]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
18 years agoAdd a new vendor ID for Personal Cummunication Systems. They make a nice
mux [Sat, 31 Dec 2005 04:41:17 +0000 (04:41 +0000)]
Add a new vendor ID for Personal Cummunication Systems.  They make a nice
USB HID device that allows to plug two PS2 controllers.  This specific
device doesn't work yet but will as soon as we support devices with
multiple report IDs.

MFC after: 3 days

18 years agoAdd support for the Wacom Graphire 3 4x5. Like the Graphire 3, it has a
mux [Sat, 31 Dec 2005 04:38:50 +0000 (04:38 +0000)]
Add support for the Wacom Graphire 3 4x5.  Like the Graphire 3, it has a
broken report descriptor.  While I'm here, make all the other report
descriptors const to match the newly added one.

Obtained from: NetBSD
MFC after: 1 week

18 years agoAdd support for Xbox 360 gamepads. They are HID-capable devices, but
mux [Sat, 31 Dec 2005 04:22:34 +0000 (04:22 +0000)]
Add support for Xbox 360 gamepads.  They are HID-capable devices, but
lack a report descriptor and don't use the standard interface class.
This patch works around these deficiencies so that the uhid(4) driver
can recognize and use those broken devices.

PR: usb/90141
Submitted by: Ed Schouten <ed@fxq.nl> (with minor mods from me)
MFC after: 1 week

18 years agoAdd basic documentation for many of the kernel USB functions that
iedowse [Sat, 31 Dec 2005 03:53:02 +0000 (03:53 +0000)]
Add basic documentation for many of the kernel USB functions that
USB device drivers use to talk to USB devices. This is probably
still a bit rough and it does not yet include the functions specific
to HID, ethernet, hubs, host controller drivers, task threads or
debugging.

18 years agoFix LOR #174.
ariff [Sat, 31 Dec 2005 01:58:40 +0000 (01:58 +0000)]
Fix LOR #174.

Tested with: WITNESS, INVARIANTS and DIAGNOSTICS

18 years agoFixing build bustage.
avatar [Sat, 31 Dec 2005 01:45:40 +0000 (01:45 +0000)]
Fixing build bustage.

18 years agoOnly clear sb.st_size if it is clearly wrong or meaningless. This fixes
des [Fri, 30 Dec 2005 23:36:26 +0000 (23:36 +0000)]
Only clear sb.st_size if it is clearly wrong or meaningless.  This fixes
mirror mode.

PR: bin/86940
MFC after: 2 weeks

18 years agoAdd a FreeBSD-specific -S flag which controls the maximum size of an argument
jmallett [Fri, 30 Dec 2005 23:25:41 +0000 (23:25 +0000)]
Add a FreeBSD-specific -S flag which controls the maximum size of an argument
having replacements done in it via -I.

18 years agoHandle maxsize==0 in such a way that we don't modify the string.
jmallett [Fri, 30 Dec 2005 23:22:50 +0000 (23:22 +0000)]
Handle maxsize==0 in such a way that we don't modify the string.

18 years agoAdd hotplug support for Marvell chips.
sos [Fri, 30 Dec 2005 23:15:02 +0000 (23:15 +0000)]
Add hotplug support for Marvell chips.

18 years agoOn i386, 7.0 requires 106MB on /usr. 6.0 requires 101MB.
ceri [Fri, 30 Dec 2005 21:45:10 +0000 (21:45 +0000)]
On i386, 7.0 requires 106MB on /usr.  6.0 requires 101MB.
Bump USR_MIN_SIZE to 128MB, so that an auto-layout install won't fail
due to /usr being too small.

MFC After: 8 days

18 years agoMake RID2BAR actually return a number that can be plugged into PCIR_BAR()
imp [Fri, 30 Dec 2005 20:25:40 +0000 (20:25 +0000)]
Make RID2BAR actually return a number that can be plugged into PCIR_BAR()
rather than a bitmask.

Submitted by: ru@

18 years agomkstemp(3) and friends use a combination of base 10 digits, along
csjp [Fri, 30 Dec 2005 20:10:21 +0000 (20:10 +0000)]
mkstemp(3) and friends use a combination of base 10 digits, along
with upper and lower case letters from the English alphabet. Change
the number of possible file names mktemp will return from 26**6
to (10+26+26)**6 instead. This keeps things consistent with mkstemp(3)

18 years agoRemove now-obsolete printf warning.
imp [Fri, 30 Dec 2005 19:38:47 +0000 (19:38 +0000)]
Remove now-obsolete printf warning.

