Warner Losh [Mon, 20 Dec 1999 06:47:38 +0000 (06:47 +0000)]
Make attach work, almost. Calling kthread_create from a
config_intrhook_establish doesn't work. Children aren't yet attached
properly, but that's ok because pccard would likely panic in its
current shape.
o Save dev of pcic early in attach process
o save dev in pcic_handle for use in pcic_create_event_thread.
o Remove direct attachments of pccard children for now
o move establishment of pcic kthread to config_intrhook.
Greg Lehey [Mon, 20 Dec 1999 02:16:55 +0000 (02:16 +0000)]
Install the crunched binaries into /sbin, not /stand. This was
originally done to track down yet another case of lost init, and is
not strictly necessary, but it seems more logical to have binaries in
/sbin than in /stand. Previously /sbin and /bin were symlinks to
/stand. Now /bin and /stand are symlinks to /sbin.
Greg Lehey [Mon, 20 Dec 1999 02:15:04 +0000 (02:15 +0000)]
Add content. This was one of the files that somehow got checked in
empty. An empty loader.rc will enable the kernel to boot, but it
won't find init (because the MFS file system hasn't been loaded).
Greg Lehey [Mon, 20 Dec 1999 02:12:27 +0000 (02:12 +0000)]
Define RELEASE_CRUNCH as a make variable. Previously it was defined
as a preprocessor variable only. This broke the build of ppp. This
problem still exists in the old-style directories.
Greg Lehey [Mon, 20 Dec 1999 02:10:16 +0000 (02:10 +0000)]
Install the crunched binaries into /sbin, not /stand. This was
originally done to track down yet another case of lost init, and is
not strictly necessary, but it seems more logical to have binaries in
/sbin than in /stand. Previously /sbin and /bin were symlinks to
/stand. Now /bin and /stand are symlinks to /sbin.
Greg Lehey [Mon, 20 Dec 1999 02:07:42 +0000 (02:07 +0000)]
Handle building individual targets better. It's still broken; the
only way to be sure the build works correctly is to do a 'make all'.
But with these changes, it's easier to test individual targets. In
particular, ensure that the vnode file systems are mounted before
writing to them.
Ensure that we don't get CVS directories on our floppies.
Use kgzip instead of kzip to compress the loader. This saves a few
kB.
David E. O'Brien [Sun, 19 Dec 1999 22:05:32 +0000 (22:05 +0000)]
Install sendmail.cf into /etc/mail.
The current offical Sendmail Inc. version uses /var/mail/ and when we upgrade
our repository to that version, we will get the change. It is best to make
the path change in 4.0-R (which may not have the latest Sendmail Inc. version,
than to change in mid-4.x stream when we may upgrade.
Bill Fumerola [Sun, 19 Dec 1999 20:33:05 +0000 (20:33 +0000)]
Borrow phk's axe and chop off the old soundcard-CDROM devices. We get
about 40k of savings from this, and these abominations are still in LINT
if anyone needs to use them.
David E. O'Brien [Sun, 19 Dec 1999 19:55:19 +0000 (19:55 +0000)]
Install sendmail.cf into /etc/mail.
The current offical Sendmail Inc. version uses /var/mail/ and when we upgrade
our repository to that version, we will get the change. It is best to make
the path change in 4.0-R (which may not have the latest Sendmail Inc. version,
than to change in mid-4.x stream when we may upgrade.
Ollivier Robert [Sun, 19 Dec 1999 15:43:19 +0000 (15:43 +0000)]
Second part of bin/3648: add -flags to search for specific flags.
I added $FreeBSD$ whicle I was here. The patch wasn't usable anymore
due to its age so I adapted it.
PR: bin/3648
Submitted by: Martin Birgmeier <mbirg@austria.ds.philips.com>
Cameron Grant [Sun, 19 Dec 1999 15:27:24 +0000 (15:27 +0000)]
move make_dev operations for audio channels to pcm_addchan(). in theory,
with modifications to MAKEDEV this will allow use of multiple output streams
on cards supporting it, eg trident 4dwave.
Brian Feldman [Sun, 19 Dec 1999 09:42:15 +0000 (09:42 +0000)]
Remove all vestiges of /usr programs (awk, sed, grep...) from MAKEDEV.
