August 2023 - Kernel - mailweb.openeuler.org

[PATCH openEuler-1.0-LTS] ip6mr: Fix skb_under_panic in ip6mr_cache_report()
by Yue Haibing 26 Aug '23

26 Aug '23

mainline inclusion from mainline-v6.5-rc3 commit 30e0191b16e8a58e4620fa3e2839ddc7b9d4281c category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I7W5X8 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… --------------------------- skbuff: skb_under_panic: text:ffffffff88771f69 len:56 put:-4 head:ffff88805f86a800 data:ffff887f5f86a850 tail:0x88 end:0x2c0 dev:pim6reg ------------[ cut here ]------------ kernel BUG at net/core/skbuff.c:192! invalid opcode: 0000 [#1] PREEMPT SMP KASAN CPU: 2 PID: 22968 Comm: kworker/2:11 Not tainted 6.5.0-rc3-00044-g0a8db05b571a #236 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014 Workqueue: ipv6_addrconf addrconf_dad_work RIP: 0010:skb_panic+0x152/0x1d0 Call Trace: <TASK> skb_push+0xc4/0xe0 ip6mr_cache_report+0xd69/0x19b0 reg_vif_xmit+0x406/0x690 dev_hard_start_xmit+0x17e/0x6e0 __dev_queue_xmit+0x2d6a/0x3d20 vlan_dev_hard_start_xmit+0x3ab/0x5c0 dev_hard_start_xmit+0x17e/0x6e0 __dev_queue_xmit+0x2d6a/0x3d20 neigh_connected_output+0x3ed/0x570 ip6_finish_output2+0x5b5/0x1950 ip6_finish_output+0x693/0x11c0 ip6_output+0x24b/0x880 NF_HOOK.constprop.0+0xfd/0x530 ndisc_send_skb+0x9db/0x1400 ndisc_send_rs+0x12a/0x6c0 addrconf_dad_completed+0x3c9/0xea0 addrconf_dad_work+0x849/0x1420 process_one_work+0xa22/0x16e0 worker_thread+0x679/0x10c0 ret_from_fork+0x28/0x60 ret_from_fork_asm+0x11/0x20 When setup a vlan device on dev pim6reg, DAD ns packet may sent on reg_vif_xmit(). reg_vif_xmit() ip6mr_cache_report() skb_push(skb, -skb_network_offset(pkt));//skb_network_offset(pkt) is 4 And skb_push declared as: void *skb_push(struct sk_buff *skb, unsigned int len); skb->data -= len; //0xffff88805f86a84c - 0xfffffffc = 0xffff887f5f86a850 skb->data is set to 0xffff887f5f86a850, which is invalid mem addr, lead to skb_push() fails. Fixes: 14fb64e1f449 ("[IPV6] MROUTE: Support PIM-SM (SSM).") Signed-off-by: Yue Haibing <yuehaibing(a)huawei.com> Reviewed-by: Eric Dumazet <edumazet(a)google.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Yue Haibing <yuehaibing(a)huawei.com> --- net/ipv6/ip6mr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index d392a72cd784..c94a06b11141 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -1070,7 +1070,7 @@ static int ip6mr_cache_report(struct mr_table *mrt, struct sk_buff *pkt, And all this only to mangle msg->im6_msgtype and to set msg->im6_mbz to "mbz" :-) */ - skb_push(skb, -skb_network_offset(pkt)); + __skb_pull(skb, skb_network_offset(pkt)); skb_push(skb, sizeof(*msg)); skb_reset_transport_header(skb); -- 2.34.1

2 1

[PATCH OLK-5.10] SUNRPC: don't pause on incomplete allocation
by ZhaoLong Wang 26 Aug '23

26 Aug '23

From: NeilBrown <neilb(a)suse.de> mainline inclusion from mainline-5.15-rc1 commit e38b3f20059426a0adbde014ff71071739ab5226 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I7RRY4 CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… ------------------------------------------------- alloc_pages_bulk_array() attempts to allocate at least one page based on the provided pages, and then opportunistically allocates more if that can be done without dropping the spinlock. So if it returns fewer than requested, that could just mean that it needed to drop the lock. In that case, try again immediately. Only pause for a time if no progress could be made. Reported-and-tested-by: Mike Javorski <mike.javorski(a)gmail.com> Reported-and-tested-by: Lothar Paltins <lopa(a)mailbox.org> Fixes: f6e70aab9dfe ("SUNRPC: refresh rq_pages using a bulk page allocator") Signed-off-by: NeilBrown <neilb(a)suse.de> Acked-by: Mel Gorman <mgorman(a)suse.com> Signed-off-by: Chuck Lever <chuck.lever(a)oracle.com> Signed-off-by: ZhaoLong Wang <wangzhaolong1(a)huawei.com> Conflicts: net/sunrpc/svc_xprt.c --- net/sunrpc/svc_xprt.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index 362487f3a6de..f56b4df6ce28 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -643,7 +643,7 @@ static int svc_alloc_arg(struct svc_rqst *rqstp) { struct svc_serv *serv = rqstp->rq_server; struct xdr_buf *arg = &rqstp->rq_arg; - unsigned long pages, filled; + unsigned long pages, filled, ret; pages = (serv->sv_max_mesg + 2 * PAGE_SIZE) >> PAGE_SHIFT; if (pages > RPCSVC_MAXPAGES) { @@ -653,11 +653,12 @@ static int svc_alloc_arg(struct svc_rqst *rqstp) pages = RPCSVC_MAXPAGES; } - for (;;) { - filled = alloc_pages_bulk_array(GFP_KERNEL, pages, - rqstp->rq_pages); - if (filled == pages) - break; + for (filled = 0; filled < pages; filled = ret) { + ret = alloc_pages_bulk_array(GFP_KERNEL, pages, + rqstp->rq_pages); + if (ret > filled) + /* Made progress, don't sleep yet */ + continue; set_current_state(TASK_INTERRUPTIBLE); if (signalled() || kthread_should_stop()) { -- 2.39.2

2 1

[PATCH openEuler-1.0-LTS] ip6mr: Fix skb_under_panic in ip6mr_cache_report()
by Yue Haibing 26 Aug '23

26 Aug '23

mainline inclusion from mainline-v6.5-rc3 commit 30e0191b16e8a58e4620fa3e2839ddc7b9d4281c category: bugfix bugzilla: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… --------------------------- skbuff: skb_under_panic: text:ffffffff88771f69 len:56 put:-4 head:ffff88805f86a800 data:ffff887f5f86a850 tail:0x88 end:0x2c0 dev:pim6reg ------------[ cut here ]------------ kernel BUG at net/core/skbuff.c:192! invalid opcode: 0000 [#1] PREEMPT SMP KASAN CPU: 2 PID: 22968 Comm: kworker/2:11 Not tainted 6.5.0-rc3-00044-g0a8db05b571a #236 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014 Workqueue: ipv6_addrconf addrconf_dad_work RIP: 0010:skb_panic+0x152/0x1d0 Call Trace: <TASK> skb_push+0xc4/0xe0 ip6mr_cache_report+0xd69/0x19b0 reg_vif_xmit+0x406/0x690 dev_hard_start_xmit+0x17e/0x6e0 __dev_queue_xmit+0x2d6a/0x3d20 vlan_dev_hard_start_xmit+0x3ab/0x5c0 dev_hard_start_xmit+0x17e/0x6e0 __dev_queue_xmit+0x2d6a/0x3d20 neigh_connected_output+0x3ed/0x570 ip6_finish_output2+0x5b5/0x1950 ip6_finish_output+0x693/0x11c0 ip6_output+0x24b/0x880 NF_HOOK.constprop.0+0xfd/0x530 ndisc_send_skb+0x9db/0x1400 ndisc_send_rs+0x12a/0x6c0 addrconf_dad_completed+0x3c9/0xea0 addrconf_dad_work+0x849/0x1420 process_one_work+0xa22/0x16e0 worker_thread+0x679/0x10c0 ret_from_fork+0x28/0x60 ret_from_fork_asm+0x11/0x20 When setup a vlan device on dev pim6reg, DAD ns packet may sent on reg_vif_xmit(). reg_vif_xmit() ip6mr_cache_report() skb_push(skb, -skb_network_offset(pkt));//skb_network_offset(pkt) is 4 And skb_push declared as: void *skb_push(struct sk_buff *skb, unsigned int len); skb->data -= len; //0xffff88805f86a84c - 0xfffffffc = 0xffff887f5f86a850 skb->data is set to 0xffff887f5f86a850, which is invalid mem addr, lead to skb_push() fails. Fixes: 14fb64e1f449 ("[IPV6] MROUTE: Support PIM-SM (SSM).") Signed-off-by: Yue Haibing <yuehaibing(a)huawei.com> Reviewed-by: Eric Dumazet <edumazet(a)google.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Yue Haibing <yuehaibing(a)huawei.com> --- net/ipv6/ip6mr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index d392a72cd784..c94a06b11141 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -1070,7 +1070,7 @@ static int ip6mr_cache_report(struct mr_table *mrt, struct sk_buff *pkt, And all this only to mangle msg->im6_msgtype and to set msg->im6_mbz to "mbz" :-) */ - skb_push(skb, -skb_network_offset(pkt)); + __skb_pull(skb, skb_network_offset(pkt)); skb_push(skb, sizeof(*msg)); skb_reset_transport_header(skb); -- 2.34.1

