]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
28 years agoWhoops - I made `make clean' chatty by mistake.
jkh [Thu, 20 Jun 1996 18:51:28 +0000 (18:51 +0000)]
Whoops - I made `make clean' chatty by mistake.

28 years agoAdd pre-world and post-world hooks so that people can do site specific
jkh [Thu, 20 Jun 1996 18:47:04 +0000 (18:47 +0000)]
Add pre-world and post-world hooks so that people can do site specific
cleanup or preinitializion of /usr/src before and after the world target
is run.

28 years agoMake pkg_install understand and use libftpio instead of its own
jkh [Thu, 20 Jun 1996 18:33:55 +0000 (18:33 +0000)]
Make pkg_install understand and use libftpio instead of its own
copies of the ftp support routines.  Also some cosmetic and minor
bug fixes I've been meaning to incorporate for awhile.

28 years agoRemove one last rip_output from inetsw (gpalmer missed it in rev 1.30)
fenner [Thu, 20 Jun 1996 17:52:32 +0000 (17:52 +0000)]
Remove one last rip_output from inetsw (gpalmer missed it in rev 1.30)

28 years agoadded fetch
adam [Thu, 20 Jun 1996 17:47:24 +0000 (17:47 +0000)]
added fetch

28 years agoUpdate net.inet.ip.forwarding example to give right value.
jkh [Thu, 20 Jun 1996 17:17:03 +0000 (17:17 +0000)]
Update net.inet.ip.forwarding example to give right value.
Indicate that gateway in /etc/sysconfig is *really* the correct way
if you have an /etc/sysconfig file.  Other small tweaks.
Requested by: Guy Helmer <ghelmer@alpha.dsu.edu>

28 years agoMake certain small things more consistent with the other stdio man pages.
jkh [Thu, 20 Jun 1996 15:49:54 +0000 (15:49 +0000)]
Make certain small things more consistent with the other stdio man pages.

28 years agoPut the 'debug' messages of the type:
nate [Thu, 20 Jun 1996 15:41:23 +0000 (15:41 +0000)]
Put the 'debug' messages of the type:
  /kernel: in_rtqtimo: adjusted rtq_reallyold to 1066
  /kernel: in_rtqtimo: adjusted rtq_reallyold to 710
inside of #ifdef DIAGNOSTIC to avoid the support questions from folks
asking what this means.

28 years agoI hate to read a man page that almost has useful information
jmacd [Thu, 20 Jun 1996 14:13:15 +0000 (14:13 +0000)]
I hate to read a man page that almost has useful information
but falls a little short.  I added a comment on the null
termination of struct group's gr_mem field.

Reviewed by: jkh

28 years agolibforms is no more
adam [Thu, 20 Jun 1996 13:38:15 +0000 (13:38 +0000)]
libforms is no more

28 years agoProperly account for non-page aligned buffers.
dg [Thu, 20 Jun 1996 08:07:30 +0000 (08:07 +0000)]
Properly account for non-page aligned buffers.

28 years agoMinor KNF formatting change to vmapbuf() and vunmapbuf().
dg [Thu, 20 Jun 1996 01:47:21 +0000 (01:47 +0000)]
Minor KNF formatting change to vmapbuf() and vunmapbuf().

28 years agoWhoops, forgot one FreeBSD specific change necessary for the lex .mk
nate [Wed, 19 Jun 1996 22:25:32 +0000 (22:25 +0000)]
Whoops, forgot one FreeBSD specific change necessary for the lex .mk
macros to work.

28 years agoDon't try to do a 'make bootstrap all install' in one commend. Break it
nate [Wed, 19 Jun 1996 21:19:56 +0000 (21:19 +0000)]
Don't try to do a 'make bootstrap all install' in one commend.  Break it
into two make commands so that make will recognize that it doesn't need
to buid scan.c w/lex.

28 years agoRemove libforms, it was never used.
jkh [Wed, 19 Jun 1996 21:19:37 +0000 (21:19 +0000)]
Remove libforms, it was never used.

28 years agoRemove libforms - it's unused.
jkh [Wed, 19 Jun 1996 21:18:24 +0000 (21:18 +0000)]
Remove libforms - it's unused.

