On Fri, Jan 30, 2026 at 9:33 PM [off-list ref] wrote:

quoted

diff --git a/kernel/bpf/bpf_local_storage.c b/kernel/bpf/bpf_local_storage.c
index 824d05325..7661319ad 100644
--- a/kernel/bpf/bpf_local_storage.c
+++ b/kernel/bpf/bpf_local_storage.c

[ ... ]

quoted

@@ -572,7 +591,9 @@ bpf_local_storage_update(void *owner, struct bpf_local_storage_map *smap,
      if (!alloc_selem)
              return ERR_PTR(-ENOMEM);

-     raw_spin_lock_irqsave(&local_storage->lock, flags);
+     err = raw_res_spin_lock_irqsave(&local_storage->lock, flags);
+     if (err)
+             return ERR_PTR(err);

If the lock acquisition fails here, doesn't this leak alloc_selem? The
allocation at the top of this function succeeded:

    alloc_selem = selem = bpf_selem_alloc(smap, owner, value, ...);
    if (!alloc_selem)
        return ERR_PTR(-ENOMEM);

But on lock failure, the early return bypasses the cleanup at the end of
the function:

    if (alloc_selem) {
        mem_uncharge(smap, owner, smap->elem_size);
        bpf_selem_free(alloc_selem, true);
    }

Should this goto unlock instead of returning directly, or add explicit
cleanup before returning?

Right. It should be "goto unlock".

[ ... ]

This isn't a bug, but the commit message has a minor wording issue:

quoted

Change bpf_local_storage::lock and bpf_local_storage_map_bucket::lock to
from raw_spin_lock to rqspinlock.

"to from" should be "from".

quoted

For, __bpf_local_storage_map_cache(), instead of handling the error,
skip updating the cache.

There's an extra comma after "For".

Ack.

---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

CI run summary: https://github.com/kernel-patches/bpf/actions/runs/21539340647

AI-authorship-score: low
AI-authorship-explanation: The commit follows standard kernel development patterns with appropriate technical detail and no AI-typical phrasing or structure.
issues-found: 1
issue-severity-score: low
issue-severity-explanation: Memory leak on error path in bpf_local_storage_update when rqspinlock times out; affects only rare lock failure scenarios.