tcp_send: avoid ip fragmentation of last segment of a tcp stream

With a LSO (large segment offload) capable network device, tcp data is
sent in segments larger than mss. This works well normally and

tcp_send: avoid ip fragmentation of last segment of a tcp stream

With a LSO (large segment offload) capable network device, tcp data is
sent in segments larger than mss. This works well normally and the
network card splits the large segments into smaller ip packages not
exceeding the mss.

The last segment of a stream, when fin flag is already set, takes a
different code path (tcp_xmit_mp()). LSO is not possible here and not
supported. If the segment is large (> mss) in this case, it is
ip fragmented. This is not wanted.

The fix is to ensure the last segment of a tcp stream is below mss
even in case of LSO.

(cherry picked from commit 72434ca23766cfa425b040bac615a51f16341693)

show more ...

5295 remove maxburst logic from TCP's send algorithm
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Eric Diven <eric.div

5295 remove maxburst logic from TCP's send algorithm
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Eric Diven <eric.diven@delphix.com>
Reviewed by: Dan McDonald <danmcd@omniti.com>
Reviewed by: Robert Mustacchi <rm@joyent.com>
Reviewed by: Gordon Ross <gordon.ross@nexenta.com>
Approved by: Gordon Ross <gordon.ross@nexenta.com>

show more ...

6970847 Special option handling can be pulled out of tcp_xmit_mp()

6963859 ipcl_conn_create() triggers panic in snv_143 on a machine with solaris10 branded zones
6965253 assertion failed: so->so_listener == NULL, file: ../../common/fs/sockfs/sockcommon.c, line: 524

6939100 convert KSSL into a socket filter

PSARC/2009/590 Socket Filter Framework
6939085 Socket Filter Framework
6802067 connect_failed kernel socket callback is not triggered
6776450 time spent in tcp_close could be reduced/deferred to a wo

PSARC/2009/590 Socket Filter Framework
6939085 Socket Filter Framework
6802067 connect_failed kernel socket callback is not triggered
6776450 time spent in tcp_close could be reduced/deferred to a worker thread
6828586 assertion failed: family == 26, file: ../../common/fs/sockfs/socksyscalls.c, line: 1608
6802078 kernel socket 'newconn' callback is passing rcv queue size as an argument

show more ...

PSARC 2010/106 DTrace TCP and UDP providers
6742331 DTrace TCP Provider
6932981 DTrace UDP provider

6940729 gate machine panic in ip.c: assertion failed: pkt_len == msgdsize(mp)
6936196 Missing IXAF_IS_IPV4 flag causes punchin to go boom

6935348 tcp_sack_info should always be there
6924317 TCP time wait processing is called unnecessarily
6924089 TCP Timer code does double time conversion

PSARC 2009/306 Brussels II - ipadm and libipadm
PSARC 2010/080 Brussels II addendum
6827318 Brussels Phase II aka ipadm(1m)
6731945 need BSD getifaddrs() API
6909065 explicitly disallow non-contiguou

PSARC 2009/306 Brussels II - ipadm and libipadm
PSARC 2010/080 Brussels II addendum
6827318 Brussels Phase II aka ipadm(1m)
6731945 need BSD getifaddrs() API
6909065 explicitly disallow non-contiguous netmasks in the next minor release
6853922 ifconfig dumps core when ether address is non-hexadecimal.
6815806 ipReasmTimeout value should be variable
6567083 nd_getset has some dead and confusing code.
6884466 remove unused tcp/sctp ndd tunables
6928813 Comments at odds with default value of tcp_time_wait_interval
6236982 ifconfig usesrc lets adapter use itself as source address
6936855 modifying the ip6_strict_src_multihoming to non-zero value will unbind V4 IREs

show more ...

PSARC 2010/042 increase max TCP_INIT_CWND
6923847 Increase TCP_INIT_CWND max
6918307 Some TCP kstats are not needed
6923858 TCP connection counter
6925635 The file tcp.c is too big