]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
25 years agoMake a dev2budev() function, and use it. This refixes pstat (working, broken,
green [Tue, 20 Jul 1999 21:29:13 +0000 (21:29 +0000)]
Make a dev2budev() function, and use it. This refixes pstat (working, broken,
working, broken, working) and savecore (working, working, broken, working,
working).

Sorta Reviewed by: phk

25 years agoConvert the xl driver to newbus. It is now possible to make this driver
wpaul [Tue, 20 Jul 1999 21:23:17 +0000 (21:23 +0000)]
Convert the xl driver to newbus. It is now possible to make this driver
into a loadable module, and all of the platform dependencies are gone
(except for the alpha_XXX_dmamap() thing, which is another issue -- I
still don't know how to use the busdma stuff with a network driver).

Also increase the delay in xl_reset(); testing on a 486/66 with a 3c905C
shows that reading the EEPROM fails immediately after a reset. Waiting
a little longer after the reset completes seems to fix it.

25 years agoXFree86 3.3.4 seems to require "XWINHOME" to be set for the setup tool
jkh [Tue, 20 Jul 1999 21:06:18 +0000 (21:06 +0000)]
XFree86 3.3.4 seems to require "XWINHOME" to be set for the setup tool
to work (fnark).

25 years agodev2udev() returns a CDEV udev_t, but we use block io in savecore. Savecore
green [Tue, 20 Jul 1999 20:55:50 +0000 (20:55 +0000)]
dev2udev() returns a CDEV udev_t, but we use block io in savecore. Savecore
also gets the device by st_rdev, which is alright except for the fact that
the sysctl kern.dumpdev passed out a char device. This is a workaround.
Sorry for not committing the fix earlier, before people started having
problems.

25 years agoCorrect the information about the return value when no device matches
nik [Tue, 20 Jul 1999 20:48:40 +0000 (20:48 +0000)]
Correct the information about the return value when no device matches
(or no information is available).

PR:             docs/12707
Submitted by:   Chris Costello <chris@calldei.com>

25 years agoCorrect some grammar and style problems with this page.
nik [Tue, 20 Jul 1999 20:46:26 +0000 (20:46 +0000)]
Correct some grammar and style problems with this page.

Submitted by:   Kris Kennaway <root@rebel.net.au>

25 years agoCorrect the alignment of some of the members in the wi_frame structure.
wpaul [Tue, 20 Jul 1999 20:03:42 +0000 (20:03 +0000)]
Correct the alignment of some of the members in the wi_frame structure.
The structure is the right length, but some of the members (notably
wi_q_info) were off a bit. This causes the received signal strength
values to appear bogus.

25 years agoMake bcd2int work (resume time reporting now works.)
green [Tue, 20 Jul 1999 15:37:28 +0000 (15:37 +0000)]
Make bcd2int work (resume time reporting now works.)

PR: 12613
Submitted by: Michael Constant <mconst@not.there.com>

25 years agoMake wall_cmos do something.
green [Tue, 20 Jul 1999 15:31:23 +0000 (15:31 +0000)]
Make wall_cmos do something.

PR: 12614

25 years agoI missed a not. Also, remove invltlb(), since it's "unncessary [sic] because
green [Tue, 20 Jul 1999 13:21:55 +0000 (13:21 +0000)]
I missed a not. Also, remove invltlb(), since it's "unncessary [sic] because
wbinvd already flushes the the TLB."

25 years agoNow a dev_t is a pointer to struct specinfo which is shared by all specdev
phk [Tue, 20 Jul 1999 09:47:55 +0000 (09:47 +0000)]
Now a dev_t is a pointer to struct specinfo which is shared by all specdev
vnodes referencing this device.

Details:
        cdevsw->d_parms has been removed, the specinfo is available
        now (== dev_t) and the driver should modify it directly
        when applicable, and the only driver doing so, does so:
        vn.c.  I am not sure the logic in checking for "<" was right
        before, and it looks even less so now.

        An intial pool of 50 struct specinfo are depleted during
        early boot, after that malloc had better work.  It is
        likely that fewer than 50 would do.

        Hashing is done from udev_t to dev_t with a prime number
        remainder hash, experiments show no better hash available
        for decent cost (MD5 is only marginally better)  The prime
        number used should not be close to a power of two, we use
        83 for now.

