openbsd
16 years ago- Use defined return values wherever possible instead inventing
mglocker [Fri, 1 Aug 2008 12:16:52 +0000 (12:16 +0000)]
- Use defined return values wherever possible instead inventing
custom ones.
- Define functions which use USB return values with usbd_status
instead int.
- Define functions which always return 0 with void instead int.
- (void) function calls which don't check the return code.

16 years agoIf no format descriptors have been found, abort attachment. Adjust some
mglocker [Fri, 1 Aug 2008 08:20:26 +0000 (08:20 +0000)]
If no format descriptors have been found, abort attachment.  Adjust some
DPRINTF and printf messages while here.

16 years agoforce route priority to RTP_DEFAULT unconditionally, effectively disabling
henning [Fri, 1 Aug 2008 05:08:07 +0000 (05:08 +0000)]
force route priority to RTP_DEFAULT unconditionally, effectively disabling
route prios. playing safe for 4.4-release, as not all parts of the system
deal correctly with route prios yet. ok claudio

16 years agosilence failures to run scsi_prevent. dmesg spam sucks.
dlg [Fri, 1 Aug 2008 01:44:20 +0000 (01:44 +0000)]
silence failures to run scsi_prevent. dmesg spam sucks.

ok krw@ marco@ miod@ deraadt@

16 years agoThis log message is an excruiciating pain in the ass when it repeats 4000 times a...
beck [Fri, 1 Aug 2008 00:32:56 +0000 (00:32 +0000)]
This log message is an excruiciating pain in the ass when it repeats 4000 times a second on a machine you need debug log on, and is actually used as a cacheing nameserver. remove just this one line of debug log to save people's sanity
ok djm@

16 years agoWait much longer for clock stabilization.
damien [Thu, 31 Jul 2008 20:14:17 +0000 (20:14 +0000)]
Wait much longer for clock stabilization.

From jcs@ and in accordance with what the Intel Linux driver does.

16 years agoUpdate to sudo 1.6.9p17
millert [Thu, 31 Jul 2008 16:44:03 +0000 (16:44 +0000)]
Update to sudo 1.6.9p17

16 years ago- Don't relay on bFormatIndex as an internal array index, since this
mglocker [Thu, 31 Jul 2008 15:26:25 +0000 (15:26 +0000)]
- Don't relay on bFormatIndex as an internal array index, since this
field is unreliable and can start with any number.  Use an own internal
array index instead.

- If the read buffer is too small, return a propper error to the calling
functions.  Just check the buffer size if we use the read(2) method since
it doesn't affect mmap(2).

Fixes kernel crashes seen with the M$ LifeCam NX-6000 and internal
(laptop) Sonix chipsets.

Tested by jcs@ (Sonix) and myself (NX-6000).

OK deraadt@

16 years agodon't allocate space for empty banners; report t8m at centrum.cz; ok deraadt
markus [Thu, 31 Jul 2008 14:48:28 +0000 (14:48 +0000)]
don't allocate space for empty banners; report t8m at centrum.cz; ok deraadt

16 years agoath(4) now in GENERIC;
jmc [Thu, 31 Jul 2008 14:03:26 +0000 (14:03 +0000)]
ath(4) now in GENERIC;

16 years agoDo not allow negative or zero delay values from the command line,
canacar [Thu, 31 Jul 2008 04:24:11 +0000 (04:24 +0000)]
Do not allow negative or zero delay values from the command line,
use the same check as the -s option. Noticed by johan@

16 years agoGive iada_modname[] room for a terminating NUL in struct ioasicdev_attach_args;
miod [Wed, 30 Jul 2008 22:02:09 +0000 (22:02 +0000)]
Give iada_modname[] room for a terminating NUL in struct ioasicdev_attach_args;
ok deraadt@

16 years agoIn the .Fn entries for timeout_add_usec/nsec, don't call the parameter
mk [Wed, 30 Jul 2008 19:41:32 +0000 (19:41 +0000)]
In the .Fn entries for timeout_add_usec/nsec, don't call the parameter
``sec''.  Instead, use usec/nsec.

Agreed with/no opposition from blambert.

16 years agoAdd ath(4).
brad [Wed, 30 Jul 2008 18:51:52 +0000 (18:51 +0000)]
Add ath(4).

ok deraadt@

16 years agoDo not print adapter target id on the attachment line, now that scsibus(4)
miod [Wed, 30 Jul 2008 18:08:02 +0000 (18:08 +0000)]
Do not print adapter target id on the attachment line, now that scsibus(4)
prints it. These should be the last offenders.

