Fight with false newlines in kernel message logs. Output a line into log only
after we read a newline, or we have nothing to read from /dev/klog. Read
/dev/klog in non-blocking mode.
markm [Sun, 2 May 1999 08:55:27 +0000 (08:55 +0000)]
This routine was "use"ing File::Basename. This commit removes that
"use" and replaces it with equivalent inline code. The reason is that
Perl has some very nasty circular dependancies, and I am trying to
get the System Perl upgraded by one maintenance level.
The basic rule, until I can find a way to solve this, is that
the build tools MAY NOT use any library code; it must all be inline.
grog [Sun, 2 May 1999 08:02:03 +0000 (08:02 +0000)]
printconfig:
split into vinum_printconfig (interactive) and printconfig (called
from vinum_printconfig and from vinum_create when starting an
interactive update.
grog [Sun, 2 May 1999 08:00:30 +0000 (08:00 +0000)]
Maintain a log file, by default /var/log/vinum.history, showing what
has been done. This name can be overridden by the value of the
VINUM_HISTORY environment variable.
Print dates in log file according to the variable VINUM_DATEFORMAT, by
default %e %b %Y %H:%M:%S.
grog [Sun, 2 May 1999 07:51:20 +0000 (07:51 +0000)]
close_drive:
If a drive has gone down and has dirty buffers associated with it,
we'll get a panic when we try to vn_close it. Check for this
situation and discard any buffers; they're toast anyway.
Only complain about usage count if DEBUG_WARNINGS is set.
check_drive:
Change parameter name from drivename to devicename.
Get the check for a referenced drive right.
If the partition isn't a vinum drive, set the last error to ENODEV.
vinum_scandisk:
Change parameter name from drivename [] to devicename [].
obrien [Sun, 2 May 1999 00:18:02 +0000 (00:18 +0000)]
Remove NOPERL. Seems to cause too many problems, as those that try don't
understand the implications and don't think to undo it when they get experience
trouble. (wonder about some of the people running -CURRENT sometimes)
n_hibma [Sat, 1 May 1999 23:30:09 +0000 (23:30 +0000)]
Move the disabling of interrupts right before the allocation of the
resource. Avoids useless interrupts occurring between the allocation
of the interrupt resource and the final initialisation of the
kernel. Cause of these interrupts is unknown (a resuming device?).
obrien [Sat, 1 May 1999 22:22:42 +0000 (22:22 +0000)]
JMZ says:
Add an emply MAIN__() function. This avoid an unresolved reference error
during link phase when using fortran subroutines with non-fortran 'main()'.
Obtained from: src/lib/libF77/MAIN.c rev 1.1 by jmz
kris [Sat, 1 May 1999 14:54:21 +0000 (14:54 +0000)]
Fixes for -t option:
* Don't buffer "Digesting..." output - flush it immediately.
* Increase the number of repetitions by a factor of 100, and the block size by
a factor of 10 so as to give meaningful results on modern machines (108
seconds on my P120, and ~26 seconds on a P-II 350).
PR: bin/10604
Submitted by: Stanislav Shalunov <shalunov@lynxhub.lz.att.com>
Turn DEFAULT_VTABLE_THUNKS back on. It caused problems with undefined symbols
in libstdc++.
Until I have a chance to look at what that problem is and to carefully consider
the upgrade issues of turning it back on at a later date if we leave it turned
off for any extended peroid of time.
Change references from "passwordperiod" to "passwordtime", since
"passwordtime" is what passwd(1) has actually been using. I suspect
passwordperiod was the original intent. I can't figure-out which,
if either, BSDi uses. If anyone knows...
Close PR #10970: exit at the end of ypproc_all_2_svc() instead of returning.
This really fixes the condition where a child creates children of its own.
I'm leaving the previous sanity tests in though, since they shouldn't hurt,
and will give an indication if this ever happens again.
- Try to unbreak what I broke by screwing with the tx queuing again.
I'm waiting for a few more people to test out this code and report back
before I move it into current. Hopefully it will be soon. Basically I
reverted to the old TX queuing strategy.
- Add experimental support for the 3c900B-FL (10mbps ST fiber). The card
should be detected properly and the 10baseFL mode supported, but again
I'm still waiting for word from a tester to see if this actually works.
It shouldn't affect the other cards though; all the differences are in
media selection.
- Set the TX start threshold register to get better performance.
- Increase the size of the RX and TX rings. UDP performance was pretty
bad because the TX ring was too small. Should be substantially better
now (I can saturate the link with either TCP or UDP now).
- Change some of the #defines to reflect proper 3Com ASIC names (boomerang,
cyclone, krakatoa, hurricane).
