openbsd
16 years agoDocument the new timeout_add_* functions.
art [Wed, 23 Jul 2008 11:45:21 +0000 (11:45 +0000)]
Document the new timeout_add_* functions.

16 years agovalidate packet length in debug dns packet logging before printing the header.
reyk [Wed, 23 Jul 2008 10:05:18 +0000 (10:05 +0000)]
validate packet length in debug dns packet logging before printing the header.

16 years agosync
mglocker [Wed, 23 Jul 2008 07:56:45 +0000 (07:56 +0000)]
sync

16 years agoUnbreak the tree.
mglocker [Wed, 23 Jul 2008 07:55:31 +0000 (07:55 +0000)]
Unbreak the tree.

16 years agodo not try to print options that have been compile-time disabled
djm [Wed, 23 Jul 2008 07:36:55 +0000 (07:36 +0000)]
do not try to print options that have been compile-time disabled
in config test mode (sshd -T); report from nix-corp AT esperi.org.uk
ok dtucker@

16 years agosync
deraadt [Wed, 23 Jul 2008 04:59:27 +0000 (04:59 +0000)]
sync

16 years agonew ati goo; giovanni
deraadt [Wed, 23 Jul 2008 04:59:18 +0000 (04:59 +0000)]
new ati goo; giovanni

16 years agomove the aml_register_notify() call below for better debugging output.
fgsch [Wed, 23 Jul 2008 00:20:35 +0000 (00:20 +0000)]
move the aml_register_notify() call below for better debugging output.
marco@ ok.

16 years agoAdd dynamic IPv6-to-IPv4 and IPv4-to-IPv6 translation inspired by
reyk [Tue, 22 Jul 2008 23:17:37 +0000 (23:17 +0000)]
Add dynamic IPv6-to-IPv4 and IPv4-to-IPv6 translation inspired by
faithd(8) by doing a similar mapping of IPv4/6 addresses with
relayd(8) and pf(4) redirections without the need of the faith(4)
interface.  The trick works in both directions, it can accept IPv6
connections and relay them to IPv4 hosts by extracting the last 4
octets from the IPv6 destination (like faithd(8)), and it can accept
IPv4 connections and relay them to IPv6 hosts by prepending the 4
octets of the original IPv4 destination to a configured IPv6 prefix.
An access list is not needed because the classification is done in
pf.conf(5).  It helps to get more faith in relayd.

manpage bits ok jmc@
yes, sounds good todd@

16 years agouse arc4random_uniform(); ok djm millert
deraadt [Tue, 22 Jul 2008 21:47:45 +0000 (21:47 +0000)]
use arc4random_uniform(); ok djm millert

16 years agoFAKE (the variable) died about 2 years ago...
espie [Tue, 22 Jul 2008 20:53:13 +0000 (20:53 +0000)]
FAKE (the variable) died about 2 years ago...

16 years agoNone of the pceb EISA alpha designs has more than 8 physical slots, so don't
miod [Tue, 22 Jul 2008 20:06:01 +0000 (20:06 +0000)]
None of the pceb EISA alpha designs has more than 8 physical slots, so don't
bother trying to probe more.

16 years agoremez -> Remes here, too. for consistency
martynas [Tue, 22 Jul 2008 19:58:40 +0000 (19:58 +0000)]
remez -> Remes here, too.  for consistency

16 years agoAdd another working cam reported by Jerome Pinot.
mglocker [Tue, 22 Jul 2008 19:45:51 +0000 (19:45 +0000)]
Add another working cam reported by Jerome Pinot.

16 years agoProper interrupt swizzling for devices behind bridges; matches what SRM
miod [Tue, 22 Jul 2008 18:47:24 +0000 (18:47 +0000)]
Proper interrupt swizzling for devices behind bridges; matches what SRM
configures when it can. ok kettenis@

16 years agoModify pci_intr_map() implementation to pass the pci_attach_args to the
miod [Tue, 22 Jul 2008 18:45:50 +0000 (18:45 +0000)]
Modify pci_intr_map() implementation to pass the pci_attach_args to the
per-platform implementation, instead of selected members of it; this allows
us to get rid of some globals, and paves the way for better bridge support
on some models.

