Handle return value from system() properly.
PR 45672 from River Tarnell.

(dholland)

attribute format printf -> __printflike

(dholland)

Trim down MV2120:
- comment pci(4)-related drivers
- comment dmover(4)-related drivers
- optimize for size
- add 5.0 compat
- misc and cosmetics

(jakllsch)

Explain the reasoning behind the load-time memory layout.
Only build kernels the MV2120 can boot.
Use mkubootimage -E.

(jakllsch)

Adjust the quota-fetching code to allow more than two object types.
(as far as we can so far, at least)

Note that quota won't actually work fully with multiple object types
as it is, but this and the last change are a good start and do fold
together a lot of duplicated code.

(dholland)

Clean up the printout code to support more than the old fixed two
object types.

(dholland)

Move a big chunk of code out of a loop into its own function.

(dholland)

implement a rump version of kpreempt_disabled

(yamt)

Increase SYMTAB_SPACE for evbmips.

(kiyohara)

make lfs another pager specific flag so that it won't be affected by
an nfs hack in genfs.

(yamt)

g/c #if 1
comment

(yamt)

Only return values when there was no error.

(mbalmer)

On LP64 host, fss_get and fss_get50 have the same size; so FSSIOCGET50
and FSSIOCGET are identical and the compiler complains about
"duplicate case value". Guard FSSIOCGET50 with #ifndef _LP64/#endif to
fix this.

(bouyer)

One semicolon is enough.

(njoly)

Complete backward compat with NetBSd 5.x: time_t changed size, so did
struct timeval, and so did struct fss_get. So we need a compat FSSIOCGET50
ioctl.

(bouyer)

- Rework and improve TCP state tracking.
- Fix regressions after IPv6 patch merge.

Note: npfctl(8) rework will come soon.

(rmind)

Fix FSSIOCSET50: needs to use the 5.0 struct fss_set, not the current one.

(bouyer)

sys/pcq.h isn't installed to userland, so only include it ifdef _KERNEL,
fixes glitch in kdump build

(drochner)

remove the option to build this against openssl - this hasn't been used
in the NetBSD build
since the libc version of MD5Final zeroes out the context, replace
the bzero introduced in the previous commit by comments telling that

(drochner)

add known bugs
fix license (copypaste)

(ahoka)

zero out hash context after use, to avoid traces in RAM
(hint from "Solar Designer")

(drochner)

fix minor typo

(drochner)

Handle some ``Quality of Service'' configuration options, to
help devices requesting them blindly succeed in connecting.

should fix a problem analysed by Nat Sloss on current-users

(plunky)

make "rs" static -- this name is too unspecific for the global namespace

(drochner)

add missing rnd_extract->cprng_fast conversion, fixes build of
FAST_IPSEC kernels

(drochner)

Explicitly cast the result of ptoa() to unsigned long for printf,
to fix build failure on xen.

(martin)

Add the chfs module for powerpc-4xx and powerpc-booke.

(he)

Take part of the KSEG2 space and use it to "almost" direct another 256MB
of memory so that N32 kernels can make use of ram outside of KSEG0.  This
allows N32 kernels to be useful on systems with 4GB of RAM or more.

(matt)

Add a macro to allow a port to control from which freelists "normal" pages
can be allocated.

(matt)

Redefine ptoa() to be the inverse of atop.  If you were using a 32-bit vaddr_t
with 64-bit paddr_t and using managed addresses > 4GB, uvm_page_init would
silently discard the upper 32-bits of the physical address possibly double
mapping pages.

(matt)

Redefine ptoa() to be the inverse of atop.  If you were using a 32-bit vaddr_t
with 64-bit paddr_t and using managed addresses > 4GB, uvm_page_init would
silently discard the upper 32-bits of the physical address possibly double
mapping pages.

(matt)

Remove rnd_extract_data from the public kernel API (it is for use by the
stream generators only).  Clean up some related minor issues.

(tls)

Add (links for) missing "cross-references" from cprng(9).  Some of the
functions replaced by cprng were never documented at all.

