]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
24 years agoThis commit was generated by cvs2svn to compensate for changes in r59024,
David E. O'Brien [Wed, 5 Apr 2000 04:09:32 +0000 (04:09 +0000)]
This commit was generated by cvs2svn to compensate for changes in r59024,
which included commits to RCS files with non-trunk default branches.

24 years agoImport the Sparc bits of GNU binutils 2.9.1.
David E. O'Brien [Wed, 5 Apr 2000 04:09:32 +0000 (04:09 +0000)]
Import the Sparc bits of GNU binutils 2.9.1.

Requested by: steve

24 years agodecrease buffer size to 4k
Cameron Grant [Wed, 5 Apr 2000 02:14:26 +0000 (02:14 +0000)]
decrease buffer size to 4k
use a seperate channel for generating irqs so we get a buffer-half-empty
interrupt, prevents repeating on underflow

24 years agoAlways give mktime a timeval with tm_isdst set to -1, otherwise
Brian Somers [Wed, 5 Apr 2000 01:59:36 +0000 (01:59 +0000)]
Always give mktime a timeval with tm_isdst set to -1, otherwise
things get a bit out-of-phase when we step backwards 1 hour from
between 0:00 and 1:00 on the first of the month following the
transition into Summer time.  This is probably actually a bug
in mktime().

PR: 10963

If mktime() fails and 68 < year < 138, assume that the reason is
because of Summer time and adjust up or down according to our
adjusting context by one hour.  This assumes that all DSTs are
multiples of 1 hour.

PR: 6223, 17750

24 years agoallow /dev/dsp to be opened seperately for reading and writing.
Cameron Grant [Wed, 5 Apr 2000 01:13:42 +0000 (01:13 +0000)]
allow /dev/dsp to be opened seperately for reading and writing.

24 years agotry increasing 2ndbuffer size to 64k
Cameron Grant [Wed, 5 Apr 2000 00:56:02 +0000 (00:56 +0000)]
try increasing 2ndbuffer size to 64k

24 years agomove buffer size definition to .c file for consistency, and reduce it to 4k
Cameron Grant [Wed, 5 Apr 2000 00:38:00 +0000 (00:38 +0000)]
move buffer size definition to .c file for consistency, and reduce it to 4k

24 years agoAdd automatic rescan on connect. This happens already if the drive is
Nick Hibma [Tue, 4 Apr 2000 22:15:13 +0000 (22:15 +0000)]
Add automatic rescan on connect. This happens already if the drive is
connected during boot and the vpo module is in the kernel/loaded.

In the case where the module is loaded at some later stage with kldload,
the CAM bus is now rescanned as well.

24 years agoFix _zget() so that it checks the return from kmem_alloc(), to avoid
Mike Smith [Tue, 4 Apr 2000 21:00:39 +0000 (21:00 +0000)]
Fix _zget() so that it checks the return from kmem_alloc(), to avoid
zttempting to bzero NULL when the kernel map fills up.  _zget() will
now return NULL as it seems it was originally intended to do.

24 years agoAdd information about the dangarous apollo MPV3 chipset. Several
Warner Losh [Tue, 4 Apr 2000 17:14:18 +0000 (17:14 +0000)]
Add information about the dangarous apollo MPV3 chipset.  Several
people have suggested the workaround to me (use only PIO mode) and
I've not seend a message saying ths was fixed.  I'll removed the bold
warning stars after the issue is corrected.

24 years agoUpgrade the bktr driver from 2.07 to 2.10.
Roger Hardiman [Tue, 4 Apr 2000 16:54:13 +0000 (16:54 +0000)]
Upgrade the bktr driver from 2.07 to 2.10.

 - The driver now uses bus_space() and runs on NetBSD 1.4.2
   Submitted by Thomas Klausner <wiz@danbala.ifoer.tuwien.ac.at>

 - Remove startup quirks for video and vbi capture for PAL users.
   PAL TV users can now run FXTV and Alevt in any order.

 - Add support for cable channels >100
   Submitted by Scott Presnell <srp@zgi.com>

 - New MSP3410/3415 setup code added. This is experimental.
   Please set the sysctl hw.bt848.slow_msp_audio to 1 for this.
   Submitted by Frank Nobis<fn@radio-do.de>

24 years agoAdd multiple #include protection
Roger Hardiman [Tue, 4 Apr 2000 16:46:19 +0000 (16:46 +0000)]
Add multiple #include protection

Submitted by:  Thomas Klausner <wiz@danbala.ifoer.tuwien.ac.at>

24 years agoWhitespace change:
Sheldon Hearn [Tue, 4 Apr 2000 15:32:37 +0000 (15:32 +0000)]
Whitespace change:

Remove single-space hard sentence break introduced in the previous
commit.

