[<prev] [next>] [day] [month] [year] [list]
Message-ID: <9ca5205a-ee1c-0ad6-3e9d-3287e92714a1@canonical.com>
Date: Mon, 12 Oct 2020 16:32:20 +0100
From: Colin Ian King <colin.king@...onical.com>
To: Bob Pearson <rpearsonhpe@...il.com>
Cc: Zhu Yanjun <yanjunz@...dia.com>,
"Doug Ledford <dledford"@redhat.com,
Jason Gunthorpe <jgg@...pe.ca>, linux-rdma@...r.kernel.org,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: re: RDMA/rxe: Fix skb lifetime in rxe_rcv_mcast_pkt()
Hi,
Static analysis with Coverity has detected a potential issue with the
following commit:
commit e7ec96fc7932f48a6d6cdd05bf82004a1a04285b
Author: Bob Pearson <rpearsonhpe@...il.com>
Date: Thu Oct 8 15:36:52 2020 -0500
RDMA/rxe: Fix skb lifetime in rxe_rcv_mcast_pkt()
The analysis is as follows:
16. Condition mce->qp_list.next != &mcg->qp_list, taking true branch.
269 if (mce->qp_list.next != &mcg->qp_list)
17. returned_null: skb_clone returns NULL (checked 153 out of 178 times).
18. var_assigned: Assigning: per_qp_skb = NULL return value from
skb_clone.
19. Falling through to end of if statement.
270 per_qp_skb = skb_clone(skb, GFP_ATOMIC);
271 else
272 per_qp_skb = skb;
273
274 per_qp_pkt = SKB_TO_PKT(per_qp_skb);
275 per_qp_pkt->qp = qp;
276 rxe_add_ref(qp);
Dereference null return value (NULL_RETURNS)
20. dereference: Dereferencing a pointer that might be NULL
per_qp_skb when calling rxe_rcv_pkt.
277 rxe_rcv_pkt(per_qp_pkt, per_qp_skb);
278 }
279
It is possible for skb_clone to return NULL, so dereferencing the NULL
return pointer per_qp_skb is a potential issue.
Colin
Powered by blists - more mailing lists