        Add new checkalias2() to get around the loss of info from
        dev2udev() in bdevvp();

        The aliased vnodes are hung on a list straight of the dev_t,
        and speclisth[SPECSZ] is unused.  The sharing of struct
        specinfo means that the v_specnext moves into the vnode
        which grows by 4 bytes.

        Don't use a VBLK dev_t which doesn't make sense in MFS, now
        we hang a dummy cdevsw on B/Cmaj 253 so that things look sane.

Storage overhead from all of this is O(50k).

        Bump __FreeBSD_version to 400009

The next step will add the stuff needed so device-drivers can start to
hang things from struct specinfo

25 years agoAlso check against chardevs when looking for root.
phk [Tue, 20 Jul 1999 08:50:54 +0000 (08:50 +0000)]
Also check against chardevs when looking for root.

25 years agoThe matcd driver is acting strange (returning a successful open even
jkh [Tue, 20 Jul 1999 08:47:35 +0000 (08:47 +0000)]
The matcd driver is acting strange (returning a successful open even
when it fails).  Disable it in sysinstall for now.

25 years agoFix a bad dhcp keyword; it's host-name not server-name.
jkh [Tue, 20 Jul 1999 08:08:24 +0000 (08:08 +0000)]
Fix a bad dhcp keyword; it's host-name not server-name.

25 years agoAdd slovakian ftp mirror.
jkh [Tue, 20 Jul 1999 07:50:26 +0000 (07:50 +0000)]
Add slovakian ftp mirror.

Submitted by: "Tomas TPS Ulej" <tps@ti.sk>

25 years agoMore Alpha ifdefage.
jkh [Tue, 20 Jul 1999 07:39:22 +0000 (07:39 +0000)]
More Alpha ifdefage.

25 years agouse vp->v_rdev;
phk [Tue, 20 Jul 1999 07:29:54 +0000 (07:29 +0000)]
use vp->v_rdev;

OK'ed by: grog

25 years agoadd debug.sizeof.specinfo
phk [Tue, 20 Jul 1999 07:19:32 +0000 (07:19 +0000)]
add debug.sizeof.specinfo

25 years agoDon't access the device with vp->v_specinfo->si_rdev, use vp->v_rdev.
phk [Tue, 20 Jul 1999 07:18:17 +0000 (07:18 +0000)]
Don't access the device with vp->v_specinfo->si_rdev, use vp->v_rdev.

25 years agoUpdate of the i686 MTRR/memory range support.
msmith [Tue, 20 Jul 1999 06:58:51 +0000 (06:58 +0000)]
Update of the i686 MTRR/memory range support.

 - Support for setting memory range attributes on SMP systems using the
   new SMP rendezvous function
 - Don't print the confusing default memory type message.
 - Allow legal overlapping range types.
 - Turn interrupts back on after setting MTRRs in UP mode (whoops)
 - Don't waste time calling invltlb() after wbinvd(); it's not
   SMP-compatible (interrupts are off) and unncessary because
   wbinvd already flushes the TLB.

This code is now essentially feature-complete.

25 years agoImplement an all-CPU shootdown-style rendezvous facility. This allows
msmith [Tue, 20 Jul 1999 06:52:35 +0000 (06:52 +0000)]
Implement an all-CPU shootdown-style rendezvous facility.  This allows
the caller to specify a function to be guarded between an entry and exit
barrier, as well as pre- and post-barrier functions.

The primary use for this function is synchronised update of per-cpu private
data.  The implementation is almost (but not quite) MI; with a better
mechanism for masking per-CPU interrupts it could probably be hoisted.

Reviewed by: peter (partially)

25 years agoAn SMP-specific change: Eliminate an unnecessary lock acquire and release
alc [Tue, 20 Jul 1999 06:09:53 +0000 (06:09 +0000)]
An SMP-specific change: Eliminate an unnecessary lock acquire and release
in setdelayed.

Submitted by: luoqi and bde

25 years agoConvert a "page not busy" warning to an assertion.
alc [Tue, 20 Jul 1999 05:46:56 +0000 (05:46 +0000)]
Convert a "page not busy" warning to an assertion.

