netdev - Re: [PATCH RFC v4 0/5] SCTP: Event skb list overhaul.

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: <20190314012344.GA13514@localhost.localdomain>
Date:   Wed, 13 Mar 2019 22:23:44 -0300
From:   Marcelo Ricardo Leitner <marcelo.leitner@...il.com>
To:     David Miller <davem@...emloft.net>
Cc:     netdev@...r.kernel.org, lucien.xin@...il.com, nhorman@...driver.com
Subject: Re: [PATCH RFC v4 0/5] SCTP: Event skb list overhaul.

On Wed, Mar 13, 2019 at 01:06:21PM -0300, Marcelo Ricardo Leitner wrote:
> On Mon, Mar 11, 2019 at 08:09:19PM -0700, David Miller wrote:
> > 
> > This patch series eliminates the explicit reference to the skb list
> > implementation via skb->prev dereferences.
> > 
> > The approach used is to pass a non-empty skb list around instead of an
> > event skb object which may or may not be on a list.
> > 
> > I'd like to thank Marcelo Leitner, Xin Long, and Neil Horman for
> > reviewing previous versions of this series.
> > 
> > Testing would be very much appreciated, in addition to the review of
> > course.
> > 
> > v3 --> v4: Fix the logic in patch #4 so that we don't miss cases
> >        	   where we should add event to the on-stack temp list.
> > 
> > Signed-off-by: David S. Miller <davem@...emloft.net>
> 
> Code review LGTM too.
> Running some tests, should be done by the end of the day.

Tests were good with the exception of one unrelated issue. ACK from my
side thus.

Xin, would you mind carrying forward this fix? I think we need to fix
the other sockopts too. Thanks!

---8<---

>From ddba48476b343dce84a82036e7914a1f7ac3f0c8 Mon Sep 17 00:00:00 2001
Message-Id: <ddba48476b343dce84a82036e7914a1f7ac3f0c8.1552526507.git.marcelo.leitner@...il.com>
From: Marcelo Ricardo Leitner <marcelo.leitner@...il.com>
Date: Wed, 13 Mar 2019 22:13:52 -0300
Subject: [PATCH] net/sctp: fix ignoring asoc_id for tcp-style sockets on
 setsockopt

Currently if the user pass an invalid asoc_id to SCTP_DEFAULT_SEND_PARAM
on a TCP-style socket, it will silently ignore the new parameters.
That's because after not finding an asoc, it is checking asoc_id against
the known values of CURRENT/FUTURE/ALL values and that fails to match.

IOW, if the user supplies an invalid asoc id or not, it should either
match the current asoc or the socket itself so that it will inherit
these later. Fixes it by forcing asoc_id to SCTP_FUTURE_ASSOC in case it
is a TCP-style socket without an asoc, so that the values get set on the
socket.

Fixes: 707e45b3dc5a ("sctp: use SCTP_FUTURE_ASSOC and add SCTP_CURRENT_ASSOC for SCTP_DEFAULT_SEND_PARAM sockopt")
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@...il.com>
---
 net/sctp/socket.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 533207dbeae955604a6304a8f67b65c9f580fd05..85205a2d9a7dc5c289a94bf4672b240040cbb546 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -3024,6 +3024,9 @@ static int sctp_setsockopt_default_send_param(struct sock *sk,
 		return 0;
 	}

+	if (sctp_style(sk, TCP))
+		info.sinfo_assoc_id = SCTP_FUTURE_ASSOC;
+
 	if (info.sinfo_assoc_id == SCTP_FUTURE_ASSOC ||
 	    info.sinfo_assoc_id == SCTP_ALL_ASSOC) {
 		sp->default_stream = info.sinfo_stream;
-- 
2.20.1