 + fs/dirty_pages.c | 457 +++++++++++++++++++++++++ 5 files changed, 473 insertions(+) create mode 100644 fs/dirty_pages.c diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 039f3496af78..66b4812050df 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -6883,6 +6883,7 @@ CONFIG_PROC_PAGE_MONITOR=y CONFIG_PROC_CHILDREN=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index f481c1c8b4d1..6133f66fde61 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -8090,6 +8090,7 @@ CONFIG_PROC_PID_ARCH_STATUS=y CONFIG_PROC_CPU_RESCTRL=y CONFIG_KERNFS=y CONFIG_SYSFS=y +CONFIG_DIRTY_PAGES=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y CONFIG_TMPFS_XATTR=y diff --git a/fs/Kconfig b/fs/Kconfig index aa7e03cc1941..a2280cf98729 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -170,6 +170,19 @@ source "fs/proc/Kconfig" source "fs/kernfs/Kconfig" source "fs/sysfs/Kconfig" +config DIRTY_PAGES + bool "Dumps the number of dirty pages of each file" + depends on PROC_FS + default y + help + This config supports the rendering of dirty page data to the user, + which may be useful to analyze the IO performance when using buffer + IO. + + It create 3 interfaces by using procfs. /proc/dirty/buffer_size for + buffer allocation and release; /proc/dirty/page_threshold to filter + result; /proc/dirty/dirty_list to get dirty pages. + config TMPFS bool "Tmpfs virtual memory file system support (former shm fs)" depends on SHMEM diff --git a/fs/Makefile b/fs/Makefile index f9541f40be4e..6246e173e1e4 100644 --- a/fs/Makefile +++ b/fs/Makefile @@ -44,6 +44,7 @@ obj-$(CONFIG_FS_POSIX_ACL) += posix_acl.o obj-$(CONFIG_NFS_COMMON) += nfs_common/ obj-$(CONFIG_COREDUMP) += coredump.o obj-$(CONFIG_SYSCTL) += drop_caches.o sysctls.o +obj-$(CONFIG_DIRTY_PAGES) += dirty_pages.o obj-$(CONFIG_FHANDLE) += fhandle.o obj-y += iomap/ diff --git a/fs/dirty_pages.c b/fs/dirty_pages.c new file mode 100644 index 000000000000..884e448d0c6c --- /dev/null +++ b/fs/dirty_pages.c @@ -0,0 +1,457 @@ +#include <linux/proc_fs.h> +#include <linux/seq_file.h> +#include <linux/uaccess.h> +#include <linux/pagemap.h> +#include <linux/pagevec.h> +#include <linux/fs.h> +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/sched.h> +#include <linux/proc_fs.h> +#include <linux/kdev_t.h> +#include <linux/vmalloc.h> +#include <linux/module.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include "internal.h" + +static char *buf_dirty; /* buffer to store number of dirty pages */ +static unsigned long buf_size; /* size of buffer in bytes */ +static long buff_num; /* size of buffer in number of pages */ +static int buff_limit; /* filter threshold of dirty pages*/ + +static struct proc_dir_entry *dirty_dir; + +static struct mutex buff_used; /* buffer is in used */ +static struct mutex buff_lock; /* lock when buffer is changed */ + +/* proc root directory */ +#define DIRTY_ROOT "dirty" +/* proc file for buffer allocation and release */ +#define DIRTY_SWITCH "buffer_size" +/* proc file to obtain diry pages of each inode */ +#define DIRTY_PAGES "dirty_list" +/* proc file to filter result */ +#define DIRTY_LIMIT "page_threshold" + +#define MAX_BUFF_SIZE 102400 +static void seq_set_overflow(struct seq_file *m) +{ + m->count = m->size; +} + +static unsigned long dump_dirtypages_inode(struct inode *inode) +{ + struct folio_batch fbatch; + unsigned long nr_dirtys = 0; + unsigned int nr_folios; + pgoff_t index = 0; + + folio_batch_init(&fbatch); + + while (1) { + nr_folios = filemap_get_folios_tag(inode->i_mapping, &index, + (pgoff_t)-1, PAGECACHE_TAG_DIRTY, &fbatch); + if (!nr_folios) + break; + + folio_batch_release(&fbatch); + cond_resched(); + + nr_dirtys += nr_folios; + } + + return nr_dirtys; +} + +static char *inode_filename(struct inode *inode, char *tmpname) +{ + struct dentry *dentry; + char *filename; + + dentry = d_find_alias(inode); + if (!dentry) + return ERR_PTR(-ENOENT); + + tmpname[PATH_MAX-1] = '\0'; + filename = dentry_path_raw(dentry, tmpname, PATH_MAX); + + dput(dentry); + + return