Submitted by: dillon@backplane.com

25 years agoGet bpf0 onto the MFS image.
jkh [Tue, 20 Jul 1999 05:00:00 +0000 (05:00 +0000)]
Get bpf0 onto the MFS image.

25 years agoMake memcontrol's internal help actually work. No substitute for a real
msmith [Tue, 20 Jul 1999 04:33:14 +0000 (04:33 +0000)]
Make memcontrol's internal help actually work.  No substitute for a real
manpage, but at least now you can get syntax help without resorting
to reading the source.

25 years agoReintroduce LBA (cyl > 1023) support in the bootblocks, enabled by
rnordier [Tue, 20 Jul 1999 01:19:23 +0000 (01:19 +0000)]
Reintroduce LBA (cyl > 1023) support in the bootblocks, enabled by
means of a build option.

25 years agoFix a page size vs. KB mixup. The extra buffers allocated at a reduced
peter [Mon, 19 Jul 1999 23:36:30 +0000 (23:36 +0000)]
Fix a page size vs. KB mixup. The extra buffers allocated at a reduced
rate is meant to kick in at 64MB, not 256MB.

Reviewed by: Matthew Dillon <dillon@backplane.com>

25 years agoDocument the "skey" command in telnet(1).
nik [Mon, 19 Jul 1999 22:24:32 +0000 (22:24 +0000)]
Document the "skey" command in telnet(1).

PR:             docs/12360
Submitted by:   kjm@rins.ryukoku.ac.jp (KOJIMA Hajime)

25 years agoCheck if an fs is mounted before checking if it is mounted read-only.
luoqi [Mon, 19 Jul 1999 21:04:25 +0000 (21:04 +0000)]
Check if an fs is mounted before checking if it is mounted read-only.
Pointed out by: Mike Smith <msmith@freebsd.org>

25 years agoActivate kgzip.
rnordier [Mon, 19 Jul 1999 20:22:44 +0000 (20:22 +0000)]
Activate kgzip.

25 years agoActivate kgzldr.
rnordier [Mon, 19 Jul 1999 20:16:34 +0000 (20:16 +0000)]
Activate kgzldr.

25 years agoUse M4FLAGS for m4.
rnordier [Mon, 19 Jul 1999 20:10:17 +0000 (20:10 +0000)]
Use M4FLAGS for m4.

25 years agoFix a problem where commands in backticks `` wont get executed in taint
markm [Mon, 19 Jul 1999 18:17:54 +0000 (18:17 +0000)]
Fix a problem where commands in backticks `` wont get executed in taint
mode because of an insecure path.

Reported by: George Cox

25 years agoThis commit was generated by cvs2svn to compensate for changes in r48907,
rnordier [Mon, 19 Jul 1999 18:07:44 +0000 (18:07 +0000)]
This commit was generated by cvs2svn to compensate for changes in r48907,
which included commits to RCS files with non-trunk default branches.

25 years agokgzldr is a kzipboot replacement for use with kgzip(8).
rnordier [Mon, 19 Jul 1999 18:07:44 +0000 (18:07 +0000)]
kgzldr is a kzipboot replacement for use with kgzip(8).

Development sponsored by Global Technology Associates, Inc.

Reviewed/tested by: abial

25 years agoThis commit was generated by cvs2svn to compensate for changes in r48905,
rnordier [Mon, 19 Jul 1999 18:00:18 +0000 (18:00 +0000)]
This commit was generated by cvs2svn to compensate for changes in r48905,
which included commits to RCS files with non-trunk default branches.

25 years agokgzip is a kzip(8) replacement able to compress and link bootable
rnordier [Mon, 19 Jul 1999 18:00:18 +0000 (18:00 +0000)]
kgzip is a kzip(8) replacement able to compress and link bootable
32-bit binaries in both ELF and a.out format.

Development sponsored by Global Technology Associates, Inc.

Reviewed/tested by: abial

25 years agoVarious cleanups.
green [Mon, 19 Jul 1999 17:37:42 +0000 (17:37 +0000)]
Various cleanups.

