defs.h:
- Re-order the md_*() definitions so they are in the order called
- Add some comments
- Remove obsolete md_set_no_x() definition
- Remove md_copy_filesystem()
- #if defined(DEBUG) declare backtowin(void) to avoid scattering
  externs in various md .c files

mbr.[ch]:
- Add a set_bios_geom_with_mbr_guess() which can be called by all
  the non x86 ports rather than copying the same code into arc,
  bebox, cobalt, evbmips, evbppc, hpcarm, hpcmips, landisk, ofppc,
  prep, sandpoint, and zaurus md.c

install.c:
- Remove now unnecessary call to md_copy_filesystem()

upgrade.c:
- Move move_aout_libs() here, and put under #ifdef AOUT2ELF
- Rather than having *every* md_update call wrefresh(curscr),
  wmove(stdscr, 0, 0), wclear(stdscr), and wrefresh(stdscr),
  move them here

arch/acorn26/md.c:
- Just include arch/acorn32/md.c, but add a note there warning
  about this, and while here add a note to i386/md.c that it
  is included by amd64

arch/*/md.c:
- By all means "Vive la difference", but we have 38 pairs of md.[ch],
  and they could not even agree on the order in which to list the
  md hooks, let alone formatting. Sort the md hooks to match the
  (now sane) order in def.sh, and try to normalise the formatting
- Also copy across some function level comments everywhere
- Ensure functions only used inside each md.c are defined as static
- Remove some now unused functions
- Some files had enable_rc_conf in #ifdef DEBUG. Add this to all
- bebox, evbppc and sandpoint were still playing broken games with
  copying the booted ramdisk to the target disk, the primary result
  of which was just to slow things down and ensure the target system
  lost the .profile extracted from the sets. Just kill this.
- For some ports md_update() called endwin(), and in others not.
  Take a cure from i386/amd64 and a few other more active ports,
  and update everything to match (no endwin())
- In a couple of cases correct port names in comments
- ANSIfy some lingering old style functions
- Consistently use "return 0;" rather than "return (0);"

More of the mbr code should be abstracted, along with the
get_ramsize() / set_swap() logic, but this is (more than)
enough for one day...

sysinst built for all ports but only runtime tested on amd64 & i386

(abs)

Fix start symbol. I have no idea how this worked before.

(skrll)

Switch vax to binutils 2.19.

(skrll)

arrr, implement softint_disestablish().  this code be needin' an enema, matey.

(pooka)

Add cobalt hp700 hpcarm landisk & ofppc to ALL_MACHINES list

(abs)

backout the changes that establish a workqueue to synchronize the addresses
for arg and gre because they cause a race condition by calling ioctl() during
interface initialization. To make this work correctly we would need to
synchronize all interface init routines.

(christos)

u_intNN_t -> uintNN_t

(tsutsui)

u_intNN_t -> uintNN_t

(tsutsui)

u_intNN_t -> uintNN_t

(tsutsui)

Set SAVENAME for rmdir and remove.

Addresses an easy part of PR kern/38188

(pooka)

Initialise index_gen_mtx before use.

(skrll)

Switch sparc and sparc64 to binutils 2.19.

Thanks to martin@ for testing.

(skrll)

Fix typo in name.

(tron)

Only generate SHA512 and MD5 checksums for releases.

(snj)

Include "ioconf.h" instead of extern struct cfdriver foo_cd decls.

(tsutsui)

Sync with reality.

(snj)

Update core group membership.

(snj)

Revert commit to wrong branch.  /brick snj

(snj)

merge xterm-248.

(snj)

Use common ether_crc32_le() for multicast hash.

(tsutsui)

Update for new xterm.

(snj)

merge xterm-248.

(snj)

Split device_t/softc.
Tested on TME emulating SS2.

XXX: bwtwo at obmem on TME emulating sun2 didn't match and cannot test

(tsutsui)

Split device_t/softc.
Tested on:
---
qec0 at sbus0 slot 3 offset 0x20000: 128K memory
be0 at qec0 slot 0 offset 0x0 level 4 (ipl 7): rev 1, address 00:80:17:xx:xx:xx

(tsutsui)

initial import of xterm-248

Status:

Vendor Tag: xorg
Release Tags: xterm-248

(snj)

Make bwtwo(4) match prior to genfb(4).
Required by bwtwo(4) on TME emulating SS2.

(tsutsui)

Explicitly include sys/exec_aout.h.

(mhitch)

vax macro magic requires Sysmap (untested).
XXX: I should add the blanket userspace pmap back for vax, its
pmap.h just has too many inlines and macros for this to ever work.

(pooka)

Split device_t/softc.  Tested on SS1+.

XXX: sparc/dev/cgthree.c in Attic had an attachment for cgthree* at obio?
XXX: and sparc/conf/files.sparc still has an entry of nonexistent (or not yet)
XXX: sparc/dev/cgthree_obio.c. Is there any sun4m machine which requires it?