16 years agoReplace GNU tutorial with public domain one, written from scratch by
kjell [Wed, 30 Jul 2008 17:39:31 +0000 (17:39 +0000)]
Replace GNU tutorial with public domain one, written from scratch by
Mayukh Bose. Minor additions from me. Thanks!

16 years agowi@pci now works.
miod [Wed, 30 Jul 2008 17:38:02 +0000 (17:38 +0000)]
wi@pci now works.

16 years agoImplement bus_space_{read,write}_raw_mult_[248] correctly, it needs
miod [Wed, 30 Jul 2008 17:37:44 +0000 (17:37 +0000)]
Implement bus_space_{read,write}_raw_mult_[248] correctly, it needs
endianness conversion on pci bridges.
ok deraadt@ jsing@

16 years agoProper interrupt values for pci interrupt pins B to D, found in some IRIX
miod [Wed, 30 Jul 2008 17:32:30 +0000 (17:32 +0000)]
Proper interrupt values for pci interrupt pins B to D, found in some IRIX
developer documentation some time ago.

16 years agoCheck and clear the bus space mapping size instead of the bus space
brad [Wed, 30 Jul 2008 16:47:09 +0000 (16:47 +0000)]
Check and clear the bus space mapping size instead of the bus space
handle in the detach function. Fixes compilation on sparc64.

ok reyk@

16 years agosync
deraadt [Wed, 30 Jul 2008 16:18:16 +0000 (16:18 +0000)]
sync

16 years agorepair comment
deraadt [Wed, 30 Jul 2008 15:53:31 +0000 (15:53 +0000)]
repair comment

16 years agosync
deraadt [Wed, 30 Jul 2008 15:51:01 +0000 (15:51 +0000)]
sync

16 years agoadd 2nd sierra MC5725 correctly
deraadt [Wed, 30 Jul 2008 15:50:52 +0000 (15:50 +0000)]
add 2nd sierra MC5725 correctly

16 years agoAdd an alternate ID for Sierra Wireless MC5725 modem, found
canacar [Wed, 30 Jul 2008 15:40:39 +0000 (15:40 +0000)]
Add an alternate ID for Sierra Wireless MC5725 modem, found
in a Lenovo T60, reported by Juan Delgado. ok you@, jsg@ deraadt@

16 years agoCall ip6_clearpktopts() on error in copypktopts() instead of freeing
canacar [Wed, 30 Jul 2008 15:07:40 +0000 (15:07 +0000)]
Call ip6_clearpktopts() on error in copypktopts() instead of freeing
every field individually.
ok claudio@ markus@ henning@ deraadt@

16 years agoenable ix(4) on amd64
reyk [Wed, 30 Jul 2008 14:41:54 +0000 (14:41 +0000)]
enable ix(4) on amd64

ok deraadt@

16 years agoPrevent a divide by zero panic if trunkproto loadbalance is
mpf [Wed, 30 Jul 2008 10:15:35 +0000 (10:15 +0000)]
Prevent a divide by zero panic if trunkproto loadbalance is
used w/out any trunkports. Patch from Dmitri Alenitchev.
OK reyk@

16 years agotweak previous; help/ok jsg yuo
jmc [Wed, 30 Jul 2008 08:32:53 +0000 (08:32 +0000)]
tweak previous; help/ok jsg yuo

16 years agoadd missing quote; from yar@freebsd
jmc [Wed, 30 Jul 2008 08:29:49 +0000 (08:29 +0000)]
add missing quote; from yar@freebsd

16 years ago- fix spelling of Alan Shepard's name
jmc [Wed, 30 Jul 2008 08:24:47 +0000 (08:24 +0000)]
- fix spelling of Alan Shepard's name
- note that he was the first american in space (not first man)

from Derek Taylor, freebsd bin/125812

16 years agovarious markup/grammar fixes for previous, and a missing MLINK
jmc [Wed, 30 Jul 2008 08:02:34 +0000 (08:02 +0000)]
various markup/grammar fixes for previous, and a missing MLINK
for remquof.3;

16 years agospacing
reyk [Wed, 30 Jul 2008 07:43:01 +0000 (07:43 +0000)]
spacing

16 years agoUpdate the RF, RFGAIN, MODE, INI, and BBGAIN initialization tables
reyk [Wed, 30 Jul 2008 07:15:39 +0000 (07:15 +0000)]
Update the RF, RFGAIN, MODE, INI, and BBGAIN initialization tables
with different versions for various ar5212 variants and add an extra
table for PCI-E devices.  This fixes support for various newer devices
(like the 1st generation MacBook, T61 variants) but it still does not
work on a number of other devices.

