CVE-2026-64600

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> xfs: resample the data fork mapping after cycling ILOCK<br /> <br /> xfs_reflink_fill_{cow_hole,delalloc} are both presented with an inode,<br /> a data fork mapping, and a cow fork mapping. Unfortunately, these two<br /> helpers cycle the ILOCK to grab a transaction, which means that the<br /> mappings are stale as soon as we reacquire the ILOCK. Currently we<br /> refresh the cow fork mapping by re-calling xfs_find_trim_cow_extent, but<br /> we don&amp;#39;t refresh the data fork mapping beforehand, which means that the<br /> xfs_bmap_trim_cow in that function queries the refcount btree about the<br /> wrong physical blocks and returns an inaccurate value in *shared.<br /> <br /> If *shared is now false, the directio write proceeds with a stale data<br /> fork mapping. Fix this by querying the data fork mapping if the<br /> sequence counter changes across the ILOCK cycle.