niansa-misc/linux
1
0
Fork 0
mirror of synced 2025-03-06 20:59:54 +01:00
Code Issues Projects Releases Packages Wiki Activity
linux/kernel
History
Shakeel Butt b69bb476de cgroup: fix race between fork and cgroup.kill 
Tejun reported the following race between fork() and cgroup.kill at [1].

Tejun:
  I was looking at cgroup.kill implementation and wondering whether there
  could be a race window. So, __cgroup_kill() does the following:

   k1. Set CGRP_KILL.
   k2. Iterate tasks and deliver SIGKILL.
   k3. Clear CGRP_KILL.

  The copy_process() does the following:

   c1. Copy a bunch of stuff.
   c2. Grab siglock.
   c3. Check fatal_signal_pending().
   c4. Commit to forking.
   c5. Release siglock.
   c6. Call cgroup_post_fork() which puts the task on the css_set and tests
       CGRP_KILL.

  The intention seems to be that either a forking task gets SIGKILL and
  terminates on c3 or it sees CGRP_KILL on c6 and kills the child. However, I
  don't see what guarantees that k3 can't happen before c6. ie. After a
  forking task passes c5, k2 can take place and then before the forking task
  reaches c6, k3 can happen. Then, nobody would send SIGKILL to the child.
  What am I missing?

This is indeed a race. One way to fix this race is by taking
cgroup_threadgroup_rwsem in write mode in __cgroup_kill() as the fork()
side takes cgroup_threadgroup_rwsem in read mode from cgroup_can_fork()
to cgroup_post_fork(). However that would be heavy handed as this adds
one more potential stall scenario for cgroup.kill which is usually
called under extreme situation like memory pressure.

To fix this race, let's maintain a sequence number per cgroup which gets
incremented on __cgroup_kill() call. On the fork() side, the
cgroup_can_fork() will cache the sequence number locally and recheck it
against the cgroup's sequence number at cgroup_post_fork() site. If the
sequence numbers mismatch, it means __cgroup_kill() can been called and
we should send SIGKILL to the newly created task.

