[PATCH] nvme-tcp: Fix H2CData PDU send accounting (again)

* [PATCH] nvme-tcp: Fix H2CData PDU send accounting (again)
@ 2021-10-24  7:43 Sagi Grimberg
  2021-10-25  7:18 ` Sagi Grimberg
  2021-10-25 14:38 ` Keith Busch
  0 siblings, 2 replies; 6+ messages in thread
From: Sagi Grimberg @ 2021-10-24  7:43 UTC (permalink / raw)
  To: linux-nvme, Christoph Hellwig, Keith Busch, Chaitanya Kulkarni

We should not access request members after the last send, even to
determine if indeed it was the last data payload send. The reason is
that a completion could have arrived and trigger a new execution of the
request which overridden these members. This was fixed by commit
825619b09ad3 ("nvme-tcp: fix possible use-after-completion").

Commit e371af033c56 broke that assumption again to address cases where
multiple r2t pdus are sent per request. To fix it, we need to record the
request data_sent and data_len and after the payload network send we
reference these counters to determine weather we should advance the
request iterator.

Fixes: e371af033c56 ("nvme-tcp: fix incorrect h2cdata pdu offset accounting")
Reported-by: Keith Busch <kbusch@kernel.org>
Cc: stable@vger.kernel.org # 5.10+ 
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
---
 drivers/nvme/host/tcp.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c
index 78966b8ddb1e..42b58eb1ba62 100644
--- a/drivers/nvme/host/tcp.c
+++ b/drivers/nvme/host/tcp.c
@@ -926,15 +926,17 @@ static void nvme_tcp_fail_request(struct nvme_tcp_request *req)
 static int nvme_tcp_try_send_data(struct nvme_tcp_request *req)
 {
 	struct nvme_tcp_queue *queue = req->queue;
+	int req_data_len = req->data_len;
 
 	while (true) {
 		struct page *page = nvme_tcp_req_cur_page(req);
 		size_t offset = nvme_tcp_req_cur_offset(req);
 		size_t len = nvme_tcp_req_cur_length(req);
-		bool last = nvme_tcp_pdu_last_send(req, len);
+		bool pdu_last = nvme_tcp_pdu_last_send(req, len);
+		int req_data_sent = req->data_sent;
 		int ret, flags = MSG_DONTWAIT;
 
-		if (last && !queue->data_digest && !nvme_tcp_queue_more(queue))
+		if (pdu_last && !queue->data_digest && !nvme_tcp_queue_more(queue))
 			flags |= MSG_EOR;
 		else
 			flags |= MSG_MORE | MSG_SENDPAGE_NOTLAST;
@@ -958,11 +960,11 @@ static int nvme_tcp_try_send_data(struct nvme_tcp_request *req)
 		 * in the request where we don't want to modify it as we may
 		 * compete with the RX path completing the request.
 		 */
-		if (req->data_sent + ret < req->data_len)
+		if (req_data_sent + ret < req_data_len)
 			nvme_tcp_advance_req(req, ret);
 
 		/* fully successful last send in current PDU */
-		if (last && ret == len) {
+		if (pdu_last && ret == len) {
 			if (queue->data_digest) {
 				nvme_tcp_ddgst_final(queue->snd_hash,
 					&req->ddgst);
-- 
2.30.2



^ permalink raw reply related	[flat|nested] 6+ messages in thread