]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
19 years ago- Implement svr4_emul_find() using kern_alternate_path(). This changes
jhb [Mon, 7 Feb 2005 21:53:42 +0000 (21:53 +0000)]
- Implement svr4_emul_find() using kern_alternate_path().  This changes
  the semantics in that the returned filename to use is now a kernel
  pointer rather than a user space pointer.  This required changing the
  arguments to the CHECKALT*() macros some and changing the various system
  calls that used pathnames to use the kern_foo() functions that can accept
  kernel space filename pointers instead of calling the system call
  directly.
- Use kern_open(), kern_access(), kern_msgctl(), kern_execve(),
  kern_mkfifo(), kern_mknod(), kern_statfs(), kern_fstatfs(),
  kern_setitimer(), kern_stat(), kern_lstat(), kern_fstat(), kern_utimes(),
  kern_pathconf(), and kern_unlink().

19 years agoImplement a kern_pathconf() wrapper for pathconf() which can take the
jhb [Mon, 7 Feb 2005 21:46:43 +0000 (21:46 +0000)]
Implement a kern_pathconf() wrapper for pathconf() which can take the
filename from either a user space or a kernel space pointer.

19 years agoIf the pointer to the new itimerval is NULL in kern_setitimer(), just
jhb [Mon, 7 Feb 2005 21:45:48 +0000 (21:45 +0000)]
If the pointer to the new itimerval is NULL in kern_setitimer(), just
read the old value via kern_getitimer().

19 years agoAvoid usage of implicit int.
stefanf [Mon, 7 Feb 2005 21:42:16 +0000 (21:42 +0000)]
Avoid usage of implicit int.

19 years agoInclude <stdlib.h> for exit().
stefanf [Mon, 7 Feb 2005 21:41:26 +0000 (21:41 +0000)]
Include <stdlib.h> for exit().

19 years agoAdd a prototype for localupdate().
stefanf [Mon, 7 Feb 2005 21:40:24 +0000 (21:40 +0000)]
Add a prototype for localupdate().

19 years agoAdd chroot. It can be a real lifesaver, and adds less than 2 kB.
des [Mon, 7 Feb 2005 21:19:18 +0000 (21:19 +0000)]
Add chroot.  It can be a real lifesaver, and adds less than 2 kB.

19 years agoAn aggressive little script I used to trim out duplicate fortunes.
ru [Mon, 7 Feb 2005 21:15:16 +0000 (21:15 +0000)]
An aggressive little script I used to trim out duplicate fortunes.

19 years agoTrim more cookies, by playing with different hash functions,
ru [Mon, 7 Feb 2005 21:00:20 +0000 (21:00 +0000)]
Trim more cookies, by playing with different hash functions,
e.g., by trimming all non-alphabet characters and whitespace,
converting to lowercase, and considering only first (or last)
N letters (maybe only consonants).  The fortune editor then
displays all fortunes that have the same hash, and allows to
remove one of them.  The rest is written to stdout.

19 years agoUse if_link_state_change() instead of rt_ifmsg(). Remove include net/route.h.
glebius [Mon, 7 Feb 2005 19:39:29 +0000 (19:39 +0000)]
Use if_link_state_change() instead of rt_ifmsg(). Remove include net/route.h.

Reviewed by: wpaul, sam

19 years ago- Use kern_{l,f,}stat() and kern_{f,}statfs() functions rather than
jhb [Mon, 7 Feb 2005 18:47:28 +0000 (18:47 +0000)]
- Use kern_{l,f,}stat() and kern_{f,}statfs() functions rather than
  duplicating the contents of the same functions inline.
- Consolidate common code to convert a BSD statfs struct to a Linux struct
  into a static worker function.

19 years agoMake linux_emul_convpath() a simple wrapper for kern_alternate_path().
jhb [Mon, 7 Feb 2005 18:46:05 +0000 (18:46 +0000)]
Make linux_emul_convpath() a simple wrapper for kern_alternate_path().