Reviewed by: jmb

24 years agoFixed prototype for setflags(). setflags() returns int, not u_long,
Bruce Evans [Tue, 4 Apr 2000 14:12:35 +0000 (14:12 +0000)]
Fixed prototype for setflags().  setflags() returns int, not u_long,
and "extern" in function prototypes is a style bug.  The type mismatch
broke chflags(1) on i386's with 64-bit longs and may have broken it on
alphas.

24 years agoRemove obsolete reference to PCVT_FREEBSD.
Hellmuth Michaelis [Tue, 4 Apr 2000 13:14:43 +0000 (13:14 +0000)]
Remove obsolete reference to PCVT_FREEBSD.

24 years agoAdd a '+' to the first LDADD, otherwise it clobbers the LDADD in
Gary Jennejohn [Tue, 4 Apr 2000 11:12:55 +0000 (11:12 +0000)]
Add a '+' to the first LDADD, otherwise it clobbers the LDADD in
Makefile.alpha such that the -lkvm disappears and the link fails.

This time I tested it, honest :)

24 years agoAdd an example of how to tweak CXXFLAGS in a make.conf file.
David E. O'Brien [Tue, 4 Apr 2000 09:27:59 +0000 (09:27 +0000)]
Add an example of how to tweak CXXFLAGS in a make.conf file.

This is in responce to several people breaking their world builds by
incorrectly assigning to CXXFLAGS.

24 years agoPass me the pointy hat.
Gary Jennejohn [Tue, 4 Apr 2000 09:20:53 +0000 (09:20 +0000)]
Pass me the pointy hat.

It was not a good idea to remove csu_header from struct cspace, it had
ramifications which I didn't notice.

Restore src/usr.sbin/ppp/slcompress.h to the way it was, since MAX_HDR
was already defined as 128 there and it's a user program anyway.

In sys/net/slcompress.h make MAX_HDR 128 intead of MLEN to avoid
bloat.

My apologies for any inconvenience.

24 years ago- implement logfile archivation to a separate, user
Hellmuth Michaelis [Tue, 4 Apr 2000 08:50:01 +0000 (08:50 +0000)]
- implement logfile archivation to a separate, user
  configurable directory
- implement alternate and more flexible way to specify
  logfile rotation time in addition to the ISO 8601
  restricted format
- cleanup the source which was a mix of several styles
  of persons who maintained it so far, ran through
  knfom script got from bde.

Reviewed by: (in part) sheldonh and garyj

24 years agoRun newsyslog source through knfom script obtained from Bruce Evans.
Hellmuth Michaelis [Tue, 4 Apr 2000 08:38:30 +0000 (08:38 +0000)]
Run newsyslog source through knfom script obtained from Bruce Evans.
Joerg Wunsch suggested to do this to make the functional changes in
the next commit to newsyslog (which were run through the same script)
better visible.

24 years agoAllow MAXDEP dependants, not MAXDEP-1.
Doug Rabson [Tue, 4 Apr 2000 07:48:04 +0000 (07:48 +0000)]
Allow MAXDEP dependants, not MAXDEP-1.

24 years agoAdd libcrypto to LDADD. This fixes problems seen with e.g. apache-modssl
Kris Kennaway [Tue, 4 Apr 2000 07:31:01 +0000 (07:31 +0000)]
Add libcrypto to LDADD. This fixes problems seen with e.g. apache-modssl

Submitted by: Jim Bloom <bloom@acm.org>

24 years agoAdd a missing PXE API call and structure.
Paul Saab [Tue, 4 Apr 2000 07:28:54 +0000 (07:28 +0000)]
Add a missing PXE API call and structure.

24 years agoFix mbuf leaking - duh!
Duncan Barclay [Tue, 4 Apr 2000 06:43:30 +0000 (06:43 +0000)]
Fix mbuf leaking - duh!

24 years agoFix pcic_detach_socket to get right pcic_handle.
Warner Losh [Tue, 4 Apr 2000 04:12:43 +0000 (04:12 +0000)]
Fix pcic_detach_socket to get right pcic_handle.
Pass sc->dev rather than a bogusly cast pccard_softc *sc.

