From: Ma Wupeng mawupeng1@huawei.com

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8USBA CVE: NA

--------------------------------

If system have no mirrored memory or use crashkernel.high while kernelcore=mirror is enabled in cmdline, during crashkernel, there will be limited mirrored memory and this usually lead to OOM.

To solve this problem, disable mirror feature during crashkernel.

Signed-off-by: Ma Wupeng mawupeng1@huawei.com --- mm/mm_init.c | 6 ++++++ 1 file changed, 6 insertions(+)

diff --git a/mm/mm_init.c b/mm/mm_init.c index fed4370b02e1..2318ae6de638 100644 --- a/mm/mm_init.c +++ b/mm/mm_init.c @@ -26,6 +26,7 @@ #include <linux/pgtable.h> #include <linux/swap.h> #include <linux/cma.h> +#include <linux/crash_dump.h> #include "internal.h" #include "slab.h" #include "shuffle.h" @@ -381,6 +382,11 @@ static void __init find_zone_movable_pfns_for_nodes(void) goto out; }

+ if (is_kdump_kernel()) { + pr_warn("The system is under kdump, ignore kernelcore=mirror.\n"); + goto out; + } + for_each_mem_region(r) { if (memblock_is_mirror(r)) continue;

From: Ma Wupeng mawupeng1@huawei.com

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8USBA CVE: NA

------------------------------------------

Adding reliable flag for user task. User task with reliable flag can only alloc memory from mirrored region. PF_RELIABLE is added to represent the task's reliable flag.

- For init task, which is regarded as special task which alloc memory from mirrored region.

- For normal user tasks, The reliable flag can be set via procfs interface shown as below and can be inherited via fork().

User can change a user task's reliable flag by

$ echo [0/1] > /proc/<pid>/reliable

and check a user task's reliable flag by

$ cat /proc/<pid>/reliable

Note, global init task's reliable file can not be accessed.

Signed-off-by: Peng Wu wupeng58@huawei.com Signed-off-by: Ma Wupeng mawupeng1@huawei.com --- Documentation/filesystems/proc.rst | 6 ++ fs/proc/Makefile | 1 + fs/proc/base.c | 9 +++ fs/proc/mem_reliable.c | 100 +++++++++++++++++++++++++++++ include/linux/mem_reliable.h | 3 + include/linux/sched.h | 5 ++ mm/page_alloc.c | 30 +++++++++ 7 files changed, 154 insertions(+) create mode 100644 fs/proc/mem_reliable.c

diff --git a/Documentation/filesystems/proc.rst b/Documentation/filesystems/proc.rst index 65e0556064d6..d7c614e3be67 100644 --- a/Documentation/filesystems/proc.rst +++ b/Documentation/filesystems/proc.rst @@ -163,6 +163,8 @@ usually fail with ESRCH. can be derived from smaps, but is faster and more convenient numa_maps An extension based on maps, showing the memory locality and binding policy as well as mem usage (in pages) of each mapping. + reliable Present with CONFIG_MEMORY_RELIABLE=y. Task reliable status + information ============= ===============================================================

For example, to get the status information of a process, all you have to do is @@ -674,6 +676,10 @@ Where: node locality page counters (N0 == node0, N1 == node1, ...) and the kernel page size, in KB, that is backing the mapping up.

+The /proc/pid/reliable is used to control user task's reliable status. +Task with this flag can only alloc memory from mirrored region. Global +init task's reliable flag can not be accessed. + 1.2 Kernel data ---------------

diff --git a/fs/proc/Makefile b/fs/proc/Makefile index bd08616ed8ba..70dca85a5861 100644 --- a/fs/proc/Makefile +++ b/fs/proc/Makefile @@ -34,3 +34,4 @@ proc-$(CONFIG_PROC_VMCORE) += vmcore.o proc-$(CONFIG_PRINTK) += kmsg.o proc-$(CONFIG_PROC_PAGE_MONITOR) += page.o proc-$(CONFIG_BOOT_CONFIG) += bootconfig.o +proc-$(CONFIG_MEMORY_RELIABLE) += mem_reliable.o diff --git a/fs/proc/base.c b/fs/proc/base.c index 96313167702d..e04b0126334f 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -2659,6 +2659,9 @@ static struct dentry *proc_pident_instantiate(struct dentry *dentry,

static bool proc_hide_pidents(const struct pid_entry *p) { + if (mem_reliable_hide_file(p->name)) + return true; + return false; }

@@ -3390,6 +3393,9 @@ static const struct pid_entry tgid_base_stuff[] = { ONE("oom_score", S_IRUGO, proc_oom_score), REG("oom_adj", S_IRUGO|S_IWUSR, proc_oom_adj_operations), REG("oom_score_adj", S_IRUGO|S_IWUSR, proc_oom_score_adj_operations), +#ifdef CONFIG_MEMORY_RELIABLE + REG("reliable", S_IRUGO|S_IWUSR, proc_reliable_operations), +#endif #ifdef CONFIG_AUDIT REG("loginuid", S_IWUSR|S_IRUGO, proc_loginuid_operations), REG("sessionid", S_IRUGO, proc_sessionid_operations), @@ -3739,6 +3745,9 @@ static const struct pid_entry tid_base_stuff[] = { ONE("oom_score", S_IRUGO, proc_oom_score), REG("oom_adj", S_IRUGO|S_IWUSR, proc_oom_adj_operations), REG("oom_score_adj", S_IRUGO|S_IWUSR, proc_oom_score_adj_operations), +#ifdef CONFIG_MEMORY_RELIABLE + REG("reliable", S_IRUGO|S_IWUSR, proc_reliable_operations), +#endif #ifdef CONFIG_AUDIT REG("loginuid", S_IWUSR|S_IRUGO, proc_loginuid_operations), REG("sessionid", S_IRUGO, proc_sessionid_operations), diff --git a/fs/proc/mem_reliable.c b/fs/proc/mem_reliable.c new file mode 100644 index 000000000000..52ac9ae79847 --- /dev/null +++ b/fs/proc/mem_reliable.c @@ -0,0 +1,100 @@ +// SPDX-License-Identifier: GPL-2.0-only +#include <linux/mem_reliable.h> + +#include "internal.h" + +static inline int reliable_check(struct task_struct *task, struct pid *pid) +{ + if (!mem_reliable_is_enabled()) + return -EACCES; + + if (is_global_init(task)) + return -EINVAL; + + if (!task->mm || (task->flags & PF_KTHREAD) || + (task->flags & PF_EXITING)) + return -EINVAL; + + return 0; +} + +static ssize_t reliable_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct task_struct *task = get_proc_task(file_inode(file)); + struct pid *pid = proc_pid(file_inode(file)); + char buffer[PROC_NUMBUF]; + size_t len; + short val; + int err; + + if (!task) + return -ESRCH; + + err = reliable_check(task, pid); + if (err) { + put_task_struct(task); + return err; + } + + val = task->flags & PF_RELIABLE ? 1 : 0; + put_task_struct(task); + len = snprintf(buffer, sizeof(buffer), "%hd\n", val); + return simple_read_from_buffer(buf, count, ppos, buffer, len); +} + +static ssize_t reliable_write(struct file *file, const char __user *buf, + size_t count, loff_t *ppos) +{ + struct task_struct *task = get_proc_task(file_inode(file)); + struct pid *pid = proc_pid(file_inode(file)); + char buffer[PROC_NUMBUF]; + int val; + int err; + + if (!task) + return -ESRCH; + + err = reliable_check(task, pid); + if (err) + goto out; + + memset(buffer, 0, sizeof(buffer)); + if (count > sizeof(buffer) - 1) + count = sizeof(buffer) - 1; + if (copy_from_user(buffer, buf, count)) { + err = -EFAULT; + goto out; + } + + err = kstrtoint(strstrip(buffer), 0, &val); + if (err) + goto out; + if (val != 0 && val != 1) { + err = -EINVAL; + goto out; + } + + if (val == 1) + task->flags |= PF_RELIABLE; + else + task->flags &= ~PF_RELIABLE; + +out: + put_task_struct(task); + return err < 0 ? err : count; +} + +struct file_operations proc_reliable_operations = { + .read = reliable_read, + .write = reliable_write, + .llseek = generic_file_llseek, +}; + +bool mem_reliable_hide_file(const char *name) +{ + if (!mem_reliable_is_enabled() && !strncmp("reliable", name, 8)) + return true; + + return false; +} diff --git a/include/linux/mem_reliable.h b/include/linux/mem_reliable.h index 5df1c157a2bd..019e17130c67 100644 --- a/include/linux/mem_reliable.h +++ b/include/linux/mem_reliable.h @@ -13,9 +13,11 @@ DECLARE_STATIC_KEY_FALSE(mem_reliable);

extern bool reliable_enabled; +extern struct file_operations proc_reliable_operations;

void mem_reliable_init(bool has_unmirrored_mem, unsigned long mirrored_sz); bool mem_reliable_status(void); +bool mem_reliable_hide_file(const char *name);

static inline bool mem_reliable_is_enabled(void) { @@ -74,6 +76,7 @@ static inline bool skip_non_mirrored_zone(gfp_t gfp, struct zoneref *z) return false; } static inline bool mem_reliable_status(void) { return false; } +static inline bool mem_reliable_hide_file(const char *name) { return false; } #endif