19 years ago- Tweak kern_msgctl() to return a copy of the requested message queue id
jhb [Mon, 7 Feb 2005 18:44:55 +0000 (18:44 +0000)]
- Tweak kern_msgctl() to return a copy of the requested message queue id
  structure in the struct pointed to by the 3rd argument for IPC_STAT and
  get rid of the 4th argument.  The old way returned a pointer into the
  kernel array that the calling function would then access afterwards
  without holding the appropriate locks and doing non-lock-safe things like
  copyout() with the data anyways.  This change removes that unsafeness and
  resulting race conditions as well as simplifying the interface.
- Implement kern_foo wrappers for stat(), lstat(), fstat(), statfs(),
  fstatfs(), and fhstatfs().  Use these wrappers to cut out a lot of
  code duplication for freebsd4 and netbsd compatability system calls.
- Add a new lookup function kern_alternate_path() that looks up a filename
  under an alternate prefix and determines which filename should be used.
  This is basically a more general version of linux_emul_convpath() that
  can be shared by all the ABIs thus allowing for further reduction of
  code duplication.

19 years agoVarious and sundry style fixes.
jhb [Mon, 7 Feb 2005 18:38:29 +0000 (18:38 +0000)]
Various and sundry style fixes.

19 years agoUse linux_emul_convpath() rather than linux_emul_find() as
jhb [Mon, 7 Feb 2005 18:37:51 +0000 (18:37 +0000)]
Use linux_emul_convpath() rather than linux_emul_find() as
linux_emul_find() is going away.

19 years agoUse the LCONVPATHEXIST() macro rather than it's exact expansion to be
jhb [Mon, 7 Feb 2005 18:37:13 +0000 (18:37 +0000)]
Use the LCONVPATHEXIST() macro rather than it's exact expansion to be
consistent.

19 years agoUse kern_setitimer() to implement linux_alarm() instead of fondling the
jhb [Mon, 7 Feb 2005 18:36:21 +0000 (18:36 +0000)]
Use kern_setitimer() to implement linux_alarm() instead of fondling the
real interval timer directly.

19 years agoUse kern_open() directly rather than a stackgap detour via open().
jhb [Mon, 7 Feb 2005 18:22:20 +0000 (18:22 +0000)]
Use kern_open() directly rather than a stackgap detour via open().

19 years agoDrop Giant before calling kthread_exit().
jhb [Mon, 7 Feb 2005 18:21:50 +0000 (18:21 +0000)]
Drop Giant before calling kthread_exit().

19 years agoAvoid using tsleep() in the resume path as it may result in the
mdodd [Mon, 7 Feb 2005 17:14:42 +0000 (17:14 +0000)]
Avoid using tsleep() in the resume path as it may result in the
system hanging if timer interrupts aren't running yet.

This allows my Thinkpad to resume successfully with APM.

Approved by:  sos
MFC after:  2 weeks

19 years agoRemove an unneeded check.
harti [Mon, 7 Feb 2005 16:33:57 +0000 (16:33 +0000)]
Remove an unneeded check.

19 years agoIntroduce a typedef for variable value modifation functions and use it
harti [Mon, 7 Feb 2005 16:27:19 +0000 (16:27 +0000)]
Introduce a typedef for variable value modifation functions and use it
where appropriate.

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoMake the intention of a loop clearer.
harti [Mon, 7 Feb 2005 15:51:51 +0000 (15:51 +0000)]
Make the intention of a loop clearer.

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoAdd VNASSERT() which is just like KASSERT() but takes a vnode argument
phk [Mon, 7 Feb 2005 12:56:19 +0000 (12:56 +0000)]
Add VNASSERT() which is just like KASSERT() but takes a vnode argument
which it will vn_printf() if it triggers.

19 years agoRemove unused defines.
nyan [Mon, 7 Feb 2005 12:23:07 +0000 (12:23 +0000)]
Remove unused defines.

