]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
25 years ago(1) Fix a spelling mistake in a comment.
billf [Tue, 17 Aug 1999 22:13:06 +0000 (22:13 +0000)]
(1) Fix a spelling mistake in a comment.
(2) Add braces to avoid ambigious if/if/else

Reviewed by: bde

25 years agoFix a printf(3) formatter to match its variable.
billf [Tue, 17 Aug 1999 22:11:02 +0000 (22:11 +0000)]
Fix a printf(3) formatter to match its variable.

Reviewed by: bde

25 years agoFix a printf() formatter to match its variable.
billf [Tue, 17 Aug 1999 22:10:00 +0000 (22:10 +0000)]
Fix a printf() formatter to match its variable.

Reviewed by: bde, luigi

25 years agoRemove an unused variable.
billf [Tue, 17 Aug 1999 22:08:55 +0000 (22:08 +0000)]
Remove an unused variable.

Reviewed by: bde

25 years agoWrap two declarations that are only used in an #ifdef in their own #ifdef.
billf [Tue, 17 Aug 1999 22:06:17 +0000 (22:06 +0000)]
Wrap two declarations that are only used in an #ifdef in their own #ifdef.

Reviewed by: bde

25 years agoTake out a reference to ccb(4). I never got around to writing it.
ken [Tue, 17 Aug 1999 22:03:40 +0000 (22:03 +0000)]
Take out a reference to ccb(4).  I never got around to writing it.

Reported by: "Alexey M. Zelkin" <phantom@cris.net>

25 years agoWelcome devtoname(), to most likely be used when printing information
billf [Tue, 17 Aug 1999 20:25:50 +0000 (20:25 +0000)]
Welcome devtoname(), to most likely be used when printing information
about a dev_t.

printf("%x", dev) now becomes printf("%s", devtoname(dev)) because
printing actual information about the device is much more useful then
printing a pointer to an address that would never help the developer debug.

Submitted by: phk, bde

25 years agoAdd a second arg to ``set timeout'' as an example of a minimum idle
brian [Tue, 17 Aug 1999 19:23:49 +0000 (19:23 +0000)]
Add a second arg to ``set timeout'' as an example of a minimum idle
period.

25 years agovm_page_free_toq:
alc [Tue, 17 Aug 1999 18:09:01 +0000 (18:09 +0000)]
vm_page_free_toq:
Update the comment to reflect the demise of PQ_ZERO and
remove a (now) useless test.

25 years agoImplement a minimum idle time value as an optional second argument
brian [Tue, 17 Aug 1999 17:22:46 +0000 (17:22 +0000)]
Implement a minimum idle time value as an optional second argument
to ``set timeout''.
This is useful for situations where your minimum call charge is (say)
5 minutes (like mine is)

25 years agoUse ``set openmode passive'' in the ssh VPN example
brian [Tue, 17 Aug 1999 15:00:15 +0000 (15:00 +0000)]
Use ``set openmode passive'' in the ssh VPN example

25 years agoSet the close-on-exec flag for all unused descriptors when
brian [Tue, 17 Aug 1999 14:59:05 +0000 (14:59 +0000)]
Set the close-on-exec flag for all unused descriptors when
exec()ing other programs.

25 years agoDocument a whole truckload of sysctl MIBs under net.inet.tcp,
csgr [Tue, 17 Aug 1999 14:54:26 +0000 (14:54 +0000)]
Document a whole truckload of sysctl MIBs under net.inet.tcp,
including tcp.blackhole, tcp.log_in_vain and others.

25 years agoAdd reference to blackhole(4)
csgr [Tue, 17 Aug 1999 14:28:07 +0000 (14:28 +0000)]
Add reference to blackhole(4)

25 years agoAdd man page for black hole sysctl MIBs.
csgr [Tue, 17 Aug 1999 13:46:38 +0000 (13:46 +0000)]
Add man page for black hole sysctl MIBs.
references to follow.