16 years agoUse cgbase() instead of doing arithmetic on fs_fpg, for the former does
miod [Tue, 22 Jul 2008 18:44:25 +0000 (18:44 +0000)]
Use cgbase() instead of doing arithmetic on fs_fpg, for the former does
proper casts to ensure the result is not truncated.
From FreeBSD via NetBSD. ok thib@

16 years agoDo not define VM_MIN_ADDRESS as __LDPGSZ, but provide its value, to prevent
miod [Tue, 22 Jul 2008 18:15:48 +0000 (18:15 +0000)]
Do not define VM_MIN_ADDRESS as __LDPGSZ, but provide its value, to prevent
uvm_map.c from requiring <machine/exec.h>.

16 years agoThere are devices out there which are UVC compatible, but do not show
mglocker [Tue, 22 Jul 2008 16:24:40 +0000 (16:24 +0000)]
There are devices out there which are UVC compatible, but do not show
up as this (we love standards).  To make those devices attach anyway
we add an explicit quirk device list.  Of course this list still needs
to be extended.

Thanks to Jerome Pinot for reporting about this issue and testing this
diff with his cam, which is such a device.

OK yuo@

16 years ago- added HUGE_VALF, HUGE_VALL, INFINITY, FP_INFINITE, FP_NAN,
martynas [Tue, 22 Jul 2008 16:01:46 +0000 (16:01 +0000)]
- added HUGE_VALF, HUGE_VALL, INFINITY, FP_INFINITE, FP_NAN,
FP_NORMAL, FP_SUBNORMAL, FP_ZERO, FP_ILOGB0, FP_ILOGBNAN definitions,
per C99
- classify functions into sections, add missing definitions, add
and some '#if 0' functions, and long double definitions, so we see
what's needed to be done
- other stuff is XXX'd, and will be uncommented soon
ok millert@

16 years agoafter pf_state_key_atach nothing must use the state keys passed to it any
henning [Tue, 22 Jul 2008 12:31:35 +0000 (12:31 +0000)]
after pf_state_key_atach nothing must use the state keys passed to it any
more, since they might have been invalidated and free'd.
one synproxy errorpath did so, however. just get the state keys from the
state itself. ok david mcbride

16 years agomore negotation -> negotiation; ok sthen@
martynas [Tue, 22 Jul 2008 11:20:10 +0000 (11:20 +0000)]
more negotation -> negotiation;  ok sthen@

16 years agoSpelling, negotation -> negotiation.
mglocker [Tue, 22 Jul 2008 11:02:17 +0000 (11:02 +0000)]
Spelling, negotation -> negotiation.

16 years agoFix uvm_fault crash in uvideo_vs_negotation() which was reported by some
mglocker [Tue, 22 Jul 2008 10:57:15 +0000 (10:57 +0000)]
Fix uvm_fault crash in uvideo_vs_negotation() which was reported by some
users.

Thanks to mpf@ for his help to debug this bug down.

16 years agoIf isakmpd is started with -4 or -6, virtual_get_default() may
bluhm [Tue, 22 Jul 2008 09:45:56 +0000 (09:45 +0000)]
If isakmpd is started with -4 or -6, virtual_get_default() may
return NULL.  This happens if isakmpd is configured for the other
address family.  Add a NULL pointer check and initialize rv.
ok hshoexer

16 years agoAssign the struct size to sin6_len instead of sin6_family.
bluhm [Tue, 22 Jul 2008 09:26:39 +0000 (09:26 +0000)]
Assign the struct size to sin6_len instead of sin6_family.
ok hshoexer claudio mpf henning

16 years agochange the bcachestats structures members to int64_t's and do the
thib [Tue, 22 Jul 2008 08:05:02 +0000 (08:05 +0000)]
change the bcachestats structures members to int64_t's and do the
printing in systat with the new print_fld_ssize() function.

"same diff I wrote" deraadt@
ok art@

16 years agouse generalised PBKDF#2 from bioctl; ok damien@
djm [Tue, 22 Jul 2008 07:37:25 +0000 (07:37 +0000)]
use generalised PBKDF#2 from bioctl; ok damien@

16 years agodeactivate the code to display location in the sensor description
mbalmer [Tue, 22 Jul 2008 06:06:47 +0000 (06:06 +0000)]
deactivate the code to display location in the sensor description

16 years agosync
deraadt [Tue, 22 Jul 2008 03:09:17 +0000 (03:09 +0000)]
sync

16 years agoAdd signed size printing functions, as requested by deraadt@.
canacar [Tue, 22 Jul 2008 03:00:23 +0000 (03:00 +0000)]
Add signed size printing functions, as requested by deraadt@.
Just use print_fld_ssize() for signed values.
Increase scale up to "Tera" while there.

