]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
26 years agoAdd -lz
jkh [Thu, 25 Dec 1997 23:08:44 +0000 (23:08 +0000)]
Add -lz

26 years agoThe ioopt code is still buggy, but wasn't fully disabled.
dyson [Thu, 25 Dec 1997 20:55:15 +0000 (20:55 +0000)]
The ioopt code is still buggy, but wasn't fully disabled.

26 years agoMake kern.ncpu reports the number of detected processors when running
gpalmer [Thu, 25 Dec 1997 13:14:21 +0000 (13:14 +0000)]
Make kern.ncpu reports the number of detected processors when running
with a SMP kernel.

26 years agoUse consistent spelling,
hoek [Thu, 25 Dec 1997 09:36:42 +0000 (09:36 +0000)]
Use consistent spelling,
writeable -> writable (recall prior debate over this? :-)
initialise -> initialize
recognise -> recognize

Merry Christmas! :)

26 years agoThe spl fixes in in_setsockaddr and in_setpeeraddr that were meant to
dg [Thu, 25 Dec 1997 06:57:36 +0000 (06:57 +0000)]
The spl fixes in in_setsockaddr and in_setpeeraddr that were meant to
fix PR#3618 weren't sufficient since malloc() can block - allowing the
net interrupts in and leading to the same problem mentioned in the
PR (a panic). The order of operations has been changed so that this
is no longer a problem.
Needs to be brought into the 2.2.x branch.
PR: 3618

26 years agoAdd a man page for REMOVING devices from devfs.
julian [Thu, 25 Dec 1997 05:54:24 +0000 (05:54 +0000)]
Add a man page for REMOVING devices from devfs.

26 years agoChanged pthread_detach to conform to POSIX, i.e. the single argument
alex [Thu, 25 Dec 1997 05:07:20 +0000 (05:07 +0000)]
Changed pthread_detach to conform to POSIX, i.e. the single argument
provided is of type pthread_t instead of pthread_t *.

PR: 4320

Return EINVAL instead of ESRCH if attempting to detach an already
detached thread.

26 years agoRemoved unnecessary initialization of hp in gethostbyaddr_r.
alex [Thu, 25 Dec 1997 04:21:08 +0000 (04:21 +0000)]
Removed unnecessary initialization of hp in gethostbyaddr_r.

26 years agoAdd overflow checks: if output size becomes bigger than INT_MAX,
ache [Thu, 25 Dec 1997 00:32:17 +0000 (00:32 +0000)]
Add overflow checks: if output size becomes bigger than INT_MAX,
just return EOF

26 years agoCorrect type of stored argument place (from previous fix)
ache [Wed, 24 Dec 1997 23:54:19 +0000 (23:54 +0000)]
Correct type of stored argument place (from previous fix)

26 years ago1) Restore back comment about snprintf()
ache [Wed, 24 Dec 1997 23:23:18 +0000 (23:23 +0000)]
1) Restore back comment about snprintf()
2) Optimize string buffer copy to call memcpy() and update pointers
only for count > 0, it makes snprintf(NULL, 0, ...) more efficient

26 years agoReturn back to BSD snprintf semantics which recent C9x standard adopts
ache [Wed, 24 Dec 1997 23:02:47 +0000 (23:02 +0000)]
Return back to BSD snprintf semantics which recent C9x standard adopts
instead of Singe Unix, thanx Bruce for explaining, I am not realize
standards war was there.

But now, fix n == 0 case to not return error and fix check for too
big n.

Things left to do: check for overflow in arguments.

26 years ago1) Oops! Insert again if (n == 0) return 0.
ache [Wed, 24 Dec 1997 20:24:08 +0000 (20:24 +0000)]
1) Oops! Insert again if (n == 0) return 0.

Final word is Bruce's quote:

C9x specifies the BSD4.4-Lite behaviour:

       [#3] ...   Thus,  the
       null-terminated  output  has  been completely written if and
       only if the returned value is less than n.

It means that if we not have any null-terminated output as for n == 0
we can't return value less than n, so we forced to return value
equal to n i.e. 0

The next good thing is glibc compatibility, of course.

2) Do check for too big n in machine-independent way.
3) Minor optimization assuming EOF is < 0

26 years agosprintf->snprintf paranoia
imp [Wed, 24 Dec 1997 19:39:57 +0000 (19:39 +0000)]
sprintf->snprintf paranoia
Obtained from: OpenBSD (?)