#endif diff --git a/include/linux/sched.h b/include/linux/sched.h index 9fdd08aa9626..97ad76e86ed6 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1756,6 +1756,11 @@ extern struct pid *cad_pid; #define PF_USER_WORKER 0x00004000 /* Kernel thread cloned from userspace thread */ #define PF_NOFREEZE 0x00008000 /* This thread should not be frozen */ #define PF__HOLE__00010000 0x00010000 +#ifdef CONFIG_MEMORY_RELIABLE +#define PF_RELIABLE PF__HOLE__00010000 /* Allocate from reliable memory */ +#else +#define PF_RELIABLE 0x00000000 +#endif #define PF_KSWAPD 0x00020000 /* I am kswapd */ #define PF_MEMALLOC_NOFS 0x00040000 /* All allocation requests will inherit GFP_NOFS */ #define PF_MEMALLOC_NOIO 0x00080000 /* All allocation requests will inherit GFP_NOIO */ diff --git a/mm/page_alloc.c b/mm/page_alloc.c index efde73e24b3b..9d158a1113f5 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -4530,6 +4530,33 @@ unsigned long __alloc_pages_bulk(gfp_t gfp, int preferred_nid, } EXPORT_SYMBOL_GPL(__alloc_pages_bulk);

+static inline void prepare_before_alloc(gfp_t *gfp_mask) +{ + bool zone_movable; + + if (!mem_reliable_is_enabled()) + return; + + /* + * memory reliable only handle memory allocation from movable zone + * (force alloc from non-movable zone or force alloc from movable + * zone) to get total isolation. + */ + zone_movable = gfp_zone(*gfp_mask & ~GFP_RELIABLE) == ZONE_MOVABLE; + if (!zone_movable) + goto clear_flag; + + if (!in_task()) + return; + + if ((current->flags & PF_RELIABLE) || is_global_init(current)) + *gfp_mask |= GFP_RELIABLE; + + return; +clear_flag: + *gfp_mask &= ~GFP_RELIABLE; +} + /* * This is the 'heart' of the zoned buddy allocator. */ @@ -4549,6 +4576,9 @@ struct page *__alloc_pages(gfp_t gfp, unsigned int order, int preferred_nid, return NULL;

gfp &= gfp_allowed_mask; + + prepare_before_alloc(&gfp); + /* * Apply scoped allocation constraints. This is mainly about GFP_NOFS * resp. GFP_NOIO which has to be inherited for all allocation requests

From: Chen Wandun chenwandun@huawei.com

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8USBA CVE: NA

--------------------------------

filemap_alloc_folio is used to alloc page cache in most file system, such as ext4, f2fs, so add GFP_RELIABLE flag to use reliable memory when alloc page cache.

Signed-off-by: Chen Wandun chenwandun@huawei.com --- include/linux/mem_reliable.h | 7 +++++++ mm/filemap.c | 2 ++ 2 files changed, 9 insertions(+)

diff --git a/include/linux/mem_reliable.h b/include/linux/mem_reliable.h index fd62c8ec38ed..bdd605bf2d45 100644 --- a/include/linux/mem_reliable.h +++ b/include/linux/mem_reliable.h @@ -69,6 +69,12 @@ static inline void shmem_prepare_alloc(gfp_t *gfp_mask) if (mem_reliable_is_enabled()) *gfp_mask |= GFP_RELIABLE; } + +static inline void filemap_prepare_alloc(gfp_t *gfp_mask) +{ + if (mem_reliable_is_enabled()) + *gfp_mask |= GFP_RELIABLE; +} #else #define reliable_enabled 0

@@ -84,6 +90,7 @@ static inline bool skip_non_mirrored_zone(gfp_t gfp, struct zoneref *z) static inline bool mem_reliable_status(void) { return false; } static inline bool mem_reliable_hide_file(const char *name) { return false; } static inline void shmem_prepare_alloc(gfp_t *gfp_mask) {} +static inline void filemap_prepare_alloc(gfp_t *gfp_mask) {} #endif

#endif diff --git a/mm/filemap.c b/mm/filemap.c index 485dfc8b0f9c..e75743ba4a98 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -963,6 +963,8 @@ struct folio *filemap_alloc_folio(gfp_t gfp, unsigned int order) int n; struct folio *folio;

+ filemap_prepare_alloc(&gfp); + if (cpuset_do_page_mem_spread()) { unsigned int cpuset_mems_cookie; do {

From: Ma Wupeng mawupeng1@huawei.com

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8USBA CVE: NA

--------------------------------

With mirrored feature enabled, memblock will prefer to alloc memory from mirrored memory in any case. Since mirrored region and non-mirrored region may have different capacity or bandwidth, memblock user may choose which region to alloc memory rather than choose the mirrored one by default.

To solve this problem, flag MEMBLOCK_NOMIRROR is introduced to alloc memory from non-mirrored region. Function memblock_alloc_range_nid_flags() is introduced to alloc memory with specify flag without fallback.

Signed-off-by: Ma Wupeng mawupeng1@huawei.com --- include/linux/memblock.h | 6 +++ mm/memblock.c | 92 +++++++++++++++++++++++++++++++++------- 2 files changed, 82 insertions(+), 16 deletions(-)

diff --git a/include/linux/memblock.h b/include/linux/memblock.h index ae3bde302f70..653c307f22c4 100644 --- a/include/linux/memblock.h +++ b/include/linux/memblock.h @@ -42,6 +42,7 @@ extern unsigned long long max_possible_pfn; * kernel resource tree. * @MEMBLOCK_RSRV_NOINIT: memory region for which struct pages are * not initialized (only for reserved regions). + * @MEMBLOCK_NOMIRROR: memory region for non-mirrored memory */ enum memblock_flags { MEMBLOCK_NONE = 0x0, /* No special request */ @@ -50,6 +51,7 @@ enum memblock_flags { MEMBLOCK_NOMAP = 0x4, /* don't add to kernel direct mapping */ MEMBLOCK_DRIVER_MANAGED = 0x8, /* always detected via a driver */ MEMBLOCK_RSRV_NOINIT = 0x10, /* don't initialize struct pages */ + MEMBLOCK_NOMIRROR = 0x100, /* alloc from non-mirrored region */ };

/** @@ -428,6 +430,10 @@ void *memblock_alloc_try_nid_raw(phys_addr_t size, phys_addr_t align, void *memblock_alloc_try_nid(phys_addr_t size, phys_addr_t align, phys_addr_t min_addr, phys_addr_t max_addr, int nid); +void *memblock_alloc_try_nid_raw_flags(phys_addr_t size, phys_addr_t align, + phys_addr_t min_addr, + phys_addr_t max_addr, int nid, + enum memblock_flags flags);

static __always_inline void *memblock_alloc(phys_addr_t size, phys_addr_t align) { diff --git a/mm/memblock.c b/mm/memblock.c index fd492e5bbdbc..e18a25f6ce04 100644 --- a/mm/memblock.c +++ b/mm/memblock.c @@ -1038,6 +1038,10 @@ static bool should_skip_region(struct memblock_type *type, if ((flags & MEMBLOCK_MIRROR) && !memblock_is_mirror(m)) return true;

+ /* skip mirror memory regions with MEMBLOCK_NOMIRROR */ + if ((flags & MEMBLOCK_NOMIRROR) && memblock_is_mirror(m)) + return true; + /* skip nomap memory unless we were asked for it explicitly */ if (!(flags & MEMBLOCK_NOMAP) && memblock_is_nomap(m)) return true; @@ -1381,13 +1385,14 @@ __next_mem_pfn_range_in_zone(u64 *idx, struct zone *zone, #endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */

/** - * memblock_alloc_range_nid - allocate boot memory block + * memblock_alloc_range_nid_flags - allocate boot memory block with specify flag * @size: size of memory block to be allocated in bytes * @align: alignment of the region and block's size * @start: the lower bound of the memory region to allocate (phys address) * @end: the upper bound of the memory region to allocate (phys address) * @nid: nid of the free area to find, %NUMA_NO_NODE for any node * @exact_nid: control the allocation fall back to other nodes + * @flags: alloc memory from specify memblock flag * * The allocation is performed from memory region limited by * memblock.current_limit if @end == %MEMBLOCK_ALLOC_ACCESSIBLE. @@ -1395,22 +1400,18 @@ __next_mem_pfn_range_in_zone(u64 *idx, struct zone *zone, * If the specified node can not hold the requested memory and @exact_nid * is false, the allocation falls back to any node in the system. * - * For systems with memory mirroring, the allocation is attempted first - * from the regions with mirroring enabled and then retried from any - * memory region. - * - * In addition, function using kmemleak_alloc_phys for allocated boot - * memory block, it is never reported as leaks. + * In addition, function sets the min_count to 0 using kmemleak_alloc_phys for + * allocated boot memory block, so that it is never reported as leaks. * * Return: * Physical address of allocated memory block on success, %0 on failure. */ -phys_addr_t __init memblock_alloc_range_nid(phys_addr_t size, +phys_addr_t __init memblock_alloc_range_nid_flags(phys_addr_t size, phys_addr_t align, phys_addr_t start, phys_addr_t end, int nid, - bool exact_nid) + bool exact_nid, + enum memblock_flags flags) { - enum memblock_flags flags = choose_memblock_flags(); phys_addr_t found;

if (WARN_ONCE(nid == MAX_NUMNODES, "Usage of MAX_NUMNODES is deprecated. Use NUMA_NO_NODE instead\n")) @@ -1471,6 +1472,41 @@ phys_addr_t __init memblock_alloc_range_nid(phys_addr_t size, return found; }

