Fix .At argument.

(joerg)

Use plain references and don't depend on multiple expansion runs.

(joerg)

Mask %U, it is a macro nowadays.

(joerg)

\\ -> \e

(joerg)

Added touchpad driver for WS003SH.

(nonaka)

Reflect canonical order by majority rule. The local version in share/man
already got this chance a while ago.

(joerg)

Reorder prologue

(joerg)

Reorder prologue

(joerg)

Remove the note about the standard recommending fgets(3);
should be clear enough from the big warning to "NEVER use gets()".

(jruoho)

Use COM_REGMAP instead of hd64461uart_{read,write}_1().

(kiyohara)

PR/19069: SUENAGA Hiroki: Original fix was defective because on big endian
the expression is not truncated to 16 bits, thus sending to port 65536 makes
the == 0 test fail and the packet ends up being sent.

(christos)

Drop trailing white space. Use .Oo / .Oc explicitly to annotate that the
empty content is intended.

(joerg)

typo in comment

(dholland)

Standby AFE Clock, if use_afeck is 0.  Don't stop AFE Clock, if PERSONA boots
on serial console.

(kiyohara)

Move the structure inside the body text. Some additional markup improvements.

(jruoho)

cast isblank(3)'s argument to unsigned char.

(tnozaki)

Build links for syslogp(3), syslogp_r(3), vsyslogp(3), and vsyslogp_r(3).

From Takahiro Kambe in PR lib/43301.

(jruoho)

moveed callout_init call from sysmonopen_wdog to sysmon_wdog_init

(cliff)

Hook kthread_join(9) to the build.

(jruoho)

In order to widen the paragraphs, move the newly added list of flags below
the list of function parameters. Add two small sentences to give few more
clues about KTHREAD_JOINABLE. Add a missing prototype for kthread_join().
Also: newpl -> newlp.

(jruoho)

Patches by Toru Nishimura:
- set up DBATs for 80000000-8fffffff and fc000000-ffffffff
- honour DHCP's filename field, use 'netbsd' when unspecified
- PCI-fixups for KuroBox
- determine PCIIDE programming mode for BAR assignment

(phx)

Put flags into a list. Bug fixes to kthread_join docs.

(wiz)

Add a spare int field to ifa_msghdr so its length is a multiple of 8.

(matt)

Decode the routing message using the new RT_ROUNDUP alignment restrictions.

(matt)

Make sure all structure lengths are rounded via RT_ROUNDUP in routing messages.
This simplies the protocol since all items will now start on a RT_ROUNDUP
aligned address independent of the structure.

(matt)

Add RTAX_NAMES macro to initialize an array of names for RTAX_*

(matt)

implement fullcmd mode. Now we need to modernize the machine specific code
to handle displays > 80 columns to make this work effectively.

(christos)

Fix grammar in a line of output.
From Hywel Mallett via FreeBSD.

(mbalmer)

set RI_NO_AUTO then attaching early, clear it otherwise

(macallan)

Update also keywords.{c,h} generator in order to reflect latest changes.

(kefren)

Force the Watchdog interrupts to NMI and don't mask them.

(matt)

hid_get_data() does not work if the size of data is less than
a byte and crosses a byte boundary, and it always returns a
sign-extended value.

fix this by using the algorithm from libusbhid to read bytes,
and provide a hid_get_udata() function to return unsigned data
values.

while here, const args

(plunky)

note const from actual API

(plunky)

Typo: DHB should be DMB (most significant bye, not high byte).

(phx)

hid_get_data() will read an extra byte if the data being read ends on
a byte boundary. This byte is subsequently discarded, but it could be
a byte from memory after the end of the report being parsed.

Fix this by rounding up and ending the loop one earlier.

(plunky)

Style.

(kiyohara)

Suply clock SAFECKE_[IO]ST, if platid is PERSONA.
  tested on HPW-50PA.

(kiyohara)

PR/40455: Mihai Chelaru: Add noreject noblackhole options

(christos)

Rewrote most of the code to make the 8245 I2C module finally work.
Tested with a RICOH RS5C372 RTC on a Synology DS-101g+ and Kurobox.
Reviewed by <nisimura>.

(phx)

Print a warning if AcpiOsSleep() is called with a value larger than two
seconds. It is known that there are systems in the field that pass bogus AML
values to the Sleep() operation code, possibly requesting delays that could
be measured in days.

Discussed with jmcneill@.

XXX: While the used mstohz(9) is documented to round to one second if the
    passed value is larger than 131072 ms, we may still need to force a
    sensible upper limit if this warning starts to appear.

(jruoho)

fix inversion: advance clock on cpu0, not the complement of cpu0

(pooka)

Fix arg of intr_establish().
  comintr() expects struct com_softc* for arg.