16 years agoimplement the fetching of a scsi devices "devid". recent hardware provides
dlg [Tue, 22 Jul 2008 01:01:31 +0000 (01:01 +0000)]
implement the fetching of a scsi devices "devid". recent hardware provides
a vpd page that uniquely identifies a device no matter what bus topology or
addressing was used to find it.

we have a workaround for old school scsi devices that do not differentiate
between luns. if the inq data for high luns is the same as the inq data
for lun 0, we assume it is one of these buggy devices.

the problem with this is that things like SANs present multiple
volumes as luns and they all have the same inq data. if you wanted
to present more than one volume to openbsd you would only ever see
the first one.

devices give us a mechanism to differentiate between luns, so now
i do get all my volumes attached in openbsde.

review and feedback by krw@ marco@ testing by todd@

16 years agotweak comment to reflect the new reality after my last change.
dlg [Tue, 22 Jul 2008 00:40:37 +0000 (00:40 +0000)]
tweak comment to reflect the new reality after my last change.

16 years agowhen probing a device the midlayer queries its inquiry data and
dlg [Mon, 21 Jul 2008 23:40:03 +0000 (23:40 +0000)]
when probing a device the midlayer queries its inquiry data and
keeps it on the stack till we attach a driver to it. then it copies
the inquiry data int the scsi_link struct.

this diff uses the scsi_link struct instead of the stack for that data,
which makes the inq data for users of the scsi_link struct available much
earlier during device probe.

review and feedback from both krw@ and marco@

16 years agovaxfp.h is VAX pendant of ieee.h. it defines VAX f fp, and VAX d
martynas [Mon, 21 Jul 2008 21:50:06 +0000 (21:50 +0000)]
vaxfp.h is VAX pendant of ieee.h.  it defines VAX f fp, and VAX d
fp;  from NetBSD
ok millert@

16 years agospelling fix from Mark Pecaut;
jmc [Mon, 21 Jul 2008 21:37:30 +0000 (21:37 +0000)]
spelling fix from Mark Pecaut;

16 years ago- add proper double_t and float_t definitions for each arch
martynas [Mon, 21 Jul 2008 20:50:54 +0000 (20:50 +0000)]
- add proper double_t and float_t definitions for each arch
- math.h shouldn't define FLT_EVAL_METHOD, but float.h should (per
C99).  remove from math.h, and add proper definitions in float.h
ok millert@

16 years agoNetBSD's C library supplies these functions -> OpenBSD's C library
martynas [Mon, 21 Jul 2008 20:30:35 +0000 (20:30 +0000)]
NetBSD's C library supplies these functions -> OpenBSD's C library
supplies these functions.  in comments

16 years agorename lround.c lroundf.c llround.c llroundf.c to s_lround.c
martynas [Mon, 21 Jul 2008 20:29:14 +0000 (20:29 +0000)]
rename lround.c lroundf.c llround.c llroundf.c to s_lround.c
s_lroundf.c s_llround.c s_llroundf.c, for naming consistency
looks fine to millert@

16 years agogives this a chance to work on architectures with strict alignment
damien [Mon, 21 Jul 2008 19:52:45 +0000 (19:52 +0000)]
gives this a chance to work on architectures with strict alignment
constraints.

16 years agosync
damien [Mon, 21 Jul 2008 19:49:06 +0000 (19:49 +0000)]
sync

16 years agorename rt2860 pci ids to match with reality.
damien [Mon, 21 Jul 2008 19:41:44 +0000 (19:41 +0000)]
rename rt2860 pci ids to match with reality.
import some fixes for 1T2R adapters from the linux vendor driver.

16 years agoadd ieee80211_priv.h file: contains definitions private to net80211.
damien [Mon, 21 Jul 2008 19:27:26 +0000 (19:27 +0000)]
add ieee80211_priv.h file: contains definitions private to net80211.
this must not be included by drivers.