2 1

[PATCH openEuler-23.09 v2 0/8] enable files_cgroup
by Lu Jialin 26 Aug '23

26 Aug '23

v2->v1: enable CONFIG_CGROUP_FILES in openeuler_defconfig for x86 and arm64 Binder Makin (1): cgroups: Resource controller for open files Hou Tao (1): cgroup/files: use task_get_css() to get a valid css during dup_fd() Lu Jialin (3): fs: fix files.usage bug when move tasks fs/filescontrol.c: fix warning:large integer implicitly truncated to unsigned type openeuler_defconfig: enable CONFIG_CGROUP_FILES in openeuler_defconfig for x86 and arm64 Yu Kuai (1): fs/filescontrol: add a switch to enable / disable accounting of open fds Zhang Xiaoxu (1): files_cgroup: fix error pointer when kvm_vm_worker_thread yangerkun (1): fs/dcache.c: avoid panic while lockref of dentry overflow arch/arm64/configs/openeuler_defconfig | 1 + arch/x86/configs/openeuler_defconfig | 1 + fs/Makefile | 1 + fs/dcache.c | 32 ++- fs/file.c | 53 ++++- fs/filescontrol.c | 312 +++++++++++++++++++++++++ include/linux/cgroup-defs.h | 8 +- include/linux/cgroup_subsys.h | 6 + include/linux/fdtable.h | 1 + include/linux/filescontrol.h | 34 +++ include/linux/fs.h | 3 + init/Kconfig | 10 + 12 files changed, 453 insertions(+), 9 deletions(-) create mode 100644 fs/filescontrol.c create mode 100644 include/linux/filescontrol.h -- 2.34.1

2 9

[PATCH v2 OLK-5.10 0/4] Synchronizing mainline HiSilicon uncore PMU patches
by Junhao He 25 Aug '23

25 Aug '23

Synchronizing mainline HiSilicon uncore PMU patches The HISI uncore UC PMU and uncore H60PA/PAv3 PMU drivers have been integrated into the main line. Compared with OLK-5.10, some modifications are made. Add support for HiSilicon UC/H60PA/PAv3 PMU driver. PAv3 PMU: Compared with the PAv2 PMU, the PAv3 PMU has different event. The PAv3 PMU removed some events which are supported by PAv2 PMU. The older PA PMU driver will probe v3 as v2. Add the HISI0275 HID for PAv3 PMU to distinguish different. H60PA PMU: The H60PA PMU and PA are two different devices. The H60PA PMU supports higher bandwidth, and the PA PMU delay is relatively low. Different HIDs are used to distinguish the delay. UC PMU: Each cluster is integrated with a unified cache (UC) PMU, which provides consistency between NUMA and UMA domains. It sits between L2 and the memory system. Junhao He (4): Revert "drivers/perf: hisi: Add support for HiSilicon UC PMU driver" Revert "drivers/perf: hisi: Add support for HiSilicon H60PA and PAv3 PMU driver" drivers/perf: hisi: Add support for HiSilicon H60PA and PAv3 PMU driver drivers/perf: hisi: Add support for HiSilicon UC PMU driver drivers/perf/hisilicon/hisi_uncore_pa_pmu.c | 107 +++++++++----------- drivers/perf/hisilicon/hisi_uncore_pmu.c | 5 +- drivers/perf/hisilicon/hisi_uncore_pmu.h | 32 +----- drivers/perf/hisilicon/hisi_uncore_uc_pmu.c | 35 +++---- 4 files changed, 72 insertions(+), 107 deletions(-) -- 2.33.0

2 5

[PATCH v2 openEuler-22.03-LTS-SP2] mm: add preferred_swap ability
by liubo 25 Aug '23