26 years agoBe extra paranoid about the length of data returned from gethostbyaddr or
imp [Wed, 24 Dec 1997 19:38:18 +0000 (19:38 +0000)]
Be extra paranoid about the length of data returned from gethostbyaddr or
gethostbyname.
Submitted by: Julian Assange

26 years agosprintf->snprintf paranoia. The one thing that looks like a hole in
imp [Wed, 24 Dec 1997 19:21:09 +0000 (19:21 +0000)]
sprintf->snprintf paranoia.  The one thing that looks like a hole in
the diff is in an ifdef that isn't enabled for FreeBSD.
Obtained from: OpenBSD(?)

26 years agoVarious sprintf -> snprintf fixes.
imp [Wed, 24 Dec 1997 19:13:23 +0000 (19:13 +0000)]
Various sprintf -> snprintf fixes.
Minor style fix (strcpy(foo,"") -> *foo = '\0')
Obtained from: OpenBSD(?)

26 years agoUse snprintf rather than printf out of paranoia
imp [Wed, 24 Dec 1997 18:56:03 +0000 (18:56 +0000)]
Use snprintf rather than printf out of paranoia
Obtained from: OpenBSD

26 years agoSupport running with inadequate swap space. Additionally, the code
dyson [Wed, 24 Dec 1997 15:05:25 +0000 (15:05 +0000)]
Support running with inadequate swap space.  Additionally, the code
will complain with a suggestion of increasing it.

26 years agoBack out part related to "return 0 if n == 0" and return EOF as before.
ache [Wed, 24 Dec 1997 14:32:40 +0000 (14:32 +0000)]
Back out part related to "return 0 if n == 0" and return EOF as before.

The main argument is that it is impossible to determine if %n evaluated or not
when snprintf return 0, because it can happens for both n == 0 and n == 1.
Although EOF here is good indication of the end of process, if n is
decreased in the loop...
Since it is already supposed in many places that EOF *is* negative, f.e.
from Single Unix specs for snprintf
"return ... a negative value if an output error was encountered"
this not makes situation worse.

26 years agoFix snprintf(...%n...)
ache [Wed, 24 Dec 1997 13:47:13 +0000 (13:47 +0000)]
Fix snprintf(...%n...)
to pass not more than buffer size to %n agrument, old variant
always assume infinite buffer.
%n is for actually transmitted characters, not for planned ones.

26 years agoRemove wrong comment about snprintf:
ache [Wed, 24 Dec 1997 13:17:13 +0000 (13:17 +0000)]
Remove wrong comment about snprintf:
"return the number of bytes needed, rather the number used"

According to Single Unix specs:

Upon successful completion, these functions return the number of bytes
transmitted excluding the terminating null

26 years agosnprintf return value fixes to conform Single Unix specs:
ache [Wed, 24 Dec 1997 12:31:32 +0000 (12:31 +0000)]
snprintf return value fixes to conform Single Unix specs:

1) if buffer size is smaller than arguments size, return buffer
size, not arguments size as before.

2) if buffer size is 0, return 0, not EOF as before.
(now it is compatible with Linux and Apache implementations too).

NOTE: Single Unix specs says:

If the value of n {buffer size} is zero on a call to snprintf(), an
unspecified value less than 1 is returned.

It means we can't return EOF since EOF can take *any* value in general
not especially < 1. Better variant will be return -1 (it is less then
1 and different with n == 1 case) but -1 value is already occuped by
EOF in our implementation, so we can't distinguish true IO error
in that case. So 0 here is only possible case still conforming
to Single Unix specs.

26 years agoSmall bugfix: store $uid_start instead hard coded value 1000
wosch [Wed, 24 Dec 1997 11:44:14 +0000 (11:44 +0000)]
Small bugfix: store $uid_start instead hard coded value 1000
in /etc/adduser.conf.

26 years agoCorrect copyright.
brian [Wed, 24 Dec 1997 10:28:38 +0000 (10:28 +0000)]
Correct copyright.

Requested by: Eivind Eklund <perhaps@yes.no>

26 years agoPut the dir file in the correct distribution.
jkh [Wed, 24 Dec 1997 10:16:49 +0000 (10:16 +0000)]
Put the dir file in the correct distribution.

