Kazutaka YOKOTA [Thu, 23 Oct 1997 03:23:50 +0000 (03:23 +0000)]
Reject unreasonable values passed to CONS_HISTORY ioctl. It did not
check the value and caused kernel panic when a large value was given.
- Move the configuration option SC_HISTORY_SIZE from syscons.h to
syscons.c.
- Define the maximum total number of history lines of all consoles.
It is SC_HISTORY_SIZE*MAXCONS or 1000*MAXCONS; whichever is larger.
CONS_HISTORY will allow the user to set the history size up to
SC_HISTORY_SIZE unconditionally (or the current height of the console
if it is larger than SC_HISTORY_SIZE). If the user requests a larger
buffer, it will be granted only if the total number of all allocated
history lines and the requested number of lines won't exceed the maximum.
- Don't free the previous history buffer and leave the history buffer
pointer holding a invalid pointer. Set the pointer to NULL first, then
free the buffer.
Nate Williams [Thu, 23 Oct 1997 02:30:39 +0000 (02:30 +0000)]
- When doing suspend/resume, only try to suspend those drivers on active
slots. Otherwise, we try to suspend drivers who have been disabled
already.
[
The only reason the drivers are still on the list is because of race
conditions where the card is removed while the driver is in use. We
leave the drivers on the slot list (leaving all of their structures in
place in case a process is using it) but set it's state to empty so that
further uses by the pccard code know not to expect active cards.
]
Nate Williams [Thu, 23 Oct 1997 02:25:14 +0000 (02:25 +0000)]
Back our V1.36, due to reports of badness. (IRQ's are again allocated
from the 'lower' interrupts to the 'higher' interrupts.) We need to
find a way to set the interrupt for the controller in the config file.
Changes in spirit of OpenGroup Singe Unix specs:
1) Limit max allowed argument to 1000000
2) Change return type from void to int to indicate premature termination
(by signal)
Peter Wemm [Wed, 22 Oct 1997 00:48:36 +0000 (00:48 +0000)]
Put in an initial %VERSREQ field. This will cause existing config(8)'s
to give a non-fatal warning about unknown directives, so there is no
hurry to rebuild config(8) yet, apart from shutting up the warning.
Peter Wemm [Wed, 22 Oct 1997 00:44:08 +0000 (00:44 +0000)]
For safety's sake, explicitly depend all objects on the configvers.h header
to make sure that it's all recompiled even if there is no 'make depend'.
This is overkill, but should be one less thing that someone can do wrong.
Peter Wemm [Wed, 22 Oct 1997 00:38:48 +0000 (00:38 +0000)]
Recognize a %VERSREQ=nnnnn string in the system Makefile. Both config(8)
and the kernel will have a 'config interface version number'. If an
incompatable change is made to the kernel that requires a rebuild of
config(8) (such as the cam devtab stuff), then the version number would be
bumped in both places. If a user neglects to rebuild config, then they
will get a nagging (but non-fatal) warning that they need to rebuild
config.
John-Mark Gurney [Tue, 21 Oct 1997 09:59:26 +0000 (09:59 +0000)]
a few code style changes:
o start function names in column 1
o sort order of flags in getopt and switch
o don't try to reference progname
o unspam some changes introduced by a 2.2.1-R build box instead of a
-current build box
doc changes:
o document when these commands first appeared
o put email address in angle brakets
o minor mdoc clean up
Mike Smith [Tue, 21 Oct 1997 07:40:22 +0000 (07:40 +0000)]
Reference the DMI table inside the SMBIOS table correctly, not using a variable
that won't be initialised until a later test.
Submitted by: bde via -Wunused
Bruce Evans [Tue, 21 Oct 1997 07:26:50 +0000 (07:26 +0000)]
Don't install mount_msdos setuid root. Lite2's mount(2) handles
permissions centrally and a setuid root mount utility just breaks
its security. There was no new breakage in practice because
mfdosfs_mount() still checks the ruid.
Bruce Evans [Tue, 21 Oct 1997 06:53:24 +0000 (06:53 +0000)]
Added -Wuninitialized to CWARNFLAGS. Warnings about uninitialialized
variables were lost when we removed -W, and 23 new ones including at
least one serious one have crept in for LINT.
Restored -Winline to CFLAGS. This gives only 3 old warnings and 1 new
for LINT.
David Greenman [Sun, 19 Oct 1997 18:41:23 +0000 (18:41 +0000)]
"Fixed" pstat -T by avoiding the vnode stats. Disabled pstat -v since
we no longer support that sysctl (in my opinion, pstat -v is a security
hole in any case).
