- Kernel - mailweb.openeuler.org

[PATCH 01/26] arm64/ascend: Add new enable_oom_killer interface for oom contrl
by Yang Yingliang 17 Aug '20

17 Aug '20

From: Weilong Chen <chenweilong(a)huawei.com> ascend inclusion category: feature bugzilla: NA CVE: NA ------------------------------------------------- Support disable oom-killer, and report oom events to bbox vm.enable_oom_killer: 0: disable oom killer 1: enable oom killer (default,compatible with mainline) 2: disable oom killer and panic on oom Signed-off-by: Weilong Chen <chenweilong(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- arch/arm64/Kconfig | 11 +++++++++ include/linux/oom.h | 11 +++++++++ kernel/sysctl.c | 12 ++++++++++ mm/memcontrol.c | 6 +++++ mm/oom_kill.c | 56 +++++++++++++++++++++++++++++++++++++++++++++ mm/util.c | 6 +++++ 6 files changed, 102 insertions(+) diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 10fabb5f633d..4412f14547af 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -1351,6 +1351,17 @@ config ASCEND_DVPP_MMAP special memory for DvPP processor, the new flag is only valid for Ascend platform. +config ASCEND_OOM + bool "Enable support for disable oom killer" + default y + help + In some cases we hopes that the oom will not kill the process when it occurs, + be able to notify the black box to report the event, and be able to trigger + the panic to locate the problem. + vm.enable_oom_killer: + 0: disable oom killer + 1: enable oom killer (default,compatible with mainline) + 2: disable oom killer and panic on oom endif endmenu diff --git a/include/linux/oom.h b/include/linux/oom.h index 69864a547663..689d32ab694b 100644 --- a/include/linux/oom.h +++ b/include/linux/oom.h @@ -117,4 +117,15 @@ extern struct task_struct *find_lock_task_mm(struct task_struct *p); extern int sysctl_oom_dump_tasks; extern int sysctl_oom_kill_allocating_task; extern int sysctl_panic_on_oom; + +#ifdef CONFIG_ASCEND_OOM +#define HISI_OOM_TYPE_NOMEM 0 +#define HISI_OOM_TYPE_OVERCOMMIT 1 +#define HISI_OOM_TYPE_CGROUP 2 + +extern int sysctl_enable_oom_killer; +extern int register_hisi_oom_notifier(struct notifier_block *nb); +extern int hisi_oom_notifier_call(unsigned long val, void *v); +extern int unregister_hisi_oom_notifier(struct notifier_block *nb); +#endif #endif /* _INCLUDE_LINUX_OOM_H */ diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 54ae74d3180b..665c9e2a8802 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -1264,6 +1264,18 @@ static struct ctl_table vm_table[] = { .extra1 = &zero, .extra2 = &two, }, +#ifdef CONFIG_ASCEND_OOM + { + /* 0: diasable, 1: enable, 2: disable and panic on oom */ + .procname = "enable_oom_killer", + .data = &sysctl_enable_oom_killer, + .maxlen = sizeof(sysctl_enable_oom_killer), + .mode = 0644, + .proc_handler = proc_dointvec_minmax, + .extra1 = &zero, + .extra2 = &two, + }, +#endif { .procname = "oom_kill_allocating_task", .data = &sysctl_oom_kill_allocating_task, diff --git a/mm/memcontrol.c b/mm/memcontrol.c index e0377bae0bf6..a63bfd73da9a 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -1729,6 +1729,9 @@ static enum oom_status mem_cgroup_oom(struct mem_cgroup *memcg, gfp_t mask, int current->memcg_in_oom = memcg; current->memcg_oom_gfp_mask = mask; current->memcg_oom_order = order; +#ifdef CONFIG_ASCEND_OOM + hisi_oom_notifier_call(HISI_OOM_TYPE_CGROUP, NULL); +#endif return OOM_ASYNC; } @@ -1802,6 +1805,9 @@ bool mem_cgroup_oom_synchronize(bool handle) mem_cgroup_out_of_memory(memcg, current->memcg_oom_gfp_mask, current->memcg_oom_order); } else { +#ifdef CONFIG_ASCEND_OOM + hisi_oom_notifier_call(HISI_OOM_TYPE_CGROUP, NULL); +#endif schedule(); mem_cgroup_unmark_under_oom(memcg); finish_wait(&memcg_oom_waitq, &owait.wait); diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 8a4570c53e83..c08041ecd286 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c @@ -52,6 +52,9 @@ int sysctl_panic_on_oom; int sysctl_oom_kill_allocating_task; int sysctl_oom_dump_tasks = 1; +#ifdef CONFIG_ASCEND_OOM +int sysctl_enable_oom_killer = 1; +#endif /* * Serializes oom killer invocations (out_of_memory()) from all contexts to @@ -1047,6 +1050,42 @@ int unregister_oom_notifier(struct notifier_block *nb) } EXPORT_SYMBOL_GPL(unregister_oom_notifier); +#ifdef CONFIG_ASCEND_OOM +static BLOCKING_NOTIFIER_HEAD(hisi_oom_notify_list); + +int register_hisi_oom_notifier(struct notifier_block *nb) +{ + return blocking_notifier_chain_register(&hisi_oom_notify_list, nb); +} +EXPORT_SYMBOL_GPL(register_hisi_oom_notifier); + +static unsigned long last_jiffies; +int hisi_oom_notifier_call(unsigned long val, void *v) +{ + /* when enable oom killer, just return */ + if (sysctl_enable_oom_killer == 1) + return 0; + + /* Print time interval to 10 seconds */ + if (time_after(jiffies, last_jiffies + 10 * HZ)) { + pr_err("OOM_NOTIFIER: oom type %lu\n", val); + dump_stack(); + show_mem(SHOW_MEM_FILTER_NODES, NULL); + dump_tasks(NULL, 0); + last_jiffies = jiffies; + } + + return blocking_notifier_call_chain(&hisi_oom_notify_list, val, v); +} +EXPORT_SYMBOL_GPL(hisi_oom_notifier_call); + +int unregister_hisi_oom_notifier(struct notifier_block *nb) +{ + return blocking_notifier_chain_unregister(&hisi_oom_notify_list, nb); +} +EXPORT_SYMBOL_GPL(unregister_hisi_oom_notifier); +#endif + /** * out_of_memory - kill the "best" process when we run out of memory * @oc: pointer to struct oom_control @@ -1060,10 +1099,27 @@ bool out_of_memory(struct oom_control *oc) { unsigned long freed = 0; enum oom_constraint constraint = CONSTRAINT_NONE; +#ifdef CONFIG_ASCEND_OOM + unsigned long oom_type; +#endif if (oom_killer_disabled) return false; +#ifdef CONFIG_ASCEND_OOM + if (sysctl_enable_oom_killer == 0 || sysctl_enable_oom_killer == 2) { + if (is_memcg_oom(oc)) + oom_type = HISI_OOM_TYPE_CGROUP; + else + oom_type = HISI_OOM_TYPE_NOMEM; + + hisi_oom_notifier_call(oom_type, NULL); + if (unlikely(sysctl_enable_oom_killer == 2)) + panic("Out of memory, panic by sysctl_enable_oom_killer"); + return false; + } +#endif + if (!is_memcg_oom(oc)) { blocking_notifier_call_chain(&oom_notify_list, 0, &freed); if (freed > 0) diff --git a/mm/util.c b/mm/util.c index 5515219168e8..ed64ef1f8387 100644 --- a/mm/util.c +++ b/mm/util.c @@ -17,6 +17,9 @@ #include <asm/sections.h> #include <linux/uaccess.h> +#ifdef CONFIG_ASCEND_OOM +#include <linux/oom.h> +#endif #include "internal.h" @@ -744,6 +747,9 @@ int __vm_enough_memory(struct mm_struct *mm, long pages, int cap_sys_admin) if (percpu_counter_read_positive(&vm_committed_as) < allowed) return 0; error: +#ifdef CONFIG_ASCEND_OOM + hisi_oom_notifier_call(HISI_OOM_TYPE_OVERCOMMIT, NULL); +#endif vm_unacct_memory(pages); return -ENOMEM; -- 2.25.1

