[PATCH] cgroup/cpuset: Use effective_xcpus in partcmd_update add/del mask calculation

[PATCH] cgroup/cpuset: Use effective_xcpus in partcmd_update add/del mask calculation

[Sun Shaojie]

When sibling CPU exclusion occurs, a partition's user_xcpus may contain
CPUs that were never actually granted to it. These CPUs are present in
user_xcpus(cs) but not in cs->effective_xcpus.

The partcmd_update path in update_parent_effective_cpumask() uses
user_xcpus(cs) (via the local variable xcpus) to compute the addmask
(CPUs to return to parent) and delmask (CPUs to request from parent).
This is incorrect:

 1) When newmask removes a CPU that was previously excluded by a
    sibling, addmask incorrectly includes that CPU and tries to return
    it to the parent even though the partition never actually owned it,
    causing CPU overlap with sibling partitions and triggering warnings
    in generate_sched_domains().

 2) When newmask adds a previously excluded CPU that is now available,
    delmask fails to request it from the parent because user_xcpus(cs)
    already includes it.

Fix this by using cs->effective_xcpus instead of user_xcpus(cs) in all
partcmd_update paths that calculate addmask or delmask, including the
PERR_NOCPUS error handling paths.

Reproducers:

  Example 1 - Removing a sibling-excluded CPU incorrectly returns it:

    # cd /sys/fs/cgroup
    # echo "0-1" > a1/cpuset.cpus
    # echo "root" > a1/cpuset.cpus.partition
    # echo "0-2" > b1/cpuset.cpus
    # echo "root" > b1/cpuset.cpus.partition
    # echo "2" > b1/cpuset.cpus
    # cat cpuset.cpus.effective
    # Actual: 0-1,3    Expected: 3

  Example 2 - Expanding to a previously excluded CPU fails to request it:

    # cd /sys/fs/cgroup
    # echo "0-1" > a1/cpuset.cpus
    # echo "root" > a1/cpuset.cpus.partition
    # echo "0-2" > b1/cpuset.cpus
    # echo "root" > b1/cpuset.cpus.partition
    # echo "member" > a1/cpuset.cpus.partition
    # echo "1-2" > b1/cpuset.cpus
    # cat cpuset.cpus.effective
    # Actual: 0-1,3    Expected: 0,3

Fixes: 2a3602030d80 ("cgroup/cpuset: Don't invalidate sibling partitions on cpuset.cpus conflict")
Signed-off-by: Sun Shaojie <sunshaojie@kylinos.cn>
---
 kernel/cgroup/cpuset.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index 1335e437098e..5a5fa2481467 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -1821,11 +1821,11 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd,
 			deleting = cpumask_and(tmp->delmask,
 					newmask, parent->effective_xcpus);
 		} else {
-			cpumask_andnot(tmp->addmask, xcpus, newmask);
+			cpumask_andnot(tmp->addmask, cs->effective_xcpus, newmask);
 			adding = cpumask_and(tmp->addmask, tmp->addmask,
 					     parent->effective_xcpus);
 
-			cpumask_andnot(tmp->delmask, newmask, xcpus);
+			cpumask_andnot(tmp->delmask, newmask, cs->effective_xcpus);
 			deleting = cpumask_and(tmp->delmask, tmp->delmask,
 					       parent->effective_xcpus);
 		}
@@ -1864,7 +1864,7 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd,
 			part_error = PERR_NOCPUS;
 			deleting = false;
 			adding = cpumask_and(tmp->addmask,
-					     xcpus, parent->effective_xcpus);
+					     cs->effective_xcpus, parent->effective_xcpus);
 		}
 	} else {
 		/*
@@ -1886,7 +1886,8 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd,
 			part_error = PERR_NOCPUS;
 			if (is_partition_valid(cs))
 				adding = cpumask_and(tmp->addmask,
-						xcpus, parent->effective_xcpus);
+						     cs->effective_xcpus,
+						     parent->effective_xcpus);
 		} else if (is_partition_invalid(cs) && !cpumask_empty(xcpus) &&
 			   cpumask_subset(xcpus, parent->effective_xcpus)) {
 			struct cgroup_subsys_state *css;
-- 
2.43.0

Re: [PATCH] cgroup/cpuset: Use effective_xcpus in partcmd_update add/del mask calculation

[Guopeng Zhang]