26 years agoCosmetic (style):
brian [Wed, 24 Dec 1997 09:29:17 +0000 (09:29 +0000)]
Cosmetic (style):
  sizeof(var) -> sizeof var
  sizeof type -> sizeof(type)

Suggested by: J Wunsch <j@uriah.heep.sax.de>

26 years agoUse sizeof cftypes128 for NCFTYPES128 macro, not sizeof cftypes !
brian [Wed, 24 Dec 1997 09:28:11 +0000 (09:28 +0000)]
Use sizeof cftypes128 for NCFTYPES128 macro, not sizeof cftypes !
Lucky that sizeof cftypes < sizeof cftypes128

26 years ago- Add prototype for adjust_timeout_calltodo().
nate [Wed, 24 Dec 1997 01:33:45 +0000 (01:33 +0000)]
- Add prototype for adjust_timeout_calltodo().

Submitted/forgotten by: Ken Key <key@cs.utk.edu>

26 years agoBe extra paranoid about trusting the length of the data returned by
imp [Wed, 24 Dec 1997 01:00:24 +0000 (01:00 +0000)]
Be extra paranoid about trusting the length of the data returned by
gethostbyaddr.
Submitted by: Julian Assange

26 years agoTwo fixes:
imp [Wed, 24 Dec 1997 00:59:49 +0000 (00:59 +0000)]
Two fixes:
1) Correct strncpy usage
2) gethostbyaddr paranoia from Julian Assange

26 years agoAdded cross references for apm, cron, inetd, lpd, moused, mrouted,
alex [Wed, 24 Dec 1997 00:59:28 +0000 (00:59 +0000)]
Added cross references for apm, cron, inetd, lpd, moused, mrouted,
nfsd, nfsiod, portmap, rpc.lockd, rpc.statd, syslogd, tickadj, and
ypbind.

26 years agoMinor style nit noticed by bde.
imp [Wed, 24 Dec 1997 00:59:02 +0000 (00:59 +0000)]
Minor style nit noticed by bde.

26 years agoBe extra paranoid about trusting the length returned by gethostbyaddr.
imp [Wed, 24 Dec 1997 00:58:39 +0000 (00:58 +0000)]
Be extra paranoid about trusting the length returned by gethostbyaddr.
Submitted by: Julian Assange

26 years agoBe extra paranoid about trusting the size of the address returned from
imp [Wed, 24 Dec 1997 00:57:41 +0000 (00:57 +0000)]
Be extra paranoid about trusting the size of the address returned from
gethostbyaddr.
Submitted by: Julian Assange

26 years agostrncpy(x,y,sizeof(x)) --> strncpy(x,y,sizeof(x)-1)
brian [Tue, 23 Dec 1997 22:38:55 +0000 (22:38 +0000)]
strncpy(x,y,sizeof(x)) --> strncpy(x,y,sizeof(x)-1)

Suggested by: Philippe Charnier <charnier@lirmm.fr>
              Theo de Raadt <deraadt@cvs.openbsd.org>

26 years agoAdd a PARITITON arg to SCSI_MKFIXED, and use it to
brian [Tue, 23 Dec 1997 19:44:45 +0000 (19:44 +0000)]
Add a PARITITON arg to SCSI_MKFIXED, and use it to
specify RAW_PART for the suopen() in sscioctl().

Now you can `scsi -f /dev/ssc -p'.

26 years agoThis patch causes the "calltodo" timer list to be decremented by the amount
nate [Tue, 23 Dec 1997 16:32:35 +0000 (16:32 +0000)]
This patch causes the "calltodo" timer list to be decremented by the amount
of time that the laptop was suspending.  Thus, select() calls that might have
suspended rather than firing  at 1hr + "time suspended" since the timer was
posted.

Adding:

    options  APM_FIXUP_CALLTODO

to the kernel config enables the patch.

[
This patch was slightly modified to use a consistant indent style and
I removed some unused local variables.  After this has been tested a
few weeks we'll make the options the default, so for now I'm now
documenting it in LINT.  Mike can later if he wants.
]

Reviewed by: Mike Smith <msmith@freebsd.org>
Submitted by: Ken Key  <key@cs.utk.edu>

26 years ago- Added support for a new MouseMan model (4 buttons, no wheel).
yokota [Tue, 23 Dec 1997 08:04:38 +0000 (08:04 +0000)]
- Added support for a new MouseMan model (4 buttons, no wheel).
  Based on the report from Dave Bodenstab.
- Turn off PnP COM device enumeration procedure if the user explicitly
  specifies a protocol type with the "-t" option.
- Accept "-t auto".  Now the user may entirely omit the "-t" option
  in the command line, or specify "-t auto" in order to make moused
  detect an appropriate protocol type automatically.  In the
  previous version, moused did so only if the "-t" option is absent
  in the command line. ("-t auto" won't disable PnP COM device
  enumeration.)
- Updated the man page.