This allows us to insert and remove cards w/o panicing the kernel.
However, the cis isn't mapped in, so the pccard_scan_cis function
fails.

24 years agoBe consistant about using the right child to get the ivars, which is
Warner Losh [Tue, 4 Apr 2000 03:47:51 +0000 (03:47 +0000)]
Be consistant about using the right child to get the ivars, which is
the pcic_handle structure for the socket.

24 years agoenable emu10k1 (sblive) driver, though still playback only
Cameron Grant [Tue, 4 Apr 2000 02:34:45 +0000 (02:34 +0000)]
enable emu10k1 (sblive) driver, though still playback only

24 years agomake the click on starting output go away
Cameron Grant [Tue, 4 Apr 2000 02:16:15 +0000 (02:16 +0000)]
make the click on starting output go away

24 years agoAdd all the PXE related structures from the 2.1 PXE spec from Intel.
Paul Saab [Tue, 4 Apr 2000 00:38:59 +0000 (00:38 +0000)]
Add all the PXE related structures from the 2.1 PXE spec from Intel.
Fix exsisting code to match the spec.

24 years agoDereference a pointer (print int instead of pointer value)
Nick Hibma [Mon, 3 Apr 2000 22:28:13 +0000 (22:28 +0000)]
Dereference a pointer (print int instead of pointer value)

24 years agoNumerous fixes brought in by Lennart Augustsson from NetBSD:
Nick Hibma [Mon, 3 Apr 2000 22:19:44 +0000 (22:19 +0000)]
Numerous fixes brought in by Lennart Augustsson from NetBSD:

- Request the maximum number of LUNs on a device if it is
  a Bulk-Only device.

- Handle NO_TEST_UNIT_READY case for SCSI

- Add NO_START_STOP quirk for LS-120

- Fix a KASSERT which was the wrong way around.

- Kickstart the LS-120 with the infamous LS-120 specific command of
  which no one knows what it doesn, apart from the fact that it appeared
  in the Linux code someday.

24 years agoDocument that the flags can be specified as an octal number.
John Polstra [Mon, 3 Apr 2000 22:10:58 +0000 (22:10 +0000)]
Document that the flags can be specified as an octal number.

24 years agoOpenBSD has a broken debugger that does not grok static. Use a
Nick Hibma [Mon, 3 Apr 2000 20:58:30 +0000 (20:58 +0000)]
OpenBSD has a broken debugger that does not grok static. Use a

        #define Static  static

that the OpenBSD folks can define it to be empty if they like.

24 years agoPlug two file descriptor leaks
Brian Somers [Mon, 3 Apr 2000 19:56:49 +0000 (19:56 +0000)]
Plug two file descriptor leaks

24 years agoNuke csu_hdr from struct cspace. csu_hdr is not used anywhere in the
Gary Jennejohn [Mon, 3 Apr 2000 19:54:33 +0000 (19:54 +0000)]
Nuke csu_hdr from struct cspace. csu_hdr is not used anywhere in the
tree. This considerably reduces unnecessary bloat in struct slcompress.

I'm running with this change right now and have seen no negative
side-effects.

On my sytem this reduced kernel BSS by about 25KB.

Submitted by: bde
Approved by: brian for user-ppp

24 years agoAdd note on SCSI floppy on DEC3000 series
Wilko Bulte [Mon, 3 Apr 2000 19:32:22 +0000 (19:32 +0000)]
Add note on SCSI floppy on DEC3000 series

24 years agoEnable etc/defaults/pccard.conf which is default configuration file
Mitsuru IWASAKI [Mon, 3 Apr 2000 19:24:25 +0000 (19:24 +0000)]
Enable etc/defaults/pccard.conf which is default configuration file
for pccardd.
Please install /etc/defaults/pccard.conf and update /etc/defaults/rc.conf
as well.
Note that old pccard.conf.sample still remains for while but
no longer to be maintained.

Reviewed by: imp, -mobile ML and nomads ML in Japan.

24 years agoRegen.
Nick Hibma [Mon, 3 Apr 2000 15:15:50 +0000 (15:15 +0000)]
Regen.

24 years agoAdd Freecom + product; Update NetBSD's copyright
Nick Hibma [Mon, 3 Apr 2000 14:47:43 +0000 (14:47 +0000)]
Add Freecom + product; Update NetBSD's copyright

24 years agoExplain static numbering.
Sheldon Hearn [Mon, 3 Apr 2000 13:07:33 +0000 (13:07 +0000)]
Explain static numbering.