18 years agoRemove debug now that I've looped back the big changes into my p4 tree.
imp [Fri, 30 Dec 2005 19:36:29 +0000 (19:36 +0000)]
Remove debug now that I've looped back the big changes into my p4 tree.

18 years agoExpose pci_add_resources to the outside world, add a 'force' flag to
imp [Fri, 30 Dec 2005 19:28:26 +0000 (19:28 +0000)]
Expose pci_add_resources to the outside world, add a 'force' flag to
force allocation of unallocated BARs (cardbus uses this to preallocate
everything).  Add a prefetchmask to allow for busses that get prefetch
hints to set them.  Addjust pci_add_map and pci_ata_maps to take a new
force flag which pci_add_resources will pass in.  Implement 'force' in
pci_add_map.  Write new value of allocated resource into the bar, if
the allocation succeeded (we should have done this before, but with
the new force the bug was very obvious).

18 years agoMore scanners supported.
brueffer [Fri, 30 Dec 2005 19:25:28 +0000 (19:25 +0000)]
More scanners supported.

18 years agoMove all of the resource allocation into the pci layer. The resource
imp [Fri, 30 Dec 2005 19:25:04 +0000 (19:25 +0000)]
Move all of the resource allocation into the pci layer.  The resource
allocation here just duplicated it (badly).

18 years agoThe RID2BAR macro returns a number, not a bitmask. Fix this.
imp [Fri, 30 Dec 2005 19:23:32 +0000 (19:23 +0000)]
The RID2BAR macro returns a number, not a bitmask.  Fix this.

Spotted by: ru, jhb

18 years agoReflect progress in process title.
des [Fri, 30 Dec 2005 18:17:11 +0000 (18:17 +0000)]
Reflect progress in process title.

Submitted by: flz
MFC after: 2 weeks

18 years agoDon't be sexist - fix a typo.
ceri [Fri, 30 Dec 2005 17:59:23 +0000 (17:59 +0000)]
Don't be sexist - fix a typo.

18 years agoWhen printing SSID's in ifconfig(8)'s scan mode, set the width of the
rwatson [Fri, 30 Dec 2005 17:27:26 +0000 (17:27 +0000)]
When printing SSID's in ifconfig(8)'s scan mode, set the width of the
ssid field to 32 characters instead of 14, as long SSID's are quite
common and hard to type in if you can't read them.

18 years agoRemove duplicate options (originals in sys/conf/NOTES).
ru [Fri, 30 Dec 2005 16:45:31 +0000 (16:45 +0000)]
Remove duplicate options (originals in sys/conf/NOTES).

Reported by: fresh config(8)

18 years ago- Avoid adding devices multiple times to the device list.
ru [Fri, 30 Dec 2005 16:28:06 +0000 (16:28 +0000)]
- Avoid adding devices multiple times to the device list.
- Avoid adding options multiple times to the option list.

Based on a patch by: Matt Emmerton <matt@gsicomp.on.ca>

18 years agoClean up most of the "XXX"-tagged items:
ru [Fri, 30 Dec 2005 15:29:50 +0000 (15:29 +0000)]
Clean up most of the "XXX"-tagged items:

- The code that creates hints.c and env.c from the skeleton files
  moved into separate functions.

- Sanity checks for missing "ident" and "cputype" directives moved
  into main(), alongside the existing check for "machine".

PR: bin/90310
Submitted by: Matt Emmerton <matt@gsicomp.on.ca>

18 years agoFix the location of the initdiskless rc script.
ceri [Fri, 30 Dec 2005 14:43:31 +0000 (14:43 +0000)]
Fix the location of the initdiskless rc script.
Lose pointers to where these files were two major releases ago.

PR: docs/90370
Submitted by: Daniel Feenberg <feenberg at nber dot org>
Harassed by: Daniel Gerzo <danger at rulez dot sk>

18 years agoAdd nofification events to the list of event types.
ceri [Fri, 30 Dec 2005 14:01:01 +0000 (14:01 +0000)]
Add nofification events to the list of event types.

PR: docs/88390
Submitted by: Niki Denev <nike_d at cytexbg dot com>
Discussed with: Daniel Gerzo <danger at rulez dot sk>
MFC After: 8 days

18 years agoinetd(8) requires that /etc/netconfig be present, and contain entries
ceri [Fri, 30 Dec 2005 13:16:15 +0000 (13:16 +0000)]
inetd(8) requires that /etc/netconfig be present, and contain entries
for each of udp and tcp (and their IPv6 equivalents when INET6 is
enabled).  Note that dependency here.