25 years agoClaim ownership
markm [Tue, 17 Aug 1999 13:39:28 +0000 (13:39 +0000)]
Claim ownership

25 years agoAdd net.inet.tcp.blackhole and net.inet.udp.blackhole
csgr [Tue, 17 Aug 1999 12:17:53 +0000 (12:17 +0000)]
Add net.inet.tcp.blackhole and net.inet.udp.blackhole
sysctl knobs.

With these knobs on, refused connection attempts are dropped
without sending a RST, or Port unreachable in the UDP case.
In the TCP case, sending of RST is inhibited iff the incoming
segment was a SYN.

Docs and rc.conf settings to follow.

25 years agoAdd support for the serial version of Interlink VersaPad.
yokota [Tue, 17 Aug 1999 12:14:59 +0000 (12:14 +0000)]
Add support for the serial version of Interlink VersaPad.

25 years agoMention Interlink VersaPad support.
yokota [Tue, 17 Aug 1999 12:14:41 +0000 (12:14 +0000)]
Mention Interlink VersaPad support.

25 years agoRecognize Interlink VersaPad. `Tap' action will be recognized
yokota [Tue, 17 Aug 1999 12:14:13 +0000 (12:14 +0000)]
Recognize Interlink VersaPad.  `Tap' action will be recognized
as the button 4.

Submitted by: Masachika ISHIZUKA <ishizuka@ish.org>

25 years agoAdd definitions for Interlink VersaPad.
yokota [Tue, 17 Aug 1999 12:12:44 +0000 (12:12 +0000)]
Add definitions for Interlink VersaPad.

25 years agoThe sysinstall man page menetions that it was due to be replaced
mpp [Tue, 17 Aug 1999 10:57:18 +0000 (10:57 +0000)]
The sysinstall man page menetions that it was due to be replaced
in a previous FreeBSD version.  That never happened.  Document that
it is due to be replaced, but leave it open-ended as to when.

Also do some mdoc cleanup.

PR: docs/13148
PR: docs/13144
Submitted by: Lee Cremeans <lcremeans@erols.com>
Alex M. Zelkim <phantom@cris.net>
Discussed with: jkh

25 years agoFix a bug in debug-printfs of struct linux_termios fields, where I forgot to
marcel [Tue, 17 Aug 1999 10:27:55 +0000 (10:27 +0000)]
Fix a bug in debug-printfs of struct linux_termios fields, where I forgot to
change the format specifier after changing the definition of the structure.

Submitted by: billf
Commented on by: bde

25 years agoFix bug in the debug-printf of the vfork syscall, where the format specifier
marcel [Tue, 17 Aug 1999 10:09:06 +0000 (10:09 +0000)]
Fix bug in the debug-printf of the vfork syscall, where the format specifier
didn't match the argument (p->p_pid).

While I'm at it, also fix the dupo in the format string and fix the annoying
inconsistency in all the debug-printfs wrt p_pid arguments. Change all of them
to use the %ld format specifier and cast the p_pid arguments to long.

Submitted by: billf

25 years agoFix description of the _thread_autoinit_dummy_decl trick.
rse [Tue, 17 Aug 1999 09:50:21 +0000 (09:50 +0000)]
Fix description of the _thread_autoinit_dummy_decl trick.

25 years ago- Added documentation
n_hibma [Tue, 17 Aug 1999 07:36:34 +0000 (07:36 +0000)]
- Added documentation
- fix bug in data toggle handling

25 years agoSearch for and interrogate the PnP BIOS if found. This code just prints
msmith [Tue, 17 Aug 1999 07:10:34 +0000 (07:10 +0000)]
Search for and interrogate the PnP BIOS if found.  This code just prints
the PnP device IDs in verbose mode; it does not (yet) save any resource
data or contribute to the PnP process nor resource management.

25 years agoMindbogglingly, many BIOS vendors expect to be able to load %ds with
msmith [Tue, 17 Aug 1999 07:09:13 +0000 (07:09 +0000)]
Mindbogglingly, many BIOS vendors expect to be able to load %ds with
0x40 and then access data stored in real-mode segment 0x40, even when
called in protected mode.  Microsoft unfortunately coddle these individuals,
and so must we if we want to run their code.

