PT-2026-152476
CVE-2026-99596
·
تاريخ النشر
2026-11-20
·
تاريخ التحديث
2026-11-20
الوصف من
الدرجة
N/A
غير متوفرلا تتوفر حاليًا تقييمات أو مقاييس للخطورة. سنحدّث المعلومات المقابلة في الصفحة فور توفرها.
In the Linux kernel, the following vulnerability has been resolved:
nvme: wait for SRCU readers before splicing the namespace list
Readers walk ctrl->namespaces under ctrl->srcu, and some of them sleep
there, e.g. nvme wait freeze() waiting for I/O to drain.
nvme remove namespaces() moves the list onto its on-stack ns list with
list splice init rcu(), which only waits for RCU readers before
relinking the entries. A reader still on the last namespace then follows
its next pointer to ns list instead of ctrl->namespaces and treats the
stack as a namespace. Deleting a controller while a Format NVM
passthrough waits for a delayed read on the last namespace gives:
BUG: KASAN: stack-out-of-bounds in nvme wait freeze+0x102/0x120
Read of size 8 at addr ff110001d2f3fc38 by task nvme/4413
Call Trace:
nvme wait freeze
nvme passthru start
nvme submit user cmd
nvme user cmd
nvme dev ioctl
BUG: unable to handle page fault for address: ffffffff8af04850
Use list splice init srcu() instead, which waits for the SRCU readers.
This also takes the RCU grace period out of every controller deletion.
Deleting 64 tcp controllers with one namespace each, median of 3 runs on
each of two boots per kernel:
before afternvme disconnect-all (ms) 3491 / 3509 2996 / 3165
nvme remove namespaces() avg (ms) 39.0 / 40.5 31.9 / 34.7
وجدت خطأً في الوصف؟ لديك ما تضيفه؟ لا تتردد في مراسلتنا 👾