(tsutsui)

Belated version bump for pmf(9) changes.

(dyoung)

Use the common ether_crc32_le() function for multicast hash.

(tsutsui)

Misc KNF.

(tsutsui)

Make local functions static.

(tsutsui)

Use uint8_t pointers for TX/RX buffers to avoid extra pointer casts.

(tsutsui)

Adjust attach messages.

(tsutsui)

Split device_t/softc.

(tsutsui)

Make be(4) is properly initialized at the first ifconfig(8) command:
- make beinit() and bestop() match struct ifnet
- if ether_ioctl() returns ENETRESET always call init function

(tsutsui)

Implement rename query.

(joerg)

- wrap a long line
- ANSIfy

(tsutsui)

hme_reset() is no longer called from bus attachments so make it static.

(tsutsui)

- use device_private(), proper types and variables for device_t/softc
- remove unnecessary casts against void pointer

XXX: many inconsistent use of aprint_*(9) and printf(9)

(tsutsui)

Of the arious list files 28 said
    PROG  usr/bin/gzip    usr/bin/gzcat usr/bin/gunzip
while 14 picked
    PROG  usr/bin/gzip    usr/bin/gunzip usr/bin/gzcat

Make them all choose the former. No functional change.

(abs)

From Tomas Mraz: Fix gssapi error checking.

(tteras)

From Tomas Mraz: Fix gssapi error checking.

(tteras)

Merge four separate copies of list2sh.awk runlist.sh (three identical
and one with slightly less options). Release build verified on amiga,
bebox and evbsh3. Vax build fails for unrelated reason.

(abs)

Ticket 1012.

(snj)

Apply patch (requested by mrg in ticket #958):
Missed a few bits in sparc64 xserver setlist and reachover Makefiles.

(snj)

Pull up following revision(s) (requested by tron in ticket #1012):
etc/mtree/special: revision 1.130
Remove "/etc/postfix/postfix-script" as the file was obsoleted by
the upgrade to Postfix 2.6.x.

(snj)

Not all platforms can do PaX stuff.  The list was derived in private
conversation with several developers.  The change was tested on
sparc64.

NetBSD ultra.example.ca 5.99.17 NetBSD 5.99.17 (MODULAR) #0: Tue Sep 15 19:08:02 PDT 2009  jnemeth@ultra.example.ca:/usr/local/current/objdir/sys/arch/sparc64/compile/MODULAR sparc64

NAME CLASS SOURCE REFS SIZE REQUIRES
...
exec_elf64      misc filesys 0 6981 coredump
...

(jnemeth)

As Jonathan Kollasch points out, elansc_attach() does not actually
use gba after filling it out.  Delete it.

(dyoung)

Change a few hda_error to hda_debug.  Fixes "WARNING: 4 errors while
detecting hardware; check system log." when I boot.
OK jmcneill.

(apb)

Move alpha to binutils 2.19.

(skrll)

Misc KNF and cosmetics, and remove unnecessary casts etc.

(tsutsui)

Use proper types or variables for device_t and softc.

(tsutsui)

Fix a message format in error path.

(tsutsui)

Split device_t/softc.
Tested on:
> lebuffer0 at sbus0 slot 3 offset 0x40000: 128K memory

(tsutsui)

Reflect actual status of cube-autoconf.

(cube)

Apply patch (requested by mrg in ticket #958):
Fix macppc xserver set list to reflect the fact that the imstt driver
is not built.

(snj)

Include "ioconf.h" instead of extern struct cfdriver foo_cd decls.

(tsutsui)

Remove struct sbusdev and related functions sbus_establish() and sbusreset().
They are derived from 4.4BSD/sparc and have been there since initial import
of NetBSD/sparc in 1993, but the struct sbusdev is almost unused for years,
nothing calls sbusreset(), and all (*sd->sd_reset)() functions look bogus.

Suggested by mrg@ and martin@, and tested on SS1+ and SS20.

(tsutsui)

fix a typo of register bit name.

(bsh)

Set timeout to 20 because for some reason this test hangs in qemu.
(Well, it *still* hangs even after the timeout has passed, but at
least there's multiple problems to debug now)

(pooka)

Use device_t, cfdriver_t, device_private(), device_xname(),
and appropriate types and variables for device_t/softc.

(tsutsui)

Remove "/etc/postfix/postfix-script" as the file was obsolete by
the  upgrade to Postfix 2.6.x.

(tron)

Update TODO message; we can format with meta data partitions, only not yet for
BD-R, i.e. UDF 2.60.

(reinoud)

Provide unwind log for bufq sysctls, since (theoretically) bufq might
not be initialized during kernel bootstrap and therefore "permanent"
nodes can be created only with an unwind log.

(pooka)

Add hdaudio headers

(sborrill)

Can't use CTLFLAG_PERMANENT here without providing a rollback log,
since accept filters aren't (necessarily) added during kernel boot
phase.

pointed out & tested by Geoff Wing

(pooka)

offer some actual information about what PaX might be, and provide
tips for further reading.