This change works around GPFs in some APM and PnP BIOS implementations.

Obtained from: Linux

25 years agoCorrect an accidental omission of one "vm_page_undirty" replacement
alc [Tue, 17 Aug 1999 05:56:00 +0000 (05:56 +0000)]
Correct an accidental omission of one "vm_page_undirty" replacement
from the previous commit.

25 years agovm_page_free_toq:
alc [Tue, 17 Aug 1999 05:08:39 +0000 (05:08 +0000)]
vm_page_free_toq:
Clear the dirty bit mask (vm_page_undirty) before adding the page
to the free page queue.

Submitted by: dillon

25 years agoUse V-FAST mode register to check whether it supports V-FST mode.
nyan [Tue, 17 Aug 1999 04:33:11 +0000 (04:33 +0000)]
Use V-FAST mode register to check whether it supports V-FST mode.

Submitted by: WATANABE Takuya <sodium@xuni.ne.jp>

25 years agoAdd the (inline) function vm_page_undirty for clearing the dirty bitmask
alc [Tue, 17 Aug 1999 04:02:34 +0000 (04:02 +0000)]
Add the (inline) function vm_page_undirty for clearing the dirty bitmask
of a vm_page.

Use it.

Submitted by: dillon

25 years agoAdded "include" feature to pccard.conf to put the default pccard.conf
hosokawa [Tue, 17 Aug 1999 03:38:51 +0000 (03:38 +0000)]
Added "include" feature to pccard.conf to put the default pccard.conf
entries in /etc/defaults/.

25 years agoEnable printing of bridging statistics.
luigi [Tue, 17 Aug 1999 03:30:34 +0000 (03:30 +0000)]
Enable printing of bridging statistics.

25 years agoMake hostname comparisons case insensitive
brian [Tue, 17 Aug 1999 01:25:16 +0000 (01:25 +0000)]
Make hostname comparisons case insensitive

25 years agoDon't assume all ctags are meant to match a whole line ("^...line...$").
hoek [Tue, 17 Aug 1999 00:49:40 +0000 (00:49 +0000)]
Don't assume all ctags are meant to match a whole line ("^...line...$").
This lets more(1) work with, for example, /usr/local/share/vim/doc/*.txt.

25 years agoMerge style- and trivial- only changes from OpenBSD (dated 1999/07/29-19:55+1).
hoek [Tue, 17 Aug 1999 00:39:26 +0000 (00:39 +0000)]
Merge style- and trivial- only changes from OpenBSD (dated 1999/07/29-19:55+1).

Obtained from: OpenBSD, sometimes indirected from NetBSD; myself

25 years agovm_pageout_clean:
alc [Tue, 17 Aug 1999 00:07:35 +0000 (00:07 +0000)]
vm_pageout_clean:
Remove dead code.

Submitted by: dillon

25 years agoBump CAM version number.
gibbs [Mon, 16 Aug 1999 23:18:24 +0000 (23:18 +0000)]
Bump CAM version number.

25 years agoAdd a cleanup handler called "atexit" to make sure we free up target
gibbs [Mon, 16 Aug 1999 22:52:17 +0000 (22:52 +0000)]
Add a cleanup handler called "atexit" to make sure we free up target
mode instances.

Print out exception information passed out from the kernel target mode
driver to this userland daemon.

25 years agoDifferentiate between aic7895C chips and their earlier brethren
gibbs [Mon, 16 Aug 1999 22:50:37 +0000 (22:50 +0000)]
Differentiate between aic7895C chips and their earlier brethren
via the PCI revid register.

25 years agoAdd support for issuing immediate notify event ccbs for bus resets, bdr
gibbs [Mon, 16 Aug 1999 22:49:29 +0000 (22:49 +0000)]
Add support for issuing immediate notify event ccbs for bus resets, bdr
messages, abort messages, and abort tag messages.

