Commit Graph

42082 Commits

Author SHA1 Message Date
julian
b7856caf0b Make it compile inthe NETGRAPH case. 1999-10-25 23:14:04 +00:00
julian
f963ed1eaf Bring ng_iface up-to-date with what has happenned to the bpf code. 1999-10-25 22:36:39 +00:00
julian
50524ff0ef Add an option for the van Jacobson Compression Netgraph node. 1999-10-25 22:07:30 +00:00
wes
60a677e4fa Add Kirk's comment about "copycenter"; it's just too good to pass up. 1999-10-25 21:59:31 +00:00
billf
16c187b39b Add $FreeBSD$ 1999-10-25 21:51:04 +00:00
wpaul
41000ec44d Small tweak: just reset the transmit block instead of doing a global reset
in xl_init(). This achieves the effect that I wanted without totally
resetting the chip.
1999-10-25 20:52:32 +00:00
dillon
3f308ed167 Add -c, -s, and -W options to nfsstat. Improve interval output and add
wide-format option to get client-side ops and cache statistics on a single
    line.

    Change client side ops to the number of attempted ops, whether cached or
    not, rather then just the number of rpc's that went through to the server.
    This brings nfsstat inline with systat -vm and vmstat and reduces
    confusion.  The combined cache percentage stats now available via
    'nfsstat -cW 1' becomes very useful.
1999-10-25 19:39:17 +00:00
dillon
47a282162e Move NFS access cache hits/misses into nfsstats structure so
/usr/bin/nfsstat can get to it easily.
1999-10-25 19:22:33 +00:00
wpaul
b01e7dfec4 Make release is now falling over in if_fe.c due to Warner's pccard changes:
if_fe.c uses PCCARD_MODULE() and is part of GENERIC. I've #ifdef'ed out
the #include of "card.h" to hopefully disable pccard support in this
driver until it can be converted. I'm not positive this will fix make
release, but it can't possibly make it any worse than it is now.

I hope this stuff settles down soon.
1999-10-25 17:04:17 +00:00
julian
ff0350df84 install netgraph.h a bit earlier.
Submitted by:	Boris Popov <bp@butya.kz>
1999-10-25 14:50:13 +00:00
brian
faf44b90cf Introduce ``set logout''; another chat script. This is in preparation
for the abstraction of ``set dial'' and ``set hangup''.
1999-10-25 13:49:44 +00:00
brian
6d6b73fcc1 Don't (unnecessarily) parse wtmp, force ID0logout() to DTRT instead. 1999-10-25 13:31:42 +00:00
tg
d5152f15d2 Bring back texindex from the dead. Some ports use it to format their docs. 1999-10-25 10:35:39 +00:00
sheldonh
5b5ad21e78 Do not rely on malloc() to zero-fill allocated memory, since it is
not documented as doing so.  This fixes a potential segmentation
violation.

PR:	14465
Reported by:	Tony Finch <fanf@demon.net>
1999-10-25 09:46:57 +00:00
alc
fde4bba974 Add text for the AMD-751 host-to-PCI and PCI-to-PCI (AGP) bridges. 1999-10-25 06:28:04 +00:00
peter
6976901e20 If running on the client side of a remote commit, don't reject root when
trying to do a commit.  The server side will use the correct name, and the
client side restriction is just an annoyance.

Requested by:  lots of folks
1999-10-25 06:24:11 +00:00
imp
3304242c3e Have only one attach routine, rather than two.
Remove debug printfs.

This should allow the ed driver to work when memory mapping works again.
1999-10-25 06:15:45 +00:00
imp
10abd67792 MIN unused, so kill it 1999-10-25 04:28:53 +00:00
wes
e72875ec25 Don't say a fscking word, any one of you. 1999-10-25 04:23:21 +00:00
imp
e102eb6fff Sony CXD1847A FireWire Host Controller 1999-10-25 04:18:20 +00:00
obrien
ff295fd857 "S" comes before "U"...
Alphabet taught to Green by:	obrien
1999-10-25 03:51:01 +00:00
imp
4db6555364 Add pccard attachment. This won't work with memory mapped ed devices
yet, but that should be resolved shortly.  Non memory mapped ed
devices should work, but I cannot test this since my only ed card is
memory mapped.

Submitted by: Matt Dodd <mdodd@freebsd.org>
1999-10-25 02:55:14 +00:00
imp
6c561ab3d9 Add newbus attachment for pccard driver. Heavily tested by me, and
I'm committing this from a laptop running this driver.  Have only one
devclass for all ep devices (at least for pccard and eisa) so unit
numbering is sane.  Might not work with both isa and non-isa devices
on the same system until ep is updated (Matt has some patches in the
pipeline which should resolve this, he wanted me to commit this so he
can resolve any conflicts against cvs rather than my patches).

Reviewed by: Matt Dodd <mdodd@freebsd.org>
1999-10-25 02:52:16 +00:00
imp
94464bc92e Add newbus pccard attachment for sio. Some of this code was written
by Peter Wemm, but I've not merged all the changes he sent to me yet.
This has not been reviewed by bde, so I'm committing to resolve any
issues he has with this when he returns from FreeBSD CON 99.

I've had four reports of this working for them.  I've been able to
communicate to both my built in modem and a pccard modem with these
patches.
1999-10-25 02:46:52 +00:00
imp
b3ebae016a Massive rewrite of pccard to convert it to newbus.
o Gut the compatibility interface, you now must attach with newbus.
o Unit numbers from pccardd are now ignored.  This may change the units
  assigned to a card.  It now uses the first available unit.
o kill old skeleton code that is now obsolete.
o Use newbus attachment code.
o cleanup interfile dependencies some.
o kill list of devices per slot.  we use the device tree for what we need.
o Remove now obsolete code.
o The ep driver (and maybe ed) may need some config file tweaks to
  allow it to attach.  See config files that were committed for examples
  on how to do this.