19 years agoBe more verbose about altq SYNOPSIS and add more linkage in the relating pf
mlaier [Mon, 7 Feb 2005 11:46:36 +0000 (11:46 +0000)]
Be more verbose about altq SYNOPSIS and add more linkage in the relating pf
documents.

Inspired by: scottl
Reviewed by: Brad Davis <so14kNOso14kSPAMcom>
MFC after: 3 days

19 years agoFix the problem with incorrect throttling level reported immediately after
sobomax [Mon, 7 Feb 2005 11:35:24 +0000 (11:35 +0000)]
Fix the problem with incorrect throttling level reported immediately after
reboot. Safter the reboot the TCC is usually in the Automatic mode, in which
reading current performance level is likely to produce bogus results make sure
to switch it to the On-Demand mode and set to some known performance level.
Unfortunately there is no reliable way to check that TCC is in the Automatic
mode. Reading bit 4 of ACPI Thermal Monitor Control Register produces 0
regardless of the current mode.

MFC after: 1 week

19 years agoIntroduce Buf_StripNewLines() and use it where appropriate.
harti [Mon, 7 Feb 2005 11:27:47 +0000 (11:27 +0000)]
Introduce Buf_StripNewLines() and use it where appropriate.

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoFix "to a file" example so it works (one must touch(1) the file first).
obrien [Mon, 7 Feb 2005 11:26:53 +0000 (11:26 +0000)]
Fix "to a file" example so it works (one must touch(1) the file first).

19 years agoImprove parsing of hook name.
glebius [Mon, 7 Feb 2005 11:16:07 +0000 (11:16 +0000)]
Improve parsing of hook name.

Submitted by: ru

19 years agoAccess vmobject via the bufobj instead of the vnode
phk [Mon, 7 Feb 2005 10:04:06 +0000 (10:04 +0000)]
Access vmobject via the bufobj instead of the vnode

19 years agoVOP_DESTROYVOBJECT() is no more.
phk [Mon, 7 Feb 2005 09:26:58 +0000 (09:26 +0000)]
VOP_DESTROYVOBJECT() is no more.

19 years agoRemove vop_stddestroyvobject()
phk [Mon, 7 Feb 2005 09:26:39 +0000 (09:26 +0000)]
Remove vop_stddestroyvobject()

19 years agoRemove vop_destroyvobject()
phk [Mon, 7 Feb 2005 09:23:34 +0000 (09:23 +0000)]
Remove vop_destroyvobject()

19 years agoFix disorder.
ru [Mon, 7 Feb 2005 09:15:52 +0000 (09:15 +0000)]
Fix disorder.

19 years agoSort SEE ALSO.
glebius [Mon, 7 Feb 2005 08:51:34 +0000 (08:51 +0000)]
Sort SEE ALSO.

Submitted by: ru

19 years agoFix bugs in the previous commit.
ru [Mon, 7 Feb 2005 08:44:34 +0000 (08:44 +0000)]
Fix bugs in the previous commit.

19 years agoDeimplement vop_destroyvobject()
phk [Mon, 7 Feb 2005 08:23:36 +0000 (08:23 +0000)]
Deimplement vop_destroyvobject()

19 years agoAdd missing isa_dmatc() function.
phk [Mon, 7 Feb 2005 08:19:53 +0000 (08:19 +0000)]
Add missing isa_dmatc() function.

This may or may not be correct,  Only the pcii driver would notice and
it doesn't support PC98 yet.

19 years agoEnglish and mdoc(7) cleanup.
glebius [Mon, 7 Feb 2005 08:17:51 +0000 (08:17 +0000)]
English and mdoc(7) cleanup.

Submitted by: ru

19 years agoRemove vop_destroyvobject() initialization.
phk [Mon, 7 Feb 2005 08:04:24 +0000 (08:04 +0000)]
Remove vop_destroyvobject() initialization.