Fix a bug in how default transfer negotiations are handled if the
user had disabled initial bus resets.

Support multi-targetid on the aic7895C.

25 years agoKill an unused INTSTAT type.
gibbs [Mon, 16 Aug 1999 22:43:08 +0000 (22:43 +0000)]
Kill an unused INTSTAT type.

25 years agoFix a few compiler nits.
gibbs [Mon, 16 Aug 1999 22:42:33 +0000 (22:42 +0000)]
Fix a few compiler nits.

Submitted by: Bill Fumerola <billf@jade.chc-chimes.com>

25 years agoAdd CAM_DEBUG_DELAY option.
gibbs [Mon, 16 Aug 1999 22:39:53 +0000 (22:39 +0000)]
Add CAM_DEBUG_DELAY option.

25 years agoCheckpoint work on handling error conditions.
gibbs [Mon, 16 Aug 1999 22:39:25 +0000 (22:39 +0000)]
Checkpoint work on handling error conditions.

Handle errors reported through immediate notify ccbs.  This includes
bus resets, BDR messages, and abort messages.

Properly setup CA conditions for errors such as uncorectable parity
in data phases, reselection failure, and IDE message received.

Add a mechanism for stalling the queue of transactions to a particular
initiator while a CA is in progress.  Since we don't yet support
tagged I/O, this is largely just for show right now.

Properly bzero our softc at init time (oops).

Add more documentation on what all of the queues in our softc do.

25 years agoxpt_release_devq now takes a count.
gibbs [Mon, 16 Aug 1999 22:23:01 +0000 (22:23 +0000)]
xpt_release_devq now takes a count.

25 years agoImplement the guts of the XPT_ABORT function code.
gibbs [Mon, 16 Aug 1999 22:22:41 +0000 (22:22 +0000)]
Implement the guts of the XPT_ABORT function code.

Clear residuals for XPT_TARGET_IO/XPT_CONT_TARGET_IO calls just
as we do for XPT_SCSI_IO calls.

Make xpt_release_devq a proper reciprocal of xpt_freeze_devq by
allowing the caller to decrement the count more than once per call.

Change the semantics of xpt_path_cmp to return 1 for a match
with a wildcard from path1 and 2 for a match with wildcards
from path2.  This change doesn't affect most callers as they
only check for >=0 (exact match or one containing a wildcard).

When cleaning up a device node, ensure that there are no timeouts
outstanding that reference it.

25 years agoAdd another flag for debugging just peripheral drivers. The whole debugging
gibbs [Mon, 16 Aug 1999 22:18:14 +0000 (22:18 +0000)]
Add another flag for debugging just peripheral drivers.  The whole debugging
mechanism should be completely  revamped, but this will serve for the time
being.

Take advantage of some new behavior in xpt_path_comp which allows the
caller to determine which path, if any, had wildcards in it that allowed
for a match to succeed.  We can now use this to filter out wildcard
nodes in the EDT.

Make the amount of time we pause after writing a debug message
configurable.

25 years agoAdd another command characteristic flag for commands that are queued
gibbs [Mon, 16 Aug 1999 22:14:47 +0000 (22:14 +0000)]
Add another command characteristic flag for commands that are queued
on physical devices as oposed to just in the XPT.  This is used to
mark XPT_SCSI_IO, XPT_RESET_DEV, XPT_ENG_EXEC, XPT_TARGET_IO, and
XPT_CONT_TARGET_IO function codes.

Change tag_id and init_id to be u_int from u_int8_t.  Some topologies
supported by CAM support much larger ID and tag ID address spaces
than parallel SCSI.

Kill some unused fields in the IMMED_NOTIFY ccb type.

25 years agoBad ref time(2) changed to time(3)
chris [Mon, 16 Aug 1999 22:11:19 +0000 (22:11 +0000)]
Bad ref time(2) changed to time(3)

25 years agoChange calling of chained event handler from always after
n_hibma [Mon, 16 Aug 1999 21:44:14 +0000 (21:44 +0000)]
Change calling of chained event handler from always after
the attach/detach to
1) MOD_LOAD before attach
2) MOD_UNLOAD after detach

The driver specific event handler can now be used to function as
driver specific init/deinit function (compare to device specific
init/deinit functions: attach & detach).