1 25

kernel branch openEuler-1.0-LTS build failed
by 李义 17 Aug '20

17 Aug '20

Hi ALL, [Description] The kernel branch  openEuler-1.0-LTS  has import the patch  f6b330acc  but has no no_refcnt and build failed [Debug] the upstream commit ad0f75  has introduce the no_refcnt to the struct sock_cgroup_data. Maybe we should the merge the patch or revert it on branch openEuler-1.0-LTS

2 1

[PATCH kernel-4.19] arm64: add pmem module for kernel update
by zhuling 17 Aug '20

17 Aug '20

Euleros inclusion Category: feature Bugzilla: NA CVE: NA Use reserved memory to create a pmem device to store the processes information that dumped before kernel update. When you want to use this feature you need to declare by "pmemmem=pmem_size:pmem_phystart" in cmdline. (exp: pmemmem=100M:0x202000000000) Signed-off-by: zhuling <zhuling8(a)huawei.com> --- arch/arm64/kernel/setup.c | 5 +++ arch/arm64/mm/init.c | 90 ++++++++++++++++++++++++++++++++++++++ drivers/nvdimm/Kconfig | 11 +++++ drivers/nvdimm/Makefile | 3 ++ drivers/nvdimm/kup_pmem.c | 107 ++++++++++++++++++++++++++++++++++++++++++++++ include/linux/ioport.h | 1 + include/linux/mm.h | 4 ++ lib/Kconfig | 6 +++ 8 files changed, 227 insertions(+) create mode 100644 drivers/nvdimm/kup_pmem.c diff --git a/arch/arm64/kernel/setup.c b/arch/arm64/kernel/setup.c index 155b8a6..e96cade 100644 --- a/arch/arm64/kernel/setup.c +++ b/arch/arm64/kernel/setup.c @@ -251,6 +251,11 @@ static void __init request_standard_resources(void) if (kernel_data.start >= res->start && kernel_data.end <= res->end) request_resource(res, &kernel_data); +#ifdef CONFIG_KUP_PMEM_MEMORY + if (pmem_res.end) + insert_resource(&iomem_resource, &pmem_res); +#endif + #ifdef CONFIG_KEXEC_CORE /* Userspace will find "Crash kernel" region in /proc/iomem. */ if (crashk_low_res.end && crashk_low_res.start >= res->start && diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c index e43764d..169d663 100644 --- a/arch/arm64/mm/init.c +++ b/arch/arm64/mm/init.c @@ -65,6 +65,18 @@ phys_addr_t arm64_dma_phys_limit __ro_after_init; struct res_mem res_mem[MAX_RES_REGIONS]; int res_mem_count; +#ifdef CONFIG_KUP_PMEM_MEMORY +static unsigned long long pmem_size, pmem_phystart; + +struct resource pmem_res = { + .name = "Kpmem Dev", + .start = 0, + .end = 0, + .flags = IORESOURCE_MEM, + .desc = IORES_DESC_KPMEM_DEV +}; +#endif + #ifdef CONFIG_BLK_DEV_INITRD static int __init early_initrd(char *p) { @@ -192,6 +204,80 @@ static void __init kexec_reserve_crashkres_pages(void) } #endif /* CONFIG_KEXEC_CORE */ +#ifdef CONFIG_KUP_PMEM_MEMORY +/* + * reserve_pmem() - reserves memory for pmem + * + * This function reserves memory area given in "pmemmem=" kernel command + * line parameter. The memory reserved is used by pmem restore progress + * when kernel update. + */ +static int __init parse_pmem(char *par) +{ + char *cur = par; + + if(!par) + return 0; + + pmem_size = 0; + pmem_phystart = 0; + + pmem_size = memparse(par, &cur); + if (par == cur) { + pr_warn("pmem: memory value expected\n"); + return -EINVAL; + } + + if (*cur == ':') + pmem_phystart = memparse(cur+1, &cur); + else if (*cur != ' ' && *cur != '\0') { + pr_warn("pmem: unrecognized char %c\n", *cur); + return -EINVAL; + } + + return 0; +} +early_param("pmemmem", parse_pmem); + +static void __init reserve_pmem(void) +{ + if (!pmem_size ||!pmem_phystart) { + return; + } + + pmem_size = PAGE_ALIGN(pmem_size); + + if (!memblock_is_region_memory(pmem_phystart, pmem_size)) { + pr_warn("cannot reserve pmem: region is not memory!\n"); + return; + } + + if (memblock_is_region_reserved(pmem_phystart, pmem_size)) { + pr_warn("cannot reserve pmem: region overlaps reserved memory!\n"); + return; + } + + if (!IS_ALIGNED(pmem_phystart, SZ_2M)) { + pr_warn("cannot reserve pmem: base address is not 2MB aligned\n"); + return; + } + memblock_reserve(pmem_phystart, pmem_size); + memblock_remove(pmem_phystart, pmem_size); + pr_info("pmem reserved: 0x%016llx - 0x%016llx (%lld MB)\n", + pmem_phystart, pmem_phystart + pmem_size, pmem_size >> 20); + + pmem_res.start = pmem_phystart; + pmem_res.end = pmem_phystart + pmem_size - 1; +} +#else +static void __init reserve_pmem(void) +{ +} +static void __init reserve_pmem_pages(void) +{ +} +#endif /*CONFIG_KUP_PMEM_MEMORY*/ + #ifdef CONFIG_CRASH_DUMP static int __init early_init_dt_scan_elfcorehdr(unsigned long node, const char *uname, int depth, void *data) @@ -584,6 +670,10 @@ void __init arm64_memblock_init(void) reserve_elfcorehdr(); +#ifdef CONFIG_KUP_PMEM_MEMORY + reserve_pmem(); +#endif + high_memory = __va(memblock_end_of_DRAM() - 1) + 1; dma_contiguous_reserve(arm64_dma_phys_limit); diff --git a/drivers/nvdimm/Kconfig b/drivers/nvdimm/Kconfig index 9d36473..1097a8d 100644 --- a/drivers/nvdimm/Kconfig +++ b/drivers/nvdimm/Kconfig @@ -112,4 +112,15 @@ config OF_PMEM Select Y if