PT-2026-152518
CVE-2026-99638
·
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:
net/sched: clamp skb->queue mapping before indexing the tx queue
This issue was caught by Sashiko (nipa). A producer-side path cap is
already merged (ea4d4b5dddb5); this patch adds the matching consumer-side
caps.
The issue:
The tx queue is picked and capped in netdev core pick tx() before the
dev xmit skb() qdisc enqueue path. A tc BPF program running inside an
egress qdisc (not clsact) can then change sk buff->queue mapping; that
store is never re-checked. bpf convert ctx access() bounds it only by
NO QUEUE MAPPING, not by the device's queue count.
Two consumers read the stale value:
-
On dequeue, qdisc restart() calls netdev get tx queue(), which indexes dev-> tx[]. netdev get tx queue() guards it with just a DEBUG NET WARN ON ONCE, so sch direct xmit() takes the out-of-bounds txq-> xmit lock and writes txq->xmit lock owner past the end of the allocation on non-lltx devices such as ifb. Drivers that index their rings from skb get queue mapping() (ex: ifb) go out of range the same way.
-
On enqueue, taprio enqueue() indexes q->qdiscs[] with the same value; the array is sized dev->num tx queues. A duplicating netem on a child qdisc clones the skb and re-enqueues the clone at the taprio root (qdisc root bh()) through a direct rootq->enqueue() call, so the value reaches taprio enqueue() before any dequeue cap runs.
The Fix:
Cap the mapping at the consumers and write the capped value back, so the
qdisc's tx[] lookup, the taprio q->qdiscs[] lookup and any later driver
read stay in range.
On the dequeue side the cap folds into the dequeue path instead of a
separate walk over the returned list: dequeue skb() caps the head, and
each bulk helper caps the element it appends. For a qdisc that is not
TCQ F ONETXQUEUE, netdev cap txqueue() logs a ratelimited notice and
selects queue 0.
A TCQ F ONETXQUEUE qdisc is bound to a single tx queue: dequeue skb() uses
q->dev queue for the frozen check, the BQL limit and the trace point, while
qdisc restart() transmits on the skb's queue. Capping an out-of-range
mapping to queue 0 would therefore send a per-queue child's skb through a
different queue than the one it was dequeued for. On mq, mqprio, taprio and
htb offload children, a stopped queue 0 could then park the list in
q->gso skb, and the later wake would schedule queue 0's own child instead
of this qdisc. Pin a TCQ F ONETXQUEUE mapping to the qdisc's own queue to
keep the documented invariant.
The pin covers every mapping, not only an out-of-range one: an in-range
mapping that a filter sets on one of those per-queue children is rewritten
to the child's own queue, and this path does not call netdev cap txqueue(),
so it logs no notice. On a single-queue device qdisc create() sets
TCQ F ONETXQUEUE, so the one-queue trigger described below also takes this
branch and is corrected without a notice.
That also keeps a bulk list on one queue by construction: every element of
a TCQ F ONETXQUEUE qdisc, including a follower changed by a filter, is
pinned to q->dev queue before it is appended. The fast bulk helper
therefore needs no head comparison and no skb bad txq requeue; the slow
helper, used only by non-ONETXQUEUE qdiscs, still compares and requeues a
mismatched follower, where a wake on any queue reschedules q.
On the enqueue side taprio enqueue() caps the mapping against
real num tx queues and writes it back before selecting q->qdiscs[queue], so
a duplicating child cannot steer the clone out of the array.
Conditions to recreate the bug:
The dequeue path: load a SCHED CLS or SCHED ACT program that stores
sk buff->queue mapping (a tc BPF action is enough), attach it to a
transmit qdisc (for example a matchall filter on a prio root), and send a
packet. With a 1-queue device a store of 1 indexes one past dev-> tx[].
Under KASAN with panic on warn=1 the unfixed kernel faults with "BUG:
KASAN: slab-out-
---truncated---
وجدت خطأً في الوصف؟ لديك ما تضيفه؟ لا تتردد في مراسلتنا 👾