diff options
author | Pavel Begunkov <asml.silence@gmail.com> | 2022-08-26 17:15:47 +0100 |
---|---|---|
committer | Jens Axboe <axboe@kernel.dk> | 2022-08-26 10:31:42 -0600 |
commit | dfb58b1796d19c8405a38eb457f97669440c59d4 (patch) | |
tree | 2eb07c0a8818e7c181983fb952ba6c4934a247ef /io_uring/net.c | |
parent | 581711c46612c1fd7f98960f9ad53f04fdb89853 (diff) | |
download | lwn-dfb58b1796d19c8405a38eb457f97669440c59d4.tar.gz lwn-dfb58b1796d19c8405a38eb457f97669440c59d4.zip |
io_uring/net: fix overexcessive retries
Length parameter of io_sg_from_iter() can be smaller than the iterator's
size, as it's with TCP, so when we set from->count at the end of the
function we truncate the iterator forcing TCP to return preliminary with
a short send. It affects zerocopy sends with large payload sizes and
leads to retries and possible request failures.
Fixes: 3ff1a0d395c00 ("io_uring: enable managed frags with register buffers")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/0bc0d5179c665b4ef5c328377c84c7a1f298467e.1661530037.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Diffstat (limited to 'io_uring/net.c')
-rw-r--r-- | io_uring/net.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/io_uring/net.c b/io_uring/net.c index 0af8a02df580..7a5468cc905e 100644 --- a/io_uring/net.c +++ b/io_uring/net.c @@ -956,7 +956,7 @@ static int io_sg_from_iter(struct sock *sk, struct sk_buff *skb, shinfo->nr_frags = frag; from->bvec += bi.bi_idx; from->nr_segs -= bi.bi_idx; - from->count = bi.bi_size; + from->count -= copied; from->iov_offset = bi.bi_bvec_done; skb->data_len += copied; |