netdev - RE: [PATCH net-next v5 1/3] net: ipv6: Unduplicate {raw,udp}v6

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]

Date:	Fri, 25 Apr 2014 11:24:21 +0000
From:	David Laight <David.Laight@...LAB.COM>
To:	'Lorenzo Colitti' <lorenzo@...gle.com>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>
CC:	"yoshfuji@...ux-ipv6.org" <yoshfuji@...ux-ipv6.org>,
	"hannes@...essinduktion.org" <hannes@...essinduktion.org>,
	"davem@...emloft.net" <davem@...emloft.net>,
	"eric.dumazet@...il.com" <eric.dumazet@...il.com>
Subject: RE: [PATCH net-next v5 1/3] net: ipv6: Unduplicate
 {raw,udp}v6_sendmsg code

From: Lorenzo Colitti
> rawv6_sendmsg and udpv6_sendmsg have ~100 lines of almost
> identical code. Move this into a new ipv6_datagram_send_common
> helper function.
> 
> Tested: black-box tested using user-mode Linux.
> 
> - Basic UDP sends using sendto work.
> - Mark routing and oif routing using SO_BINDTODEVICE work.
...
> diff --git a/include/net/ipv6.h b/include/net/ipv6.h
> index d640925..f1a247a 100644
> --- a/include/net/ipv6.h
> +++ b/include/net/ipv6.h
> @@ -785,6 +785,13 @@ int compat_ipv6_getsockopt(struct sock *sk, int level, int optname,
>  int ip6_datagram_connect(struct sock *sk, struct sockaddr *addr, int addr_len);
>  int ip6_datagram_connect_v6_only(struct sock *sk, struct sockaddr *addr,
>  				 int addr_len);
> +int ip6_datagram_send_common(struct sock *sk, struct msghdr *msg,
> +			     struct sockaddr_in6 *sin6, int addr_len,
> +			     struct flowi6 *fl6, struct dst_entry **dstp,
> +			     struct ipv6_txoptions **optp,
> +			     struct ipv6_txoptions *opt_space,
> +			     int *hlimit, int *tclass, int *dontfrag,
> +			     int *connected);

That is a lot of parameters! and udp send is probably an important path.

You also repeatedly dereference some of them (eg hlimit).
I think that all of hlimit, tclass, dontfrag and connected could be shadowed
by locals and only written out on success (maybe into a structure?).
It might also be possible to reduce the number of arguments by using a sequence of calls.
(getting the args 'passed by register' helps performance.)

	David



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html