netdev - [PATCH net-next] net/tls: Corrected enabling of zero-copy mode

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20180723153006.24601-1-vakul.garg@nxp.com>
Date:   Mon, 23 Jul 2018 21:00:06 +0530
From:   Vakul Garg <vakul.garg@....com>
To:     netdev@...r.kernel.org
Cc:     borisp@...lanox.com, aviadye@...lanox.com, davejwatson@...com,
        davem@...emloft.net, Vakul Garg <vakul.garg@....com>
Subject: [PATCH net-next] net/tls: Corrected enabling of zero-copy mode

Zero-copy mode was left enabled even when zerocopy_from_iter() failed.
In this case, control falls back to regular non zero-copy mode. This
caused skipping call of datagram copy function skb_copy_datagram_msg()
and required an extra iteration of record decryption loop for decrypted
record to be copied into user space provided buffer. Hence zero-copy
mode should be enabled/disabled as per the success/failure of
zerocopy_from_iter().

Fixes: c46234ebb4d1 ("tls: RX path for ktls")
Signed-off-by: Vakul Garg <vakul.garg@....com>
---

The patch does not need to be applied to 'net' branch as it does not fix
any functional bug. The patch achieves better run-time efficiency and
code readability.

 net/tls/tls_sw.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c
index 0c2d029c9d4c..9ae57bec4927 100644
--- a/net/tls/tls_sw.c
+++ b/net/tls/tls_sw.c
@@ -787,7 +787,7 @@ int tls_sw_recvmsg(struct sock *sk,
 	target = sock_rcvlowat(sk, flags & MSG_WAITALL, len);
 	timeo = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
 	do {
-		bool zc = false;
+		bool zc;
 		int chunk = 0;
 
 		skb = tls_wait_data(sk, flags, timeo, &err);
@@ -824,7 +824,6 @@ int tls_sw_recvmsg(struct sock *sk,
 				struct scatterlist sgin[MAX_SKB_FRAGS + 1];
 				int pages = 0;
 
-				zc = true;
 				sg_init_table(sgin, MAX_SKB_FRAGS + 1);
 				sg_set_buf(&sgin[0], ctx->rx_aad_plaintext,
 					   TLS_AAD_SPACE_SIZE);
@@ -836,6 +835,7 @@ int tls_sw_recvmsg(struct sock *sk,
 				if (err < 0)
 					goto fallback_to_reg_recv;
 
+				zc = true;
 				err = decrypt_skb_update(sk, skb, sgin, &zc);
 				for (; pages > 0; pages--)
 					put_page(sg_page(&sgin[pages]));
@@ -845,6 +845,7 @@ int tls_sw_recvmsg(struct sock *sk,
 				}
 			} else {
 fallback_to_reg_recv:
+				zc = false;
 				err = decrypt_skb_update(sk, skb, NULL, &zc);
 				if (err < 0) {
 					tls_err_abort(sk, EBADMSG);
-- 
2.13.6