PR: docs/90435
Submitted by: Dmitry Kazarov <kazarov at mcm dot ru>
Pointed out by: Daniel Gerzo <danger at rulez dot sk>
MFC after: 8 days

18 years agoUpdate manual page after last memguard changes.
pjd [Fri, 30 Dec 2005 12:28:19 +0000 (12:28 +0000)]
Update manual page after last memguard changes.

Glanced at by: simon

18 years agoAdd printproto.c to libipf
guido [Fri, 30 Dec 2005 11:56:57 +0000 (11:56 +0000)]
Add printproto.c to libipf

18 years agoAdd mcopywrap prototype to ip_compat.h
guido [Fri, 30 Dec 2005 11:55:37 +0000 (11:55 +0000)]
Add mcopywrap prototype to ip_compat.h
Remove h323 proxy from ip_proxy (copyright issue)

18 years agoResolve conflicts (and believe me...you don't want to know).
guido [Fri, 30 Dec 2005 11:52:26 +0000 (11:52 +0000)]
Resolve conflicts (and believe me...you don't want to know).

18 years agoImprove memguard a bit:
pjd [Fri, 30 Dec 2005 11:45:07 +0000 (11:45 +0000)]
Improve memguard a bit:
- Provide tunable vm.memguard.desc, so one can specify memory type without
  changing the code and recompiling the kernel.
- Allow to use memguard for kernel modules by providing sysctl
  vm.memguard.desc, which can be changed to short description of memory
  type before module is loaded.
- Move as much memguard code as possible to memguard.c.
- Add sysctl node vm.memguard. and move memguard-specific sysctl there.
- Add malloc_desc2type() function for finding memory type based on its
  short description (ks_shortdesc field).
- Memory type can be changed (via vm.memguard.desc sysctl) only if it
  doesn't exist (will be loaded later) or when no memory is allocated yet.
  If there is allocated memory for the given memory type, return EBUSY.
- Implement two ways of memory types comparsion and make safer/slower the
  default.

18 years agoThis commit was generated by cvs2svn to compensate for changes in r153877,
guido [Fri, 30 Dec 2005 11:34:54 +0000 (11:34 +0000)]
This commit was generated by cvs2svn to compensate for changes in r153877,
which included commits to RCS files with non-trunk default branches.

18 years agoImport IP Filter 4.1.10
guido [Fri, 30 Dec 2005 11:34:54 +0000 (11:34 +0000)]
Import IP Filter 4.1.10

18 years agoResolve conflicts
guido [Fri, 30 Dec 2005 11:32:23 +0000 (11:32 +0000)]
Resolve conflicts

18 years ago- Retire BARBIT in favor of new PCI_RID2BAR.
glebius [Fri, 30 Dec 2005 11:31:47 +0000 (11:31 +0000)]
- Retire BARBIT in favor of new PCI_RID2BAR.
- Fix build.

18 years agoThis commit was generated by cvs2svn to compensate for changes in r153872,
guido [Fri, 30 Dec 2005 11:22:11 +0000 (11:22 +0000)]
This commit was generated by cvs2svn to compensate for changes in r153872,
which included commits to RCS files with non-trunk default branches.

18 years agoImport IP Filter version 4.1.10
guido [Fri, 30 Dec 2005 11:22:11 +0000 (11:22 +0000)]
Import IP Filter version 4.1.10

18 years ago1. Add missing semicolon between "warn" and "return" to make sure
rse [Fri, 30 Dec 2005 09:16:23 +0000 (09:16 +0000)]
1. Add missing semicolon between "warn" and "return" to make sure
   the line continuation backslash doesn't cause "warn" to print "return".
2. Group "warn" and "return" together as the "return 1" should be
   performed only if the "kldload nfsclient" also failed (and not
   already if the "vfs.nfs" sysctl(8) check failed).

MFC after: 3 days

18 years agoRemove superfluous line continuation backslash.
rse [Fri, 30 Dec 2005 08:57:33 +0000 (08:57 +0000)]
Remove superfluous line continuation backslash.

MFC after: 3 days

18 years agoDisable frantic DMA update within few SNDCTL_DSP_* ioctl.
ariff [Fri, 30 Dec 2005 07:33:28 +0000 (07:33 +0000)]
Disable frantic DMA update within few SNDCTL_DSP_* ioctl.
This should reduce huge playback / recording latency for
applications that try to act smarter and manage their own
buffering (XMMS, Skype, etc.).

Note to Skype + via8xxx users: Remove previous hackish
"hint.pcm.<unit>.via_dxs_disabled" from kernel hint and see
whether this changes cure all those annoying sound issues.

