1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-24 11:29:10 +00:00
Commit Graph

2865 Commits

Author SHA1 Message Date
Andre Oppermann
104ebb2a45 Make the handling of the tcp window explicit for the SYN_SENT case
in tcp_outout().  This is currently not strictly necessary but paves
the way to simplify the entire SYN options handling quite a bit.
Clarify comment.  No change in effective behavour with this commit.

RFC1323 requires the window field in a SYN (i.e., a <SYN> or
<SYN,ACK>) segment itself never be scaled.
2007-06-09 21:19:12 +00:00
Andre Oppermann
5396d0f8d8 Remove some bogosity from the SYN_SENT case in tcp_do_segment
and simplify handling of the send/receive window scaling.  No
change in effective behavour.

RFC1323 requires the window field in a SYN (i.e., a <SYN> or
<SYN,ACK>) segment itself never be scaled.

Noticed by:	yar
2007-06-09 21:09:49 +00:00
Andre Oppermann
b7de7d87a0 Don't send pure window updates when the peer has closed the connection
and won't ever send more data.
2007-06-09 19:39:14 +00:00
Andre Oppermann
f58747375d Handle a race condition on >2 core machines in tcp_timer() when
a timer issues a shutdown and a simultaneous close on the socket
happens.  This race condition is inherent in the current socket/
inpcb life cycle system but can be handled well.

Reported by:	kris
Tested by:	kris (on 8-core machine)
2007-06-09 17:49:39 +00:00
Randall Stewart
2bf083e4c9 - Opps.. takes out debug printfs I accidentally left in :-( 2007-06-09 13:53:27 +00:00
Randall Stewart
d00aff5d79 - fix send_failed notification contents
- Reorder send failed to be in correct order.
- Fixed calulation of init-ack to be right off
  mbuf lengths instead of the precalculated value. This
  will fix one 64 bit platform issue.
2007-06-09 13:46:57 +00:00
Yaroslav Tykhiy
22b971db87 Replace a constant with an already defined symbolic name for it.
Tested with: md5(1)
2007-06-08 13:43:28 +00:00
Yaroslav Tykhiy
dba3c50842 Add a sysctl for the purge run interval so that it can
be tuned along with the rest of hostcache parameters.
The new sysctl name is `net.inet.tcp.hostcache.prune'.
2007-06-08 13:35:51 +00:00
Randall Stewart
108df27c0b - RTO was not being initialized to 0, thus the rtt calculation
algoritm would not go through the proper initialization.
- The initialization was incorrect as well, causing problems in
  sat networks with > 1sec RTT
- Get rid of magic numbers in RTT calculations.
2007-06-08 10:57:11 +00:00
Andre Oppermann
45024be06f In tcp_hc_insert() we may have the case where we have hit the global
cache size limit but this bucket row is empty.  Normally we want to
recycle the oldest entry in the bucket row.  If there isn't any the
TAILQ_REMOVE leads to a panic by trying to remove a non-existing
element.  Fix this by just returning NULL and failing the insert.
This is not a problem as the TCP hostache is only advisory.

Submitted by:	jhb
2007-06-07 21:41:50 +00:00
Andre Oppermann
1f939165ce Correctly print SEQ and IRS in the corresponding log message in
syncache_expand().
2007-06-06 22:10:12 +00:00
Gleb Smirnoff
e9bf9fb67c Do not leak lock in the case of EEXIST error.
PR:		kern/92776
Submitted by:	Ed Schouten <Ed.Schouten tunix.nl>
2007-06-06 14:21:49 +00:00
Randall Stewart
5f26a41d17 - Fixes a case where doing a sysctl would leave locks held
when coping out association data.
- Fixes a small bug that prevented the SCTP_UNORDERED indication
  from going up to the app on a recv in the sinfo_flags field.
2007-06-06 00:40:41 +00:00
David Malone
041b706b2f Despite several examples in the kernel, the third argument of
sysctl_handle_int is not sizeof the int type you want to export.
The type must always be an int or an unsigned int.

Remove the instances where a sizeof(variable) is passed to stop
people accidently cut and pasting these examples.