Tested by many
ok deraadt@

16 years agoFree the correct buffer list on failure.
damien [Wed, 30 Jul 2008 06:25:23 +0000 (06:25 +0000)]
Free the correct buffer list on failure.

From Andrew Thompson at FreeBSD.

16 years agofix up a few free -> drm_free() cases that were missed before.
oga [Wed, 30 Jul 2008 01:30:45 +0000 (01:30 +0000)]
fix up a few free -> drm_free() cases that were missed before.

16 years agoSun stopped distributing patches through FTP :(.
kettenis [Tue, 29 Jul 2008 22:24:25 +0000 (22:24 +0000)]
Sun stopped distributing patches through FTP :(.

16 years agoUpdate to DRM git.
oga [Tue, 29 Jul 2008 22:23:50 +0000 (22:23 +0000)]
Update to DRM git.

Some stability fixes for radeon. The most part of this diff is related
to fixing up the VBLANK (vertical blank interrupt) handling. Now, if the
X driver supports the DRM_IOCTL_MODESET_CTL ioctl, (to be used when
changing the video modes), then allow the vblank to be disabled once
that ioctl has been called. Otherwise, keep the interrupt enabled at all
time, since disabling it otherwise will lead to problems.

Tested by a few. "no problem" on API/ABI deraadt@.

16 years agoUpdate list of supported machines.
kettenis [Tue, 29 Jul 2008 22:20:04 +0000 (22:20 +0000)]
Update list of supported machines.

16 years agoadd descriptions for fpclassify, isfinite, isnormal, nan, remquo,
martynas [Tue, 29 Jul 2008 22:12:29 +0000 (22:12 +0000)]
add descriptions for fpclassify, isfinite, isnormal, nan, remquo,
signbit

16 years agosync
deraadt [Tue, 29 Jul 2008 22:10:18 +0000 (22:10 +0000)]
sync

16 years ago- add man pages for nan and remainder, from freebsd
martynas [Tue, 29 Jul 2008 22:08:43 +0000 (22:08 +0000)]
- add man pages for nan and remainder, from freebsd
- mlink drem, dremf, remainderf to remainder;  nanf to nan
- describe that drem and dremf are obsolete aliases
- remove descriptions for remainder and remainderf from ieee
looked over by jmc@, but it's easier for him to work on when they
are in tree

16 years agosync
deraadt [Tue, 29 Jul 2008 22:04:38 +0000 (22:04 +0000)]
sync

16 years agodescribe exp2, exp2f and mlink them
martynas [Tue, 29 Jul 2008 21:57:13 +0000 (21:57 +0000)]
describe exp2, exp2f and mlink them

16 years agowhen detaching pcbs, also free the v6 output options so that we do not
deraadt [Tue, 29 Jul 2008 21:46:27 +0000 (21:46 +0000)]
when detaching pcbs, also free the v6 output options so that we do not
leak kernel memory for every closed v6 descriptor with certain options set
ok otto jsing todd claudio

16 years agoIn ahc_free(), which is invoked if attachment fails, do not free()
miod [Tue, 29 Jul 2008 21:18:53 +0000 (21:18 +0000)]
In ahc_free(), which is invoked if attachment fails, do not free()
pointers we don't own, but the autoconf machinery does.

16 years agosync
deraadt [Tue, 29 Jul 2008 20:19:36 +0000 (20:19 +0000)]
sync

16 years agoclarify EMSGSIZE
gilles [Tue, 29 Jul 2008 20:15:53 +0000 (20:15 +0000)]
clarify EMSGSIZE

rewording by jmc@, "lovely" deraadt@

16 years agotweak previous;
jmc [Tue, 29 Jul 2008 20:11:45 +0000 (20:11 +0000)]
tweak previous;

16 years agoAdd an ugly hack to disable DMA on the Tadpole SPARCLE laptop. It doesn't
kettenis [Tue, 29 Jul 2008 20:09:11 +0000 (20:09 +0000)]
Add an ugly hack to disable DMA on the Tadpole SPARCLE laptop.  It doesn't
work and this way we fall back on PIO which does work.  I'm pretty sure
this is caused by bugs in our code, but I cannot find them.

ok deraadt@, marco@

16 years agoSwitch all instances of malloc/free in the DRM to drm_alloc, drm_free
oga [Tue, 29 Jul 2008 19:44:13 +0000 (19:44 +0000)]
Switch all instances of malloc/free in the DRM to drm_alloc, drm_free
and drm_calloc.

