Implement the XuhostPWREN/XuhostOVERCUR register selection
Update the usb2 isolation. It ought to work for Exynos5 too.

Only issue now is ohci not working

(reinoud)

Add usb20 phy host link enable
insert extra space

(reinoud)

Remove extra line

(reinoud)

Xuhost bits for Exynos5 are located in the two top bits

(reinoud)

disable dma on cubietruck for now

(jmcneill)

if "dma" property is set to false, fallback to pio mode

(jmcneill)

Eliminate IFAREF() and IFAFREE() macros in favour of functions.

(rmind)

do BUS_DMASYNC_POSTWRITE after transfer completes

(jmcneill)

Update CPUID signature values from the latest Intel SDM.
- Core M-5xxx
- Core i7 extreme
- Future Core (0x4e)
- Future Xeon (0x56)

(msaitoh)

Add new cache descriptor (0xc3) from the latest Intel SDM.

(msaitoh)

Update CPUID(EAX=0x0d, ECXD=1) from Intel SDM:
- XSAVEC(bit1)
- XGETBV(bit2)
- XSAVES(bit3)

(msaitoh)

add motg

(jmcneill)

initial A31 support

(jmcneill)

turn on ntpd

(jmcneill)

Add missing defines for 16, 24 and 32 bit NE and OE formats.
Add the ability to set 24 and 32 bit precision.

This commit was approved by wiz@

(nat)

Mention ossaudio(3) inital OSSv4 support from May.
This commit was approved by wiz@

(nat)

Add a test for backslash-newline at the very end of a command script.

(apb)

Change @echo to echo in several places in escape.mk,
to make it easier to spot changes in what make sends to the shell.
Adjust escape.out to match.

(apb)

Remove debug printf, now that MII status changes work

(martin)

Make the MII clock variable and passed in from the frontend.

(martin)

Better safe than sorry: do the full clock setup, and pass the MII clock
used (depending on board configuration, we use RGMII mode) to the
core attach function.

(martin)

Ticket 1008.

(msaitoh)

Ticket 1008.

(msaitoh)

Ticket 1083, 1108-1110.

(msaitoh)

Pull up following revision(s) (requested by martin in ticket #1110):
sys/arch/sparc/sparc/autoconf.c: revision 1.258
PR port-sparc/49103: add device name mapping for "qfe" (quad hme) cards.
From MOCHIDA Shuji.

(msaitoh)

Pull up following revision(s) (requested by tls in ticket #1109):
external/gpl2/xcvs/dist/src/server.c: revision 1.7
Move the arbitrary limit on number of files per commit into
1000000 earlier today.

(msaitoh)

Pull up following revision(s) (requested by tls in ticket #1108):
external/gpl2/xcvs/dist/src/server.c: revision 1.6
Relax arbitrary limit of 10,000 changed files per commit idiotically
introduced in CVS 1.11.17.

(msaitoh)