unsure. +config KUP_PMEM + tristate "Persistent memory for kernel update" + depends on LIBNVDIMM + depends on KUP_PMEM_MEMORY + default LIBNVDIMM + help + Allows regions of persistent memory to be described in the + device-tree. + + Select Y if unsure. + endif diff --git a/drivers/nvdimm/Makefile b/drivers/nvdimm/Makefile index e884704..853cbcb 100644 --- a/drivers/nvdimm/Makefile +++ b/drivers/nvdimm/Makefile @@ -5,6 +5,7 @@ obj-$(CONFIG_ND_BTT) += nd_btt.o obj-$(CONFIG_ND_BLK) += nd_blk.o obj-$(CONFIG_X86_PMEM_LEGACY) += nd_e820.o obj-$(CONFIG_OF_PMEM) += of_pmem.o +obj-$(CONFIG_KUP_PMEM) += nd_kup_pmem.o nd_pmem-y := pmem.o @@ -14,6 +15,8 @@ nd_blk-y := blk.o nd_e820-y := e820.o +nd_kup_pmem-y := kup_pmem.o + libnvdimm-y := core.o libnvdimm-y += bus.o libnvdimm-y += dimm_devs.o diff --git a/drivers/nvdimm/kup_pmem.c b/drivers/nvdimm/kup_pmem.c new file mode 100644 index 00000000..d9b0633 --- /dev/null +++ b/drivers/nvdimm/kup_pmem.c @@ -0,0 +1,107 @@ +/* + * Copyright (c) Huawei Technologies Co., Ltd. 2020. All rights reserved. + * + * This source code is licensed under the GNU General Public License, + * Version 2. See the file COPYING for more details. + * + * kup_pmem.c - kernel update support code. + * create a pmem device to store the processes information that is dumped + * when we want to kernel update. + */ + +#include <linux/platform_device.h> +#include <linux/memory_hotplug.h> +#include <linux/libnvdimm.h> +#include <linux/module.h> +#include <asm/io.h> + +static const struct attribute_group *kup_pmem_attribute_groups[] = { + &nvdimm_bus_attribute_group, + NULL, +}; + +static const struct attribute_group *kup_pmem_region_attribute_groups[] = { + &nd_region_attribute_group, + &nd_device_attribute_group, + NULL, +}; + +static int kup_pmem_remove(struct platform_device *pdev) +{ + struct nvdimm_bus *nvdimm_bus = platform_get_drvdata(pdev); + + nvdimm_bus_unregister(nvdimm_bus); + + return 0; +} + +static int kup_register_one(struct resource *res, void *data) +{ + struct nd_region_desc ndr_desc; + struct nvdimm_bus *nvdimm_bus = data; + + memset(&ndr_desc, 0, sizeof(ndr_desc)); + ndr_desc.res = res; + ndr_desc.attr_groups = kup_pmem_region_attribute_groups; + ndr_desc.numa_node = NUMA_NO_NODE; + set_bit(ND_REGION_PAGEMAP, &ndr_desc.flags); + if (!nvdimm_pmem_region_create(nvdimm_bus, &ndr_desc)) + return -ENXIO; + return 0; +} + +static int kup_pmem_probe(struct platform_device *pdev) +{ + static struct nvdimm_bus_descriptor nd_desc; + struct device *dev = &pdev->dev; + struct nvdimm_bus *nvdimm_bus; + int rc = -ENXIO; + + nd_desc.attr_groups = kup_pmem_attribute_groups; + nd_desc.provider_name = "kup_pmem"; + nd_desc.module = THIS_MODULE; + nvdimm_bus = nvdimm_bus_register(dev, &nd_desc); + if (!nvdimm_bus) + goto err; + platform_set_drvdata(pdev, nvdimm_bus); + + rc = walk_iomem_res_desc(IORES_DESC_KPMEM_DEV, + IORESOURCE_MEM, 0, -1, nvdimm_bus, kup_register_one); + if (rc) + goto err; + + return 0; +err: + nvdimm_bus_unregister(nvdimm_bus); + dev_err(dev, "kup_pmem: failed to register legacy persistent memory ranges\n"); + return rc; +} + +static struct platform_driver kup_pmem_driver = { + .probe = kup_pmem_probe, + .remove = kup_pmem_remove, + .driver = { + .name = "kup_pmem", + }, +}; +static struct platform_device *pdev; + +static __init int register_kup_pmem(void) +{ + platform_driver_register(&kup_pmem_driver); + pdev = platform_device_alloc("kup_pmem", -1); + + return platform_device_add(pdev); +} + +static __exit void unregister_kup_pmem(void) +{ + platform_device_del(pdev); + platform_driver_unregister(&kup_pmem_driver); +} + +module_init(register_kup_pmem); +module_exit(unregister_kup_pmem); +MODULE_ALIAS("platform:kup_pmem*"); +MODULE_LICENSE("GPL v2"); +MODULE_AUTHOR("Huawei Corporation"); diff --git a/include/linux/ioport.h b/include/linux/ioport.h index 5330288..c5f59b9 100644 --- a/include/linux/ioport.h +++ b/include/linux/ioport.h @@ -139,6 +139,7 @@ enum { IORES_DESC_PERSISTENT_MEMORY_LEGACY = 5, IORES_DESC_DEVICE_PRIVATE_MEMORY = 6, IORES_DESC_DEVICE_PUBLIC_MEMORY = 7, + IORES_DESC_KPMEM_DEV = 8, }; /* helpers to define resources */ diff --git a/include/linux/mm.h b/include/linux/mm.h index b985af8..d84b0f0 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -38,6 +38,10 @@ struct bdi_writeback; void init_mm_internals(void); +#ifdef CONFIG_KUP_PMEM_MEMORY +extern struct resource pmem_res; +#endif + #ifndef CONFIG_NEED_MULTIPLE_NODES /* Don't use mapnrs, do it properly */ extern unsigned long max_mapnr; diff --git a/lib/Kconfig b/lib/Kconfig index a3928d4..cc49a86 100644 --- a/lib/Kconfig +++ b/lib/Kconfig @@ -621,3 +621,9 @@ config GENERIC_LIB_CMPDI2 config GENERIC_LIB_UCMPDI2 bool + +config KUP_PMEM_MEMORY + bool "reserve memory for kup pmem to store image" + default y + help + Say y here to enable this feature -- 2.9.5