25 Aug '23

From: Ni Cunshu <nicunshu(a)huawei.com> euleros inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I7U78A CVE: NA ------------------------------------------------ add a new attribute preferred_swap for each process, which indicates the swapfile or swap device that can be used by the process. Each process can only have one preferred_swap. Signed-off-by: Ni Cunshu <nicunshu(a)huawei.com> --- fs/proc/base.c | 113 +++++++++++++++++++++++++++++++++++++++ include/linux/mm_types.h | 4 ++ include/linux/swap.h | 13 +++++ kernel/fork.c | 4 ++ mm/Kconfig | 7 +++ mm/swap_slots.c | 68 ++++++++++++++++++++++- mm/swapfile.c | 86 ++++++++++++++++++++++++++++- 7 files changed, 292 insertions(+), 3 deletions(-) diff --git a/fs/proc/base.c b/fs/proc/base.c index 24c70ff923b8..34537572db57 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -3330,6 +3330,116 @@ static const struct file_operations proc_preferred_cpuset_operations = { }; #endif +#ifdef CONFIG_PREFERRED_SWAP +static ssize_t proc_preferred_swap_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct task_struct *task = get_proc_task(file_inode(file)); + struct mm_struct *mm; + int ret; + int err = 0; + size_t len; + char *buffer; + char *path; + + if (!task) + return -ESRCH; + + ret = 0; + mm = get_task_mm(task); + if (!mm) + goto out_no_mm; + + if (mm->preferred_swap) { + buffer = (char *)kmalloc(PATH_MAX + 1, GFP_KERNEL); + if (buffer == NULL) { + err = -ENOMEM; + goto out; + } + + path = d_path(&mm->preferred_swap->swap_file->f_path, + buffer, PATH_MAX); + + if (IS_ERR(path)) { + err = PTR_ERR(path); + kfree(buffer); + goto out; + } + + len = snprintf(buffer, strlen(path) + 2, "%s\n", path); + ret = simple_read_from_buffer(buf, count, ppos, buffer, len); + kfree(buffer); + } +out: + mmput(mm); +out_no_mm: + put_task_struct(task); + if (err < 0) + return err; + return ret; +} + +static ssize_t proc_preferred_swap_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos) +{ + struct task_struct *task; + struct mm_struct *mm; + char *buffer; + int err = 0; + + if (count >= PATH_MAX) + return -EINVAL; + + buffer = (char *)kmalloc(PATH_MAX, GFP_KERNEL); + if (buffer == NULL) { + err = -ENOMEM; + goto out_no_buffer; + } + if (copy_from_user(buffer, buf, count)) { + err = -EFAULT; + goto out_no_task; + } + buffer[count] = '\0'; + strreplace(buffer, '\n', '\0'); + task = get_proc_task(file_inode(file)); + if (!task) { + err = -ESRCH; + goto out_no_task; + } + mm = get_task_mm(task); + if (!mm) { + err = -ESRCH; + goto out_no_mm; + } + if (strlen(buffer) == 0) { + mm->preferred_swap = NULL; + goto out; + } + err = find_swap_info(buffer, mm); + if (err) { + pr_info("%s failed to be found as swap\n", buffer); + goto out; + } + + preferred_swap_used = 1; +out: + mmput(mm); +out_no_mm: + put_task_struct(task); +out_no_task: + kfree(buffer); +out_no_buffer: + if (err < 0) + return err; + return count; +} +static const struct file_operations proc_preferred_swap_operations = { + .write = proc_preferred_swap_write, + .read = proc_preferred_swap_read, + .llseek = generic_file_llseek, +}; +#endif /* CONFIG_PREFERRED_SWAP */ + static int proc_pid_personality(struct seq_file *m, struct pid_namespace *ns, struct pid *pid, struct task_struct *task) { @@ -3526,6 +3636,9 @@ static const struct pid_entry tgid_base_stuff[] = { ONE("ksm_merging_pages", S_IRUSR, proc_pid_ksm_merging_pages), ONE("ksm_stat", S_IRUSR, proc_pid_ksm_stat), #endif +#ifdef CONFIG_PREFERRED_SWAP + REG("preferred_swap", S_IRUGO|S_IWUSR, proc_preferred_swap_operations), +#endif }; static int proc_tgid_base_readdir(struct file *file, struct dir_context *ctx) diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index e3eaf458787a..2edd816fba7f 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -638,7 +638,11 @@ struct mm_struct { KABI_RESERVE(2) KABI_RESERVE(3) #endif +#ifdef CONFIG_PREFERRED_SWAP + KABI_USE(4, struct swap_info_struct *preferred_swap) +#else KABI_RESERVE(4) +#endif KABI_RESERVE(5) KABI_RESERVE(6) KABI_RESERVE(7) diff --git a/include/linux/swap.h b/include/linux/swap.h index 7f49964f27d2..998f10e97601 100644 --- a/include/linux/swap.h +++ b/include/linux/swap.h @@ -389,6 +389,10 @@ extern int vm_swappiness; extern int remove_mapping(struct address_space *mapping, struct page *page); extern unsigned long reclaim_pages(struct list_head *page_list); +#ifdef CONFIG_PREFERRED_SWAP +extern int find_swap_info(char *filename, struct mm_struct *mm); +extern void clear_tasks_mm_preferred_swap(struct swap_info_struct *p); +#endif #ifdef CONFIG_ETMEM enum etmem_swapcache_watermark_en { ETMEM_SWAPCACHE_WMARK_LOW, @@ -496,6 +500,10 @@ extern long total_swap_pages; extern atomic_t nr_rotate_swap; extern bool has_usable_swap(void); +#ifdef CONFIG_PREFERRED_SWAP +extern int preferred_swap_used; +#endif + /* Swap 50% full? Release swapcache more aggressively.. */ static inline bool vm_swap_full(void) { @@ -511,7 +519,12 @@ extern void si_swapinfo(struct sysinfo *); extern swp_entry_t get_swap_page(struct page *page); extern void put_swap_page(struct page *page, swp_entry_t entry); extern swp_entry_t get_swap_page_of_type(int); +#ifdef CONFIG_PREFERRED_SWAP +extern int get_swap_pages(int n, swp_entry_t swp_entries[], int entry_size, + struct swap_info_struct *preferred_swap); +#else extern int get_swap_pages(int n, swp_entry_t swp_entries[], int entry_size); +#endif extern int add_swap_count_continuation(swp_entry_t, gfp_t); extern void swap_shmem_alloc(swp_entry_t); extern int swap_duplicate(swp_entry_t); diff --git a/kernel/fork.c b/kernel/fork.c index 2547c6a6e5b1..734882ed98c6 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1072,6 +1072,10 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, mm_init_uprobes_state(mm); hugetlb_count_init(mm); +#ifdef CONFIG_PREFERRED_SWAP + mm->preferred_swap = NULL; +#endif + if (current->mm) { mm->flags = current->mm->flags & MMF_INIT_MASK; mm->def_flags = current->mm->def_flags & VM_INIT_DEF_MASK; diff --git a/mm/Kconfig b/mm/Kconfig index f66457168de9..48c37c3bda8c 100644 --- a/mm/Kconfig +++ b/mm/Kconfig @@ -999,6 +999,13 @@ config EXTEND_HUGEPAGE_MAPPING help Introduce vmalloc/vmap/remap interfaces that handle only hugepages. +config PREFERRED_SWAP + bool "make task only swapping memory to assigned swap" + depends on SYSFS + default y + help + Say Y here to enable kernel to limit tasks only swapping memory to assigned swap. + source "mm/damon/Kconfig" endmenu diff --git a/mm/swap_slots.c b/mm/swap_slots.c index 0357fbe70645..79a3164bf393 100644 --- a/mm/swap_slots.c +++ b/mm/swap_slots.c @@ -33,6 +33,9 @@ #include <linux/vmalloc.h> #include <linux/mutex.h> #include <linux/mm.h> +#ifdef CONFIG_PREFERRED_SWAP +#include <linux/rmap.h> +#endif static DEFINE_PER_CPU(struct swap_slots_cache, swp_slots); static bool swap_slot_cache_active; @@ -265,8 +268,13 @@ static int refill_swap_slots_cache(struct swap_slots_cache *cache) cache->cur = 0; if (swap_slot_cache_active) +#ifdef CONFIG_PREFERRED_SWAP + cache->nr = get_swap_pages(SWAP_SLOTS_CACHE_SIZE, + cache->slots, 1, NULL); +#else cache->nr = get_swap_pages(SWAP_SLOTS_CACHE_SIZE, cache->slots, 1); +#endif return cache->nr; } @@ -303,16 +311,66 @@ int free_swap_slot(swp_entry_t entry) return 0; } +#ifdef CONFIG_PREFERRED_SWAP +static bool has_preferred_swap(struct page *page, struct vm_area_struct *vma, + unsigned long addr, void *arg) +{ + struct swap_info_struct **si = (struct swap_info_struct **)arg; + + if (*si) { + *si = ERR_PTR(-EPERM); + return false; + } + if (vma->vm_file) { + *si = ERR_PTR(-EPERM); + return false; + } + if (vma->vm_mm->preferred_swap) { + *si = vma->vm_mm->preferred_swap; + } else { + *si = ERR_PTR(-EPERM); + return false; + } + return true; +} + +static struct swap_info_struct *page_preferred_swap(struct page *page) +{ + struct swap_info_struct *preferred_swap = NULL; + struct rmap_walk_control rwc = { + .rmap_one = has_preferred_swap, + .arg = (void *)&preferred_swap, + .anon_lock = page_lock_anon_vma_read, + }; + rmap_walk(page, &rwc); + if (IS_ERR_OR_NULL(preferred_swap)) + preferred_swap = NULL; + + return preferred_swap; +} +#endif /* CONFIG_PREFERRED_SWAP */ + swp_entry_t get_swap_page(struct page *page) { swp_entry_t entry; struct swap_slots_cache *cache; +#ifdef CONFIG_PREFERRED_SWAP + struct swap_info_struct *preferred_swap = NULL; + + if (!page_mapping(page)) + preferred_swap = page_preferred_swap(page); +#endif + entry.val = 0; if (PageTransHuge(page)) { if (IS_ENABLED(CONFIG_THP_SWAP)) +#ifdef CONFIG_PREFERRED_SWAP + get_swap_pages(1, &entry, HPAGE_PMD_NR, preferred_swap); +#else get_swap_pages(1, &entry, HPAGE_PMD_NR); +#endif goto out; } @@ -327,7 +385,11 @@ swp_entry_t get_swap_page(struct page *page) */ cache = raw_cpu_ptr(&swp_slots); - if (likely(check_cache_active() && cache->slots)) { +#ifdef CONFIG_PREFERRED_SWAP + if (likely(check_cache_active() && cache->slots && !preferred_swap)) { +#else + if (likely(check_cache_active() && cache->slots &&)) { +#endif mutex_lock(&cache->alloc_lock); if (cache->slots) { repeat: @@ -344,7 +406,11 @@ swp_entry_t get_swap_page(struct page *page) goto out; } +#ifdef CONFIG_PREFERRED_SWAP + get_swap_pages(1, &entry, 1, preferred_swap); +#else get_swap_pages(1, &entry, 1); +#endif out: if (mem_cgroup_try_charge_swap(page, entry)) { put_swap_page(page, entry); diff --git a/mm/swapfile.c b/mm/swapfile.c index 14e2396fa8a3..835f8a8913b6 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -62,7 +62,10 @@ EXPORT_SYMBOL_GPL(nr_swap_pages); /* protected with swap_lock. reading in vm_swap_full() doesn't need lock */ long total_swap_pages; static int least_priority = -1; - +#ifdef CONFIG_PREFERRED_SWAP +DEFINE_SPINLOCK(preferred_swap_lock); +int preferred_swap_used; +#endif static const char Bad_file[] = "Bad swap file entry "; static const char Unused_file[] = "Unused swap file entry "; static const char Bad_offset[] = "Bad swap offset entry "; @@ -1056,7 +1059,13 @@ static unsigned long scan_swap_map(struct swap_info_struct *si, } -int get_swap_pages(int n_goal, swp_entry_t swp_entries[], int entry_size) +#ifdef CONFIG_PREFERRED_SWAP +int get_swap_pages(int n_goal, swp_entry_t swp_entries[], + int entry_size, struct swap_info_struct *preferred_swap) +#else +int get_swap_pages(int n_goal, swp_entry_t swp_entries[], + int entry_size) +#endif { unsigned long size = swap_entry_size(entry_size); struct swap_info_struct *si, *next; @@ -1082,6 +1091,10 @@ int get_swap_pages(int n_goal, swp_entry_t swp_entries[], int entry_size) start_over: node = numa_node_id(); plist_for_each_entry_safe(si, next, &swap_avail_heads[node], avail_lists[node]) { +#ifdef CONFIG_PREFERRED_SWAP + if (preferred_swap && preferred_swap != si) + goto nextsi; +#endif /* requeue si to after same-priority siblings */ plist_requeue(&si->avail_lists[node], &swap_avail_heads[node]); spin_unlock(&swap_avail_lock); @@ -2643,6 +2656,10 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile) p->flags &= ~SWP_WRITEOK; spin_unlock(&p->lock); spin_unlock(&swap_lock); +#ifdef CONFIG_PREFERRED_SWAP + if (preferred_swap_used) + clear_tasks_mm_preferred_swap(p); +#endif disable_swap_slots_cache_lock(); @@ -3878,6 +3895,71 @@ void cgroup_throttle_swaprate(struct page *page, gfp_t gfp_mask) } #endif +#ifdef CONFIG_PREFERRED_SWAP +int find_swap_info(char *filename, struct mm_struct *mm) +{ + struct file *chosen_file; + unsigned int type; + struct path path; + int err = 0; + struct swap_info_struct *found_swap = NULL; + + err = kern_path(filename, LOOKUP_FOLLOW, &path); + if (err) + goto out_no_path; + chosen_file = dentry_open(&path, O_RDWR|O_LARGEFILE, current_cred()); + if (IS_ERR(chosen_file)) { + err = PTR_ERR(chosen_file); + goto out_no_file; + } + spin_lock(&preferred_swap_lock); + spin_lock(&swap_lock); + for (type = 0; type < nr_swapfiles; type++) { + if ((swap_info[type]->flags & SWP_USED) && + (swap_info[type]->flags & SWP_WRITEOK)) { + if (swap_info[type]->swap_file->f_mapping == + chosen_file->f_mapping) { + found_swap = swap_info[type]; + break; + } + } + } + + spin_unlock(&swap_lock); + if (type >= nr_swapfiles) { + err = -EINVAL; + goto out; + } + + mm->preferred_swap = found_swap; +out: + spin_unlock(&preferred_swap_lock); + filp_close(chosen_file, NULL); +out_no_file: + path_put(&path); +out_no_path: + return err; +} + +void clear_tasks_mm_preferred_swap(struct swap_info_struct *p) +{ + struct task_struct *task; + struct mm_struct *mm; + + spin_lock(&preferred_swap_lock); + read_lock(&tasklist_lock); + for_each_process(task) { + mm = get_task_mm(task); + if (mm) { + if (mm->preferred_swap == p) + mm->preferred_swap = NULL; + mmput(mm); + } + } + read_unlock(&tasklist_lock); + spin_unlock(&preferred_swap_lock); +} +#endif static int __init swapfile_init(void) { int nid; -- 2.27.0

