]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
19 years agoAdd a note to indicate that the path set in named_pidfile must
dougb [Fri, 24 Sep 2004 04:45:16 +0000 (04:45 +0000)]
Add a note to indicate that the path set in named_pidfile must
also be set in named.conf. Our default named.conf has this already.

Update the note for named_symlink_enable to indicate that ndc is gone.

19 years agoReplace misuse of NULL with 0UL.
marcel [Fri, 24 Sep 2004 04:40:28 +0000 (04:40 +0000)]
Replace misuse of NULL with 0UL.

19 years agoRemove the dependency on the Intel EFI headers in sys/boot/efi. Instead
marcel [Fri, 24 Sep 2004 04:35:07 +0000 (04:35 +0000)]
Remove the dependency on the Intel EFI headers in sys/boot/efi. Instead
use <machine/efi.h> for the necessary definitions. This makes the EFI
code in sys/boot/efi totally unused, except for pure EFI loaders. As
such, maintenance and porting (to IA-32) of the EFI code is made as easy
as possible.

19 years agoUnhook libski and skiload and hook up ski.
marcel [Fri, 24 Sep 2004 04:21:14 +0000 (04:21 +0000)]
Unhook libski and skiload and hook up ski.

19 years agoPost repocopy build fixes.
marcel [Fri, 24 Sep 2004 04:06:22 +0000 (04:06 +0000)]
Post repocopy build fixes.

19 years agoForced commit to note that this file was repocopied from ../libski.
marcel [Fri, 24 Sep 2004 03:53:50 +0000 (03:53 +0000)]
Forced commit to note that this file was repocopied from ../libski.

Repocopy by: peter

19 years agoForced commit to note that this file was repocopied from ../skiload.
marcel [Fri, 24 Sep 2004 03:51:06 +0000 (03:51 +0000)]
Forced commit to note that this file was repocopied from ../skiload.

Repocopy by: peter

19 years agoFix-ups for a few problems reported with how ports are handled in the
kensmith [Fri, 24 Sep 2004 02:10:06 +0000 (02:10 +0000)]
Fix-ups for a few problems reported with how ports are handled in the
release:

- run 'make index' as part of release build, INDEX* files in
  CVS no longer updated and likely to be removed from CVS soon
- don't include README.html files in ports.tgz tarball because
  they cause cvsup problems for users later
- put a copy of /usr/ports on disc2 that does include README.html
  files so users have access to them

Reviewed by: ru (found one bug, several clean-up suggestions)
Tested by: 'make release'
MFC before: BETA6 build (this isn't likely to be tested in -current
anyway...)

19 years agoImplement the KDENABIO etc ioctl's for amd64 as well.
peter [Fri, 24 Sep 2004 01:19:11 +0000 (01:19 +0000)]
Implement the KDENABIO etc ioctl's for amd64 as well.

19 years agoImplement the syscons mouse pointer characters for amd64. It was only
peter [Fri, 24 Sep 2004 01:18:22 +0000 (01:18 +0000)]
Implement the syscons mouse pointer characters for amd64.  It was only
implemented for i386 and alpha.

19 years agoLike on i386, use the definition of struct bios_smap from machine/pc/bios.h
peter [Fri, 24 Sep 2004 01:11:11 +0000 (01:11 +0000)]
Like on i386, use the definition of struct bios_smap from machine/pc/bios.h
again.

19 years agoConverge towards i386. I originally resisted creating <machine/pc/bios.h>
peter [Fri, 24 Sep 2004 01:08:34 +0000 (01:08 +0000)]
Converge towards i386.  I originally resisted creating <machine/pc/bios.h>
because it was mostly irrelevant - except for the silly BIOS_PADDRTOVADDR
etc macros.  Along the way of working around this, I missed a few things.

* Make syscons properly inherit the bios capslock/shiftlock/etc state like
  i386 does.  Note that we cannot inherit the bios key repeat rate because
  that requires a bios call (which is impossible for us).
* Give syscons the ability to beep on amd64.  Oops.

