Commit Graph

144 Commits

Author SHA1 Message Date
Julian Elischer
7ccbb17bb3 Swap the order of two tags in the pppoe PADI and PADS packets
as there are apparently some buggy switches that need them in that order.
(I hope there aren't any that require them in the old order!)
2000-10-31 14:40:23 +00:00
Brian Somers
76a70671fc Change the format of ngpppoe_init_data so that the provider is NUL
terminated and the data_len field is no longer necessary.

Add ASCII2BINARY and BINARY2ASCII capabilities.

The old format is still understood and dealt with, but can't do
the ASCII2BINARY and BINARY2ASCII stuff.

Approved by: archie
2000-10-31 02:45:24 +00:00
Poul-Henning Kamp
cf9fa8e725 Move suser() and suser_xxx() prototypes and a related #define from
<sys/proc.h> to <sys/systm.h>.

Correctly document the #includes needed in the manpage.

Add one now needed #include of <sys/systm.h>.
Remove the consequent 48 unused #includes of <sys/proc.h>.
2000-10-29 16:06:56 +00:00
Julian Elischer
cc3bbd68c5 Since neither archie nor I work at Whistle any more, change our email
addresses to be the more usefu @freebsd.org ones
so we can keep getting bug-reports.
- man pages to follow..
2000-10-24 17:32:45 +00:00
Poul-Henning Kamp
db7e3af111 Remove unneeded #include <machine/clock.h> 2000-10-15 14:19:01 +00:00
Archie Cobbs
034d9dac34 Calling untimeout(9) leads to a race window where memory could be leaked.
Close this window by simply not calling untimeout(9).
2000-10-11 20:29:12 +00:00
Archie Cobbs
be731c30f3 Fix memory leak.
Submitted by:	Christopher N. Harrell <cnh@ivmg.net>
2000-10-11 19:04:34 +00:00
Archie Cobbs
a89effcde1 Fix typo in NETGRAPH_INIT() macro. 2000-10-09 18:37:11 +00:00
Archie Cobbs
6f16db8137 More complete fix for multi-link sequence number handling bugs.
Add a new control message for querying the sequence number state.
2000-10-06 23:42:02 +00:00
Archie Cobbs
1cd643640f Fix bug in handling of multi-link sequence numbers.
Reported by:	Becca Anderson <becca@worldint.com>
2000-10-06 20:36:17 +00:00
Archie Cobbs
763fcb92c5 Use m_dup() instead of m_copypacket() for the time being. Not all
of the code in the kernel properly checks for read-onlyness before
writing into an mbuf data area. When that code is fixed, the m_dup()
can go back to being m_copypacket().

Requested by:	nsayer
2000-09-23 23:22:27 +00:00
Archie Cobbs
ee79f58e34 Remove unnecessary #include's as reported by phk's script. 2000-09-22 16:51:14 +00:00
Archie Cobbs
65b9a0da90 Allocate all memory (including within node constructors) with M_NOWAIT
instead of M_WAITOK, to allow for maximum flexibility.
2000-09-21 18:01:23 +00:00
Archie Cobbs
840f71272d Allocate memory with M_NOWAIT instead of M_WAITOK because we could
be called in an interrupt context.
2000-09-21 17:33:33 +00:00
Julian Elischer
a62b20c4a1 When sending a packet back to a network interface to simulate an arrived
packet, make sure that the packet has the interface marked in the first mbuf,
the same a truely arrived packets would have.
2000-09-19 08:35:44 +00:00
Archie Cobbs
2b9cf2f709 Rename "struct session" to "struct sess_con" to avoid conflict with
upcoming "struct session" in proc.h.

Requested by:	jasone
2000-09-19 03:22:06 +00:00
Jason Evans
0384fff8c5 Major update to the way synchronization is done in the kernel. Highlights
include:

* Mutual exclusion is used instead of spl*().  See mutex(9).  (Note: The
  alpha port is still in transition and currently uses both.)

* Per-CPU idle processes.