With the recent change to check overflow in drm_calloc, this means that
all allocations that require multiplication are now checked. Also use
drm_calloc() when zeroing is needed and drop the bzero/memset
afterwards.  Finally, make drm_free() check for NULL, so we don't need
to do so every time.

ok miod@, deraadt@

16 years agoSplit SHA1Transform() back into smaller pieces when compiling on landisk,
miod [Tue, 29 Jul 2008 19:32:50 +0000 (19:32 +0000)]
Split SHA1Transform() back into smaller pieces when compiling on landisk,
to defeat gcc optimization bugs. Similar to what had been done on sparc64
in the past.
ok millert@ deraadt@

16 years agoStricter bounds checking for values controlling loops or memory allocations,
miod [Tue, 29 Jul 2008 19:13:56 +0000 (19:13 +0000)]
Stricter bounds checking for values controlling loops or memory allocations,
which may come from userland via ioctls. ok oga@

16 years agoCheck for multiplication overflow in drm_calloc(). ok oga@
miod [Tue, 29 Jul 2008 19:13:19 +0000 (19:13 +0000)]
Check for multiplication overflow in drm_calloc(). ok oga@

16 years agosync
deraadt [Tue, 29 Jul 2008 18:43:28 +0000 (18:43 +0000)]
sync

16 years agoGive the eisa device a valid bus_dma_tag_t.
miod [Tue, 29 Jul 2008 18:40:23 +0000 (18:40 +0000)]
Give the eisa device a valid bus_dma_tag_t.

16 years agoFix range check in sio_intr_establish().
miod [Tue, 29 Jul 2008 18:39:45 +0000 (18:39 +0000)]
Fix range check in sio_intr_establish().

16 years agoremove finite from ieee.3, and unmlink from makefile
martynas [Tue, 29 Jul 2008 18:38:21 +0000 (18:38 +0000)]
remove finite from ieee.3, and unmlink from makefile

16 years agoremove
martynas [Tue, 29 Jul 2008 18:25:51 +0000 (18:25 +0000)]
remove

16 years agoan enum specifier is more elegant than a set of #defines;
sobrado [Tue, 29 Jul 2008 18:25:28 +0000 (18:25 +0000)]
an enum specifier is more elegant than a set of #defines;
storing the program mode variable (pmode) as a global let us have
a more consistent prototype for usage().

changes suggested by pyr@.

ok millert@, pyr@

16 years agomd5(1), sha1(1), rmd160(1), cksum(1) and sum(1) have different options;
sobrado [Tue, 29 Jul 2008 18:24:31 +0000 (18:24 +0000)]
md5(1), sha1(1), rmd160(1), cksum(1) and sum(1) have different options;
while here, some KNF improvements.

based on a diff for compress(1) written by millert@,
fixes from guenther@ and pyr@.

ok millert@, pyr@; (for the manual page tweaks) jmc@

16 years agos/affect/assign/ in wmm option.
damien [Tue, 29 Jul 2008 18:20:02 +0000 (18:20 +0000)]
s/affect/assign/ in wmm option.

16 years ago- man pages for fpclassify, isfinite, isinf, isnan, isnormal, signbit,
martynas [Tue, 29 Jul 2008 18:19:17 +0000 (18:19 +0000)]
- man pages for fpclassify, isfinite, isinf, isnan, isnormal, signbit,
mostly from freebsd
- also describe deprecated aliases for isinff, isnanf, finite,
finitef
- describe behavior on vax
- add mlinks for finite, finitef, isfinite, isinf, isinff, isnan,
isnanf, isnormal, signbit
- add man pages for isgreater, isgreaterequal, isless, islessequal,
islessgreater, isunordered;  from freebsd
looked over by jmc@, it's better for him to work on when they are in tree

16 years agotweak the wep section somewhat, to bring it into line with what
jmc [Tue, 29 Jul 2008 17:03:35 +0000 (17:03 +0000)]
tweak the wep section somewhat, to bring it into line with what
we say for those devices which support wpa;

16 years ago- beef up the text on wep/wpa
jmc [Tue, 29 Jul 2008 16:44:19 +0000 (16:44 +0000)]
- beef up the text on wep/wpa
- simplify wpa example

help/ok damien

16 years agodisable these targets since this will not be fixed until post-release
deraadt [Tue, 29 Jul 2008 15:40:36 +0000 (15:40 +0000)]
disable these targets since this will not be fixed until post-release