+/** + * memblock_alloc_range_nid - allocate boot memory block + * @size: size of memory block to be allocated in bytes + * @align: alignment of the region and block's size + * @start: the lower bound of the memory region to allocate (phys address) + * @end: the upper bound of the memory region to allocate (phys address) + * @nid: nid of the free area to find, %NUMA_NO_NODE for any node + * @exact_nid: control the allocation fall back to other nodes + * + * The allocation is performed from memory region limited by + * memblock.current_limit if @end == %MEMBLOCK_ALLOC_ACCESSIBLE. + * + * If the specified node can not hold the requested memory and @exact_nid + * is false, the allocation falls back to any node in the system. + * + * For systems with memory mirroring, the allocation is attempted first + * from the regions with mirroring enabled and then retried from any + * memory region. + * + * In addition, function using kmemleak_alloc_phys for allocated boot + * memory block, it is never reported as leaks. + * + * Return: + * Physical address of allocated memory block on success, %0 on failure. + */ +phys_addr_t __init memblock_alloc_range_nid(phys_addr_t size, + phys_addr_t align, phys_addr_t start, + phys_addr_t end, int nid, + bool exact_nid) +{ + return memblock_alloc_range_nid_flags(size, align, start, end, nid, + exact_nid, + choose_memblock_flags()); +} + /** * memblock_phys_alloc_range - allocate a memory block inside specified range * @size: size of memory block to be allocated in bytes @@ -1522,6 +1558,7 @@ phys_addr_t __init memblock_phys_alloc_try_nid(phys_addr_t size, phys_addr_t ali * @max_addr: the upper bound of the memory region to allocate (phys address) * @nid: nid of the free area to find, %NUMA_NO_NODE for any node * @exact_nid: control the allocation fall back to other nodes + * @flags: alloc memory from specify memblock flag * * Allocates memory block using memblock_alloc_range_nid() and * converts the returned physical address to virtual. @@ -1534,10 +1571,11 @@ phys_addr_t __init memblock_phys_alloc_try_nid(phys_addr_t size, phys_addr_t ali * Return: * Virtual address of allocated memory block on success, NULL on failure. */ -static void * __init memblock_alloc_internal( +static void * __init __memblock_alloc_internal( phys_addr_t size, phys_addr_t align, phys_addr_t min_addr, phys_addr_t max_addr, - int nid, bool exact_nid) + int nid, bool exact_nid, + enum memblock_flags flags) { phys_addr_t alloc;

@@ -1552,13 +1590,13 @@ static void * __init memblock_alloc_internal( if (max_addr > memblock.current_limit) max_addr = memblock.current_limit;

- alloc = memblock_alloc_range_nid(size, align, min_addr, max_addr, nid, - exact_nid); + alloc = memblock_alloc_range_nid_flags(size, align, min_addr, max_addr, nid, + exact_nid, flags);

/* retry allocation without lower limit */ if (!alloc && min_addr) - alloc = memblock_alloc_range_nid(size, align, 0, max_addr, nid, - exact_nid); + alloc = memblock_alloc_range_nid_flags(size, align, 0, max_addr, nid, + exact_nid, flags);

if (!alloc) return NULL; @@ -1566,6 +1604,15 @@ static void * __init memblock_alloc_internal( return phys_to_virt(alloc); }

+static void * __init memblock_alloc_internal( + phys_addr_t size, phys_addr_t align, + phys_addr_t min_addr, phys_addr_t max_addr, + int nid, bool exact_nid) +{ + return __memblock_alloc_internal(size, align, min_addr, max_addr, nid, + exact_nid, choose_memblock_flags()); +} + /** * memblock_alloc_exact_nid_raw - allocate boot memory block on the exact node * without zeroing memory @@ -1629,6 +1676,19 @@ void * __init memblock_alloc_try_nid_raw( false); }

+void * __init memblock_alloc_try_nid_raw_flags( + phys_addr_t size, phys_addr_t align, + phys_addr_t min_addr, phys_addr_t max_addr, + int nid, enum memblock_flags flags) +{ + memblock_dbg("%s: %llu bytes align=0x%llx nid=%d from=%pa max_addr=%pa %pS\n", + __func__, (u64)size, (u64)align, nid, &min_addr, + &max_addr, (void *)_RET_IP_); + + return __memblock_alloc_internal(size, align, min_addr, max_addr, nid, + false, flags); +} + /** * memblock_alloc_try_nid - allocate boot memory block * @size: size of memory block to be allocated in bytes

From: Peng Wu wupeng58@huawei.com

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8USBA CVE: NA

--------------------------------

Add kernel param reliable_debug to enable separate memory reliable features on demand: - Page cache will not use reliable memory when passing option "P" to reliable_debug in cmdline. - Shmem will not use reliable memory when passing opiton "S" to reliable_debug in cmdline.

Signed-off-by: Peng Wu wupeng58@huawei.com Signed-off-by: Ma Wupeng mawupeng1@huawei.com --- .../admin-guide/kernel-parameters.txt | 7 ++++ include/linux/mem_reliable.h | 29 +++++++++++++- mm/mem_reliable.c | 40 +++++++++++++++++++ mm/shmem.c | 3 ++ 4 files changed, 77 insertions(+), 2 deletions(-)

diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt index 01fc38217459..53b820d51822 100644 --- a/Documentation/admin-guide/kernel-parameters.txt +++ b/Documentation/admin-guide/kernel-parameters.txt @@ -5518,6 +5518,13 @@ [KNL, SMP] Set scheduler's default relax_domain_level. See Documentation/admin-guide/cgroup-v1/cpusets.rst.

+ reliable_debug= [ARM64] + Format: [P][,S] + Only works with CONFIG_MEMORY_RELIABLE and + "kernelcore=reliable" is configured. + P: Page cache does not use the reliable memory. + S: The shmem does not use the reliable memory. + reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory Format: <base1>,<size1>[,<base2>,<size2>,...] Reserve I/O ports or memory so the kernel won't use diff --git a/include/linux/mem_reliable.h b/include/linux/mem_reliable.h index bdd605bf2d45..c47f928cfa59 100644 --- a/include/linux/mem_reliable.h +++ b/include/linux/mem_reliable.h @@ -14,10 +14,13 @@ DECLARE_STATIC_KEY_FALSE(mem_reliable);

extern bool reliable_enabled; extern struct file_operations proc_reliable_operations; +extern bool shmem_reliable; +extern bool pagecache_reliable;

void mem_reliable_init(bool has_unmirrored_mem, unsigned long mirrored_sz); bool mem_reliable_status(void); bool mem_reliable_hide_file(const char *name); +void shmem_reliable_init(void);

static inline bool mem_reliable_is_enabled(void) { @@ -46,6 +49,16 @@ static inline bool folio_reliable(struct folio *folio) return folio_zonenum(folio) < ZONE_MOVABLE; }

+static inline bool shmem_reliable_is_enabled(void) +{ + return shmem_reliable; +} + +static inline bool filemap_reliable_is_enabled(void) +{ + return pagecache_reliable; +} + static inline bool skip_non_mirrored_zone(gfp_t gfp, struct zoneref *z) { if (!mem_reliable_is_enabled()) @@ -66,19 +79,30 @@ static inline bool skip_non_mirrored_zone(gfp_t gfp, struct zoneref *z)

static inline void shmem_prepare_alloc(gfp_t *gfp_mask) { - if (mem_reliable_is_enabled()) + if (!mem_reliable_is_enabled()) + return; + + if (shmem_reliable_is_enabled()) *gfp_mask |= GFP_RELIABLE; + else + *gfp_mask &= ~GFP_RELIABLE; }

static inline void filemap_prepare_alloc(gfp_t *gfp_mask) { - if (mem_reliable_is_enabled()) + if (!mem_reliable_is_enabled()) + return; + + if (filemap_reliable_is_enabled()) *gfp_mask |= GFP_RELIABLE; + else + *gfp_mask &= ~GFP_RELIABLE; } #else #define reliable_enabled 0

static inline bool mem_reliable_is_enabled(void) { return false; } +static inline bool filemap_reliable_is_enabled(void) { return false; } static inline void mem_reliable_init(bool has_unmirrored_mem, unsigned long mirrored_sz) {} static inline bool page_reliable(struct page *page) { return false; } @@ -91,6 +115,7 @@ static inline bool mem_reliable_status(void) { return false; } static inline bool mem_reliable_hide_file(const char *name) { return false; } static inline void shmem_prepare_alloc(gfp_t *gfp_mask) {} static inline void filemap_prepare_alloc(gfp_t *gfp_mask) {} +static inline void shmem_reliable_init(void) {} #endif

#endif diff --git a/mm/mem_reliable.c b/mm/mem_reliable.c index 53d11d48555e..c94e62407263 100644 --- a/mm/mem_reliable.c +++ b/mm/mem_reliable.c @@ -13,6 +13,8 @@ DEFINE_STATIC_KEY_FALSE(mem_reliable); EXPORT_SYMBOL_GPL(mem_reliable);

bool reliable_enabled; +bool shmem_reliable __read_mostly = true; +bool pagecache_reliable __read_mostly = true;

bool mem_reliable_status(void) { @@ -63,3 +65,41 @@ void mem_reliable_init(bool has_unmirrored_mem, unsigned long mirrored_sz)

pr_info("init succeed, mirrored memory size(%lu)\n", mirrored_sz); } + +void shmem_reliable_init(void) +{ + if (!mem_reliable_is_enabled() || !shmem_reliable_is_enabled()) + shmem_reliable = false; +} + +static int __init setup_reliable_debug(char *str) +{ + if (*str++ != '=' || !*str) + /* + * No options specified. + */ + goto out; + + /* + * Determine which debug features should be switched on + */ + for (; *str && *str != ','; str++) { + switch (*str) { + case 'P': + pagecache_reliable = false; + pr_info("disable page cache use reliable memory\n"); + break; + case 'S': + shmem_reliable = false; + pr_info("disable shmem use reliable memory\n"); + break; + default: + pr_err("reliable_debug option '%c' unknown. skipped\n", + *str); + } + } + +out: + return 1; +} +__setup("reliable_debug", setup_reliable_debug); diff --git a/mm/shmem.c b/mm/shmem.c index 71fd12ad473f..33f4a5d36f62 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -4627,6 +4627,9 @@ void __init shmem_init(void) else shmem_huge = SHMEM_HUGE_NEVER; /* just in case it was patched */ #endif + + shmem_reliable_init(); + return;

out1:

From: Ma Wupeng mawupeng1@huawei.com

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8USBA CVE: NA

--------------------------------

Count reliable shmem usage.

