This way systemd in container on cgroup-v2 can enable the controllers it
requires.
https://jira.vzint.dev/browse/PSBM-149975
Signed-off-by: Pavel Tikhomirov
---
kernel/cgroup/cgroup.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgrou
In cgroup-v2 memsw is not used anymore, instead we have swap and memory
separated from one another. So memory limit should be read from
memory.max not from memsw.max.
https://jira.vzint.dev/browse/PSBM-149975
Fixes: 9417436f33d1 ("ve/mm: add heuristic check for memory overcommit")
Signed-off-by: P
In vfs_get_tree we explicitly check that fc->root is NULL before calling
->get_tree helper. So when mounting cgroup2 filesystem in the begining
of cgroup_get_tree the fc->root is uninitializled. We were lucky that
ve_hide_cgroups never dereferenced it on this code path, as mounting
cgroup2 from con
The change in is_virtualized_cgroot makes cgroup-v2 "virtualized" only
when it has enabled subsystems in it. And "virtualized" means that it
will be:
a) shown in different cgroup related proc files in container,
b) allowed to be mounted in container and
c) required to have separate root directory f
To test it need to boot with systemd.unified_cgroup_hierarchy=1 option.
https://jira.vzint.dev/browse/PSBM-149975
Signed-off-by: Pavel Tikhomirov
Pavel Tikhomirov (4):
cgroup: fix uninitiallized use of ctx->root
mm/memcontrol: fix mem_cgroup_enough_memory on cgroup-v2
cgroup: allow to writ
Lockdep reports circular lock dependency here
&p->lock --> &plo->ctl_mutex --> sb_writers.
this is only triggered by trinity and it did not happen
in real world for the last 10 years. fixing it at this stage is not feasible
since it could break so many things.
So take the safe route and just disabl
Before this patch .get_stats() for venetdev was not defended from
racing, Documentation/networking/netdevices.txt states ndo_get_stats()
is to be syncronized by dev_base_lock rwlock, but we do READ stats, so
can go in parallel.
At the same time the algorithm of gathering per-cpu stats into a singl
On 28.08.2023 23:54, Konstantin Khorenko wrote:
Before this patch .get_stats() for venetdev was not defended from
racing, Documentation/networking/netdevices.txt states ndo_get_stats()
is to be syncronized by dev_base_lock rwlock, but we do READ stats, so
can go in parallel.
At the same time