25 years agoSync with GENERIC (place of xe0 driver)
hosokawa [Mon, 19 Jul 1999 15:18:21 +0000 (15:18 +0000)]
Sync with GENERIC (place of xe0 driver)

25 years agosync with boot_crunch.conf.generic
hosokawa [Mon, 19 Jul 1999 12:23:33 +0000 (12:23 +0000)]
sync with boot_crunch.conf.generic
(added arp, sed, and test)

25 years agoCatch an XFree86 3.3.4 update I missed.
jkh [Mon, 19 Jul 1999 11:58:01 +0000 (11:58 +0000)]
Catch an XFree86 3.3.4 update I missed.

25 years agoUpdate for XFree86 3.3.4
jkh [Mon, 19 Jul 1999 11:49:22 +0000 (11:49 +0000)]
Update for XFree86 3.3.4

25 years agoWait for dhclient in a far saner fashion.
jkh [Mon, 19 Jul 1999 11:00:56 +0000 (11:00 +0000)]
Wait for dhclient in a far saner fashion.

Submitted by: obrien

25 years agoFix handling of the cd command inside evaluations. It was changing PWD,
sheldonh [Mon, 19 Jul 1999 11:00:33 +0000 (11:00 +0000)]
Fix handling of the cd command inside evaluations. It was changing PWD,
which it should not do.

PR: 12578
Reported by: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
Submitted by: Niall Smart <niall@pobox.com>

25 years agoAdd /bin/test to the boot floppy. Guess we'll have to sync this up
jkh [Mon, 19 Jul 1999 10:53:54 +0000 (10:53 +0000)]
Add /bin/test to the boot floppy.  Guess we'll have to sync this up
with boot_crunch.conf.pccard at some point?

25 years agocleanup pass over new stuff.
jkh [Mon, 19 Jul 1999 10:18:52 +0000 (10:18 +0000)]
cleanup pass over new stuff.

25 years agoAnother batch of fixes for dhcp support in sysinstall, now dragging
jkh [Mon, 19 Jul 1999 10:06:18 +0000 (10:06 +0000)]
Another batch of fixes for dhcp support in sysinstall, now dragging
in some code from C. Stone to parse the lease information.  This is still
a WIP and this commit is largely intended to allow others to sync up; the
dhclient code still only works when doing dhcp configuration post-install
and requires a bit more work on the boot floppy before it will truly
work in the minimal bootstrapping role.

25 years ago[click] Now all dev_t's in the kernel have their char device major.
phk [Mon, 19 Jul 1999 09:37:59 +0000 (09:37 +0000)]
[click]  Now all dev_t's in the kernel have their char device major.

Only know casualy of this is swapinfo/pstat which should be fixes
the right way:  Store the actual pathname in the kernel like mount
does.  [Volounteers sought for this task]

The road map from here is roughly:  expand struct specinfo into struct
based dev_t.  Add dev_t registration facilities for device drivers and
start to use them.

25 years agoAdd sysctl tree debug.sizeof to tell us how big things are. First two
phk [Mon, 19 Jul 1999 09:13:12 +0000 (09:13 +0000)]
Add sysctl tree debug.sizeof to tell us how big things are.  First two
entries are struct proc and struct vnode.

25 years agoFix type of host-name option. Partial fix for referenced PR.
jkh [Mon, 19 Jul 1999 01:58:46 +0000 (01:58 +0000)]
Fix type of host-name option.  Partial fix for referenced PR.

Submitted by: John Baldwin <jobaldwi@vt.edu>
PR: 12205

25 years agoUpdated acquire_timer2()'s state machine to work when the i8254 is
bde [Sun, 18 Jul 1999 18:32:42 +0000 (18:32 +0000)]
Updated acquire_timer2()'s state machine to work when the i8254 is
being used for timecounting.  Fixed a race or two in it.  Undisabled
it.

PR: 10455

25 years agoDon't let the machdep.tsc_freq sysctl proceed if the TSC is present
bde [Sun, 18 Jul 1999 15:19:29 +0000 (15:19 +0000)]
Don't let the machdep.tsc_freq sysctl proceed if the TSC is present
but broken, since tsc_timecounter is not initialised in that case,
and updating an uninitialised timecounter is fatal.

Fixed style bugs in the machdep.i8254_freq and machdep.tsc_freq
sysctls.

Reviewed by: phk

25 years agoAdded a sysctl "kern.timecounter.hardware" for selecting the hardware
bde [Sun, 18 Jul 1999 15:07:20 +0000 (15:07 +0000)]
Added a sysctl "kern.timecounter.hardware" for selecting the hardware
used for timecounting.  The possible values are the names of the
physically present harware timecounters ("i8254" and "TSC" on i386's).