Joerg Wunsch [Sun, 19 Oct 1997 16:48:50 +0000 (16:48 +0000)]
Declare the various 2-byte values in TFTP as being `unsigned short'.
RFC 1350 isn't very explicit about this, but the use of signed
integers doesn't seem to make sense.
Should be in 2.2.5, but is probably too late to be verified completely.
PR: bin/4502
Submitted by: Dirk Nehrling <nerle@pdv.de> via Lars Koeller
Joerg Wunsch [Sun, 19 Oct 1997 16:40:01 +0000 (16:40 +0000)]
Introduce a -N option that disables the use of reserved ports, now
that -P is on by default. Remove do-nothing code in the -P case (but
leave the option itself for backward compatibility).
Joerg Wunsch [Sun, 19 Oct 1997 13:12:02 +0000 (13:12 +0000)]
Introduce a device flags value of 0x1 to always pretend a 1.44 MB
floppy drive #0, regardless of what the CMOS says. This is intended
as a bandaid for those plagued with Compaq's idea to not announce the
floppy drive on their `Aero' notebook.
Using the device flags is not very nice (in particular since they
aren't per-drive but per-controller), but still looks a lot better to
me than the disgusting guesswork hack that was recently posted to
-hackers.
Peter Wemm [Sat, 18 Oct 1997 18:15:03 +0000 (18:15 +0000)]
Merge in changes from Matt. Accton EN12xx support, bugfixes. There is
a change that might have an effect on the problems some have seen
with older chips, it looks like the driver may have mistakenly thought
there was an SIA when there isn't.
Bruce Evans [Sat, 18 Oct 1997 13:59:48 +0000 (13:59 +0000)]
Removed unused file. It just forces a return value of 0 on success
(no carry), but mount() in the kernel has returned 0 on success since
prehistoric times.
Peter Wemm [Sat, 18 Oct 1997 13:23:35 +0000 (13:23 +0000)]
Import Matt's current de driver... He appears to be maintaining it in
the NetBSD source tree now. This is slightly newer than the updated
.tar.gz on the 3am-software web site.
Peter Wemm [Sat, 18 Oct 1997 13:23:35 +0000 (13:23 +0000)]
Import Matt's current de driver... He appears to be maintaining it in
the NetBSD source tree now. This is slightly newer than the updated
.tar.gz on the 3am-software web site.
Joerg Wunsch [Sat, 18 Oct 1997 10:59:47 +0000 (10:59 +0000)]
Make all the documented (in pcvt(4)) options supported options. While
i was at it, do no longer insist on `PCVT_FREEBSD' being declared in
the config file, but default it to a reasonable value.
More cleanup to follow, but this part is safe for RELENG_2_2, too.
Joerg Wunsch [Sat, 18 Oct 1997 10:40:02 +0000 (10:40 +0000)]
Something that should have been done 2.5 years ago: install pcvt's
accompanying doc files into /usr/share/pcvt/. (This goes to the `doc'
distribution, not `bin'.)
- do string-magic on the right variable when trying
to find dependent packages on the local disk (i.e. package given as
filename, not URL)
- document $PKG_PATH
- use -v on pkg_add for dependent packages
Peter Wemm [Sat, 18 Oct 1997 01:37:52 +0000 (01:37 +0000)]
Update Makefile to know about PPP_FILTER and fix the PPP_BSDCOMP and
PPP_DEFLATE options... The code uses #if defined(PPP_DEFLATE) etc, so the
original method of "#define PPP_DEFLATE 0" did not actually disable
anything. It was not possible to configure out bsdcomp or zlib
compression. Also, join the settings for PPP_FILTER and NBPFILTER together
since they are related here (can't have PPP_FILTER without NBPFILTER).
Peter Wemm [Sat, 18 Oct 1997 01:29:18 +0000 (01:29 +0000)]
Don't say that the filtering options are only available under NetBSD, and
don't claim that _any_ serial speed is settable, mention that it has to
be supported by the device driver. Some have very flexible baud rate
selection (ie: based on divisors etc, such as the com ports), while others
have explicit baud rate choices.
Peter Wemm [Sat, 18 Oct 1997 00:56:23 +0000 (00:56 +0000)]
Better fix for the bpf dependency that doesn't have such a large impact
on the code and pppd in userland. PPP_FILTER is meant to be an option (or
negatable option).