While here, make bios.c compile and add it to files.amd64.

19 years agoSeverely strip down the repocopied i386/bios.c and bios.h files. It turns
peter [Fri, 24 Sep 2004 00:42:36 +0000 (00:42 +0000)]
Severely strip down the repocopied i386/bios.c and bios.h files.  It turns
out that bios_sigsearch() etc is useful for finding tables in roms.

19 years agoA modest collection of various and sundry style, spelling, and whitespace
jhb [Fri, 24 Sep 2004 00:38:15 +0000 (00:38 +0000)]
A modest collection of various and sundry style, spelling, and whitespace
fixes.

Submitted by: bde (mostly)

19 years agoOn arm, set the default elf brand to FreeBSD, until the binutils do it for us.
cognet [Thu, 23 Sep 2004 23:29:24 +0000 (23:29 +0000)]
On arm, set the default elf brand to FreeBSD, until the binutils do it for us.

19 years agoArbitraly choose the Variant II for arm.
cognet [Thu, 23 Sep 2004 23:14:09 +0000 (23:14 +0000)]
Arbitraly choose the Variant II for arm.

19 years agoMake setjmp()/longjmp() use sigprocmask.
cognet [Thu, 23 Sep 2004 23:13:46 +0000 (23:13 +0000)]
Make setjmp()/longjmp() use sigprocmask.

19 years agoAdd a dummy set_tp().
cognet [Thu, 23 Sep 2004 23:12:57 +0000 (23:12 +0000)]
Add a dummy set_tp().

19 years agoAlways use bx for returning on Xscale.
cognet [Thu, 23 Sep 2004 23:11:32 +0000 (23:11 +0000)]
Always use bx for returning on Xscale.

Obtained from: NetBSD

19 years agoFix the align-to-32-bits code.
cognet [Thu, 23 Sep 2004 23:10:56 +0000 (23:10 +0000)]
Fix the align-to-32-bits code.

Obtained from: NetBSD

19 years agokgmon works on arm, so make it parts of the build for arm as well.
cognet [Thu, 23 Sep 2004 23:06:29 +0000 (23:06 +0000)]
kgmon works on arm, so make it parts of the build for arm as well.

19 years agoAdd stubs for TLS.
cognet [Thu, 23 Sep 2004 23:04:52 +0000 (23:04 +0000)]
Add stubs for TLS.
Arbitraly choose the 2nd variant until I figure out which one I should use.

19 years agoImplement .init and .fini.
cognet [Thu, 23 Sep 2004 23:00:51 +0000 (23:00 +0000)]
Implement .init and .fini.

19 years agoAdd CPUCFLAGS for the strongarm and xscale CPUs.
cognet [Thu, 23 Sep 2004 22:59:31 +0000 (22:59 +0000)]
Add CPUCFLAGS for the strongarm and xscale CPUs.

19 years agoRe-route interrupts on arm as well.
cognet [Thu, 23 Sep 2004 22:58:43 +0000 (22:58 +0000)]
Re-route interrupts on arm as well.

19 years agoDo not use the IO-mapping to issue the reset on the 82546 on arm. For some
cognet [Thu, 23 Sep 2004 22:57:53 +0000 (22:57 +0000)]
Do not use the IO-mapping to issue the reset on the 82546 on arm. For some
reason, it results in corrupted descriptors.

19 years agoAdd the config file for the IQ31244 board.
cognet [Thu, 23 Sep 2004 22:55:00 +0000 (22:55 +0000)]
Add the config file for the IQ31244 board.

19 years agoUse the new KERNVIRTADDR and PHYSADDR options.
cognet [Thu, 23 Sep 2004 22:53:50 +0000 (22:53 +0000)]
Use the new KERNVIRTADDR and PHYSADDR options.
Add KDB.

19 years agoUse -O2 for the arm kernel as well
cognet [Thu, 23 Sep 2004 22:53:22 +0000 (22:53 +0000)]
Use -O2 for the arm kernel as well

19 years agoUse the address provided in the conf file, instead of hardcode 0xc0000000.
cognet [Thu, 23 Sep 2004 22:52:55 +0000 (22:52 +0000)]
Use the address provided in the conf file, instead of hardcode 0xc0000000.