19 years agoSort branches in correct numeric order.
des [Mon, 7 Feb 2005 07:55:00 +0000 (07:55 +0000)]
Sort branches in correct numeric order.

19 years agoInvent a Buf_AppendRange function that appends a non-NUL-terminated string
harti [Mon, 7 Feb 2005 07:54:23 +0000 (07:54 +0000)]
Invent a Buf_AppendRange function that appends a non-NUL-terminated string
given by a pointer to the start of the string and a pointer one behind
the end.

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoInvent a Buf_Append function to append a NUL-terminated string
harti [Mon, 7 Feb 2005 07:49:16 +0000 (07:49 +0000)]
Invent a Buf_Append function to append a NUL-terminated string
and use it thoughout the code.

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoDon't call VOP_DESTROYVOBJECT(), trust that VOP_RECLAIM() did what
phk [Mon, 7 Feb 2005 07:48:03 +0000 (07:48 +0000)]
Don't call VOP_DESTROYVOBJECT(), trust that VOP_RECLAIM() did what
was necessary.

19 years agoAdd a missing prefix to a struct field for consistency.
phk [Mon, 7 Feb 2005 07:40:39 +0000 (07:40 +0000)]
Add a missing prefix to a struct field for consistency.

19 years agoRevamp several Makefile nits:
delphij [Mon, 7 Feb 2005 05:45:38 +0000 (05:45 +0000)]
Revamp several Makefile nits:
- .PATH: is spelled ``.PATH: ''
- Don't forget to use DPADD so ``make checkdpadd'' is
  not broken.
- LDADD should not come with + as it's empty in the
  first place

Suggested by: ru

19 years agoUndo the VCS tag move to reduce diff hunks.
delphij [Mon, 7 Feb 2005 05:34:35 +0000 (05:34 +0000)]
Undo the VCS tag move to reduce diff hunks.

Pointed out by: ru

19 years agoRemove /stand when we are done with it.
obrien [Mon, 7 Feb 2005 04:25:34 +0000 (04:25 +0000)]
Remove /stand when we are done with it.

Submitted by: Ryan Sommers <ryans@gamersimpact.com>

19 years agoRemove handling _PSS notifies from acpi_cpu and let acpi_perf handle them.
njl [Mon, 7 Feb 2005 04:03:06 +0000 (04:03 +0000)]
Remove handling _PSS notifies from acpi_cpu and let acpi_perf handle them.

19 years agoAdd a mechanism for associating a mutex with a callout when the
iedowse [Mon, 7 Feb 2005 02:47:33 +0000 (02:47 +0000)]
Add a mechanism for associating a mutex with a callout when the
callout is first initialised, using a new function callout_init_mtx().
The callout system will acquire this mutex before calling the callout
function and release it on return.

In addition, the callout system uses the mutex to avoid most of the
complications and race conditions inherent in asynchronous timer
facilities, so mutex-protected callouts have much simpler semantics.
As long as the mutex is held when invoking callout_stop() or
callout_reset(), then these functions will guarantee that the callout
will be stopped, even if softclock() had already begun to process
the callout.

Existing Giant-locked callouts will automatically pick up the new
race-free semantics. This should close a number of race conditions
in the USB code and probably other areas of the kernel too.

There should be no change in behaviour for "MP-safe" callouts; these
still need to use the techniques mentioned in timeout(9) to avoid
race conditions.

19 years agoEnable interrupt routing as first choice on alpha.
ticso [Mon, 7 Feb 2005 00:43:14 +0000 (00:43 +0000)]
Enable interrupt routing as first choice on alpha.
The alpha default handler knows how to trigger a fallback.

19 years agoAllocate more space for softc. Amazing my machine survived this.
phk [Sun, 6 Feb 2005 22:45:12 +0000 (22:45 +0000)]
Allocate more space for softc.  Amazing my machine survived this.