2 1

[PATCH openEuler-22.03-LTS-SP2] mm: add preferred_swap ability
by liubo 24 Aug '23

24 Aug '23

From: Ni Cunshu <nicunshu(a)huawei.com> euleros inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I7U78A CVE: NA ------------------------------------------------ add a new attribute preferred_swap for each process, which indicates the swapfile or swap device that can be used by the process. Each process can only have one preferred_swap. Signed-off-by: Ni Cunshu <nicunshu(a)huawei.com> --- fs/proc/base.c | 113 +++++++++++++++++++++++++++++++++++++++ include/linux/mm_types.h | 4 ++ include/linux/swap.h | 13 +++++ kernel/fork.c | 4 ++ mm/Kconfig | 7 +++ mm/swap_slots.c | 68 ++++++++++++++++++++++- mm/swapfile.c | 86 ++++++++++++++++++++++++++++- 7 files changed, 292 insertions(+), 3 deletions(-) diff --git a/fs/proc/base.c b/fs/proc/base.c index 24c70ff923b8..34537572db57 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -3330,6 +3330,116 @@ static const struct file_operations proc_preferred_cpuset_operations = { }; #endif +#ifdef CONFIG_PREFERRED_SWAP +static ssize_t proc_preferred_swap_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct task_struct *task = get_proc_task(file_inode(file)); + struct mm_struct *mm; + int ret; + int err = 0; + size_t len; + char *buffer; + char *path; + + if (!task) + return -ESRCH; + + ret = 0; + mm = get_task_mm(task); + if (!mm) + goto out_no_mm; + + if (mm->preferred_swap) { + buffer = (char *)kmalloc(PATH_MAX + 1, GFP_KERNEL); + if (buffer == NULL) { + err = -ENOMEM; + goto out; + } + + path = d_path(&mm->preferred_swap->swap_file->f_path, + buffer, PATH_MAX); + + if (IS_ERR(path)) { + err = PTR_ERR(path); + kfree(buffer); + goto out; + } + + len = snprintf(buffer, strlen(path) + 2, "%s\n", path); + ret = simple_read_from_buffer(buf, count, ppos, buffer, len); + kfree(buffer); + } +out: + mmput(mm); +out_no_mm: + put_task_struct(task); + if (err < 0) + return err; + return ret; +} + +static ssize_t proc_preferred_swap_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos) +{ + struct task_struct *task; + struct mm_struct *mm; + char *buffer; + int err = 0; + + if (count >= PATH_MAX) + return -EINVAL; + + buffer = (char *)kmalloc(PATH_MAX, GFP_KERNEL); + if (buffer == NULL) { + err = -ENOMEM; + goto out_no_buffer; + } + if (copy_from_user(buffer, buf, count)) { + err = -EFAULT; + goto out_no_task; + } + buffer[count] = '\0'; + strreplace(buffer, '\n', '\0'); + task = get_proc_task(file_inode(file)); + if (!task) { + err = -ESRCH; + goto out_no_task; + } + mm = get_task_mm(task); + if (!mm) { + err = -ESRCH; + goto out_no_mm; + } + if (strlen(buffer) == 0) { + mm->preferred_swap = NULL; + goto out; + } + err = find_swap_info(buffer, mm); + if (err) { + pr_info("%s failed to be found as swap\n", buffer); + goto out; + } + + preferred_swap_used = 1; +out: + mmput(mm); +out_no_mm: + put_task_struct(task); +out_no_task: + kfree(buffer); +out_no_buffer: + if (err < 0) + return err; + return count; +} +static const struct file_operations proc_preferred_swap_operations = { + .write = proc_preferred_swap_write, + .read = proc_preferred_swap_read, + .llseek = generic_file_llseek, +}; +#endif /* CONFIG_PREFERRED_SWAP */ + static int proc_pid_personality(struct seq_file *m, struct pid_namespace *ns, struct pid *pid, struct task_struct *task) { @@ -3526,6 +3636,9 @@ static const struct pid_entry tgid_base_stuff[] = { ONE("ksm_merging_pages", S_IRUSR, proc_pid_ksm_merging_pages), ONE("ksm_stat", S_IRUSR, proc_pid_ksm_stat), #endif +#ifdef CONFIG_PREFERRED_SWAP + REG("preferred_swap", S_IRUGO|S_IWUSR, proc_preferred_swap_operations), +#endif }; static int proc_tgid_base_readdir(struct file *file, struct dir_context *ctx) diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index e3eaf458787a..2edd816fba7f 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -638,7 +638,11 @@ struct mm_struct { KABI_RESERVE(2) KABI_RESERVE(3) #endif +#ifdef CONFIG_PREFERRED_SWAP + KABI_USE(4, struct swap_info_struct *preferred_swap) +#else KABI_RESERVE(4) +#endif KABI_RESERVE(5) KABI_RESERVE(6) KABI_RESERVE(7) diff --git a/include/linux/swap.h b/include/linux/swap.h index 7f49964f27d2..998f10e97601 100644 --- a/include/linux/swap.h +++ b/include/linux/swap.h @@ -389,6 +389,10 @@ extern int vm_swappiness; extern int remove_mapping(struct address_space *mapping, struct page *page); extern unsigned long reclaim_pages(struct list_head *page_list); +#ifdef CONFIG_PREFERRED_SWAP +extern int find_swap_info(char *filename, struct mm_struct *mm); +extern void clear_tasks_mm_preferred_swap(struct swap_info_struct *p); +#endif #ifdef CONFIG_ETMEM enum etmem_swapcache_watermark_en { ETMEM_SWAPCACHE_WMARK_LOW, @@ -496,6 +500,10 @@ extern long total_swap_pages; extern atomic_t nr_rotate_swap; extern bool has_usable_swap(void); +#ifdef CONFIG_PREFERRED_SWAP +extern int preferred_swap_used; +#endif + /* Swap 50% full? Release swapcache more aggressively.. */ static inline bool vm_swap_full(void) { @@ -511,7 +519,12 @@ extern void si_swapinfo(struct sysinfo *); extern swp_entry_t get_swap_page(struct page *page); extern void put_swap_page(struct page *page, swp_entry_t entry); extern swp_entry_t get_swap_page_of_type(int); +#ifdef CONFIG_PREFERRED_SWAP +extern int get_swap_pages(int n, swp_entry_t swp_entries[], int entry_size, + struct swap_info_struct *preferred_swap); +#else extern int get_swap_pages(int n, swp_entry_t swp_entries[], int entry_size); +#endif extern int add_swap_count_continuation(swp_entry_t, gfp_t); extern void swap_shmem_alloc(swp_entry_t); extern int swap_duplicate(swp_entry_t); diff --git a/kernel/fork.c b/kernel/fork.c index 2547c6a6e5b1..734882ed98c6 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1072,6 +1072,10 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, mm_init_uprobes_state(mm); hugetlb_count_init(mm); +#ifdef CONFIG_PREFERRED_SWAP + mm->preferred_swap = NULL; +#endif + if (current->mm) { mm->flags = current->mm->flags & MMF_INIT_MASK; mm->def_flags = current->mm->def_flags & VM_INIT_DEF_MASK; diff --git a/mm/Kconfig b/mm/Kconfig index f66457168de9..48c37c3bda8c 100644 --- a/mm/Kconfig +++ b/mm/Kconfig @@ -999,6 +999,13 @@ config EXTEND_HUGEPAGE_MAPPING help Introduce vmalloc/vmap/remap interfaces that handle only hugepages. +config PREFERRED_SWAP + bool "make task only swapping memory to assigned swap" + depends on SYSFS + default y + help + Say Y here to enable kernel to limit