28 years agoRemove ncftp
jkh [Wed, 19 Jun 1996 21:13:45 +0000 (21:13 +0000)]
Remove ncftp

28 years agoNuke this thing.
jkh [Wed, 19 Jun 1996 21:13:05 +0000 (21:13 +0000)]
Nuke this thing.

28 years agoAdded lex to the bootstrap target.
nate [Wed, 19 Jun 1996 20:48:48 +0000 (20:48 +0000)]
Added lex to the bootstrap target.

28 years agoFinish the import and merge in FreeBSD specific changes.
nate [Wed, 19 Jun 1996 20:48:06 +0000 (20:48 +0000)]
Finish the import and merge in FreeBSD specific changes.

Add a 'bootstrap' target which *must* be run before building the new
version, since the new scanner relies on the current version of flex to
build itself otherwise.

28 years agoAdd a newline - it seems to want it.
jkh [Wed, 19 Jun 1996 20:36:35 +0000 (20:36 +0000)]
Add a newline - it seems to want it.

28 years agoThis commit was generated by cvs2svn to compensate for changes in r16514,
nate [Wed, 19 Jun 1996 20:26:48 +0000 (20:26 +0000)]
This commit was generated by cvs2svn to compensate for changes in r16514,
which included commits to RCS files with non-trunk default branches.

28 years agoFlex version 2.5.3 from Vern Paxson at LBL.
nate [Wed, 19 Jun 1996 20:26:48 +0000 (20:26 +0000)]
Flex version 2.5.3 from Vern Paxson at LBL.

28 years agoPrint out the IRQ rather than the IRQ mask when printing out the
nate [Wed, 19 Jun 1996 17:27:55 +0000 (17:27 +0000)]
Print out the IRQ rather than the IRQ mask when printing out the
resources allocated to the driver when DEBUG is defined.

28 years agoSet IFF_RUNNING on the loopback interface.
wollman [Wed, 19 Jun 1996 16:24:10 +0000 (16:24 +0000)]
Set IFF_RUNNING on the loopback interface.

28 years agoMacro expressions should be fully parenthesized! Fix the MAXMEM
nate [Wed, 19 Jun 1996 15:37:52 +0000 (15:37 +0000)]
Macro expressions should be fully parenthesized!   Fix the MAXMEM
definition although it would work as it was written.

options        "MAXMEM=(128*1024)"

Suggested by: bde

28 years agoFix from Keith Bostic <bostic@bsdi.com> for bug in sed dealing with
dg [Wed, 19 Jun 1996 11:20:07 +0000 (11:20 +0000)]
Fix from Keith Bostic <bostic@bsdi.com> for bug in sed dealing with
continuation lines.

Submitted by: Keith Bostic via Kirk McKusick

28 years agoMerged in 4.4Lite2 versions, added $Id$'s. The ktrace changes should
jraynard [Wed, 19 Jun 1996 09:56:30 +0000 (09:56 +0000)]
Merged in 4.4Lite2 versions, added $Id$'s. The ktrace changes should
fix the garbage error messages printed out under certain circumstances.

28 years agoThe changes to go with ncftp -> fetch. Besides changing the names,
asami [Wed, 19 Jun 1996 09:33:31 +0000 (09:33 +0000)]
The changes to go with ncftp -> fetch.  Besides changing the names,
standardize the variables a little, and simplify some parts because
we can now trust the fetcher's return value.

Reviewed by: the list

28 years agoThis commit was generated by cvs2svn to compensate for changes in r16504,
asami [Wed, 19 Jun 1996 09:32:11 +0000 (09:32 +0000)]
This commit was generated by cvs2svn to compensate for changes in r16504,
which included commits to RCS files with non-trunk default branches.

28 years agoJean-Marc's url fetch program, with Josh MacDonald's patches and
asami [Wed, 19 Jun 1996 09:32:11 +0000 (09:32 +0000)]
Jean-Marc's url fetch program, with Josh MacDonald's patches and
Jordan's ftpio library.

Submitted by: jmz, jkh, jmacd (three-j!)