(kiyohara)

When scanning the device "capabilities" (power, wake-up, etc.),
include all device nodes, regardless of the status of the device.

XXX: It is known that some systems implement the _STA method incorrectly.
    If needed in the future, attachment based on the values from this
    method may need revisiting. Same goes for ACPI_ACTIVATE_DEV.

(jruoho)

Initial pass for issues observed by dyoung@ on HP Pavilion N3270. In this
laptop there is a PNP0C0B ("ACPI fan") with the following properties:

        _PSC : Power state for D3 (alone).
        _PR0 : Power resources for D0.
        _PSx : Power state switch for D0 and D3.

Thus, it is impossible to get or set the D3 power state via power resources
alone; there is only a single PowerResource() and it is for D0.

To tackle this:

  1. Evaluate the direct _PSC control method if and only if there is no
    given _PRx. The order is important; it is known that some other
    systems implement the _PSC method (like _STA) incorrectly.

  2. If no _PRx is available (and thus no _ON or _OFF), do not error out.
    Instead, if we have AE_NOT_FOUND, continue to evaluate the power state
    switch method, _PSx, which (on this laptop) should alone suffice for
    the D0 -> D3 transition.

(jruoho)

Document kthread_join in kthread(9).

(haad)

Fix bogus KASSERT, typo and KNFize my previous patch.

(haad)

Add PSL_HIGHIPL for recent subr.S change

(matt)

move MKDEBUG entries for libs from md.shark to ad.arm to match ../base/ad.arm

(lukem)

regen

(christos)

Use device names from the OpenBSD iwn driver, adding and modifying device
names as needed. From Sverre Froyen

(christos)

Remove obsolete comment about a data corruption on IDE controllers
in Blade 100/Netra X1.

Workaround for a data corruption was commited as rev. 1.16 of
src/sys/dev/pci/aceride.c in 2005.

(nakayama)

Use callout_halt() instead of callout_stop() - otherwise gem_detach() might
be too fast and run callout_destroy() while another cpu is still busy
in the callout.

(martin)

pool_{cache_}get: improve previous diagnostic by checking for panicstr,
so it wont trigger the assert while trying to dump core on crash.

(rmind)

make this work again

(macallan)

Fix gimplish.

(snj)

- Sprinkle asserts to catch calls from interrupt context on IPL_NONE pools.
- Add diagnostic drain attempt.

(rmind)

execve1: move few PNBUF_PUT() outside the lock paths.

(rmind)

Make sure the softint lwp runs at IPL_HIGH.

(matt)

Add support for kthread_join in our kernel thread implementation. This is used
by zfs but I think that it can be generaly usefull. Thread need to be created
with KTHREAD_JOINABLE flag and can be joined only once.
When joinable thread was created it will not e automatically reaped from system
and kthread_join must be called on it to reap it.

Ok by ad@.

(haad)

Actually, push defining _RUMPKERNEL down to libkern, since it's
not needed elsewhere.

(pooka)

Fix userspace cv_timedwait implementation to work on NetBSD.

(haad)

Use assembly since deref a 64bit value as a pointer does not make a
32bit compiler happy.

(matt)

Need to turn KX for N32 kernels with mips3_lw_a64 and mips3_sw_a64

(matt)

fix typo

(pooka)

grammar police!

(matt)

Limit visibility of _RUMPKERNEL to prevent abuse.

(pooka)

Add siisata/wd for LannerInc MR-460

(matt)

Fix printf for u_quad_t route changes.

(matt)

add __HAVE_CPU_COUNTER stubs where possible (i.e. where the arch
doesn't think inlines are the second compiling)

(pooka)

Fixup for u_quad_t route changes.

(matt)

Fixup for u_quad_t change for route interface.

(matt)

Use RT_ROUNDUP which will always do the right thing for alignment now.

(matt)

A few changes that make the route interface and related sysctls 32/64 bit
independent so the netbsd32 userland can use them.

(matt)

Properly pad some structures to 64bit.  A N32 ktruss on a N64 kernel now
works (this might not work for sparc/x86 if their register_t for 32bit isn't
the same size as register_t).

(matt)

If icache_sync_index_range is > way size, limit to way size.

(matt)

Change to not use t8.  Any former use of t8 is now t9.
Any former use of t9 is going to either AT or a to-be-overwritten register.

(matt)

Test for barebones rump network config (i.e. just rumpnet and kernel base).

(pooka)

adjust comment in previous.
XXX: should make that (and physmem) mean something here

(pooka)

remove unnecessary #ifdef

(pooka)

regen: _RUMPKERNEL -> _KERNEL

(pooka)

_RUMPKERNEL -> _KERNEL

(pooka)

_RUMPKERNEL -> _KERNEL

(pooka)

This not working on ews4800mips has nothing to do with rump, so:
_RUMPKERNEL -> _MACHINE != ews4800mips

(pooka)

Fix YOW entry while I am here...