16 years agofix ugly debug dump output. found with bluhm@
rainer [Mon, 21 Jul 2008 19:14:15 +0000 (19:14 +0000)]
fix ugly debug dump output. found with bluhm@

ok thib@, bluhm@

16 years agomove processing of EAPOL frames away from ieee80211_{input,output}.c
damien [Mon, 21 Jul 2008 19:05:21 +0000 (19:05 +0000)]
move processing of EAPOL frames away from ieee80211_{input,output}.c

16 years agoinstead of passing rx tstamp and rssi to the ieee80211_input function,
damien [Mon, 21 Jul 2008 18:43:18 +0000 (18:43 +0000)]
instead of passing rx tstamp and rssi to the ieee80211_input function,
pass a pointer to an ieee80211_rxinfo structure containing those two
fields plus an extra flags field that indicates whether the frame was
decrypted by hardware or not.
required for a future fix.

16 years agoExtra sanity checking for afree(); OK deraadt@ and pvalchev@
millert [Mon, 21 Jul 2008 17:30:08 +0000 (17:30 +0000)]
Extra sanity checking for afree();  OK deraadt@ and pvalchev@

16 years agoInitial stab at DHCPINFORM support. OK phessler@
millert [Mon, 21 Jul 2008 16:51:18 +0000 (16:51 +0000)]
Initial stab at DHCPINFORM support.  OK phessler@

16 years agosome whitespace cleanup I did while looking through the code
david [Mon, 21 Jul 2008 15:58:59 +0000 (15:58 +0000)]
some whitespace cleanup I did while looking through the code
ok mcbride@ henning@

16 years agofix typo that broke rdr rules (without pass) with non-TCP/UDP/ICMP protocols
david [Mon, 21 Jul 2008 15:56:55 +0000 (15:56 +0000)]
fix typo that broke rdr rules (without pass) with non-TCP/UDP/ICMP protocols
ok mcbride@ henning@

16 years agoAdd curproc to ioctl call.
marco [Mon, 21 Jul 2008 14:49:40 +0000 (14:49 +0000)]
Add curproc to ioctl call.

Noticed by todd,  proded by miod.

16 years agoFree the rules in the rule_queue also if ipsecctl is called with
bluhm [Mon, 21 Jul 2008 14:37:53 +0000 (14:37 +0000)]
Free the rules in the rule_queue also if ipsecctl is called with
the -n switch.  This triggers malloc related bugs during the regress
tests.
ok hshoexer

16 years agoImplement the cpu_yield hypervisor call. Use it in the idle loop for
art [Mon, 21 Jul 2008 13:30:04 +0000 (13:30 +0000)]
Implement the cpu_yield hypervisor call. Use it in the idle loop for
SUN4V to let it suspend strands (why does everyone invent own words for
hyperthreads?). This gives a huge performance boost when most of the
cpus are idle.

kettenis@ ok

16 years agoIPL_NONE doesn't really exist. The corresponding bit in the
kettenis [Mon, 21 Jul 2008 10:07:14 +0000 (10:07 +0000)]
IPL_NONE doesn't really exist.  The corresponding bit in the
per-processor soft interrupt register is used for hardware clock-tick
counter interrupts.  So make smp_signotify() use IPL_SOFTINT instead
such that we don't unintentionally trigger a spurious clock interrupt.
This seems to fix the timekeeping anomalies on the t1k.

tested by art@

16 years agoLeave PATH alone.
ajacoutot [Mon, 21 Jul 2008 08:28:55 +0000 (08:28 +0000)]
Leave PATH alone.

noticed by and ok deraadt@

16 years agoopenssh-5.1
djm [Mon, 21 Jul 2008 08:19:07 +0000 (08:19 +0000)]
openssh-5.1

16 years agoAdd acphy/amphy.
brad [Mon, 21 Jul 2008 07:41:37 +0000 (07:41 +0000)]
Add acphy/amphy.

16 years agonam2blk[] needs a "vnd" entry for some things (like softraid)
todd [Mon, 21 Jul 2008 04:35:53 +0000 (04:35 +0000)]
nam2blk[] needs a "vnd" entry for some things (like softraid)
ok marco@ no objection miod@ need this for regress djm@ no objection krw@

16 years agocleanup - we don't have resource structure
kevlo [Mon, 21 Jul 2008 04:12:21 +0000 (04:12 +0000)]
cleanup - we don't have resource structure
ok jsg@

16 years agono, ipmi is still not fine on some machines (since acpi got enabled)
deraadt [Sun, 20 Jul 2008 22:17:50 +0000 (22:17 +0000)]
no, ipmi is still not fine on some machines (since acpi got enabled)