1 0

[PATCH kernel-4.19 1/2] kexec: Add quick kexec support for kernel
by Sang Yan 13 Aug '20

13 Aug '20

euleros inclusion category: feature bugzilla: NA issues: #I1RC8Z CVE: NA In normal kexec, relocating kernel may cost 5 ~ 10 seconds, to copy all segments from vmalloced memory to kernel boot memory, because of disabled mmu. We introduce quick kexec to save time of copying memory as above, just like kdump(kexec on crash), by using reserved memory "Quick Kexec". Constructing quick kimage as the same as crash kernel, then simply copy all segments of kimage to reserved memroy. We also add this support in syscall kexec_load using flags of KEXEC_QUICK. Signed-off-by: Sang Yan <sangyan(a)huawei.com> --- arch/Kconfig | 7 +++++++ include/linux/ioport.h | 3 +++ include/linux/kexec.h | 13 +++++++++++- include/uapi/linux/kexec.h | 3 +++ kernel/kexec.c | 10 ++++++++++ kernel/kexec_core.c | 41 +++++++++++++++++++++++++++++--------- 6 files changed, 67 insertions(+), 10 deletions(-) diff --git a/arch/Kconfig b/arch/Kconfig index b54e485e47ae..1a9d00a6e122 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -18,6 +18,13 @@ config KEXEC_CORE select CRASH_CORE bool +config QUICK_KEXEC + bool "Support for quick kexec" + depends on KEXEC_CORE + help + Say y here to enable this feature. + It use reserved memory to accelerate kexec. + config HAVE_IMA_KEXEC bool diff --git a/include/linux/ioport.h b/include/linux/ioport.h index 5330288da8db..a42a1800c133 100644 --- a/include/linux/ioport.h +++ b/include/linux/ioport.h @@ -139,6 +139,9 @@ enum { IORES_DESC_PERSISTENT_MEMORY_LEGACY = 5, IORES_DESC_DEVICE_PRIVATE_MEMORY = 6, IORES_DESC_DEVICE_PUBLIC_MEMORY = 7, +#ifdef CONFIG_QUICK_KEXEC + IORES_DESC_QUICK_KEXEC = 8, +#endif }; /* helpers to define resources */ diff --git a/include/linux/kexec.h b/include/linux/kexec.h index d6b8d0a69720..d5e42a44f2ca 100644 --- a/include/linux/kexec.h +++ b/include/linux/kexec.h @@ -233,9 +233,12 @@ struct kimage { unsigned long control_page; /* Flags to indicate special processing */ - unsigned int type : 1; + unsigned int type : 2; #define KEXEC_TYPE_DEFAULT 0 #define KEXEC_TYPE_CRASH 1 +#ifdef CONFIG_QUICK_KEXEC +#define KEXEC_TYPE_QUICK 2 +#endif unsigned int preserve_context : 1; /* If set, we are using file mode kexec syscall */ unsigned int file_mode:1; @@ -296,6 +299,11 @@ extern int kexec_load_disabled; #define KEXEC_FLAGS (KEXEC_ON_CRASH | KEXEC_PRESERVE_CONTEXT) #endif +#ifdef CONFIG_QUICK_KEXEC +#undef KEXEC_FLAGS +#define KEXEC_FLAGS (KEXEC_ON_CRASH | KEXEC_QUICK) +#endif + /* List of defined/legal kexec file flags */ #define KEXEC_FILE_FLAGS (KEXEC_FILE_UNLOAD | KEXEC_FILE_ON_CRASH | \ KEXEC_FILE_NO_INITRAMFS) @@ -305,6 +313,9 @@ extern int kexec_load_disabled; extern struct resource crashk_res; extern struct resource crashk_low_res; extern note_buf_t __percpu *crash_notes; +#ifdef CONFIG_QUICK_KEXEC +extern struct resource quick_kexec_res; +#endif /* flag to track if kexec reboot is in progress */ extern bool kexec_in_progress; diff --git a/include/uapi/linux/kexec.h b/include/uapi/linux/kexec.h index 6d112868272d..dcf9857452da 100644 --- a/include/uapi/linux/kexec.h +++ b/include/uapi/linux/kexec.h @@ -12,6 +12,9 @@ /* kexec flags for different usage scenarios */ #define KEXEC_ON_CRASH 0x00000001 #define KEXEC_PRESERVE_CONTEXT 0x00000002 +#ifdef CONFIG_QUICK_KEXEC +#define KEXEC_QUICK 0x00000004 +#endif #define KEXEC_ARCH_MASK 0xffff0000 /* diff --git a/kernel/kexec.c b/kernel/kexec.c index 68559808fdfa..47dfad722b7c 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -46,6 +46,9 @@ static int kimage_alloc_init(struct kimage **rimage, unsigned long entry, int ret; struct kimage *image; bool kexec_on_panic = flags & KEXEC_ON_CRASH; +#ifdef CONFIG_QUICK_KEXEC + bool kexec_on_quick = flags & KEXEC_QUICK; +#endif if (kexec_on_panic) { /* Verify we have a valid entry point */ @@ -71,6 +74,13 @@ static int kimage_alloc_init(struct kimage **rimage, unsigned long entry, image->type = KEXEC_TYPE_CRASH; } +#ifdef CONFIG_QUICK_KEXEC + if (kexec_on_quick) { + image->control_page = quick_kexec_res.start; + image->type = KEXEC_TYPE_QUICK; + } +#endif + ret = sanity_check_segment_list(image); if (ret) goto out_free_image; diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index b36c9c46cd2c..595a757af656 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -74,6 +74,16 @@ struct resource crashk_low_res = { .desc = IORES_DESC_CRASH_KERNEL }; +#ifdef CONFIG_QUICK_KEXEC +struct resource quick_kexec_res = { + .name = "Quick kexec", + .start = 0, + .end = 0, + .flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM, + .desc = IORES_DESC_QUICK_KEXEC +}; +#endif + int kexec_should_crash(struct task_struct *p) { /* @@ -470,8 +480,10 @@ static struct page *kimage_alloc_normal_control_pages(struct kimage *image, return pages; } -static struct page *kimage_alloc_crash_control_pages(struct kimage *image, - unsigned int order) + +static struct page *kimage_alloc_special_control_pages(struct kimage *image, + unsigned int order, + unsigned long end) { /* Control pages are special, they are the intermediaries * that are needed while we copy the rest of the pages @@ -501,7 +513,7 @@ static struct page *kimage_alloc_crash_control_pages(struct kimage *image, size = (1 << order) << PAGE_SHIFT; hole_start = (image->control_page + (size - 1)) & ~(size - 1); hole_end = hole_start + size - 1; - while (hole_end <= crashk_res.end) { + while (hole_end <= end) { unsigned long i; cond_resched(); @@ -536,7 +548,6 @@ static struct page *kimage_alloc_crash_control_pages(struct kimage *image, return pages; } - struct page *kimage_alloc_control_pages(struct kimage *image, unsigned int order) { @@ -547,8 +558,15 @@ struct page *kimage_alloc_control_pages(struct kimage *image, pages = kimage_alloc_normal_control_pages(image, order); break; case KEXEC_TYPE_CRASH: - pages = kimage_alloc_crash_control_pages(image, order); + pages = kimage_alloc_special_control_pages(image, order, + crashk_res.end); + break; +#ifdef CONFIG_QUICK_KEXEC + case KEXEC_TYPE_QUICK: + pages = kimage_alloc_special_control_pages(image, order, + quick_kexec_res.end); break; +#endif } return pages; @@ -898,11 +916,11 @@ static int kimage_load_normal_segment(struct kimage *image, return result; } -static int kimage_load_crash_segment(struct kimage *image, +static int kimage_load_special_segment(struct kimage *image, struct kexec_segment *segment) { - /* For crash dumps kernels we simply copy the data from - * user space to it's destination. + /* For crash dumps kernels and quick kexec kernels + * we simply copy the data from user space to it's destination. * We do things a page at a time for the sake of kmap. */ unsigned long maddr; @@ -976,8 +994,13 @@ int kimage_load_segment(struct kimage *image, result = kimage_load_normal_segment(image, segment); break; case KEXEC_TYPE_CRASH: - result = kimage_load_crash_segment(image, segment); + result = kimage_load_special_segment(image, segment); + break; +#ifdef CONFIG_QUICK_KEXEC + case KEXEC_TYPE_QUICK: + result = kimage_load_special_segment(image, segment); break; +#endif } return result; -- 2.19.1

1 1

[PATCH kernel-4.19 1/2] kexec: Add quick kexec support for kernel
by Sang Yan 13 Aug '20