MAKEDEV can now run (with /sbin/chown and /bin/chgrp, still) without
any /usr. The default PATH for MAKEDEV is now "/sbin:/bin" to prevent
against further spammage of /usr programs in MAKEDEV. In the event
of an emergency, the variable "MAKEDEVPATH" can be used to replace
PATH, but I don't know what that emergency could be.
Let me know if you have any problems. I've tested it as well as
I could, which basically means everything except for RocketPorts.
Warner Losh [Sun, 19 Dec 1999 07:39:10 +0000 (07:39 +0000)]
o Hide pcic_chip functions inside file
o Attach children in a better way than before.
o Remove some dead code
o collapse the socket attach routines into pcic_attach
Robert Watson [Sun, 19 Dec 1999 06:08:07 +0000 (06:08 +0000)]
Second pass commit to introduce new ACL and Extended Attribute system
calls, vnops, vfsops, both in /kern, and to individual file systems that
require a vfsop_ array entry.
Robert Watson [Sun, 19 Dec 1999 05:54:46 +0000 (05:54 +0000)]
First pass commit to introduce new ACL and Extended Attribute system calls.
The second pass commit with all the supporting code will happen shortly
afterwards.
Eivind Eklund [Sat, 18 Dec 1999 23:01:52 +0000 (23:01 +0000)]
Since VOP_LOCK can be used to up and downgrade locks, it is not possible
to say anything about the lockstate before and after it. Thus, change the
lockspec from U L U to ? ? ?.
Eivind Eklund [Sat, 18 Dec 1999 20:50:48 +0000 (20:50 +0000)]
Correct the documentation of LOCKPARENT - it does not reliably result in the
parent being locked, but rather plays some hide and seek (does not lock if
dvp == vp).
Also add a BUGS section noting that this is undesired behaviour.
Eivind Eklund [Sat, 18 Dec 1999 17:12:15 +0000 (17:12 +0000)]
Make NDFREE unlock VP if appropriate and not asked not to, as documented in
the manpage for namei(). This capability is not yet used attempted used
anywhere in the source tree.
Peter Wemm [Sat, 18 Dec 1999 13:55:17 +0000 (13:55 +0000)]
Revert the libcrypt/libmd stuff back to how it was. This should not have
happened as it was working around problems elsewhere (ie: binutils/ld
not doing the right thing according to the ELF design). libcrypt has
been adjusted to not need the runtime -lmd. It's still not quite right
(ld is supposed to work damnit) but at least it doesn't impact all the
users of libcrypt in Marcel's cross-build model.
Brian Feldman [Sat, 18 Dec 1999 13:52:44 +0000 (13:52 +0000)]
I've been Bruceified! As penance, I've fixed my error (improper ordering,
moving #ifdefs to the non-historically-correct place, indentation bugs),
and I've also completely restyle(9)d mbuf.h. There shouldn't be any
KNF issues with it anymore, nor should there be any internal
discrepencies.
(Oh, I also fixed a small bug, but it was benign. The spl was not
released in MCLALLOC if M_NOWAIT and there were no free mclusters.)
Warner Losh [Sat, 18 Dec 1999 06:11:22 +0000 (06:11 +0000)]
Driver for the smc91xx series of ethernet chips. Ported from PAO to
3.3R and then to -current. The pccard support has been left in the
driver, but is presently non-functional because we are using the
isa_compat layer for the moment.
Yet more cosmetic fixes (I have a little time to kill while waiting for
something else, so I might as well tweak 3.4's look-and-feel for the better
while I'm at it - final release build should be sometime later on tonite).
Peter Wemm [Fri, 17 Dec 1999 22:07:43 +0000 (22:07 +0000)]
Be more consistant in using perl vs. perl5. We were using perl5 in the
kernel builds so as not to confuse with perl4 when bootstrapping from old
systems. I don't know if this is still applicable but it shouldn't hurt
to be consistant at least.
Also copy vnode_if.sh to vnode_if.pl. Doing a 'sh vnode_if.sh' when it
was a perl script was kinda silly.
Peter Wemm [Fri, 17 Dec 1999 20:04:01 +0000 (20:04 +0000)]
Remove -lmd. Use dlopen() and dlsym() instead for calls to the MD5* and
SHA* routines so that callers of libcrypt are not exposed to the internal
implementation.