Pull up following revision(s) (requested by dholland in ticket #1083):
usr.sbin/ypbind/ypbind.c: revision 1.91
usr.sbin/ypbind/ypbind.c: revision 1.92
usr.sbin/ypbind/ypbind.c: revision 1.93
usr.sbin/ypbind/ypbind.c: revision 1.94
usr.sbin/ypbind/ypbind.c: revision 1.95
usr.sbin/ypbind/ypbind.c: revision 1.96
usr.sbin/ypbind/ypbind.c: revision 1.97
usr.sbin/ypbind/ypbind.c: revision 1.98
usr.sbin/ypbind/ypbind.8: revision 1.20
usr.sbin/ypbind/ypbind.8: revision 1.19
Don't store the default domain name in a global. While running we
really don't care which domain is the system's default domain.
Factor out some rpc validation code.
While there are times it's appropriate to call a state variable
"evil", this isn't one of them. Since the logic involved is to wait
until the default domain binds before backgrounding, call the variable
"started" instead.
Don't rake up the default domain until after processing arguments.
Processing arguments just sets flags -- may as well do it first, and
this way detection of silly errors isn't contingent on having things
fully configured and operating.
Load up with comments.
Instead of using magic numbers in what looks like a boolean
(dom_alive), create a state enumeration (domainstates) and use it
instead.
Instead of three states (new, alive, and, effectively, 'troubled') go
to five: new, alive, pinging, lost, and dead.
Domains start in the NEW state. When we get a reply from a server, the
state goes to ALIVE. The state is set to PINGING when we ping the
server (once a minute normally) and if the ping times out, it goes to
LOST. If we stay lost for a minute, go to DEAD, and in DEAD, do
exponential backoff of nag_servers calls.
Getting rid of the broken logic attached to the 'troubled' state fixes
PR 15355 (ypbind defeats disk idle spindown) -- it will now only
rewrite the binding file when the binding changes.
Also, fix the HEURISTIC code so it doesn't trigger except in ALIVE
state. I think this was the source of a lot of the spamming behavior
seen in PR 32519, which is now fixed.
Might also fix PR 23135 (broadcast ypbind sometimes fails to find
servers).
Add a SIGHUP handler; upon SIGHUP do an extra nag_servers on any
domain that's in DEAD state. This lets you explicitly rescue ypbind
from its exponential backoff when you know the world's back up.
Log state transitions.
Don't store the default domain name in a global. While running we
really don't care which domain is the system's default domain.
Factor out some rpc validation code.
While there are times it's appropriate to call a state variable
"evil", this isn't one of them. Since the logic involved is to wait
until the default domain binds before backgrounding, call the variable
"started" instead.
Don't rake up the default domain until after processing arguments.
Processing arguments just sets flags -- may as well do it first, and
this way detection of silly errors isn't contingent on having things
fully configured and operating.
Load up with comments.
Instead of using magic numbers in what looks like a boolean
(dom_alive), create a state enumeration (domainstates) and use it
instead.
Instead of three states (new, alive, and, effectively, 'troubled') go
to five: new, alive, pinging, lost, and dead.
Domains start in the NEW state. When we get a reply from a server, the
state goes to ALIVE. The state is set to PINGING when we ping the
server (once a minute normally) and if the ping times out, it goes to
LOST. If we stay lost for a minute, go to DEAD, and in DEAD, do
exponential backoff of nag_servers calls.
Getting rid of the broken logic attached to the 'troubled' state fixes
PR 15355 (ypbind defeats disk idle spindown) -- it will now only
rewrite the binding file when the binding changes.
Also, fix the HEURISTIC code so it doesn't trigger except in ALIVE
state. I think this was the source of a lot of the spamming behavior
seen in PR 32519, which is now fixed.
Might also fix PR 23135 (broadcast ypbind sometimes fails to find
servers).
Add a SIGHUP handler; upon SIGHUP do an extra nag_servers on any
domain that's in DEAD state. This lets you explicitly rescue ypbind
from its exponential backoff when you know the world's back up.
Log state transitions.
Document exponential backoff behavior and SIGHUP support, plus a couple
other minor edits.
Use more markup.

(msaitoh)

Fix MII setup and interrupt handling, from Robert Swindell.

(martin)

Mark v7fs_lstat_symlink as an expected failure, referencing PR kern/48864.

(gson)

Restore -r1.234 of make.1, which was an unrelated fix from sjg pertaining
to the docs for the -w option.

(dholland)

Revert the man page as well. (hi joerg)

Revert it to 20140823 (-r1.230), before the controversial commits,
which changed it a good deal.

(dholland)

__USE(id) in the !PCIBIOSVERBOSE case.
Fixes PR 49181.

(apb)

- record the sector size of the disk
- correct confusion with end_cylinder and end_sector in MBRs

(jnemeth)

Restore apb's 20140820 commit (-r1.228 of main.c):
It should not be an error to have VAR != command that prints no output

Joerg reverted a bit too enthusiastically.

(dholland)

cubieboard2 uses awge not awe

(jmcneill)

add BUFQ_PRIOCSCAN

(jmcneill)

eliminate most of the polling from awin_mmc_exec_command

(jmcneill)

With PMAP_NEED_ALLOC_POOLPAGE, before using pmap_map_poolpage make sure
the page comes from the same freelist that pmap_map_poolpage uses.

(matt)

Replace random with cprng_fast32. Reorganise computation to replace
(32bit) division with (long) multiplication.

(joerg)

Prefer cprng_fast32 over random. A good distribution even in the lower
bits beat any minor performance advantage randomo(9) might have,
especially given the disk IO involved.

(joerg)

Pull up following revision(s) (requested by tsutsui in ticket #1008):
sys/arch/next68k/next68k/pmap_bootstrap.c: revision 1.43
Move physmem calculations before nptpage initialization.
Should fix next68k specific part of PR port-m68k/45915
(panic: pmap_enter_ptpage: can't get KPT page), and
this is the last possible m68k MD part of this PR.
Note this change is not tested on the actual machine (yet),
but as noted in comment next68k/pmap_bootstrap.c is based on
the mvme68k one which has been fixed by the similar diff.
Should be pulled up to all netbsd-6 branches.