19 years agoRemove a comment from the raw IPv6 output function regarding
rwatson [Sun, 6 Feb 2005 21:43:55 +0000 (21:43 +0000)]
Remove a comment from the raw IPv6 output function regarding
M_TRYWAIT allocations: M_PREPEND() now uses M_DONTWAIT.

MFC after: 3 days

19 years agoNote changes to throttling sysctls and new power_profile(8) cpufrq support.
njl [Sun, 6 Feb 2005 21:31:41 +0000 (21:31 +0000)]
Note changes to throttling sysctls and new power_profile(8) cpufrq support.

19 years agoNote the cpufreq import and acpi throttling changes.
njl [Sun, 6 Feb 2005 21:24:50 +0000 (21:24 +0000)]
Note the cpufreq import and acpi throttling changes.

19 years agoMove the $NetBSD$ tag to the top of the file, per standard NetBSD
imp [Sun, 6 Feb 2005 21:15:19 +0000 (21:15 +0000)]
Move the $NetBSD$ tag to the top of the file, per standard NetBSD
practice (which we seem to mostly follow in the tree).  Move the
$FreeBSD$ tag to its more proper place after all copyright and license
notices.  Add '-' to the copyright notice for Christian E. Hopps so my
copyright script picks it up.

19 years agoHook acpi_throttle(4) up to the build. It's currently part of acpi_perf.ko
njl [Sun, 6 Feb 2005 21:13:41 +0000 (21:13 +0000)]
Hook acpi_throttle(4) up to the build.  It's currently part of acpi_perf.ko
although this may change.

19 years agoAdd support for cpufreq to power_profile(8). Values for on/offline cpu
njl [Sun, 6 Feb 2005 21:12:25 +0000 (21:12 +0000)]
Add support for cpufreq to power_profile(8).  Values for on/offline cpu
frequencies are specified with performance_cpu_freq and economy_cpu_freq.
Of course, special values LOW and HIGH are also supported.  Also, remove
old throttling support.

19 years agoccs is a size_t. RAY_CCS_INDEX takes ccs and does math on it,
imp [Sun, 6 Feb 2005 21:12:22 +0000 (21:12 +0000)]
ccs is a size_t.  RAY_CCS_INDEX takes ccs and does math on it,
resulting in a size_t due to C's rules of arithmetic.  Rather than
bogusly cast the result to a uint8_t, fix the printf format specifier
to have a 'z' modifier which tells the compiler that the sizes really
do match.

It turns out that change 1.75 was incorrect to assume that this
'really' was a 8bit quantity.  It isn't.  Although the hardware
appears to limit things to < 256, it would be a bug that should be
caught by debug printf it it were.  Casting it to uint8_t would have
lost this useful information.

Aslo add 'z' to a nearby debug statement that's never compiled in.

19 years agoRemove acpi throttling support from the acpi_cpu(4) driver now that this
njl [Sun, 6 Feb 2005 21:10:19 +0000 (21:10 +0000)]
Remove acpi throttling support from the acpi_cpu(4) driver now that this
is supported by acpi_throttle(4).

19 years agoBreak out acpi throttling support into a new relative cpufreq driver,
njl [Sun, 6 Feb 2005 21:09:51 +0000 (21:09 +0000)]
Break out acpi throttling support into a new relative cpufreq driver,
acpi_throttle(4).

19 years agoAdd support for relative cpufreq drivers. Such drivers modulate clock
njl [Sun, 6 Feb 2005 21:08:35 +0000 (21:08 +0000)]
Add support for relative cpufreq drivers.  Such drivers modulate clock
frequency as a percentage of the base rate and do not change the base
rate directly.  The cpufreq framework combines these with absolute drivers
to produce synthesized levels made of one or more settings.

