sync with HEAD

(matt)

- Support for select/poll.
- Convert pool to pool-cache.
- Wrap long lines, adjust the license.

(rmind)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

sync with HEAD

(matt)

remove myself from releng

(pavel)

Check the "shared-pins" property for the value "serdes".
Makes SBUS GEM 1000Base-SX cards attach.

(jdc)

Remove unused _BSD_OFF_T_.

(simonb)

Update for select/poll changes.

(ad)

Remove computed source files that may confuse mkdep.

(mlelstv)

Use C99 initializer.

(tsutsui)

Make <cobalt/leds.h> obsolete. (no userland stuff in it anyway)

(tsutsui)

Reorganize interrupt handling on cobalt and allow multiple interrupts on
each icu irq to support multi function PCI devices in the Qube's PCI slot:

- split out interrupt stuff from machdep.c into new interrupt.c
- prepare struct *_intrhead to handle cpu and icu interrupts
- use LIST and malloc(9) to handle multiple interrupts per icu irq
  (mostly taken from algor/pci/pcib.c and evbmips/malta/pci/pcib.c)
- move evcnt(9) stuff into *_intrhead from cobalt_intrhand
  (i.e. make it per interrupt rather than per handler)
- in pci_intr_map(9) handle CPU interrupt separately and add a sanity check
- fixup interrupt lines for devices on the PCI slot in pci_conf_interrupt(9)
- move some device address definitions into <machine/cpu.h> and
  remove <machine/leds.h>
- misc cosmetics

No objection on port-cobalt for three weeks.

(tsutsui)

4.99.57 - struct selinfo changed

(ad)

Commit the "per-CPU" select patch. This is the result of much work and
testing by rmind@ and myself.

Which approach to use is still being discussed, but I would like to get
this out of my working tree. If we decide to use a different approach
there is no problem with revisiting this.

(ad)

Cheat and add inlines for _atomic_cas_ptr() to work around gcc emitting
unneeded PIC stuff in mutex_lock() and mutex_unlock(), when a thread
register is used.

(ad)

LWP_CACHE_CREDS: instead of testing (l_cred != p_cred), use a per-LWP
flag bit to indicate a pending cred update. Avoids touching one item of
shared state in the syscall path.

(ad)

remove unnecessary include file

(plunky)

Add more assertion.

(nakayama)

Fix esym calculation to match comments.  Reported by Gao Ya'nan in
PR port-i386/38170.

(jmmv)

Remove computed source files that may confuse mkdep.

(mlelstv)

bring some stuff from time_t=64...
- add sysalign parameter to syscalls.conf
- add compat_50

(christos)

In example, change dd count size and also chmod so others can't read swap.

(reed)

Make it compile.

(ad)

vn_close no longer takes a 'struct lwp' argument.

(ad)

Regen.

(ad)

Unmark kevent/kqueue as MPSAFE. There seems to be some kind of deadlock
involving kernel_lock.

(ad)

Back out previous. It seems to expose another bug in libpthread/libc,
potentially errno being used before threading is up and running.

(ad)

Add obsolete entries for Heimdal-1.1 import.  Hi, mlelstv@!

(jmmv)

match whitespace after RCSID

(mlelstv)

NetBSD uses __RCSID

(mlelstv)

Heimdal on NetBSD uses __RCSID.

(mlelstv)

Beleatedly document the MODULAR kernel option and mk.conf's MKMODULAR.
Per hubertf@'s request.

(jmmv)

add USE_* to params

(lukem)

Import Heimdal-1.1
one more missing file

(mlelstv)

call setlocale so that nl_langinfo makes sense.

(yamt)

wrap a long line.

(yamt)

Import Heimdal-1.1
more files

(mlelstv)

Import Heimdal-1.1
more files

(mlelstv)

Import Heimdal-1.1
more missing files

(mlelstv)

Import Heimdal-1.1
Files missing from first commit.

(mlelstv)

document the import of Heimdal-1.1

(mlelstv)

Import Heimdal-1.1

(mlelstv)

Allow this to compile if LISTQ is undefined:

- Put '#ifdef LISTQ' ... '#endif' pairs around pageq_insert_head()
  and clockpro_insert_head().

- Add missing argument to printf statement.

(bjs)

Lock also on freeing own context, and add assertion.

(nakayama)

Insert a delay with enabling interrupts to avoid deadlocks
when sending IPI is failed.

(nakayama)

Revert interrupt stack changes.  There are some issues and, at least
in the current form, they complicate implementation of fast software
interrupts.  I'd rather get fast softints in now and may be revisit
interrupt stack later.

(uwe)

Catch up with recent file descriptor changes.

(ad)

fdputfile -> fd_putfile

(bjs)

Oops, delete commented out line forgotten in previous.

(uwe)

On SH4 we must be running on P2 to access memory-mapped TLB arrays.
Print DA2 again - that messenger was shot in 1.34

(uwe)

unp_gc: unlock filelist_lock in a case of restart.

(rmind)

Don't coredump on out of memory conditions. This solution leaks, but gdtoa
is too complicated to fix. Try printf %99999999999.9999999999f 2

(christos)

Er, regen...