25 years agoRe-add -Wmissing-prototypes to `COPTS'. I mistook -W for -Wno in this case.
chris [Mon, 16 Aug 1999 21:22:39 +0000 (21:22 +0000)]
Re-add -Wmissing-prototypes to `COPTS'.  I mistook -W for -Wno in this case.

25 years agoFix some bad references.
chris [Mon, 16 Aug 1999 20:56:31 +0000 (20:56 +0000)]
Fix some bad references.

25 years agoClarify and cleanup some CAM queueing breakages.
mjacob [Mon, 16 Aug 1999 20:11:45 +0000 (20:11 +0000)]
Clarify and cleanup some CAM queueing breakages.

25 years agoadd in SIMQFRZ_TIMED && CMD_RQLATER defines
mjacob [Mon, 16 Aug 1999 20:04:06 +0000 (20:04 +0000)]
add in SIMQFRZ_TIMED && CMD_RQLATER defines

25 years agoadd in new ICB structure variant (for 2200)
mjacob [Mon, 16 Aug 1999 20:02:09 +0000 (20:02 +0000)]
add in new ICB structure variant (for 2200)

25 years agoClarify lun limits for FC && SCSI.
mjacob [Mon, 16 Aug 1999 20:01:14 +0000 (20:01 +0000)]
Clarify lun limits for FC && SCSI.

25 years agoMore code cleanup. Go back to using FULL_LOGIN Fibre Chan if f/w is less than
mjacob [Mon, 16 Aug 1999 19:59:55 +0000 (19:59 +0000)]
More code cleanup. Go back to using FULL_LOGIN Fibre Chan if f/w is less than
1.17.0 level. Change where we do the loop database init. Add in the CMD_RQLATER
return. Add some register debounce.

25 years agoSet some correct return values. Prefer I/O map all the time unless configured
mjacob [Mon, 16 Aug 1999 19:52:29 +0000 (19:52 +0000)]
Set some correct return values. Prefer I/O map all the time unless configured
otherwise.

25 years agoAdd virtual MAINTAINER line.
markm [Mon, 16 Aug 1999 19:05:02 +0000 (19:05 +0000)]
Add virtual MAINTAINER line.

25 years agoAccording to Mark Murray, Makefiles do not belong here. I guess we're
nsayer [Mon, 16 Aug 1999 18:59:05 +0000 (18:59 +0000)]
According to Mark Murray, Makefiles do not belong here. I guess we're
going to have to figure something else out.

25 years agovm_map_lock*:
alc [Mon, 16 Aug 1999 18:21:09 +0000 (18:21 +0000)]
vm_map_lock*:
Remove semicolons or add "do { } while (0)" as necessary
to enable the use of these macros in arbitrary statements.
(There are no functional changes.)

Submitted by: dillon

25 years agoexpand_name:
billf [Mon, 16 Aug 1999 18:13:39 +0000 (18:13 +0000)]
expand_name:
use pid_t and uid_t in the declaration as that is what we are passed
fix printf formatters accordingly.

Reviewed by: green

25 years agoPut in a limit (8) for lun width- do not exceed probing this unless a device
mjacob [Mon, 16 Aug 1999 17:47:39 +0000 (17:47 +0000)]
Put in a limit (8) for lun width- do not exceed probing this unless a device
is quirked for allowing it. This will allow SCSI cards to specify their true
correct lun width (32 for Qlogic SCSI, 16 or 65536 for Qlogic Fibre Channel)
without crippling the system because of f/w breakage of a lot of drives.
Reviewed by: gibbs@freebsd.org

25 years agoUse a colon (:) instead of a period (.) to distinguish the group name
sheldonh [Mon, 16 Aug 1999 16:00:23 +0000 (16:00 +0000)]
Use a colon (:) instead of a period (.) to distinguish the group name
from the user name when using chown(8).