(tls)

Remove quotes from "hash" and "tree" in the table examples so
the example stands a chance of actually working as written.  There appear
to be other problems, too.

(riz)

Spelling.

(wiz)

Whitespace fixes; new sentence, new line; better macro usage.
Sort SEE ALSO.

(wiz)

The secmodel(9)s init, start and stop routines are managed by each
secmodel module(7), so there is no point in calling suser/securelevel
routines from bsd44. This leads to unwanted cross-secmodel dependencies.

Do not call secmodel_bsd44_init() from secmodel_overlay_init(). Doing so
resets all curtain/securelevel values, which is not really needed when
loading an overlay filter.

Remove the secmodel_register/deregister comments, they will be
implemented differently in an upcoming patch.

ok elad@ (via private mail).

(jym)

Whitespace fixes.

(jym)

Add cprng(9) manual page, remove arc4random(9) manual page

(tls)

There is problem with 82576 chips (and probably 82575 too) with hardware vlan
tagging: some packets are sent untagged on the wire.
Follow OpenBSD and disable hardware vlan tagging for these chips
(I couldn't find a hint in other open-source drivers at what could be
wrong ...)

(bouyer)

Don't build postscreen(8) if "MKCRYPTO" is set to "no". It cannot be
built without TLS support, at least not without major surgery.
I've only tested this by building with "MKCRYPTO" set to "yes"
because the build fails much ealier otherwise.

Problem reported by Nick Hudson in private e-mail.

(tron)

Fix one last dangling use of arc4randbytes().

(tls)

comments

(yamt)

cleanup, some style and remove leftover code

(ahoka)

Kill a bunch of manual __attribute__ usages and replace them with the
equivalent sys/cdefs.h invocation.

(joerg)

PR# kern/45664: audio panic at detach

Pass the size of the structure to kmem_free, not the size of the DMA
buffer.

(jmcneill)

Remove unused header. From Henning Petersen in PR 45658.

(wiz)

New sentence, new line. Use more markup. Bump date for previous.

(wiz)

Remove arc4random() and arc4randbytes() from the kernel API.  Replace
arc4random() hacks in rump with stubs that call the host arc4random() to
get numbers that are hopefully actually random (arc4random() keyed with
stack junk is not).  This should fix some of the currently failing anita
tests -- we should no longer generate duplicate "random" MAC addresses in
the test environment.

(tls)

Change to rnd.h missed in previous commit.

(tls)

Add support for passing saved entropy (random seed file) to the kernel
from the bootloader.  This can fix the problem of poor quality keys
for other kernel modules which call arc4random() early in kernel startup
(NFS startup, in particular, causes this).

We continue to rely on the etc/rc.d/random_seed script to save entropy
to the seed file at shutdown and erase the seed file at startup.

Boot loader support implemented only for i386 and amd64 ports for now but
it should be easy for other ports to do the same or similar.

(tls)

get rid of an old merge conflict which managed to creep through

(agc)

Note kern-SHEEVAPLUG.

(sekiya)

Enable SHEEVAPLUG for build.

(sekiya)

- Add missing ENOENT or ESTALL when accessing deleted node
- Fix a warning, fix style (80 chars for a line)

(manu)

Move "unloaded module" debug message higher as it appears that name
sometimes points to memory that gets freed causing the system to crash.

(jnemeth)

SVN r220966 from FreeBSD:

Fix a corner-case of interrupt handling which resulted in potentially
spurious (and fatal) interrupt errors.

(jmcneill)

Big patch that changes the signal stack usage of urkel significantly.

Formerly, all signals came on the signal stack and the two important ones were
then forwared to either the system call or the pagefault handler. This worked
fine but the signal stack remains that, a stack. When we go multi-process this
stack gets corrupted and out-of-order with all kind of nastyness since a
userland process switch can occure when a system call is called or when a
process gets a page fault.