(mbalmer)

Language fix: upto date -> up to date.

(mbalmer)

Cache directory entry name length.  This brings kernel bootstrap
time down: 14ms -> 12ms.  Further hashing etc. did not seem to have
any noticable effect.
(without /dev node creation bootstrap time is 8ms, so it's still
the bottleneck)

(pooka)

Cosmetic: remove superfluous parentheses from return statements.  Don't
test truth of pointers, or compare with 0, but compare with NULL,
instead.

The compiler generated the same assembly language before & after this
change.

(dyoung)

Correctly spell 'acknowledged'.

Found by Christian Brueffer.

(mbalmer)

Initialize p_pgrp when creating a new process structure (and not
only for proc0).  This makes something work.  I just can't remember
what it was anymore.

(pooka)

Set default number of vnodes to 1k instead of 64k: a large default
reserves a large amount of memory by default and this is not
desirable in a rump kernel where the typical usage is minimal.
Maybe I should write a few lines to autoscale desiredvnodes up to
a hard limit after the soft limit is reached?

(pooka)

Fix reclaim locking so that we don't attempt lock reentry if making
a new rumpfs vnode triggers a reclaim for a rumpfs vnode.

(pooka)

uvm_object_printit() should be wrapped in DEBUGPRINT

(pooka)

Add SSP SSSR register definitions.

(nonaka)

update slightly

(pooka)

Added some CKEN_* definition for pxa270.

(nonaka)

Sync with md.h rev 1.17:
> convert sparc64 to MULTIPROCESSOR kernel by default, and introduce
> s/MP/UP/ kernels were otherwise in place.

(tsutsui)

drop silly backronym.  just rump.

(pooka)

the crowd went crazy
as typo left the stage

(pooka)

Reclaim spec-type vnodes properly.

(pooka)

Fix some typos in comments.

(njoly)

libalpha.so.0.0.debug

(lukem)

Add MASK_SOFT_FLOAT to TARGET_CPU_DEFAULT for mips64

(matt)

Fix indentation

(joerg)

use RI_NO_AUTO for the gdium's early rasops console so we don't try to
allocate memory before kmem is ready

(macallan)

Fix tunnelling (ssh -w).  The kernel expects a device to be passed
to the socked, not a path.

(jnemeth)

Provide pmap_enter_ma(), pmap_extract_ma(), pmap_kenter_ma() in all x86
kernels, and use them in the bus_space(9) implementation instead of ugly
Xen #ifdef-age.  In a non-Xen kernel, the _ma() functions either call or
alias the equivalent _pa() functions.

Reviewed on port-xen@netbsd.org and port-i386@netbsd.org.  Passes
rmind@'s and bouyer@'s inspection.  Tested on i386 and on Xen DOMU /
DOM0.

(dyoung)

Add WZERO3 kernel to MD kernel set menu for
possible newer users of these smartphones:
http://www.willcom-inc.com/ja/lineup/ws/004sh/
http://www.willcom-inc.com/ja/lineup/ws/007sh/
http://www.willcom-inc.com/ja/lineup/ws/011sh/
etc.

XXX: sysinst/util.c:init_set_status() calls md_init_set_status()
XXX: to allow MD hook choose the right kernel, but it can do nothing
XXX: because set_status[] was changed to static. Is it okay to export it?

(tsutsui)

Add MASK_SOFT_FLOAT to G_target_cpu_default

(matt)

PrintOnError: run .ERROR last, so even if it causes problems we
still print MAKE_PRINT_VAR_ON_ERROR.

(sjg)

Kill a left-over break from the for-loop that vanished in 1.13.
Issue noticed by tnn.

(joerg)

Include <bsd.own.mk> before ${NETBSDSRCDIR}.

(tsutsui)

merge 20100510 properly

(mrg)

re-do this with the right sources this time.

Status:

Vendor Tag: bozohttpd
Release Tags: bozohttpd-20100510

(mrg)

merge bozohttpd 20100510.

(mrg)

import bozohttpd 20100510 -- just fixes some compile errors.

Status:

Vendor Tag: bozohttpd
Release Tags: bozohttpd-20100510

(mrg)

Make this compile on sparc64.

(tsutsui)

Use kmem(9) instead of malloc(9).

(kiyohara)

Regen.

(matt)

Add Marvell 88SB2211

(matt)

Regen.

(matt)

Add MARVELL 88SB2211 PCIe-PCI bridge

(matt)

Disable building of stub_stand.c we build all locking primitives in libzpool.

(haad)

If thread_create was called with thread_id = NULL use our local thread id
as replacement otherwise we will crash.

(haad)

Fix spa_thread to compile under the NetBSD.