28 years agoSpit out the a proper port path for the describe target.
jkh [Wed, 19 Jun 1996 07:07:16 +0000 (07:07 +0000)]
Spit out the a proper port path for the describe target.
Submitted-By: Chuck Robey <chuckr@Glue.umd.edu>
28 years agoClean up vmapbuf and vunmapbuf significantly. The previous code was
dyson [Wed, 19 Jun 1996 03:39:24 +0000 (03:39 +0000)]
Clean up vmapbuf and vunmapbuf significantly.  The previous code was
very rough.

28 years agoSubmitted by: archie@whistle.com
julian [Wed, 19 Jun 1996 01:50:10 +0000 (01:50 +0000)]
Submitted by: archie@whistle.com

gary went a little overboard on commenting out unused variables.
Variables needed for ISO, LLC and NETATALK
were only enabled for ISO &  LLC.. so NETATALK bombed.

28 years agoAdd #include <err.h>.
alex [Wed, 19 Jun 1996 01:49:01 +0000 (01:49 +0000)]
Add #include <err.h>.

28 years agoRemove the hardcoding of the IRQ, and instead have it determine a free
nate [Wed, 19 Jun 1996 01:28:07 +0000 (01:28 +0000)]
Remove the hardcoding of the IRQ, and instead have it determine a free
IRQ.

For now only enable on the 3C589 until other cards have been tested.

28 years agoOk, now cleandir: is correct.
jkh [Wed, 19 Jun 1996 01:26:18 +0000 (01:26 +0000)]
Ok, now cleandir: is correct.

28 years agoEnable code that allows the daemon to select a free IRQ from the list if
nate [Wed, 19 Jun 1996 01:25:37 +0000 (01:25 +0000)]
Enable code that allows the daemon to select a free IRQ from the list if
none is selected.  The code was already there but needed to be updated
to use the correct structure element.

28 years agoUse irq_tok() instead of num_tok() for IRQ's. This will allow us to use
nate [Wed, 19 Jun 1996 01:08:58 +0000 (01:08 +0000)]
Use irq_tok() instead of num_tok() for IRQ's.  This will allow us to use
the '?' (undefined) token for stating IRQ's in the driver config line.

This will allow the card to automatically select and unused IRQ when
that code is enabled.

28 years agoAdd some unobtrusive changes to handle sources which aren't ports
jkh [Wed, 19 Jun 1996 01:04:23 +0000 (01:04 +0000)]
Add some unobtrusive changes to handle sources which aren't ports
(SRC_ENCAPSUATION).  Stick in some cd ${.CURDIR} directives which have
been (benignly) missing all this time.  Allow more types of targets to be
selectively disabled.

28 years agoAdd a better TEST script. I was surprised how much can be wrong
wosch [Wed, 19 Jun 1996 00:44:44 +0000 (00:44 +0000)]
Add a better TEST script. I was surprised how much can be wrong
with program test(1) or builtin test(1).

28 years agoStaticize and removed unused code.
nate [Tue, 18 Jun 1996 23:50:51 +0000 (23:50 +0000)]
Staticize and removed unused code.

28 years agoDocument MAXMEM option.
nate [Tue, 18 Jun 1996 23:21:28 +0000 (23:21 +0000)]
Document MAXMEM option.

[ Closes PR#1334, slightly modified by me ]

Submitted by: James Raynard <james@jraynard.demon.co.uk>

28 years agoUse CFLAGS += instead of setting it directly.
nate [Tue, 18 Jun 1996 22:54:25 +0000 (22:54 +0000)]
Use CFLAGS += instead of setting it directly.

28 years agoStaticize and remove unused function.
nate [Tue, 18 Jun 1996 21:58:55 +0000 (21:58 +0000)]
Staticize and remove unused function.

Inspired by: The Nomad code

28 years agoObtained from: Netatalk distribution.
julian [Tue, 18 Jun 1996 20:55:13 +0000 (20:55 +0000)]
Obtained from: Netatalk  distribution.

copyright for the appletalk stack  just to keep legal

28 years agoShut up a message which now appears to be a natural consequence of RIP2 and
jkh [Tue, 18 Jun 1996 19:59:35 +0000 (19:59 +0000)]
Shut up a message which now appears to be a natural consequence of RIP2 and
RIP1 routeds on the same subnet and I *don't* want filling up my syslog.