19 years agoMove resource allocation routines from cardbus_cis.c to cardbus.c.
imp [Sun, 6 Feb 2005 21:03:13 +0000 (21:03 +0000)]
Move resource allocation routines from cardbus_cis.c to cardbus.c.
They have nothing at all to do with CIS parsing.

Remove some unused funce parsing: nothing used the results.

Use more of pccard_cis.h's deifnitions for the cardbus specific cis
parsing we do.  More work is needed in this area.

This reduces the size of the cardbus module by 380 bytes or so...

19 years agoNotify the OS that we're taking over Px states in acpi_perf(4) instead of
njl [Sun, 6 Feb 2005 20:12:28 +0000 (20:12 +0000)]
Notify the OS that we're taking over Px states in acpi_perf(4) instead of
doing it in the cpu driver.  The previous code was incorrect anyway since
this value controls Px states, not throttling as the comment said.  Since
we didn't support Px states before, there was no impact.  Also, note that
we delay the write to SMI_CMD until after booting is complete since it
sometimes triggers a change in the frequency and we want to have all
drivers ready to detect/handle this.

19 years agoWhitespace.
glebius [Sun, 6 Feb 2005 19:24:59 +0000 (19:24 +0000)]
Whitespace.

19 years agoRemove comment which left after removal of ng_rcvdataq_t.
glebius [Sun, 6 Feb 2005 19:20:16 +0000 (19:20 +0000)]
Remove comment which left after removal of ng_rcvdataq_t.

MFC after: 1 week

19 years agoWhen running Linux binaries, set up the initial FPU state as Linux
das [Sun, 6 Feb 2005 17:29:20 +0000 (17:29 +0000)]
When running Linux binaries, set up the initial FPU state as Linux
would.

PR: 28966

19 years agoCheck that we have at least a 586-class CPU before calling do_cpuid().
iedowse [Sun, 6 Feb 2005 16:55:52 +0000 (16:55 +0000)]
Check that we have at least a 586-class CPU before calling do_cpuid().
This fixes booting on a number of 486 processors.

PR: i386/75686
Reviewed by: markm
MFC after: 1 week

19 years agoDocument NO_GPIB
phk [Sun, 6 Feb 2005 16:46:49 +0000 (16:46 +0000)]
Document NO_GPIB

19 years agoTie libgpib into the build (unless NO_GPIB)
phk [Sun, 6 Feb 2005 16:45:57 +0000 (16:45 +0000)]
Tie libgpib into the build (unless NO_GPIB)

19 years agoAdd libgpib containing the glue to get ibfoo() calls into the kernel.
phk [Sun, 6 Feb 2005 16:45:20 +0000 (16:45 +0000)]
Add libgpib containing the glue to get ibfoo() calls into the kernel.

19 years agoSince it is not un-common for a process's resident set size (rss)
csjp [Sun, 6 Feb 2005 16:34:49 +0000 (16:34 +0000)]
Since it is not un-common for a process's resident set size (rss)
to exceed 10 megabytes in size (especially in X), bump the max
column width from 4 bytes to 5. This will make the ps auxw output
uniform again when a process's rss exceeds 10 megs.

It should be noted that when 5 digits becomes to small, other
solutions should be explored such as displaying them in megabytes
or having ps automatically re-size column widths.

Discussed with: gad
MFC after: 1 week

19 years agoAdd instructions about upgrading netcat(1).
delphij [Sun, 6 Feb 2005 15:26:44 +0000 (15:26 +0000)]
Add instructions about upgrading netcat(1).

19 years agoFurther elaborate the GPIB driver. We now support a minimal subset of
phk [Sun, 6 Feb 2005 15:22:23 +0000 (15:22 +0000)]
Further elaborate the GPIB driver.  We now support a minimal subset of
the ibfoo() API.