13 Aug '20

euleros inclusion category: feature bugzilla: NA issues: #I1RC8Z CVE: NA In normal kexec, relocating kernel may cost 5 ~ 10 seconds, to copy all segments from vmalloced memory to kernel boot memory, because of disabled mmu. We introduce quick kexec to save time of copying memory as above, just like kdump(kexec on crash), by using reserved memory "Quick Kexec". Constructing quick kimage as the same as crash kernel, then simply copy all segments of kimage to reserved memroy. We also add this support in syscall kexec_load using flags of KEXEC_QUICK. Signed-off-by: Sang Yan <sangyan(a)huawei.com> --- arch/Kconfig | 7 +++++++ include/linux/ioport.h | 3 +++ include/linux/kexec.h | 13 +++++++++++- include/uapi/linux/kexec.h | 3 +++ kernel/kexec.c | 10 ++++++++++ kernel/kexec_core.c | 41 +++++++++++++++++++++++++++++--------- 6 files changed, 67 insertions(+), 10 deletions(-) diff --git a/arch/Kconfig b/arch/Kconfig index b54e485e47ae..1a9d00a6e122 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -18,6 +18,13 @@ config KEXEC_CORE select CRASH_CORE bool +config QUICK_KEXEC + bool "Support for quick kexec" + depends on KEXEC_CORE + help + Say y here to enable this feature. + It use reserved memory to accelerate kexec. + config HAVE_IMA_KEXEC bool diff --git a/include/linux/ioport.h b/include/linux/ioport.h index 5330288da8db..a42a1800c133 100644 --- a/include/linux/ioport.h +++ b/include/linux/ioport.h @@ -139,6 +139,9 @@ enum { IORES_DESC_PERSISTENT_MEMORY_LEGACY = 5, IORES_DESC_DEVICE_PRIVATE_MEMORY = 6, IORES_DESC_DEVICE_PUBLIC_MEMORY = 7, +#ifdef CONFIG_QUICK_KEXEC + IORES_DESC_QUICK_KEXEC = 8, +#endif }; /* helpers to define resources */ diff --git a/include/linux/kexec.h b/include/linux/kexec.h index d6b8d0a69720..d5e42a44f2ca 100644 --- a/include/linux/kexec.h +++ b/include/linux/kexec.h @@ -233,9 +233,12 @@ struct kimage { unsigned long control_page; /* Flags to indicate special processing */ - unsigned int type : 1; + unsigned int type : 2; #define KEXEC_TYPE_DEFAULT 0 #define KEXEC_TYPE_CRASH 1 +#ifdef CONFIG_QUICK_KEXEC +#define KEXEC_TYPE_QUICK 2 +#endif unsigned int preserve_context : 1; /* If set, we are using file mode kexec syscall */ unsigned int file_mode:1; @@ -296,6 +299,11 @@ extern int kexec_load_disabled; #define KEXEC_FLAGS (KEXEC_ON_CRASH | KEXEC_PRESERVE_CONTEXT) #endif +#ifdef CONFIG_QUICK_KEXEC +#undef KEXEC_FLAGS +#define KEXEC_FLAGS (KEXEC_ON_CRASH | KEXEC_QUICK) +#endif + /* List of defined/legal kexec file flags */ #define KEXEC_FILE_FLAGS (KEXEC_FILE_UNLOAD | KEXEC_FILE_ON_CRASH | \ KEXEC_FILE_NO_INITRAMFS) @@ -305,6 +313,9 @@ extern int kexec_load_disabled; extern struct resource crashk_res; extern struct resource crashk_low_res; extern note_buf_t __percpu *crash_notes; +#ifdef CONFIG_QUICK_KEXEC +extern struct resource quick_kexec_res; +#endif /* flag to track if kexec reboot is in progress */ extern bool kexec_in_progress; diff --git a/include/uapi/linux/kexec.h b/include/uapi/linux/kexec.h index 6d112868272d..dcf9857452da 100644 --- a/include/uapi/linux/kexec.h +++ b/include/uapi/linux/kexec.h @@ -12,6 +12,9 @@ /* kexec flags for different usage scenarios */ #define KEXEC_ON_CRASH 0x00000001 #define KEXEC_PRESERVE_CONTEXT 0x00000002 +#ifdef CONFIG_QUICK_KEXEC +#define KEXEC_QUICK 0x00000004 +#endif #define KEXEC_ARCH_MASK 0xffff0000 /* diff --git a/kernel/kexec.c b/kernel/kexec.c index 68559808fdfa..47dfad722b7c 100644 --- a/kernel/kexec.c +++ b/kernel/kexec.c @@ -46,6 +46,9 @@ static int kimage_alloc_init(struct kimage **rimage, unsigned long entry, int ret; struct kimage *image; bool kexec_on_panic = flags & KEXEC_ON_CRASH; +#ifdef CONFIG_QUICK_KEXEC + bool kexec_on_quick = flags & KEXEC_QUICK; +#endif if (kexec_on_panic) { /* Verify we have a valid entry point */ @@ -71,6 +74,13 @@ static int kimage_alloc_init(struct kimage **rimage, unsigned long entry, image->type = KEXEC_TYPE_CRASH; } +#ifdef CONFIG_QUICK_KEXEC + if (kexec_on_quick) { + image->control_page = quick_kexec_res.start; + image->type = KEXEC_TYPE_QUICK; + } +#endif + ret = sanity_check_segment_list(image); if (ret) goto out_free_image; diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c index b36c9c46cd2c..595a757af656 100644 --- a/kernel/kexec_core.c +++ b/kernel/kexec_core.c @@ -74,6 +74,16 @@ struct resource crashk_low_res = { .desc = IORES_DESC_CRASH_KERNEL }; +#ifdef CONFIG_QUICK_KEXEC +struct resource quick_kexec_res = { + .name = "Quick kexec", + .start = 0, + .end = 0, + .flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM, + .desc = IORES_DESC_QUICK_KEXEC +}; +#endif + int kexec_should_crash(struct task_struct *p) { /* @@ -470,8 +480,10 @@ static struct page *kimage_alloc_normal_control_pages(struct kimage *image, return pages; } -static struct page *kimage_alloc_crash_control_pages(struct kimage *image, - unsigned int order) + +static struct page *kimage_alloc_special_control_pages(struct kimage *image, + unsigned int order, + unsigned long end) { /* Control pages are special, they are the intermediaries * that are needed while we copy the rest of the pages @@ -501,7 +513,7 @@ static struct page *kimage_alloc_crash_control_pages(struct kimage *image, size = (1 << order) << PAGE_SHIFT; hole_start = (image->control_page + (size - 1)) & ~(size - 1); hole_end = hole_start + size - 1; - while (hole_end <= crashk_res.end) { + while (hole_end <= end) { unsigned long i; cond_resched(); @@ -536,7 +548,6 @@ static struct page *kimage_alloc_crash_control_pages(struct kimage *image, return pages; } - struct page *kimage_alloc_control_pages(struct kimage *image, unsigned int order) { @@ -547,8 +558,15 @@ struct page *kimage_alloc_control_pages(struct kimage *image, pages = kimage_alloc_normal_control_pages(image, order); break; case KEXEC_TYPE_CRASH: - pages = kimage_alloc_crash_control_pages(image, order); + pages = kimage_alloc_special_control_pages(image, order, + crashk_res.end); + break; +#ifdef CONFIG_QUICK_KEXEC + case KEXEC_TYPE_QUICK: + pages = kimage_alloc_special_control_pages(image, order, + quick_kexec_res.end); break; +#endif } return pages; @@ -898,11 +916,11 @@ static int kimage_load_normal_segment(struct kimage *image, return result; } -static int kimage_load_crash_segment(struct kimage *image, +static int kimage_load_special_segment(struct kimage *image, struct kexec_segment *segment) { - /* For crash dumps kernels we simply copy the data from - * user space to it's destination. + /* For crash dumps kernels and quick kexec kernels + * we simply copy the data from user space to it's destination. * We do things a page at a time for the sake of kmap. */ unsigned long maddr; @@ -976,8 +994,13 @@ int kimage_load_segment(struct kimage *image, result = kimage_load_normal_segment(image, segment); break; case KEXEC_TYPE_CRASH: - result = kimage_load_crash_segment(image, segment); + result = kimage_load_special_segment(image, segment); + break; +#ifdef CONFIG_QUICK_KEXEC + case KEXEC_TYPE_QUICK: + result = kimage_load_special_segment(image, segment); break; +#endif } return result; -- 2.19.1