在 2026/5/22 15:53, Sun Shaojie 写道:
> When sibling CPU exclusion occurs, a partition's user_xcpus may contain
> CPUs that were never actually granted to it. These CPUs are present in
> user_xcpus(cs) but not in cs->effective_xcpus.
> 
> The partcmd_update path in update_parent_effective_cpumask() uses
> user_xcpus(cs) (via the local variable xcpus) to compute the addmask
> (CPUs to return to parent) and delmask (CPUs to request from parent).
> This is incorrect:
> 
>  1) When newmask removes a CPU that was previously excluded by a
>     sibling, addmask incorrectly includes that CPU and tries to return
>     it to the parent even though the partition never actually owned it,
>     causing CPU overlap with sibling partitions and triggering warnings
>     in generate_sched_domains().
> 
>  2) When newmask adds a previously excluded CPU that is now available,
>     delmask fails to request it from the parent because user_xcpus(cs)
>     already includes it.
> 
> Fix this by using cs->effective_xcpus instead of user_xcpus(cs) in all
> partcmd_update paths that calculate addmask or delmask, including the
> PERR_NOCPUS error handling paths.
> 
> Reproducers:
> 
>   Example 1 - Removing a sibling-excluded CPU incorrectly returns it:
> 
>     # cd /sys/fs/cgroup
>     # echo "0-1" > a1/cpuset.cpus
>     # echo "root" > a1/cpuset.cpus.partition
>     # echo "0-2" > b1/cpuset.cpus
>     # echo "root" > b1/cpuset.cpus.partition
>     # echo "2" > b1/cpuset.cpus
>     # cat cpuset.cpus.effective
>     # Actual: 0-1,3    Expected: 3
> 
>   Example 2 - Expanding to a previously excluded CPU fails to request it:
> 
>     # cd /sys/fs/cgroup
>     # echo "0-1" > a1/cpuset.cpus
>     # echo "root" > a1/cpuset.cpus.partition
>     # echo "0-2" > b1/cpuset.cpus
>     # echo "root" > b1/cpuset.cpus.partition
>     # echo "member" > a1/cpuset.cpus.partition
>     # echo "1-2" > b1/cpuset.cpus
>     # cat cpuset.cpus.effective
>     # Actual: 0-1,3    Expected: 0,3
> 
> Fixes: 2a3602030d80 ("cgroup/cpuset: Don't invalidate sibling partitions on cpuset.cpus conflict")
> Signed-off-by: Sun Shaojie <sunshaojie@kylinos.cn>
> ---
>  kernel/cgroup/cpuset.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
> index 1335e437098e..5a5fa2481467 100644
> --- a/kernel/cgroup/cpuset.c
> +++ b/kernel/cgroup/cpuset.c
> @@ -1821,11 +1821,11 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd,
>  			deleting = cpumask_and(tmp->delmask,
>  					newmask, parent->effective_xcpus);
>  		} else {
> -			cpumask_andnot(tmp->addmask, xcpus, newmask);
> +			cpumask_andnot(tmp->addmask, cs->effective_xcpus, newmask);
>  			adding = cpumask_and(tmp->addmask, tmp->addmask,
>  					     parent->effective_xcpus);
>  
> -			cpumask_andnot(tmp->delmask, newmask, xcpus);
> +			cpumask_andnot(tmp->delmask, newmask, cs->effective_xcpus);
>  			deleting = cpumask_and(tmp->delmask, tmp->delmask,
>  					       parent->effective_xcpus);
>  		}
> @@ -1864,7 +1864,7 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd,
>  			part_error = PERR_NOCPUS;
>  			deleting = false;
>  			adding = cpumask_and(tmp->addmask,
> -					     xcpus, parent->effective_xcpus);
> +					     cs->effective_xcpus, parent->effective_xcpus);
>  		}
>  	} else {
>  		/*
> @@ -1886,7 +1886,8 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd,
>  			part_error = PERR_NOCPUS;
>  			if (is_partition_valid(cs))
>  				adding = cpumask_and(tmp->addmask,
> -						xcpus, parent->effective_xcpus);
> +						     cs->effective_xcpus,
> +						     parent->effective_xcpus);
>  		} else if (is_partition_invalid(cs) && !cpumask_empty(xcpus) &&
>  			   cpumask_subset(xcpus, parent->effective_xcpus)) {
>  			struct cgroup_subsys_state *css;
Hi, Shaojie

The code change looks reasonable to me, but I think the comment above
the partcmd_update calculation should be updated as well.

Maybe it can be updated like this:

		 * Compute add/delete mask to/from effective_cpus
		 *
		 * For valid partition:
-		 *   addmask = exclusive_cpus & ~newmask
+		 *   addmask = cs->effective_xcpus & ~newmask
		 *			      & parent->effective_xcpus
-		 *   delmask = newmask & ~exclusive_cpus
+		 *   delmask = newmask & ~cs->effective_xcpus
		 *		       & parent->effective_xcpus
		 *
		 * For invalid partition:

Does this look reasonable to you?

Tested-by: Guopeng Zhang <zhangguopeng@kylinos.cn>

Thanks.

Re: [PATCH] cgroup/cpuset: Use effective_xcpus in partcmd_update add/del mask calculation

[Waiman Long]

