在 2022/3/16 22:27, Greg Kroah-Hartman 写道:
> On Mon, Mar 14, 2022 at 12:19:41PM +0100, Michal Koutný wrote:
>> Hello.
>>
>> In my opinion there are two approaches:
>> a) drop this backport (given other races present),
>
> I have no problem with that, want to send a revert patch?
>
>> b) swap the locks compatible with v4.19 as this patch proposes.
>>
>> On Mon, Mar 14, 2022 at 05:11:50PM +0800, Zhang Qiao <zhangqiao22@xxxxxxxxxx> wrote:
>>> + /*
>>> + * It should hold cpus lock because a cpu offline event can
>>> + * cause set_cpus_allowed_ptr() failed.
>>> + */
>>> + cpus_read_lock();
>>
>> Maybe just a nit, the old kernels before commit c5c63b9a6a2e ("cgroup:
>> Replace deprecated CPU-hotplug functions.") v5.15-rc1~159^2~5
>> would be more consistent with get_online_cpus() here (but they're
>> equivalent functionally so the locking order is correct).
>
> A fixed up patch would also be appreciated :)
>
Fixed up patch as follows, replace cpus_read_lock() with get_online_cpus().
thanks.
--------
[PATCH] cpuset: Fix unsafe lock order between cpuset lock and cpuslock
The backport commit 4eec5fe1c680a ("cgroup/cpuset: Fix a race
between cpuset_attach() and cpu hotplug") looks suspicious since
it comes before commit d74b27d63a8b ("cgroup/cpuset: Change
cpuset_rwsem and hotplug lock order") v5.4-rc1~176^2~30 when
the locking order was: cpuset lock, cpus lock.
Fix it with the correct locking order and reduce the cpus locking
range because only set_cpus_allowed_ptr() needs the protection of
cpus lock.
Fixes: 4eec5fe1c680a ("cgroup/cpuset: Fix a race between cpuset_attach() and cpu hotplug")
Reported-by: Michal Koutný <mkoutny@xxxxxxxx>
Signed-off-by: Zhang Qiao <zhangqiao22@xxxxxxxxxx>
---
kernel/cgroup/cpuset.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index d43d25acc..4e1c4232e 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -1528,9 +1528,13 @@ static void cpuset_attach(struct cgroup_taskset *tset)
cgroup_taskset_first(tset, &css);
cs = css_cs(css);
- cpus_read_lock();
mutex_lock(&cpuset_mutex);
+ /*
+ * It should hold cpus lock because a cpu offline event can
+ * cause set_cpus_allowed_ptr() failed.
+ */
+ get_online_cpus();
/* prepare for attach */
if (cs == &top_cpuset)
cpumask_copy(cpus_attach, cpu_possible_mask);
@@ -1549,6 +1553,7 @@ static void cpuset_attach(struct cgroup_taskset *tset)
cpuset_change_task_nodemask(task, &cpuset_attach_nodemask_to);
cpuset_update_task_spread_flag(cs, task);
}
+ put_online_cpus();
/*
* Change mm for all threadgroup leaders. This is expensive and may
@@ -1584,7 +1589,6 @@ static void cpuset_attach(struct cgroup_taskset *tset)
wake_up(&cpuset_attach_wq);
mutex_unlock(&cpuset_mutex);
- cpus_read_unlock();
}
/* The various types of files and directories in a cpuset file system */
--
2.18.0
> thanks,
>
> greg k-h
> .
>