16 years agoThere is currently no CD-ROM install available for OpenBSD/sgi.
jsing [Tue, 29 Jul 2008 15:06:51 +0000 (15:06 +0000)]
There is currently no CD-ROM install available for OpenBSD/sgi.

ok deraadt@

16 years agoexample code fix from Will Maier; ok otto
jmc [Tue, 29 Jul 2008 13:53:14 +0000 (13:53 +0000)]
example code fix from Will Maier; ok otto

16 years agoThe stream header length and flags field are uint8_t not int.
mglocker [Tue, 29 Jul 2008 13:45:04 +0000 (13:45 +0000)]
The stream header length and flags field are uint8_t not int.

16 years agocheck for m_devget() returning NULL, increment ierror count
thib [Tue, 29 Jul 2008 10:05:38 +0000 (10:05 +0000)]
check for m_devget() returning NULL, increment ierror count
and just return;

OK mglocker@

16 years agoadd description for the modem which has multiple com ports.
yuo [Tue, 29 Jul 2008 06:51:54 +0000 (06:51 +0000)]
add description for the modem which has multiple com ports.

checked and ok jsg@

16 years agoWhen paused (or overrun), the record ring pointers are not incremented
ratchov [Tue, 29 Jul 2008 05:59:11 +0000 (05:59 +0000)]
When paused (or overrun), the record ring pointers are not incremented
properly in audio_rint(), the periodic boundary conditions aren't met. This
causes, later read(2) to return EFAULT while trying to access unmapped
regions of the kernel address space. Fix this by using the correct pointer
arithmetic.

ok jakemsr@

16 years agoReenable the AUXWRITE dance to try and find the aux port in pckbc(4), based
miod [Tue, 29 Jul 2008 04:20:35 +0000 (04:20 +0000)]
Reenable the AUXWRITE dance to try and find the aux port in pckbc(4), based
on an attachment dependent flag; the Tadople SPARCLE needs it.
ok kettenis@ deraadt@

16 years agoAdd rebuild mark so that we can restart a rebuild upon clean shutdown.
marco [Tue, 29 Jul 2008 01:18:02 +0000 (01:18 +0000)]
Add rebuild mark so that we can restart a rebuild upon clean shutdown.
Fix a typo while here.

The rebuild field was appended to the metadata so that we can automatically
update from v2 to v3.  The only visible difference will be that in the dmesg
the SCSI version number will go from 002 to 003.

ok deraad.

16 years ago- fix FIFO overruns on PCI-E chipsets by setting the DMA size
reyk [Tue, 29 Jul 2008 00:18:25 +0000 (00:18 +0000)]
- fix FIFO overruns on PCI-E chipsets by setting the DMA size
RX/TX configuration registers to 128 instead of 512 bytes.
- add a few more MAC/RF id strings for the dmesg.
- check for PCI-E instead of single chip variants in a few places.

ok deraadt@

16 years agosync
deraadt [Tue, 29 Jul 2008 00:12:18 +0000 (00:12 +0000)]
sync

16 years agoInstall man pages for the BIO_* libcrypto functions, but not bio.3
jsg [Mon, 28 Jul 2008 23:44:10 +0000 (23:44 +0000)]
Install man pages for the BIO_* libcrypto functions, but not bio.3
as the page doesn't directly describe any functions.

ok deraadt@

16 years agoOnly optimize away plt relocation when obj_base == 0. Corrects
kurt [Mon, 28 Jul 2008 22:11:00 +0000 (22:11 +0000)]
Only optimize away plt relocation when obj_base == 0. Corrects
pie relocation. Assembly debugging help from kettenis@.

okay kettenis@ miod@

16 years agozap some extra spaces and tabs.
fgsch [Mon, 28 Jul 2008 20:58:41 +0000 (20:58 +0000)]
zap some extra spaces and tabs.

16 years agoWhen getting the string descriptors, fetch 2 bytes (size and type) first,
fgsch [Mon, 28 Jul 2008 20:49:28 +0000 (20:49 +0000)]
When getting the string descriptors, fetch 2 bytes (size and type) first,
and then the whole string based on the returned size, not just one byte.
The Lexar card reader kindly loaned by
Denis Doroshenko (denis dot doroshenko at gmail.com) it's happy now, and
after further inspection it's what other OSs do.
Discussed with yuo@. krw@ yuo@ ok.