tasks only swapping memory to assigned swap. + source "mm/damon/Kconfig" endmenu diff --git a/mm/swap_slots.c b/mm/swap_slots.c index 0357fbe70645..79a3164bf393 100644 --- a/mm/swap_slots.c +++ b/mm/swap_slots.c @@ -33,6 +33,9 @@ #include <linux/vmalloc.h> #include <linux/mutex.h> #include <linux/mm.h> +#ifdef CONFIG_PREFERRED_SWAP +#include <linux/rmap.h> +#endif static DEFINE_PER_CPU(struct swap_slots_cache, swp_slots); static bool swap_slot_cache_active; @@ -265,8 +268,13 @@ static int refill_swap_slots_cache(struct swap_slots_cache *cache) cache->cur = 0; if (swap_slot_cache_active) +#ifdef CONFIG_PREFERRED_SWAP + cache->nr = get_swap_pages(SWAP_SLOTS_CACHE_SIZE, + cache->slots, 1, NULL); +#else cache->nr = get_swap_pages(SWAP_SLOTS_CACHE_SIZE, cache->slots, 1); +#endif return cache->nr; } @@ -303,16 +311,66 @@ int free_swap_slot(swp_entry_t entry) return 0; } +#ifdef CONFIG_PREFERRED_SWAP +static bool has_preferred_swap(struct page *page, struct vm_area_struct *vma, + unsigned long addr, void *arg) +{ + struct swap_info_struct **si = (struct swap_info_struct **)arg; + + if (*si) { + *si = ERR_PTR(-EPERM); + return false; + } + if (vma->vm_file) { + *si = ERR_PTR(-EPERM); + return false; + } + if (vma->vm_mm->preferred_swap) { + *si = vma->vm_mm->preferred_swap; + } else { + *si = ERR_PTR(-EPERM); + return false; + } + return true; +} + +static struct swap_info_struct *page_preferred_swap(struct page *page) +{ + struct swap_info_struct *preferred_swap = NULL; + struct rmap_walk_control rwc = { + .rmap_one = has_preferred_swap, + .arg = (void *)&preferred_swap, + .anon_lock = page_lock_anon_vma_read, + }; + rmap_walk(page, &rwc); + if (IS_ERR_OR_NULL(preferred_swap)) + preferred_swap = NULL; + + return preferred_swap; +} +#endif /* CONFIG_PREFERRED_SWAP */ + swp_entry_t get_swap_page(struct page *page) { swp_entry_t entry; struct swap_slots_cache *cache; +#ifdef CONFIG_PREFERRED_SWAP + struct swap_info_struct *preferred_swap = NULL; + + if (!page_mapping(page)) + preferred_swap = page_preferred_swap(page); +#endif + entry.val = 0; if (PageTransHuge(page)) { if (IS_ENABLED(CONFIG_THP_SWAP)) +#ifdef CONFIG_PREFERRED_SWAP + get_swap_pages(1, &entry, HPAGE_PMD_NR, preferred_swap); +#else get_swap_pages(1, &entry, HPAGE_PMD_NR); +#endif goto out; } @@ -327,7 +385,11 @@ swp_entry_t get_swap_page(struct page *page) */ cache = raw_cpu_ptr(&swp_slots); - if (likely(check_cache_active() && cache->slots)) { +#ifdef CONFIG_PREFERRED_SWAP + if (likely(check_cache_active() && cache->slots && !preferred_swap)) { +#else + if (likely(check_cache_active() && cache->slots &&)) { +#endif mutex_lock(&cache->alloc_lock); if (cache->slots) { repeat: @@ -344,7 +406,11 @@ swp_entry_t get_swap_page(struct page *page) goto out; } +#ifdef CONFIG_PREFERRED_SWAP + get_swap_pages(1, &entry, 1, preferred_swap); +#else get_swap_pages(1, &entry, 1); +#endif out: if (mem_cgroup_try_charge_swap(page, entry)) { put_swap_page(page, entry); diff --git a/mm/swapfile.c b/mm/swapfile.c index 14e2396fa8a3..c40037f29647 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -62,7 +62,10 @@ EXPORT_SYMBOL_GPL(nr_swap_pages); /* protected with swap_lock. reading in vm_swap_full() doesn't need lock */ long total_swap_pages; static int least_priority = -1; - +#ifdef CONFIG_EULEROS_PREFERRED_SWAP +DEFINE_SPINLOCK(preferred_swap_lock); +int preferred_swap_used; +#endif static const char Bad_file[] = "Bad swap file entry "; static const char Unused_file[] = "Unused swap file entry "; static const char Bad_offset[] = "Bad swap offset entry "; @@ -1056,7 +1059,13 @@ static unsigned long scan_swap_map(struct swap_info_struct *si, } -int get_swap_pages(int n_goal, swp_entry_t swp_entries[], int entry_size) +#ifdef CONFIG_EULEROS_PREFERRED_SWAP +int get_swap_pages(int n_goal, swp_entry_t swp_entries[], + int entry_size, struct swap_info_struct *preferred_swap) +#else +int get_swap_pages(int n_goal, swp_entry_t swp_entries[], + int entry_size) +#endif { unsigned long size = swap_entry_size(entry_size); struct swap_info_struct *si, *next; @@ -1082,6 +1091,10 @@ int get_swap_pages(int n_goal, swp_entry_t swp_entries[], int entry_size) start_over: node = numa_node_id(); plist_for_each_entry_safe(si, next, &swap_avail_heads[node], avail_lists[node]) { +#ifdef CONFIG_EULEROS_PREFERRED_SWAP + if (preferred_swap && preferred_swap != si) + goto nextsi; +#endif /* requeue si to after same-priority siblings */ plist_requeue(&si->avail_lists[node], &swap_avail_heads[node]); spin_unlock(&swap_avail_lock); @@ -2643,6 +2656,10 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile) p->flags &= ~SWP_WRITEOK; spin_unlock(&p->lock); spin_unlock(&swap_lock); +#ifdef CONFIG_EULEROS_PREFERRED_SWAP + if (preferred_swap_used) + clear_tasks_mm_preferred_swap(p); +#endif disable_swap_slots_cache_lock(); @@ -3878,6 +3895,71 @@ void cgroup_throttle_swaprate(struct page *page, gfp_t gfp_mask) } #endif +#ifdef CONFIG_EULEROS_PREFERRED_SWAP +int find_swap_info(char *filename, struct mm_struct *mm) +{ + struct file *chosen_file; + unsigned int type; + struct path path; + int err = 0; + struct swap_info_struct *found_swap = NULL; + + err = kern_path(filename, LOOKUP_FOLLOW, &path); + if (err) + goto out_no_path; + chosen_file = dentry_open(&path, O_RDWR|O_LARGEFILE, current_cred()); + if (IS_ERR(chosen_file)) { + err = PTR_ERR(chosen_file); + goto out_no_file; + } + spin_lock(&preferred_swap_lock); + spin_lock(&swap_lock); + for (type = 0; type < nr_swapfiles; type++) { + if ((swap_info[type]->flags & SWP_USED) && + (swap_info[type]->flags & SWP_WRITEOK)) { + if (swap_info[type]->swap_file->f_mapping == + chosen_file->f_mapping) { + found_swap = swap_info[type]; + break; + } + } + } + + spin_unlock(&swap_lock); + if (type >= nr_swapfiles) { + err = -EINVAL; + goto out; + } + + mm->preferred_swap = found_swap; +out: + spin_unlock(&preferred_swap_lock); + filp_close(chosen_file, NULL); +out_no_file: + path_put(&path); +out_no_path: + return err; +} + +void clear_tasks_mm_preferred_swap(struct swap_info_struct *p) +{ + struct task_struct *task; + struct mm_struct *mm; + + spin_lock(&preferred_swap_lock); + read_lock(&tasklist_lock); + for_each_process(task) { + mm = get_task_mm(task); + if (mm) { + if (mm->preferred_swap == p) + mm->preferred_swap = NULL; + mmput(mm); + } + } + read_unlock(&tasklist_lock); + spin_unlock(&preferred_swap_lock); +} +#endif static int __init swapfile_init(void) { int nid; -- 2.27.0

