In the Linux kernel, the following vulnerability has been resolved:
pfifotailenqueue: Drop new packet when sch->limit == 0
Expected behaviour:
In case we reach scheduler's limit, pfifotailenqueue() will drop a
packet in scheduler's queue and decrease scheduler's qlen by one.
Then, pfifotailenqueue() enqueue new packet and increase
scheduler's qlen by one. Finally, pfifotailenqueue() return
NET_XMIT_CN
status code.
Weird behaviour:
In case we set sch->limit == 0
and trigger pfifotailenqueue() on a
scheduler that has no packet, the 'drop a packet' step will do nothing.
This means the scheduler's qlen still has value equal 0.
Then, we continue to enqueue new packet and increase scheduler's qlen by
one. In summary, we can leverage pfifotailenqueue() to increase qlen by
one and return NET_XMIT_CN
status code.
The problem is:
Let's say we have two qdiscs: QdiscA and QdiscB.
- QdiscA's type must have '->graft()' function to create parent/child relationship.
Let's say QdiscA's type is hfsc
. Enqueue packet to this qdisc will trigger hfsc_enqueue
.
- QdiscB's type is pfifoheaddrop. Enqueue packet to this qdisc will trigger pfifo_tail_enqueue
.
- QdiscB is configured to have sch->limit == 0
.
- QdiscA is configured to route the enqueued's packet to QdiscB.
Enqueue packet through QdiscA will lead to:
- hfscenqueue(QdiscA) -> pfifotailenqueue(QdiscB)
- QdiscB->q.qlen += 1
- pfifotailenqueue() return NET_XMIT_CN
- hfscenqueue() check for NET_XMIT_SUCCESS
and see NET_XMIT_CN
=> hfscenqueue() don't increase qlen of QdiscA.
The whole process lead to a situation where QdiscA->q.qlen == 0 and QdiscB->q.qlen == 1. Replace 'hfsc' with other type (for example: 'drr') still lead to the same problem. This violate the design where parent's qlen should equal to the sum of its childrens'qlen.
Bug impact: This issue can be used for user->kernel privilege escalation when it is reachable.