filename; +} + +static inline bool is_sb_writable(struct super_block *sb) +{ + if (sb_rdonly(sb)) + return false; + + if (sb->s_writers.frozen == SB_FREEZE_COMPLETE) + return false; + + return true; +} + +/* + * dump_dirtypages_sb - dump the dirty pages of each inode in the sb + * @sb the super block + * @m the seq_file witch is initialized in proc_dpages_open + * + * For each inode in the sb, call dump_dirtypages_pages to get the number + * of dirty pages. And use seq_printf to store the result in the buffer + * if it's not less than the threshold. The inode in unusual state will + * be skipped. + */ +static void dump_dirtypages_sb(struct super_block *sb, struct seq_file *m) +{ + struct inode *inode, *toput_inode = NULL; + unsigned long nr_dirtys; + const char *fstype; + char *filename; + char *tmpname; + int limit = READ_ONCE(buff_limit); + + if (!is_sb_writable(sb)) + return; + + tmpname = kmalloc(PATH_MAX, GFP_KERNEL); + if (!tmpname) + return; + + spin_lock(&sb->s_inode_list_lock); + list_for_each_entry(inode, &sb->s_inodes, i_sb_list) { + spin_lock(&inode->i_lock); + + /* + * We must skip inodes in unusual state. We may also skip + * inodes without pages but we deliberately won't in case + * we need to reschedule to avoid softlockups. + */ + if ((inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) || + (inode->i_mapping->nrpages == 0 && !need_resched())) { + spin_unlock(&inode->i_lock); + continue; + } + __iget(inode); + spin_unlock(&inode->i_lock); + spin_unlock(&sb->s_inode_list_lock); + + cond_resched(); + + nr_dirtys = dump_dirtypages_inode(inode); + if (!nr_dirtys || nr_dirtys < limit) + goto skip; + + filename = inode_filename(inode, tmpname); + if (IS_ERR_OR_NULL(filename)) + filename = "unknown"; + + if (sb->s_type && sb->s_type->name) + fstype = sb->s_type->name; + else + fstype = "unknown"; + /* + * seq_printf return nothing, if the buffer is exhausted + * (m->size <= m->count), seq_printf will not store + * anything, just set m->count = m->size and return. In + * that case, log a warn message in buffer to remind users. + */ + if (m->size <= m->count) { + seq_set_overflow(m); + strncpy(m->buf + m->count - 12, "terminated\n\0", 12); + iput(inode); + goto done; + } + seq_printf(m, "FSType: %s, Dev ID: %u(%u:%u) ino %lu, dirty pages %lu, path %s\n", + fstype, sb->s_dev, MAJOR(sb->s_dev), + MINOR(sb->s_dev), inode->i_ino, + nr_dirtys, filename); +skip: + iput(toput_inode); + toput_inode = inode; + spin_lock(&sb->s_inode_list_lock); + } + spin_unlock(&sb->s_inode_list_lock); +done: + iput(toput_inode); + kfree(tmpname); +} + +static int proc_dpages_show(struct seq_file *m, void *v) +{ + iterate_supers((void *)dump_dirtypages_sb, (void *)m); + return 0; +} + + +static ssize_t seq_read_dirty( + struct file *file, + char __user *buf, + size_t size, + loff_t *ppos) +{ + struct seq_file *m = (struct seq_file *)file->private_data; + size_t copied = 0; + size_t n; + int err = 0; + + if (!mutex_trylock(&buff_used)) + return -EBUSY; + /* don't allow buffer to change during read */ + mutex_lock(&buff_lock); + if (m->count == 0) { + memset(buf_dirty, 0, buf_size); + if (!m->buf) { + m->size = buf_size; + m->buf = buf_dirty; + } + err = m->op->show(m, NULL); + if (err < 0) + goto done; + } + + /* if buffer changed somehow */ + if (m->size != buf_size) { + mutex_unlock(&buff_lock); + mutex_unlock(&buff_used); + return -EFAULT; + } + + n = min(m->count - m->from, size); + /* check if this is the last read */ + if (n == 0) + goto done; + n -= copy_to_user(buf, m->buf + m->from, n); + if (unlikely(!n)) { + err = -EFAULT; + goto done; + } + m->from += n; + copied += n; +done: + if (!copied) + copied = err; + else + *ppos += copied; + + mutex_unlock(&buff_lock); + mutex_unlock(&buff_used); + return copied; +} + +static void free_buf_dirty(void) +{ + if (buf_dirty != NULL) { + vfree(buf_dirty); + buf_dirty = NULL; + buf_size = 0; + } +} + +static ssize_t write_proc( + struct file *filp, + const char *buf, + size_t count, + loff_t *offp) +{ + char *msg; + int ret = 0; + long old_buff_num; + + if (!mutex_trylock(&buff_used)) + return -EBUSY; + + if (count > PAGE_SIZE) { + ret = -EINVAL; + goto error; + } + + msg = kzalloc(PAGE_SIZE, GFP_KERNEL); + if (!msg) { + ret = -ENOMEM; + goto error; + } + + if (copy_from_user(msg, buf, count)) { + ret = -EINVAL; + goto free; + } + old_buff_num = buff_num; + ret = kstrtol(msg, 10, &buff_num); + if (ret != 0 || buff_num < 0 || buff_num > MAX_BUFF_SIZE) { + buff_num = old_buff_num; + ret = -EINVAL; + goto free; + } + + mutex_lock(&buff_lock); + + ret = count; + if (buff_num == 0) { + free_buf_dirty(); + goto out; + } + if (buff_num == old_buff_num) + goto out; + + free_buf_dirty(); + buf_size = PAGE_SIZE * buff_num; + buf_dirty = vzalloc(buf_size); + + if (!buf_dirty) { + ret = -ENOMEM; + goto out; + } +out: + mutex_unlock(&buff_lock); +free: + kfree(msg); +error: + mutex_unlock(&buff_used); + return ret; +} + +static int proc_dpages_open(struct inode *inode, struct file *filp) +{ + if (buf_dirty == NULL || buf_size == 0) { + pr_warn("please allocate buffer before getting dirty pages\n"); + return -ENOMEM; + } + + return single_open(filp, proc_dpages_show, NULL); +} + +static int seq_release_dirty(struct inode *inode, struct file *file) +{ + struct seq_file *m = file->private_data; + + /* we don't want to free the buf */ + m->buf = NULL; + single_release(inode, file); + return 0; +} + +static const struct proc_ops proc_dpages_operations = { + .proc_open = proc_dpages_open, + .proc_read = seq_read_dirty, + .proc_release = seq_release_dirty, +}; + +static int proc_switch_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%ld\n", buff_num); + return 0; +} + +static int proc_limit_show(struct seq_file *m, void *v) +{ + seq_printf(m, "%d\n", READ_ONCE(buff_limit)); + return 0; +} + +static int proc_switch_open(struct inode *inode, struct file *filp) +{ + return single_open(filp, proc_switch_show, NULL); +} + +static int proc_limit_open(struct inode *inode, struct file *filp) +{ + return single_open(filp, proc_limit_show, NULL); +} + +static ssize_t write_limit_proc( + struct file *filp, + const char *buf, + size_t count, + loff_t *offp) +{ + char *msg; + int ret = 0; + long temp; + + if (count > PAGE_SIZE) { + ret = -EINVAL; + goto error; + } + + msg = kzalloc(PAGE_SIZE, GFP_KERNEL); + if (!msg) { + ret = -ENOMEM; + goto error; + } + + if (copy_from_user(msg, buf, count)) { + ret = -EINVAL; + goto free; + } + ret = kstrtol(msg, 10, &temp); + if (ret != 0 || temp < 0 || temp > INT_MAX) { + ret = -EINVAL; + goto free; + } + + WRITE_ONCE(buff_limit, temp); + ret = count; + +free: + kfree(msg); +error: + return ret; +} + + +static const struct proc_ops proc_switch_operations = { + .proc_open = proc_switch_open, + .proc_read = seq_read, + .proc_write = write_proc, + .proc_lseek = seq_lseek, + .proc_release = single_release, +}; + +static const struct proc_ops proc_limit_operations = { + .proc_open = proc_limit_open, + .proc_read = seq_read, + .proc_write = write_limit_proc, + .proc_lseek = seq_lseek, + .proc_release = single_release, +}; + + +static int __init dpages_proc_init(void) +{ + static struct proc_dir_entry *proc_file; + + dirty_dir = proc_mkdir(DIRTY_ROOT, NULL); + if (!dirty_dir) + goto fail_dir; + + proc_file = proc_create(DIRTY_PAGES, 0440, + dirty_dir, &proc_dpages_operations); + if (!proc_file) + goto fail_pages; + + proc_file = proc_create(DIRTY_SWITCH, 0640, + dirty_dir, &proc_switch_operations); + if (!proc_file) + goto fail_switch; + + proc_file = proc_create(DIRTY_LIMIT, 0640, + dirty_dir, &proc_limit_operations); + if (!proc_file) + goto fail_limit; + + mutex_init(&buff_used); + mutex_init(&buff_lock); + return 0; + +fail_limit: + remove_proc_entry(DIRTY_SWITCH, dirty_dir); +fail_switch: + remove_proc_entry(DIRTY_PAGES, dirty_dir); +fail_pages: + remove_proc_entry(DIRTY_ROOT, NULL); +fail_dir: + return -ENOMEM; +} + +subsys_initcall(dpages_proc_init); -- 2.39.2