18 years agoUnderruns counting logic should be based on bufhard free space
ariff [Fri, 30 Dec 2005 07:33:01 +0000 (07:33 +0000)]
Underruns counting logic should be based on bufhard free space
and must be done after sndbuf_feed(), or any attempt to fill
up bufhard. This should fix false underruns counter.

18 years agoDocument the LOGIN_SETMAC setusercontext(3) flag. While we are here, drop
csjp [Fri, 30 Dec 2005 06:16:20 +0000 (06:16 +0000)]
Document the LOGIN_SETMAC setusercontext(3) flag. While we are here, drop
in an external reference to mac_set_proc(3).

18 years agoAdd a bunch of ipod entries, one as suggested by Bryan Liesner. We
imp [Fri, 30 Dec 2005 05:06:26 +0000 (05:06 +0000)]
Add a bunch of ipod entries, one as suggested by Bryan Liesner.  We
should likely fix this better...

Submitted by: Bryan Liesner (ipod mini)

18 years agoFew codec such as Conexant CX20468-21 does have this control
ariff [Fri, 30 Dec 2005 01:06:29 +0000 (01:06 +0000)]
Few codec such as Conexant CX20468-21 does have this control
register, although the only usable part is the mute bit.

Noticed by: Hans Petter Selasky <hselasky@c2i.net>

18 years agocorrect checking for turbo channels: rev 1.24 fixed static turbo channels
sam [Fri, 30 Dec 2005 00:40:36 +0000 (00:40 +0000)]
correct checking for turbo channels: rev 1.24 fixed static turbo channels
but broke handling of the turboG channel; since we aren't ready to revamp
the channel list just check for turboA channels for now so channel 6 is
considered in auto mode

Noticed by: gibbs

18 years agoRetire BARBIT in favor of new PCI_RID2BAR.
imp [Thu, 29 Dec 2005 23:41:29 +0000 (23:41 +0000)]
Retire BARBIT in favor of new PCI_RID2BAR.

18 years agoAdd a macro to map from BAR to rid # (inverse of PCIR_BARS).
imp [Thu, 29 Dec 2005 23:40:41 +0000 (23:40 +0000)]
Add a macro to map from BAR to rid # (inverse of PCIR_BARS).

18 years agoSimplify the opening of the resources for cardbus cards. Before we'd
imp [Thu, 29 Dec 2005 23:38:45 +0000 (23:38 +0000)]
Simplify the opening of the resources for cardbus cards.  Before we'd
try very hard to be perfect.  However, these attempts broke down when
there were large numbers of resources.  We'd not be able to map them all.

Instead, accept that we might pass more range to thse subbus than
might be optimal be able to compute.  However, there's little harm in
this and it allows us to pass greater resources through.

# it has been suggested that we allocate a fixed amount of resources
# on attach and give it out upon request.  This might not be a bad idea...

18 years agoPrint a warning when we miss vinactive() call, because of race in vget().
pjd [Thu, 29 Dec 2005 22:52:09 +0000 (22:52 +0000)]
Print a warning when we miss vinactive() call, because of race in vget().
The race is very real, but conditions needed for triggering it are rather
hard to meet now.
When gjournal will be committed (where it is quite easy to trigger) we need
to fix it.

For now, verify if it is really hard to trigger.

Discussed with: kan

18 years agoThis is the style-fix for my previous commit. Sorry for the delay, I
cracauer [Thu, 29 Dec 2005 21:34:49 +0000 (21:34 +0000)]
This is the style-fix for my previous commit.  Sorry for the delay, I
forgot about it.

18 years agopatch(1) and I aren't friends today. Axe a duplicate copy of
jhb [Thu, 29 Dec 2005 21:15:32 +0000 (21:15 +0000)]
patch(1) and I aren't friends today.  Axe a duplicate copy of
the msleep_spin() function definition.

Spotted by: pjd

18 years agoAdd a new function msleep_spin() which is a slightly stripped down version
jhb [Thu, 29 Dec 2005 20:57:45 +0000 (20:57 +0000)]
Add a new function msleep_spin() which is a slightly stripped down version
of msleep().  msleep_spin() doesn't support changing the priority of the
thread while it is asleep nor does it support interruptible sleeps (PCATCH)
or the PDROP flag.  It does support timeouts however.  It differs from
msleep() in that the passed in mutex is a spin mutex.  This means one can
use msleep_spin() and wakeup() with a spin mutex similar to msleep() and
wakeup() with a regular mutex.  Note that the spin mutex in question needs
to come before sched_lock and the sleepq locks in lock order.

