From e7fd37ba12170cc414be8b639dfc2c5f7172fac2 Mon Sep 17 00:00:00 2001 From: Ma Shimiao Date: Tue, 12 Dec 2017 09:43:49 +0800 Subject: cgroup: avoid copying strings longer than the buffers cgroup root name and file name have max length limit, we should avoid copying longer name than that to the name. tj: minor update to $SUBJ. Signed-off-by: Ma Shimiao Signed-off-by: Tejun Heo --- kernel/cgroup/cgroup.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c index 0b1ffe147f24..18d71fbd3923 100644 --- a/kernel/cgroup/cgroup.c +++ b/kernel/cgroup/cgroup.c @@ -1397,7 +1397,7 @@ static char *cgroup_file_name(struct cgroup *cgrp, const struct cftype *cft, cgroup_on_dfl(cgrp) ? ss->name : ss->legacy_name, cft->name); else - strncpy(buf, cft->name, CGROUP_FILE_NAME_MAX); + strscpy(buf, cft->name, CGROUP_FILE_NAME_MAX); return buf; } @@ -1864,9 +1864,9 @@ void init_cgroup_root(struct cgroup_root *root, struct cgroup_sb_opts *opts) root->flags = opts->flags; if (opts->release_agent) - strcpy(root->release_agent_path, opts->release_agent); + strscpy(root->release_agent_path, opts->release_agent, PATH_MAX); if (opts->name) - strcpy(root->name, opts->name); + strscpy(root->name, opts->name, MAX_CGROUP_ROOT_NAMELEN); if (opts->cpuset_clone_children) set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->cgrp.flags); } -- cgit v1.2.3 From 50034ed49645463a16327cad05694e201e6b4126 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Fri, 15 Dec 2017 05:09:47 -0800 Subject: cgroup: use strlcpy() instead of strscpy() to avoid spurious warning As long as cft->name is guaranteed to be NUL-terminated, using strlcpy() would work just as well and avoid that warning, so the change below could be folded into that commit. Signed-off-by: Arnd Bergmann Signed-off-by: Tejun Heo --- kernel/cgroup/cgroup.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c index 18d71fbd3923..f4c2f8cb5748 100644 --- a/kernel/cgroup/cgroup.c +++ b/kernel/cgroup/cgroup.c @@ -1397,7 +1397,7 @@ static char *cgroup_file_name(struct cgroup *cgrp, const struct cftype *cft, cgroup_on_dfl(cgrp) ? ss->name : ss->legacy_name, cft->name); else - strscpy(buf, cft->name, CGROUP_FILE_NAME_MAX); + strlcpy(buf, cft->name, CGROUP_FILE_NAME_MAX); return buf; } @@ -1864,9 +1864,9 @@ void init_cgroup_root(struct cgroup_root *root, struct cgroup_sb_opts *opts) root->flags = opts->flags; if (opts->release_agent) - strscpy(root->release_agent_path, opts->release_agent, PATH_MAX); + strlcpy(root->release_agent_path, opts->release_agent, PATH_MAX); if (opts->name) - strscpy(root->name, opts->name, MAX_CGROUP_ROOT_NAMELEN); + strlcpy(root->name, opts->name, MAX_CGROUP_ROOT_NAMELEN); if (opts->cpuset_clone_children) set_bit(CGRP_CPUSET_CLONE_CHILDREN, &root->cgrp.flags); } -- cgit v1.2.3 From 116d2f7496c51b2e02e8e4ecdd2bdf5fb9d5a641 Mon Sep 17 00:00:00 2001 From: Prateek Sood Date: Tue, 19 Dec 2017 12:56:57 +0530 Subject: cgroup: Fix deadlock in cpu hotplug path Deadlock during cgroup migration from cpu hotplug path when a task T is being moved from source to destination cgroup. kworker/0:0 cpuset_hotplug_workfn() cpuset_hotplug_update_tasks() hotplug_update_tasks_legacy() remove_tasks_in_empty_cpuset() cgroup_transfer_tasks() // stuck in iterator loop cgroup_migrate() cgroup_migrate_add_task() In