we're at it, eliminate obsolete exposure of `struct llinfo_arp' to
the world. (This dates back to when ARP entries were not stored in
the routing table, and there was no other way for the `arp' program
to read the whole table than to grovel around in /dev/kmem.)
match the board setting.
2) Fixed the warning message to properly print the irq mismatch (as opposed
to the bit encoded value).
3) Add irq autodetection if the kernel has "?" or no irq specified.
4) Add missing splimp protection in ixioctl().
functions. The application header <sys/user.h> includes <vm/vm.h>
which includes <vm/lock.h>...
vm.h:
Don't include <machine/cpufunc.h>. It is already included by
<sys/systm.h> in the kernel and isn't designed to be included by
applications (the 2.1 version causes a syntax error in C++ and the
current version has initializers that are invalid in strict C++).
lock.h:
Only declare kernel functions if KERNEL is defined.
for the same reasons as in slclose().
Free the cblock in the canonical queue in pppclose(). This is a no-op in
the usual cases where the tty is being closed or the line discipline is
being switched back to the standard discipline, but it saves a cblock if
the line discipline is being switched to one that doesn't use the canonical
queue.
Add prototypes. I use `extern' in prototypes for functions with bogus
linkage. This should be fixed someday.
Continue cleaning up new init stuff yet again.
and here is the patch. Submit it please. Thank you!
BTW, there is a new option "NDGBPORTS". By default it is equal to
NDGB*16 and means the number of ports of all Digiboards for which the
tty structures are reserved. It can be set to the real value in config-file
like:
options "NDGBPORTS=8"
Submitted by: Serge A. Babkin <babkin@hq.icb.chel.su>
to be no ill effects, and so far as Iknow none of the variables in
question depend on 16-bit wraparound behavior. (The sizes are in
many cases relics from when a PCB had to fit inside a 128-byte mbuf. PCBs
are no longer stored in that way, and the old structure would not have
fit, either.)
it was referenced from.. stops a rather annoying panic, but
introduces a rather interesting but "I can live with it" bug
`ln a b ; mv a b; echo ?`
returns a rather than b..
I know why but I need to think of the 'correct' answer. at least this is 'safe'
Obtained from: other people on the net ?
1. stepping over syscalls (gdb ni) sends you to DDB, and returned
to the wrong address afterwards, with or without DDB. patch in
i386/i386/trap.c below.
2. the linux emulator (modload'ed) still causes panics with DIAGNOSTIC,
re-applied a patch posted to one of the lists...
for testing.. (for cdevsw_add and bdevsw_add)
make the lkm code use the new generic devsw add routines (if so required)
define these routines in conf.h so we can use them
This is really just a token gesture at this point. We can't really turn
on swapping here anymore because (I think) we need to do a namei() lookup
in order to get a vnode which we can pass to swaponvp(). But to do a
namei(), we need a properly initialized proc structure, and we don't
have one. (I was hopping to toss it proc0, but that isn't initialized
until later.) We could conceiveably fabricate a dummy proc structure
to make namei() happy, but that's too much work for too little payoff.
For now, swapgeneric's only remaining saving grace is that it lets
you set the root device with -a.
This is a place for all things to do with conf.c and conf.h
that are not machine specific.
Other things that are at present in i386/isa/conf.c might
migrate into here..
It's the first small step in cleaning up the device interface
to make it more dynamic and to assist in more modular drivers
(i.e. both loadable via LKMs and linked in..
e.g able to add a device without having to edit conf.c)
this code is not yet used and the whole thing will be conditionally
compiled in for a while till proven useful :)
matching IP options..Check and test this - i made only a couple
of rough tests and this could be buggy.. Ipaccounting can't use
IP Options (and i don't see any need to cound packets with specific
options either..)
More to come...
1542 revs work by assuming the next few sequential ID codes are
new Adaptec boards and enabling them after printing a warning.
Conditionalize the informational boot messages with "if (bootverbose)".
1) "obj" was't initialized properly, resulting in an important vm_page_lookup
always failing (resulting in a panic).
2) busy pages could be put on the cache queue or freed (resulting in a panic).
were making and were no-brainers. However, the xcdplayer
eject button didn't work because xcdplayer gave up and didn't bother
to try the EJECT ioctl anyway when CDIOCALLOW was not available.
This all works now.
o A change so that xcdplayer gets TOC entry data for the lead-out area
on the disc. Xcdplayer can now play the last track on a CD, which
it would not do in earlier versions (but no one reported).
Cdplayer gets the TOC data differently and it works in old and new
versions.
o Eliminated a race condition that caused the driver to sleep forever
on very slow and heavily loaded systems on rare occasions when
the system was doing lots of audio-related ioctls to the drive.
o Fixed a problem where a locked drive could be unlocked by accessing
one of the non-locking devs. Door locking now follows the documented
rules.
o Made all wait channel strings unique to matcd.
Submitted by: Frank Durda IV <uhclem%nemesis@fw.ast.com>
scheduled for demolition. This is a first step: get rid of if_zereg.h,
by adding the five extra definitions to if_edreg.h.
Also add some definitions which will become needed when if_ze.c gets
replaced entirely by pccard and if_ed.c. (this is a 2.1.0 candidate)
is swapped in. Also, remove unnecessary map locking/unlocking during
selection of processes to be swapped out.
This code might afford proper panics as opposed to spontaneous reboots
on certain systems. This should allow us to debug these problems better.
David Greenman, it has been determined that the more sophisticated code
only made a very minor difference in fault performance. Therefore, this
code eliminates some of the complication of the fault code, decreasing
the amount of CPU used to scan shadow chains.
support for EXT2FS. Note that the Sig-11 problems appear to be caused by
this, but there is still probably an underlying VM problem that let this
clustering bug cause vnode objects to appear to be corrupted.
The direct manifestation of this bug would have been severely mis-read
files. It is possible that processes would Sig-11 on very damaged
input files and might explain the mysterious differences in system
behaviour when phk's malloc is being used.