* Interrupts are run in their own separate kernel threads and can be
  preempted (i386 only).

Partially contributed by:	BSDi (BSD/OS)
Submissions by (at least):	cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Archie Cobbs
ed2dbd316a New netgraph node type for Ethernet bridging.
No ipfw support yet.
2000-09-01 01:37:13 +00:00
Archie Cobbs
7f98dc0449 Fix wrong offset bug in ng_enaddr_unparse(). 2000-09-01 00:28:03 +00:00
Archie Cobbs
d99b0733c7 Avoid free'ing a NULL pointer. 2000-08-31 23:08:52 +00:00
Archie Cobbs
f30a8c449f Export ng_ether_enaddr_type for other nodes that want to use it. 2000-08-30 18:39:29 +00:00
Archie Cobbs
56045c6673 Add three more control messages to complement their opposites:
NGM_ETHER_SET_ENADDR, NGM_ETHER_GET_PROMISC, and NGM_ETHER_GET_AUTOSRC.
Alter parsing algorithm so the EN address really looks like one.
2000-08-15 01:05:50 +00:00
Archie Cobbs
a604c8808e Oops, previous commit fixed a bug that was already fixed before.
Back it out.
2000-08-10 23:04:46 +00:00
Archie Cobbs
0521578550 Add new control message to atomically get and clear statistics. 2000-08-10 22:52:40 +00:00
Archie Cobbs
a8f6d55e95 Increase the maximum allowable datagram length. 2000-08-10 22:51:57 +00:00
Archie Cobbs
cac2a7de76 "u_int32_t" should have been "int32_t". 2000-08-10 22:51:26 +00:00
Archie Cobbs
7133ac27ef Use a bigger buffer for NGM_BINARY2ASCII conversion, to handle really
long ASCII control messages.
2000-08-10 22:50:38 +00:00
Archie Cobbs
656485a2c6 - Add new control message NGM_VJC_GET_CONFIG
- Implement control message ASCII'fication for all control messages
2000-08-10 22:48:03 +00:00
Archie Cobbs
57b57be3ec Take advantage of the new unsigned and hex integer types. 2000-08-10 22:45:54 +00:00
Archie Cobbs
8fccb7e954 - Make statistics unsigned.
- Add new control message to atomically get and clear statistics.
2000-08-10 22:44:41 +00:00
Archie Cobbs
7b3bf3f9aa - Add new unsigned and hex integer parse types
- Fix bug in commented example code
2000-08-10 22:43:38 +00:00
Archie Cobbs
8db3c6cd29 - Add new unsigned and hex integer parse types; this allows simplifying
the bytearray parse type.
- Allocate (larger) temporary work buffer dynamically instead of on the
  stack when comparing to the default value.
2000-08-10 22:42:25 +00:00
Archie Cobbs
7b38c4e4d9 RFC 1661 requires that all LCP packets are sent with no address and
control field compression. The ng_ppp(4) node correctly follows this
rule. However, PPPoE is an exception: when doing PPPoE *all* frames
are sent with address and control field compression.