(msaitoh)

Pull up following revision(s) (requested by tsutsui in ticket #1008):
sys/arch/next68k/next68k/pmap_bootstrap.c: revision 1.43
Move physmem calculations before nptpage initialization.
Should fix next68k specific part of PR port-m68k/45915
(panic: pmap_enter_ptpage: can't get KPT page), and
this is the last possible m68k MD part of this PR.
Note this change is not tested on the actual machine (yet),
but as noted in comment next68k/pmap_bootstrap.c is based on
the mvme68k one which has been fixed by the similar diff.
Should be pulled up to all netbsd-6 branches.

(msaitoh)

Add rgephy.

(rjs)

Ticket 74-76, 78, 80 and 81.

(msaitoh)

Pull up following revision(s) (requested by tron in ticket #81):
external/bsd/cron/dist/crontab.c: revision 1.8
Rewind the file before we install it so that we can check it.
Before this fix, no checks were made and you could install crap.
XXX: pullup 7

(msaitoh)

Pull up following revision(s) (requested by rmind in ticket #80):
sys/netinet6/in6_pcb.c: revision 1.129
sys/netinet/in_pcb.c: revision 1.152
in_pcbdetach: move ip_freemoptions() under softnet_lock for now (this will
be changed back once other IP paths become MP-safe).  Same for IPv6 routine.
This partially reverts 1.150 of in_pcb.c and 1.127 of in6_pcb.c changes.

(msaitoh)

Pull up following revision(s) (requested by manu in ticket #78):
lib/libperfuse/ops.c: revision 1.73
rmdir dir/.. must return an error. Use ENOTEMPTY like FFS does.

(msaitoh)

Pull up following revision(s) (requested by skrll in ticket #76):
sys/arch/evbarm/conf/RPI: revision 1.50
Copy the list of PHYs from elsewhere.
XXX sys/dev/mii/miidevices.config anyone?

(msaitoh)

Pull up following revision(s) (requested by skrll in ticket #75):
sys/external/bsd/dwc2/dist/dwc2_hcd.c: revision 1.14
sys/external/bsd/dwc2/dist/dwc2_hcdqueue.c: revision 1.11
sys/external/bsd/dwc2/dist/dwc2_hcdintr.c: revision 1.10
sys/external/bsd/dwc2/dist/dwc2_hcd.h: revision 1.9
Adapt the NAK holdoff scheme for FS/LS devices from the Raspberry Pi
Foundation driver.
Should fix PR/49019 - RPI: interrupt storm when url0 is up

(msaitoh)

add null and empty tests.

(christos)

Pull up following revision(s) (requested by he in ticket #74):
sys/ufs/chfs/chfs_vnode.c: revision 1.11
sys/ufs/chfs/chfs_readinode.c: revision 1.9
sys/ufs/chfs/chfs_scan.c: revision 1.5
sys/ufs/chfs/chfs_gc.c: revision 1.6
sys/ufs/chfs/ebh.c: revision 1.4
Plug leak in chfs_scan_eraseblock() of the allocated buffer.
Make sure to release it both on success and failure returns.
OK'ed by ttoth@
Plug memory leak in a corner case in chfs_get_data_nodes().
Plug memory leaks in error returns in chfs_readvnode().
Plug memory leak in error returns and normal operation in
chfs_gcollect_pristine().
Plug memory leak in add_peb_to_free() and add_peb_to_in_use()
in case there's a duplicate in the tree.

(msaitoh)

Timestamps are bad sources of entropy, so just use cprng_fast32.

(joerg)

Always use cprng_fast32, even during initialisation. No point in using
random(9).

(joerg)

get rid of MB_STRZ as suggested by joerg.

(christos)

PR/49185: Conrad Meyer: strvisx: Handle zero-length input strings gracefully.
(don't abuse 0 to mean compute string length internally)

(christos)

Store symlinks without a NUL terminator so that lstat(2) returns the
correct length.  Fixes the tmpfs part of PR kern/48864.

(gson)

Config glue for the Designware GMAC core.
Switch allwinner specific awge driver over to use it.

(martin)

Add work-in-progress driver for the Designware GMAC core, found on some
allwinner chips.

(martin)

Revert 1.33 and 1.34 for now, until either .INVISIBLE is reinstanciated
or random.S dropped.

(joerg)

awinrtc(4): Add support for AllWinner A10/A20 real-time clock.

(jmcneill)

awinmmc(4): Use DMA for MMC transfers.