28 years ago- Removed dead code (if you need it you can get it out of the
nate [Tue, 18 Jun 1996 19:52:30 +0000 (19:52 +0000)]
- Removed dead code (if you need it you can get it out of the
  Repository).
- Added some comments, and moved some code around to make flow more
  obvious.

No functional changes.

28 years agoAdd new parts, kernel not booted in other case
ache [Tue, 18 Jun 1996 16:58:44 +0000 (16:58 +0000)]
Add new parts, kernel not booted in other case

28 years agoAdd tcl.
phk [Tue, 18 Jun 1996 16:53:59 +0000 (16:53 +0000)]
Add tcl.

28 years agoAdd a public domain notice to these files.
wollman [Tue, 18 Jun 1996 15:35:06 +0000 (15:35 +0000)]
Add a public domain notice to these files.

Requested by: jkh, 03 Sep 1995

28 years agoWhen bringing the netkey stuff over, I forgot that I had decided to change
wollman [Tue, 18 Jun 1996 15:22:26 +0000 (15:22 +0000)]
When bringing the netkey stuff over, I forgot that I had decided to change
AF_KEY into pseudo_AF_KEY, and defined PF_KEY incorrectly.  Fix.

Noticed by: pst

28 years agoAs we have appletalk protocol support we might as well show
julian [Tue, 18 Jun 1996 10:20:16 +0000 (10:20 +0000)]
As we have appletalk protocol support we might as well show
how to get it..

28 years agoremove some debugging printfs left in by mistake
julian [Tue, 18 Jun 1996 10:16:29 +0000 (10:16 +0000)]
remove some debugging printfs left in by mistake

28 years agoAdd procfs_type.c to the repository.
dyson [Tue, 18 Jun 1996 05:22:45 +0000 (05:22 +0000)]
Add procfs_type.c to the repository.

28 years agoAdd the file procfs_type.c to procfs.
dyson [Tue, 18 Jun 1996 05:19:45 +0000 (05:19 +0000)]
Add the file procfs_type.c to procfs.

28 years agoClean-up the new VM map procfs code, and also add support for executable
dyson [Tue, 18 Jun 1996 05:16:00 +0000 (05:16 +0000)]
Clean-up the new VM map procfs code, and also add support for executable
format file "etype".  It contains a description of the binary type for
a process.

28 years agoSet the program name before trying to use it.
alex [Tue, 18 Jun 1996 01:46:34 +0000 (01:46 +0000)]
Set the program name before trying to use it.

Found by: Aage Robekk <aagero@aage.priv.no>

28 years agoRemoved unused #includes of <i386/isa/icu.h> and <i386/isa/icu.h>. icu.h
bde [Tue, 18 Jun 1996 01:22:40 +0000 (01:22 +0000)]
Removed unused #includes of <i386/isa/icu.h> and <i386/isa/icu.h>.  icu.h
is only used by the icu support modules and by a few drivers that know
too much about the icu (most only use it to convert `n' to `IRQn').  isa.h
is only used by ioconf.c and by a few drivers that know too much about
isa addresses (a few have to, because config is deficient).

28 years agoManage control connections a little better for the URL routines.
jkh [Mon, 17 Jun 1996 23:16:04 +0000 (23:16 +0000)]
Manage control connections a little better for the URL routines.
This will do as a stop-gap until I figure out a more fault-tolerant
way of having deferred closes against the control connection work
without blocking.

28 years agoack - left in a typo. Time for bed!
jkh [Mon, 17 Jun 1996 23:04:24 +0000 (23:04 +0000)]
ack - left in a typo.  Time for bed!

28 years agoThis file is the "meat" of the process address space capability. If you
dyson [Mon, 17 Jun 1996 22:53:27 +0000 (22:53 +0000)]
This file is the "meat" of the process address space capability.  If you
would like other things added, just ask!!!  It might be pretty easy to add.