Reviewed by: asmodai

24 years agoEntry for Zip drive no longer needed. We do automatic rescans on connect
Nick Hibma [Mon, 3 Apr 2000 11:16:41 +0000 (11:16 +0000)]
Entry for Zip drive no longer needed. We do automatic rescans on connect
now.

24 years agoAdd a hack to cam that makes the cam_xpt available to the rest of the
Nick Hibma [Mon, 3 Apr 2000 11:11:12 +0000 (11:11 +0000)]
Add a hack to cam that makes the cam_xpt available to the rest of the
kernel. Justin agress that there is no other reasonable alternative to
do automatic rescans on connect.

The problem is that when a new device attaches to a SIM (SCSI host
controller) we need to send a XPT_SCAN_BUS command to the SIM using
xpt_action. This requires however that there is a peripheral available
to take the command (otherwise xpt_done and later bomb). The RESCAN
ioctl uses the same periph.

This enables a USB mass storage drive to do an automatic rescan on
connection of the drive.

The automatic dropping of a CAM entry on disconnection was already
working (asynchronous event).

The next thing to do is find someone to commit a change to vpo to do the
same thing. Just port umass_cam_rescan and friends across to that
driver.

Approved by:   gibbs

24 years agoSilence some warnings.
Poul-Henning Kamp [Mon, 3 Apr 2000 10:09:18 +0000 (10:09 +0000)]
Silence some warnings.

24 years agoFix unsafe tempfile creation. This file is already off the vendor
Sheldon Hearn [Mon, 3 Apr 2000 09:49:49 +0000 (09:49 +0000)]
Fix unsafe tempfile creation.  This file is already off the vendor
branch.  Although this problem has been reported to the GNU folks,
it's unlikely that any solution they may come up with will involve
the use of mktemp(1).

PR: 16942
Submitted by: Colin Phipps <crp22@cam.ac.uk>

24 years agoBrush up the last commit.
Kazutaka YOKOTA [Mon, 3 Apr 2000 09:40:04 +0000 (09:40 +0000)]
Brush up the last commit.

- Change some expressions.
- Wrap "resume" in single quotes to maintain consistency.

Submitted by: sheldonh

24 years agoBrush up the last commit.
Kazutaka YOKOTA [Mon, 3 Apr 2000 09:38:52 +0000 (09:38 +0000)]
Brush up the last commit.

- Expand contractions.
- Change some expressions.
- Reword the sections that describe the reporting of down events.

Submitted by: sheldonh

24 years agoUnbreak LINT.
Kazutaka YOKOTA [Mon, 3 Apr 2000 09:24:16 +0000 (09:24 +0000)]
Unbreak LINT.

24 years agoregenerate with MPSAFE from syscalls.master
Alfred Perlstein [Mon, 3 Apr 2000 06:36:57 +0000 (06:36 +0000)]
regenerate with MPSAFE from syscalls.master

24 years agoMake makesyscalls.sh parse an optional field 'MPSAFE' that specifies
Alfred Perlstein [Mon, 3 Apr 2000 06:36:14 +0000 (06:36 +0000)]
Make makesyscalls.sh parse an optional field 'MPSAFE' that specifies
that a syscall does not want the BGL to be grabbed automatically.

Add the new MPSAFE flag to the syscalls that dillon has determined to
be MPSAFE.

24 years agoRemove the !(I386 & SMP) tests; we don't run SMP on an i386 system, and
Mike Smith [Mon, 3 Apr 2000 06:02:00 +0000 (06:02 +0000)]
Remove the !(I386 & SMP) tests; we don't run SMP on an i386 system, and
they break the LINT build.

24 years agoAdd support for sunos_big_vec to support 'make world's on Solaris
Steve Price [Mon, 3 Apr 2000 03:51:46 +0000 (03:51 +0000)]
Add support for sunos_big_vec to support 'make world's on Solaris
boxes.

24 years agoAdd support for the sunos_big_vec format as well.
Steve Price [Mon, 3 Apr 2000 03:48:49 +0000 (03:48 +0000)]
Add support for the sunos_big_vec format as well.

24 years agoBuild and install the elf32_sparc ld(1) scripts and make elf32_sparc
Steve Price [Mon, 3 Apr 2000 03:42:07 +0000 (03:42 +0000)]
Build and install the elf32_sparc ld(1) scripts and make elf32_sparc
the default emulation.