18 years agoTeach WITNESS_SAVE() and WITNESS_RESTORE() to work with spin locks instead
jhb [Thu, 29 Dec 2005 20:54:25 +0000 (20:54 +0000)]
Teach WITNESS_SAVE() and WITNESS_RESTORE() to work with spin locks instead
of only sleep locks.

18 years agoFix a deadlock I introduced with the recently added printf to warn about
jhb [Thu, 29 Dec 2005 20:53:01 +0000 (20:53 +0000)]
Fix a deadlock I introduced with the recently added printf to warn about
spin locks that are not in the static order list.  It is not safe to call
printf while holding the witness spin mutex since the console drivers that
back printf may need to use their own spin locks which would try to talk
to witness when they were locked.  Given this, it is possible for one
CPU to lock a console driver lock (such as sio) which then tries to lock
the witness lock while another CPU is doing the printf while holding the
witness lock.  Fix this by moving the printf outside of the witness lock.
All other printf's in witness are already correct.

MFC after: 3 days

18 years agoReveal sacred spells necessary for putting kbdmux(4) to real use.
sobomax [Thu, 29 Dec 2005 20:11:46 +0000 (20:11 +0000)]
Reveal sacred spells necessary for putting kbdmux(4) to real use.

Hopefully gods of RTFS and RTFCL will not go after me.

18 years agoAn i386 machine with 32MB of RAM requires a 700MB disk for auto-layout to
ceri [Thu, 29 Dec 2005 19:25:17 +0000 (19:25 +0000)]
An i386 machine with 32MB of RAM requires a 700MB disk for auto-layout to
succeed, whilst an i386 with 256MB of RAM requires more.  Up the amount
of disk hinted at as a minimum requisite for auto-layout.

MFC after: 1 week

18 years agocorrect (reversed) ms<->tu macros
sam [Thu, 29 Dec 2005 19:03:45 +0000 (19:03 +0000)]
correct (reversed) ms<->tu macros

Reviewed by: apatti, kcyu

18 years agoSync the type (and size, compare mousestatus_t in /usr/include/sys/mouse.h)
netchild [Thu, 29 Dec 2005 18:35:28 +0000 (18:35 +0000)]
Sync the type (and size, compare mousestatus_t in /usr/include/sys/mouse.h)
of a variable according to the usage (after increasing the number of max
buttons this may matter).

Noticed by: flz

18 years agoAdd support for the HP 8200C/8250C/8290C scanners.
netchild [Thu, 29 Dec 2005 18:25:25 +0000 (18:25 +0000)]
Add support for the HP 8200C/8250C/8290C scanners.

PR: 90467
Submitted by: Adam McDougall <mcdouga9@egr.msu.edu>

18 years agoAdd support for the Canon CanoSvan LIDE 25.
netchild [Thu, 29 Dec 2005 18:15:03 +0000 (18:15 +0000)]
Add support for the Canon CanoSvan LIDE 25.

PR: 89509
Submitted by: David Gilbert <dgilbert@daveg.ca>

18 years agoFix some kind of "off by one"-error: the min or max sample rate the
netchild [Thu, 29 Dec 2005 18:11:11 +0000 (18:11 +0000)]
Fix some kind of "off by one"-error: the min or max sample rate the
device is able to reproduce should be usable too instead of failing
in such a case.

PR: 89269
Submitted by: Don L. Belcher <don@siad.net>

18 years agoIncrement kobj_lookup_misses on a miss rather than decrementing it.
jhb [Thu, 29 Dec 2005 18:00:42 +0000 (18:00 +0000)]
Increment kobj_lookup_misses on a miss rather than decrementing it.
Otherwise, the miss count is actually -kobj_lookup_misses.  Mostly a
pedantic change as KOBJ_STATS isn't on by default.

18 years agoDocument Epson Perfection 2480 support.
brueffer [Thu, 29 Dec 2005 17:56:56 +0000 (17:56 +0000)]
Document Epson Perfection 2480 support.

18 years agoMerge NetBSD rev. 1.61:
netchild [Thu, 29 Dec 2005 17:44:40 +0000 (17:44 +0000)]
Merge NetBSD rev. 1.61:
 - Support more than 7 buttons for USB mice. Patch from NetBSD kern/30248.

PR: 83353
Submitted by: Seán Farley <sean-freebsd@farley.org>

18 years agoAdd support for Epson 2480 scanner.
netchild [Thu, 29 Dec 2005 17:38:03 +0000 (17:38 +0000)]
Add support for Epson 2480 scanner.

PR: 86094
Submitted by: Erik Norgaard <norgaard@locolomo.org>

18 years agoEmit USB_EVENT_DRIVER_DETACH on detach.
netchild [Thu, 29 Dec 2005 17:29:18 +0000 (17:29 +0000)]
Emit USB_EVENT_DRIVER_DETACH on detach.