cgroup_migrate_add_task() it checks for PF_EXITING flag of task T. Task T will not migrate to destination cgroup. css_task_iter_start() will keep pointing to task T in loop waiting for task T cg_list node to be removed. Task T do_exit() exit_signals() // sets PF_EXITING exit_task_namespaces() switch_task_namespaces() free_nsproxy() put_mnt_ns() drop_collected_mounts() namespace_unlock() synchronize_rcu() _synchronize_rcu_expedited() schedule_work() // on cpu0 low priority worker pool wait_event() // waiting for work item to execute Task T inserted a work item in the worklist of cpu0 low priority worker pool. It is waiting for expedited grace period work item to execute. This work item will only be executed once kworker/0:0 complete execution of cpuset_hotplug_workfn(). kworker/0:0 ==> Task T ==>kworker/0:0 In case of PF_EXITING task being migrated from source to destination cgroup, migrate next available task in source cgroup. Signed-off-by: Prateek Sood Signed-off-by: Tejun Heo --- kernel/cgroup/cgroup-v1.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/kernel/cgroup/cgroup-v1.c b/kernel/cgroup/cgroup-v1.c index 024085daab1a..a2c05d2476ac 100644 --- a/kernel/cgroup/cgroup-v1.c +++ b/kernel/cgroup/cgroup-v1.c @@ -123,7 +123,11 @@ int cgroup_transfer_tasks(struct cgroup *to, struct cgroup *from) */ do { css_task_iter_start(&from->self, 0, &it); - task = css_task_iter_next(&it); + + do { + task = css_task_iter_next(&it); + } while (task && (task->flags & PF_EXITING)); + if (task) get_task_struct(task); css_task_iter_end(&it); -- cgit v1.2.3 From 74d0833c659a8a54735e5efdd44f4b225af68586 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 20 Dec 2017 07:09:19 -0800 Subject: cgroup: fix css_task_iter crash on CSS_TASK_ITER_PROC While teaching css_task_iter to handle skipping over tasks which aren't group leaders, bc2fb7ed089f ("cgroup: add @flags to css_task_iter_start() and implement CSS_TASK_ITER_PROCS") introduced a silly bug. CSS_TASK_ITER_PROCS is implemented by repeating css_task_iter_advance() while the advanced cursor is pointing to a non-leader thread. However, the cursor variable, @l, wasn't updated when the iteration has to advance to the next css_set and the following repetition would operate on the terminal @l from the previous iteration which isn't pointing to a valid task leading to oopses like the following or infinite looping. BUG: unable to handle kernel NULL pointer dereference at 0000000000000254 IP: __task_pid_nr_ns+0xc7/0xf0 PGD 0 P4D 0 Oops: 0000 [#1] SMP ... CPU: 2 PID: 1 Comm: systemd Not tainted 4.14.4-200.fc26.x86_64 #1 Hardware name: System manufacturer System Product Name/PRIME B350M-A, BIOS 3203 11/09/2017 task: ffff88c4baee8000 task.stack: ffff96d5c3158000 RIP: 0010:__task_pid_nr_ns+0xc7/0xf0 RSP: 0018:ffff96d5c315bd50 EFLAGS: 00010206 RAX: 0000000000000000 RBX: ffff88c4b68c6000 RCX: 0000000000000250 RDX: ffffffffa5e47960 RSI: 0000000000000000 RDI: ffff88c490f6ab00 RBP: ffff96d5c315bd50 R08: 0000000000001000 R09: 0000000000000005 R10: ffff88c4be006b80 R11: ffff88c42f1b8004 R12: ffff96d5c315bf18 R13: ffff88c42d7dd200 R14: ffff88c490f6a510 R15: ffff88c4b68c6000 FS: 00007f9446f8ea00(0000) GS:ffff88c4be680000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000254 CR3: 00000007f956f000 CR4: 00000000003406e0 