1 1

[PATCH 0/7] Improvements and fixes
by Roberto Sassu 07 Aug '20

07 Aug '20

This patch set provides some improvements and fixes. The first three patches solve a permission problem caused by EVM. EVM denies xattr operations also for files that are not appraised by IMA. If only executables are appraised, xattr operations on the other files should be allowed, even if metadata verification fails (for example due to missing security.evm). At the moment, in openEuler we use EVM_ALLOW_METADATA_WRITES to avoid this problem (EVM does not check metadata integrity), but it would be useful to do the verification for example to prevent accidental changes on immutable metadata. The fourth patch enables the choice of the algorithm for the HMAC and ensures that the parameters passed to the functions which handle the HMAC are consistent with the algorithm chosen. The last three patches are simple bug fixes. Roberto Sassu (7): evm: Move hooks outside LSM infrastructure evm: Extend API of post hooks to pass the result of pre hooks evm: Return -EAGAIN to ignore verification failures evm: Propagate choice of HMAC algorithm in evm_crypto.c ima: Fix datalen check in ima_write_data() evm: Fix validation of fake xattr passed by IMA evm: Initialize saved_evm_status fs/attr.c | 7 ++- fs/xattr.c | 64 +++++++++++++++++--------- include/linux/evm.h | 18 +++++--- security/integrity/evm/Kconfig | 32 +++++++++++++ security/integrity/evm/evm.h | 1 + security/integrity/evm/evm_crypto.c | 15 ++++-- security/integrity/evm/evm_main.c | 71 +++++++++++++++++++++-------- security/integrity/ima/ima_fs.c | 2 +- security/integrity/integrity.h | 2 +- security/security.c | 18 ++------ 10 files changed, 158 insertions(+), 72 deletions(-) -- 2.27.GIT

1 7

[PATCH 01/23] ima: Use buffer large enough to store fake IMA xattr for appraisal
by Roberto Sassu 06 Aug '20

06 Aug '20

hulk inclusion category: feature feature: digest-lists --------------------------- A fake IMA xattr is created to perform EVM verification even if security.ima is not present. Appraisal could succeed if EVM status is unknown and the file digest is found in a digest list. This patch allocates a larger buffer to store fake IMA xattrs (struct evm_ima_xattr_data can be used only for SHA1 digests). Signed-off-by: Roberto Sassu <roberto.sassu(a)huawei.com> --- security/integrity/ima/ima_appraise.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/security/integrity/ima/ima_appraise.c b/security/integrity/ima/ima_appraise.c index a11577147022..c6376ec28ccd 100644 --- a/security/integrity/ima/ima_appraise.c +++ b/security/integrity/ima/ima_appraise.c @@ -193,19 +193,19 @@ int ima_appraise_measurement(enum ima_hooks func, struct dentry *dentry = file_dentry(file); struct inode *inode = d_backing_inode(dentry); enum integrity_status status = INTEGRITY_UNKNOWN; - struct evm_ima_xattr_data digest_list_value; + char _buf[sizeof(struct evm_ima_xattr_data) + SHA512_DIGEST_SIZE]; int rc = xattr_len, hash_start = 0; if (!(inode->i_opflags & IOP_XATTR)) return INTEGRITY_UNKNOWN; if (rc == -ENODATA && found_digest && - !(file->f_mode && FMODE_CREATED)) { - digest_list_value.type = EVM_IMA_XATTR_DIGEST_LIST; - digest_list_value.digest[0] = found_digest->algo; - memcpy(&digest_list_value.digest[1], found_digest->digest, + !(file->f_mode & FMODE_CREATED)) { + xattr_value = (struct evm_ima_xattr_data *)_buf; + xattr_value->type = IMA_XATTR_DIGEST_NG; + xattr_value->digest[0] = found_digest->algo; + memcpy(&xattr_value->digest[1], found_digest->digest, hash_digest_size[found_digest->algo]); - xattr_value = &digest_list_value; rc = hash_digest_size[found_digest->algo] + 2; } @@ -283,7 +283,6 @@ int ima_appraise_measurement(enum ima_hooks func, status = INTEGRITY_PASS; break; } - if (!ima_appraise_no_metadata) { cause = "IMA-xattr-untrusted"; status = INTEGRITY_FAIL; -- 2.27.GIT

3 26

[PATCH 01/35] selinux: allow labeling before policy is loaded
by Roberto Sassu 06 Aug '20

06 Aug '20

From: Jonathan Lebon <jlebon(a)redhat.com> mainline inclusion from mainline-v5.5-rc1 commit 3e3e24b42043eceb97ed834102c2d094dfd7aaa6 category: bugfix --------------------------- Currently, the SELinux LSM prevents one from setting the `security.selinux` xattr on an inode without a policy first being loaded. However, this restriction is problematic: it makes it impossible to have newly created files with the correct label before actually loading the policy. This is relevant in distributions like Fedora, where the policy is loaded by systemd shortly after pivoting out of the initrd. In such instances, all files created prior to pivoting will be unlabeled. One then has to relabel them after pivoting, an operation which inherently races with other processes trying to access those same files. Going further, there are use cases for creating the entire root filesystem on first boot from the initrd (e.g. Container Linux supports this today[1], and we'd like to support it in Fedora CoreOS as well[2]). One can imagine doing this in two ways: at the block device level (e.g. laying down a disk image), or at the filesystem level. In the former, labeling can simply be part of the image. But even in the latter scenario, one still really wants to be able to set the right labels when populating the new filesystem. This patch enables this by changing behaviour in the following two ways: 1. allow `setxattr` on mounts without `SBLABEL_MNT` (which is all of them if no policy is loaded yet) 2. don't try to set the in-core inode SID if we're not initialized; instead leave it as `LABEL_INVALID` so that revalidation may be attempted at a later time Note the first hunk of this patch is functionally the same as a previously discussed one[3], though it was part of a larger series which wasn't accepted. Co-developed-by: Victor Kamensky <kamensky(a)cisco.com> Signed-off-by: Victor Kamensky <kamensky(a)cisco.com> Signed-off-by: Jonathan Lebon <jlebon(a)redhat.com> [1] https://coreos.com/os/docs/latest/root-filesystem-placement.html [2] https://github.com/coreos/fedora-coreos-tracker/issues/94 [3] https://www.spinics.net/lists/linux-initramfs/msg04593.html --- security/selinux/hooks.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index 452254fd89f8..931546d80211 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -3305,7 +3305,7 @@ static int selinux_inode_setxattr(struct dentry *dentry, const char *name, } sbsec = inode->i_sb->s_security; - if (!(sbsec->flags & SBLABEL_MNT)) + if (!(sbsec->flags & SBLABEL_MNT) && selinux_state.initialized) return -EOPNOTSUPP; if (!inode_owner_or_capable(inode)) @@ -3387,6 +3387,15 @@ static void selinux_inode_post_setxattr(struct dentry *dentry, const char *name, return; } + if (!selinux_state.initialized) { + /* If we haven't even been initialized, then we can't validate + * against a policy, so leave the label as invalid. It may + * resolve to a valid label on the next revalidation try if + * we've since initialized. + */ + return; + } + rc = security_context_to_sid_force(&selinux_state, value, size, &newsid); if (rc) { -- 2.27.GIT

3 40

[PATCH 1/5] cgroup: memcg: net: do not associate sock with unrelated cgroup
by Yang Yingliang 04 Aug '20