24 years agofix missing defines and prototype for emu_vdump()
Cameron Grant [Mon, 3 Apr 2000 00:28:43 +0000 (00:28 +0000)]
fix missing defines and prototype for emu_vdump()

24 years agoSync with NetBSD: Add a few id's and correct numerous typo's
Nick Hibma [Mon, 3 Apr 2000 00:07:43 +0000 (00:07 +0000)]
Sync with NetBSD: Add a few id's and correct numerous typo's

24 years ago- Add HPFS and NWFS vnode tags.
Chris Costello [Sun, 2 Apr 2000 23:53:57 +0000 (23:53 +0000)]
- Add HPFS and NWFS vnode tags.
- Update ``struct vnode'' changes from revisions 1.91, 1.94, 1.98, 1.101,
  1.102.

Reviewed by: alfred

24 years agoSync with NetBSD: Keep the bit position even when the report descriptor says POP.
Nick Hibma [Sun, 2 Apr 2000 23:15:08 +0000 (23:15 +0000)]
Sync with NetBSD: Keep the bit position even when the report descriptor says POP.

24 years agoEnable Haifa on sparc64 as well.
Steve Price [Sun, 2 Apr 2000 22:51:02 +0000 (22:51 +0000)]
Enable Haifa on sparc64 as well.

24 years agoFirst cut at the (non-contrib) bits needed to build the binutils
Steve Price [Sun, 2 Apr 2000 22:36:56 +0000 (22:36 +0000)]
First cut at the (non-contrib) bits needed to build the binutils
distribution for FreeBSD/Sparc64.

Reviewed by: obrien

24 years agoShoot. Did a compile in the wrong directory and missed this compiler
Nick Hibma [Sun, 2 Apr 2000 22:04:18 +0000 (22:04 +0000)]
Shoot. Did a compile in the wrong directory and missed this compiler
error.

Could someone pass me the pointy hat, please?

24 years agoRemove a static variable, a uniquifier for a SCSI command. It might have
Nick Hibma [Sun, 2 Apr 2000 21:59:27 +0000 (21:59 +0000)]
Remove a static variable, a uniquifier for a SCSI command. It might have
produced inadvertently failed transfers in the case were multiple
Bulk-Only devices are connected and a command is prepared at the same
time.

24 years agoAdd a comment about the fact that the usb module is automatically
Nick Hibma [Sun, 2 Apr 2000 21:09:19 +0000 (21:09 +0000)]
Add a comment about the fact that the usb module is automatically
loaded.

24 years agoThe ld directory is in ${SRCDIR} not ${BINUTILSDISTDIR}.
Steve Price [Sun, 2 Apr 2000 20:23:58 +0000 (20:23 +0000)]
The ld directory is in ${SRCDIR} not ${BINUTILSDISTDIR}.

24 years agoCorrect Charles Mott's email address
Brian Somers [Sun, 2 Apr 2000 20:23:34 +0000 (20:23 +0000)]
Correct Charles Mott's email address
Requested by: cmott@scientech.com

24 years agoThe FreeBSD/Sparc64 port will also have an __ELF_WORD_SIZE of 64.
Steve Price [Sun, 2 Apr 2000 20:22:04 +0000 (20:22 +0000)]
The FreeBSD/Sparc64 port will also have an __ELF_WORD_SIZE of 64.

24 years agoCorrect Charles Mott's email address
Brian Somers [Sun, 2 Apr 2000 20:16:45 +0000 (20:16 +0000)]
Correct Charles Mott's email address

Requested by: Charles Mott <cmott@scientech.com>

24 years agoClone bio versions of certain bits of infrastructure:
Poul-Henning Kamp [Sun, 2 Apr 2000 19:08:05 +0000 (19:08 +0000)]
Clone bio versions of certain bits of infrastructure:
        devstat_end_transaction_bio()
        bioq_* versions of bufq_* incl bioqdisksort()
the corresponding "buf" versions will disappear when no longer used.

Move b_offset, b_data and b_bcount to struct bio.

Add BIO_FORMAT as a hack for fd.c etc.

We are now largely ready to start converting drivers to use struct
bio instead of struct buf.