19 years agoInstall GPIB related includefiles (unless NO_GPIB)
phk [Sun, 6 Feb 2005 15:18:24 +0000 (15:18 +0000)]
Install GPIB related includefiles (unless NO_GPIB)

19 years agoAdd directories for GPIB support
phk [Sun, 6 Feb 2005 15:17:21 +0000 (15:17 +0000)]
Add directories for GPIB support

19 years agoDocument NO_NETCAT.
delphij [Sun, 6 Feb 2005 14:59:03 +0000 (14:59 +0000)]
Document NO_NETCAT.

19 years agoFinish the import of nc(1) from OpenBSD. This includes:
delphij [Sun, 6 Feb 2005 14:44:27 +0000 (14:44 +0000)]
Finish the import of nc(1) from OpenBSD.  This includes:
- Bring IPsec support from the ports collection [1].
- Bring -o ("once only") option from the ports
  collection [2].
- Adopt the Makefile framework into
  usr.bin/nc/Makefile.
- Add a knob to control whether to build nc(1),
  NO_NETCAT.
- Bump __FreeBSD_version so ports collection can
  detect this change.

Original patchset are contributed to the ports collection by:
[1] nectar, [2] joerg.

Note: WARNS?=6 patchset spined off in this commit, in order not
to take too many files off the vendor branch.

19 years agoRemove code, inherited from ipacctctl, which sets socket to non-blocking
glebius [Sun, 6 Feb 2005 14:00:57 +0000 (14:00 +0000)]
Remove code, inherited from ipacctctl, which sets socket to non-blocking
mode, since this introduces problems on SMP.

19 years agoReplace RELENG_4_9 (no longer supported) with RELENG_4_11.
des [Sun, 6 Feb 2005 13:56:12 +0000 (13:56 +0000)]
Replace RELENG_4_9 (no longer supported) with RELENG_4_11.

19 years agoSince we are quite unlikely to ever face another platform which
phk [Sun, 6 Feb 2005 13:46:39 +0000 (13:46 +0000)]
Since we are quite unlikely to ever face another platform which
uses the i8237 without trying to emulate the PC architecture move
the register definitions for the i8237 chip into the central include
file for the chip, except for the PC98 case which is magic.

Add new isa_dmatc() function which tells us as cheaply as possible
if the terminal count has been reached for a given channel.

19 years agoSort includes and remove duplicate include.
nyan [Sun, 6 Feb 2005 13:31:22 +0000 (13:31 +0000)]
Sort includes and remove duplicate include.

19 years agoMFi386: revision 1.606.
nyan [Sun, 6 Feb 2005 13:23:20 +0000 (13:23 +0000)]
MFi386: revision 1.606.

19 years agoRemove the useless "version" number output.
obrien [Sun, 6 Feb 2005 12:54:54 +0000 (12:54 +0000)]
Remove the useless "version" number output.
We don't give a version number each userland binary.

19 years agoFix problem with some logitec usb wireless mice.
obrien [Sun, 6 Feb 2005 12:41:00 +0000 (12:41 +0000)]
Fix problem with some logitec usb wireless mice.

Submitted by: Markus <mw@kpnqwest.ch>
Tested by: Randy Bush <randy@psg.com>

19 years agoJump to common action checks after doing specific once. This fixes adding
glebius [Sun, 6 Feb 2005 11:13:59 +0000 (11:13 +0000)]
Jump to common action checks after doing specific once. This fixes adding
of divert rules, which I break in previous commit.

Pointy hat to: glebius

19 years agoList the LSI 320-2E PCI-Express controller as supported.
rushani [Sun, 6 Feb 2005 10:49:28 +0000 (10:49 +0000)]
List the LSI 320-2E PCI-Express controller as supported.

Reviewed by: scottl@

19 years agoo Implement net.inet.tcp.drop sysctl and userland part, tcpdrop(8)
maxim [Sun, 6 Feb 2005 10:47:12 +0000 (10:47 +0000)]
o Implement net.inet.tcp.drop sysctl and userland part, tcpdrop(8)
utility:

    The tcpdrop command drops the TCP connection specified by the
    local address laddr, port lport and the foreign address faddr,
    port fport.