26 years agoDocument `flags' for the psm driver.
yokota [Tue, 23 Dec 1997 08:01:16 +0000 (08:01 +0000)]
Document `flags' for the psm driver.

26 years agoRemoved unnecessary setting of 'error' -- binding to a privileged port
alex [Tue, 23 Dec 1997 01:40:40 +0000 (01:40 +0000)]
Removed unnecessary setting of 'error' -- binding to a privileged port
by a non-root user always returns EACCES.

26 years agoYTK fix.
alex [Mon, 22 Dec 1997 23:00:35 +0000 (23:00 +0000)]
YTK fix.

26 years agoReviewed by: William LeFebvre <wnl@groupsys.com>
jb [Mon, 22 Dec 1997 19:39:20 +0000 (19:39 +0000)]
Reviewed by: William LeFebvre <wnl@groupsys.com>
Change errno -> error in local structure to avoid a clash with the
thread-aware version of errno which is required for a thread-safe libc.

Have discussed this with the author and he has agreed to this change. 8-)

26 years agoFix sysctl path /usr/sbin->/sbin
ache [Mon, 22 Dec 1997 19:11:28 +0000 (19:11 +0000)]
Fix sysctl path /usr/sbin->/sbin

26 years agoRemove unused include files
eivind [Mon, 22 Dec 1997 13:46:08 +0000 (13:46 +0000)]
Remove unused include files

Submitted by: bde

26 years agoImprove my copyright.
dyson [Mon, 22 Dec 1997 11:54:00 +0000 (11:54 +0000)]
Improve my copyright.

26 years agoImprove my copyright.
dyson [Mon, 22 Dec 1997 11:48:13 +0000 (11:48 +0000)]
Improve my copyright.

26 years agoCorrect my previous fix for the UPAGES problem.
dyson [Mon, 22 Dec 1997 10:06:09 +0000 (10:06 +0000)]
Correct my previous fix for the UPAGES problem.

26 years agoRemove obsolete /etc/hostname.ep? reference.
charnier [Mon, 22 Dec 1997 07:37:50 +0000 (07:37 +0000)]
Remove obsolete /etc/hostname.ep? reference.
PR: kern/5310
Submitted by: tom@sdf.com

26 years agoUse .Tn, .Bx Free and .An.
charnier [Mon, 22 Dec 1997 07:35:40 +0000 (07:35 +0000)]
Use .Tn, .Bx Free and .An.

26 years agoDon't forget to ntohs(sin.sin_port) (DUH!)
brian [Mon, 22 Dec 1997 02:28:18 +0000 (02:28 +0000)]
Don't forget to ntohs(sin.sin_port) (DUH!)

26 years agoHopefully fix the problem with the TLB not being updated correctly.
dyson [Mon, 22 Dec 1997 00:36:48 +0000 (00:36 +0000)]
Hopefully fix the problem with the TLB not being updated correctly.
Problem tracked down by bde@freebsd.org, but this is an attempted
efficient fix.

26 years agoProperly clean out the SI_MOUNTEDON flag iff the mount attempt fails
joerg [Sun, 21 Dec 1997 21:40:02 +0000 (21:40 +0000)]
Properly clean out the SI_MOUNTEDON flag iff the mount attempt fails
half the way down.  Otherwise, further attempts to mount the device
will be rejected with BUSY.

IMHO, this flag can completely go away for cd9660.  There's no reason
you need to prevent CDs from being mounted multiple times, and in case
of multisession CDs it can even make sense to mount two different
sessions by the same time (to different mount points, otherwise it
would be pointless ;).

26 years agoWhen called as `zgrep', the -Z argument is assumed. Make a
wosch [Sun, 21 Dec 1997 19:15:12 +0000 (19:15 +0000)]
When called as `zgrep', the -Z argument is assumed. Make a
link from zgrep to grep.

Pointed out by: Tim Vanderhoek <hoek@hwcn.org> and
                Mike Smith <mike@smith.net.au>

26 years agoMoved some declarations from <sys/socket.h> to the correct places, and
bde [Sun, 21 Dec 1997 16:35:12 +0000 (16:35 +0000)]
Moved some declarations from <sys/socket.h> to the correct places, and
fixed everything that depended on them being misplaced.