PR: 13157
Reported by: Patrick Powell <papowell@astart.com>

25 years agoFix core dump with `fortune -f'.
tg [Mon, 16 Aug 1999 13:47:15 +0000 (13:47 +0000)]
Fix core dump with `fortune -f'.

25 years agoDescribe proper signals behavior rather than just removing it.
ru [Mon, 16 Aug 1999 12:36:45 +0000 (12:36 +0000)]
Describe proper signals behavior rather than just removing it.
Remove obsoleted reference to alarm(3).

Submitted by: bde

25 years agoImplement linux_vfork() syscall by calling vfork(). Analogous to the
marcel [Mon, 16 Aug 1999 11:49:30 +0000 (11:49 +0000)]
Implement linux_vfork() syscall by calling vfork(). Analogous to the
linux_fork() implementation.

25 years agoImplement vfork() syscall.
marcel [Mon, 16 Aug 1999 11:47:21 +0000 (11:47 +0000)]
Implement vfork() syscall.

25 years ago- Add an OPTIONAL_MANPATH directive (same as MANDATORY_MANPATH,
ru [Mon, 16 Aug 1999 11:34:57 +0000 (11:34 +0000)]
- Add an OPTIONAL_MANPATH directive (same as MANDATORY_MANPATH,
  except an absence of the directory is not considered an error
  and doesn't produce a warning).
  Put /usr/local/lib/perl5/*/man under OPTIONAL_MANPATH.

- An order of directives in manpath.config is now irrelevant.

- Get rid of infinite loop when PATH is unset or NULL, and
  MANDATORY_MANPATH directory doesn't exist.

- mdoc(9)ify manpage.

Reviewed by: des, markm, sheldonh

25 years agoAdd SRA authentication to src/crypto/telnet.
nsayer [Mon, 16 Aug 1999 11:24:29 +0000 (11:24 +0000)]
Add SRA authentication to src/crypto/telnet.

SRA does a Diffie-Hellmen exchange and then DES-encrypts the
authentication data. If the authentication is successful, it also
sets up a session key for DES encryption.

SRA was originally developed at Texas A&M University.

This code is probably export restricted (despite the fact that I
originally found it at a University in Germany).

SRA is not perfect. It is vulnerable to monkey-in-the-middle attacks
and does not use tremendously large DH constants (and thus an individual
exchange probably could be factored in a few days on modern CPU
horsepower). It does not, however, require any changes in user or
administrative behavior and foils session hijacking and sniffing.
The goal of this commit is that telnet and telnetd end up in the DES
distribution and that therefore an encrypted session telnet becomes
standard issue for FreeBSD.

25 years agoRev 1.36 also includes functional chagnes I intended to commit seperately.
obrien [Mon, 16 Aug 1999 09:55:16 +0000 (09:55 +0000)]
Rev 1.36 also includes functional chagnes I intended to commit seperately.
Configuration header inclusion has been moved around to reduce diffs from
the offical GCC distribution.  We now generate the same ``tm.h'' produced by
gcc's `configure' script [minus all the "#ifdef IN_GCC"'s].

Jeff Law of EGCS/Cygus says the new "approved" way of doing configure-related
includes is to list them all in ``tm.h'' rather than having the architure
config headers include large numbers of other configure headers.

25 years agoReplace our test(1) with NetBSD's pdksh-derived version. The code is
sheldonh [Mon, 16 Aug 1999 09:44:09 +0000 (09:44 +0000)]
Replace our test(1) with NetBSD's pdksh-derived version. The code is
significantly easier to read and extend and offers a few new tests.
A few style changes taken from style(9) and OpenBSD, as well as
whitespace cleanups.

This change was discussed on freebsd-committers and freebsd-hackers
and met with approval from at least des, eivind and brian.

PR: 13091
Obtained from: NetBSD

25 years agoDocument -i, -k, and -K options.
yokota [Mon, 16 Aug 1999 09:25:01 +0000 (09:25 +0000)]
Document -i, -k, and -K options.

