CVE-2026-93268

*** Pendiente de traducción *** In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> ext4: skip extra isize expansion during mount to prevent deadlock<br /> <br /> ext4_try_to_expand_extra_isize() is called from __ext4_mark_inode_dirty()<br /> while holding an active jbd2 handle. During mount (!SB_ACTIVE), the<br /> expand path may move xattrs to external blocks and release ea_inodes via<br /> iput(). When !SB_ACTIVE, iput() calls write_inode_now() which acquires<br /> s_writepages_rwsem, creating a circular lock dependency:<br /> <br /> s_writepages_rwsem --&gt; jbd2_handle --&gt; xattr_sem --&gt; s_writepages_rwsem<br /> <br /> This can be triggered via:<br /> <br /> ext4_process_orphan() -&gt; ext4_truncate() -&gt; ext4_mark_inode_dirty()<br /> -&gt; ext4_try_to_expand_extra_isize()<br /> <br /> or:<br /> <br /> ext4_evict_inode() -&gt; ext4_mark_inode_dirty()<br /> -&gt; ext4_try_to_expand_extra_isize()<br /> <br /> Skip expansion when !SB_ACTIVE. This is a minor loss of functionality<br /> (extra isize won&amp;#39;t grow for these inodes during mount), which e2fsck<br /> can resolve later if needed.