(haad)

Get new vnode when creating zfs share dir for it's znode.

(haad)

note bozohttpd 20100509

(mrg)

merge bozohttpd 20100509.

(mrg)

import bozohttpd 20100509.  it has these changes:
o  major rework and clean up of internal interfaces.  move the main
  program into main.c, the remaining parts are useable as library.
  add bindings for lua.  by Alistair G. Crooks <agc@netbsd.org>
o  fix http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=566325


special thanks to al for the majority of these changes.

Status:

Vendor Tag: bozohttpd
Release Tags: bozohttpd-20100509

(mrg)

The interrupt handler uses the softc, so it's probably a good idea to
initialize it before causing interrupts. Prevents a panic on VAXstation
4000/9x with SPX console.

(hans)

Make console keyboard work on VAXstation 4000/9x. Seems not to break
anything on other VAXstations (tested on 4000 VLC).

(hans)

Make SPX work as console device on VAXstation 4000/9x.

(hans)

Drop x86 MD package/core/smt IDs and use MI.

(rmind)

Pull-up rev 1.30 from HEAD.  OK matt@.

(rmind)

Add node/core/smt ID numbers to struct cpu_data, will be filled by MD.
OK matt@.

(rmind)

Don't try to build xldscripts during build.sh modules

(morr)

Remove trailing whitespace and dot in Nd.

(wiz)

Add missing license

(degroote)

MKDEBUG support for X11R6 shlibs

(lukem)

libm68k debug

(lukem)

Added touchpad driver for WS007SH.

(nonaka)

pfs.debug

(lukem)

Added 2 options for X.

(nonaka)

Fix tools build on Solaris with MKCROSSGDB=YES.

(hans)

document IPFILTER_LOOKUP and IPFILTER_COMPAT.
use "IP-Filter" instead of "ip-filter".
bump date.

(mrg)

regenerate these configurations from GENERIC.in.

(mrg)

enable IPFILTER_COMPAT in all kernels that have ipfilter already.
canonicalise several of the ipf option segments in various files
(this mostly means adding commented out IPFILTER_DEFAULT_BLOCK,
or adding commented or uncommented IPFILTER_LOG or IPFILTER_LOOKUP
option statements.)

i built about 20 of these kernels to check, but not all of them.

(mrg)

Regen from GENERIC.in rev 1.89:
> - don't enable TMPFS for SMALL030
> - remove dup PTYFS

(tsutsui)

- don't enable TMPFS for SMALL030
- remove dup PTYFS

(tsutsui)

PCI-ISA Bridge paired with M5229 UDMA IDE Controller rev. 0xC3 is
M1533.

Follow the fix in rev 1.1004 of pcidevs.

(nakayama)

Support for IOMEGA Storcenter, by Toru Nishimura.
Kurobox console is 57600.

(phx)

Synology uses sk(4) NIC.
Included information about Iomega StorCenter. Note that the EPIC mapping
is not verified.

(phx)

Rework the way interrupts are decided on.  Don't use pa_bus since that's
arbitrary.  Instead grab the device from pa_intrtag since that corresponds
to the PCIe bus we are actually attached to.

While I'm here, compact some switch statements into a few simple assignments.

(matt)

Synology boxes need sk(4), not msk(4).

(phx)

Make the .pc file rules depend on the Makefile so that when rebuilding
the tree with MKUPDATE=yes set after an atf upgrade, the .pc files get
the correct version in them.

Suggested by njoly@ in private mail.

(jmmv)

Print "Kuro Box" when detected. Print "Unknown board" when the hardware was
not identified.

(phx)

Calculate busclock and cpuclock for 8245-based boards from PLL_CFG and HID1.
Print the clocks after netboot's welcome-message.
Make sure DUART-mode is enabled for eumb-systems.
Support for communication with the satellite processor via the 2nd UART.
Currently only KuroBox and Synology make use of it. Synology also resets the
board via its satellite.
Fixed console speed for KuroBox and QNAP (it is 115200).

(phx)

Remove trailing whitespace; new sentence, new line; sort sections;
standardize SYNOPSIS; use serial comma.

(wiz)

Wording.

(wiz)

Sort options, standardize SYNOPSIS, slight rewordings. Use more markup.

(wiz)

Remove double quotes around missing function in BUGS section.

(wiz)

Adjust paths to new location of the atf-compile sources in atf 0.8.
(You may be happy to know that this tool will probably disappear in
the next atf release!)

(jmmv)

Add missing debug libraries for MKDEBUG=yes.

(jmmv)

Note update of atf to 0.8.

(jmmv)

Update file lists to match atf 0.8.

(jmmv)

Update directories to match the atf 0.8 layout.

(jmmv)

Update reachover Makefiles for atf 0.8.

(jmmv)

Merge atf 0.8.

(jmmv)