Shmem usage is counted during NR_SHMEM.

Signed-off-by: Ma Wupeng mawupeng1@huawei.com --- include/linux/mem_reliable.h | 10 ++++++++++ mm/filemap.c | 9 +++++++-- mm/mem_reliable.c | 7 ++++++- mm/migrate.c | 5 +++++ mm/shmem.c | 4 ++++ 5 files changed, 32 insertions(+), 3 deletions(-)

diff --git a/include/linux/mem_reliable.h b/include/linux/mem_reliable.h index 7e36ac6b4dd4..b2be7010c520 100644 --- a/include/linux/mem_reliable.h +++ b/include/linux/mem_reliable.h @@ -18,6 +18,7 @@ extern bool shmem_reliable; extern bool pagecache_reliable; extern struct percpu_counter pagecache_reliable_pages; extern struct percpu_counter anon_reliable_pages; +extern struct percpu_counter shmem_reliable_pages;

void mem_reliable_init(bool has_unmirrored_mem, unsigned long mirrored_sz); bool mem_reliable_status(void); @@ -113,6 +114,13 @@ static inline unsigned long task_reliable_used_pages(void)

return nr_pages; } + +static inline void shmem_reliable_folio_add(struct folio *folio, int nr_page) +{ + if (shmem_reliable_is_enabled() && folio_reliable(folio)) + percpu_counter_add(&shmem_reliable_pages, nr_page); +} + #else #define reliable_enabled 0

@@ -136,6 +144,8 @@ static inline void reliable_lru_add(enum lru_list lru, struct folio *folio, static inline void reliable_lru_add_batch(int zid, enum lru_list lru, int val) {} static inline bool mem_reliable_counter_initialized(void) { return false; } +static inline void shmem_reliable_folio_add(struct folio *folio, + int nr_page) {} #endif

#endif diff --git a/mm/filemap.c b/mm/filemap.c index e75743ba4a98..571a5c5cd372 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -185,6 +185,7 @@ static void filemap_unaccount_folio(struct address_space *mapping, __lruvec_stat_mod_folio(folio, NR_FILE_PAGES, -nr); if (folio_test_swapbacked(folio)) { __lruvec_stat_mod_folio(folio, NR_SHMEM, -nr); + shmem_reliable_folio_add(folio, -nr); if (folio_test_pmd_mappable(folio)) __lruvec_stat_mod_folio(folio, NR_SHMEM_THPS, -nr); } else if (folio_test_pmd_mappable(folio)) { @@ -830,10 +831,14 @@ void replace_page_cache_folio(struct folio *old, struct folio *new) __lruvec_stat_sub_folio(old, NR_FILE_PAGES); if (!folio_test_hugetlb(new)) __lruvec_stat_add_folio(new, NR_FILE_PAGES); - if (folio_test_swapbacked(old)) + if (folio_test_swapbacked(old)) { __lruvec_stat_sub_folio(old, NR_SHMEM); - if (folio_test_swapbacked(new)) + shmem_reliable_folio_add(old, -folio_nr_pages(old)); + } + if (folio_test_swapbacked(new)) { __lruvec_stat_add_folio(new, NR_SHMEM); + shmem_reliable_folio_add(new, folio_nr_pages(new)); + } xas_unlock_irq(&xas); if (free_folio) free_folio(old); diff --git a/mm/mem_reliable.c b/mm/mem_reliable.c index be2f3fdfcf1c..9251f82255db 100644 --- a/mm/mem_reliable.c +++ b/mm/mem_reliable.c @@ -19,6 +19,7 @@ bool shmem_reliable __read_mostly = true; bool pagecache_reliable __read_mostly = true; struct percpu_counter pagecache_reliable_pages; struct percpu_counter anon_reliable_pages; +struct percpu_counter shmem_reliable_pages;

bool mem_reliable_counter_initialized(void) { @@ -78,8 +79,12 @@ void mem_reliable_init(bool has_unmirrored_mem, unsigned long mirrored_sz)

void shmem_reliable_init(void) { - if (!mem_reliable_is_enabled() || !shmem_reliable_is_enabled()) + if (!mem_reliable_is_enabled() || !shmem_reliable_is_enabled()) { shmem_reliable = false; + return; + } + + percpu_counter_init(&shmem_reliable_pages, 0, GFP_KERNEL); }

void reliable_lru_add_batch(int zid, enum lru_list lru, int val) diff --git a/mm/migrate.c b/mm/migrate.c index 5aab4994c4b5..322c63e6f9be 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -465,6 +465,11 @@ int folio_migrate_mapping(struct address_space *mapping, xas_unlock(&xas); /* Leave irq disabled to prevent preemption while updating stats */

+ if (folio_test_swapbacked(folio) && !folio_test_swapcache(folio)) { + shmem_reliable_folio_add(folio, -nr); + shmem_reliable_folio_add(newfolio, nr); + } + /* * If moved to a different zone then also account * the page for that zone. Other VM counters will be diff --git a/mm/shmem.c b/mm/shmem.c index 33f4a5d36f62..0d6807b608ed 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -808,6 +808,7 @@ static int shmem_add_to_page_cache(struct folio *folio, mapping->nrpages += nr; __lruvec_stat_mod_folio(folio, NR_FILE_PAGES, nr); __lruvec_stat_mod_folio(folio, NR_SHMEM, nr); + shmem_reliable_folio_add(folio, nr); unlock: xas_unlock_irq(&xas); } while (xas_nomem(&xas, gfp)); @@ -839,6 +840,7 @@ static void shmem_delete_from_page_cache(struct folio *folio, void *radswap) mapping->nrpages -= nr; __lruvec_stat_mod_folio(folio, NR_FILE_PAGES, -nr); __lruvec_stat_mod_folio(folio, NR_SHMEM, -nr); + shmem_reliable_folio_add(folio, -nr); xa_unlock_irq(&mapping->i_pages); folio_put(folio); BUG_ON(error); @@ -1756,8 +1758,10 @@ static int shmem_replace_folio(struct folio **foliop, gfp_t gfp, mem_cgroup_migrate(old, new); __lruvec_stat_mod_folio(new, NR_FILE_PAGES, 1); __lruvec_stat_mod_folio(new, NR_SHMEM, 1); + shmem_reliable_folio_add(new, 1); __lruvec_stat_mod_folio(old, NR_FILE_PAGES, -1); __lruvec_stat_mod_folio(old, NR_SHMEM, -1); + shmem_reliable_folio_add(old, -1); } xa_unlock_irq(&swap_mapping->i_pages);

From: Ma Wupeng mawupeng1@huawei.com

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8USBA CVE: NA

--------------------------------

Add reliable limit for reliable user task, page cache and shmem.

1. For reliable user task:

There is a upper limit for all memory allocation if the following conditions are met: - gfp_zone(gfp & ~ GFP_RELIABLE) == ZONE_MOVABLE - gfp & GFP_RELIABLE is true

Init tasks will alloc memory from non-mirrored region if their allocation trigger limit.

The limit can be set or access via /proc/sys/vm/task_reliable_limit

2. For page cache:

This limit's default value is ULONG_MAX. User can update this value between current user used reliable memory size and total reliable memory size.

Add interface /proc/sys/vm/pagecache_reliable_limit to set the max size for reliable page cache, the max size can not beyond total reliable ram.

the whole reliable memory feature depend on kernelcore=mirror, and which depend on NUMA, so remove redundant code in UMA.

3. For shmem:

This limit is used to restrict the amount of mirrored memory by shmem. This memory allocation will return no memory if reliable fallback is off or fallback to non-mirrored region if reliable fallback on.

This limit can be set or access via /proc/sys/vm/shmem_reliable_bytes_limit. The default value of this limit is LONG_MAX. This limit can be set from 0 to the total size of mirrored memory.

