julian [Mon, 14 Dec 1998 21:14:11 +0000 (21:14 +0000)]
The OHCI interfaces I have access to map their control regs etc. into
memory address space rather than IO space.. reflect this when looking for the
interface revision register.
If this is not true for them all then we probably need some smarter code.
brian [Mon, 14 Dec 1998 19:24:30 +0000 (19:24 +0000)]
Rather than interrupting 10 times per second then checking
to see if there's anything to do, schedule the next alarm
based on the next required timeout.
This decreases the load when there are lots of relatively
idle ppp processes.
While I'm in there, handle the possibility that a timeout
makes the timer element go out of scope by grabbing the
enext pointer before executing the timer function.
bde [Mon, 14 Dec 1998 16:29:58 +0000 (16:29 +0000)]
Fixed LINT breakage in previous commit. Option FDC_YE enabled a
syntax error. Options FDC_YE and DEVFS together enabled references
to a nonexistent variable and calls of a nonexistent function.
mckay [Mon, 14 Dec 1998 13:27:48 +0000 (13:27 +0000)]
Restore traditional behaviour by allowing tabs in strings. In deference
to the new age of internationalisation, I used isblank() instead of a
literal '\t'.
dillon [Mon, 14 Dec 1998 08:34:55 +0000 (08:34 +0000)]
Fixed problems with kernel config file overrides of sysv semaphore
parameters. Prior to this fix a kernel config override would effect
only some of the kernel files, resulting in panics.
dillon [Mon, 14 Dec 1998 05:47:29 +0000 (05:47 +0000)]
pci_device pd_probe function changed from returning char * to returning
const char *. Originally I was going to add casts from const char * to
char * in some of the pci device drivers, but the reality is that the
pci device probes return constant quoted strings.
dillon [Mon, 14 Dec 1998 05:37:37 +0000 (05:37 +0000)]
Fix -Wuninitialized warning regarding zero-length var-args ctl element.
( this isn't really an error, but I think it is important to fix the
warning ).
dillon [Mon, 14 Dec 1998 05:35:56 +0000 (05:35 +0000)]
author was assuming that nextpaddr declared *inside* the do loop would
survive within the loop. This is not guarenteed by C. I have moved
the nextpaddr declaration to outside the do loop.
dillon [Mon, 14 Dec 1998 02:25:32 +0000 (02:25 +0000)]
Reviewed by: freebsd-current
Add bounds checking to netbios NS packet resolving code. This should
prevent natd from crashing on badly formed netbios packets (as might be
heard when the machine is sitting on a cable modem or certain DSL
networks), and also closes potential security holes that might have
exploited the lack of bounds checking in the previous version of the
code.
n_hibma [Sun, 13 Dec 1998 22:27:42 +0000 (22:27 +0000)]
Updated USB kernel sources to NetBSD sources of 1998-12-09.
1 bug fix and several textual changes.
Preparing to feed back changes for port into NetBSD to create one source base.
truckman [Sun, 13 Dec 1998 07:07:51 +0000 (07:07 +0000)]
getpgid() and getsid() were doing a comparision rather than an assignment,
which is fortunate, because otherwise another bug would allow them to be
used to stomp on the syscall return value of another process.
dillon [Sun, 13 Dec 1998 06:40:18 +0000 (06:40 +0000)]
PR: bin/9016
Fix bug with od/hd/hexdump. "*" lines are supposed to indicate one or
duplicates of the previous line, but a small file with less then 16
characters of zeros in it will be falsy identified as a repeat of
the (non-existant) previous line. i.e. the first line of output winds
up being a "*". Added a bit of code to handle the degenerate 'there is
no previous line' case for the first line.
dillon [Sun, 13 Dec 1998 06:07:38 +0000 (06:07 +0000)]
PR: bin/5572
Prevent cron from going crazy if the time steps. For example, if you
have a system with hundreds of users and lots of different crontabs
and your time steps back an hour, the old cron would then attempt to
run an hours worth of cron jobs in a few seconds.
dillon [Sun, 13 Dec 1998 03:42:51 +0000 (03:42 +0000)]
PR: bin/8680
Increase the size of a number of buffers and replace strcpy/strcat/sprintf
with snprintf(). There was a minor and possibly not exploitable security
hole related to one sprintf().
dillon [Sun, 13 Dec 1998 02:47:00 +0000 (02:47 +0000)]
PR: docs/9045
Update chpass(1) manual page. Change the definition of the class
field and also reorganize the field list to match that of the
/etc/master.passwd file, to avoid confusion.
dillon [Sun, 13 Dec 1998 01:53:50 +0000 (01:53 +0000)]
PR: bin/3478
Have pwd_mkdb lock the source file while rebuilding the database. When
called by programs such as vipw, the source file is a temporary file and
this does not conflict with the lock on /etc/master.passwd already held
by vipw. When run manually, however, master.passwd is typically specified
as the argument and the locking prevents other programs from messing with
master.passwd during the database rebuild.
Also pwd_mkdb uses a blocking exclusive lock as it may be called from
a script. The -N option was added to cause pwd_mkdb to get the lock
non-blocking and exit with an error if the attempt fails, again useful
for scripts.
dillon [Sun, 13 Dec 1998 01:36:45 +0000 (01:36 +0000)]
Handle the race condition where vipw may lock a password file which has
just been replaced. After our lock succeeds we check if st_nlink is 0
and if it is we close the descriptor and retry our open/lock sequence.
dillon [Sat, 12 Dec 1998 23:05:22 +0000 (23:05 +0000)]
Reviewed by: cvs-current
Delete rc.local from CVS tree, its remaining functionality has been
moved to /etc/rc. /etc/rc still supports an rc.local but it is now
a 100% user-controlled file.
dillon [Sat, 12 Dec 1998 23:04:21 +0000 (23:04 +0000)]
Reviewed by: cvs-current
Commit changes to rc and rc.local, removing the remaining minimal
functionality of rc.local into rc and commenting it out of rc.local
prior to the deletion of rc.local from the CVS tree.
jb [Sat, 12 Dec 1998 22:00:49 +0000 (22:00 +0000)]
Back out revs 1.181 and 1.182 which upset a few people. I hope those
(3?) people will make an effort to help those who would have benefitted from
this change. And just telling them that they should read and understand
the significance of each message posted to -current is not really good
enough IMHO.
dillon [Sat, 12 Dec 1998 21:45:49 +0000 (21:45 +0000)]
PR: kern/8990
If timer calculation results in degenerate value (0), force it to 1
to avoid divide-by-zero panic later on in calls to IGMP_RANDOM_DELAY().
I considered simply adding 1 to the timer calculation, but was unsure
if the calculation was part of the IGMP standard or not so did not want
to mess with it for all cases.
dillon [Sat, 12 Dec 1998 21:07:09 +0000 (21:07 +0000)]
PR: kern/8965
Obtained from: Stephen Clawson <sclawson@cs.utah.edu>
Wakeup anyone waiting on a mount point prior to returning from umount,
whether an error occurs or not. Fixes a stat/NFS-umount race and other
potential future problems. Fix taken from bug/pr which also indicated
that the same fix has already been applied to OpenBSD and NetBSD.