Reported-by: Tejun Heo <tj@kernel.org>
Closes: https://lore.kernel.org/all/Z5QHE2Qn-QZ6M-KW@slm.duckdns.org/ [1]
Fixes: 661ee62809 ("cgroup: introduce cgroup.kill")
Cc: stable@vger.kernel.org # v5.14+
Signed-off-by: Shakeel Butt <shakeel.butt@linux.dev>
Reviewed-by: Michal Koutný <mkoutny@suse.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
2025-02-02 06:54:51 -10:00
..
bpf Summary: 2025-01-29 10:35:40 -08:00
cgroup cgroup: fix race between fork and cgroup.kill 2025-02-02 06:54:51 -10:00
configs configs/debug: make sure PROVE_RCU_LIST=y takes effect 2024-10-28 10:21:09 -07:00
debug kdb: Remove unused flags stack 2025-01-25 08:22:26 +00:00
dma dma-debug: fix physical address calculation for struct dma_debug_entry 2024-11-28 10:19:16 +01:00
entry sched: Add TIF_NEED_RESCHED_LAZY infrastructure 2024-11-05 12:55:37 +01:00
events kernel: be more careful about dup_mmap() failures and uprobe registering 2025-02-01 03:53:25 -08:00
futex Mainly individually changelogged singleton patches. The patch series in 2025-01-26 17:50:53 -08:00
gcov gcov: clang: use correct function param names 2025-01-24 22:47:27 -08:00
irq Updates for the interrupt subsystem: 2025-01-21 13:51:07 -08:00
kcsan kcsan: Remove redundant call of kallsyms_lookup_name() 2024-10-14 16:44:56 +02:00
livepatch livepatch: Add stack_order sysfs attribute 2024-12-09 11:44:03 +01:00
locking treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
module Kbuild updates for v6.14 2025-01-31 12:07:07 -08:00
power More power management updates for 6.14-rc1 2025-01-30 15:10:34 -08:00
printk treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
rcu The various patchsets are summarized below. Plus of course many 2025-01-26 18:36:23 -08:00
sched More power management updates for 6.14-rc1 2025-01-30 15:10:34 -08:00
time treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
trace treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
.gitignore
acct.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
async.c
audit.c audit: Initialize lsmctx to avoid memory allocation error 2025-01-29 20:02:04 -05:00
audit.h audit: change context data from secid to lsm_prop 2024-10-11 14:34:16 -04:00
audit_fsnotify.c
audit_tree.c fsnotify: create a wrapper fsnotify_find_inode_mark() 2024-04-04 16:24:16 +02:00
audit_watch.c fsnotify: create a wrapper fsnotify_find_inode_mark() 2024-04-04 16:24:16 +02:00
auditfilter.c audit: fix suffixed '/' filename matching 2024-12-05 19:22:38 -05:00
auditsc.c lsm/stable-6.14 PR 20250121 2025-01-21 20:03:04 -08:00
backtracetest.c backtracetest: add MODULE_DESCRIPTION() 2024-06-24 22:24:55 -07:00
bounds.c bounds: Use the right number of bits for power-of-two CONFIG_NR_CPUS 2024-04-29 08:29:29 -07:00
capability.c kernel: remove get_task_comm() and print task comm directly 2025-01-12 20:21:15 -08:00
cfi.c
compat.c
configs.c
context_tracking.c context_tracking, rcu: Rename rcu_dyntick trace event into rcu_watching 2024-08-15 21:30:43 +05:30
cpu.c The various patchsets are summarized below. Plus of course many 2025-01-26 18:36:23 -08:00
cpu_pm.c
crash_core.c kexec/crash: no crash update when kexec in progress 2024-11-05 17:12:27 -08:00
crash_reserve.c crash: fix crash memory reserve exceed system memory bug 2024-09-01 20:43:30 -07:00
cred.c cred: remove old {override,revert}_creds() helpers 2024-12-02 11:25:09 +01:00
delayacct.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
dma.c
elfcorehdr.c
exec_domain.c
exit.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
exit.h
extable.c
fail_function.c
fork.c kernel: be more careful about dup_mmap() failures and uprobe registering 2025-02-01 03:53:25 -08:00
freezer.c sched/fair: Fix external p->on_rq users 2024-10-14 09:14:35 +02:00
gen_kheaders.sh Kbuild updates for v6.14 2025-01-31 12:07:07 -08:00
groups.c
hung_task.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
iomem.c
irq_work.c kasan: make kasan_record_aux_stack_noalloc() the default behaviour 2025-01-13 22:40:36 -08:00
jump_label.c jump_label: Fix static_key_slow_dec() yet again 2024-09-10 11:57:27 +02:00
kallsyms.c kallsyms: Match symbols exactly with CONFIG_LTO_CLANG 2024-08-15 09:33:35 -07:00
kallsyms_internal.h kallsyms: get rid of code for absolute kallsyms 2024-07-20 16:33:21 +09:00
kallsyms_selftest.c kallsyms: Use kthread_run_on_cpu() 2025-01-02 22:12:12 +01:00
kallsyms_selftest.h
kcmp.c get rid of ...lookup...fdget_rcu() family 2024-10-07 13:34:41 -04:00
Kconfig.freezer
Kconfig.hz
Kconfig.kexec crash, powerpc: default to CRASH_DUMP=n on PPC_BOOK3S_32 2024-11-14 22:43:48 -08:00
Kconfig.locks
Kconfig.preempt sched: No PREEMPT_RT=y for all{yes,mod}config 2024-11-07 15:25:05 +01:00