Signed-off-by: Chen Wandun chenwandun@huawei.com Signed-off-by: Ma Wupeng mawupeng1@huawei.com --- include/linux/mem_reliable.h | 77 +++++++++++++++++++++--- mm/mem_reliable.c | 113 +++++++++++++++++++++++++++++++++++ mm/page_alloc.c | 87 +++++++++++++++++++++++++++ mm/shmem.c | 4 +- 4 files changed, 271 insertions(+), 10 deletions(-)

diff --git a/include/linux/mem_reliable.h b/include/linux/mem_reliable.h index 8dba07dc6983..22a62deb8274 100644 --- a/include/linux/mem_reliable.h +++ b/include/linux/mem_reliable.h @@ -7,8 +7,10 @@ #include <linux/stddef.h> #include <linux/gfp.h> #include <linux/mmzone.h> +#include <linux/oom.h> #include <linux/mm_types.h> #include <linux/sched.h> +#include <linux/percpu_counter.h>

DECLARE_STATIC_KEY_FALSE(mem_reliable);

@@ -19,6 +21,9 @@ extern bool pagecache_reliable; extern struct percpu_counter pagecache_reliable_pages; extern struct percpu_counter anon_reliable_pages; extern struct percpu_counter shmem_reliable_pages; +extern unsigned long task_reliable_limit __read_mostly; +extern unsigned long shmem_reliable_limit __read_mostly; +extern unsigned long pagecache_reliable_limit __read_mostly;

void mem_reliable_init(bool has_unmirrored_mem, unsigned long mirrored_sz); bool mem_reliable_status(void); @@ -28,6 +33,8 @@ void reliable_lru_add(enum lru_list lru, struct folio *folio, int val); void reliable_lru_add_batch(int zid, enum lru_list lru, int val); bool mem_reliable_counter_initialized(void); void reliable_report_meminfo(struct seq_file *m); +void mem_reliable_out_of_memory(gfp_t gfp_mask, unsigned int order, + int preferred_nid, nodemask_t *nodemask);

static inline bool mem_reliable_is_enabled(void) { @@ -84,26 +91,53 @@ static inline bool skip_non_mirrored_zone(gfp_t gfp, struct zoneref *z) return false; }

-static inline void shmem_prepare_alloc(gfp_t *gfp_mask) +static inline bool mem_reliable_shmem_limit_check(void) +{ + return percpu_counter_read_positive(&shmem_reliable_pages) < + (shmem_reliable_limit >> PAGE_SHIFT); +} + +/* + * Check if this memory allocation for shmem is allowed. + * Return false if limit is triggered. + */ +static inline bool shmem_prepare_alloc(gfp_t *gfp_mask) { if (!mem_reliable_is_enabled()) - return; + return true;

- if (shmem_reliable_is_enabled()) - *gfp_mask |= GFP_RELIABLE; - else + if (!shmem_reliable_is_enabled()) { *gfp_mask &= ~GFP_RELIABLE; + return true; + } + + if (mem_reliable_shmem_limit_check()) { + *gfp_mask |= GFP_RELIABLE; + return true; + } + + return false; }

static inline void filemap_prepare_alloc(gfp_t *gfp_mask) { + s64 nr_reliable = 0; + if (!mem_reliable_is_enabled()) return;

- if (filemap_reliable_is_enabled()) - *gfp_mask |= GFP_RELIABLE; - else + if (!filemap_reliable_is_enabled()) { *gfp_mask &= ~GFP_RELIABLE; + return; + } + + nr_reliable = percpu_counter_read_positive(&pagecache_reliable_pages); + if (nr_reliable > pagecache_reliable_limit >> PAGE_SHIFT) { + *gfp_mask &= ~GFP_RELIABLE; + return; + } + + *gfp_mask |= GFP_RELIABLE; }

static inline unsigned long task_reliable_used_pages(void) @@ -122,6 +156,21 @@ static inline void shmem_reliable_folio_add(struct folio *folio, int nr_page) percpu_counter_add(&shmem_reliable_pages, nr_page); }

+ +static inline bool reliable_mem_limit_check(unsigned long nr_page) +{ + return (task_reliable_used_pages() + nr_page) <= + (task_reliable_limit >> PAGE_SHIFT); +} + +static inline bool mem_reliable_should_reclaim(void) +{ + if (percpu_counter_sum_positive(&pagecache_reliable_pages) >= + MAX_ORDER_NR_PAGES) + return true; + + return false; +} #else #define reliable_enabled 0

@@ -137,7 +186,7 @@ static inline bool skip_non_mirrored_zone(gfp_t gfp, struct zoneref *z) } static inline bool mem_reliable_status(void) { return false; } static inline bool mem_reliable_hide_file(const char *name) { return false; } -static inline void shmem_prepare_alloc(gfp_t *gfp_mask) {} +static inline bool shmem_prepare_alloc(gfp_t *gfp_mask) { return true; } static inline void filemap_prepare_alloc(gfp_t *gfp_mask) {} static inline void shmem_reliable_init(void) {} static inline void reliable_lru_add(enum lru_list lru, struct folio *folio, @@ -148,6 +197,16 @@ static inline bool mem_reliable_counter_initialized(void) { return false; } static inline void shmem_reliable_folio_add(struct folio *folio, int nr_page) {} static inline void reliable_report_meminfo(struct seq_file *m) {} +static inline bool mem_reliable_shmem_limit_check(void) { return true; } +static inline bool reliable_mem_limit_check(unsigned long nr_page) +{ + return false; +} +static inline bool mem_reliable_should_reclaim(void) { return false; } +static inline void mem_reliable_out_of_memory(gfp_t gfp_mask, + unsigned int order, + int preferred_nid, + nodemask_t *nodemask) {} #endif

#endif diff --git a/mm/mem_reliable.c b/mm/mem_reliable.c index c046d8016228..ae654d6bb047 100644 --- a/mm/mem_reliable.c +++ b/mm/mem_reliable.c @@ -20,6 +20,10 @@ bool pagecache_reliable __read_mostly = true; struct percpu_counter pagecache_reliable_pages; struct percpu_counter anon_reliable_pages; struct percpu_counter shmem_reliable_pages; +unsigned long pagecache_reliable_limit = ULONG_MAX; +/* reliable user limit for user tasks with reliable flag */ +unsigned long task_reliable_limit = ULONG_MAX; +unsigned long shmem_reliable_limit = ULONG_MAX;

bool mem_reliable_counter_initialized(void) { @@ -117,11 +121,103 @@ void reliable_lru_add(enum lru_list lru, struct folio *folio, int val) } }

+static int reliable_pagecache_max_bytes_write(struct ctl_table *table, + int write, void __user *buffer, + size_t *length, loff_t *ppos) +{ + unsigned long old_value = pagecache_reliable_limit; + int ret; + + ret = proc_doulongvec_minmax(table, write, buffer, length, ppos); + if (!ret && write) { + if (pagecache_reliable_limit > + PAGES_TO_B(total_reliable_pages())) { + pagecache_reliable_limit = old_value; + return -EINVAL; + } + } + + return ret; +} + +static int reliable_limit_handler(struct ctl_table *table, int write, + void __user *buffer, size_t *length, + loff_t *ppos) +{ + unsigned long old = task_reliable_limit; + int ret; + + ret = proc_doulongvec_minmax(table, write, buffer, length, ppos); + if (!ret && write) { + if (task_reliable_limit > PAGES_TO_B(total_reliable_pages()) || + task_reliable_limit < + (task_reliable_used_pages() << PAGE_SHIFT)) { + task_reliable_limit = old; + return -EINVAL; + } + } + + return ret; +} + +#ifdef CONFIG_SHMEM +static int reliable_shmem_bytes_limit_handler(struct ctl_table *table, + int write, void __user *buffer, + size_t *length, loff_t *ppos) +{ + unsigned long *data_ptr = (unsigned long *)(table->data); + unsigned long old = *data_ptr; + int ret; + + ret = proc_doulongvec_minmax(table, write, buffer, length, ppos); + if (!ret && write) { + if (*data_ptr > PAGES_TO_B(total_reliable_pages())) { + *data_ptr = old; + return -EINVAL; + } + } + + return ret; +} +#endif + +static struct ctl_table reliable_ctl_table[] = { + { + .procname = "reliable_pagecache_max_bytes", + .data = &pagecache_reliable_limit, + .maxlen = sizeof(pagecache_reliable_limit), + .mode = 0644, + .proc_handler = reliable_pagecache_max_bytes_write, + }, + { + .procname = "task_reliable_limit", + .data = &task_reliable_limit, + .maxlen = sizeof(task_reliable_limit), + .mode = 0644, + .proc_handler = reliable_limit_handler, + }, +#ifdef CONFIG_SHMEM + { + .procname = "shmem_reliable_bytes_limit", + .data = &shmem_reliable_limit, + .maxlen = sizeof(shmem_reliable_limit), + .mode = 0644, + .proc_handler = reliable_shmem_bytes_limit_handler, + }, +#endif + {} +}; + static int __init reliable_sysctl_init(void) { if (!mem_reliable_is_enabled()) return 0;

+ if (!register_sysctl("vm", reliable_ctl_table)) { + pr_err("register sysctl failed."); + return -ENOMEM; + } + percpu_counter_init(&pagecache_reliable_pages, 0, GFP_KERNEL); percpu_counter_init(&anon_reliable_pages, 0, GFP_KERNEL);

@@ -167,6 +263,23 @@ void reliable_report_meminfo(struct seq_file *m) } }

+void mem_reliable_out_of_memory(gfp_t gfp, unsigned int order, + int preferred_nid, nodemask_t *nodemask) +{ + struct oom_control oc = { + .zonelist = node_zonelist(preferred_nid, gfp), + .nodemask = nodemask, + .memcg = NULL, + .gfp_mask = gfp, + .order = order, + }; + + if (!mutex_trylock(&oom_lock)) + return; + out_of_memory(&oc); + mutex_unlock(&oom_lock); +} + static int __init setup_reliable_debug(char *str) { if (*str++ != '=' || !*str) diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 9d158a1113f5..33d38ac47105 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -4557,6 +4557,89 @@ static inline void prepare_before_alloc(gfp_t *gfp_mask) *gfp_mask &= ~GFP_RELIABLE; }