2 1

[PATCH OLK-5.10] mm: add preferred_swap ability
by liubo 24 Aug '23

24 Aug '23

From: Ni Cunshu <nicunshu(a)huawei.com> euleros inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I7VR4N CVE: NA ------------------------------------------------ add a new attribute preferred_swap for each process, which indicates the swapfile or swap device that can be used by the process. Each process can only have one preferred_swap. Signed-off-by: Ni Cunshu <nicunshu(a)huawei.com> --- fs/proc/base.c | 113 +++++++++++++++++++++++++++++++++++++++ include/linux/mm_types.h | 4 ++ include/linux/swap.h | 13 +++++ kernel/fork.c | 4 ++ mm/Kconfig | 7 +++ mm/swap_slots.c | 68 ++++++++++++++++++++++- mm/swapfile.c | 85 ++++++++++++++++++++++++++++- 7 files changed, 291 insertions(+), 3 deletions(-) diff --git a/fs/proc/base.c b/fs/proc/base.c index 24c70ff923b8..34537572db57 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -3330,6 +3330,116 @@ static const struct file_operations proc_preferred_cpuset_operations = { }; #endif +#ifdef CONFIG_PREFERRED_SWAP +static ssize_t proc_preferred_swap_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct task_struct *task = get_proc_task(file_inode(file)); + struct mm_struct *mm; + int ret; + int err = 0; + size_t len; + char *buffer; + char *path; + + if (!task) + return -ESRCH; + + ret = 0; + mm = get_task_mm(task); + if (!mm) + goto out_no_mm; + + if (mm->preferred_swap) { + buffer = (char *)kmalloc(PATH_MAX + 1, GFP_KERNEL); + if (buffer == NULL) { + err = -ENOMEM; + goto out; + } + + path = d_path(&mm->preferred_swap->swap_file->f_path, + buffer, PATH_MAX); + + if (IS_ERR(path)) { + err = PTR_ERR(path); + kfree(buffer); + goto out; + } + + len = snprintf(buffer, strlen(path) + 2, "%s\n", path); + ret = simple_read_from_buffer(buf, count, ppos, buffer, len); + kfree(buffer); + } +out: + mmput(mm); +out_no_mm: + put_task_struct(task); + if (err < 0) + return err; + return ret; +} + +static ssize_t proc_preferred_swap_write(struct file *file, + const char __user *buf, size_t count, loff_t *ppos) +{ + struct task_struct *task; + struct mm_struct *mm; + char *buffer; + int err = 0; + + if (count >= PATH_MAX) + return -EINVAL; + + buffer = (char *)kmalloc(PATH_MAX, GFP_KERNEL); + if (buffer == NULL) { + err = -ENOMEM; + goto out_no_buffer; + } + if (copy_from_user(buffer, buf, count)) { + err = -EFAULT; + goto out_no_task; + } + buffer[count] = '\0'; + strreplace(buffer, '\n', '\0'); + task = get_proc_task(file_inode(file)); + if (!task) { + err = -ESRCH; + goto out_no_task; + } + mm = get_task_mm(task); + if (!mm) { + err = -ESRCH; + goto out_no_mm; + } + if (strlen(buffer) == 0) { + mm->preferred_swap = NULL; + goto out; + } + err = find_swap_info(buffer, mm); + if (err) { + pr_info("%s failed to be found as swap\n", buffer); + goto out; + } + + preferred_swap_used = 1; +out: + mmput(mm); +out_no_mm: + put_task_struct(task); +out_no_task: + kfree(buffer); +out_no_buffer: + if (err < 0) + return err; + return count; +} +static const struct file_operations proc_preferred_swap_operations = { + .write = proc_preferred_swap_write, + .read = proc_preferred_swap_read, + .llseek = generic_file_llseek, +}; +#endif /* CONFIG_PREFERRED_SWAP */ + static int proc_pid_personality(struct seq_file *m, struct pid_namespace *ns, struct pid *pid, struct task_struct *task) { @@ -3526,6 +3636,9 @@ static const struct pid_entry tgid_base_stuff[] = { ONE("ksm_merging_pages", S_IRUSR, proc_pid_ksm_merging_pages), ONE("ksm_stat", S_IRUSR, proc_pid_ksm_stat), #endif +#ifdef CONFIG_PREFERRED_SWAP + REG("preferred_swap", S_IRUGO|S_IWUSR, proc_preferred_swap_operations), +#endif }; static int proc_tgid_base_readdir(struct file *file, struct dir_context *ctx) diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index d1be389c0468..07ccbcd4e65a 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -640,7 +640,11 @@ struct mm_struct { KABI_RESERVE(2) KABI_RESERVE(3) #endif +#ifdef CONFIG_PREFERRED_SWAP + KABI_USE(4, struct swap_info_struct *preferred_swap) +#else KABI_RESERVE(4) +#endif KABI_RESERVE(5) KABI_RESERVE(6) KABI_RESERVE(7) diff --git a/include/linux/swap.h b/include/linux/swap.h index 524e3ca9d0f5..403e37c5641c 100644 --- a/include/linux/swap.h +++ b/include/linux/swap.h @@ -393,6 +393,10 @@ extern int vm_swappiness; extern int remove_mapping(struct address_space *mapping, struct page *page); extern unsigned long reclaim_pages(struct list_head *page_list); +#ifdef CONFIG_PREFERRED_SWAP +extern int find_swap_info(char *filename, struct mm_struct *mm); +extern void clear_tasks_mm_preferred_swap(struct swap_info_struct *p); +#endif #ifdef CONFIG_ETMEM enum etmem_swapcache_watermark_en { ETMEM_SWAPCACHE_WMARK_LOW, @@ -500,6 +504,10 @@ extern long total_swap_pages; extern atomic_t nr_rotate_swap; extern bool has_usable_swap(void); +#ifdef CONFIG_PREFERRED_SWAP +extern int preferred_swap_used; +#endif + /* Swap 50% full? Release swapcache more aggressively.. */ static inline bool vm_swap_full(void) { @@ -517,8 +525,13 @@ extern void si_swapinfo(struct sysinfo *); extern swp_entry_t get_swap_page(struct page *page); extern void put_swap_page(struct page *page, swp_entry_t entry); extern swp_entry_t get_swap_page_of_type(int); +#ifdef CONFIG_PREFERRED_SWAP +extern int get_swap_pages(int n, swp_entry_t swp_entries[], int entry_size, + struct swap_info_struct *preferred_swap, int type); +#else extern int get_swap_pages(int n, swp_entry_t swp_entries[], int entry_size, int type); +#endif extern int add_swap_count_continuation(swp_entry_t, gfp_t); extern void swap_shmem_alloc(swp_entry_t); extern int swap_duplicate(swp_entry_t); diff --git a/kernel/fork.c b/kernel/fork.c index 2547c6a6e5b1..734882ed98c6 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1072,6 +1072,10 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, mm_init_uprobes_state(mm); hugetlb_count_init(mm); +#ifdef CONFIG_PREFERRED_SWAP + mm->preferred_swap = NULL; +#endif + if (current->mm) { mm->flags = current->mm->flags & MMF_INIT_MASK; mm->def_flags = current->mm->def_flags & VM_INIT_DEF_MASK; diff --git a/mm/Kconfig b/mm/Kconfig index 0fe459e79336..61fb6cbaab6b 100644 --- a/mm/Kconfig +++ b/mm/Kconfig @@ -1008,6 +1008,13 @@ config EXTEND_HUGEPAGE_MAPPING help Introduce vmalloc/vmap/remap interfaces that handle only hugepages. +config PREFERRED_SWAP + bool "make task only swapping memory to assigned swap" + depends on SYSFS + default y + help + Say Y here to enable kernel to limit tasks only swapping memory to assigned swap. + source "mm/damon/Kconfig" endmenu diff --git a/mm/swap_slots.c b/mm/swap_slots.c index b7958ca276c2..b77af28cdceb 100644 --- a/mm/swap_slots.c +++ b/mm/swap_slots.c @@ -33,6 +33,9 @@ #include <linux/vmalloc.h> #include <linux/mutex.h> #include <linux/mm.h> +#ifdef CONFIG_PREFERRED_SWAP +#include <linux/rmap.h> +#endif static DEFINE_PER_CPU(struct swap_slots_cache, swp_slots); #ifdef CONFIG_MEMCG_SWAP_QOS @@ -385,8 +388,13 @@ static int refill_swap_slots_cache(struct swap_slots_cache *cache, int type) cache->cur = 0; if (swap_slot_cache_active) +#ifdef CONFIG_PREFERRED_SWAP + cache->nr = get_swap_pages(SWAP_SLOTS_CACHE_SIZE, + cache->slots, 1, NULL, type); +#else cache->nr = get_swap_pages(SWAP_SLOTS_CACHE_SIZE, cache->slots, 1, type); +#endif return cache->nr; } @@ -423,12 +431,58 @@ int free_swap_slot(swp_entry_t entry) return 0; } +#ifdef CONFIG_PREFERRED_SWAP +static bool has_preferred_swap(struct page *page, struct vm_area_struct *vma, + unsigned long addr, void *arg) +{ + struct swap_info_struct **si = (struct swap_info_struct **)arg; + + if (*si) { + *si = ERR_PTR(-EPERM); + return false; + } + if (vma->vm_file) { + *si = ERR_PTR(-EPERM); + return false; + } + if (vma->vm_mm->preferred_swap) { + *si = vma->vm_mm->preferred_swap; + } else { + *si = ERR_PTR(-EPERM); + return false; + } + return true; +} + +static struct swap_info_struct *page_preferred_swap(struct page *page) +{ + struct swap_info_struct *preferred_swap = NULL; + struct rmap_walk_control rwc = { + .rmap_one = has_preferred_swap, + .arg = (void *)&preferred_swap, + .anon_lock = page_lock_anon_vma_read, + }; + rmap_walk(page, &rwc); + if (IS_ERR_OR_NULL(preferred_swap)) + preferred_swap = NULL; + + return preferred_swap; +} +#endif /* CONFIG_PREFERRED_SWAP */ + swp_entry_t get_swap_page(struct page *page) { swp_entry_t entry; struct swap_slots_cache *cache; int type; +#ifdef CONFIG_PREFERRED_SWAP + struct swap_info_struct *preferred_swap = NULL; + + if (!page_mapping(page)) + preferred_swap = page_preferred_swap(page); +#endif + entry.val = 0; type = memcg_get_swap_type(page); @@ -437,7 +491,11 @@ swp_entry_t get_swap_page(struct page *page) if (PageTransHuge(page)) { if (IS_ENABLED(CONFIG_THP_SWAP) && arch_thp_swp_supported()) +#ifdef CONFIG_PREFERRED_SWAP + get_swap_pages(1, &entry, HPAGE_PMD_NR, preferred_swap, type); +#else get_swap_pages(1, &entry, HPAGE_PMD_NR, type); +#endif goto out; } @@ -452,7 +510,11 @@ swp_entry_t get_swap_page(struct page *page) */ cache = get_slots_cache(type); - if (likely(check_cache_active() && cache->slots)) { +#ifdef CONFIG_PREFERRED_SWAP + if (likely(check_cache_active() && cache->slots && !preferred_swap)) { +#else + if (likely(check_cache_active() && cache->slots &&)) { +#endif mutex_lock(&cache->alloc_lock); if (cache->slots) { repeat: @@ -469,7 +531,11 @@ swp_entry_t get_swap_page(struct page *page) goto out; } +#ifdef CONFIG_PREFERRED_SWAP + get_swap_pages(1, &entry, 1, preferred_swap, type); +#else get_swap_pages(1, &entry, 1, type); +#endif out: if (mem_cgroup_try_charge_swap(page, entry)) { put_swap_page(page, entry); diff --git a/mm/swapfile.c b/mm/swapfile.c index f5a15511f45a..223af9c9fadb 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -62,7 +62,10 @@ EXPORT_SYMBOL_GPL(nr_swap_pages); /* protected with swap_lock. reading in vm_swap_full() doesn't need lock */ long total_swap_pages; static int least_priority = -1; - +#ifdef CONFIG_PREFERRED_SWAP +DEFINE_SPINLOCK(preferred_swap_lock); +int preferred_swap_used; +#endif static const char Bad_file[] = "Bad swap file entry "; static const char Unused_file[] = "Unused swap file entry "; static const char Bad_offset[] = "Bad swap offset entry "; @@ -1145,8 +1148,13 @@ static inline bool should_skip_swap_type(int swap_type, int type) } #endif +#ifdef CONFIG_PREFERRED_SWAP +int get_swap_pages(int n_goal, swp_entry_t swp_entries[], int entry_size, + struct swap_info_struct *preferred_swap, int type) +#else int get_swap_pages(int n_goal, swp_entry_t swp_entries[], int entry_size, int type) +#endif { unsigned long size = swap_entry_size(entry_size); struct swap_info_struct *si, *next; @@ -1172,11 +1180,15 @@ int get_swap_pages(int n_goal, swp_entry_t swp_entries[], int entry_size, start_over: node = numa_node_id(); plist_for_each_entry_safe(si, next, &swap_avail_heads[node], avail_lists[node]) { +#ifdef CONFIG_PREFERRED_SWAP + if (preferred_swap && preferred_swap != si) + goto nextsi; +#endif /* requeue si to after same-priority siblings */ plist_requeue(&si->avail_lists[node], &swap_avail_heads[node]); spin_unlock(&swap_avail_lock); spin_lock(&si->lock); - if (should_skip_swap_type(si->type, type)) { + if (!preferred_swap && should_skip_swap_type(si->type, type)) { spin_unlock(&si->lock); spin_lock(&swap_avail_lock); goto nextsi; @@ -2738,6 +2750,10 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile) p->flags &= ~SWP_WRITEOK; spin_unlock(&p->lock); spin_unlock(&swap_lock); +#ifdef CONFIG_PREFERRED_SWAP + if (preferred_swap_used) + clear_tasks_mm_preferred_swap(p); +#endif disable_swap_slots_cache_lock(); @@ -3974,6 +3990,71 @@ void cgroup_throttle_swaprate(struct page *page, gfp_t gfp_mask) } #endif +#ifdef CONFIG_PREFERRED_SWAP +int find_swap_info(char *filename, struct mm_struct *mm) +{ + struct file *chosen_file; + unsigned int type; + struct path path; + int err = 0; + struct swap_info_struct *found_swap = NULL; + + err = kern_path(filename, LOOKUP_FOLLOW, &path); + if (err) + goto out_no_path; + chosen_file = dentry_open(&path, O_RDWR|O_LARGEFILE, current_cred()); + if (IS_ERR(chosen_file)) { + err = PTR_ERR(chosen_file); + goto out_no_file; + } + spin_lock(&preferred_swap_lock); + spin_lock(&swap_lock); + for (type = 0; type < nr_swapfiles; type++) { + if ((swap_info[type]->flags & SWP_USED) && + (swap_info[type]->flags & SWP_WRITEOK)) { + if (swap_info[type]->swap_file->f_mapping == + chosen_file->f_mapping) { + found_swap = swap_info[type]; + break; + } + } + } + + spin_unlock(&swap_lock); + if (type >= nr_swapfiles) { + err = -EINVAL; + goto out; + } + + mm->preferred_swap = found_swap; +out: + spin_unlock(&preferred_swap_lock); + filp_close(chosen_file, NULL); +out_no_file: + path_put(&path); +out_no_path: + return err; +} + +void clear_tasks_mm_preferred_swap(struct swap_info_struct *p) +{ + struct task_struct *task; + struct mm_struct *mm; + + spin_lock(&preferred_swap_lock); + read_lock(&tasklist_lock); + for_each_process(task) { + mm = get_task_mm(task); + if (mm) { + if (mm->preferred_swap == p) + mm->preferred_swap = NULL; + mmput(mm); + } + } + read_unlock(&tasklist_lock); + spin_unlock(&preferred_swap_lock); +} +#endif static int __init swapfile_init(void) { int nid; -- 2.27.0

2 1

[PATCH openEuler-23.09 0/7] enable filescgroup and bugfix
by Lu Jialin 24 Aug '23