24 years ago Make the sigprocmask() and geteuid() system calls MP SAFE. Expand
Matthew Dillon [Sun, 2 Apr 2000 17:52:43 +0000 (17:52 +0000)]
Make the sigprocmask() and geteuid() system calls MP SAFE.  Expand
    commentary for copyin/copyout to indicate that they are MP SAFE as
    well.

Reviewed by: msmith

24 years agoI got the CAM numbers (path id's, buses, unit numbers) and newbus
Nick Hibma [Sun, 2 Apr 2000 16:59:16 +0000 (16:59 +0000)]
I got the CAM numbers (path id's, buses, unit numbers) and newbus
unit numbers all wrong. This did not show up because most of them where
zero anyway.

Use a separate buffer for command transforms instead of fiddling with
the existing cdb_bytes.

Take CAM_CDB_POINTER into account. Nobody is using it, but someone might
in the future.

Be more picky about what to accept in the UFI command set.

First attempt at implementing the ATAPI command protocol transforms.
This should at least make Imation Superdisk and other e-Shuttle based
devices show as attached. Maybe they even work to some extent.

24 years agoPrint out more hints at illegal cksum len value.
Yoshinobu Inoue [Sun, 2 Apr 2000 16:23:16 +0000 (16:23 +0000)]
Print out more hints at illegal cksum len value.

24 years agoMove htons() ip_len to after the in_delayed_cksum() call.
Yoshinobu Inoue [Sun, 2 Apr 2000 16:18:26 +0000 (16:18 +0000)]
Move htons() ip_len to after the in_delayed_cksum() call.
This should stop cksum error messages on IPsec communication
which was reported on freebsd-current.

Reviewed by: jlemon

24 years agoMake sure to use IPv4 mapped IPv6 address when mapped address is
Hajimu UMEMOTO [Sun, 2 Apr 2000 16:11:14 +0000 (16:11 +0000)]
Make sure to use IPv4 mapped IPv6 address when mapped address is
requested in /etc/inetd.conf.

Reviewed by: shin

24 years agoMove B_ERROR flag to b_ioflags and call it BIO_ERROR.
Poul-Henning Kamp [Sun, 2 Apr 2000 15:24:56 +0000 (15:24 +0000)]
Move B_ERROR flag to b_ioflags and call it BIO_ERROR.

(Much of this done by script)

Move B_ORDERED flag to b_ioflags and call it BIO_ORDERED.

Move b_pblkno and b_iodone_chain to struct bio while we transition, they
will be obsoleted once bio structs chain/stack.

Add bio_queue field for struct bio aware disksort.

Address a lot of stylistic issues brought up by bde.

24 years ago Add support for the LinkSys USB10T network adapter that
Jonathan M. Bresler [Sun, 2 Apr 2000 12:43:34 +0000 (12:43 +0000)]
Add support for the LinkSys USB10T network adapter that
contains the ADMtek Pegasus AN986 USB chipset.  The
adapter supports both 10BaseT and 100BaseT (including
full-duplex).  The product code for these adapters is
0x2206.

24 years agoAdd and install BUS_SETUP_INTR.9.
Jeroen Ruigrok van der Werven [Sun, 2 Apr 2000 12:00:52 +0000 (12:00 +0000)]
Add and install BUS_SETUP_INTR.9.
This effectively replaces BUS_CREATE_INTR.9 and BUS_CONNECT_INTR.9.

Reviewed by: n_hibma

24 years agoAvoid rawname() stupidly prepending an `r' before the device name even
Joerg Wunsch [Sun, 2 Apr 2000 10:16:40 +0000 (10:16 +0000)]
Avoid rawname() stupidly prepending an `r' before the device name even
in cases where the supplied name was already pointing to a character
special device.  This fixes the breakage that occured when trying to
dump a filesystem by name (e. g. /usr), with an fstab already
mentioning the raw device name (like /dev/rda0g) where dump attempted
to use /dev/rrda0g then.

Also removed the now obsolete remark that fstab were carrying block
special names.

24 years agoThis new driver version uses the DMA mapping interface
Gerard Roudier [Sun, 2 Apr 2000 10:07:54 +0000 (10:07 +0000)]
This new driver version uses the DMA mapping interface
introduced in FreeBSD-4.0. The driver is now full up-to-date
with regards to the current kernel interfaces.

Another significant change in this driver version applies
to the checking of the data direction. The driver is now
able to check against the expected data direction in any
circumstance and will not hang either if direction is wrong
at the start of the IO, or if for some weird reason,
the device changes to the wrong direction during the IO.