28 years agoAdd a feature to procfs to allow display of the process address map
dyson [Mon, 17 Jun 1996 22:43:36 +0000 (22:43 +0000)]
Add a feature to procfs to allow display of the process address map
with multiple entries as follows:

start address, end address, resident pages in range, private pages
in range, RW/RO, COW or not, (vnode/device/swap/default).

28 years agoPrepend the CIS string length define with CIS_ per the rest of the
nate [Mon, 17 Jun 1996 22:30:30 +0000 (22:30 +0000)]
Prepend the CIS string length define with CIS_ per the rest of the
constants.

Obtained from: Nomad code

28 years agoUse the defined path to the card devices instead of hard-coding it's
nate [Mon, 17 Jun 1996 22:14:19 +0000 (22:14 +0000)]
Use the defined path to the card devices instead of hard-coding it's
value in every file.

No functional changes.

28 years agoMake binary mode the default.
jkh [Mon, 17 Jun 1996 22:10:15 +0000 (22:10 +0000)]
Make binary mode the default.

28 years agoInit network at proper time.
jkh [Mon, 17 Jun 1996 21:48:33 +0000 (21:48 +0000)]
Init network at proper time.

28 years agoStamp out a potential memory leak.
jkh [Mon, 17 Jun 1996 20:36:57 +0000 (20:36 +0000)]
Stamp out a potential memory leak.
Make ftpChdir return the server status again - it was more convenient.

28 years agoBring in libftpio.
jkh [Mon, 17 Jun 1996 20:18:11 +0000 (20:18 +0000)]
Bring in libftpio.

28 years agoAdd RCS Id and fix date.
wollman [Mon, 17 Jun 1996 20:12:53 +0000 (20:12 +0000)]
Add RCS Id and fix date.

28 years agoCross-reference addr2ascii(3).
wollman [Mon, 17 Jun 1996 20:11:12 +0000 (20:11 +0000)]
Cross-reference addr2ascii(3).

28 years agoDo not print herror message, if w -n gethostname failed to be 100%
ache [Mon, 17 Jun 1996 20:07:44 +0000 (20:07 +0000)]
Do not print herror message, if w -n gethostname failed to be 100%
compatible with any program
Suggested by: peter

28 years agoCorrect date and add $Id$ to reflect previous modification.
wollman [Mon, 17 Jun 1996 20:04:06 +0000 (20:04 +0000)]
Correct date and add $Id$ to reflect previous modification.

28 years agoCross-reference addr2ascii(3) and inlcude a note about how those
wollman [Mon, 17 Jun 1996 20:01:05 +0000 (20:01 +0000)]
Cross-reference addr2ascii(3) and inlcude a note about how those
functions are preferred but are not widely available as yet.

28 years agoAdd keyadmin directory.
wollman [Mon, 17 Jun 1996 19:49:17 +0000 (19:49 +0000)]
Add keyadmin directory.

28 years agoThis commit was generated by cvs2svn to compensate for changes in r16449,
wollman [Mon, 17 Jun 1996 19:47:57 +0000 (19:47 +0000)]
This commit was generated by cvs2svn to compensate for changes in r16449,
which included commits to RCS files with non-trunk default branches.

28 years agoThis is the `key' program from NRL's IPv6 distribution, heavily
wollman [Mon, 17 Jun 1996 19:47:57 +0000 (19:47 +0000)]
This is the `key' program from NRL's IPv6 distribution, heavily
hacked up by me to remove the IPv6 stuff (fow now).  I renamed
it `keyadmin' since `key' was already taken by the S/Key calculator.
Its purpose is to act as a command-driven interface to the `PF_KEY'
socket domain, analogously to thefunction of route(8) in the
`PF_ROUTE' domain.  This program is believed to be exportable, since
it does no actual cryptography itself.

28 years agoWhups! Slight adjustment.
jkh [Mon, 17 Jun 1996 19:33:27 +0000 (19:33 +0000)]
Whups!  Slight adjustment.

28 years agoAdd more items to the glossary menu and reformat it slightly.
jkh [Mon, 17 Jun 1996 19:31:34 +0000 (19:31 +0000)]
Add more items to the glossary menu and reformat it slightly.