PR: 83247
Submitted by: Sangwoo Shim <ssw@neo.redjade.org>
MFC after: 2 weeks

18 years agoAdd a new extensible GSS-API layer which can support GSS-API plugins,
dfr [Thu, 29 Dec 2005 14:40:22 +0000 (14:40 +0000)]
Add a new extensible GSS-API layer which can support GSS-API plugins,
similar the the Solaris implementation. Repackage the krb5 GSS mechanism
as a plugin library for the new implementation. This also includes a
comprehensive set of manpages for the GSS-API functions with text mostly
taken from the RFC.

Reviewed by: Love Hörnquist Åstrand <lha@it.su.se>, ru (build system), des (openssh parts)

18 years agoFix semantics in the documentation of the "jail_socket_unixiproute_only"
rse [Thu, 29 Dec 2005 13:39:46 +0000 (13:39 +0000)]
Fix semantics in the documentation of the "jail_socket_unixiproute_only"
variable ("YES" vs. "NO") and be more precise and intuitive on what is
actually allowed/restricted by this variable ("UNIX/IP/route sockets"
vs. "TCP/IP protocols").

MFC after: 3 days

18 years agoRemove pcb_switchout, it has not been used for a long time.
davidxu [Thu, 29 Dec 2005 13:23:48 +0000 (13:23 +0000)]
Remove pcb_switchout, it has not been used for a long time.

18 years agoAdd code to report zombie state.
davidxu [Thu, 29 Dec 2005 13:00:42 +0000 (13:00 +0000)]
Add code to report zombie state.

PR: threads/91044
MFC after: 3 days

18 years agoHelp Warner with merge from p4.
glebius [Thu, 29 Dec 2005 10:38:42 +0000 (10:38 +0000)]
Help Warner with merge from p4.

18 years agoWhen pfil(9) is enabled the bridge only considers ETHERTYPE_ARP, ETHERTYPE_IP and
thompsa [Thu, 29 Dec 2005 09:39:15 +0000 (09:39 +0000)]
When pfil(9) is enabled the bridge only considers ETHERTYPE_ARP, ETHERTYPE_IP and
ETHERTYPE_IPV6 frames. Change this to be a sysctl knob so that is able to still
bridge non-IP packets if desired.

Also return early if all pfil_* sysctls are turned off, the user obviously does
not want to filter on the bridge.

18 years agoMinor updates relative to the 9.2.3 import
dougb [Thu, 29 Dec 2005 04:41:09 +0000 (04:41 +0000)]
Minor updates relative to the 9.2.3 import

18 years agoBuild gprof for powerpc now.
grehan [Thu, 29 Dec 2005 04:30:03 +0000 (04:30 +0000)]
Build gprof for powerpc now.

18 years agoUpdated versions of header files generated per the instructions
dougb [Thu, 29 Dec 2005 04:29:03 +0000 (04:29 +0000)]
Updated versions of header files generated per the instructions
in src/contrib/bind9/FREEBSD-Upgrade for the 9.2.3 import

18 years agoRemove files no longer in the BIND 9 distribution
dougb [Thu, 29 Dec 2005 04:26:13 +0000 (04:26 +0000)]
Remove files no longer in the BIND 9 distribution

18 years agoclone i386.h for powerpc
grehan [Thu, 29 Dec 2005 04:23:54 +0000 (04:23 +0000)]
clone i386.h for powerpc

18 years agoThis commit was generated by cvs2svn to compensate for changes in r153816,
dougb [Thu, 29 Dec 2005 04:22:58 +0000 (04:22 +0000)]
This commit was generated by cvs2svn to compensate for changes in r153816,
which included commits to RCS files with non-trunk default branches.

18 years agoVendor import of BIND 9.3.2
dougb [Thu, 29 Dec 2005 04:22:58 +0000 (04:22 +0000)]
Vendor import of BIND 9.3.2

18 years agogmon now supported on powerpc
grehan [Thu, 29 Dec 2005 04:10:52 +0000 (04:10 +0000)]
gmon now supported on powerpc

18 years agoThe minbrk symbol is hidden the same on powerpc as other FreeBSD platforms.
grehan [Thu, 29 Dec 2005 04:09:38 +0000 (04:09 +0000)]
The minbrk symbol is hidden the same on powerpc as other FreeBSD platforms.

18 years agoAdd user-space profiling support. Kernel profiling still todo.
grehan [Thu, 29 Dec 2005 04:07:36 +0000 (04:07 +0000)]
Add user-space profiling support. Kernel profiling still todo.