Obtained from: OpenBSD
Reviewed by: rwatson (locking), ru (man page), -current
MFC after: 1 month

19 years agoStaticize the legacy cpu devclasses and revert the name for the acpi_cpu
njl [Sun, 6 Feb 2005 07:36:08 +0000 (07:36 +0000)]
Staticize the legacy cpu devclasses and revert the name for the acpi_cpu
devclass.  As pointed out by dfr@, devclasses don't have to share the same
linkage if multiple drivers have the same name.  Newbus should match the
devclasses based on name and allocate non-conflicting unit numbers.

19 years agoUpdate my email address.
das [Sun, 6 Feb 2005 03:23:31 +0000 (03:23 +0000)]
Update my email address.

19 years agoFinish the job of sorting all includes and fix the build by including
njl [Sun, 6 Feb 2005 01:55:08 +0000 (01:55 +0000)]
Finish the job of sorting all includes and fix the build by including
malloc.h before proc.h on sparc64.  Noticed by das@

Compiled on: alpha, amd64, i386, pc98, sparc64

19 years agoNotice that NG_VERSION has been increased.
glebius [Sat, 5 Feb 2005 23:25:59 +0000 (23:25 +0000)]
Notice that NG_VERSION has been increased.

19 years agoIncrease size of arglen to uint32_t, since uint16_t proved to be
glebius [Sat, 5 Feb 2005 23:23:14 +0000 (23:23 +0000)]
Increase size of arglen to uint32_t, since uint16_t proved to be
not enough (e.g. listing 911 nodes). Bump NG_VERSION.

Reviewed by: julian, archie, ru

19 years agoMake cpu_est_clockrate() more accurate by disabling interrupts for the
njl [Sat, 5 Feb 2005 23:16:27 +0000 (23:16 +0000)]
Make cpu_est_clockrate() more accurate by disabling interrupts for the
millisecond it is calibrating.  Suggested by jhb@ and bde@.  Don't clobber
the tsc_freq with the new value since it isn't accurate enough for
timecounters and the timecounter system as a whole needs support for
changing rates before we do this.  Subtract 0.5% from our measurement
to account for overhead in DELAY.  Note that this interface is for
estimating the clockrate and needs to work well at runtime so doing a full
calibration including disabling interrupts for a second is not feasible.

19 years agoConvert to the new GAS APIs to allow for detach in the future. Also, check
njl [Sat, 5 Feb 2005 22:30:57 +0000 (22:30 +0000)]
Convert to the new GAS APIs to allow for detach in the future.  Also, check
the PERF_CTRL register in our probe method so that we can tell earlier
that another driver should handle this device due to FFixedHW.  This avoids
scaring users when attach failed when we really wanted probe to fail.

19 years agoConvert to the new GAS API so that we can free registers in the future.
njl [Sat, 5 Feb 2005 22:29:03 +0000 (22:29 +0000)]
Convert to the new GAS API so that we can free registers in the future.

19 years agoConvert the acpi_bus_alloc_gas() and acpi_PkgGas() APIs to output the memory
njl [Sat, 5 Feb 2005 22:28:36 +0000 (22:28 +0000)]
Convert the acpi_bus_alloc_gas() and acpi_PkgGas() APIs to output the memory
type.  This is needed if the resource is to be released later.  The RID is
still also present, though less necessary since rman_get_rid() can be used
to obtain it from the resource.

19 years agoAcquire the source pmap's lock in pmap_copy().
alc [Sat, 5 Feb 2005 22:06:47 +0000 (22:06 +0000)]
Acquire the source pmap's lock in pmap_copy().

19 years agoBuild cpufreq and acpi_perf on platforms that are likely to be able to
njl [Sat, 5 Feb 2005 21:01:09 +0000 (21:01 +0000)]
Build cpufreq and acpi_perf on platforms that are likely to be able to
use them.