19 years agoAdd new options :
cognet [Thu, 23 Sep 2004 22:52:25 +0000 (22:52 +0000)]
Add new options :
PHYSADDR : Address of the physical memory
KERNPHYSADDR : Physical address where the kernel starts
KERNVIRTADDR : Virtual address of the kernel
STARTUP_PAGETABLE_ADDR : Where to put the page table at bootstrap
+ Xscale specific options

19 years agoAdd the possibility to specify the kernel virtual address and the kernel
cognet [Thu, 23 Sep 2004 22:47:57 +0000 (22:47 +0000)]
Add the possibility to specify the kernel virtual address and the kernel
physical address in the kernel config file, as it varies from CPU to CPU.

19 years agoImport partial support for the IQ31244 eval board (i80321 CPU). IQ80321 might
cognet [Thu, 23 Sep 2004 22:45:36 +0000 (22:45 +0000)]
Import partial support for the IQ31244 eval board (i80321 CPU). IQ80321 might
work out of the box too, but I have no hardware to test.
It works well enough to go multiuser. Network works, SATA does not, as I have
no drive to test.
Thanks to Intel for sending such a board.

Obtained from:  NetBSD

19 years agoAdd Xscale common headers.
cognet [Thu, 23 Sep 2004 22:36:13 +0000 (22:36 +0000)]
Add Xscale common headers.

19 years agoBig cleanup: get ride of the whole spl level logic, as FreeBSD doesn't use
cognet [Thu, 23 Sep 2004 22:33:38 +0000 (22:33 +0000)]
Big cleanup: get ride of the whole spl level logic, as FreeBSD doesn't use
it anymore.

19 years agoNow that we have pmap_growkernel(), use more KVA.
cognet [Thu, 23 Sep 2004 22:32:33 +0000 (22:32 +0000)]
Now that we have pmap_growkernel(), use more KVA.

19 years agoRemove the empty definition of struct osigcontext, as it will never be used.
cognet [Thu, 23 Sep 2004 22:31:49 +0000 (22:31 +0000)]
Remove the empty definition of struct osigcontext, as it will never be used.

19 years agoRemove the pcb32_cstate field of struct pcb.
cognet [Thu, 23 Sep 2004 22:31:08 +0000 (22:31 +0000)]
Remove the pcb32_cstate field of struct pcb.

19 years agoDeclare sigcode and szsigcode.
cognet [Thu, 23 Sep 2004 22:30:05 +0000 (22:30 +0000)]
Declare sigcode and szsigcode.

19 years agoDefine VM_PROT_READ_IS_EXEC.
cognet [Thu, 23 Sep 2004 22:29:43 +0000 (22:29 +0000)]
Define VM_PROT_READ_IS_EXEC.

19 years agoImplement _mcount().
cognet [Thu, 23 Sep 2004 22:29:18 +0000 (22:29 +0000)]
Implement _mcount().

Obtained from: NetBSD

19 years agoDefine STACKALIGNBYTES and STACKALIGN.
cognet [Thu, 23 Sep 2004 22:27:42 +0000 (22:27 +0000)]
Define STACKALIGNBYTES and STACKALIGN.

19 years agoWe are using _mcount, not __mcount.
cognet [Thu, 23 Sep 2004 22:26:29 +0000 (22:26 +0000)]
We are using _mcount, not __mcount.
Remove the !__ELF__ case.

19 years agoUse sf_bufs for uiomove_fromphys().
cognet [Thu, 23 Sep 2004 22:25:20 +0000 (22:25 +0000)]
Use sf_bufs for uiomove_fromphys().

19 years agoOn Xscale, use the minicache for the kernel stack.
cognet [Thu, 23 Sep 2004 22:24:12 +0000 (22:24 +0000)]
On Xscale, use the minicache for the kernel stack.