Obtained from: NetBSD

18 years agoRefer readers to the hwpmc(4) manual page where the additional
jkoshy [Thu, 29 Dec 2005 02:12:54 +0000 (02:12 +0000)]
Refer readers to the hwpmc(4) manual page where the additional
(architecture specific) kernel configuration options needed for
hwpmc are documented.

PR: kern/83738
MFC after: 3 days

18 years agoImplement /dev/cardbus%d.cis, same thing as /dev/pccard%d.cis. There
imp [Thu, 29 Dec 2005 01:43:47 +0000 (01:43 +0000)]
Implement /dev/cardbus%d.cis, same thing as /dev/pccard%d.cis.  There
are some rough edges with this still, but it seems to work well enough
to commit.

18 years agoInitialize the event tailq.
scottl [Wed, 28 Dec 2005 21:18:55 +0000 (21:18 +0000)]
Initialize the event tailq.

Submitted by: Frank Mayhar
PR: kern/90882
MFC After: 1 day

18 years agoAdd support for printing IPSEC protocol stats if the kernel was compiled
kbyanc [Wed, 28 Dec 2005 20:36:55 +0000 (20:36 +0000)]
Add support for printing IPSEC protocol stats if the kernel was compiled
with FAST_IPSEC rather than the KAME IPSEC stack.

Note that the output of "netstat -s -p ipsec" differs depending on which
stack is compiled into the kernel since they each keep different stats.
This delta also adds the "esp", "ah", and "ipcomp" protocol stats, which
are also available when the kernel is compiled with the FAST_IPSEC stack
(e.g. "netstat -s -p esp").

Submitted by: Matt Titus <titus at nttmcl dot com>
MFC after: 3 days

18 years agoAdd missing parens.
glebius [Wed, 28 Dec 2005 18:00:37 +0000 (18:00 +0000)]
Add missing parens.

Submitted by: njl

18 years agoFix the order of the stereo channels (left <-> right).
netchild [Wed, 28 Dec 2005 17:57:36 +0000 (17:57 +0000)]
Fix the order of the stereo channels (left <-> right).

From the PR:
---snip---
 I think I have found the change which reversed the channels.
 Revision 1.44 of emu10k1.c, which added Audigy support, has the line

 emu_wrptr(sc, v->vnum, FXRT, 0xd01c0000);

 replaced with the following lines:

 if (sc->audigy) {
         emu_wrptr(sc, v->vnum, A_FXRT1, v->fxrt1);
         emu_wrptr(sc, v->vnum, A_FXRT2, v->fxrt2);
         emu_wrptr(sc, v->vnum, A_SENDAMOUNTS, 0);
 }
 else
         emu_wrptr(sc, v->vnum, FXRT, v->fxrt1 << 16);

 where v->fxrt1 << 16 == 0xd10c0000

 I don't have Audigy, so I'm not sure if the problem affects Audigy cards
 too. The order of the channels can't be tested by just altering mixer
 settings. Here's a small program to test if the channels are reversed on
 your sound card:

 #include <sys/soundcard.h>
 #include <fcntl.h>
 #include <unistd.h>

 int main(int argc, char **argv)
 {
  int fd = open("/dev/dsp", O_WRONLY), format = AFMT_S16_LE;
   int channels = 2, rate = 22050, i;

   /* 450 Hz sine wave on left channel, right channel silent */
   unsigned char samples[] = {0, 0, 0, 0, 94, 16, 0, 0, 120, 32, 0, 0,
     9, 48, 0, 0, 208, 62, 0, 0, 143, 76, 0, 0, 12, 89, 0, 0, 19, 100,
     0, 0, 117, 109, 0, 0, 11, 117, 0, 0, 182, 122, 0, 0, 92, 126, 0,
     0, 239, 127, 0, 0, 105, 127, 0, 0, 202, 124, 0, 0, 32, 120, 0, 0,
     124, 113, 0, 0, 251, 104, 0, 0, 193, 94, 0, 0, 249, 82, 0, 0,
     212, 69, 0, 0, 138, 55, 0, 0, 85, 40, 0, 0, 120, 24, 0, 0, 51, 8,
     0, 0, 205, 247, 0, 0, 136, 231, 0, 0, 171, 215, 0, 0, 118, 200,
     0, 0, 44, 186, 0, 0, 7, 173, 0, 0, 63, 161, 0, 0, 5, 151, 0, 0,
     132, 142, 0, 0, 224, 135, 0, 0, 54, 131, 0, 0, 151, 128, 0, 0,
     17, 128, 0, 0, 164, 129, 0, 0, 74, 133, 0, 0, 245, 138, 0, 0,
     139, 146, 0, 0, 237, 155, 0, 0, 244, 166, 0, 0, 113, 179, 0, 0,
     48, 193, 0, 0, 247, 207, 0, 0, 136, 223, 0, 0, 162, 239, 0, 0};