28 years agoFix xterm/screen/(any program using just :xxx form) compatibility
ache [Mon, 17 Jun 1996 19:23:38 +0000 (19:23 +0000)]
Fix xterm/screen/(any program using just :xxx form) compatibility

28 years agoRemove advertising actions requirement, they are too bothering for anyone
ache [Mon, 17 Jun 1996 19:05:38 +0000 (19:05 +0000)]
Remove advertising actions requirement, they are too bothering for anyone

28 years agoThis time, get rid of the struct copies that were really causing gcc to
peter [Mon, 17 Jun 1996 18:52:53 +0000 (18:52 +0000)]
This time, get rid of the struct copies that were really causing gcc to
call memcpy..  It seems that gcc would not inline the implicit call
when copying from a volatile...

28 years agoDon't try convert "-" to numeric form for -n option
ache [Mon, 17 Jun 1996 18:35:16 +0000 (18:35 +0000)]
Don't try convert "-" to numeric form for -n option

28 years agoUpdate Rod's comments on chipsets to include the Triton II.
jkh [Mon, 17 Jun 1996 17:44:01 +0000 (17:44 +0000)]
Update Rod's comments on chipsets to include the Triton II.

28 years agoStomp another compiler warning: sc_devfs_token[] should only be declared
wpaul [Mon, 17 Jun 1996 17:21:35 +0000 (17:21 +0000)]
Stomp another compiler warning: sc_devfs_token[] should only be declared
if DEVFS is #defined.

28 years agoAdd a couple of #ifdef DEVFS/#endif clauses to slence the following
wpaul [Mon, 17 Jun 1996 16:54:03 +0000 (16:54 +0000)]
Add a couple of #ifdef DEVFS/#endif clauses to slence the following
compiler warnings which occur if you don't have 'options DEVFS' in
your kernel config file:

../../kern/kern_descrip.c: In function `fildesc_drvinit':
../../kern/kern_descrip.c:1103: warning: unused variable `fd'
../../kern/kern_descrip.c: At top level:
../../kern/kern_descrip.c:1095: warning: `devfs_token_stdin' defined but not use
d
../../kern/kern_descrip.c:1096: warning: `devfs_token_stdout' defined but not us
ed
../../kern/kern_descrip.c:1097: warning: `devfs_token_stderr' defined but not us
ed
../../kern/kern_descrip.c:1098: warning: `devfs_token_fildesc' defined but not u
sed

28 years agoReduced nesting of #includes in random.h and adjusted isa/random_machdep.c
bde [Mon, 17 Jun 1996 16:47:43 +0000 (16:47 +0000)]
Reduced nesting of #includes in random.h and adjusted isa/random_machdep.c
to match (pc98/random_machdep.c probably requires a similar change).  This
is a problem area for the PC98 merge - all PC98 ifdefs in <machine/*.h> are
kludges to work around incorrect layering.

28 years agoLocalize time
ache [Mon, 17 Jun 1996 16:44:16 +0000 (16:44 +0000)]
Localize time

28 years agoAdd support for ${CLEANDIRS} that will be rm -rf'ed during clean & cleandir.
phk [Mon, 17 Jun 1996 15:59:52 +0000 (15:59 +0000)]
Add support for ${CLEANDIRS} that will be rm -rf'ed during clean & cleandir.

28 years agoFix w -n to really show numeric addresses, close PR 402
ache [Mon, 17 Jun 1996 15:46:25 +0000 (15:46 +0000)]
Fix w -n to really show numeric addresses, close PR 402
Partially submitted by: henrich@msu.edu (Charles Henrich)

28 years agoRethink and reimpliment the way RESTARTS are handled. The method I inheirited
jkh [Mon, 17 Jun 1996 15:28:08 +0000 (15:28 +0000)]
Rethink and reimpliment the way RESTARTS are handled.  The method I inheirited
from jmz was a hopeless kludge (sorry Jean-Marc :) and handled the problem
in the wrong way.  ftpRestart() has now gone away and ftpGet() has grown a
new parameter.