In a few places this was sysctl_handle_int was being used on 64 bit
types, which would truncate the value to be exported.  In these
cases use sysctl_handle_quad to export them and change the format
to Q so that sysctl(1) can still print them.
2007-06-04 18:25:08 +00:00
Randall Stewart
f4c93d2405 - fix initial pcb vrf setting when the initial vrf is not the
default_vrf_id
- Missing lock/unlock of inp added as well in the v6 side.
- IFN hash table moves to sctppcbinfo since indexes are
  unique across systems (including different VRFs) this makes it easier
  to do ifn lookups.
2007-06-02 11:05:08 +00:00
Randall Stewart
ad21a36485 - Take out the broken table-id concept. Panda Routers have a M-VRF
concept that is NOT well thought out for a multi-homed transport
  protocol. So the useless table-id entries passed around need to
  be removed.
- Add a event timer for the zero copy api.
- Fix a bug in sctp_timer.c when searching for an alternate
  with the largest ssthresh (the compare was wrong).
2007-06-01 11:19:54 +00:00
Jeff Roberson
1c4bcd050a - Move rusage from being per-process in struct pstats to per-thread in
td_ru.  This removes the requirement for per-process synchronization in
   statclock() and mi_switch().  This was previously supported by
   sched_lock which is going away.  All modifications to rusage are now
   done in the context of the owning thread.  reads proceed without locks.
 - Aggregate exiting threads rusage in thread_exit() such that the exiting
   thread's rusage is not lost.
 - Provide a new routine, rufetch() to fetch an aggregate of all rusage
   structures from all threads in a process.  This routine must be used
   in any place requiring a rusage from a process prior to it's exit.  The
   exited process's rusage is still available via p_ru.
 - Aggregate tick statistics only on demand via rufetch() or when a thread
   exits.  Tick statistics are kept in the thread and protected by sched_lock
   until it exits.

Initial patch by:	attilio
Reviewed by:		attilio, bde (some objections), arch (mostly silent)
2007-06-01 01:12:45 +00:00
Robert Watson
abc7d91030 (1) In tcp_usrclosed(), tp can never become NULL, so don't test for NULL
before handling the socket disconnection case.

(2) Clean up surrounding comments and formatting.

Found with:	Coverity Prevent(tm) (1)
CID:		2203
2007-05-31 12:06:02 +00:00
Randall Stewart
4c9179ad6c - Fixed (Apple) compiler warnings in sctp_input.c, sctputil.c, sctp_output.c
- Fixed a LOR in handling a cookie. Turns out create lock is applied.
  And if we abort processing, this causes LOR. Changed to force the
  timer to clean up, that way create lock is released.
2007-05-30 22:34:21 +00:00
Randall Stewart
0696e1203e - Fix a memory overwrite when the mapping array
is expanded, size of expansion was not taken int consideration.
-  Fix so vtag hash is 1 bigger so that it modulo's out
   correctly, avoids a panic when restart with right modulo happens.
-  do not dereference stcb when control->do_not_ref_stcb is set
-  Fix up packet logging to not often use a lock and also to
   add to options.
-  Fix some logging option duplication in the sctputil.h
2007-05-30 17:39:45 +00:00
Randall Stewart
3c6f353630 Adds gcc attribute to prevent inlining of a function. If
it goes inline we may well blow the stack if witness and
such are enabled.
2007-05-29 14:17:47 +00:00
Randall Stewart
6b4ae3566a - Fix spelling errors in comments per Ruslan (.. thanks... ) 2007-05-29 11:53:27 +00:00
Randall Stewart
207304d4b7 - Fixes so we won't try to start a timer when we
hold a wq lock for the iterator. Panda uses a
  silly recursive lock they hold through the timer.
- Add poor mans wireshark compile option..
- Allocate and start using SCTP_M_XXX for all SCTP_MALLOC() calls.
- sysctl now will get back the refcnt for viewing by onlookers.

Reviewed by:	gnn
2007-05-29 09:29:03 +00:00
Andre Oppermann
8d573cc158 Make log messages more verbose and simpler to understand for non-experts.
Update comments to be more conscious, verbose and fully reflect reality.
2007-05-28 23:27:44 +00:00
Andre Oppermann
e885b205c6 Fix indentation of the syncache_expand() section in tcp_input(). 2007-05-28 11:35:40 +00:00
Randall Stewart
d61a0ae066 - fixed autclose to not allow setting on 1-2-1 model.
- bounded cookie-life to 1 second minimum in socket option set.
- Delayed_ack_time becomes delayed_ack per new socket api document.
- Improve port number selection, we now use low/high bounds and
  no chance of a endless loop. Only one call to random per bind
  as well.
- fixes so set_peer_primary pre-screens addresses to be
  valid to this host.
- maxseg did not allow setting on an assoc basis. We needed
  to thus track and use an association value instead of a inp value.
- Fixed ep get of HB status to report back properly.
- use settings flag to tell if assoc level hb is on off not
  the timer.. since the timer may still run if unconf address
  are present.
- check for crazy ENABLE/DISABLE conditions.
- set and get of pmtud (fixed path mtu) not always taking into account ovh.
- Getting PMTU info on stcb only needs to return PMTUD_ENABLED if
  any net is doing PMTU discovery.
- Panic or warning fixed to not do so when a valid ip frag is
  taking place.
- sndrcvinfo appearing in both inp and stcb was full size, instead
  of the non-pad version. This saves about 92 bytes from each struct
  by carefully converting to use the smaller version.
- one-2-one model get(maxseg) would always get ep value, never the
  tcb's value.
- The delayed ack time could be under a tick, this fixes so
  it bounds it to at least 1 tick for platforms whos tick
  is more than a ms.
- Fragment interleave level set to wrong default value.
- Fragment interleave could not set level 0.
- Defered stream reset was broken due to a guard check and ntohl issue.
- Found two lock order reversals and fixed.
- Tighten up address checking, if the user gives an address the sa_len
  had better be set properly.
- Get asoc by assoc-id would return a locked tcb when it was asked
  not to if the tcb was in the restart hash.
- sysctl to dig down and get more association details

