sctp: fix incorrect overflow check on autoclose
authorXi Wang <xi.wang@gmail.com>
Fri, 16 Dec 2011 12:44:15 +0000 (12:44 +0000)
committerRohan Somvanshi <rsomvanshi@nvidia.com>
Wed, 11 Jan 2012 18:24:20 +0000 (10:24 -0800)
commitc6f9bd8da0485859671812915ddd8019f639c988
tree6a509b3b18969dcdb4667855f8d82ee8262c2917
parent30bbf4c0b17c5b066efb2df30e3a9cb0e660286f
sctp: fix incorrect overflow check on autoclose

[ Upstream commit 2692ba61a82203404abd7dd2a027bda962861f74 ]

Commit 8ffd3208 voids the previous patches f6778aab and 810c0719 for
limiting the autoclose value.  If userspace passes in -1 on 32-bit
platform, the overflow check didn't work and autoclose would be set
to 0xffffffff.

This patch defines a max_autoclose (in seconds) for limiting the value
and exposes it through sysctl, with the following intentions.

1) Avoid overflowing autoclose * HZ.

2) Keep the default autoclose bound consistent across 32- and 64-bit
   platforms (INT_MAX / HZ in this patch).

3) Keep the autoclose value consistent between setsockopt() and
   getsockopt() calls.

Suggested-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Change-Id: I74631fca684d7a475f1eed5f6823c11b62bbd2bc
Reviewed-on: http://git-master/r/74219
Reviewed-by: Varun Wadekar <vwadekar@nvidia.com>
Tested-by: Varun Wadekar <vwadekar@nvidia.com>
include/net/sctp/structs.h
net/sctp/associola.c
net/sctp/protocol.c
net/sctp/socket.c
net/sctp/sysctl.c