(jmcneill)

use dma for mmc data transfers

(jmcneill)

Adapt formats for debug printfs to clock_subr type changes.

(martin)

Add awinrtc.  Hi Jared.

(skrll)

Change enum pipe to enum i915_pipe
(enum pipe uses too global of a tag and conflicts with struct pipe).

(matt)

According to http://eab.abime.net/showpost.php?p=782840&postcount=115
there are two methods for a Gayle PCMCIA soft-reset.
pccard_attach() only implemented a method with works on A600 and some
older (?) A1200 models. I added the second method as well, to make the
reset work in any case. Tested on the latest A1200 model, which failed before.

(phx)

print chip name and id on attach, and use chip id instead of cpu id to determine device attachment

(jmcneill)

add chip id functions

(jmcneill)

wrap awin_mmc_idma_descriptor in ifndef _LOCORE

(jmcneill)

add MMC IDMA descriptor

(jmcneill)

Revert all make changes except the unit tests to the state of three
weeks ago. Individual changes can be reapplied after review.

(joerg)

Add BPI to the list of kernels to build.

(skrll)

add awinrtc

(jmcneill)

a10/a20 rtc driver

(jmcneill)

replace mmc magic bits with symbolic names

(jmcneill)

Grab the DMA channel mask from the firmware.

(skrll)

DMA channel interrupt names

(skrll)

fix a stupid bug at attach

(jmcneill)

add bcmdmac0, commented out until something uses it

(jmcneill)

bcm2835 dma controller driver

(jmcneill)

add DMA0 int

(jmcneill)

Make crontab(1) work again which got broken by the last commit.

(tron)

Mark usage and quit as dead.

(joerg)

Avoid overflowing the "year" value by making the field uint64_t. Adapt
arguments and local variables accordingly.
This now fixes PR 49144 for real.

(martin)

Add a test of symlinks with long targets.

(gson)

Add a Banana Pi kernel config and fixup mmc0detect assignment for BPI.

(skrll)

Fix BOARDTYPE handling

(skrll)

in_pcbdetach: move ip_freemoptions() under softnet_lock for now (this will
be changed back once other IP paths become MP-safe).  Same for IPv6 routine.

This partially reverts 1.150 of in_pcb.c and 1.127 of in6_pcb.c changes.

(rmind)

change order of dma setup and use the correct dma channel in awinac_trigger_input

(jmcneill)

capture support

(jmcneill)

Sort sections. Remove trailing whitespace.

(wiz)

sun4v: New function prom_startcpu_by_cpuid() necessary for starting cpus on sun4v

(palle)

mmap support

(jmcneill)

initialize if_xname, if_capabilities, if_flags, if_snd

(jmcneill)

Clarify the usage message. For PR 49177.

(dholland)

On reflection change "outputname" to "headername" so it's clear it's not
the output of uuencode.

(dholland)

Clarify and generally improve. The expansion of files is not a bug,
it's a consequence of what uuencode does. Also note for a more hostile
internet environment that blindly uudecoding files without inspecting
the header is dangerous.

Pursuant to PR 49177.

(dholland)

test some pointers and return EINVAL instead of blindly assuming
they're valid.  converts kernel crashes in to app failures.

GL is still not working on evergreen for me.

(mrg)

Fix wrong use of argument identifiers; pursuant to PR 49177.

(dholland)

Initialize ec_mii

(matt)

Counting leap years was fine while we had 32bit time_t - but now it
is not a good idea for dates far away in the future.
For dates in the year 2000 or later, use arithmetic instead (since the
repeating periods are well aligned). Should fix PR 49144.

(martin)

instead of writing 0 to cfg reg, just clear the loading bit -- otherwise the dma controller might go crazy when shutting down

(jmcneill)

awindma(4): Add support for AllWinner A10/A20 DMA controller.
awinac(4): Add support for AllWinner A10/A20 audio codec.

(jmcneill)

remove duplicate awindma entry

(jmcneill)

enable audio driver now that it works

(jmcneill)

use the right dma tag

(jmcneill)

seems I have to enable ADC PA for sound to come out

(jmcneill)

use bus_dma instead of pmap_extract to get phys addr of audio buffer

(jmcneill)

add awinac_dump_regs if DDB is defined

(jmcneill)

setup DRQ_CLR_CNT reg, clear FIR_VER bit, be more relaxed about blocksize, and set SRC DRQ TYPE for dma channel

(jmcneill)

fix off by one in intr handler

(jmcneill)

add awin_dma_dump_regs func if DDB is defined, and print the irq # at attach time

(jmcneill)

store lastch to an int so that EOF is handled properly.