Reviewed by:	gnn
2007-05-28 11:17:24 +00:00
Andre Oppermann
a160e6302c Refactor and rewrite in parts the SYN handling code on listen sockets
in tcp_input():

 o tighten the checks on allowed TCP flags to be RFC793 and
   tcp-secure conform
 o log check failures to syslog at LOG_DEBUG level
 o rearrange the code flow to be easier to follow
 o add KASSERTs to validate assumptions of the code flow

Add sysctl net.inet.tcp.syncache.rst_on_sock_fail defaulting to enable
that controls the behavior on socket creation failure for a otherwise
successful 3-way handshake.  The socket creation can fail due to global
memory shortage, listen queue limits and file descriptor limits.  The
sysctl allows to chose between two options to deal with this.  One is
to send a reset to the other endpoint to notify it about the failure
(default).  The other one is to ignore and treat the failure as a
transient error and have the other endpoint retransmit for another try.

Reviewed by:	rwatson (in general)
2007-05-28 11:03:53 +00:00
Robert Watson
e487a5e2a0 Normalize spelling and grammar in TCP hostcache comments. 2007-05-27 19:39:26 +00:00
Robert Watson
c214db75f2 In tcp_timer_2msl(), tp can never become NULL, so don't check it for
NULL before entering tcp_trace().

Found with:	Coverity Prevent(tm)
CID:		1840
2007-05-27 17:52:02 +00:00
Robert Watson
b312d4b0ba Don't assign sp to the value of s when we're about to assign it instead to
s + strlen(s).

Found with:	Coverity Prevent(tm)
CID:		2243
2007-05-27 17:02:54 +00:00
Andre Oppermann
faedb66c2a The printf %b list in PRINT_TH_FLAGS has to be in octal numbering.
Thus convert \8 to \10 and the warnings go away.

Pointed out by:	sam, ru, thompsa
2007-05-25 21:28:49 +00:00
Andre Oppermann
a250f3820c Add CWR back into the PRINT_TH_FLAGS list as gcc42 doesn't complain
about \8 in a string anymore.
2007-05-23 19:16:21 +00:00
Andre Oppermann
ec05a17370 In tcp_log_addrs():
o add the hex output of the th_flags field to the example log
   line in comments
 o simplify the log line length calculation and make it less
   evil
 o correct the test for the length panic; the line isn't on
   the stack but malloc'ed
2007-05-23 19:07:53 +00:00
Andre Oppermann
d2ddf5d4b0 Be more restrictive with segment validity checks in syncache_expand()
and log check failures to syslog at LOG_DEBUG level.

Always prefill the sc->sc_ts field to use it in the checks.
2007-05-18 21:42:25 +00:00
Andre Oppermann
5df429a002 o Add syslog logging under LOG_DEBUG to various failures caused by
bogus segments
o Add more KASSERT()s
o Update comments
2007-05-18 21:13:01 +00:00
Andre Oppermann
df541e5fc1 Add tcp_log_addrs() function to generate and standardized TCP log line
for use thoughout the tcp subsystem.

It is IPv4 and IPv6 aware creates a line in the following format:

 "TCP: [1.2.3.4]:50332 to [1.2.3.4]:80 tcpflags <RST>"

A "\n" is not included at the end.  The caller is supposed to add
further information after the standard tcp log header.

The function returns a NUL terminated string which the caller has
to free(s, M_TCPLOG) after use.  All memory allocation is done
with M_NOWAIT and the return value may be NULL in memory shortage
situations.

Either struct in_conninfo || (struct tcphdr && (struct ip || struct
ip6_hdr) have to be supplied.

