kernel.git/include/linux/socket.h, branch linux-2.6.32.y

tcp: allow splice() to build full TSO packets

2013-06-10T09:43:32Z

[ This combines upstream commit 2f53384424251c06038ae612e56231b96ab610ee and the follow-on bug fix commit 35f9c09fe9c72eb8ca2b8e89a593e1c151f28fc2 ] vmsplice()/splice(pipe, socket) call do_tcp_sendpages() one page at a time, adding at most 4096 bytes to an skb. (assuming PAGE_SIZE=4096) The call to tcp_push() at the end of do_tcp_sendpages() forces an immediate xmit when pipe is not already filled, and tso_fragment() try to split these skb to MSS multiples. 4096 bytes are usually split in a skb with 2 MSS, and a remaining sub-mss skb (assuming MTU=1500) This makes slow start suboptimal because many small frames are sent to qdisc/driver layers instead of big ones (constrained by cwnd and packets in flight of course) In fact, applications using sendmsg() (adding an additional memory copy) instead of vmsplice()/splice()/sendfile() are a bit faster because of this anomaly, especially if serving small files in environments with large initial [c]wnd. Call tcp_push() only if MSG_MORE is not set in the flags parameter. This bit is automatically provided by splice() internals but for the last page, or on all pages if user specified SPLICE_F_MORE splice() flag. In some workloads, this can reduce number of sent logical packets by an order of magnitude, making zero-copy TCP actually faster than one-copy :) Reported-by: Tom Herbert Cc: Nandita Dukkipati Cc: Neal Cardwell Cc: Tom Herbert Cc: Yuchung Cheng Cc: H.K. Jerry Chu Cc: Maciej Å»enczykowski Cc: Mahesh Bandewar Cc: Ilpo JÃ¤rvinen Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman Signed-off-by: Willy Tarreau

net: Limit socket I/O iovec total length to INT_MAX.

2010-12-09T21:27:13Z

commit 8acfe468b0384e834a303f08ebc4953d72fb690a upstream. This helps protect us from overflow issues down in the individual protocol sendmsg/recvmsg handlers. Once we hit INT_MAX we truncate out the rest of the iovec by setting the iov_len members to zero. This works because: 1) For SOCK_STREAM and SOCK_SEQPACKET sockets, partial writes are allowed and the application will just continue with another write to send the rest of the data. 2) For datagram oriented sockets, where there must be a one-to-one correspondance between write() calls and packets on the wire, INT_MAX is going to be far larger than the packet size limit the protocol is going to check for and signal with -EMSGSIZE. Based upon a patch by Linus Torvalds. Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman

tcp: Fix >4GB writes on 64-bit.

2010-10-29T04:44:07Z

[ Upstream commit 01db403cf99f739f86903314a489fb420e0e254f ] Fixes kernel bugzilla #16603 tcp_sendmsg() truncates iov_len to an 'int' which a 4GB write to write zero bytes, for example. There is also the problem higher up of how verify_iovec() works. It wants to prevent the total length from looking like an error return value. However it does this using 'int', but syscalls return 'long' (and thus signed 64-bit on 64-bit machines). So it could trigger false-positives on 64-bit as written. So fix it to use 'long'. Reported-by: Olaf Bonorden Reported-by: Daniel Büse Reported-by: Andrew Morton Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman

net: Support inclusion of before

2009-10-05T07:24:36Z

The following user-space program fails to compile: #include #include int main() { return 0; } The reason is that tests __GLIBC__ to decide whether it should define various structures and macros that are now defined for user-space by , but __GLIBC__ is not defined if no libc headers have yet been included. It seems safe to drop support for libc 5 now. Signed-off-by: Ben Hutchings Signed-off-by: Bastian Blank Signed-off-by: David S. Miller

Add constants for the ieee 802.15.4 stack

2009-06-09T12:25:30Z

IEEE 802.15.4 stack requires several constants to be defined/adjusted. Signed-off-by: Dmitry Eremin-Solenikov Signed-off-by: Sergey Lapin Signed-off-by: David S. Miller

af_iucv: add sockopt() to enable/disable use of IPRM_DATA msgs

2009-04-23T11:04:32Z

Provide the socket operations getsocktopt() and setsockopt() to enable/disable sending of data in the parameter list of IUCV messages. The patch sets respective flag only. Signed-off-by: Hendrik Brueckner Signed-off-by: Ursula Braun Signed-off-by: David S. Miller

tun: fix tun_chr_aio_write so that aio works

2009-04-21T12:42:46Z

aio_write gets const struct iovec * but tun_chr_aio_write casts this to struct iovec * and modifies the iovec. As a result, attempts to use io_submit to send packets to a tun device fail with weird errors such as EINVAL. Since tun is the only user of skb_copy_datagram_from_iovec, we can fix this simply by changing the later so that it does not touch the iovec passed to it. Signed-off-by: Michael S. Tsirkin Signed-off-by: David S. Miller

net: skb_copy_datagram_const_iovec()

2009-04-21T12:42:44Z

There's an skb_copy_datagram_iovec() to copy out of a paged skb, but it modifies the iovec, and does not support starting at an offset in the destination. We want both in tun.c, so let's add the function. It's a carbon copy of skb_copy_datagram_iovec() with enough changes to be annoying. Signed-off-by: Michael S. Tsirkin Signed-off-by: David S. Miller

Merge branch 'header-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

2009-03-26T23:11:41Z

* 'header-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (50 commits) x86: headers cleanup - setup.h emu101k1.h: fix duplicate include of compiler-gcc4: conditionalize #error on __KERNEL__ remove __KERNEL_STRICT_NAMES make netfilter use strict integer types make drm headers use strict integer types make MTD headers use strict integer types make most exported headers use strict integer types make exported headers use strict posix types unconditionally include asm/types.h from linux/types.h make linux/types.h as assembly safe Neither asm/types.h nor linux/types.h is required for arch/ia64/include/asm/fpu.h headers_check fix cleanup: linux/reiserfs_fs.h headers_check fix cleanup: linux/nubus.h headers_check fix cleanup: linux/coda_psdev.h headers_check fix: x86, setup.h headers_check fix: x86, prctl.h headers_check fix: linux/reinserfs_fs.h headers_check fix: linux/socket.h headers_check fix: linux/nubus.h ... Manually fix trivial conflicts in: include/linux/netfilter/xt_limit.h include/linux/netfilter/xt_statistic.h

RDS: Add AF and PF #defines for RDS sockets

2009-02-27T07:41:38Z

RDS is a reliable datagram protocol used for IPC on Oracle database clusters. This adds address and protocol family numbers for it. Signed-off-by: Andy Grover Signed-off-by: David S. Miller