(christos)

Add WARNS=6

(christos)

Trailing whitespace.

(skrll)

transfer to TXDATA reg, not DMA irq en reg

(jmcneill)

dma support

(jmcneill)

Install altivec.h for ppc64 too.

(joerg)

add awindma

(jmcneill)

add driver for DMA controller

(jmcneill)

Fix previous.  There is no uap.

(nakayama)

WARNS=6

(christos)

implement WSCONSIO_[G|S]VIDEO by hijacking bcm2835_genfb's ioctl
next step: hardware cursor

(macallan)

allow board-specific ioctl() handlers

(macallan)

Rewind the file before we install it so that we can check it.
Before this fix, no checks were made and you could install crap.
XXX: pullup 7

(christos)

The OpenRISC calling standard only passes 6 parameters by register.  Any
others are passed on the stack.  NetBSD syscalls use up to 8 parameters
so that could mean two left on the stack for the kernel to access.  It's
easier for the kernel if userland could fetch into registers beforehand.
So we generate sysassym.h which contains the number of arguments for each
syscall and then use that to determine how many arguments need to be fetched
from the stack.

(matt)

When changing a directory content, update the ctime/mtime in kernel cache,
otherwise the updated ctime/mtime appears after the cached entry expire.

(manu)

rmdir dir/.. must return an error. Use ENOTEMPRY like FFS does.

(manu)

Copy the list of PHYs from elsewhere.

XXX sys/dev/mii/miidevices.config anyone?

(skrll)

make this compile again

(christos)

The comment about toxicity was correct, restore VNON setting code and
then set the proper type in lookup.

(christos)

Rename enum pipe to enum pipe so it won't conflcit with struct pipe.

(matt)

Try not to use f_data, use f_{vnode,socket,pipe,mqueue,kqueue,ksem} to get
a correctly typed pointer.

(matt)

Try not to use f_data, use f_{vnode,socket,pipe,mqueue,kqueue,ksem} to get
a correctly typed pointer.

(matt)

Don't use C++ keyword new

(matt)

Try not to use f_data, use f_{vnode,socket,pipe,mqueue,kqueue,ksem} to get
a correctly typed pointer.

(matt)

Use f_vnode instead of f_data

(matt)

Don't nest structure definitions.

(matt)

Try not to use f_data, use f_fcrypt to get a correctly typed pointer.

(matt)

Try not to use f_data, use f_rndctx to get a correctly typed pointer.

(matt)

Try not to use f_data, use f_{vnode,socket,pipe,mqueue,kqueue,ksem} to get
a correctly typed pointer.

(matt)

Try not to use f_data, use f_{vnode,socket,pipe,mqueue,kqueue,ksem} to get
a correctly typed pointer.

(matt)

Try not to use f_data, use f_{vnode,socket,pipe,mqueue,kqueue,ksem} to get
a correctly typed pointer.

(matt)

Try not to use f_data, use f_{vnode,socket,pipe,mqueue,kqueue,ksem} to get
a correctly typed pointer.

(matt)

Avoid casting f_data, make it a union of the various pointers it can be.

(matt)

RCSId police

(skrll)

Reduce pirate slang.

(wiz)

Add a commented out CWARNFLAGS to enable -Wc++-compat but don't treat them
as errors.

(matt)

Missed a C++ keyword (DEBUG case)

(matt)

Don't nest structure definitions.

(matt)

Don't use C++ keyword as variable.
Use different prefix for nd6_prefixctl members than for nd6_prefix members.

(matt)

Don't nest structure definitions.

(matt)

Don't use C++ keyword (template) as variable.

(matt)

Don't use C++ keywords (class, template) as variables

(matt)

Deanonymize structure for llinfo_arp.

(matt)

Cast return value of _NET_STAT_GETREF

(matt)

Don't use C++ new keyword

(matt)

Don't next structure and enum definitions.
Don't use C++ keywords new, try, class, private, etc.

(matt)

Don't nest struct undo and change it to sem_undo_entry

(matt)

Change rndsource_setcb to use void (*)(size_t, void *) for cb.

(matt)

const mbuftypes
explicitly cast assignments to ext_buf

(matt)

Don't nest structure definitions.

(matt)

Don't nest enum in structures.

(matt)

Add explicit casts for (const uint8_t *)

(matt)

Change type of cpu_uvm to struct uvm_cpu *

(matt)

Don't use or in __RENAME because it triggers -Wc++-compat

(matt)