16 years agofill in variant and version components of the UUID
djm [Sun, 20 Jul 2008 21:57:51 +0000 (21:57 +0000)]
fill in variant and version components of the UUID
pretty-print the UUID in RFC4122 form

16 years agosync
deraadt [Sun, 20 Jul 2008 21:12:53 +0000 (21:12 +0000)]
sync

16 years agosync
deraadt [Sun, 20 Jul 2008 20:32:01 +0000 (20:32 +0000)]
sync

16 years agoOn TI cardbus bridges, enable PCI interrupt routing if the device has not
kettenis [Sun, 20 Jul 2008 18:58:07 +0000 (18:58 +0000)]
On TI cardbus bridges, enable PCI interrupt routing if the device has not
been initialized by the firmware.  This was probably the intention of the
origional code, however, it used the wrong default value for the register
in question.

Add TI PCI1510 to the list of 12XX-compatible bridges.

"be brave" deraadt@

16 years agoHuawei E618 and E620 is not compatible with ubsa(4).
yuo [Sun, 20 Jul 2008 14:24:49 +0000 (14:24 +0000)]
Huawei E618 and E620 is not compatible with ubsa(4).
This patch moves these device to umsm(4).

reported by Aleksander Piotrowski, ok jsg@

16 years agoNuke unused readbsdlabel() function.
krw [Sun, 20 Jul 2008 13:46:16 +0000 (13:46 +0000)]
Nuke unused readbsdlabel() function.

ok jsing@

16 years agoUse NULL instead of 0 for pointer parameters.
krw [Sun, 20 Jul 2008 13:40:27 +0000 (13:40 +0000)]
Use NULL instead of 0 for pointer parameters.

ok jsing@

16 years agoReme algorithm -> Remes algorithm. in comments
martynas [Sun, 20 Jul 2008 13:11:00 +0000 (13:11 +0000)]
Reme algorithm -> Remes algorithm.  in comments

16 years agoget in the bug fix for pow, from fdlibm 5.3. also adapt it for
martynas [Sun, 20 Jul 2008 13:07:27 +0000 (13:07 +0000)]
get in the bug fix for pow, from fdlibm 5.3.  also adapt it for
powf, the float version of it
>     1. e_pow.c incorrect results when
>         x is very close to -1.0 and y is very large, e.g.
>         pow(-1.0000000000000002e+00,4.5035996273704970e+15) = 0
>         pow(-9.9999999999999978e-01,4.5035996273704970e+15) = 0
>         Correct results are close to -e and -1/e.
while here merge more changes for pow, fixes sign in oflow/uflow
cases
ok millert@

16 years agouse more accurate algorithm for tan. from fdlibm 5.3:
martynas [Sun, 20 Jul 2008 13:03:13 +0000 (13:03 +0000)]
use more accurate algorithm for tan.  from fdlibm 5.3:
>     2. k_tan.c error was > 1 ulp target for FDLIBM
>         5.2: Worst error at least 1.45 ulp at
>         tan(1.7765241907548024E+269) = 1.7733884462610958E+16
>         5.3: Worst error 0.96 ulp
ok millert@

16 years agoProvide a timecounter based on the UltraSPARC IIe STICK logic.
kettenis [Sun, 20 Jul 2008 10:37:43 +0000 (10:37 +0000)]
Provide a timecounter based on the UltraSPARC IIe STICK logic.

16 years agoAdd DIOCGPDINFO support. 'disklabel -d svnd0' now works.
krw [Sun, 20 Jul 2008 01:53:43 +0000 (01:53 +0000)]
Add DIOCGPDINFO support. 'disklabel -d svnd0' now works.

16 years agolink to apache web server, not the project group. found while
martynas [Sat, 19 Jul 2008 23:35:44 +0000 (23:35 +0000)]
link to apache web server, not the project group.  found while
catching up w/ the old email.  submitted by Pierre Riteau on tech@
a year ago.
"no objection" jmc@.  ok, "that's clearly correct" okan@

16 years agoFull rewrite of metadata handling. This fixes power failures and crashes
marco [Sat, 19 Jul 2008 22:41:58 +0000 (22:41 +0000)]
Full rewrite of metadata handling.  This fixes power failures and crashes
that caused illegal checksums.  The new metadata code is more or less ready
to deal with other vendor's metadata formats.

While here clean up the name space.