Fixed some nearby bitrot in comments in <sys/time.h>.

Reviewed by: phk

25 years ago fix comment re: RST received in TIME_WAIT to match the code.
jmb [Sun, 18 Jul 1999 14:42:48 +0000 (14:42 +0000)]
fix comment re: RST received in TIME_WAIT to match the code.

25 years agoUse the vn_todev() function, rather than VOP_GETATTR
phk [Sun, 18 Jul 1999 14:31:01 +0000 (14:31 +0000)]
Use the vn_todev() function, rather than VOP_GETATTR

25 years agoIntroduce the vn_todev(struct vnode*) function, which returns the dev_t
phk [Sun, 18 Jul 1999 14:30:37 +0000 (14:30 +0000)]
Introduce the vn_todev(struct vnode*) function, which returns the dev_t
corresponding to a VBLK or VCHR node, or NODEV.

25 years agoReset SA_NOCLDWAIT on exec().
peter [Sun, 18 Jul 1999 13:40:11 +0000 (13:40 +0000)]
Reset SA_NOCLDWAIT on exec().

PR: kern/12669
Submitted by: Doug Ambrisko <ambrisko@whistle.com>

25 years agoMake devname(3) return something more intelligent than NULL if it doesn't
phk [Sun, 18 Jul 1999 10:19:48 +0000 (10:19 +0000)]
Make devname(3) return something more intelligent than NULL if it doesn't
find anything in the database.

25 years agoSome additional optimizations for using DHCP.
jkh [Sun, 18 Jul 1999 10:18:06 +0000 (10:18 +0000)]
Some additional optimizations for using DHCP.

25 years agoUse a less annoying default hostname.
jkh [Sun, 18 Jul 1999 09:58:01 +0000 (09:58 +0000)]
Use a less annoying default hostname.

25 years ago- Clear relevant variables when the keyboard is not available.
yokota [Sun, 18 Jul 1999 06:16:53 +0000 (06:16 +0000)]
- Clear relevant variables when the keyboard is not available.
- Do not touch the `tty' struct when it's not yet available.
- Initialize correct bits in `fonts_loaded'.

25 years ago- Move the `return' statement the correct place so that the keyboard
yokota [Sun, 18 Jul 1999 06:16:25 +0000 (06:16 +0000)]
- Move the `return' statement the correct place so that the keyboard
  won't be initialized if `atkbd?' is disabled.

25 years agoPrint out the fields that are set in fs_flags.
mpp [Sun, 18 Jul 1999 02:33:05 +0000 (02:33 +0000)]
Print out the fields that are set in fs_flags.

25 years agoRemove a duplicate comment.
jdp [Sun, 18 Jul 1999 02:26:02 +0000 (02:26 +0000)]
Remove a duplicate comment.

25 years agoMiscellaneous fixes for dhcp client support.
jkh [Sun, 18 Jul 1999 02:20:56 +0000 (02:20 +0000)]
Miscellaneous fixes for dhcp client support.

25 years agoRemove four no-op casts.
jdp [Sun, 18 Jul 1999 01:35:26 +0000 (01:35 +0000)]
Remove four no-op casts.

25 years agoChange many asserts into normal errors. They were all for conditions
jdp [Sun, 18 Jul 1999 00:02:19 +0000 (00:02 +0000)]
Change many asserts into normal errors.  They were all for conditions
caused by invalid shared objects rather than by internal errors.

Enable format string mismatch checking for _rtld_error().

25 years agoRemove trailing whitespace from a couple of lines. It made its way
jdp [Sat, 17 Jul 1999 23:22:13 +0000 (23:22 +0000)]
Remove trailing whitespace from a couple of lines.  It made its way
into the .cf file and gave sendmail indigestion.

25 years agoCentralize dumpdev handling.
phk [Sat, 17 Jul 1999 20:47:52 +0000 (20:47 +0000)]
Centralize dumpdev handling.

25 years agoReverse the sense of a test, dev2udev() will be much cheaper than
phk [Sat, 17 Jul 1999 20:29:10 +0000 (20:29 +0000)]
Reverse the sense of a test, dev2udev() will be much cheaper than
udev2dev().