(ad)

Regen.

(ad)

Mark kqueue/kevent MPSAFE.

(ad)

4.99.56 - file descriptor and kqueue changes

(ad)

Catch up with descriptor handling changes. See kern_descrip.c revision
1.173 for details.

(ad)

File descriptor changes, discussed on tech-kern:

- Redo reference counting to be sane.  LWPs accessing files take a short
  term reference on the local file descriptor.  This is the most common
  case.  While a file is in a process descriptor table, a reference is
  held to the file.  The file reference count only changes during control
  operations like open() or close().  Code that comes at files from an
  unusual direction (i.e. foreign to the process) like procfs or sysctl
  takes a reference on the file (f_count), and not on a descriptor.

- Remove knowledge of reference counting and locking from most code that
  deals with files.

- Make the usual case of file descriptor lookup lockless.

- Make kqueue MP and MT safe. PR kern/38098, PR kern/38137.

- Fix numerous file handling bugs, and bugs in the descriptor code that
  affected multithreaded processes.

- Split descriptor system calls out into sys_descrip.c.

- A few stylistic changes: KNF, remove unused casts now that caddr_t is
  gone. Replace dumb gotos with loop control in a few places.

- Don't do redundant pointer passing (struct proc, lwp, filedesc *) unless
  the routine is likely to be inlined.  Most of the time it's about the
  current process.

(ad)

Add five more tips.

(reed)

Move pthread__errno() into pthread_specific.c so it gets the "no stack
frame" treatment.

(ad)

Fix reference to sshd_config manpage.

(reed)

Add "netbsd-tips" fortune database with 23 initial fortunes --
actually hints or suggestions or tips on using NetBSD.
Please share your brief NetBSD usage tips.
If you want to discuss or share suggestions, have a look
at http://wiki.netbsd.se/netbsd-tips wiki page.

(reed)

use devsw_name2chr() to look up the major number of the midi device as
it is a character device

(plunky)

add devsw_name2chr() function to look up character devices

(plunky)

DCTL_NEW_MOUNT events are also placed onto a separate "mount event" list.
They are removed from this list when a DCTL_UNMOUNT event for the
corresponding mount gets sent to dctl(4).

This allows devfsd(8) to always know about devfs mounts on a system,
even if one was mounted whilst devfsd was not running.

(mjf)

when queueing a message back to Venus, prioritise signals by inserting
them at the head of the queue.

The original code had a comment saying to do this but the INSQUE did
not have any way to do that so it didn't actually

(plunky)

convert more insque/remque style queuing to use a queue(3) TAILQ,
this necessitates using a flag to indicate vc_open instead of abusing
the queue pointer but apart from that there is no functional
difference.

(plunky)

Add -s option which allows devfsd to run in the foreground and not wait
for new events. When there are no more events to be handled, it exits.

(mjf)

Put CONS* in opt_com.h

(matt)

- Use atomic_inc_uint() to maintain the global count.
- Use a barrier to sync thread start.

Ok yamt@.

(ad)

Sync to head.

(chris)

Increment libcurses minor version (termattrs() and term_attrs() added).

(jdc)

Add termattrs() and term_attrs().
Order the descriptions so that functions handling normal and functions
handling wide attributes are grouped together.

(jdc)

Add termattrs() and term_attrs().

(jdc)

Add termattrs() and term_attrs().
Conditionally define HAVE_WCHAR, so that userland programs can see wide
definitions and functions.

(jdc)

Add termattrs() and term_attrs() (requested by wiz@).
Only allow standout and underline if the terminal can do them.

(jdc)

Sync with head.

(chris)

Split device_t/softc for all hpet consumers.

(xtraeme)

Split device_t/softc for ichlpcib(4) and all hpet consumers, plus
other related cosmetic changes.

(xtraeme)

Add PB_PREV(), RB_FOREACH_REVERSE(). From FreeBSD.

(ad)

Split device_t/softc for amdpcib and the hpet attachment, plus other
related cosmetic changes.

(xtraeme)

test select(2).

(yamt)

when the char being searched for is not found strchr() must return NULL
not a pointer to the terminating '\0'. the only time we return a pointer
to the terminating '\0' is when the char being searched for is '\0'.

fixes problem observed booting -current install
http://mail-index.netbsd.org/current-users/2008/03/20/msg001445.html

(rtr)

Add alphanbsd-nat.c and it's dependencies.

(skrll)

Add missing bits of NetBSD/alpha native support.

(skrll)

default itojuns responsibilities to core

(spz)

fix typo noticed by Christoph Egger

(spz)

Don't forget to set the atac dev pointer in the softc; ahcisata at jmide
works again after that.

Also use CFATTACH_DECL_NEW, device_t and related changes for the NJMAHCI
block.

(xtraeme)

Use aprint_*_dev() instead of accessing struct device member dv_xname
directly.

(dyoung)

Use aprint_*_dev(), and device_xname().  Constify a bit.

(dyoung)

pci_activate() expects for its void * argument to be a device_t,
so change the type of the argument to device_t.  Update each use
of pci_activate().

Use device_t and accessors.  Use aprint_*_dev().

(dyoung)

Use device_t.

(dyoung)