- Simplify and reorganize interrupt handler; ack all interrupts right
away and then process them. This avoids a potential race condition.
(Noted by Matt Dillon.)
- Reorganize the bridging code to eliminate using a goto to jump into
the middle of an if() {} clause. Sorry, that just made my brain itch.
- Use m_adj() in xl_rxeof().
- Make the payload alignment in xl_newbuf() the default (instead of
just conditionally defined for the alpha) to improve NFS performance
(avoids need for nfs_realign()).
Tcopy uses 32 bit unsigned to accumulate a count of bytes read/written.
That doesn't work well for tapes over 4G.
I use tcopy a lot to write images of a tape to tape as tape to tape
copying is terribly slow. Slower than it should be. Quickly found out
tcopy can not rewind a file when doing copy/verify.
PR: 11386
Submitted by: David Kelly dkelly@hiwaay.net
Reviewed by: phk
This patch also moves the bogus comment (the comment is still not quite
right) and (as a side effect) removes some verbose initialisations (we
depend on static initialisation to 0 for almost everything in proc0).
The alpha kernels are bootable again. The change won't affect i386's
until machdep.c is changed.
Add some extra checks to make sure that a child ypserv process doesn't
try to fork() a child of its own, which could result in several children
ypservs running at once. I'm still not sure exactly what leads to this
condition, but these fixes should stop it from causing trouble. A new
function, yp_fork() checks to see if the current process is already a
child of the parent ypserv, and returns failure (and logs an error message)
rather than spawning another child.
Address a performance problem in getnewbuf:
In heavy-writing situations, QUEUE_LRU can contain a large number
of DELWRI buffers at its head. These buffers must be moved
to the tail if they cannot be written async in order to reduce
the scanning time required to skip past these buffers in later
getnewbuf() calls.
Submitted by: Matthew Dillon <dillon@apollo.backplane.com>
release/Makefile:
Build a separate NOCRYPT version of pppd, to keep This Great
Nation's top-secret cryptographic tools out of the filthy hands
of those evil furriners.
Add a test to ti_encap() to try and prevent the transmit producer index
from ever catching up to the transmit consumer index. We can't let this
happen because ti_txeof() depends on the assumption that producer == consumer
means the ring is empty, and producer != consumer means the ring has some
number of active descriptors in it.
roger [Thu, 29 Apr 1999 10:15:55 +0000 (10:15 +0000)]
Added initial code for VBI capture based on work by
Hiroki Mori <mori@infocity.co.jp> and reworked by myself.
This allows software decoding of teletext, intercast and
subtitles via /dev/vbi.
roger [Thu, 29 Apr 1999 05:48:32 +0000 (05:48 +0000)]
Fix crashes caused by rows=0 or columns=0.
Add new #ifdef. By defining BKTR_NO_MSP_RESET you can prevent the
MSP34xx being reset by the bt848 driver. This is handy
if you pre-initialise the MSP34xx stereo audio chip in another
operating system first (eg MS Windows).
Suggested by: Randal Hopper<aa8vb@ipass.net>
Suggested by: Yuri Gindin <yuri@xpert.com>
- Handle mixer read ioctls correctly. They have the same group, number and
argument size as their write counterparts and were handled as write ioctls.
- Emulate some cdrom ioctls.
This is a seriously beefed up chroot kind of thing. The process
is jailed along the same lines as a chroot does it, but with
additional tough restrictions imposed on what the superuser can do.
For all I know, it is safe to hand over the root bit inside a
prison to the customer living in that prison, this is what
it was developed for in fact: "real virtual servers".
Each prison has an ip number associated with it, which all IP
communications will be coerced to use and each prison has its own
hostname.
Needless to say, you need more RAM this way, but the advantage is
that each customer can run their own particular version of apache
and not stomp on the toes of their neighbors.
It generally does what one would expect, but setting up a jail
still takes a little knowledge.
A few notes:
I have no scripts for setting up a jail, don't ask me for them.
The IP number should be an alias on one of the interfaces.
mount a /proc in each jail, it will make ps more useable.
/proc/<pid>/status tells the hostname of the prison for
jailed processes.
Quotas are only sensible if you have a mountpoint per prison.
There are no privisions for stopping resource-hogging.
Some "#ifdef INET" and similar may be missing (send patches!)
If somebody wants to take it from here and develop it into
more of a "virtual machine" they should be most welcome!
Tools, comments, patches & documentation most welcome.
Have fun...
Sponsored by: http://www.rndassociates.com/
Run for almost a year by: http://www.servetheweb.com/
projects / FreeBSD / FreeBSD.git / log