19 years agoMake sure to call cred_update_thread() if needed.
cognet [Thu, 23 Sep 2004 22:22:33 +0000 (22:22 +0000)]
Make sure to call cred_update_thread() if needed.
Add partial support for KTRACE.

19 years agoImplement cpu_throw().
cognet [Thu, 23 Sep 2004 22:20:59 +0000 (22:20 +0000)]
Implement cpu_throw().

Obtained from: NetBSD

19 years agoRemove unused macroes.
cognet [Thu, 23 Sep 2004 22:18:56 +0000 (22:18 +0000)]
Remove unused macroes.
Add user, btrap, etrap, bintr and eintrt in the GPROF case.

19 years agoImplement sigreturn().
cognet [Thu, 23 Sep 2004 22:12:28 +0000 (22:12 +0000)]
Implement sigreturn().

19 years agoAdd the hw.machine sysctl.
cognet [Thu, 23 Sep 2004 22:11:43 +0000 (22:11 +0000)]
Add the hw.machine sysctl.

19 years agoRemove definitions related to the pmap cache state, and add TDF_NEEDRESCHED.
cognet [Thu, 23 Sep 2004 22:11:06 +0000 (22:11 +0000)]
Remove definitions related to the pmap cache state, and add TDF_NEEDRESCHED.

19 years agoAdd new functions to know which irqs are pending, and to mask and unmask
cognet [Thu, 23 Sep 2004 22:09:57 +0000 (22:09 +0000)]
Add new functions to know which irqs are pending, and to mask and unmask
interrupts, as these are CPU specific.
If the interrupt handler is not marked as INTR_FAST, don't unmask the
interrupt until it as been serviced.

19 years agoRename macroes, as we don't need to mess with alignment faults.
cognet [Thu, 23 Sep 2004 22:05:40 +0000 (22:05 +0000)]
Rename macroes, as we don't need to mess with alignment faults.
Call ast() if TDF_NEEDRESCHED is set too, not just TDF_ASTPENDING.

19 years agoUse sigcode.
cognet [Thu, 23 Sep 2004 22:03:25 +0000 (22:03 +0000)]
Use sigcode.

19 years agoIn db_stack_trace_cmd, remove the "pc" variable, we don't need it.
cognet [Thu, 23 Sep 2004 22:02:59 +0000 (22:02 +0000)]
In db_stack_trace_cmd, remove the "pc" variable, we don't need it.

19 years agoUse the right path for xscale files.
cognet [Thu, 23 Sep 2004 21:59:43 +0000 (21:59 +0000)]
Use the right path for xscale files.

19 years agoRemove bus_space_vaddr(), it does not exists in FreeBSD.
cognet [Thu, 23 Sep 2004 21:59:14 +0000 (21:59 +0000)]
Remove bus_space_vaddr(), it does not exists in FreeBSD.

19 years agoDon't attempt to manage our own segment list, and just remember the buffers
cognet [Thu, 23 Sep 2004 21:57:47 +0000 (21:57 +0000)]
Don't attempt to manage our own segment list, and just remember the buffers
provided.

Obtained from: NetBSD

19 years agoUse the right path for the bcopyinout_xscale.S file.
cognet [Thu, 23 Sep 2004 21:56:36 +0000 (21:56 +0000)]
Use the right path for the bcopyinout_xscale.S file.

19 years agoAdd MD syscalls to sync the icache and to drain the write buffer.
cognet [Thu, 23 Sep 2004 21:56:01 +0000 (21:56 +0000)]
Add MD syscalls to sync the icache and to drain the write buffer.

Obtained from: NetBSD

19 years agoImplement pmap_growkernel() and pmap_extract_and_hold().
cognet [Thu, 23 Sep 2004 21:54:25 +0000 (21:54 +0000)]
Implement pmap_growkernel() and pmap_extract_and_hold().
Remove the cache state logic : right now, it provides more problems than it
helps.
Add helper functions for mapping devices while bootstrapping.
Reorganize the code a bit, and remove dead code.

