From: Herbert Xu <herbert@gondor.apana.org.au>
To: Eric Dumazet <eric.dumazet@gmail.com>
Cc: syzbot <syzbot+c2f1558d49e25cc36e5e@syzkaller.appspotmail.com>,
linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
syzkaller-bugs@googlegroups.com, viro@zeniv.linux.org.uk,
"David S. Miller" <davem@davemloft.net>,
Linux Crypto Mailing List <linux-crypto@vger.kernel.org>
Subject: [PATCH] crypto: af_alg - Use bh_lock_sock in sk_destruct
Date: Thu, 5 Dec 2019 13:45:05 +0800 [thread overview]
Message-ID: <20191205054505.wulhkajz64lwwffc@gondor.apana.org.au> (raw)
In-Reply-To: <f7009e8d-a488-c6df-6875-e872265efec0@gmail.com>
On Wed, Dec 04, 2019 at 08:59:11PM -0800, Eric Dumazet wrote:
>
> crypto layer (hash_sock_destruct()) is called from rcu callback (this in BH context) but tries to grab a socket lock.
>
> A socket lock can schedule, which is illegal in BH context.
Fair enough. Although I was rather intrigued as to how the RCU call
occured in the first place. After some digging my theory is that
this is due to a SO_ATTACH_REUSEPORT_CBPF or SO_ATTACH_REUSEPORT_EBPF
setsockopt on the crypto socket.
What are these filters even suppposed to do on an af_alg socket?
Anyhow, this is a bug that could have been triggered even without
this, but it would have been almost impossible to do it through
syzbot as you need to have an outstanding async skcipher/aead request
that is freed in BH context.
---8<---
As af_alg_release_parent may be called from BH context (most notably
due to an async request that only completes after socket closure,
or as reported here because of an RCU-delayed sk_destruct call), we
must use bh_lock_sock instead of lock_sock.
Reported-by: syzbot+c2f1558d49e25cc36e5e@syzkaller.appspotmail.com
Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
Fixes: c840ac6af3f8 ("crypto: af_alg - Disallow bind/setkey/...")
Cc: <stable@vger.kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
diff --git a/crypto/af_alg.c b/crypto/af_alg.c
index 0dceaabc6321..3d8e53010cda 100644
--- a/crypto/af_alg.c
+++ b/crypto/af_alg.c
@@ -134,11 +134,13 @@ void af_alg_release_parent(struct sock *sk)
sk = ask->parent;
ask = alg_sk(sk);
- lock_sock(sk);
+ local_bh_disable();
+ bh_lock_sock(sk);
ask->nokey_refcnt -= nokey;
if (!last)
last = !--ask->refcnt;
- release_sock(sk);
+ bh_unlock_sock(sk);
+ local_bh_enable();
if (last)
sock_put(sk);
--
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
next prev parent reply other threads:[~2019-12-05 5:45 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-12-05 4:35 BUG: sleeping function called from invalid context in lock_sock_nested syzbot
2019-12-05 4:59 ` Eric Dumazet
2019-12-05 5:45 ` Herbert Xu [this message]
2020-01-17 9:39 ` syzbot
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20191205054505.wulhkajz64lwwffc@gondor.apana.org.au \
--to=herbert@gondor.apana.org.au \
--cc=davem@davemloft.net \
--cc=eric.dumazet@gmail.com \
--cc=linux-crypto@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=syzbot+c2f1558d49e25cc36e5e@syzkaller.appspotmail.com \
--cc=syzkaller-bugs@googlegroups.com \
--cc=viro@zeniv.linux.org.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).