The new scheme only uses the signal stack as a jumpboard. It swaps states and
then returns from the signal, clearing the stack but instead of returning to
the code it now jumpt to the handler and that handler then returns to the code
when its finished.

(reinoud)

Create two panics that signal inconsistent states.

(reinoud)

Add some config settings as default and make a note that DDB is far from
working yet :-/

(reinoud)

Add thunk_sigprocmask()

(reinoud)

Use AIO by default

(reinoud)

Fix in-code declaration of the sigaction structure.

(reinoud)

Implement the choice between using AIO or not.

(reinoud)

Compile in syscall name table so urkel debugging code can print out the names
of the system calls so its more clear where it strands.`

(reinoud)

more whitespace and comment spelling fixes

(jakllsch)

whitespace consistency adjustment

(jakllsch)

Fix -Wshadow.  (should be committed before
src/external/mit/xorg/server/drivers/xf86-input-mouse/Makefile rev 1.5. Sorry)

(tsutsui)

Add the Avance Logic ALC203 code, found on e.g. alix3d3 boards.

(mbalmer)

We need a cv_destroy() here too.  Fixes LOCKDEBUG panic on interface detachment.

(jakllsch)

G/C unnecessary IPL_HARDUSB

(rmind)

Use __dead, not attribute noreturn.

(dholland)

The proper type for a variable that can be either a uid or gid is id_t,
not uint32_t.

(dholland)

No <stdbool.h> if _STANDALONE

(martin)

Typo.

(mbalmer)

If not kernel, include <stdbool.h>

(martin)

New Rawrite32 release is out.

(martin)

Typo

(skrll)

Note powsw(4) for x68k.

(isaki)

Add powsw(4) for x68k power switch.  It substitutes for a part about
power switch handler of pow(4) deleted before.
Benefits than pow(4):
- separate a front switch (= powsw0) and an EXPWON line (= powsw1)
  completely. Only powsw0 is enabled in GENERIC by default.
- prevent chattering in some hardware individuals.
  thank you for a report and a test: Yasushi Oshima and Y.Sugahara.

(isaki)

also apply the previous spin mutex hack to uaudio_trigger_input()

(mrg)

drop sc_intr_lock when calling into uaudio_chan_ptransfer().

(mrg)

re-write most of the debug messages using c99 varaidic macros and always
include the __func__ and pid:lid.  remove the handful of other __func__
uses and the 100+ other hard coded function name uses.

(mrg)

Make KERNEL_LOCKED_P() work for !defined(MULTIPROCESSOR) && !defined(_MODULE).

(jakllsch)

assert kernel_lock in usbd_transfer, usb_transfer_complete,
and usb_insert_transfer

(jmcneill)

add KERNEL_LOCKED_P() macro

(jmcneill)

Set WARNS?=2.

(tsutsui)

Remove no longer necessary -Wno-error and set WARNS?=2 .

(tsutsui)

Revert "stopcap fix" for rump by christos, which causes build failure on
most non-x86 ports and seems unnecessary. (caused by wrong rump_namei.h?)

(tsutsui)

Disable a DIAGNOSTIC test which causes kernel stack overflow on all arm ports.

Boot failure is reported from several arm users (nonaka@, ryo@, and others),
and too large auto variable in arc4_randrekey() is pointed out by enami@.

(tsutsui)

- consistency with types
- prefer sizeof(*var)

(christos)

If the ioctl to play tracks returned EINVAL, switch to digital mode and
try again.
XXX: Should we do this for the other ioctls? Should we default to digital mode?
Definitely the error message is completely unfriendly and should be fixed.

(christos)

Add _npf_config_setsubmit() function.

(rmind)

Improve wording, sort sections, remove ls(1) escapee.

(wiz)

Sort sections, remove line that must have crept in here from ls(1).

(wiz)

Various improvements.

(wiz)

Do not call cprng_fast32() before locks init. Makes rump build with
RUMP_LOCKDEBUG=yes work again.