04 Aug '20

From: Shakeel Butt <shakeelb(a)google.com> [ Upstream commit e876ecc67db80dfdb8e237f71e5b43bb88ae549c ] We are testing network memory accounting in our setup and noticed inconsistent network memory usage and often unrelated cgroups network usage correlates with testing workload. On further inspection, it seems like mem_cgroup_sk_alloc() and cgroup_sk_alloc() are broken in irq context specially for cgroup v1. mem_cgroup_sk_alloc() and cgroup_sk_alloc() can be called in irq context and kind of assumes that this can only happen from sk_clone_lock() and the source sock object has already associated cgroup. However in cgroup v1, where network memory accounting is opt-in, the source sock can be unassociated with any cgroup and the new cloned sock can get associated with unrelated interrupted cgroup. Cgroup v2 can also suffer if the source sock object was created by process in the root cgroup or if sk_alloc() is called in irq context. The fix is to just do nothing in interrupt. WARNING: Please note that about half of the TCP sockets are allocated from the IRQ context, so, memory used by such sockets will not be accouted by the memcg. The stack trace of mem_cgroup_sk_alloc() from IRQ-context: CPU: 70 PID: 12720 Comm: ssh Tainted: 5.6.0-smp-DEV #1 Hardware name: ... Call Trace: <IRQ> dump_stack+0x57/0x75 mem_cgroup_sk_alloc+0xe9/0xf0 sk_clone_lock+0x2a7/0x420 inet_csk_clone_lock+0x1b/0x110 tcp_create_openreq_child+0x23/0x3b0 tcp_v6_syn_recv_sock+0x88/0x730 tcp_check_req+0x429/0x560 tcp_v6_rcv+0x72d/0xa40 ip6_protocol_deliver_rcu+0xc9/0x400 ip6_input+0x44/0xd0 ? ip6_protocol_deliver_rcu+0x400/0x400 ip6_rcv_finish+0x71/0x80 ipv6_rcv+0x5b/0xe0 ? ip6_sublist_rcv+0x2e0/0x2e0 process_backlog+0x108/0x1e0 net_rx_action+0x26b/0x460 __do_softirq+0x104/0x2a6 do_softirq_own_stack+0x2a/0x40 </IRQ> do_softirq.part.19+0x40/0x50 __local_bh_enable_ip+0x51/0x60 ip6_finish_output2+0x23d/0x520 ? ip6table_mangle_hook+0x55/0x160 __ip6_finish_output+0xa1/0x100 ip6_finish_output+0x30/0xd0 ip6_output+0x73/0x120 ? __ip6_finish_output+0x100/0x100 ip6_xmit+0x2e3/0x600 ? ipv6_anycast_cleanup+0x50/0x50 ? inet6_csk_route_socket+0x136/0x1e0 ? skb_free_head+0x1e/0x30 inet6_csk_xmit+0x95/0xf0 __tcp_transmit_skb+0x5b4/0xb20 __tcp_send_ack.part.60+0xa3/0x110 tcp_send_ack+0x1d/0x20 tcp_rcv_state_process+0xe64/0xe80 ? tcp_v6_connect+0x5d1/0x5f0 tcp_v6_do_rcv+0x1b1/0x3f0 ? tcp_v6_do_rcv+0x1b1/0x3f0 __release_sock+0x7f/0xd0 release_sock+0x30/0xa0 __inet_stream_connect+0x1c3/0x3b0 ? prepare_to_wait+0xb0/0xb0 inet_stream_connect+0x3b/0x60 __sys_connect+0x101/0x120 ? __sys_getsockopt+0x11b/0x140 __x64_sys_connect+0x1a/0x20 do_syscall_64+0x51/0x200 entry_SYSCALL_64_after_hwframe+0x44/0xa9 The stack trace of mem_cgroup_sk_alloc() from IRQ-context: Fixes: 2d7580738345 ("mm: memcontrol: consolidate cgroup socket tracking") Fixes: d979a39d7242 ("cgroup: duplicate cgroup reference when cloning sockets") Signed-off-by: Shakeel Butt <shakeelb(a)google.com> Reviewed-by: Roman Gushchin <guro(a)fb.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- kernel/cgroup/cgroup.c | 4 ++++ mm/memcontrol.c | 4 ++++ 2 files changed, 8 insertions(+) diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c index 78ef274b036e..94f4d6a73d70 100644 --- a/kernel/cgroup/cgroup.c +++ b/kernel/cgroup/cgroup.c @@ -5928,6 +5928,10 @@ void cgroup_sk_alloc(struct sock_cgroup_data *skcd) return; } + /* Don't associate the sock with unrelated interrupted task's cgroup. */ + if (in_interrupt()) + return; + rcu_read_lock(); while (true) { diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 5d62c1fada01..f90672e7b4e6 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -6328,6 +6328,10 @@ void mem_cgroup_sk_alloc(struct sock *sk) return; } + /* Do not associate the sock with unrelated interrupted task's memcg. */ + if (in_interrupt()) + return; + rcu_read_lock(); memcg = mem_cgroup_from_task(current); if (memcg == root_mem_cgroup) -- 2.25.1

1 4

[PATCH 1/6] net/hinic: Set net device link down when the chip fault
by Yang Yingliang 04 Aug '20