+static inline long mem_reliable_direct_reclaim(int nr_pages, struct alloc_context *ac) +{ + long nr_reclaimed = 0; + + while (nr_reclaimed < nr_pages) { + /* try to free cache from reliable region */ + long progress = __perform_reclaim(GFP_KERNEL, 0, ac); + + nr_reclaimed += progress; + if (progress < SWAP_CLUSTER_MAX) + break; + } + + return nr_reclaimed; +} + +/* + * return true means memory allocation need retry and flag GFP_RELIABLE + * must be cleared. + */ +static inline bool check_after_alloc(gfp_t *gfp, unsigned int order, + int preferred_nid, + struct alloc_context *ac, + struct page **_page) +{ + int retry_times = MAX_RECLAIM_RETRIES; + int nr_pages; + + if (!mem_reliable_is_enabled()) + return false; + + if (!(*gfp & GFP_RELIABLE)) + return false; + + if (!*_page) + goto out_retry; + + if (*gfp & __GFP_NOFAIL || current->flags & PF_MEMALLOC) + goto out; + + /* percpu counter is not initialized, ignore limit check */ + if (!mem_reliable_counter_initialized()) + goto out; + +limit_check: + /* user task is limited by task_reliable_limit */ + if (!reliable_mem_limit_check(1 << order)) + goto out_free_page; + + goto out; + +out_free_page: + if (mem_reliable_should_reclaim() && retry_times--) { + nr_pages = mem_reliable_direct_reclaim(1 << order, ac); + if (nr_pages) + goto limit_check; + } + + __free_pages(*_page, order); + *_page = NULL; + +out_retry: + if (is_global_init(current)) { + *gfp &= ~GFP_RELIABLE; + return true; + } + + if (*gfp & (__GFP_NORETRY | __GFP_RETRY_MAYFAIL | __GFP_THISNODE)) + goto out; + + /* Coredumps can quickly deplete all memory reserves */ + if (current->flags & PF_DUMPCORE) + goto out; + /* The OOM killer will not help higher order allocs */ + if (order > PAGE_ALLOC_COSTLY_ORDER) + goto out; + + /* oom here */ + mem_reliable_out_of_memory(*gfp, order, preferred_nid, ac->nodemask); +out: + return false; +} + /* * This is the 'heart' of the zoned buddy allocator. */ @@ -4579,6 +4662,7 @@ struct page *__alloc_pages(gfp_t gfp, unsigned int order, int preferred_nid,

prepare_before_alloc(&gfp);

+retry: /* * Apply scoped allocation constraints. This is mainly about GFP_NOFS * resp. GFP_NOIO which has to be inherited for all allocation requests @@ -4621,6 +4705,9 @@ struct page *__alloc_pages(gfp_t gfp, unsigned int order, int preferred_nid, page = NULL; }

+ if (check_after_alloc(&gfp, order, preferred_nid, &ac, &page)) + goto retry; + trace_mm_page_alloc(page, order, alloc_gfp, ac.migratetype); kmsan_alloc_page(page, order, alloc_gfp);

diff --git a/mm/shmem.c b/mm/shmem.c index 0d6807b608ed..b44bfad90f8d 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -1679,7 +1679,8 @@ static struct folio *shmem_alloc_and_acct_folio(gfp_t gfp, struct inode *inode, if (err) goto failed;

- shmem_prepare_alloc(&gfp); + if (!shmem_prepare_alloc(&gfp)) + goto no_mem;

if (huge) folio = shmem_alloc_hugefolio(gfp, info, index); @@ -1691,6 +1692,7 @@ static struct folio *shmem_alloc_and_acct_folio(gfp_t gfp, struct inode *inode, return folio; }

+no_mem: err = -ENOMEM; shmem_inode_unacct_blocks(inode, nr); failed:

From: Ma Wupeng mawupeng1@huawei.com

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8USBA CVE: NA

--------------------------------

Counting reliable memory allocated by the reliable user tasks.

The policy of counting reliable memory usage is based on RSS statistics. Anywhere with counter of mm need count reliable pages too. Reliable page which is checked by page_reliable() need to update the reliable page counter by calling reliable_page_counter().

Updating the reliable pages should be considered if the following logic is added: - add_mm_counter - dec_mm_counter - inc_mm_counter_fast - dec_mm_counter_fast - rss[mm_counter(page)]

Signed-off-by: Ma Wupeng mawupeng1@huawei.com --- Documentation/filesystems/proc.rst | 2 ++ fs/proc/task_mmu.c | 1 + include/linux/mem_reliable.h | 39 ++++++++++++++++++++++++++++++ include/linux/mm_types.h | 4 +++ kernel/events/uprobes.c | 2 ++ mm/huge_memory.c | 4 +++ mm/khugepaged.c | 3 +++ mm/ksm.c | 2 ++ mm/mem_reliable.c | 11 +++++++++ mm/memory.c | 10 ++++++++ mm/migrate_device.c | 1 + mm/rmap.c | 7 ++++++ mm/swapfile.c | 1 + mm/userswap.c | 4 +++ 14 files changed, 91 insertions(+)

diff --git a/Documentation/filesystems/proc.rst b/Documentation/filesystems/proc.rst index 3bbae7b35bb0..56d9ba24bd14 100644 --- a/Documentation/filesystems/proc.rst +++ b/Documentation/filesystems/proc.rst @@ -197,6 +197,7 @@ read the file /proc/PID/status:: VmPTE: 20 kb VmSwap: 0 kB HugetlbPages: 0 kB + Reliable: 1608 kB CoreDumping: 0 THP_enabled: 1 Threads: 1 @@ -280,6 +281,7 @@ It's slow but very precise. VmSwap amount of swap used by anonymous private data (shmem swap usage is not included) HugetlbPages size of hugetlb memory portions + Reliable size of reliable memory used CoreDumping process's memory is currently being dumped (killing the process may lead to a corrupted core) THP_enabled process is allowed to use THP (returns 0 when diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index fe12b057d077..fac21bcba8a6 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -77,6 +77,7 @@ void task_mem(struct seq_file *m, struct mm_struct *mm) SEQ_PUT_DEC(" kB\nVmSwap:\t", swap); seq_puts(m, " kB\n"); hugetlb_report_usage(m, mm); + reliable_report_usage(m, mm); } #undef SEQ_PUT_DEC

diff --git a/include/linux/mem_reliable.h b/include/linux/mem_reliable.h index 09621a4d5ae5..15f69349a2a8 100644 --- a/include/linux/mem_reliable.h +++ b/include/linux/mem_reliable.h @@ -36,6 +36,7 @@ bool mem_reliable_counter_initialized(void); void reliable_report_meminfo(struct seq_file *m); void mem_reliable_out_of_memory(gfp_t gfp_mask, unsigned int order, int preferred_nid, nodemask_t *nodemask); +void reliable_report_usage(struct seq_file *m, struct mm_struct *mm);

static inline bool mem_reliable_is_enabled(void) { @@ -180,6 +181,38 @@ static inline bool mem_reliable_should_reclaim(void)

return false; } + +static inline void reliable_page_counter_inner(struct mm_struct *mm, int val) +{ + atomic_long_add(val, &mm->reliable_nr_page); + + /* + * Update reliable page counter to zero if underflows. + * + * Since reliable page counter is used for debug purpose only, + * there is no real function problem by doing this. + */ + if (unlikely(atomic_long_read(&mm->reliable_nr_page) < 0)) + atomic_long_set(&mm->reliable_nr_page, 0); +} + +static inline void add_reliable_folio_counter(struct folio *folio, + struct mm_struct *mm, int val) +{ + if (!folio_reliable(folio)) + return; + + reliable_page_counter_inner(mm, val); +} + +static inline void add_reliable_page_counter(struct page *page, + struct mm_struct *mm, int val) +{ + if (!page_reliable(page)) + return; + + reliable_page_counter_inner(mm, val); +} #else #define reliable_enabled 0

@@ -217,6 +250,12 @@ static inline void mem_reliable_out_of_memory(gfp_t gfp_mask, int preferred_nid, nodemask_t *nodemask) {} static inline bool reliable_allow_fb_enabled(void) { return false; } +static inline void add_reliable_page_counter(struct page *page, + struct mm_struct *mm, int val) {} +static inline void add_reliable_folio_counter(struct folio *folio, + struct mm_struct *mm, int val) {} +static inline void reliable_report_usage(struct seq_file *m, + struct mm_struct *mm) {} #endif

#endif diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 0bc3c7c191a5..a077f60819d9 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -936,6 +936,10 @@ struct mm_struct { #endif /* CONFIG_LRU_GEN */ #ifdef CONFIG_SHARE_POOL struct sp_group_master *sp_group_master; +#endif +#ifdef CONFIG_MEMORY_RELIABLE + /* total used reliable pages */ + atomic_long_t reliable_nr_page; #endif } __randomize_layout;

diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c index 3048589e2e85..789966263996 100644 --- a/kernel/events/uprobes.c +++ b/kernel/events/uprobes.c @@ -181,6 +181,7 @@ static int __replace_page(struct vm_area_struct *vma, unsigned long addr,

if (new_page) { folio_get(new_folio); + add_reliable_folio_counter(new_folio, mm, folio_nr_pages(new_folio)); page_add_new_anon_rmap(new_page, vma, addr); folio_add_lru_vma(new_folio, vma); } else @@ -198,6 +199,7 @@ static int __replace_page(struct vm_area_struct *vma, unsigned long addr, set_pte_at_notify(mm, addr, pvmw.pte, mk_pte(new_page, vma->vm_page_prot));