Alter this node's behavior so that when an outgoing frame is received,
any leading address and control field bytes are removed. This makes
this node compatible with ng_ppp(4).
2000-08-10 20:05:12 +00:00
Archie Cobbs
1baeddb81b In a struct sockaddr, sa->sa_len can be zero if uninitialized.
Make sure that this doesn't cause a problem when parsing.
2000-08-09 23:57:44 +00:00
Archie Cobbs
3b49655c07 Fix bug where bundle-level receive statistics were not getting updated. 2000-08-09 01:43:21 +00:00
Archie Cobbs
8a660c773a Fix a bug where we were accessing already free'd memory during node shutdown.
Detected via:	0xdeadc0de
2000-08-07 22:41:12 +00:00
Archie Cobbs
4b39c3c7b3 Add three new control messages to the ng_ether(4) netgraph node type:
NGM_ETHER_GET_ENADDR:	Get the device's Ethernet address
    NGM_ETHER_SET_PROMISC:	Enable/disable promiscuous mode
    NGM_ETHER_SET_AUTOSRC:	Enable/disable packet source address override
2000-08-07 18:52:26 +00:00
Archie Cobbs
9dcab5306e Fix misspelling. 2000-08-05 20:17:04 +00:00
Archie Cobbs
da010626df Followup to previous commit..
- It's worthwhile to use untimeout(9), even though we must still protect
  against "false" timeouts, because most of the time it saves having to
  handle a dummy timeout event.
- Slight tweaks to the delayed ACK algorithm paramters.
2000-07-25 18:57:20 +00:00
Archie Cobbs
678f9e335e Several fixes:
- Fix slowness when operating over fast connections, where the timeout(9)
  granularity is on the same order of magnitude as the round trip time.
  timeout(9) can happen up to 1 tick early, which was causing receive
  ack timeouts to happen too early, causing bogus "lost" packets.
- Increase the local time counter to 64 bits to avoid roll-over.
- Keep statistics on memory allocation failures.
- Add a new option to always include the ack when sending data packets.
  Might be useful in high packet loss situations. Might not.
2000-07-25 00:23:19 +00:00
Archie Cobbs
d2ea40c23f Allocate memory with M_NOWAIT instead of M_WAITOK, because it's possible
for these routines to be called from an interrupt context.

PR:		kern/20057
2000-07-20 17:23:49 +00:00
Archie Cobbs
7c24617ca0 Remove node's name reference when the interface is detached;
otherwise, the ng_ether.ko KLD will never be unloadable after
all Ethernet interfaces are detached, as it should be, because
of the lingering extra reference.

Submitted by:	"Yevmenkin, Maksim N, CSCIO" <myevmenkin@att.com>
2000-07-19 17:33:53 +00:00
Archie Cobbs
840e97785b Set NG_INVALID flag when destroying node. 2000-07-14 22:35:13 +00:00
Jeroen Ruigrok van der Werven
514baf3f99 Fix typo, teh -> the. 2000-07-14 11:17:16 +00:00
Archie Cobbs
051fbf7d66 Fix race condition caused by using NG_SEND_DATAQ() where we meant
to use ng_queue_data().

Reported by:	Udo Erdelhoff <ue@nathan.ruhr.de>
2000-07-12 23:55:07 +00:00
Julian Elischer
561e4fb9f6 Don't forget to set our MAC address into packets we wre sending out via
netgraph. Eventually we may need to have a separate hook for packets
that already have a source AMC address but for now just drop it in.
Should fix PPPoE.
2000-07-06 15:35:59 +00:00
Poul-Henning Kamp
7095e0970e Experiemntal ascii based device configuration mechanism.
This may or may not survive, decision will be made well before 5.0-R
2000-07-03 13:34:18 +00:00
Archie Cobbs
e962a82308 Fix incorrectly implemented receive ACK timeout algorithm:
instead of bumping the recvAck counter by one, pretend that
all outstanding xmit packets are acknowleged, and restart
transmitting anew, with an empty (but halved) transmit window.

Put a lower bound on the adaptive timeout value.
2000-06-28 19:43:34 +00:00
Archie Cobbs
e1e1452d61 Make the ng_ether(4) node type dynamically loadable like the rest.
This means 'options NETGRAPH' is no longer necessary in order to get
netgraph-enabled Ethernet interfaces. This supports loading/unloading
the ng_ether.ko and attaching/detaching the Ethernet interface in any
order.

Add two new hooks 'upper' and 'lower' to allow access to the protocol
demux engine and the raw device, respectively. This enables bridging
to be defined as a netgraph node, if so desired.

Reviewed by:	freebsd-net@freebsd.org
2000-06-26 23:34:54 +00:00
Archie Cobbs
3cd7db2297 - Start sequence numbers at zero instead of one; the rest of the
world seems to interpret the spec this way
- Initialize transmit window to two instead of one; helps get things
  going initially when the first packet may get dropped
- Really fix the shutdown + timeout race condition this time
2000-06-26 19:43:24 +00:00