linux-kernel - Re: [PATCH -next] cgroup/cpuset: Do not clear xcpus when clearing cpus

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <a3cee760-398f-4661-b4b5-f2fcfd5de7b7@redhat.com>
Date: Thu, 1 Aug 2024 12:31:44 -0400
From: Waiman Long <longman@...hat.com>
To: Chen Ridong <chenridong@...wei.com>, tj@...nel.org,
 lizefan.x@...edance.com, hannes@...xchg.org, adityakali@...gle.com,
 sergeh@...nel.org
Cc: bpf@...r.kernel.org, cgroups@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH -next] cgroup/cpuset: Do not clear xcpus when clearing
 cpus


On 7/31/24 23:22, Waiman Long wrote:
> On 7/31/24 05:21, Chen Ridong wrote:
>> After commit 737bb142a00d ("cgroup/cpuset: Make cpuset.cpus.exclusive
>> independent of cpuset.cpus"), cpuset.cpus.exclusive and cpuset.cpus
>> became independent. However we found that 
>> cpuset.cpus.exclusive.effective
>> is cleared when cpuset.cpus is clear. To fix this issue, just remove 
>> xcpus
>> clearing when cpuset.cpus is being cleared.
>>
>> It can be reproduced as below:
>> cd /sys/fs/cgroup/
>> mkdir test
>> echo +cpuset > cgroup.subtree_control
>> cd test
>> echo 3 > cpuset.cpus.exclusive
>> cat cpuset.cpus.exclusive.effective
>> 3
>> echo > cpuset.cpus
>> cat cpuset.cpus.exclusive.effective // was cleared
>>
>> Signed-off-by: Chen Ridong <chenridong@...wei.com>
>> ---
>>   kernel/cgroup/cpuset.c | 5 ++---
>>   1 file changed, 2 insertions(+), 3 deletions(-)
>>
>> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
>> index a9b6d56eeffa..248c39bebbe9 100644
>> --- a/kernel/cgroup/cpuset.c
>> +++ b/kernel/cgroup/cpuset.c
>> @@ -2523,10 +2523,9 @@ static int update_cpumask(struct cpuset *cs, 
>> struct cpuset *trialcs,
>>        * that parsing.  The validate_change() call ensures that cpusets
>>        * with tasks have cpus.
>>        */
>> -    if (!*buf) {
>> +    if (!*buf)
>>           cpumask_clear(trialcs->cpus_allowed);
>> -        cpumask_clear(trialcs->effective_xcpus);
>> -    } else {
>> +    else {
>>           retval = cpulist_parse(buf, trialcs->cpus_allowed);
>>           if (retval < 0)
>>               return retval;
>
> Yes, that is a corner case bug that has not been properly handled.
>
> Reviewed-by: Waiman Long <longman@...hat.com>
>
With a second thought, I think we should keep the clearing of 
effective_xcpus if exclusive_cpus is empty. IOW

diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index 6ba8313f1fc3..2023cd68d9bc 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -2516,7 +2516,8 @@ static int update_cpumask(struct cpuset *cs, 
struct cpuset *trialcs,
          */
         if (!*buf) {
                 cpumask_clear(trialcs->cpus_allowed);
-               cpumask_clear(trialcs->effective_xcpus);
+               if (cpumask_empty(trialcs->exclusive_cpus))
+ cpumask_clear(trialcs->effective_xcpus);
         } else {
                 retval = cpulist_parse(buf, trialcs->cpus_allowed);
                 if (retval < 0)

Thanks,
Longman