04 Aug '20

From: Chiqijun <chiqijun(a)huawei.com> driver inclusion category: bugfix bugzilla: 4472 ----------------------------------------------------------------------- Set the fault device link down so that the bond device can fail over to the backup device. Signed-off-by: Chiqijun <chiqijun(a)huawei.com> Reviewed-by: Zengweiliang <zengweiliang.zengweiliang(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/net/ethernet/huawei/hinic/hinic_hw.h | 4 +- .../net/ethernet/huawei/hinic/hinic_hwdev.c | 22 +++-- drivers/net/ethernet/huawei/hinic/hinic_lld.c | 83 +++++++++++++++++-- .../net/ethernet/huawei/hinic/hinic_main.c | 8 ++ 4 files changed, 100 insertions(+), 17 deletions(-) diff --git a/drivers/net/ethernet/huawei/hinic/hinic_hw.h b/drivers/net/ethernet/huawei/hinic/hinic_hw.h index 6afe31c9b80e..c4d48ccf8f2a 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_hw.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_hw.h @@ -565,7 +565,8 @@ union hinic_fault_hw_mgmt { struct hinic_fault_event { /* enum hinic_fault_type */ u8 type; - u8 rsvd0[3]; + u8 fault_level; /* sdk write fault level for uld event */ + u8 rsvd0[2]; union hinic_fault_hw_mgmt event; }; @@ -653,6 +654,7 @@ enum hinic_event_type { HINIC_EVENT_MCTP_GET_HOST_INFO, HINIC_EVENT_MULTI_HOST_MGMT, HINIC_EVENT_INIT_MIGRATE_PF, + HINIC_EVENT_MGMT_WATCHDOG_EVENT, }; struct hinic_event_info { diff --git a/drivers/net/ethernet/huawei/hinic/hinic_hwdev.c b/drivers/net/ethernet/huawei/hinic/hinic_hwdev.c index 058939d05346..2d6a547c4d34 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_hwdev.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_hwdev.c @@ -3537,9 +3537,10 @@ static void fault_event_handler(struct hinic_hwdev *hwdev, void *buf_in, struct hinic_cmd_fault_event *fault_event; struct hinic_event_info event_info; struct hinic_fault_info_node *fault_node; + u8 fault_level; if (in_size != sizeof(*fault_event)) { - sdk_err(hwdev->dev_hdl, "Invalid fault event report, length: %d, should be %ld.\n", + sdk_err(hwdev->dev_hdl, "Invalid fault event report, length: %d, should be %ld\n", in_size, sizeof(*fault_event)); return; } @@ -3547,11 +3548,16 @@ static void fault_event_handler(struct hinic_hwdev *hwdev, void *buf_in, fault_event = buf_in; fault_report_show(hwdev, &fault_event->event); + if (fault_event->event.type == HINIC_FAULT_SRC_HW_MGMT_CHIP) + fault_level = fault_event->event.event.chip.err_level; + else + fault_level = FAULT_LEVEL_FATAL; + if (hwdev->event_callback) { event_info.type = HINIC_EVENT_FAULT; memcpy(&event_info.info, &fault_event->event, sizeof(event_info.info)); - + event_info.info.fault_level = fault_level; hwdev->event_callback(hwdev->event_pri_handle, &event_info); } @@ -3567,11 +3573,7 @@ static void fault_event_handler(struct hinic_hwdev *hwdev, void *buf_in, else if (fault_event->event.type == FAULT_TYPE_PHY_FAULT) fault_node->info.fault_src = HINIC_FAULT_SRC_HW_PHY_FAULT; - if (fault_node->info.fault_src == HINIC_FAULT_SRC_HW_MGMT_CHIP) - fault_node->info.fault_lev = - fault_event->event.event.chip.err_level; - else - fault_node->info.fault_lev = FAULT_LEVEL_FATAL; + fault_node->info.fault_lev = fault_level; memcpy(&fault_node->info.fault_data.hw_mgmt, &fault_event->event.event, sizeof(union hinic_fault_hw_mgmt)); @@ -3811,10 +3813,16 @@ static void mgmt_watchdog_timeout_event_handler(struct hinic_hwdev *hwdev, void *buf_out, u16 *out_size) { struct hinic_fault_info_node *fault_node; + struct hinic_event_info event_info = { 0 }; sw_watchdog_timeout_info_show(hwdev, buf_in, in_size, buf_out, out_size); + if (hwdev->event_callback) { + event_info.type = HINIC_EVENT_MGMT_WATCHDOG_EVENT; + hwdev->event_callback(hwdev->event_pri_handle, &event_info); + } + /* refresh history fault info */ fault_node = kzalloc(sizeof(*fault_node), GFP_KERNEL); if (!fault_node) { diff --git a/drivers/net/ethernet/huawei/hinic/hinic_lld.c b/drivers/net/ethernet/huawei/hinic/hinic_lld.c index 044d0df996e5..412f4c8a93ea 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_lld.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_lld.c @@ -1793,18 +1793,11 @@ static void __multi_host_mgmt(struct hinic_pcidev *dev, } } -void hinic_event_process(void *adapter, struct hinic_event_info *event) +static void send_uld_dev_event(struct hinic_pcidev *dev, + struct hinic_event_info *event) { - struct hinic_pcidev *dev = adapter; enum hinic_service_type type; - if (event->type == HINIC_EVENT_FMW_ACT_NTC) - return hinic_sync_time_to_fmw(dev); - else if (event->type == HINIC_EVENT_MCTP_GET_HOST_INFO) - return __mctp_get_host_info(dev, &event->mctp_info); - else if (event->type == HINIC_EVENT_MULTI_HOST_MGMT) - return __multi_host_mgmt(dev, &event->mhost_mgmt); - for (type = SERVICE_T_NIC; type < SERVICE_T_MAX; type++) { if (test_and_set_bit(type, &dev->state)) { sdk_warn(&dev->pcidev->dev, "Event: 0x%x can't handler, %s is in detach\n", @@ -1819,6 +1812,78 @@ void hinic_event_process(void *adapter, struct hinic_event_info *event) } } +static void send_event_to_all_pf(struct hinic_pcidev *dev, + struct hinic_event_info *event) +{ + struct hinic_pcidev *des_dev = NULL; + + lld_dev_hold(); + list_for_each_entry(des_dev, &dev->chip_node->func_list, node) { + if (test_bit(HINIC_FUNC_IN_REMOVE, &des_dev->flag)) + continue; + + if (hinic_func_type(des_dev->hwdev) == TYPE_VF) + continue; + + send_uld_dev_event(des_dev, event); + } + lld_dev_put(); +} + +static void send_event_to_dst_pf(struct hinic_pcidev *dev, u16 func_id, + struct hinic_event_info *event) +{ + struct hinic_pcidev *des_dev = NULL; + + lld_dev_hold(); + list_for_each_entry(des_dev, &dev->chip_node->func_list, node) { + if (test_bit(HINIC_FUNC_IN_REMOVE, &des_dev->flag)) + continue; + + if (hinic_func_type(des_dev->hwdev) == TYPE_VF) + continue; + + if (hinic_global_func_id(des_dev->hwdev) == func_id) { + send_uld_dev_event(des_dev, event); + break; + } + } + lld_dev_put(); +} + +void hinic_event_process(void *adapter, struct hinic_event_info *event) +{ + struct hinic_pcidev *dev = adapter; + u16 func_id; + + switch (event->type) { + case HINIC_EVENT_FMW_ACT_NTC: + hinic_sync_time_to_fmw(dev); + break; + case HINIC_EVENT_MCTP_GET_HOST_INFO: + __mctp_get_host_info(dev, &event->mctp_info); + break; + case HINIC_EVENT_MULTI_HOST_MGMT: + __multi_host_mgmt(dev, &event->mhost_mgmt); + break; + case HINIC_EVENT_FAULT: + if (event->info.fault_level == FAULT_LEVEL_SERIOUS_FLR && + event->info.event.chip.func_id < HINIC_MAX_PF_NUM) { + func_id = event->info.event.chip.func_id; + send_event_to_dst_pf(adapter, func_id, event); + } else { + send_uld_dev_event(adapter, event); + } + break; + case HINIC_EVENT_MGMT_WATCHDOG_EVENT: + send_event_to_all_pf(adapter, event); + break; + default: + send_uld_dev_event(adapter, event); + break; + } +} + static int mapping_bar(struct pci_dev *pdev, struct hinic_pcidev *pci_adapter) { u32 db_dwqe_size; diff --git a/drivers/net/ethernet/huawei/hinic/hinic_main.c b/drivers/net/ethernet/huawei/hinic/hinic_main.c index 8dc7b555f0bf..e0635287bf05 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_main.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_main.c @@ -3026,8 +3026,13 @@ void nic_event(struct hinic_lld_dev *lld_dev, void *adapter, break; case HINIC_EVENT_HEART_LOST: hinic_heart_lost(nic_dev); + hinic_link_status_change(nic_dev, false); break; case HINIC_EVENT_FAULT: + if (event->info.fault_level == FAULT_LEVEL_SERIOUS_FLR && + event->info.event.chip.func_id == + hinic_global_func_id(lld_dev->hwdev)) + hinic_link_status_change(nic_dev, false); break; case HINIC_EVENT_DCB_STATE_CHANGE: if (nic_dev->default_cos_id == event->dcb_state.default_cos) @@ -3047,6 +3052,9 @@ void nic_event(struct hinic_lld_dev *lld_dev, void *adapter, case HINIC_EVENT_PORT_MODULE_EVENT: hinic_port_module_event_handler(nic_dev, event); break; + case HINIC_EVENT_MGMT_WATCHDOG_EVENT: + hinic_link_status_change(nic_dev, false); + break; default: break; } -- 2.25.1

1 5

2024

2023

2022

2021

2020

2019

Kernel