CVE-2026-89771

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

ring-buffer: Fix subbuf resize race with ring buffer readers

trace buffer subbuf size is read lockless in ring buffer read page() and ring buffer read start(), while it can simultaneously be resized with ring buffer subbuf order set().

Instead of trace buffer::subbuf size, use bpage::order in ring buffer read start() and ring buffer read page().

In ring buffer read start(), even with resize disabled, there is still a possibility of a race with a buffer modification. Hold the trace buffer mutex to synchronise with any pending ring buffer order modification.

trace buffer::subbuf size is now actually useless, remove it. Also, create accessors rb subbuf capacity() and rb page capacity() which return the actual size available for storing events, while rb subbuf size() returns the actual subbuf page-size.