(njoly)

Add "Novatel Wireless Expedite EU850D/EU860D/EU870D" entry.

requested by http://mail-index.netbsd.org/tech-kern/2011/11/26/msg012002.html

(nonaka)

regen.

(nonaka)

Add "Novatel Wireless Expedite EU850D/EU860D/EU870D" entry.

requested by http://mail-index.netbsd.org/tech-kern/2011/11/26/msg012002.html.
model name from http://www.nvtl.com/index.php?option=com_content&view=article&id=76

(nonaka)

clct(4)/cs4281 confirmed to work (thanks herdware@irc)

(tnn)

G/c empty #ifndef __OLD_INTERRUPT_CODE

(skrll)

man page

(yamt)

Only set symbols not provided from the linker script as regular so
that we can find their version from verdef not vertree. This bug
showed up on the hp700 build of heimdal's libkrb5, where we got
the "end" symbol with a bad version from libhx509. The hp700 linker
script defines that end symbol.

(christos)

whitespace

(yamt)

stopgap fix to avoid panic due to recursive locking if the keyboard beep
is activated through a tty (which it usually is)
IMO it was no good idea to abuse tty_lock here - it is already
problematic in the tty subsystem

(drochner)

Slight merge mistake. Now no difference to verndor branch.

(skrll)

Also add vioif(4) and viomb(4) to a set list.

(tsutsui)

update after radixtree.h api changes

(yamt)

- uvm_page_array_fill: add some more parameters
- uvn_findpages: use gang-lookup
- genfs_putpages: re-enable backward clustering
- mechanical changes after the recent radixtree.h api changes

(yamt)

Add vioif(4) and viomb(4).

(minoura)

Add missing semicolon to make this compile.

(tron)

Make sure to call pflusha before (not after) enabling MMU on
68030/68851 initialization paths.

Pointed out by isaki@ per observation of ATC status on XM6i emulator.
Tested on TT030, HP362, NWS-1750, and X68030 by me.

XXX: probably we can remove existing pflusha or TBIA calls after MMU turned
XXX: on (and before enabling cache) but I leave them as is for now to avoid
XXX: unexcpected side effects.

(tsutsui)

Use exact 32bit integral type to declare variable to store
32bit random value.  This is necessary since it is overwritten
by four bytes.

(enami)

Fix white space usage.

(enami)

avoid taking the thread lock when the kernel lock is needed.
now umidi detaches successfully.

(mrg)

convert to kmem.  also, zero all contents properly.  now umidi begins
to work.  crashed when i unplugged the synth.  for a later time.

(mrg)

nick-net80211-sync went dormant a LONG time ago.

(skrll)

harmony(4) tested on port-hp700.

(skrll)

try to speed up symbol lookup

(yamt)

man page

(yamt)

Bump (commented out) SYMTAB_SPACE.

(tsutsui)

Reorganize IPL for devices.

(tsutsui)

Use pmap_kenter_pa(9) to map msgbuf memory which is not managed by VM.

(tsutsui)

Clear innmihand flag (for recursion check) when the NMI hander is returned
so that we can abort kernel multiple times even after continue.

(tsutsui)

Move KASSERT down to the point where the endpoint is known to exist.

(joerg)

More use of uninitialized memory

(joerg)

Add Corega Ether CF-TD LAN Card.

(nonaka)

regen.

(nonaka)

Add Corega Ether CF-TD LAN Card.

(nonaka)

Mention that "pseudo-device rnd" should be removed from custom kernel
configs.  Thanks to Julio for pointing this out.

(tls)

Make the rndsave structure public -- the kernel will learn to read it
and sysinst may learn to write it (since, on some systems, most of
the keyboard input they ever get happens to be during install).  Fix a
couple of minor problems with the random_seed rc script addition.

(tls)

Match on DP83849.  From Kurt Lidl in PR 41256.

(jakllsch)

regen. (PR 41256)

(jakllsch)

Add National DP83849.  Prompted by PR 41256.

(jakllsch)

KASSERT(!mutex_owned(...)) is not kosher.