Fix thib's pool mess by adding removing bad flags in interrupt context.

tested on macppc, amd64, i386, sparc64 & hppa

sparc64 has issues with crypto however those do not seem to be softraid
specific.

help from okan@ ckuethe@ Will Backman and others

16 years agoUse errx instead of fprintf, exit combo.
claudio [Sat, 19 Jul 2008 21:31:39 +0000 (21:31 +0000)]
Use errx instead of fprintf, exit combo.
Diff from Gleydson Soares gsoares (at) gmail (dot) com
OK henning@

16 years agoIn dec_1000a_intr_map(), use the SRM hint if available. This allows cards
miod [Sat, 19 Jul 2008 19:25:18 +0000 (19:25 +0000)]
In dec_1000a_intr_map(), use the SRM hint if available. This allows cards
plugged to the bottom 4 PCI slots of AlphaServer 1000A (attaching to pci1
behind a ppb) to get interrupts.
No regressions on AlphaServer 800 (which do not have these extra slots).

16 years agoAdd code to walk the eisa configuration, from NetBSD. This configuration
miod [Sat, 19 Jul 2008 18:13:04 +0000 (18:13 +0000)]
Add code to walk the eisa configuration, from NetBSD. This configuration
is not used yet, but this seems to ``warm up'' the eisa chips so that
accesses to the eisa bus later do not cause machine checks.

16 years agoIn _io_map(), make sure the whole range fits in one of the io windows,
miod [Sat, 19 Jul 2008 17:10:03 +0000 (17:10 +0000)]
In _io_map(), make sure the whole range fits in one of the io windows,
instead of only the starting address. From NetBSD.

16 years agominor wording tweaks;
jmc [Sat, 19 Jul 2008 16:35:50 +0000 (16:35 +0000)]
minor wording tweaks;

16 years agoCorrect missed capitalization of MPLS.
brad [Sat, 19 Jul 2008 15:55:07 +0000 (15:55 +0000)]
Correct missed capitalization of MPLS.

16 years agoAdd acx(4) at cardbus(4).
kettenis [Sat, 19 Jul 2008 12:36:06 +0000 (12:36 +0000)]
Add acx(4) at cardbus(4).

16 years agoprint the host parent id if available
reyk [Sat, 19 Jul 2008 12:10:07 +0000 (12:10 +0000)]
print the host parent id if available

16 years agono need for using a TAILQ queue for the host children list, use a
reyk [Sat, 19 Jul 2008 11:38:54 +0000 (11:38 +0000)]
no need for using a TAILQ queue for the host children list, use a
singly-linked SLIST instead.  the only noticeable change is the
reversed order to notify the children but it does not really matter
here.  also only walk through the children host list if the host
itself is a potential parent.

16 years agoIf a resolution which has been requested over VIDEO_TRY_FMT or
mglocker [Sat, 19 Jul 2008 11:30:55 +0000 (11:30 +0000)]
If a resolution which has been requested over VIDEO_TRY_FMT or
VIDEO_S_FMT doesn't exactly match the devices available resolutions,
return the next best matching resolution which we have.

Makes some V4L2 apps happy when running them with the default resolution
(no options).

16 years agoIf the new 'parent' keyword is specified for a host in a table,
reyk [Sat, 19 Jul 2008 10:52:32 +0000 (10:52 +0000)]
If the new 'parent' keyword is specified for a host in a table,
inherit the state from another host with the specified Id; no
additional check will be for the inheriting host.  This helps in
scenarios with lots of IP aliases that all point to the same service
on the same host (like web hosting with many SSL domains).

discussed with pyr, tested in different setups

16 years agoUse the IPv6 documentation prefix 2001:db8::/32 from RFC 3849 instead
reyk [Sat, 19 Jul 2008 10:35:31 +0000 (10:35 +0000)]
Use the IPv6 documentation prefix 2001:db8::/32 from RFC 3849 instead
of the deprecated 3ffe::/16 6bone test address space.  Deprecated or
public IPv6 allocations should not be used for documentation; we have
the same approach for IPv4 by using RFC 1918 private addresses.

ok jmc@ todd@

16 years agoUnbreak the tree. Pointed out by James Hartley.
kettenis [Sat, 19 Jul 2008 10:02:09 +0000 (10:02 +0000)]
Unbreak the tree.  Pointed out by James Hartley.