+ add_reliable_page_counter(old_page, mm, -1); page_remove_rmap(old_page, vma, false); if (!folio_mapped(old_folio)) folio_free_swap(old_folio); diff --git a/mm/huge_memory.c b/mm/huge_memory.c index a13418df1115..65421d751a9d 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -704,6 +704,7 @@ static vm_fault_t __do_huge_pmd_anonymous_page(struct vm_fault *vmf, set_pmd_at(vma->vm_mm, haddr, vmf->pmd, entry); update_mmu_cache_pmd(vma, vmf->address, vmf->pmd); add_mm_counter(vma->vm_mm, MM_ANONPAGES, HPAGE_PMD_NR); + add_reliable_page_counter(page, vma->vm_mm, HPAGE_PMD_NR); mm_inc_nr_ptes(vma->vm_mm); spin_unlock(vmf->ptl); count_vm_event(THP_FAULT_ALLOC); @@ -1143,6 +1144,7 @@ int copy_huge_pmd(struct mm_struct *dst_mm, struct mm_struct *src_mm, return -EAGAIN; } add_mm_counter(dst_mm, MM_ANONPAGES, HPAGE_PMD_NR); + add_reliable_page_counter(src_page, dst_mm, HPAGE_PMD_NR); out_zero_page: mm_inc_nr_ptes(dst_mm); pgtable_trans_huge_deposit(dst_mm, dst_pmd, pgtable); @@ -1687,6 +1689,7 @@ int zap_huge_pmd(struct mmu_gather *tlb, struct vm_area_struct *vma,

if (pmd_present(orig_pmd)) { page = pmd_page(orig_pmd); + add_reliable_page_counter(page, tlb->mm, -HPAGE_PMD_NR); page_remove_rmap(page, vma, true); VM_BUG_ON_PAGE(page_mapcount(page) < 0, page); VM_BUG_ON_PAGE(!PageHead(page), page); @@ -2103,6 +2106,7 @@ static void __split_huge_pmd_locked(struct vm_area_struct *vma, pmd_t *pmd, set_page_dirty(page); if (!PageReferenced(page) && pmd_young(old_pmd)) SetPageReferenced(page); + add_reliable_page_counter(page, mm, -HPAGE_PMD_NR); page_remove_rmap(page, vma, true); put_page(page); } diff --git a/mm/khugepaged.c b/mm/khugepaged.c index e28b79d02f31..6f2787d3b682 100644 --- a/mm/khugepaged.c +++ b/mm/khugepaged.c @@ -717,6 +717,7 @@ static void __collapse_huge_page_copy_succeeded(pte_t *pte, */ spin_lock(ptl); ptep_clear(vma->vm_mm, address, _pte); + add_reliable_page_counter(src_page, vma->vm_mm, 1); page_remove_rmap(src_page, vma, false); spin_unlock(ptl); free_page_and_swap_cache(src_page); @@ -1224,6 +1225,7 @@ static int collapse_huge_page(struct mm_struct *mm, unsigned long address,

spin_lock(pmd_ptl); BUG_ON(!pmd_none(*pmd)); + add_reliable_page_counter(hpage, vma->vm_mm, HPAGE_PMD_NR); page_add_new_anon_rmap(hpage, vma, address); lru_cache_add_inactive_or_unevictable(hpage, vma); pgtable_trans_huge_deposit(mm, pmd, pgtable); @@ -1631,6 +1633,7 @@ int collapse_pte_mapped_thp(struct mm_struct *mm, unsigned long addr, */ ptep_clear(mm, addr, pte); page_remove_rmap(page, vma, false); + add_reliable_page_counter(page, mm, -1); nr_ptes++; }

diff --git a/mm/ksm.c b/mm/ksm.c index 981af9c72e7a..7401a6c87a4b 100644 --- a/mm/ksm.c +++ b/mm/ksm.c @@ -1233,6 +1233,7 @@ static int replace_page(struct vm_area_struct *vma, struct page *page, */ if (!is_zero_pfn(page_to_pfn(kpage))) { get_page(kpage); + add_reliable_page_counter(kpage, mm, 1); page_add_anon_rmap(kpage, vma, addr, RMAP_NONE); newpte = mk_pte(kpage, vma->vm_page_prot); } else { @@ -1262,6 +1263,7 @@ static int replace_page(struct vm_area_struct *vma, struct page *page, */ ptep_clear_flush(vma, addr, ptep); set_pte_at_notify(mm, addr, ptep, newpte); + add_reliable_page_counter(page, mm, -1);

folio = page_folio(page); page_remove_rmap(page, vma, false); diff --git a/mm/mem_reliable.c b/mm/mem_reliable.c index a77d0cc4100d..de90d9b97243 100644 --- a/mm/mem_reliable.c +++ b/mm/mem_reliable.c @@ -316,3 +316,14 @@ static int __init setup_reliable_debug(char *str) return 1; } __setup("reliable_debug", setup_reliable_debug); + +#define SEQ_PUT_DEC(str, val) \ + seq_put_decimal_ull_width(m, str, (val) << (PAGE_SHIFT-10), 8) +void reliable_report_usage(struct seq_file *m, struct mm_struct *mm) +{ + if (!mem_reliable_is_enabled()) + return; + + SEQ_PUT_DEC("Reliable:\t", atomic_long_read(&mm->reliable_nr_page)); + seq_puts(m, " kB\n"); +} diff --git a/mm/memory.c b/mm/memory.c index 6569c9e97c9d..944c2ce2756b 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -963,6 +963,7 @@ copy_present_pte(struct vm_area_struct *dst_vma, struct vm_area_struct *src_vma, folio_get(folio); page_dup_file_rmap(page, false); rss[mm_counter_file(page)]++; + add_reliable_folio_counter(folio, dst_vma->vm_mm, 1); }

/* @@ -1463,6 +1464,7 @@ static unsigned long zap_pte_range(struct mmu_gather *tlb, mark_page_accessed(page); } rss[mm_counter(page)]--; + add_reliable_page_counter(page, mm, -1); if (!delay_rmap) { page_remove_rmap(page, vma, false); if (unlikely(page_mapcount(page) < 0)) @@ -1490,6 +1492,7 @@ static unsigned long zap_pte_range(struct mmu_gather *tlb, */ WARN_ON_ONCE(!vma_is_anonymous(vma)); rss[mm_counter(page)]--; + add_reliable_page_counter(page, mm, -1); if (is_device_private_entry(entry)) page_remove_rmap(page, vma, false); put_page(page); @@ -3166,10 +3169,13 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf) dec_mm_counter(mm, mm_counter_file(&old_folio->page)); inc_mm_counter(mm, MM_ANONPAGES); } + add_reliable_folio_counter(old_folio, mm, -1); } else { ksm_might_unmap_zero_page(mm, vmf->orig_pte); inc_mm_counter(mm, MM_ANONPAGES); } + + add_reliable_folio_counter(new_folio, mm, 1); flush_cache_page(vma, vmf->address, pte_pfn(vmf->orig_pte)); entry = mk_pte(&new_folio->page, vma->vm_page_prot); entry = pte_sw_mkyoung(entry); @@ -4023,6 +4029,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf) folio_free_swap(folio);

inc_mm_counter(vma->vm_mm, MM_ANONPAGES); + add_reliable_folio_counter(folio, vma->vm_mm, 1); dec_mm_counter(vma->vm_mm, MM_SWAPENTS); pte = mk_pte(page, vma->vm_page_prot);

@@ -4198,6 +4205,7 @@ static vm_fault_t do_anonymous_page(struct vm_fault *vmf) }

inc_mm_counter(vma->vm_mm, MM_ANONPAGES); + add_reliable_folio_counter(folio, vma->vm_mm, 1); folio_add_new_anon_rmap(folio, vma, vmf->address); folio_add_lru_vma(folio, vma); setpte: @@ -4340,6 +4348,7 @@ vm_fault_t do_set_pmd(struct vm_fault *vmf, struct page *page) entry = maybe_pmd_mkwrite(pmd_mkdirty(entry), vma);

add_mm_counter(vma->vm_mm, mm_counter_file(page), HPAGE_PMD_NR); + add_reliable_page_counter(page, vma->vm_mm, HPAGE_PMD_NR); page_add_file_rmap(page, vma, true);

/* @@ -4396,6 +4405,7 @@ void set_pte_range(struct vm_fault *vmf, struct folio *folio, if (unlikely(uffd_wp)) entry = pte_mkuffd_wp(entry); /* copy-on-write page */ + add_reliable_folio_counter(folio, vma->vm_mm, nr); if (write && !(vma->vm_flags & VM_SHARED)) { add_mm_counter(vma->vm_mm, MM_ANONPAGES, nr); VM_BUG_ON_FOLIO(nr != 1, folio); diff --git a/mm/migrate_device.c b/mm/migrate_device.c index 8ac1f79f754a..2e5bce2f1cb9 100644 --- a/mm/migrate_device.c +++ b/mm/migrate_device.c @@ -652,6 +652,7 @@ static void migrate_vma_insert_page(struct migrate_vma *migrate, goto unlock_abort;

inc_mm_counter(mm, MM_ANONPAGES); + add_reliable_page_counter(page, mm, 1); page_add_new_anon_rmap(page, vma, addr); if (!is_zone_device_page(page)) lru_cache_add_inactive_or_unevictable(page, vma); diff --git a/mm/rmap.c b/mm/rmap.c index 9f795b93cf40..93ea81fe5180 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -1636,6 +1636,7 @@ static bool try_to_unmap_one(struct folio *folio, struct vm_area_struct *vma, hsz); } else { dec_mm_counter(mm, mm_counter(&folio->page)); + add_reliable_page_counter(&folio->page, mm, -1); set_pte_at(mm, address, pvmw.pte, pteval); }

