linux-kernel - Re: [PATCH -next RFC 06/11] cpuset: introduce cpus_excl_conflict and mems_excl

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: <27beb9ee-114f-4071-88e5-cb8889826a53@huaweicloud.com>
Date: Sat, 30 Aug 2025 09:27:15 +0800
From: Chen Ridong <chenridong@...weicloud.com>
To: Waiman Long <llong@...hat.com>, tj@...nel.org, hannes@...xchg.org,
 mkoutny@...e.com
Cc: cgroups@...r.kernel.org, linux-kernel@...r.kernel.org,
 lujialin4@...wei.com, chenridong@...wei.com
Subject: Re: [PATCH -next RFC 06/11] cpuset: introduce cpus_excl_conflict and
 mems_excl_conflict helpers



On 2025/8/30 3:29, Waiman Long wrote:
> 
> On 8/28/25 8:56 AM, Chen Ridong wrote:
>> From: Chen Ridong <chenridong@...wei.com>
>>
>> This patch adds cpus_excl_conflict() and mems_excl_conflict() helper
>> functions to improve code readability and maintainability. The exclusive
>> conflict checking follows these rules:
>>
>> 1. If either cpuset has the 'exclusive' flag set, their user_xcpus must
>>     not have any overlap.
>> 2. If both cpusets are non-exclusive, their 'cpuset.cpus.exclusive' values
>>     must not intersect.
> Do you mean "both cpusets are exclusive"?

Thank you Longman.

I meant that neither of the two cpusets has the "exclusive" flag set. Case 1 already handles
situations where one or both cpusets have the exclusive flag enabled.

>> 3. The 'cpuset.cpus' of one cpuset must not form a subset of another
>>     cpuset's 'cpuset.cpus.exclusive'.
>>
>> Signed-off-by: Chen Ridong <chenridong@...wei.com>
>> ---
>>   kernel/cgroup/cpuset.c | 62 ++++++++++++++++++++++--------------------
>>   1 file changed, 32 insertions(+), 30 deletions(-)
>>
>> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
>> index 5dd1e9552000..5cfc53fe717c 100644
>> --- a/kernel/cgroup/cpuset.c
>> +++ b/kernel/cgroup/cpuset.c
>> @@ -584,6 +584,35 @@ static inline bool cpusets_are_exclusive(struct cpuset *cs1, struct cpuset *cs2)
>>       return true;
>>   }
>>   +static inline bool cpus_excl_conflict(struct cpuset *cs1, struct cpuset *cs2)
>> +{
>> +    /* One is exclusive, they must be exclusive */
> The comment is hard to understand. Basically, if one cpuset has exclusive flag set (a v1 feature),
> they must be exclusive wrt each other.

My apologies for the earlier comment.

To quote your remark:
  "Basically, if one cpuset has the exclusive flag set (a v1 feature), they must be exclusive with
wrt to each other."

That is exactly what I meant. Although cpuset v2 does not expose a user interface to explicitly set
the exclusive flag, the flag will still be set internally when a partition is enabled with
update_partition_exclusive_flag function.

>> +    if (is_cpu_exclusive(cs1) || is_cpu_exclusive(cs2))
>> +        return !cpusets_are_exclusive(cs1, cs2);
>> +
>> +    /* Exclusive_cpus can not have intersects*/
> Grammatical mistake, better wording - "exclusive_cpus cannot intersect"
>> +    if (cpumask_intersects(cs1->exclusive_cpus, cs2->exclusive_cpus))
>> +        return true;
>> +
>> +    /* One cpus_allowed can not be a subset of another's cpuset.effective_cpus */
> "cpus_allowed of one cpuset cannot be a subset of another cpuset's exclusive_cpus"

Will update with next version.

>> +    if (!cpumask_empty(cs1->cpus_allowed) &&
>> +        cpumask_subset(cs1->cpus_allowed, cs2->exclusive_cpus))
>> +        return true;
>> +
>> +    if (!cpumask_empty(cs2->cpus_allowed) &&
>> +        cpumask_subset(cs2->cpus_allowed, cs1->exclusive_cpus))
>> +        return true;
>> +
>> +    return false;
>> +}
>> +
> Cheers,
> Longman

-- 
Best regards,
Ridong