Obtained from: NetBSD (partially)

19 years agoMap the kernel very early if needed.
cognet [Thu, 23 Sep 2004 21:49:10 +0000 (21:49 +0000)]
Map the kernel very early if needed.
Implement sigcode.

19 years agoBackout the hack to install mtree, newsyslog, and touch in the cloned
brooks [Thu, 23 Sep 2004 21:17:31 +0000 (21:17 +0000)]
Backout the hack to install mtree, newsyslog, and touch in the cloned
root in the previous commit.  We'll fix /etc/rc.d/var directly (someday).

19 years agoImplement terminating the worker thread when the driver is about to
joerg [Thu, 23 Sep 2004 21:12:21 +0000 (21:12 +0000)]
Implement terminating the worker thread when the driver is about to
be deregistered.

Not yet tested, since by now, GEOM doesn't want us to deregister.  PHK
wants to fix that RSN.

19 years agoAdd locking notes on the members of rusage. Note that these notes reflect
jhb [Thu, 23 Sep 2004 21:08:14 +0000 (21:08 +0000)]
Add locking notes on the members of rusage.  Note that these notes reflect
what is already present in the tree and apply to the rusage struct for
current process usage rather than child process usage.

19 years agoUpdate locking notes on several fields to reflect locking already in the
jhb [Thu, 23 Sep 2004 21:07:13 +0000 (21:07 +0000)]
Update locking notes on several fields to reflect locking already in the
tree:
- td_standin is (k + a) as it is only touched by either curthread or when
  a thread is being created.
- td_upcall is (k + j)
- td_sticks is (k) rather than the earlier (j) note.
- td_uuticks and td_usticks are both (k).
- td_intrval is (j)
- Neither kg_nextupcall or kg_upquantum seem to be locked and that seems
  to be on purpose, so mark those as (n).

19 years agoDon't try to protect td_sticks with sched_lock. It doesn't need it as it
jhb [Thu, 23 Sep 2004 21:03:58 +0000 (21:03 +0000)]
Don't try to protect td_sticks with sched_lock.  It doesn't need it as it
is only accessed by curthread.

19 years ago- Assert sched_lock in upcall_remove() since it is needed there and all
jhb [Thu, 23 Sep 2004 21:03:16 +0000 (21:03 +0000)]
- Assert sched_lock in upcall_remove() since it is needed there and all
  callers already lock it there.
- Lock sched_lock slightly earlier in kse_create() so that it covers
  kg_numupcalls.

19 years ago- Don't try to unlock Giant if single threading fails since we don't have
jhb [Thu, 23 Sep 2004 21:01:50 +0000 (21:01 +0000)]
- Don't try to unlock Giant if single threading fails since we don't have
  it locked.
- Unlock Giant before calling exit1() since exit1() does not require Giant.

19 years agoRe-add namedb. I misunderstood what Doug said about using ISC's
des [Thu, 23 Sep 2004 19:33:54 +0000 (19:33 +0000)]
Re-add namedb.  I misunderstood what Doug said about using ISC's
layout: he meant for binaries, not configuration files.

19 years agoRedefine a PTE as a 64-bit integral type instead of a struct of
marcel [Thu, 23 Sep 2004 18:37:36 +0000 (18:37 +0000)]
Redefine a PTE as a 64-bit integral type instead of a struct of
bit-fields. Unify the PTE defines accordingly and update all
uses.

19 years agoPrefer C99's __func__ over GCC's __FUNCTION__.
stefanf [Thu, 23 Sep 2004 18:25:46 +0000 (18:25 +0000)]
Prefer C99's __func__ over GCC's __FUNCTION__.

Approved by: alfred

19 years agoNew release notes:
hrs [Thu, 23 Sep 2004 16:59:14 +0000 (16:59 +0000)]
New release notes:
SA-04:14.

19 years agoSplit the ioctl function in control and slave side, this eliminated
phk [Thu, 23 Sep 2004 16:13:46 +0000 (16:13 +0000)]
Split the ioctl function in control and slave side, this eliminated
a troublesome devsw() call.