kcov.c kcov: mark in_softirq_really() as __always_inline 2024-12-30 17:59:08 -08:00
kexec.c crash: add a new kexec flag for hotplug support 2024-04-23 14:59:01 +10:00
kexec_core.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
kexec_elf.c
kexec_file.c kexec_file: fix elfcorehdr digest exclusion when CONFIG_CRASH_HOTPLUG=y 2024-09-01 17:59:01 -07:00
kexec_internal.h kexec: use atomic_try_cmpxchg_acquire() in kexec_trylock() 2024-09-01 20:43:23 -07:00
kheaders.c kheaders: Simplify attribute through __BIN_ATTR_SIMPLE_RO() 2024-12-24 09:46:49 +01:00
kprobes.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
ksyms_common.c
ksysfs.c kernel/ksysfs.c: simplify bin_attribute definition 2025-01-07 16:59:15 +01:00
kthread.c Mainly individually changelogged singleton patches. The patch series in 2025-01-26 17:50:53 -08:00
latencytop.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
Makefile mm: move kernel/numa.c to mm/ 2024-09-03 21:15:26 -07:00
module_signature.c
notifier.c reboot: move reboot_notifier_list to kernel/reboot.c 2024-11-05 17:12:31 -08:00
nsproxy.c fdget(), trivial conversions 2024-11-03 01:28:06 -05:00
padata.c padata: avoid UAF for reorder_work 2025-01-19 12:44:28 +08:00
panic.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
params.c module: Constify 'struct module_attribute' 2025-01-26 13:05:23 +01:00
pid.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
pid_namespace.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
pid_sysctl.h treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
profile.c profiling: remove profile=sleep support 2024-08-04 13:36:28 -07:00
ptrace.c
range.c
reboot.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
regset.c regset: use kvzalloc() for regset_get_alloc() 2024-04-25 21:07:03 -07:00
relay.c [tree-wide] finally take no_llseek out 2024-09-27 08:18:43 -07:00
resource.c kernel/resource: simplify API __devm_release_region() implementation 2025-01-12 20:20:58 -08:00
resource_kunit.c resource, kunit: fix user-after-free in resource_test_region_intersects() 2024-10-09 12:47:19 -07:00
rseq.c rseq: Fix rseq unregistration regression 2025-01-21 08:10:51 +01:00
scftorture.c scftorture: Handle NULL argument passed to scf_add_to_free_list(). 2024-11-14 16:09:51 -08:00
scs.c
seccomp.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
signal.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
smp.c CSD-lock pull request for v6.14 2025-01-28 11:34:03 -08:00
smpboot.c
smpboot.h
softirq.c softirq: Allow raising SCHED_SOFTIRQ from SMP-call-function on RT kernel 2024-12-02 12:01:27 +01:00
stackleak.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
stacktrace.c
static_call.c
static_call_inline.c x86/static-call: provide a way to do very early static-call updates 2024-12-13 09:28:32 +01:00
stop_machine.c stop_machine: Fix rcu_momentary_eqs() call in multi_cpu_stop() 2024-12-11 20:50:47 -08:00
sys.c tracing: Add task_prctl_unknown tracepoint 2024-12-22 20:28:11 -08:00
sys_ni.c Probes updates for v6.11: 2024-07-18 12:19:20 -07:00
sysctl-test.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
sysctl.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
task_work.c kasan: make kasan_record_aux_stack_noalloc() the default behaviour 2025-01-13 22:40:36 -08:00
taskstats.c fdget(), more trivial conversions 2024-11-03 01:28:06 -05:00
torture.c torture: Add MODULE_DESCRIPTION() 2024-05-30 15:31:38 -07:00
tracepoint.c tracing: Fix syscall tracepoint use-after-free 2024-11-01 14:37:31 -04:00
tsacct.c tsacct: replace strncpy() with strscpy() 2024-07-12 16:39:53 -07:00
ucount.c ucounts: move kfree() out of critical zone protected by ucounts_lock 2025-01-12 20:21:00 -08:00
uid16.c
uid16.h
umh.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
up.c
user-return-notifier.c
user.c uidgid: make sure we fit into one cacheline 2024-09-12 12:16:09 +02:00
user_namespace.c user_namespace: use kmemdup_array() instead of kmemdup() for multiple allocation 2024-09-09 16:47:42 -07:00
usermode_driver.c
utsname.c
utsname_sysctl.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
vhost_task.c vhost_task: Handle SIGKILL by flushing work and exiting 2024-05-22 08:31:15 -04:00
vmcore_info.c mm: support only one page_type per page 2024-09-03 21:15:43 -07:00
watch_queue.c watch_queue: Use page->private instead of page->index 2024-12-22 11:29:51 +01:00
watchdog.c treewide: const qualify ctl_tables where applicable 2025-01-28 13:48:37 +01:00
watchdog_buddy.c
watchdog_perf.c watchdog/perf: properly initialize the turbo mode timestamp and rearm counter 2024-07-17 21:11:34 -07:00
workqueue.c The various patchsets are summarized below. Plus of course many 2025-01-26 18:36:23 -08:00
workqueue_internal.h