This driver version is still usable under FreeBSD 3.2/3.3,
since it only requires CAM, other kernel interface dependencies
being #if'ed in the sources according to kernel version.
But, in order to use the driver under those early kernel
versions, user has to move the driver sources by hand and make
appropriate tiny changes to let the kernel know about the driver.

Other changes:
- Remove the debugging stuff for WSR bit.
- Get rid of some warnings about volatile being discarded.

24 years agoDraw the outline of "struct bio".
Poul-Henning Kamp [Sun, 2 Apr 2000 09:26:51 +0000 (09:26 +0000)]
Draw the outline of "struct bio".

Struct bio is the future carrier of I/O requests for "struct buf".

24 years agoGet rid of the manpages for BUS_CREATE_INTR() and BUS_CONNECT_INTR().
Jeroen Ruigrok van der Werven [Sun, 2 Apr 2000 09:25:15 +0000 (09:25 +0000)]
Get rid of the manpages for BUS_CREATE_INTR() and BUS_CONNECT_INTR().
These got replaced by BUS_SETUP_INTR().

This once again illustrates an API change without informing -doc, so
that these sort post cleanup actions could've been avoided.

And then people wonder why the docs suck so much at times.

Reviewed by: peter

24 years ago- Add temporary workaround to fool some "clever" KVM switch products
Kazutaka YOKOTA [Sun, 2 Apr 2000 08:28:21 +0000 (08:28 +0000)]
- Add temporary workaround to fool some "clever" KVM switch products
  which think they know the IntelliMouse 4-byte packet and believe,
  wrongly, that any other protocols use 3-byte packets.
- Update a couple of comment lines for A4 Tech mice.

24 years ago- Reflect recent update.
Kazutaka YOKOTA [Sun, 2 Apr 2000 08:25:34 +0000 (08:25 +0000)]
- Reflect recent update.

24 years agoFix two problems in the ch(4) driver.
Kenneth D. Merry [Sun, 2 Apr 2000 07:49:09 +0000 (07:49 +0000)]
Fix two problems in the ch(4) driver.

- Mike Smith discovered a panic in the changer probe code if the probe
  command (mode sense) fails.  So we need to release the CCB used in the
  probe before we unlock the peripheral.  (i.e. the same fix mjacob put in
  the CD and DA drivers)

- A newline was missing in a warning message.  (PR kern/17512)

PR: kern/17512
Submitted by: Louis Mamakos <louie@uu.net> (newline fix)

24 years agounfinished sblive driver, playback/mixer only for now - not enabled in
Cameron Grant [Sun, 2 Apr 2000 07:41:17 +0000 (07:41 +0000)]
unfinished sblive driver, playback/mixer only for now - not enabled in
conf/files

i don't seem to be clearing the cache right resulting in a short initial
burst of noise, despite doing the same as creative and alsa.  i'm committing
now so more eyes can pore over the code.

24 years agoOne in a while, something happens so uncannily that you get knocked off
Brian Feldman [Sun, 2 Apr 2000 07:02:48 +0000 (07:02 +0000)]
One in a while, something happens so uncannily that you get knocked off
your feet.  The conversion of the "snp" device to a dynamically growing
device driver was done just a few days ago by Brooks Davis!  Shame on
me for not finding that PR :(

This is a forced commit of tty_snoop.c to give the submitter proper credit,
as most of the patch submitted is actually exactly the same code (by some
large amount of entropy).  Brooks also submitted the change to LINT to
set the example of "snp" usage to not include a number, as that number is
now deprecated, so that is also in this commit.

PR: 17629
Submitted by: Brooks Davis <brooks@one-eyed-alien.net>

24 years agoReflect the recent update in moused.c.
Kazutaka YOKOTA [Sun, 2 Apr 2000 06:34:49 +0000 (06:34 +0000)]
Reflect the recent update in moused.c.

24 years ago if USB support is not compiled into the kernel,
Jonathan M. Bresler [Sun, 2 Apr 2000 06:28:40 +0000 (06:28 +0000)]
if USB support is not compiled into the kernel,
and has not been loaded via a kldload,
running usbd(8) will autoload the "usb.ko" kld.

thanks to Peter Wemm for enlightening me on the
differences between kldfind(2) and modfind(2).

24 years agoReplace the prefix for yylval to less confusing one.
Yoshinobu Inoue [Sun, 2 Apr 2000 05:08:07 +0000 (05:08 +0000)]
Replace the prefix for yylval to less confusing one.
(Sorry, this should be committed with previous commit to Makefile.)

