Commit 101260f5 authored by Bob Pearson's avatar Bob Pearson Committed by Liu Jian
Browse files

RDMA/rxe: Fix seg fault in rxe_comp_queue_pkt

mainline inclusion
from mainline-v6.10-rc1
commit 2b23b6097303ed0ba5f4bc036a1c07b6027af5c6
category: bugfix
bugzilla: https://gitee.com/src-openeuler/kernel/issues/IA72Y8
CVE: CVE-2024-38544

Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=2b23b6097303ed0ba5f4bc036a1c07b6027af5c6

---------------------------

In rxe_comp_queue_pkt() an incoming response packet skb is enqueued to the
resp_pkts queue and then a decision is made whether to run the completer
task inline or schedule it. Finally the skb is dereferenced to bump a 'hw'
performance counter. This is wrong because if the completer task is
already running in a separate thread it may have already processed the skb
and freed it which can cause a seg fault.  This has been observed
infrequently in testing at high scale.

This patch fixes this by changing the order of enqueuing the packet until
after the counter is accessed.

Link: https://lore.kernel.org/r/20240329145513.35381-4-rpearsonhpe@gmail.com


Signed-off-by: default avatarBob Pearson <rpearsonhpe@gmail.com>
Fixes: 0b1e5b99 ("IB/rxe: Add port protocol stats")
Signed-off-by: default avatarJason Gunthorpe <jgg@nvidia.com>

Conflicts:
	drivers/infiniband/sw/rxe/rxe_comp.c
[Did not backport dccb23f6.]
Signed-off-by: default avatarLiu Jian <liujian56@huawei.com>
parent 5b8dd9be
Loading
Loading
Loading
Loading
+3 −3
Original line number Diff line number Diff line
@@ -123,12 +123,12 @@ void rxe_comp_queue_pkt(struct rxe_qp *qp, struct sk_buff *skb)
{
	int must_sched;

	skb_queue_tail(&qp->resp_pkts, skb);

	must_sched = skb_queue_len(&qp->resp_pkts) > 1;
	must_sched = skb_queue_len(&qp->resp_pkts) > 0;
	if (must_sched != 0)
		rxe_counter_inc(SKB_TO_PKT(skb)->rxe, RXE_CNT_COMPLETER_SCHED);

	skb_queue_tail(&qp->resp_pkts, skb);

	rxe_run_task(&qp->comp.task, must_sched);
}