   ioctl(fd, SNDCTL_DSP_SETFMT,&format);
   ioctl(fd, SNDCTL_DSP_CHANNELS,&channels);
   ioctl(fd, SNDCTL_DSP_SPEED,&rate);

   for(i=0;i<500;i++)
     write(fd, &samples, sizeof(samples));
   write(fd, &samples, 2); /* swap channels */
   for(i=0;i<500;i++)
     write(fd, &samples, sizeof(samples));

   return 0;
 }

 You should hear a sound on the left channel followed by a sound on the
 right channel. If you hear a sound on the right channel first, the
 channels are reversed.
---snip---

Owners of an audigy card should verify if it DTRT and report back.

Noticed by: Matthias Buelow <mkb@mukappabeta.de>
Submitted by: Juha-Matti Tilli <juhis@nallukka.net>
PR: 72221

18 years agoTrim trailing whitespace.
kan [Wed, 28 Dec 2005 17:13:31 +0000 (17:13 +0000)]
Trim trailing whitespace.

18 years agoo Fix typos in the comments.
maxim [Wed, 28 Dec 2005 15:18:18 +0000 (15:18 +0000)]
o Fix typos in the comments.

Submitted by: Wojciech A. Koszek

18 years agoCheck for IFF_DRV_RUNNING in the interrupt loop.
glebius [Wed, 28 Dec 2005 13:21:05 +0000 (13:21 +0000)]
Check for IFF_DRV_RUNNING in the interrupt loop.

Reported & tested by: Martin P. Hansen <mph lima.dyndns.dk>

18 years agoIn ng_netflow_disconnect() check whether we are working with "iface"
glebius [Wed, 28 Dec 2005 12:56:59 +0000 (12:56 +0000)]
In ng_netflow_disconnect() check whether we are working with "iface"
or with "out" hook, and clear the right pointer.

Reported by: Vitaliy Ovsyannikov <V.Ovsyannikov kr.ru>

18 years agoAdd 'juggle', a simple IPC micro-benchmark that uses various IPC types to
rwatson [Wed, 28 Dec 2005 12:42:12 +0000 (12:42 +0000)]
Add 'juggle', a simple IPC micro-benchmark that uses various IPC types to
perform varying number of small IPC operations.  It runs using a single
process and one thread, a single process and two threads, and using
multiple processes.  Critical to its performance measure are the cost and
frequency of context switches, locking overhead, and threading
performance.  The benchmark generates .csv output appropriate for reading
into a spreadsheet to generate summary statistics and perform statistical
tests easily.

18 years agoFurther updates to the Marvell support code.
sos [Wed, 28 Dec 2005 11:55:43 +0000 (11:55 +0000)]
Further updates to the Marvell support code.
Should be reliable enough for actual use by now.

18 years ago When in rev. 1.47 cardbus_alloc_resources() function was moved from
glebius [Wed, 28 Dec 2005 10:15:01 +0000 (10:15 +0000)]
  When in rev. 1.47 cardbus_alloc_resources() function was moved from
cardbus_cis.c to this file, some code was not merged and thus resource
list entries were invalid. They didn't have a resources attached to
them.
  However, the problem was masked for some time later, because newer
resources list entries were added to the head of the list, and
resource_list_find() always returned the first matching resource list
entry. Usually the underlying driver allocated a valid resource and
added it to the head of the list, and invalid one wasn't used.
  In rev. 1.174 of subr_bus.c the sorting of resource list entries was
reversed demasking the problem in cardbus_alloc_resources().
  This commit fixes the problem returning back some code from
cardbus_cis.c, pre-1.49 revisions.

PR: kern/87114
PR: kern/90441
Hardware provided by: Vasily Olekhov <olekhov yandex.ru>
Reviewed by: imp

18 years agoCorrect a typo
delphij [Wed, 28 Dec 2005 10:03:48 +0000 (10:03 +0000)]
Correct a typo

18 years agoA style nit.
glebius [Wed, 28 Dec 2005 09:37:04 +0000 (09:37 +0000)]
A style nit.

18 years agoDocument new support for Marvell chipsets.
brueffer [Wed, 28 Dec 2005 09:35:29 +0000 (09:35 +0000)]
Document new support for Marvell chipsets.