nbd: only set MSG_MORE when we have more to send

A user noticed that write performance was horrible over loopback and we
traced it to an inversion of when we need to set MSG_MORE.  It should be
set when we have more bvec's to send, not when we are on the last bvec.
This patch made the test go from 20 iops to 78k iops.

Signed-off-by: Josef Bacik <jbacik@fb.com>
Fixes: 429a787be679 ("nbd: fix use-after-free of rq/bio in the xmit path")
Signed-off-by: Jens Axboe <axboe@fb.com>

diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
index 50a2020b5b724fd657c01b6d1a27c33a68a27fd0..9fd06eeb1a17b3880e2ad574f4ca081903a67961 100644 (file)
--- a/drivers/block/nbd.c
+++ b/drivers/block/nbd.c
@@ -271,7 +271,7 @@ static inline int sock_send_bvec(struct nbd_device *nbd, int index,
 static int nbd_send_cmd(struct nbd_device *nbd, struct nbd_cmd *cmd, int index)
 {
        struct request *req = blk_mq_rq_from_pdu(cmd);
-       int result, flags;
+       int result;
        struct nbd_request request;
        unsigned long size = blk_rq_bytes(req);
        struct bio *bio;
@@ -310,7 +310,6 @@ static int nbd_send_cmd(struct nbd_device *nbd, struct nbd_cmd *cmd, int index)
        if (type != NBD_CMD_WRITE)
                return 0;
 
-       flags = 0;
        bio = req->bio;
        while (bio) {
                struct bio *next = bio->bi_next;
@@ -319,9 +318,8 @@ static int nbd_send_cmd(struct nbd_device *nbd, struct nbd_cmd *cmd, int index)
 
                bio_for_each_segment(bvec, bio, iter) {
                        bool is_last = !next && bio_iter_last(bvec, iter);
+                       int flags = is_last ? 0 : MSG_MORE;
 
-                       if (is_last)
-                               flags = MSG_MORE;
                        dev_dbg(nbd_to_dev(nbd), "request %p: sending %d bytes data\n",
                                cmd, bvec.bv_len);
                        result = sock_send_bvec(nbd, index, &bvec, flags);