Call Trace: cgroup_procs_show+0x19/0x30 cgroup_seqfile_show+0x4c/0xb0 kernfs_seq_show+0x21/0x30 seq_read+0x2ec/0x3f0 kernfs_fop_read+0x134/0x180 __vfs_read+0x37/0x160 ? security_file_permission+0x9b/0xc0 vfs_read+0x8e/0x130 SyS_read+0x55/0xc0 entry_SYSCALL_64_fastpath+0x1a/0xa5 RIP: 0033:0x7f94455f942d RSP: 002b:00007ffe81ba2d00 EFLAGS: 00000293 ORIG_RAX: 0000000000000000 RAX: ffffffffffffffda RBX: 00005574e2233f00 RCX: 00007f94455f942d RDX: 0000000000001000 RSI: 00005574e2321a90 RDI: 000000000000002b RBP: 0000000000000000 R08: 00005574e2321a90 R09: 00005574e231de60 R10: 00007f94458c8b38 R11: 0000000000000293 R12: 00007f94458c8ae0 R13: 00007ffe81ba3800 R14: 0000000000000000 R15: 00005574e2116560 Code: 04 74 0e 89 f6 48 8d 04 76 48 8d 04 c5 f0 05 00 00 48 8b bf b8 05 00 00 48 01 c7 31 c0 48 8b 0f 48 85 c9 74 18 8b b2 30 08 00 00 <3b> 71 04 77 0d 48 c1 e6 05 48 01 f1 48 3b 51 38 74 09 5d c3 8b RIP: __task_pid_nr_ns+0xc7/0xf0 RSP: ffff96d5c315bd50 Fix it by moving the initialization of the cursor below the repeat label. While at it, rename it to @next for readability. Signed-off-by: Tejun Heo Fixes: bc2fb7ed089f ("cgroup: add @flags to css_task_iter_start() and implement CSS_TASK_ITER_PROCS") Cc: stable@vger.kernel.org # v4.14+ Reported-by: Laura Abbott Reported-by: Bronek Kozicki Reported-by: George Amanakis Signed-off-by: Tejun Heo --- kernel/cgroup/cgroup.c | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-) diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c index f4c2f8cb5748..2cf06c274e4c 100644 --- a/kernel/cgroup/cgroup.c +++ b/kernel/cgroup/cgroup.c @@ -4125,26 +4125,24 @@ static void css_task_iter_advance_css_set(struct css_task_iter *it) static void css_task_iter_advance(struct css_task_iter *it) { - struct list_head *l = it->task_pos; + struct list_head *next; lockdep_assert_held(&css_set_lock); - WARN_ON_ONCE(!l); - repeat: /* * Advance iterator to find next entry. cset->tasks is consumed * first and then ->mg_tasks. After ->mg_tasks, we move onto the * next cset. */ - l = l->next; + next = it->task_pos->next; - if (l == it->tasks_head) - l = it->mg_tasks_head->next; + if (next == it->tasks_head) + next = it->mg_tasks_head->next; - if (l == it->mg_tasks_head) + if (next == it->mg_tasks_head) css_task_iter_advance_css_set(it); else - it->task_pos = l; + it->task_pos = next; /* if PROCS, skip over tasks which aren't group leaders */ if ((it->flags & CSS_TASK_ITER_PROCS) && it->task_pos && -- cgit v1.2.3 From 4f58424da3deead2605e39a9df65f5f06107a3cb Mon Sep 17 00:00:00 2001 From: Roman Gushchin Date: Wed, 10 Jan 2018 04:35:12 -0800 Subject: cgroup: make cgroup.threads delegatable Make cgroup.threads file delegatable. The behavior of cgroup.threads should follow the behavior of cgroup.procs. Signed-off-by: Roman Gushchin Discovered-by: Michael Kerrisk Cc: Tejun Heo Signed-off-by: Tejun Heo --- kernel/cgroup/cgroup.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c index 2cf06c274e4c..7e4c44538119 100644 --- a/kernel/cgroup/cgroup.c +++ b/kernel/cgroup/cgroup.c @@ -4447,6 +4447,7 @@ static struct cftype cgroup_base_files[] = { }, { .name = "cgroup.threads", + .flags = CFTYPE_NS_DELEGATABLE, .release = cgroup_procs_release, .seq_start = cgroup_threads_start, .seq_next = cgroup_procs_next, -- cgit v1.2.3