26 years agoTypo fix in the message displayed.
max [Sun, 21 Dec 1997 15:40:11 +0000 (15:40 +0000)]
Typo fix in the message displayed.

26 years agoRemove unused private header.
eivind [Sun, 21 Dec 1997 15:35:22 +0000 (15:35 +0000)]
Remove unused private header.

Submitted by: max

26 years agoCharles Mott created these (and told me via email that
brian [Sun, 21 Dec 1997 14:28:24 +0000 (14:28 +0000)]
Charles Mott created these (and told me via email that
they were BSD copyright).  Use his name, not mine.

26 years agoRemove unused private include files from slattach.
eivind [Sun, 21 Dec 1997 13:42:27 +0000 (13:42 +0000)]
Remove unused private include files from slattach.

Submitted by: bde

26 years agoAdd (BSD) copyright headers.
brian [Sun, 21 Dec 1997 12:11:13 +0000 (12:11 +0000)]
Add (BSD) copyright headers.

26 years agoI added vfs_ioopt prematurely, disabled.
dyson [Sun, 21 Dec 1997 10:41:19 +0000 (10:41 +0000)]
I added vfs_ioopt prematurely, disabled.

26 years agoRemove private install rule in favor of INTERNALLIB and
cwt [Sun, 21 Dec 1997 08:31:50 +0000 (08:31 +0000)]
Remove private install rule in favor of INTERNALLIB and
INTERNALSTATICLIB.

Suggested by: bde

26 years agoCosmetic: Test for comment lines after the lenght has been established.
brian [Sun, 21 Dec 1997 03:41:23 +0000 (03:41 +0000)]
Cosmetic: Test for comment lines after the lenght has been established.
Pointed out by: Theo de Raadt <deraadt@cvs.openbsd.org>

26 years agoAllow multiple (comma seperated) devices on the "set device" line.
brian [Sun, 21 Dec 1997 03:16:19 +0000 (03:16 +0000)]
Allow multiple (comma seperated) devices on the "set device" line.
Submitted by: Derek Inksetter <derek@saidev.com>

26 years agoCosmetic: Tidy up.
brian [Sun, 21 Dec 1997 02:37:49 +0000 (02:37 +0000)]
Cosmetic: Tidy up.
Submitted by: Theo de Raadt <deraadt@cvs.openbsd.org>

26 years agoMention that leading whitespace is ignored when identifying comments.
brian [Sun, 21 Dec 1997 02:34:27 +0000 (02:34 +0000)]
Mention that leading whitespace is ignored when identifying comments.

26 years agoZero struct sockaddr_??s before filling them in.
brian [Sun, 21 Dec 1997 02:27:32 +0000 (02:27 +0000)]
Zero struct sockaddr_??s before filling them in.
Suggested by: Theo de Raadt <deraadt@cvs.openbsd.org>

26 years agoAllow leading whitespace before '#' when identifying comments.
brian [Sun, 21 Dec 1997 02:11:48 +0000 (02:11 +0000)]
Allow leading whitespace before '#' when identifying comments.
Submitted by: Dave Bodenstab <imdave@mcs.net>

26 years agoCorrect the "how to find out if the line is up" example.
brian [Sun, 21 Dec 1997 01:21:40 +0000 (01:21 +0000)]
Correct the "how to find out if the line is up" example.

26 years agoCosmetic: Remove blank lines, add .Pp where necessary.
brian [Sun, 21 Dec 1997 01:07:13 +0000 (01:07 +0000)]
Cosmetic: Remove blank lines, add .Pp where necessary.
Submitted by: Theo de Raadt <deraadt@cvs.openbsd.org>

26 years agoFixed overflow in chkrange(). Some out of bounds block numbers,
bde [Sun, 21 Dec 1997 00:00:44 +0000 (00:00 +0000)]
Fixed overflow in chkrange().  Some out of bounds block numbers,
e.g. -1, were not detected.  Use a bulletproof check that doesn't
depend on special properties of the args or the limit.

PR: 3528

26 years agoDuplicate the entry for the Plasmon CD-R device, so both possibilities
joerg [Sat, 20 Dec 1997 23:05:14 +0000 (23:05 +0000)]
Duplicate the entry for the Plasmon CD-R device, so both possibilities
(read/only and write/once) will be covered.  This is necessary in order
to fill in the proper device information record.