28 years agoSynced with Makefile.i386: added -Wunused.
bde [Mon, 17 Jun 1996 15:11:10 +0000 (15:11 +0000)]
Synced with Makefile.i386: added -Wunused.

28 years agoIf hostname > UT_HOSTSIZE, use its numerical address instead to keep
ache [Mon, 17 Jun 1996 14:59:07 +0000 (14:59 +0000)]
If hostname > UT_HOSTSIZE, use its numerical address instead to keep
valid utmp and wtmp entries

28 years agoIf hostname > UT_HOSTSIZE, use its numeric address instead to keep
ache [Mon, 17 Jun 1996 14:49:02 +0000 (14:49 +0000)]
If hostname > UT_HOSTSIZE, use its numeric address instead to keep
valid entries into utmp and wtmp

28 years agoMoved initialization of defaults for the label for the whole disk from
bde [Mon, 17 Jun 1996 14:43:54 +0000 (14:43 +0000)]
Moved initialization of defaults for the label for the whole disk from
disklabel(8) to the kernel (dsopen()).  Drivers should initialize the
hardware values (rpm, interleave, skews).  Drivers currently don't do
this, but it usually doesn't matter since rotational position stuff is
normally disabled.

28 years agoAdded support for TIOCDCDTIMESTAMP (enable/get timestamp of last DCD rise).
bde [Mon, 17 Jun 1996 14:23:39 +0000 (14:23 +0000)]
Added support for TIOCDCDTIMESTAMP (enable/get timestamp of last DCD rise).
Original version by John Hay.

Simplified timestamp code by reading the time exactly when necessary.
This may slow down the interrupt handler with extra calls to microtime(),
but only in bad configurations - the input fifo should normally be
disabled if timestamps on input are being used, since otherwise the
timestamp won't be precisely associated with any particular input event.
The interrupt handler remains slowed down by one test and branch for
each input (and now DCD change) event - avoiding this is not practical
yet.

The simplifications also fixed:
- timestamps for input sometimes being clobbered by output and modem
  status interrupts.
- valid timestamps not being available unless the port is configured with
  vector siointrts.  siointrts no longer exists.
- compiler warnings about siointr* in some configurations.

Simplified timestamp and probe code by depending on recent changes in
microtime() and DELAY() to preserve the interrupt enable flag.

28 years agoAdded TIOCDCDTIMESTAMP (enable/get timestamp of last DCD rise).
bde [Mon, 17 Jun 1996 13:08:09 +0000 (13:08 +0000)]
Added TIOCDCDTIMESTAMP (enable/get timestamp of last DCD rise).

Reformatted FreeBSD additions in a consistent style.

28 years agoIn getit(), use read_eflags()/write_eflags() to preserve the interrupt
bde [Mon, 17 Jun 1996 12:50:22 +0000 (12:50 +0000)]
In getit(), use read_eflags()/write_eflags() to preserve the interrupt
enable flag instead of enable_intr() to restore it to its usual state.
getit() is only called from DELAY() so there is no point in optimising
its speed (this wasn't so clear when it was extern), and using
enable_intr() made it inconvenient to call DELAY() from probes that need
to run with interrupts disabled.

28 years agoIn microtime(), use pushfl/popfl to preserve the interrupt enable flag
bde [Mon, 17 Jun 1996 12:43:04 +0000 (12:43 +0000)]
In microtime(), use pushfl/popfl to preserve the interrupt enable flag
instead of sti to it restore to its usual state.  pushfl/popfl is
actually faster in protected mode on Pentiums (4+3 cycles instead of 9),
and using sti made it extremely inconvenient to call microtime() from
fast interrupt handlers.  pushfl/popfl is a couple of cycles slower than
sti on 486's and a couple more cycles slower on 386's, but the relative
cost of using it is not large since microtime() has to use slow i/o
instructions on the old cpus.

28 years agoAdd a feature: If the environment variable FTP_PASSIVE_MODE is defined
jkh [Mon, 17 Jun 1996 12:42:33 +0000 (12:42 +0000)]
Add a feature:  If the environment variable FTP_PASSIVE_MODE is defined
(the convention as established by pkg_install(1)), select passive mode
FTP automatically.