netdev - [PATCH net 4/5] tls: fix race condition in async decryption of corrupted records

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

Message-Id: <e094325019f7fd960470c10efda41c1b7f9bc54f.1694018970.git.sd@queasysnail.net>
Date: Wed,  6 Sep 2023 19:08:34 +0200
From: Sabrina Dubroca <sd@...asysnail.net>
To: netdev@...r.kernel.org
Cc: Sabrina Dubroca <sd@...asysnail.net>,
	Dave Watson <davejwatson@...com>,
	Jakub Kicinski <kuba@...nel.org>,
	Vakul Garg <vakul.garg@....com>,
	Boris Pismenny <borisp@...dia.com>,
	John Fastabend <john.fastabend@...il.com>
Subject: [PATCH net 4/5] tls: fix race condition in async decryption of corrupted records

In case a corrupted record is decrypted asynchronously, the error can
be reported in two ways:
 - via the completion's error mechanism
 - via sk->sk_err

If all asynchronous decrypts finish before we reach the end of
tls_sw_recvmsg, decrypt_pending will be 0 and we don't check the
completion for errors. We should still check sk->sk_err, otherwise
we'll miss the error and return garbage to userspace. This is visible
in the bad_auth test case.

Fixes: 94524d8fc965 ("net/tls: Add support for async decryption of tls records")
Signed-off-by: Sabrina Dubroca <sd@...asysnail.net>
---
 net/tls/tls_sw.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c
index babbd43d41ed..f80a2ea1dd7e 100644
--- a/net/tls/tls_sw.c
+++ b/net/tls/tls_sw.c
@@ -2116,6 +2116,9 @@ int tls_sw_recvmsg(struct sock *sk,
 		ret = 0;
 		if (pending)
 			ret = crypto_wait_req(-EINPROGRESS, &ctx->async_wait);
+		/* Crypto completion could have run before us, check sk_err */
+		if (ret == 0)
+			ret = -sk->sk_err;
 		__skb_queue_purge(&ctx->async_hold);
 
 		if (ret) {
-- 
2.40.1