Drivers to be commited shortly.

This is an interrum fix until the new pccard.  ed, ep and sio will be
supported by me with this release, although others are welcome to try
to support other devices before new pccard is working.

I plan on doing minimal further work on this code base.  Be careful
when upgrading, since this code is known to work on my laptop and
those of a couple others as well, but your milage may vary.

BUGS TO BE FIXED:

o system memory isn't allocated yet, it will be soon.
o No devices actually have a pccard newbus attach in the tree.

BUGS THAT MIGHT BE FIXED:

o card removal, including suspend, usually hangs the system.

Many thanks to Peter Wemm and Doug Rabson for helping me to fill in
the missing bits of New Bus understanding at FreeBSD Con '99.
1999-10-25 02:41:58 +00:00
alc
5c654c4743 Add text for the Athlon's MMX and 3DNow! (DSP) instruction extensions
to print_AMD_features.
1999-10-24 23:36:20 +00:00
nyan
0e060fbfb1 Fixed to compile a kernel with scbus0 and without da0.
Added $FreeBSD$.

Submitted by:   chi@bd.mbn.or.jp (Chiharu Shibata)
1999-10-24 14:54:12 +00:00
ache
490fc835e8 Remove UNSAFE_WARN ifdef for mktemp warning (never defined)
Use _mktemp internally
1999-10-24 11:57:24 +00:00
green
ad5952ca34 Remove a KASSERT() that has fulfilled its purpose. Note that it did
cause problems by tripping on shutdown (reboot(), not the socket
operation :).  Cause is still uncertain, but the panic isn't really
necessary here.
1999-10-24 08:37:21 +00:00
mdodd
3094f99fa7 Fix compiling PCCARD kernels.
This is a hack and I don't like it.

Hopefully the new PCCARD stuff should be in working order in a day or so.
1999-10-24 07:17:26 +00:00
wes
ec908db6f5 Moderate the "public" fortunes file somewhat; add the original to
the -offensive database.  Also provide some whimsy in the $FBSD$
headers.

Suggested by: The Kibbutz squad.
1999-10-24 06:39:58 +00:00
ache
a9f235047c Cosmetique: use standard prototypes scheme
Back out prev change: toupper is more compatible with sort -f
1999-10-24 04:47:57 +00:00
ache
d8f93ede38 Cosmetique: use standard prototypes scheme
Back out prev. change: toupper is more compatible with sort -f
1999-10-24 04:41:31 +00:00
ache
c870ad3624 toupper->tolower to match what strcasecmp does 1999-10-24 04:21:42 +00:00
ache
9fbc2f9470 Use strcoll to provide the same results as sort and comm
Use LINE_MAX for max line size (as comm does)
1999-10-24 04:08:15 +00:00
ache
c594629213 Localize it
PR:		11221
Submitted by:	Grigoriy Strokin <grg@philol.msu.ru>
1999-10-24 03:42:35 +00:00
dillon
16d1a2d1fd Adjust the buffer cache to better handle small-memory machines. A
slightly older version of this code was tested by BDE and I.

    Also fixes a lockup situation when kva gets too fragmented.

    Remove the maxvmiobufspace variable and sysctl, they are no longer
    used.  Also cleanup (remove) #if 0 sections from prior commits.

    This code is more of a hack, but presumably the whole buffer cache
    implementation is going to be rewritten in the next year so it's no
    big deal.
1999-10-24 03:27:28 +00:00
dillon
00d0ab1ce1 Put tty back in, problem was due to cvsup mess on freefall. 1999-10-24 03:01:41 +00:00
dillon
8c8443ab96 Put include <netgraph.h> back in - problem was due to a cvsup mess on
freefall.  There may also be a real problem w/ buildworld but the fix
    would go somewhere else, not here.
1999-10-24 02:58:39 +00:00
nsayer
cbd46ae245 Add commented entry to the lo0 section inviting bridge users to
enable ARP on filtering bridges.
1999-10-24 00:26:49 +00:00
julian
48db5a679e more fixes, braino's, typo's, etc. 1999-10-23 22:46:38 +00:00
dillon
b08edd3233 Temporarily remove 'tty' netgraph sub module until it can be made
to compile.
1999-10-23 18:53:46 +00:00
chris
805cb5a69e Fix a typo: `maintain compatible'' -> `maintain compatibility'' 1999-10-23 18:15:14 +00:00
dillon
405fc1165a Cleanup missing includes, stale includes, and a few printf format
inconsistancies.
1999-10-23 17:45:11 +00:00
green
b6ebb73004 Add the -F option. This is for forcing restarts with -r by inhibiting
transmission of the If-Range HTTP header field.
1999-10-23 15:22:53 +00:00
julian
732f8d1ced A version of the pppoe code ellicits a response from the ISP end
(but still not quite right)
1999-10-23 15:15:42 +00:00
julian
41ef3662ca dang, deleted a line 1999-10-23 04:52:54 +00:00
julian
bd67127489 Now that Netgraph is in the system there are some cleanups we can do.
Also save a slightly closer to completion version of the PPPOE code.

Submitted by: Archie Cobbs <archie@freebsd.org>
1999-10-23 04:28:11 +00:00
green
309540210a Small bugfixes (point not getting marked in one case, string not NUL and
wrong size in the other.)

Submitted by:	Martin Blapp <mb@imp.ch>
1999-10-23 00:54:58 +00:00
msmith
1e1a86c642 Use the physical block number, not the logical block number, for I/O
operations.  The latter only works where the partion begins at the bottom
of the disk.  Whoops.

Submitted by:	Chris Csanady <cc@137.org>
1999-10-22 20:55:15 +00:00