2 1

[PATCH OLK-6.6 00/11] LoongArch: fix some pci problems
by Hongchen Zhang 18 Dec '23

18 Dec '23

Baoqi Zhang (2): LS7A2000: Add quirk for OHCI device rev 0x02 PCI: LS7A2000: fix GPU card error Hongchen Zhang (1): PCI: fix kabi error caused by pm_suspend_target_state Jianmin Lv (2): PCI: PM: fix pcie mrrs restoring PCI: LS7A2000: fix pm transition of devices under pcie port Juxin Gao (1): PCI: irq: Add early_param pci_irq_limit to limit pci irq numbers Tianli Xiong (4): PCI: Check if entry->offset already exist for mem resource PCI: Check if the pci controller can use both CFG0 and CFG1 mode to access configuration space LoongArch: fix some PCIE card not scanning properly LoongArch: pci root bridige set acpi companion only when not acpi_disabled. suijingfeng (1): PCI: fix X server auto probe fail when both ast and etnaviv drm present arch/loongarch/pci/acpi.c | 23 ++-- drivers/pci/controller/pci-loongson.c | 147 +++++++++++++++++++++++++- drivers/pci/msi/msi.c | 25 +++++ drivers/pci/pci.c | 20 +++- 4 files changed, 200 insertions(+), 15 deletions(-) -- 2.33.0

2 12

[PATCH OLK-5.10 00/10] Backport md mainline patch
by Li Nan 18 Dec '23

18 Dec '23

Christoph Hellwig (9): md: extend disks_mutex coverage md: convert to blk_alloc_disk/blk_cleanup_disk md: add the bitmap group to the default groups for the md kobject md: properly unwind when failing to add the kobject in md_alloc md: fix mddev->kobj lifetime md: fix error handling in md_alloc md: factor out a mddev_alloc_unit helper from mddev_find md: refactor mddev_find_or_alloc md: do not return existing mddevs from mddev_find_or_alloc Luis Chamberlain (1): md: add error handling support for add_disk() drivers/md/md.c | 231 +++++++++++++++++++++++++----------------------- 1 file changed, 121 insertions(+), 110 deletions(-) -- 2.39.2