25 years ago-Wmissing-prototypes is unnecessary here. There are no warnings produced
chris [Mon, 16 Aug 1999 08:13:17 +0000 (08:13 +0000)]
-Wmissing-prototypes is unnecessary here.  There are no warnings produced
with this option disabled.

25 years agoAdd French-Canadian keymap.
yokota [Mon, 16 Aug 1999 07:05:58 +0000 (07:05 +0000)]
Add French-Canadian keymap.
PR: misc/12615
Submitted by: Alexandre Normand <outcast@globetrotter.net>

25 years agoAdd prototypes for strl* functions
imp [Mon, 16 Aug 1999 06:53:13 +0000 (06:53 +0000)]
Add prototypes for strl* functions

Forgotten by: imp
Reported by: bde

25 years agoBad reference: tcpdump(8) -> tcpdump(1)
chris [Mon, 16 Aug 1999 05:36:11 +0000 (05:36 +0000)]
Bad reference: tcpdump(8) -> tcpdump(1)

25 years agofree_drive: mark drive down. This may fix a panic in the 'start' command.
grog [Mon, 16 Aug 1999 05:09:39 +0000 (05:09 +0000)]
free_drive: mark drive down.  This may fix a panic in the 'start' command.

Reported-by: peter

25 years agoClean up some comments.
grog [Mon, 16 Aug 1999 05:08:46 +0000 (05:08 +0000)]
Clean up some comments.

format_config: code preening.

vinum_scandisk: If we find a partition in the first pass over a drive,
note the fact so we don't grab the compatibility partition as well.

Submitted-by: peter
25 years ago{EGCSDIR}->{SRCDIR} + associated tweaks.
obrien [Mon, 16 Aug 1999 05:00:59 +0000 (05:00 +0000)]
{EGCSDIR}->{SRCDIR} + associated tweaks.
This is in prepreparation of the gcc-2.95 upgrade.

Also seems I was -I'ing several GCC directories I did not need to.

25 years ago{EGCSDIR}->{SRCDIR} + associated tweaks.
obrien [Mon, 16 Aug 1999 04:58:10 +0000 (04:58 +0000)]
{EGCSDIR}->{SRCDIR} + associated tweaks.
This is in prepreparation of the gcc-2.95 upgrade.

25 years agoPurely cosmetic changes -- fix Id's
obrien [Mon, 16 Aug 1999 04:27:28 +0000 (04:27 +0000)]
Purely cosmetic changes -- fix Id's

25 years agoPurely cosmetic cleanups.
obrien [Mon, 16 Aug 1999 04:10:48 +0000 (04:10 +0000)]
Purely cosmetic cleanups.
 - Id tag layout wrong.
 - "FOO =" -> "FOO=".
 - 4 space indention on continuation lines rather than a <tab>

Basic idea checked by: bde

25 years agoFix a long standing bug in the auto-termination control settings for
gibbs [Mon, 16 Aug 1999 02:33:46 +0000 (02:33 +0000)]
Fix a long standing bug in the auto-termination control settings for
the aic7890/91/96/97 cards.  This could cause the system to go into
a long retry/recovery loop during probe.

Fix the alignment argument to bus_dma_tag_create().

Don't set the CACHETHEN bit in dscommand0 for Ultra2 controllers
until we know more about its behavior.  The description for this
bit makes it sound like it could cause problems with certain
PCI chipsets.

25 years agoFix a long standing bug in the camq teardown code.
gibbs [Mon, 16 Aug 1999 02:23:42 +0000 (02:23 +0000)]
Fix a long standing bug in the camq teardown code.

25 years agoFix int function without return (make consistent with neighbors)
nsayer [Mon, 16 Aug 1999 02:15:29 +0000 (02:15 +0000)]
Fix int function without return (make consistent with neighbors)

25 years agoProperly set the alignment argument to bus_dma_tag_create(). If we
gibbs [Mon, 16 Aug 1999 01:52:21 +0000 (01:52 +0000)]
Properly set the alignment argument to bus_dma_tag_create().  If we
don't care about the alignment, set it to 1, meaning single byte alignment.