25 years agoAdd a field to struct swdevt to avoid a bogus udev2dev() call.
phk [Sat, 17 Jul 1999 19:59:55 +0000 (19:59 +0000)]
Add a field to struct swdevt to avoid a bogus udev2dev() call.

25 years agoFix 2nd arg to udev2dev() call in ccd.c
phk [Sat, 17 Jul 1999 19:58:37 +0000 (19:58 +0000)]
Fix 2nd arg to udev2dev() call in ccd.c

25 years agoUse 256 as magic in bmaj2cmaj[]. Treat BLK/CHR dev_t more correctly.
phk [Sat, 17 Jul 1999 19:57:25 +0000 (19:57 +0000)]
Use 256 as magic in bmaj2cmaj[].  Treat BLK/CHR dev_t more correctly.

25 years agoFix 2nd arg to udev2dev().
phk [Sat, 17 Jul 1999 19:38:00 +0000 (19:38 +0000)]
Fix 2nd arg to udev2dev().

25 years agoAdd arp to the boot floppy.
jkh [Sat, 17 Jul 1999 19:30:01 +0000 (19:30 +0000)]
Add arp to the boot floppy.

25 years agoAdd braces to appease the egcs -Wall mom.
jdp [Sat, 17 Jul 1999 19:03:20 +0000 (19:03 +0000)]
Add braces to appease the egcs -Wall mom.

25 years agoMake gcore work again. It was broken by a format change in the
jdp [Sat, 17 Jul 1999 18:58:32 +0000 (18:58 +0000)]
Make gcore work again.  It was broken by a format change in the
procfs map file when object IDs were eliminated in the mega-commit
that included procfs_map.c revision 1.19.

The map file is a terrible hodge-podge.  The fields that are used
mainly for kernel debugging should be moved out of it into a
separate file, so that the interface presented by the map file to
applications can remain stable in the face of VM system changes.

25 years agoI have not one single time remembered the name of this function correctly
phk [Sat, 17 Jul 1999 18:43:50 +0000 (18:43 +0000)]
I have not one single time remembered the name of this function correctly
so obviously I gave it the wrong name.  s/umakedev/makeudev/g

25 years agoAn SMP-specific change: eliminate another check on bsp_apic_ready.
alc [Sat, 17 Jul 1999 18:34:32 +0000 (18:34 +0000)]
An SMP-specific change: eliminate another check on bsp_apic_ready.
(See rev 1.23.)

25 years agoslight change in the way we hi-jack the wd drivers cdevsw.
phk [Sat, 17 Jul 1999 17:55:53 +0000 (17:55 +0000)]
slight change in the way we hi-jack the wd drivers cdevsw.

25 years agoFix a typo
brian [Sat, 17 Jul 1999 10:33:57 +0000 (10:33 +0000)]
Fix a typo
Submitted by: Rich Wood <rich@chugaboom.net>

25 years agoDecode new ptrace requests PT_GETDBREGS and PT_SETDBREGS.
bde [Sat, 17 Jul 1999 09:27:41 +0000 (09:27 +0000)]
Decode new ptrace requests PT_GETDBREGS and PT_SETDBREGS.

25 years agoImplementation of TCXONC.
marcel [Sat, 17 Jul 1999 08:24:57 +0000 (08:24 +0000)]
Implementation of TCXONC.

Reviewed by: bde

25 years agoMerge some patches from RedHat-6.0. The most important change fixes a bug
dfr [Sat, 17 Jul 1999 08:04:58 +0000 (08:04 +0000)]
Merge some patches from RedHat-6.0. The most important change fixes a bug
when linking large shared libraries where the linker missed out some of
the dynamic relocations from the output. This prevents libqt from linking
properly among other packages.

Obtained from: RedHat-6.0

25 years agoAdd in a hack to turn off unaligned access warnings for alpha.
jkh [Fri, 16 Jul 1999 22:07:09 +0000 (22:07 +0000)]
Add in a hack to turn off unaligned access warnings for alpha.

Submitted by: msmith

25 years agoDon't offer compat dists on alpha since it makes no sense to do so.
jkh [Fri, 16 Jul 1999 22:03:27 +0000 (22:03 +0000)]
Don't offer compat dists on alpha since it makes no sense to do so.