(riastradh)

aucc(4) and vidcaudio(4) confirmed to be working.  thanks frank/mike.

(mrg)

Switch hppa to gdb 7.

(skrll)

Unbreak RTLD_NEXT and co on platforms that don't inline or tail call
do_dlsym.

(joerg)

Add subr_open_disk.c for getdiskinfo(). Once we get rid of getdiskinfo,
this will not be needed.

(christos)

stopgap fix for rump build

(christos)

If we have to cast, do it properly.

(joerg)

Rename struct ufs_quota_entry -> struct quotaval.

(dholland)

Regen.

(dholland)

Revert revs. 1.78/1.79 of namei.h and apply the change to namei.src instead.
(hi christos)

(dholland)

Use CHAR_BIT; don't hardwire 8. Fix logic slightly to work with arbitrary
CHAR_BIT. Compiler output unchanged (on amd64).

(dholland)

Rollback previous change as it breaks autmates tests

(manu)

Replace if() NADA else if() NADA else continue logic with one explicit
block. Split out the MIPS handling to make it explicit and readable.

(joerg)

Add SHT_GNU_HASH.

(joerg)

Make locking in pcppi_pckbd_bell() work.  Works around locking problems
triggered when an X server rings the console bell.

(jakllsch)

fix non-DIAGNOSTIC build

(skrll)

radix_tree_gang_lookup_node and its variants: add a option to stop on a hole.

(yamt)

Replace broken memset with M_ZERO

(joerg)

Don't ignore the gpiobus_close() return value.  Found by joerg@ using clang.

(mbalmer)

fix non-DIAGNOSTIC build

(jmcneill)

gcscaudio(4) tested on port-i386.

(mbalmer)

Use __dead

(joerg)

static + __dead

(joerg)

Use static + __dead

(joerg)

remove an unnecessary call to vtbl->unlock

(jmcneill)

Spell restrict with __.

(joerg)

Replace memset with incorrect size argument with M_ZERO.

(joerg)

need to hold thread lock when using AC97_IS_* macros

(jmcneill)

When union_allocvp() finds a node being cleaned out and the caller holds
a lock, ignore the node and continue.  To allow the cleaning to succeed
the current threadmust make progress.
For a brief time the cache may contain more than one vnode referring to
a lower node.

Don't unlock the hash mutex if getnewvnode fails -- we don't hold it.

(hannken)

Don't shadow some stupid function defined globally in random platforms.

(ahoka)

Add explicit casts to make lint happy and fix the NetBSD/i386 build.

(tron)

+mount_chfs.debug

(njoly)

Correct copy/paste/modify issue in audio_suspend().

(jakllsch)

Prevent poll and kqueue on the control endpoint.
Addresses PR#33352.

(The control endpoint doesn't support pipes, so this would be of dubious
usefulness even if it didn't expose bugs.)

(jakllsch)

- accept "midi0" as an alias for "0" to the -d flag
- verify that the requested unit exists before playing
- if the verbose flag is set, print the selected device name and unit #

(jmcneill)

note umidi status.

(mrg)

vs(4) is tested working on port-x68k by isaki@.  thanks!

(mrg)

several steps towards making umidi appear to work:

- split out_solicit() into locked and unlocked front end, and use the
  unlocked version from the softintr

- kill sc_intr_lock, midi doesn't really use it (XXX finish this?
  change this? in the midi* code)

- convert tsleep()/wakeup() to cv

- move some free's out of the lock

- KASSERT() lock is held in a few more places

TODO:

- malloc -> kmem

- crashes in midiclose(), doesn't actually play things to the right
  device.  "midiplay -d midi1 -xv" plays out my midi@pcppi speaker,
  and then the above crash.  clearly something is calling the wrong
  sub-device callbacks!

(mrg)

begin to document the per-midi call backs and locking.

XXX: lots more to do.

(mrg)

add chfs.kmod

(ahoka)

add /sbin/mount_chfs

(ahoka)

add mount_chfs.8

(ahoka)