24 years agoEnable installation of picobsd.8
Greg Lehey [Sun, 2 Apr 2000 04:32:44 +0000 (04:32 +0000)]
Enable installation of picobsd.8

24 years agoTry and make the kernel build again without INET6.
Paul Saab [Sun, 2 Apr 2000 03:49:25 +0000 (03:49 +0000)]
Try and make the kernel build again without INET6.

24 years agoMake mergemaster only take single letters for options as well as loop
Alfred Perlstein [Sun, 2 Apr 2000 02:47:15 +0000 (02:47 +0000)]
Make mergemaster only take single letters for options as well as loop
asking a question again if given an invalid input instead of assuming
what the user wants.  /etc is not the place to make assumptions when
given invalid input.

Reformat some of the more convoluted code into seperate functions instead
of being inline using tabs instead of space indents.

Allow the user to view merged files they created with sdiff.

Allow the user to redisplay the diff between the installed file and
the new file again.

Time wasted waiting for review: 1 month 2 weeks

24 years ago Change the write-behind code to take more care when starting
Matthew Dillon [Sun, 2 Apr 2000 00:55:28 +0000 (00:55 +0000)]
Change the write-behind code to take more care when starting
    async I/O's.  The sequential read heuristic has been extended to
    cover writes as well.  We continue to call cluster_write() normally,
    thus blocks in the file will still be reallocated for large (but still
    random) I/O's, but I/O will only be initiated for truely sequential
    writes.

    This solves a number of annoying situations, especially with DBM (hash
    method) writes, and also has the side effect of fixing a number of
    (stupid) benchmarks.

Reviewed-by: mckusick
24 years agoUnstaticize this driver. You can have as many snoop devices as you can
Brian Feldman [Sun, 2 Apr 2000 00:35:37 +0000 (00:35 +0000)]
Unstaticize this driver.  You can have as many snoop devices as you can
mknod :)

Clean things up a lot while I'm here.  A lot of KNF changes.

24 years agoSupport per socket based IPv4 mapped IPv6 addr enable/disable control.
Yoshinobu Inoue [Sat, 1 Apr 2000 22:35:47 +0000 (22:35 +0000)]
Support per socket based IPv4 mapped IPv6 addr enable/disable control.

Submitted by: ume

24 years agoReplace the prefix for yylval to less confusing one.
Yoshinobu Inoue [Sat, 1 Apr 2000 22:28:36 +0000 (22:28 +0000)]
Replace the prefix for yylval to less confusing one.

Suggested by: bruce

24 years agofail in attach if we seem to have no ac97 codec
Cameron Grant [Sat, 1 Apr 2000 22:24:03 +0000 (22:24 +0000)]
fail in attach if we seem to have no ac97 codec

24 years agoDon't compile in IPsec support when NOIPSEC is defined.
Yoshinobu Inoue [Sat, 1 Apr 2000 21:48:07 +0000 (21:48 +0000)]
Don't compile in IPsec support when NOIPSEC is defined.

PR: bin/17739
Submitted by: Omachonu Ogali <oogali@intranova.net>

24 years agobump the buffer size from 4k to 16k. should improve performance under load.
Cameron Grant [Sat, 1 Apr 2000 21:28:09 +0000 (21:28 +0000)]
bump the buffer size from 4k to 16k.  should improve performance under load.

24 years agoRemove spurious .Fc which caused this manpage to be closed
Jeroen Ruigrok van der Werven [Sat, 1 Apr 2000 20:31:07 +0000 (20:31 +0000)]
Remove spurious .Fc which caused this manpage to be closed
prematurely.

24 years agoCalculate any delayed checksums before handing an mbuf off to a
Jonathan Lemon [Sat, 1 Apr 2000 18:51:03 +0000 (18:51 +0000)]
Calculate any delayed checksums before handing an mbuf off to a
divert socket.  This fixes a problem with ppp/natd.

Reviewed by: bsd (Brian Dean, gotta love that login name)

24 years agoAdd additional path for info to /usr/X11R6/info. This will allow
Jeroen Ruigrok van der Werven [Sat, 1 Apr 2000 15:30:57 +0000 (15:30 +0000)]
Add additional path for info to /usr/X11R6/info.  This will allow
people to actually look at the info files for gdk and gtk for example.