24 Aug '23

Binder Makin (1): cgroups: Resource controller for open files Hou Tao (1): cgroup/files: use task_get_css() to get a valid css during dup_fd() Lu Jialin (2): fs: fix files.usage bug when move tasks fs/filescontrol.c: fix warning:large integer implicitly truncated to unsigned type Yu Kuai (1): fs/filescontrol: add a switch to enable / disable accounting of open fds Zhang Xiaoxu (1): files_cgroup: fix error pointer when kvm_vm_worker_thread yangerkun (1): fs/dcache.c: avoid panic while lockref of dentry overflow fs/Makefile | 1 + fs/dcache.c | 32 +++- fs/file.c | 53 +++++- fs/filescontrol.c | 312 ++++++++++++++++++++++++++++++++++ include/linux/cgroup-defs.h | 8 +- include/linux/cgroup_subsys.h | 6 + include/linux/fdtable.h | 1 + include/linux/filescontrol.h | 34 ++++ include/linux/fs.h | 3 + init/Kconfig | 10 ++ 10 files changed, 451 insertions(+), 9 deletions(-) create mode 100644 fs/filescontrol.c create mode 100644 include/linux/filescontrol.h -- 2.34.1

2 8

[PATCH openEuler-1.0-LTS] nbd: pass nbd_sock to nbd_read_reply() instead of index
by Li Nan 24 Aug '23

24 Aug '23

hulk inclusion category: bugfix bugzilla: 189113, https://gitee.com/openeuler/kernel/issues/I7T755 CVE: NA -------------------------------- If a socket is processing ioctl 'NBD_SET_SOCK', config->socks might be krealloc in nbd_add_socket(), and a garbage request is received now, a UAF may occurs. T1 nbd_ioctl __nbd_ioctl nbd_add_socket blk_mq_freeze_queue T2 recv_work nbd_read_reply sock_xmit krealloc config->socks def config->socks Pass nbd_sock to nbd_read_reply(). And introduce a new function sock_xmit_recv(), which differs from sock_xmit only in the way it get socket. ================================================================== BUG: KASAN: use-after-free in sock_xmit+0x525/0x550 Read of size 8 at addr ffff8880188ec428 by task kworker/u12:1/18779 Workqueue: knbd4-recv recv_work Call Trace: __dump_stack dump_stack+0xbe/0xfd print_address_description.constprop.0+0x19/0x170 __kasan_report.cold+0x6c/0x84 kasan_report+0x3a/0x50 sock_xmit+0x525/0x550 nbd_read_reply+0xfe/0x2c0 recv_work+0x1c2/0x750 process_one_work+0x6b6/0xf10 worker_thread+0xdd/0xd80 kthread+0x30a/0x410 ret_from_fork+0x22/0x30 Allocated by task 18784: kasan_save_stack+0x1b/0x40 kasan_set_track set_alloc_info __kasan_kmalloc __kasan_kmalloc.constprop.0+0xf0/0x130 slab_post_alloc_hook slab_alloc_node slab_alloc __kmalloc_track_caller+0x157/0x550 __do_krealloc krealloc+0x37/0xb0 nbd_add_socket +0x2d3/0x880 __nbd_ioctl nbd_ioctl+0x584/0x8e0 __blkdev_driver_ioctl blkdev_ioctl+0x2a0/0x6e0 block_ioctl+0xee/0x130 vfs_ioctl __do_sys_ioctl __se_sys_ioctl+0x138/0x190 do_syscall_64+0x33/0x40 entry_SYSCALL_64_after_hwframe+0x61/0xc6 Freed by task 18784: kasan_save_stack+0x1b/0x40 kasan_set_track+0x1c/0x30 kasan_set_free_info+0x20/0x40 __kasan_slab_free.part.0+0x13f/0x1b0 slab_free_hook slab_free_freelist_hook slab_free kfree+0xcb/0x6c0 krealloc+0x56/0xb0 nbd_add_socket+0x2d3/0x880 __nbd_ioctl nbd_ioctl+0x584/0x8e0 __blkdev_driver_ioctl blkdev_ioctl+0x2a0/0x6e0 block_ioctl+0xee/0x130 vfs_ioctl __do_sys_ioctl __se_sys_ioctl+0x138/0x190 do_syscall_64+0x33/0x40 entry_SYSCALL_64_after_hwframe+0x61/0xc6 Signed-off-by: Li Nan <linan122(a)huawei.com> --- drivers/block/nbd.c | 35 ++++++++++++++++++++++------------- 1 file changed, 22 insertions(+), 13 deletions(-) diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c index fb4e9cb3d5d6..3b974988b0a9 100644 --- a/drivers/block/nbd.c +++ b/drivers/block/nbd.c @@ -62,6 +62,7 @@ struct nbd_sock { struct recv_thread_args { struct work_struct work; struct nbd_device *nbd; + struct nbd_sock *nsock; int index; }; @@ -453,15 +454,9 @@ static enum blk_eh_timer_return nbd_xmit_timeout(struct request *req, return BLK_EH_DONE; } -/* - * Send or receive packet. Return a positive value on success and - * negtive value on failue, and never return 0. - */ -static int sock_xmit(struct nbd_device *nbd, int index, int send, - struct iov_iter *iter, int msg_flags, int *sent) +static int __sock_xmit(struct nbd_device *nbd, struct socket *sock, int send, + struct iov_iter *iter, int msg_flags, int *sent) { - struct nbd_config *config = nbd->config; - struct socket *sock = config->socks[index]->sock; int result; struct msghdr msg; unsigned int noreclaim_flag; @@ -503,6 +498,19 @@ static int sock_xmit(struct nbd_device *nbd, int index, int send, return result; } +/* + * Send or receive packet. Return a positive value on success and + * negtive value on failure, and never return 0. + */ +static int sock_xmit(struct nbd_device *nbd, int index, int send, + struct iov_iter *iter, int msg_flags, int *sent) +{ + struct nbd_config *config = nbd->config; + struct socket *sock = config->socks[index]->sock; + + return __sock_xmit(nbd, sock, send, iter, msg_flags, sent); +} + /* * Different settings for sk->sk_sndtimeo can result in different return values * if there is a signal pending when we enter sendmsg, because reasons? @@ -664,7 +672,7 @@ static int nbd_send_cmd(struct nbd_device *nbd, struct nbd_cmd *cmd, int index) return 0; } -static int nbd_read_reply(struct nbd_device *nbd, int index, +static int nbd_read_reply(struct nbd_device *nbd, struct socket *sock, struct nbd_reply *reply) { struct kvec iov = {.iov_base = reply, .iov_len = sizeof(*reply)}; @@ -673,7 +681,7 @@ static int nbd_read_reply(struct nbd_device *nbd, int index, reply->magic = 0; iov_iter_kvec(&to, READ | ITER_KVEC, &iov, 1, sizeof(*reply)); - result = sock_xmit(nbd, index, 0, &to, MSG_WAITALL, NULL); + result = __sock_xmit(nbd, sock, 0, &to, MSG_WAITALL, NULL); if (result < 0) { if (!nbd_disconnected(nbd->config)) dev_err(disk_to_dev(nbd->disk), @@ -798,13 +806,13 @@ static void recv_work(struct work_struct *work) struct nbd_device *nbd = args->nbd; struct nbd_config *config = nbd->config; struct request_queue *q = nbd->disk->queue; - struct nbd_sock *nsock; + struct nbd_sock *nsock = args->nsock; struct nbd_cmd *cmd; while (1) { struct nbd_reply reply; - if (nbd_read_reply(nbd, args->index, &reply)) + if (nbd_read_reply(nbd, nsock->sock, &reply)) break; /* @@ -829,7 +837,6 @@ static void recv_work(struct work_struct *work) percpu_ref_put(&q->q_usage_counter); } - nsock = config->socks[args->index]; mutex_lock(&nsock->tx_lock); nbd_mark_nsock_dead(nbd, nsock, 1); mutex_unlock(&nsock->tx_lock); @@ -1164,6 +1171,7 @@ static int nbd_reconnect_socket(struct nbd_device *nbd, unsigned long arg) INIT_WORK(&args->work, recv_work); args->index = i; args->nbd = nbd; + args->nsock = nsock; nsock->cookie++; mutex_unlock(&nsock->tx_lock); sockfd_put(old); @@ -1354,6 +1362,7 @@ static int nbd_start_device(struct nbd_device *nbd) refcount_inc(&nbd->config_refs); INIT_WORK(&args->work, recv_work); args->nbd = nbd; + args->nsock = config->socks[i]; args->index = i; queue_work(nbd->recv_workq, &args->work); } -- 2.39.2

2 1