26 years agoProtect against a null pointer dereferencation in the case of an
joerg [Sat, 20 Dec 1997 23:03:49 +0000 (23:03 +0000)]
Protect against a null pointer dereferencation in the case of an
unknown drive.  Such a drive will be configured by worm(4)
nevertheless (albeit with a warning), but cannot be opened except of
the SCSI control device (so scsi(8) or cdrecord will continue to
work).

26 years agoYear 2000 fixes
jraynard [Sat, 20 Dec 1997 22:31:37 +0000 (22:31 +0000)]
Year 2000 fixes
PR: 4930
Submitted by:  Stephen J. Roznowski <sjr@home.net>

26 years agoYear 2000 fixes for groff
jraynard [Sat, 20 Dec 1997 22:29:55 +0000 (22:29 +0000)]
Year 2000 fixes for groff
PR: 4930
Submitted by: Stephen J. Roznowski

26 years agoFixed style bugs in the printing of statistics after preening. Use
bde [Sat, 20 Dec 1997 22:24:32 +0000 (22:24 +0000)]
Fixed style bugs in the printing of statistics after preening.  Use
floating point better in the percentage calculation there to avoid
overflow when there are more than about 20 million fragments.  Start
using floating point in the other percentage calculation to avoid
overflow when there are more than about 2 million fragments.

Fixed printf format strings.

Converted sccsid to rcsid.

26 years agoDon't attempt to print the statistics for a "clean" preened filesystem
bde [Sat, 20 Dec 1997 21:36:58 +0000 (21:36 +0000)]
Don't attempt to print the statistics for a "clean" preened filesystem
when there isn't even a filesystem.  Attempting to print them tended
to cause SIGSEGV or SIGFPE depending on how far setup() got before it
returned 0.  This was broken in the previous revision by removing a
return statement that the previous case depended on falling into.

PR: 4840 (fixed by this commit)
PR: 2537 (possibly fixed by Lite2 merge and later changes.  setup()
      does more checking now)

26 years agoAdded zgrep.libz. This version of zgrep(1) depend on a
wosch [Sat, 20 Dec 1997 19:20:33 +0000 (19:20 +0000)]
Added zgrep.libz. This version of zgrep(1) depend on a
grep(1) linked with libz.

26 years agoAdded a sparc category to the list.
wosch [Sat, 20 Dec 1997 18:54:22 +0000 (18:54 +0000)]
Added a sparc category to the list.

26 years agoAdded builtin decompression using zlib library, option -Z.
wosch [Sat, 20 Dec 1997 18:46:09 +0000 (18:46 +0000)]
Added builtin decompression using zlib library, option -Z.

Enabled this new feature with the makefile variable GREP_LIBZ. If
you don't like it, compile with `make GREP_LIBZ='.

grep + zlib has several advantages:

- the shell script zgrep(1) will be basically a one line
  exec grep -Z "$@"

- no shell script, no bugs. The current zgrep implementations
  have many bugs and some grep options are no supported.

- no shell script, no security risks.

- it is a magnitude faster than a shell script

Also fixed:
0 -> STDIN_FILENO
Close a file descriptor only if the open call was successfully. It does
not hurt for the open(2) function, but the gzclose(3) function
died in free() to free up (not) allocated memory.

26 years agoAdd some copyright and license terms, at Jordan's request. Note that
sef [Sat, 20 Dec 1997 18:40:43 +0000 (18:40 +0000)]
Add some copyright and license terms, at Jordan's request.  Note that
syscall.h just refers to another file for the copyright notice.

26 years agoAdd a copyright and license notice, on Jordan's request.
sef [Sat, 20 Dec 1997 18:36:02 +0000 (18:36 +0000)]
Add a copyright and license notice, on Jordan's request.

26 years agoAdd a copyright and license statement, at the request of Jordan.
sef [Sat, 20 Dec 1997 18:34:49 +0000 (18:34 +0000)]
Add a copyright and license statement, at the request of Jordan.

26 years agoRemove bogus #ifdef INET - SLIP doesn't compile without INET.
eivind [Sat, 20 Dec 1997 16:44:51 +0000 (16:44 +0000)]
Remove bogus #ifdef INET - SLIP doesn't compile without INET.