19 years agoInvalidate cache after changing pte entry.
rik [Thu, 23 Sep 2004 16:06:27 +0000 (16:06 +0000)]
Invalidate cache after changing pte entry.

Discussed with: jhp and njl
MFC after: 5 days

19 years agoHack around the fact that the current diskless scripts require some
brooks [Thu, 23 Sep 2004 15:52:42 +0000 (15:52 +0000)]
Hack around the fact that the current diskless scripts require some
binaries in /usr before /usr is mounted by copying those files to the
root.

Stop making devices.  devfs is manditory.

19 years agoSort the SEE ALSO list correctly. Update the document date.
brooks [Thu, 23 Sep 2004 15:26:01 +0000 (15:26 +0000)]
Sort the SEE ALSO list correctly.  Update the document date.

Reported by: ru

19 years agoEliminate DEV_STRATEGY() macro: call dev_strategy() directly.
phk [Thu, 23 Sep 2004 14:45:04 +0000 (14:45 +0000)]
Eliminate DEV_STRATEGY() macro: call dev_strategy() directly.

Make dev_strategy() handle errors and departing devices properly.

19 years agoIn fullpath_from_shell(), move the nologin detection before the cat | while
roam [Thu, 23 Sep 2004 13:09:42 +0000 (13:09 +0000)]
In fullpath_from_shell(), move the nologin detection before the cat | while
loop to avoid an incorrect display of the nologin path twice.

PR: 71786
Submitted by: Andrew Hayden <andrew.hayden@gmail.com>
Reviewed by: mtm
MFC after: 3 days

19 years agoProtect sockaddr_union definitions with a protecting define. This allows to
mlaier [Thu, 23 Sep 2004 12:44:40 +0000 (12:44 +0000)]
Protect sockaddr_union definitions with a protecting define. This allows to
build kernels with FAST_IPSEC and PF. This is the least disruptive fix.

PR: kern/71836
Reviewed by: bms, various mailing lists
MFC after: 3 days

19 years agoRemove bogus cdevsw frobbing code which tries to prevent double
phk [Thu, 23 Sep 2004 12:21:11 +0000 (12:21 +0000)]
Remove bogus cdevsw frobbing code which tries to prevent double
loading of modules.

MODULE_VERSION() should be used for this I belive.

19 years agoDo not use devsw() but si_devsw direction. This is still bogus but a
phk [Thu, 23 Sep 2004 12:19:24 +0000 (12:19 +0000)]
Do not use devsw() but si_devsw direction.  This is still bogus but a
fair bit less so.

19 years agoIntroduce new /boot/loader.conf variable: root_mount_delay.
pjd [Thu, 23 Sep 2004 10:13:18 +0000 (10:13 +0000)]
Introduce new /boot/loader.conf variable: root_mount_delay.
It can be used to delay mounting root partition to give a chance to GEOM
providers to show up.
Now, when there is no needed provider, vfs_rootmount() function will look
for it every second and if it can't be find in defined time, it'll ask
for root device name (before this change it was done immediately).

This will allow to boot from gmirror device in degraded mode.

19 years agoPer recent HEADSUP: Disconnect (old)vinum from the kernel build.
phk [Thu, 23 Sep 2004 08:34:50 +0000 (08:34 +0000)]
Per recent HEADSUP:  Disconnect (old)vinum from the kernel build.

Users should move to the new geom_vinum implementation instead.

The refcount logic which is being added to devices to enable safe module
unloading and the buf/vm work also in progress would require a major rework
of the (old)-vinum code to comply with the new semantics.

The actual source files will not be removed until I have coordinated with
the geomvinum people if they need any bits repo-copied etc.

19 years agoStyle fix.
des [Thu, 23 Sep 2004 08:00:07 +0000 (08:00 +0000)]
Style fix.

Submitted by: ru@

19 years agoMore remnants of BIND 8 to de-orbit.
des [Thu, 23 Sep 2004 07:54:01 +0000 (07:54 +0000)]
More remnants of BIND 8 to de-orbit.