Suggested by: msmith

25 years agoI think the last revision got lost here. Identd needs to be run as root,
green [Fri, 16 Jul 1999 16:24:13 +0000 (16:24 +0000)]
I think the last revision got lost here. Identd needs to be run as root,
at least for now. I relegated the getcred sysctls to only root, but if
they're deemed to be "allowable" to export to users, I'll do so and
revert this change.

25 years agoDocument the new {auth,ident,tap} service and provide examples in the
sheldonh [Fri, 16 Jul 1999 15:41:14 +0000 (15:41 +0000)]
Document the new {auth,ident,tap} service and provide examples in the
configuration file.

Requested by: green

25 years agoRe-enable DHCP client support again (but optional and turned off by default
jkh [Fri, 16 Jul 1999 11:13:09 +0000 (11:13 +0000)]
Re-enable DHCP client support again (but optional and turned off by default
for the time being) for debugging purposes.

Fix bug in options selection.

25 years agoAllow DHCP to be used in an ifconfig variable instead of the usual
jkh [Fri, 16 Jul 1999 09:26:53 +0000 (09:26 +0000)]
Allow DHCP to be used in an ifconfig variable instead of the usual
address information, producing the obvious effect (dhcp configuration).

Submitted by:   "Sean O'Connell" <sean@stat.Duke.EDU>

25 years agoDocument dhcp flags.
jkh [Fri, 16 Jul 1999 09:22:01 +0000 (09:22 +0000)]
Document dhcp flags.

Submitted by: "Sean O'Connell" <sean@stat.Duke.EDU>

25 years agoHandle R_ALPHA_NONE relocations in KLD.
dfr [Fri, 16 Jul 1999 09:16:11 +0000 (09:16 +0000)]
Handle R_ALPHA_NONE relocations in KLD.

25 years agoEnable gprof on alpha.
simokawa [Fri, 16 Jul 1999 07:22:10 +0000 (07:22 +0000)]
Enable gprof on alpha.
* alpha.{c,h} are same as i386.{c,h}.
* Force address calculation to be done in long precision(64bit on alpha)
  rather than double precision(52bit).

25 years agoEnable gmon/mcount on alpha.
simokawa [Fri, 16 Jul 1999 07:05:34 +0000 (07:05 +0000)]
Enable gmon/mcount on alpha.

25 years agoEnable gcrt1.o.
simokawa [Fri, 16 Jul 1999 06:59:27 +0000 (06:59 +0000)]
Enable gcrt1.o.

25 years ago* Restore register correctly.
simokawa [Fri, 16 Jul 1999 06:45:52 +0000 (06:45 +0000)]
* Restore register correctly.
* Add missing bits for profiling.

25 years agoMake profil() 64bit-safe for alpha.
simokawa [Fri, 16 Jul 1999 06:28:55 +0000 (06:28 +0000)]
Make profil() 64bit-safe for alpha.
uintfptr_t may be better for offset, but we must wait until
the definition of uintfptr_t moves from machine/profile.h.

Reviewed by: bde

25 years agoRemove vm_object::last_read. It is used by the old swap pager, but
alc [Fri, 16 Jul 1999 05:11:37 +0000 (05:11 +0000)]
Remove vm_object::last_read.  It is used by the old swap pager, but
not by the new one, i.e., vm/swap_pager.c rev 1.108.

Reviewed by: dillon@backplane.com

25 years agoAdd support for multiple PCI busses directly connected to the nexus.
msmith [Fri, 16 Jul 1999 01:00:30 +0000 (01:00 +0000)]
Add support for multiple PCI busses directly connected to the nexus.
This is only partially complete, but allows 450NX-based systems with
more than one PCI bus to be used again.

Submitted by: dfr

25 years agoAdd a knob to avoid DES code when making crunched binary. This isn't a beauty
abial [Thu, 15 Jul 1999 21:47:52 +0000 (21:47 +0000)]
Add a knob to avoid DES code when making crunched binary. This isn't a beauty
incarnated, it just matches other deficiencies related to crunchgen
and friends... and we already have similar code in ppp/Makefile.

RELEASE_CRUNCH should be axed, but for now let's be consistent.

Submitted by: Patrick Powell <papowell@astart.com>