25 years agoFix a bug in busdma_mem_free() where we were improperly checking
gibbs [Mon, 16 Aug 1999 01:51:38 +0000 (01:51 +0000)]
Fix a bug in busdma_mem_free() where we were improperly checking
the map associated with the region to free.

25 years agoRemove the declarations for "vm_map_t io_map". It's been unused
alc [Sun, 15 Aug 1999 23:55:46 +0000 (23:55 +0000)]
Remove the declarations for "vm_map_t io_map".  It's been unused
since i386/i386/machdep rev 1.310, i.e., the demise of BOUNCE_BUFFERS.

25 years agoMove initialization of announce_buf up to make
mjacob [Sun, 15 Aug 1999 23:34:40 +0000 (23:34 +0000)]
Move initialization of announce_buf up to make
sure it is initialized in all cases.

25 years agoReduce the number of -I's. I have no idea why I thought I needed extra ones.
obrien [Sun, 15 Aug 1999 22:36:41 +0000 (22:36 +0000)]
Reduce the number of -I's.  I have no idea why I thought I needed extra ones.

Submitted by: bde

25 years agoMove "WORD_SWITCH_TAKES_ARG" & "SWITCH_TAKES_ARG" to the general spec file
obrien [Sun, 15 Aug 1999 22:02:50 +0000 (22:02 +0000)]
Move "WORD_SWITCH_TAKES_ARG" & "SWITCH_TAKES_ARG" to the general spec file
since their definitions were identical.  FreeBSD/MIPS will need to override
these two defintions, but it can #undef them and define them approapiately.

25 years agoMove "LIBGCC_SPEC" to freebsd-native.h as it is not approapiate for
obrien [Sun, 15 Aug 1999 21:59:46 +0000 (21:59 +0000)]
Move "LIBGCC_SPEC" to freebsd-native.h as it is not approapiate for
submitting to the GCC mainainers for folding into the stock GCC.

25 years agoRemove the declarations for "vm_map_t u_map". It's been unused
alc [Sun, 15 Aug 1999 21:55:20 +0000 (21:55 +0000)]
Remove the declarations for "vm_map_t u_map".  It's been unused
since i386/i386/pmap rev 1.190.  (The alpha never used it.)

25 years agoReverted to revision 1.8 as previous fix causes fts_open with with a
pho [Sun, 15 Aug 1999 19:21:29 +0000 (19:21 +0000)]
Reverted to revision 1.8 as previous fix causes fts_open with with a
path name argument with a trailing '/' to fail.

Reviewed by: phk

25 years agoMajor rewrite.
marcel [Sun, 15 Aug 1999 18:59:44 +0000 (18:59 +0000)]
Major rewrite.

The first reason for this rewrite is KNF conformance.

The second reason is to avoid redundancy. Each function printed the same
string, with only the syscall name being different. The actual printing is now
performed by a single function, which gets the syscall name as an argument.

The third reason is that of convenience. It's now very easy to add a new
dummy implementation. Just add ``DUMMY(foo);'' to the file. It's also a lot
easier now to see if a syscall has a dummy implementation or not.

The dummies are ordered on syscall number. Please maintain this when adding
new dummies (there're 32 candidates at the time of writing :-)

Reviewed by: bde

25 years agoProvide wrappers for sched_{s|g}etscheduler. We need to convert the policy
marcel [Sun, 15 Aug 1999 17:28:40 +0000 (17:28 +0000)]
Provide wrappers for sched_{s|g}etscheduler. We need to convert the policy
argument.

PR: 12006
Originator: Jean-Claude MICHOT <jcmichot@teaser.fr>

25 years agoProvide wrappers for sched_{s|g}etscheduler. We need to convert the policy
marcel [Sun, 15 Aug 1999 17:25:56 +0000 (17:25 +0000)]
Provide wrappers for sched_{s|g}etscheduler. We need to convert the policy
argument.

PR: 12006