2 11

[PATCH openEuler-1.0-LTS 0/2] Fix data-races around
by Zhengchao Shao 18 Dec '23

18 Dec '23

Fix data-races around sysctl_ip_fwd_update_priority and sysctl_ip_no_pmtu_disc. Kuniyuki Iwashima (2): ip: Fix data-races around sysctl_ip_fwd_update_priority. ip: Fix data-races around sysctl_ip_no_pmtu_disc. drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 3 ++- net/ipv4/af_inet.c | 2 +- net/ipv4/icmp.c | 2 +- net/ipv4/ip_forward.c | 2 +- net/ipv6/af_inet6.c | 2 +- 5 files changed, 6 insertions(+), 5 deletions(-) -- 2.34.1

2 3

[PATCH OLK-6.6 v2] blk-throttle: enable hierarchical throttle in cgroup v1
by Yu Kuai 18 Dec '23

18 Dec '23

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I4R7RE CVE: NA ----------------------------------------------- blkio subsytem is not under default hierarchy in cgroup v1 by default, which means configurations will only be effective on current cgroup for io throttle. This patch introduces a new feature that enable default hierarchy for io throttle, which means configurations will be effective on child cgroups. Such feature is disabled by default, and can be enabled by: 1) set CONFIG_BLK_DEV_SUPPORT_LEGACY_GLOBAL_LIMIT, then 2) adding "blkcg_global_limit=1" or "blkcg_global_limit=Y" or "blkcg_global_limit=y" in boot cmd; Signed-off-by: Yu Kuai <yukuai3(a)huawei.com> --- block/Kconfig | 11 +++++++++++ block/blk-throttle.c | 26 +++++++++++++++++++++++++- 2 files changed, 36 insertions(+), 1 deletion(-) diff --git a/block/Kconfig b/block/Kconfig index f1364d1c0d93..95378563a05a 100644 --- a/block/Kconfig +++ b/block/Kconfig @@ -111,6 +111,17 @@ config BLK_DEV_THROTTLING_LOW Note, this is an experimental interface and could be changed someday. +config BLK_DEV_SUPPORT_LEGACY_GLOBAL_LIMIT + bool "Block layer global limit in cgroup v1" + depends on BLK_DEV_THROTTLING + default n + help + Blkio subsytem is not under default hierarchy in cgroup v1 by default, + enabling this will support globlal limit in cgroup v1. + + Note, a cmdline "blkcg_global_limit=1" is still required to enabled this + feature. + config BLK_WBT bool "Enable support for block device writeback throttling" help diff --git a/block/blk-throttle.c b/block/blk-throttle.c index 13e4377a8b28..3a03cb4a1866 100644 --- a/block/blk-throttle.c +++ b/block/blk-throttle.c @@ -41,6 +41,29 @@ /* A workqueue to queue throttle related work */ static struct workqueue_struct *kthrotld_workqueue; +#ifdef CONFIG_BLK_DEV_SUPPORT_LEGACY_GLOBAL_LIMIT +/* True if global limit is enabled in cgroup v1 */ +static bool global_limit; + +static inline bool blkcg_global_limit_enabled(void) +{ + return global_limit; +} + +static int __init setup_global_limit(char *str) +{ + if (!strcmp(str, "1") || !strcmp(str, "Y") || !strcmp(str, "y")) + global_limit = true; + + return 1; +} + +__setup("blkcg_global_limit=", setup_global_limit); + +#else +static inline bool blkcg_global_limit_enabled(void) + return false; +#endif #define rb_entry_tg(node) rb_entry((node), struct throtl_grp, rb_node) /* We measure latency for request size from <= 4k to >= 1M */ @@ -405,7 +428,8 @@ static void throtl_pd_init(struct blkg_policy_data *pd) * regardless of the position of the group in the hierarchy. */ sq->parent_sq = &td->service_queue; - if (cgroup_subsys_on_dfl(io_cgrp_subsys) && blkg->parent) + if ((cgroup_subsys_on_dfl(io_cgrp_subsys) || + blkcg_global_limit_enabled()) && blkg->parent) sq->parent_sq = &blkg_to_tg(blkg->parent)->service_queue; tg->td = td; } -- 2.39.2

2 1