[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180712150343.GA19138@macbook-pro-6.local.dhcp.thefacebook.com>
Date: Thu, 12 Jul 2018 08:03:43 -0700
From: Dave Watson <davejwatson@...com>
To: Boris Pismenny <borisp@...lanox.com>,
"davem@...emloft.net" <davem@...emloft.net>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
"aviadye@...lanox.com" <aviadye@...lanox.com>,
Daniel Borkmann <daniel@...earbox.net>, <doronrk@...com>,
Vakul Garg <vakul.garg@....com>
Subject: [PATCH net] tls: Stricter error checking in zerocopy sendmsg path
In the zerocopy sendmsg() path, there are error checks to revert
the zerocopy if we get any error code. syzkaller has discovered
that tls_push_record can return -ECONNRESET, which is fatal, and
happens after the point at which it is safe to revert the iter,
as we've already passed the memory to do_tcp_sendpages.
Previously this code could return -ENOMEM and we would want to
revert the iter, but AFAIK this no longer returns ENOMEM after
a447da7d004 ("tls: fix waitall behavior in tls_sw_recvmsg"),
so we fail for all error codes.
Reported-by: syzbot+c226690f7b3126c5ee04@...kaller.appspotmail.com
Reported-by: syzbot+709f2810a6a05f11d4d3@...kaller.appspotmail.com
Signed-off-by: Dave Watson <davejwatson@...com>
Fixes: 3c4d7559159b ("tls: kernel TLS support")
---
net/tls/tls_sw.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c
index 7818011fd250..4618f1c31137 100644
--- a/net/tls/tls_sw.c
+++ b/net/tls/tls_sw.c
@@ -440,7 +440,7 @@ int tls_sw_sendmsg(struct sock *sk, struct msghdr *msg, size_t size)
ret = tls_push_record(sk, msg->msg_flags, record_type);
if (!ret)
continue;
- if (ret == -EAGAIN)
+ if (ret < 0)
goto send_end;
copied -= try_to_copy;
--
2.17.1
Powered by blists - more mailing lists