Hi, Michal,
On Tue, 18 Nov 2025 18:52:24 +0100, Michal Koutný wrote:
On Mon, Nov 17, 2025 at 09:57:08AM +0800, Sun Shaojie sunshaojie@kylinos.cn wrote:
This patch ensures that for sibling cpusets A1 (exclusive) and B1 (non-exclusive), change B1 cannot affect A1's exclusivity.
for example. Assume a machine has 4 CPUs (0-3).
root cgroup / \ A1 B1
Case 1: Table 1.1: Before applying the patch Step | A1's prstate | B1'sprstate | #1> echo "0-1" > A1/cpuset.cpus | member | member | #2> echo "root" > A1/cpuset.cpus.partition | root | member | #3> echo "0" > B1/cpuset.cpus | root invalid | member |
After step #3, A1 changes from "root" to "root invalid" because its CPUs (0-1) overlap with those requested by B1 (0-3). However, B1 can actually use CPUs 2-3(from B1's parent), so it would be more reasonable for A1 to remain as "root."
Table 1.2: After applying the patch Step | A1's prstate | B1'sprstate | #1> echo "0-1" > A1/cpuset.cpus | member | member | #2> echo "root" > A1/cpuset.cpus.partition | root | member | #3> echo "0" > B1/cpuset.cpus | root | member |
OK, this looks fine to me, based on this statement from the docs about cpuset.cpus.effective:
subset of "cpuset.cpus" unless none of the CPUs listed in "cpuset.cpus" can be granted. In this case, it will be treated just like an empty "cpuset.cpus".
I was likely confused by the eventual switch of B1 to root in your previous example. (Because if you continue, it should result in (after patch too): #4> echo "root" > B1/cpuset.partition | root invalid | root invalid | and end state should be invariant wrt A1,B1 or B1,A1 config order.)
This patch is based on a version after v6.18.0-rc5. Whether or not this patch is applied, modifications to cpuset.partition do not affect the state of sibling partitions.
If continue, the result should be as follows: #4> echo "root" > B1/cpuset.partition | root | root invalid |
I've updated patch v5 with some new ideas and look forward to your feedback.
patch v5 : https://lore.kernel.org/cgroups/20251119105749.1385946-1-sunshaojie@kylinos....
Thanks, Sun Shaojie