@@ -1651,6 +1652,7 @@ static bool try_to_unmap_one(struct folio *folio, struct vm_area_struct *vma, * copied pages. */ dec_mm_counter(mm, mm_counter(&folio->page)); + add_reliable_page_counter(&folio->page, mm, -1); } else if (folio_test_anon(folio)) { swp_entry_t entry = page_swap_entry(subpage); pte_t swp_pte; @@ -1693,6 +1695,7 @@ static bool try_to_unmap_one(struct folio *folio, struct vm_area_struct *vma, if (ref_count == 1 + map_count && !folio_test_dirty(folio)) { dec_mm_counter(mm, MM_ANONPAGES); + add_reliable_folio_counter(folio, mm, -1); goto discard; }

@@ -1737,6 +1740,7 @@ static bool try_to_unmap_one(struct folio *folio, struct vm_area_struct *vma, spin_unlock(&mmlist_lock); } dec_mm_counter(mm, MM_ANONPAGES); + add_reliable_folio_counter(folio, mm, -1); inc_mm_counter(mm, MM_SWAPENTS); swp_pte = swp_entry_to_pte(entry); if (anon_exclusive) @@ -1759,6 +1763,7 @@ static bool try_to_unmap_one(struct folio *folio, struct vm_area_struct *vma, * See Documentation/mm/mmu_notifier.rst */ dec_mm_counter(mm, mm_counter_file(&folio->page)); + add_reliable_folio_counter(folio, mm, -1); } discard: page_remove_rmap(subpage, vma, folio_test_hugetlb(folio)); @@ -2033,6 +2038,7 @@ static bool try_to_migrate_one(struct folio *folio, struct vm_area_struct *vma, hsz); } else { dec_mm_counter(mm, mm_counter(&folio->page)); + add_reliable_page_counter(&folio->page, mm, -1); set_pte_at(mm, address, pvmw.pte, pteval); }

@@ -2048,6 +2054,7 @@ static bool try_to_migrate_one(struct folio *folio, struct vm_area_struct *vma, * copied pages. */ dec_mm_counter(mm, mm_counter(&folio->page)); + add_reliable_page_counter(&folio->page, mm, -1); } else { swp_entry_t entry; pte_t swp_pte; diff --git a/mm/swapfile.c b/mm/swapfile.c index 4bc70f459164..99b54ff74973 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -1792,6 +1792,7 @@ static int unuse_pte(struct vm_area_struct *vma, pmd_t *pmd,

dec_mm_counter(vma->vm_mm, MM_SWAPENTS); inc_mm_counter(vma->vm_mm, MM_ANONPAGES); + add_reliable_page_counter(page, vma->vm_mm, 1); get_page(page); if (page == swapcache) { rmap_t rmap_flags = RMAP_NONE; diff --git a/mm/userswap.c b/mm/userswap.c index 18c99c2a0fc7..e76e9d7a40de 100644 --- a/mm/userswap.c +++ b/mm/userswap.c @@ -162,6 +162,7 @@ static int uswap_unmap_anon_page(struct mm_struct *mm, SWP_USERSWAP_ENTRY, page_to_pfn(page))));

dec_mm_counter(mm, MM_ANONPAGES); + add_reliable_page_counter(page, mm, -1); page_remove_rmap(page, vma, false); page->mapping = NULL;

@@ -192,6 +193,7 @@ static unsigned long vm_insert_anon_page(struct vm_area_struct *vma, }

inc_mm_counter(mm, MM_ANONPAGES); + add_reliable_page_counter(page, mm, 1); page_add_new_anon_rmap(page, vma, addr); dst_pte = mk_pte(page, vma->vm_page_prot); if (vma->vm_flags & VM_WRITE) @@ -217,6 +219,7 @@ static void uswap_map_anon_page(struct mm_struct *mm, flush_cache_page(vma, addr, pte_pfn(*pte)); set_pte_at(mm, addr, pte, old_pte); inc_mm_counter(mm, MM_ANONPAGES); + add_reliable_page_counter(page, mm, 1); page_add_new_anon_rmap(page, vma, addr); pte_unmap_unlock(pte, ptl); } @@ -531,6 +534,7 @@ int mfill_atomic_pte_nocopy(struct mm_struct *mm, pmd_t *dst_pmd, }

inc_mm_counter(mm, MM_ANONPAGES); + add_reliable_page_counter(page, mm, 1); page_add_new_anon_rmap(page, dst_vma, dst_addr); set_pte_at(mm, dst_addr, pte, dst_pte);

From: Ma Wupeng mawupeng1@huawei.com

hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I8USBA CVE: NA

--------------------------------

Show debug info about memory reliable if oom occurs.

Signed-off-by: Ma Wupeng mawupeng1@huawei.com --- mm/mem_reliable.c | 41 ++++++++++++++++++++++++++++++++++------- mm/show_mem.c | 1 + 2 files changed, 35 insertions(+), 7 deletions(-)

diff --git a/mm/mem_reliable.c b/mm/mem_reliable.c index c20a8e72b3e2..c28b4ebe71c9 100644 --- a/mm/mem_reliable.c +++ b/mm/mem_reliable.c @@ -346,26 +346,49 @@ static void mem_reliable_feature_disable(int idx) pr_info("%s is disabled\n", str); }

+#define TO_KB(bytes) ((bytes) >> 10) #define PAGES_TO_KB(n_pages) ((n_pages) << (PAGE_SHIFT - 10))

+#define SEQ_printf(m, x...) \ +do { \ + if (m) \ + seq_printf(m, x); \ + else \ + pr_info(x); \ +} while (0) + +static void reliable_show_limits(void) +{ + SEQ_printf(NULL, "ReliableTaskLimit: %lu kB\n", + TO_KB(task_reliable_limit)); + + if (shmem_reliable_is_enabled()) + SEQ_printf(NULL, "ReliableShmemLimit: %lu kB\n", + TO_KB(shmem_reliable_limit)); + + if (filemap_reliable_is_enabled()) + SEQ_printf(NULL, "ReliableFileCacheLimit: %lu kB\n", + TO_KB(pagecache_reliable_limit)); +} + void reliable_report_meminfo(struct seq_file *m) { if (!mem_reliable_is_enabled()) return;

- seq_printf(m, "ReliableTotal: %8lu kB\n", + SEQ_printf(m, "ReliableTotal: %8lu kB\n", PAGES_TO_KB(total_reliable_pages())); - seq_printf(m, "ReliableUsed: %8lu kB\n", + SEQ_printf(m, "ReliableUsed: %8lu kB\n", PAGES_TO_KB(used_reliable_pages())); - seq_printf(m, "ReliableTaskUsed: %8lu kB\n", + SEQ_printf(m, "ReliableTaskUsed: %8lu kB\n", PAGES_TO_KB(task_reliable_used_pages())); - seq_printf(m, "ReliableBuddyMem: %8lu kB\n", + SEQ_printf(m, "ReliableBuddyMem: %8lu kB\n", PAGES_TO_KB(free_reliable_pages()));

if (shmem_reliable_is_enabled()) { unsigned long shmem_pages = (unsigned long)percpu_counter_sum( &shmem_reliable_pages); - seq_printf(m, "ReliableShmem: %8lu kB\n", + SEQ_printf(m, "ReliableShmem: %8lu kB\n", PAGES_TO_KB(shmem_pages)); }

@@ -375,13 +398,17 @@ void reliable_report_meminfo(struct seq_file *m)

num += global_node_page_state(NR_LRU_BASE + LRU_ACTIVE_FILE); num += global_node_page_state(NR_LRU_BASE + LRU_INACTIVE_FILE); - seq_printf(m, "FileCache: %8lu kB\n", PAGES_TO_KB(num)); + SEQ_printf(m, "FileCache: %8lu kB\n", PAGES_TO_KB(num));

nr_reliable_pages = percpu_counter_sum_positive(&pagecache_reliable_pages); - seq_printf(m, "ReliableFileCache: %8lu kB\n", + SEQ_printf(m, "ReliableFileCache: %8lu kB\n", PAGES_TO_KB(nr_reliable_pages)); } + + /* show limit info during oom */ + if (!m) + reliable_show_limits(); }

void mem_reliable_out_of_memory(gfp_t gfp, unsigned int order, diff --git a/mm/show_mem.c b/mm/show_mem.c index 4b888b18bdde..5604925fb0b4 100644 --- a/mm/show_mem.c +++ b/mm/show_mem.c @@ -426,4 +426,5 @@ void __show_mem(unsigned int filter, nodemask_t *nodemask, int max_zone_idx) #ifdef CONFIG_MEMORY_FAILURE printk("%lu pages hwpoisoned\n", atomic_long_read(&num_poisoned_pages)); #endif + reliable_report_meminfo(NULL); }