16 years agosome SEE ALSO fixes;
jmc [Mon, 28 Jul 2008 20:19:32 +0000 (20:19 +0000)]
some SEE ALSO fixes;

16 years agodocument WPA-PSK support.
damien [Mon, 28 Jul 2008 20:10:28 +0000 (20:10 +0000)]
document WPA-PSK support.

"i would really appreciate just something going in" jmc@

16 years agoman page bits for new malloc; ok jmc@
otto [Mon, 28 Jul 2008 19:53:03 +0000 (19:53 +0000)]
man page bits for new malloc; ok jmc@

16 years agomention newer versions.
brad [Mon, 28 Jul 2008 19:52:21 +0000 (19:52 +0000)]
mention newer versions.

16 years agoignore PS mode changes and PS-Poll from non-associated STAs.
damien [Mon, 28 Jul 2008 19:42:13 +0000 (19:42 +0000)]
ignore PS mode changes and PS-Poll from non-associated STAs.
keep track of the number of associated STAs in PS mode.

16 years agoAlign the route message length to the next natural boundary via ALIGN() to
claudio [Mon, 28 Jul 2008 19:38:07 +0000 (19:38 +0000)]
Align the route message length to the next natural boundary via ALIGN() to
ensure that the next message is nicely aligned as well. Some archs (sparc, m88k)have strict alignement issues since the inclusion of 64bit counters in the
messages because those values could end up on non aligned addresses.
Found and fix tested by miod@, OK deraadt@

16 years agoNo longer clear ci_want_resched within cpu_switchto(), now that it's done
miod [Mon, 28 Jul 2008 19:08:43 +0000 (19:08 +0000)]
No longer clear ci_want_resched within cpu_switchto(), now that it's done
in the MI code.

16 years agoregen
miod [Mon, 28 Jul 2008 18:51:11 +0000 (18:51 +0000)]
regen

16 years agoPut XBOW in the generated constants, not PCI...
miod [Mon, 28 Jul 2008 18:50:59 +0000 (18:50 +0000)]
Put XBOW in the generated constants, not PCI...

16 years agoIn addition to cop0 sr, prevent cop0 ic from being modified in
miod [Mon, 28 Jul 2008 18:48:25 +0000 (18:48 +0000)]
In addition to cop0 sr, prevent cop0 ic from being modified in
process_write_regs().

16 years agoUpdate to tzcode2008e from elsie.nci.nih.gov
millert [Mon, 28 Jul 2008 18:22:24 +0000 (18:22 +0000)]
Update to tzcode2008e from elsie.nci.nih.gov

16 years agoUpdate to tzdata2008e from elsie.nci.nih.gov
millert [Mon, 28 Jul 2008 18:19:57 +0000 (18:19 +0000)]
Update to tzdata2008e from elsie.nci.nih.gov

16 years agosync
deraadt [Mon, 28 Jul 2008 17:57:49 +0000 (17:57 +0000)]
sync

16 years agoRemove dead code.
miod [Mon, 28 Jul 2008 17:57:25 +0000 (17:57 +0000)]
Remove dead code.

16 years agoIn process_write_regs() and sigreturn(), be more strict about the bits
miod [Mon, 28 Jul 2008 17:50:09 +0000 (17:50 +0000)]
In process_write_regs() and sigreturn(), be more strict about the bits
userland is allowed to change in psr.

16 years agoAccording to the manual, delay slot kill the benefits of being superscalar,
miod [Mon, 28 Jul 2008 17:49:38 +0000 (17:49 +0000)]
According to the manual, delay slot kill the benefits of being superscalar,
so don't use any in the 88110-specific parts of locore.

16 years agosync
reyk [Mon, 28 Jul 2008 16:56:05 +0000 (16:56 +0000)]
sync

16 years agorevert AR5416 PCIE to AR5418, the product description is clear about
reyk [Mon, 28 Jul 2008 16:53:18 +0000 (16:53 +0000)]
revert AR5416 PCIE to AR5418, the product description is clear about
it while atheros' driver says something wrong (even if it is
technically correct).  also rename AR9280 PCIE to AR9281 (guessed).

triggered by kettenis@

16 years agodocument which functions use which YP maps; Ingo Schwarze
deraadt [Mon, 28 Jul 2008 16:27:39 +0000 (16:27 +0000)]
document which functions use which YP maps; Ingo Schwarze

16 years agosome cleanup; Gleydson Soares
deraadt [Mon, 28 Jul 2008 16:22:11 +0000 (16:22 +0000)]
some cleanup; Gleydson Soares