CVE-2026-98414

In the Linux kernel, the following vulnerability has been resolved:

memcg: avoid charging the root memcg from obj cgroup charge pages()

obj cgroup charge pages() resolves the objcg to its memcg and calls try charge memcg(), which does not short circuit the root memcg. That memcg can be the root memcg: obj cgroup is root() reflects the memcg the objcg was created for and is never updated, while memcg reparent objcgs() does redirect objcg->memcg to the parent on rmdir. An objcg of a dying child of root therefore passes every obj cgroup is root() filter but resolves to the root memcg.

Folios keep the objcg they were charged with, so this is easy to reach through zswap: allocate anon memory in a cgroup, move the task out, remove the cgroup, then write to the root cgroup's memory.reclaim. The reclaimed folios are charged through the reparented objcg and end up in refill stock() with the root memcg:

WARNING: mm/memcontrol.c:2198 at refill stock+0x644/0x940 refill stock+0x644/0x940 try charge memcg+0x12d6/0x1570 obj cgroup charge+0x35/0xf0 obj cgroup charge+0x1de/0x210 obj cgroup charge zswap+0x83/0x270 zswap store+0x1620/0x2000 swap writeout+0x94c/0x14c0 shrink folio list+0x3388/0x52b0 [...] try to free mem cgroup pages+0x30d/0x830 user proactive reclaim+0x504/0x840 memory reclaim+0x1f/0x30

Beyond the warning, the charge is asymmetric: obj cgroup uncharge pages() skips refill stock() for the root memcg, so the root's page counter grows and is never uncharged. It is not user visible, since memory.current is not exposed on the root, but it is a leak.

Use try charge(), which returns early for the root memcg, restoring the symmetry with obj cgroup uncharge pages().

The above sequence was scripted into a standalone reproducer (zswap on, swap on a virtio disk, 512MB of anon memory faulted in inside a child of the root cgroup, the task then migrated to the root cgroup, the child removed, followed by "echo 600M swappiness=max > memory.reclaim" on the root) and run in a CONFIG DEBUG VM=y VM. It reproduces the splat on the first zswap store of a reparented folio, with the same call chain as the report. With this patch applied the splat is gone while the zswap store count over the run is unchanged, so the same path is still exercised. cgroup selftests test zswap, test kmem and test memcontrol show no new failures.