Reminded by: ru@

19 years agoInstead of hardcoding the BIND version, deduce it from ${BIND_DIR}/version.
des [Thu, 23 Sep 2004 07:52:51 +0000 (07:52 +0000)]
Instead of hardcoding the BIND version, deduce it from ${BIND_DIR}/version.

19 years agolibpthread is always libpthread, even when it's libc_r.
des [Thu, 23 Sep 2004 07:37:19 +0000 (07:37 +0000)]
libpthread is always libpthread, even when it's libc_r.

Reminded by: ru@

19 years agoDe-orbit more remnants of BIND 8.
des [Thu, 23 Sep 2004 07:36:31 +0000 (07:36 +0000)]
De-orbit more remnants of BIND 8.

Prodded by: ru@

19 years agoBump version number after vendor import of 9.3.0.
des [Thu, 23 Sep 2004 07:21:02 +0000 (07:21 +0000)]
Bump version number after vendor import of 9.3.0.

19 years agoThis commit was generated by cvs2svn to compensate for changes in r135601,
des [Thu, 23 Sep 2004 07:18:50 +0000 (07:18 +0000)]
This commit was generated by cvs2svn to compensate for changes in r135601,
which included commits to RCS files with non-trunk default branches.

19 years agoVendor import of BIND 9.3.0.
des [Thu, 23 Sep 2004 07:18:50 +0000 (07:18 +0000)]
Vendor import of BIND 9.3.0.

19 years agoDo not refcount the cdevsw, but rather maintain a cdev->si_threadcount
phk [Thu, 23 Sep 2004 07:17:41 +0000 (07:17 +0000)]
Do not refcount the cdevsw, but rather maintain a cdev->si_threadcount
of the number of threads which are inside whatever is behind the
cdevsw for this particular cdev.

Make the device mutex visible through dev_lock() and dev_unlock().
We may want finer granularity later.

Replace spechash_mtx use with dev_lock()/dev_unlock().

19 years ago- Wrapped BIND 9 libraries defines into !defined(NO_BIND).
ru [Thu, 23 Sep 2004 07:13:36 +0000 (07:13 +0000)]
- Wrapped BIND 9 libraries defines into !defined(NO_BIND).

- Added forgotten LIBLWRES to fix missing dependencies
  revealed by "make checkdpadd".

19 years agoMFen 1.258
den [Thu, 23 Sep 2004 06:49:40 +0000 (06:49 +0000)]
MFen 1.258

MFC after: 2 days

19 years agoPAE seems to work for isp- at least under mimimal testing.
mjacob [Thu, 23 Sep 2004 05:26:19 +0000 (05:26 +0000)]
PAE seems to work for isp- at least under mimimal testing.

19 years agoPAE support changes that included at least some minimal actual testing
mjacob [Thu, 23 Sep 2004 05:25:22 +0000 (05:25 +0000)]
PAE support changes that included at least some minimal actual testing
with a kernel that booted.

19 years agoRefrence bsdlabel.8 and bsdlabel.5 rather then the obsolete disklabel.8
brooks [Thu, 23 Sep 2004 04:57:58 +0000 (04:57 +0000)]
Refrence bsdlabel.8 and bsdlabel.5 rather then the obsolete disklabel.8
and disklabel.5.  Refrencing bsdlabel.5 is somewhat bogus as it is not
connected to the build and is mostly unwritten at this point.

Reported by: Matthias Andree <matthias.andree at web.de>
PR: docs/72020

19 years agoLOCALSTATEDIR should be /var, since the BIND 9 source appends
dougb [Thu, 23 Sep 2004 04:46:56 +0000 (04:46 +0000)]
LOCALSTATEDIR should be /var, since the BIND 9 source appends
things like "run/named.pid" to it.

19 years agoAdd a knob 'daily_status_security_diff_flags' controlling the
jkoshy [Thu, 23 Sep 2004 02:00:52 +0000 (02:00 +0000)]
Add a knob 'daily_status_security_diff_flags' controlling the
format of the 'diff' output generated during periodic(8) scripts.

Submitted by: keramida (script changes)
Reviewed by: keramida (man page changes)