{"id":21654,"date":"2026-09-12T04:31:28","date_gmt":"2026-09-11T20:31:28","guid":{"rendered":"https:\/\/www.54iter.cn\/?vuln=cve-2026-89732"},"modified":"2026-09-12T04:31:28","modified_gmt":"2026-09-11T20:31:28","slug":"cve-2026-89732","status":"publish","type":"vuln","link":"https:\/\/www.54iter.cn\/?vuln=cve-2026-89732","title":{"rendered":"CVE-2026-89732 &#8211; In the Linux kernel, the following vulnerability has been resolved:\n\nusb: gadget: f_fs: Pr"},"content":{"rendered":"<p>In the Linux kernel, the following vulnerability has been resolved:<\/p>\n<p>usb: gadget: f_fs: Prevent deadlock during ep0 read loop<\/p>\n<p>Currently, ffs_ep0_read() holds ffs-&gt;mutex when it prepares to go to<br \/>\nsleep waiting for an event. When no setup events are pending, it calls<br \/>\nwait_event_interruptible_exclusive_locked_irq() with the mutex still<br \/>\nheld. The wait macro deliberately drops the waitqueue spinlock before<br \/>\nsleeping but does not drop the mutex.<\/p>\n<p>If a userspace daemon is polling ep0 via read() and the gadget is<br \/>\nasynchronously torn down via configfs (e.g., echo &#8220;&#8221; &gt; UDC), a<br \/>\ndeadlock can occur:<\/p>\n<p>1. The configfs teardown calls functionfs_unbind(), which queues a<br \/>\n   FUNCTIONFS_UNBIND event.<br \/>\n2. The daemon wakes up, consumes the event, and drops the mutex.<br \/>\n3. However, if the daemon loops and immediately issues another read()<br \/>\n   before exiting, it reacquires ffs-&gt;mutex and again goes into an<br \/>\n   interruptible sleep.<br \/>\n4. Meanwhile, functionfs_unbind() continues execution and attempts to<br \/>\n   acquire ffs-&gt;mutex to tear down ep0req.<br \/>\n5. The kernel deadlocks because the configfs thread is stuck in an<br \/>\n   uninterruptible sleep waiting for the mutex, while the userspace<br \/>\n   daemon is in an interruptible sleep holding the mutex forever<br \/>\n   because no more events will arrive.<\/p>\n<p>To fix this, we drop both the waitqueue spinlock and ffs-&gt;mutex before<br \/>\ngoing to sleep, and use wait_event_interruptible_exclusive() instead.<br \/>\nUpon waking up, we jump back to the `retry` label to safely reacquire<br \/>\nthe mutex and re-evaluate the state machine. By not sleeping with<br \/>\nffs-&gt;mutex held, we natively decouple gadget teardowns (which require<br \/>\nthe mutex) from userspace polling.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>In the Linux kernel, the following vulnerability has been resolved:<\/p>\n<p>usb: gadget: f_fs: Prevent deadlock during ep0 read loop<\/p>\n<p>Currently, ffs_ep0_read() holds ffs-&gt;mutex when it prepares to go to<br \/>\nslee<\/p>\n","protected":false},"template":"","meta":[],"categories":[],"tags":[],"class_list":["post-21654","vuln","type-vuln","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/www.54iter.cn\/index.php?rest_route=\/wp\/v2\/vuln\/21654","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.54iter.cn\/index.php?rest_route=\/wp\/v2\/vuln"}],"about":[{"href":"https:\/\/www.54iter.cn\/index.php?rest_route=\/wp\/v2\/types\/vuln"}],"wp:attachment":[{"href":"https:\/\/www.54iter.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=21654"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.54iter.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=21654"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.54iter.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=21654"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}