Due to ip[6].h header inclusion limitations and ordering issues the
struct ip and struct ip6_hdr parameters have to be casted and passed
as void * pointers.

tcp_log_addrs(struct in_conninfo *inc, struct tcphdr *th, void *ip4hdr,
    void *ip6hdr)

Usage example:

 struct ip *ip;
 char *tcplog;

 if (tcplog = tcp_log_addrs(NULL, th, (void *)ip, NULL)) {
	log(LOG_DEBUG, "%s; %s: Connection attempt to closed port\n",
	    tcplog, __func__);
	free(s, M_TCPLOG);
 }
2007-05-18 19:58:37 +00:00
John Baldwin
0ba5d2eedb Fix statistical accounting for bytes and packets during sack retransmits.
MFC after:	1 week
Submitted by:	mohans
2007-05-18 19:56:24 +00:00
JINMEI Tatuya
187069853c - Disabled responding to NI queries from a global address by default as
specified in RFC4620.  A new flag for icmp6_nodeinfo was added to enable the
  feature.
- Also cleaned up the code so that the semantics of the icmp6_nodeinfo
  flags is clearer (i.e., defined specific macro names instead of using
  hard-coded values).

Approved by:	gnn (mentor)
MFC after:	1 week
2007-05-17 21:20:24 +00:00
Randall Stewart
3c503c28da - Fixed 1-2-1 model to not worry about associd in sockopts
- Fixed RTOinfo for bounding.
- Fixed connect() to return ECONNREFUSED when an ABORT is received.
- Added comments to direct Static Analysis not to look at some things
  it does not understand (comments are /* sa_ignore XXXXX */)
- Bind when colliding was broken, missing not_found = 1 before
  checking to see if the port was in use caused endless bind loop.
- Cookie life needs to be in milliseconds to conform to socket api.
- Cookie life is not supposed to change if its 0, On the assoc
  level set we changed it to 0 opps.
- Two more static analysis issues identified by the cisco
  tool. Null checks needed.
- An issue for sendfile(). Need to validate the correct
  input argument.
- When sending failed due to a no route to host, we leaked
  the mbuf chain failing to call m_freem().
- Fix #ifdef issue for getting hash block len when HAVE_SHA2 is NOT defined
Reviewed by:	gnn
2007-05-17 12:16:24 +00:00
Oleg Bulyzhin
7e17f8b864 Unbreak IPv4 kernel build. 2007-05-17 00:05:13 +00:00
Robert Watson
6751f8364e Remove leading spaces before tabs spotted thanks to silby using
kwrite to read ip_input.c.
2007-05-16 20:46:58 +00:00
Andre Oppermann
abb91d889a Remove now unused stuff forgotten in the previous commit. 2007-05-16 17:55:22 +00:00
Andre Oppermann
2104448fe7 Move TIME_WAIT related functions and timer handling from files
other than repo copied tcp_subr.c into tcp_timewait.c#1.284:

 tcp_input.c#1.350 tcp_timewait() -> tcp_twcheck()

 tcp_timer.c#1.92 tcp_timer_2msl_reset() -> tcp_tw_2msl_reset()
 tcp_timer.c#1.92 tcp_timer_2msl_stop() -> tcp_tw_2msl_stop()
 tcp_timer.c#1.92 tcp_timer_2msl_tw() -> tcp_tw_2msl_scan()

This is a mechanical move with appropriate renames and making
them static if used only locally.

The tcp_tw_2msl_scan() cleanup function is still run from the
tcp_slowtimo() in tcp_timer.c.
2007-05-16 17:14:25 +00:00
David Malone
39629c92cc When verifying the IPv4 UDP checksum, don't overwrite the checksum
value in the mbuf with the result of the calculation. Previously,
if we chose to return an ICMP message, the quoted UDP checksum bytes
would be different to what was sent.

PR:		112471
Submitted by:	Matthew Luckie <mluckie@cs.waikato.ac.nz>
MFC after:	3 weeks
2007-05-16 09:12:16 +00:00
Andre Oppermann
ec9c755352 Complete the (mechanical) move of the TCP reassembly and timewait
functions from their origininal place to their own files.

TCP Reassembly from tcp_input.c -> tcp_reass.c
TCP Timewait   from tcp_subr.c  -> tcp_timewait.c
2007-05-13 22:16:13 +00:00
Andre Oppermann
57615c7e86 Drop everything that doesn't belong into this new file.
It's neither functional not connected to the build yet.
2007-05-11 21:17:53 +00:00
Andre Oppermann
1433541aa4 Drop everything that doesn't belong into this new file.
It's neither functional nor connected to the build yet.
2007-05-11 21:04:57 +00:00
Andre Oppermann
0489b64c5e Make the TCP timer callout obtain Giant if the network stack is marked
as non-mpsafe.

This change is to be removed when all protocols are mp-safe.
2007-05-11 20:52:47 +00:00
Andre Oppermann
504abdc6e6 Add the timestamp offset to struct tcptw so we can generate proper
ACKs in TIME_WAIT state that don't get dropped by the PAWS check
on the receiver.
2007-05-11 18:29:39 +00:00
Robert Watson
632bbf0f5b Coalesce two identical UCB licenses into a single license instance with
one set of copyright years.

White space and comment cleanup.

Export $FreeBSD$ via __FBSDID.
2007-05-11 11:21:43 +00:00