16 years agoIn interrupt handler change printf -> DRPINTF.
mglocker [Sat, 19 Jul 2008 06:39:00 +0000 (06:39 +0000)]
In interrupt handler change printf -> DRPINTF.

16 years agoAdd a macro that clears the want_resched flag that need_resched sets.
art [Fri, 18 Jul 2008 23:43:31 +0000 (23:43 +0000)]
Add a macro that clears the want_resched flag that need_resched sets.
Right now when mi_switch picks up the same proc, we didn't clear the
flag which would mean that every time we service an AST we would attempt
a context switch. For some architectures, amd64 being probably the
most extreme, that meant attempting to context switch for every
trap and interrupt.

Now we clear_resched explicitly after every context switch, even if it
didn't do anything. Which also allows us to remove some more code
in cpu_switchto (not done yet).

miod@ ok

16 years agono need for .Pp before or after .Sh;
jmc [Fri, 18 Jul 2008 22:51:01 +0000 (22:51 +0000)]
no need for .Pp before or after .Sh;

16 years ago+.Xr crash 8 ,
jmc [Fri, 18 Jul 2008 22:44:21 +0000 (22:44 +0000)]
+.Xr crash 8 ,
from Stephan A. Rickauer

16 years agoAdd ohci(4) and ehci(4) at cardbus(4).
kettenis [Fri, 18 Jul 2008 22:27:50 +0000 (22:27 +0000)]
Add ohci(4) and ehci(4) at cardbus(4).

16 years agoDetermine the free address space by looking at the "available" property of
kettenis [Fri, 18 Jul 2008 22:05:29 +0000 (22:05 +0000)]
Determine the free address space by looking at the "available" property of
the PCI host bridge if we're not running on an UltraBook.  Fix allocation of
bus number such that it works on machines that have OpenBoot 4.x.

16 years agoFor negotation first get the devices default values (GET_DEF) instead
mglocker [Fri, 18 Jul 2008 21:45:24 +0000 (21:45 +0000)]
For negotation first get the devices default values (GET_DEF) instead
asking for (GET_CUR) directly.  This gets us better negotation values.

16 years agoProtect dma operations with splvm instead of splhigh, that's faster than
miod [Fri, 18 Jul 2008 21:39:52 +0000 (21:39 +0000)]
Protect dma operations with splvm instead of splhigh, that's faster than
splbio and won't delay clock interrupts.

16 years agoMake IPL_VM level 5, which is guaranteed to be above bio tty and net, instead
miod [Fri, 18 Jul 2008 21:39:14 +0000 (21:39 +0000)]
Make IPL_VM level 5, which is guaranteed to be above bio tty and net, instead
of making it dynamic and the smallest value above the former three. Idea
from NetBSD.

16 years agoImplement VIDIOC_S_FMT and VIDIOC_TRY_FMT. Now V4L2 applications can
mglocker [Fri, 18 Jul 2008 18:49:11 +0000 (18:49 +0000)]
Implement VIDIOC_S_FMT and VIDIOC_TRY_FMT.  Now V4L2 applications can
set their desired image size, and therefore users can manipulate the
image size, too via the application.

Also tested by brad@

16 years agoAdd support for the kernel to recognize, load and execute position
kurt [Fri, 18 Jul 2008 16:58:06 +0000 (16:58 +0000)]
Add support for the kernel to recognize, load and execute position
independent executables using the uvm_map_pie() function to
randomize the load address. okay miod@, kettenis@, drahn@

16 years agoAdd new uvm function called uvm_map_pie() which takes align as a
kurt [Fri, 18 Jul 2008 16:40:17 +0000 (16:40 +0000)]
Add new uvm function called uvm_map_pie() which takes align as a
parameter and returns an aligned random load address for position
independent executables to use. This also adds three new vmparam.h
defines to specify the maximum address, minimum address and minimum
allowed alignment for uvm_map_pie() to use. The PIE address range
for i386 was carefully selected to work well within the i386 W^X
framework.

With much help and feedback from weingart@.
okay weingart@, miod@, kettenis@, drahn@

16 years agoAdd OmniVision OV7670 cam to the supported devices which comes in the
mglocker [Fri, 18 Jul 2008 14:45:50 +0000 (14:45 +0000)]
Add OmniVision OV7670 cam to the supported devices which comes in the
Dell XPS M1330 Notebooks.

Tested by kettenis@