26 years agoMake the class code checks in function pci_cfgcheck less strict.
se [Sat, 20 Dec 1997 09:04:25 +0000 (09:04 +0000)]
Make the class code checks in function pci_cfgcheck less strict.
It failed to recognize the PCI bus in a system that had only an
old chip-set (class code 000000) and a Cyclom multiport serial
card on PCI bus 0, but no VGA card or disk or network controller.

PR: i386/5300
Submitted by: Nickolay N. Dudorov <nnd@itfs.nsk.su>

26 years agoChange errno usage as a field in a structure and as an argument to a
jb [Sat, 20 Dec 1997 04:06:06 +0000 (04:06 +0000)]
Change errno usage as a field in a structure and as an argument to a
function from 'errno' to 'error' so that there is no conflict with the
thread-safe definition of errno in errno.h.

26 years agoClear the p_stops field on change of user/group id, unless the correct
sef [Sat, 20 Dec 1997 03:05:47 +0000 (03:05 +0000)]
Clear the p_stops field on change of user/group id, unless the correct
flag is set in the p_pfsflags field.  This, essentially, prevents an SUID
proram from hanging after being traced.  (E.g., "truss /usr/bin/rlogin" would
fail, but leave rlogin in a stopevent state.)  Yet another case where procctl
is (hopefully ;)) no longer needed in the general case.

Reviewed by: bde (thanks bruce :))

26 years agoSync with sys/i386/conf/Makefile.i386 revision 1.106.
kato [Sat, 20 Dec 1997 02:37:55 +0000 (02:37 +0000)]
Sync with sys/i386/conf/Makefile.i386 revision 1.106.

26 years agoNever create the destination directory at install time.
bde [Sat, 20 Dec 1997 00:40:05 +0000 (00:40 +0000)]
Never create the destination directory at install time.
Don't build things at install time.
Don't override the (correct) clean rule.
Rewrote all the rules and cleaned up.

26 years agoIfdefed some conditionally used declarations.
bde [Sat, 20 Dec 1997 00:28:49 +0000 (00:28 +0000)]
Ifdefed some conditionally used declarations.

26 years agoRemoved a stray semicolon.
bde [Sat, 20 Dec 1997 00:27:03 +0000 (00:27 +0000)]
Removed a stray semicolon.

26 years agoAdded a used include.
bde [Sat, 20 Dec 1997 00:25:01 +0000 (00:25 +0000)]
Added a used include.

Fixed a gratuitous ANSIism and nearby KNF violations.

26 years agoFixed a sloppy common-style declaration.
bde [Sat, 20 Dec 1997 00:13:51 +0000 (00:13 +0000)]
Fixed a sloppy common-style declaration.

natm_pcb.c:
Include <sys/socketvar.h> which will be used when M_PCB is declared
in the right place.

26 years agoFixed gratuitous ANSIisms.
bde [Sat, 20 Dec 1997 00:07:11 +0000 (00:07 +0000)]
Fixed gratuitous ANSIisms.

26 years agoHandle the ``maintenance'' case better.
bde [Fri, 19 Dec 1997 23:56:18 +0000 (23:56 +0000)]
Handle the ``maintenance'' case better.

Fixed some style bugs.

26 years agoDon't use ANSI string concatenation to misformat a string.
bde [Fri, 19 Dec 1997 23:46:21 +0000 (23:46 +0000)]
Don't use ANSI string concatenation to misformat a string.

26 years agoRemoved a stale comment. (We don't declare ip_len and ip_offset as
bde [Fri, 19 Dec 1997 23:33:08 +0000 (23:33 +0000)]
Removed a stale comment.  (We don't declare ip_len and ip_offset as
short.  I guess we depend on bogus ANSI value-preserving extension
of u_short to int to avoid unsigned comparison bugs.)

26 years agoUse __inline instead of inline to prevent pedantic compiler warnings.
bde [Fri, 19 Dec 1997 23:25:16 +0000 (23:25 +0000)]
Use __inline instead of inline to prevent pedantic compiler warnings.

26 years agoRemoved some bogus casts.
bde [Fri, 19 Dec 1997 23:18:37 +0000 (23:18 +0000)]
Removed some bogus casts.

26 years agoFixed some FUD (don't blame current msdosfs for old msdosfs+vfs_bio bugs).
bde [Fri, 19 Dec 1997 23:03:00 +0000 (23:03 +0000)]
Fixed some FUD (don't blame current msdosfs for old msdosfs+vfs_bio bugs).