On 5/22/26 4:41 AM, Guopeng Zhang wrote:
>
> 在 2026/5/22 15:53, Sun Shaojie 写道:
>> When sibling CPU exclusion occurs, a partition's user_xcpus may contain
>> CPUs that were never actually granted to it. These CPUs are present in
>> user_xcpus(cs) but not in cs->effective_xcpus.
>>
>> The partcmd_update path in update_parent_effective_cpumask() uses
>> user_xcpus(cs) (via the local variable xcpus) to compute the addmask
>> (CPUs to return to parent) and delmask (CPUs to request from parent).
>> This is incorrect:
>>
>>   1) When newmask removes a CPU that was previously excluded by a
>>      sibling, addmask incorrectly includes that CPU and tries to return
>>      it to the parent even though the partition never actually owned it,
>>      causing CPU overlap with sibling partitions and triggering warnings
>>      in generate_sched_domains().
>>
>>   2) When newmask adds a previously excluded CPU that is now available,
>>      delmask fails to request it from the parent because user_xcpus(cs)
>>      already includes it.
>>
>> Fix this by using cs->effective_xcpus instead of user_xcpus(cs) in all
>> partcmd_update paths that calculate addmask or delmask, including the
>> PERR_NOCPUS error handling paths.
>>
>> Reproducers:
>>
>>    Example 1 - Removing a sibling-excluded CPU incorrectly returns it:
>>
>>      # cd /sys/fs/cgroup
>>      # echo "0-1" > a1/cpuset.cpus
>>      # echo "root" > a1/cpuset.cpus.partition
>>      # echo "0-2" > b1/cpuset.cpus
>>      # echo "root" > b1/cpuset.cpus.partition
>>      # echo "2" > b1/cpuset.cpus
>>      # cat cpuset.cpus.effective
>>      # Actual: 0-1,3    Expected: 3
>>
>>    Example 2 - Expanding to a previously excluded CPU fails to request it:
>>
>>      # cd /sys/fs/cgroup
>>      # echo "0-1" > a1/cpuset.cpus
>>      # echo "root" > a1/cpuset.cpus.partition
>>      # echo "0-2" > b1/cpuset.cpus
>>      # echo "root" > b1/cpuset.cpus.partition
>>      # echo "member" > a1/cpuset.cpus.partition
>>      # echo "1-2" > b1/cpuset.cpus
>>      # cat cpuset.cpus.effective
>>      # Actual: 0-1,3    Expected: 0,3
>>
>> Fixes: 2a3602030d80 ("cgroup/cpuset: Don't invalidate sibling partitions on cpuset.cpus conflict")
>> Signed-off-by: Sun Shaojie <sunshaojie@kylinos.cn>
>> ---
>>   kernel/cgroup/cpuset.c | 9 +++++----
>>   1 file changed, 5 insertions(+), 4 deletions(-)
>>
>> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
>> index 1335e437098e..5a5fa2481467 100644
>> --- a/kernel/cgroup/cpuset.c
>> +++ b/kernel/cgroup/cpuset.c
>> @@ -1821,11 +1821,11 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd,
>>   			deleting = cpumask_and(tmp->delmask,
>>   					newmask, parent->effective_xcpus);
>>   		} else {
>> -			cpumask_andnot(tmp->addmask, xcpus, newmask);
>> +			cpumask_andnot(tmp->addmask, cs->effective_xcpus, newmask);
>>   			adding = cpumask_and(tmp->addmask, tmp->addmask,
>>   					     parent->effective_xcpus);
>>   
>> -			cpumask_andnot(tmp->delmask, newmask, xcpus);
>> +			cpumask_andnot(tmp->delmask, newmask, cs->effective_xcpus);
>>   			deleting = cpumask_and(tmp->delmask, tmp->delmask,
>>   					       parent->effective_xcpus);
>>   		}
>> @@ -1864,7 +1864,7 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd,
>>   			part_error = PERR_NOCPUS;
>>   			deleting = false;
>>   			adding = cpumask_and(tmp->addmask,
>> -					     xcpus, parent->effective_xcpus);
>> +					     cs->effective_xcpus, parent->effective_xcpus);
>>   		}
>>   	} else {
>>   		/*
>> @@ -1886,7 +1886,8 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd,
>>   			part_error = PERR_NOCPUS;
>>   			if (is_partition_valid(cs))
>>   				adding = cpumask_and(tmp->addmask,
>> -						xcpus, parent->effective_xcpus);
>> +						     cs->effective_xcpus,
>> +						     parent->effective_xcpus);
>>   		} else if (is_partition_invalid(cs) && !cpumask_empty(xcpus) &&
>>   			   cpumask_subset(xcpus, parent->effective_xcpus)) {
>>   			struct cgroup_subsys_state *css;
> Hi, Shaojie
>
> The code change looks reasonable to me, but I think the comment above
> the partcmd_update calculation should be updated as well.
>
> Maybe it can be updated like this:
>
> 		 * Compute add/delete mask to/from effective_cpus
> 		 *
> 		 * For valid partition:
> -		 *   addmask = exclusive_cpus & ~newmask
> +		 *   addmask = cs->effective_xcpus & ~newmask
> 		 *			      & parent->effective_xcpus
> -		 *   delmask = newmask & ~exclusive_cpus
> +		 *   delmask = newmask & ~cs->effective_xcpus
> 		 *		       & parent->effective_xcpus
> 		 *
> 		 * For invalid partition:
>
> Does this look reasonable to you?
>
> Tested-by: Guopeng Zhang <zhangguopeng@kylinos.cn>

Shaojie, thanks for the fix. The changes look good to me. I also like 
the suggested changes in the comment as suggested by Guopeng. Would you 
mind sending a v2 with this additional change as well?

Thanks,
Longman