openslx-ng/ipxe.git/src/net/tcp.c, branch v0.9.7

[xfer] Make consistent assumptions that xfer metadata can never be NULL

2009-02-15T08:44:22+00:00

The documentation in xfer.h and xfer.c does not say that the metadata parameter is optional in calls such as xfer_deliver_iob_meta() and the deliver_iob() method. However, some code in net/ is prepared to accept a NULL pointer, and xfer_deliver_as_iob() passes a NULL pointer directly to the deliver_iob() method. Fix this mess of conflicting assumptions by making everything assume that the metadata parameter is mandatory, and fixing xfer_deliver_as_iob() to pass in a dummy metadata structure (as is already done in xfer_deliver_iob()).

[tcp] Always set PUSH flag on TCP transmissions

2009-01-21T04:22:34+00:00

Apparently this can cause a major speedup on some iSCSI targets, which will otherwise wait for a timer to expire before responding. It doesn't seem to hurt other simple TCP test cases (e.g. HTTP downloads). Problem and solution identified by Shiva Shankar <802.11e@gmail.com>

[tcpip] Allow for transmission to multicast IPv4 addresses

2009-01-21T03:40:39+00:00

When sending to a multicast address, it may be necessary to specify the source address explicitly, since the multicast destination address does not provide enough information to deduce the source address via the miniroute table. Allow the source address specified via the data-xfer metadata to be passed down through the TCP/IP stack to the IPv4 layer, which can use it as a default source address.

[x86_64] Fix assorted 64-bit compilation errors and warnings

2008-11-19T19:33:05+00:00

Remove various 32-bit assumptions scattered throughout the codebase. The code is still not necessarily 64-bit clean, but will at least compile.

[i386] Change [u]int32_t to [unsigned] int, rather than [unsigned] long

2008-11-19T19:15:44+00:00

This brings us in to line with Linux definitions, and also simplifies adding x86_64 support since both platforms have 2-byte shorts, 4-byte ints and 8-byte long longs.

[TCP] Avoid shrinking TCP window

2008-06-04T23:28:17+00:00

Maintain state for the advertised window length, and only ever increase it (instead of calculating it afresh on each transmit). This avoids triggering "treason uncloaked" messages on Linux peers. Respond to zero-length TCP keepalives (i.e. empty data packets transmitted outside the window). Even if the peer wouldn't otherwise expect an ACK (because its packet consumed no sequence space), force an ACK if it was outside the window. We don't yet generate TCP keepalives. It could be done, but it's unclear what benefit this would have. (Linux, for example, doesn't start sending keepalives until the connection has been idle for two hours.)

Modify gPXE core and drivers to work with the new timer subsystem

2008-03-02T00:41:10+00:00

Signed-off-by: Alexey Zaytsev

Make seek information part of the xfer metadata, rather than an entirely

2008-01-08T16:46:55+00:00

separate xfer method. Add missing .alloc_iob entries to several xfer_interface_operations structures.

Various warnings fixups for OpenBSD with gcc-3.3.5.

2007-12-07T00:11:43+00:00

Use start_timer_nodelay() in protocols which rely on the retry timer

2007-08-13T18:03:33+00:00

to generate the initial transmission; this cuts off around 0.3s per instantiated connection.