(plunky)

add hdaudio files.

(mrg)

Ticket 958.

(snj)

Apply patch (requested by mrg in ticket #958):
Sync X.Org reachover build framework and setlists with the new X.Org.

(snj)

Apply patch (requested by mrg in ticket #958):
Bring over X.Org updates from HEAD.

(snj)

Don't install config_activate links such as config_activate.9 ->
autoconf.9.  In the set list, mark the config_activate links
obsolete.

(dyoung)

Delete documentation for config_activate(9), it does not exist any
more.

(dyoung)

Nothing calls config_activate(9) any longer, so delete it.

(dyoung)

Nothing else handles DVACT_ACTIVATE, so why should umass(4)? Get
rid of the DVACT_ACTIVATE case in umass_activate().  This eliminates
the only call to config_activate() in the entire tree.

(dyoung)

Pass lmc(4) through 'unifdef -U__FreeBSD__ -U__bsdi__ -U__OpenBSD__
-U__linux__ -D__NetBSD__ -UNETGRAPH' to shorten this driver by
>2500 lines.  objdump -D does not reveal any differences in the
compiled code.

(dyoung)

Install hdaudio includes required by upcoming userland toolS

(sborrill)

Minor tweaks to the key generation for numeric fields.
Use 1's compliment for -ve numbers to avoid confitionals.

(dsl)

Deal with some LP64 issues with only O32 console.
Make bootloader invoke kernel compatible with both O32 and N32

(matt)

Fix various typos and grammatical errors

(sborrill)

In pmf(9), improve the implementation of device self-suspension
and make suspension by self, by drvctl(8), and by ACPI system sleep
play nice together.  Start solidifying some temporary API changes.

1. Extract a new header file, <sys/device_if.h>, from <sys/device.h> and
  #include it from <sys/pmf.h> instead of <sys/device.h> to break the
  circular dependency between <sys/device.h> and <sys/pmf.h>.

2. Introduce pmf_qual_t, an aggregate of qualifications on a PMF
  suspend/resume call.  Start to replace instances of PMF_FN_PROTO,
  PMF_FN_ARGS, et cetera, with a pmf_qual_t.

3. Introduce the notion of a "suspensor," an entity that holds a
  device in suspension.  More than one suspensor may hold a device
  at once.  A device stays suspended as long as at least one
  suspensor holds it.  A device resumes when the last suspensor
  releases it.

  Currently, the kernel defines three suspensors,

  3a the system-suspensor: for system suspension, initiated
      by 'sysctl -w machdep.sleep_state=3', by lid closure, by
      power-button press, et cetera,

  3b the drvctl-suspensor: for device suspension by /dev/drvctl
      ioctl, e.g., drvctl -S sip0.

  3c the system self-suspensor: for device drivers that suspend
      themselves and their children.  Several drivers for network
      interfaces put the network device to sleep while it is not
      administratively up, that is, after the kernel calls if_stop(,
      1).  The self-suspensor should not be used directly.  See
      the description of suspensor delegates, below.

  A suspensor can have one or more "delegates".  A suspensor can
  release devices that its delegates hold suspended.  Right now,
  only the system self-suspensor has delegates.  For each device
  that a self-suspending driver attaches, it creates the device's
  self-suspensor, a delegate of the system self-suspensor.

  Suspensors stop a system-wide suspend/resume cycle from waking
  devices that the operator put to sleep with drvctl before the cycle.
  They also help self-suspension to work more simply, safely, and in
  accord with expectations.

4. Add the notion of device activation level, devact_level_t,
  and a routine for checking the current activation level,
  device_activation().  Current activation levels are DEVACT_LEVEL_BUS,
  DEVACT_LEVEL_DRIVER, and DEVACT_LEVEL_CLASS, which respectively
  indicate that the device's bus is active, that the bus and device are
  active, and that the bus, device, and the functions of the device's
  class (network, audio) are active.

  Suspend/resume calls can be qualified with a devact_level_t.
  The power-management framework treats a devact_level_t that
  qualifies a device suspension as the device's current activation
  level; it only runs hooks to reduce the activation level from
  the presumed current level to the fully suspended state.  The
  framework treats a devact_level_t qualifying device resumption
  as the target activation level; it only runs hooks to raise the
  activation level to the target.

5. Use pmf_qual_t, devact_level_t, and self-suspensors in several
  drivers.

6. Temporarily add an unused power-management workqueue that I will
  remove or replace, soon.

(dyoung)

Replace a large number of link set based sysctl node creations with
calls from subsystem constructors.  Benefits both future kernel
modules and rump.

no change to sysctl nodes on i386/MONOLITHIC & build tested i386/ALL

(pooka)

include init_sysctl_base.c

(pooka)

call ifinit{,1}()

(pooka)

- no space needed between -I and dir per gcc(1)
- it's still better to use ${NETBSDSRCDIR} per other existing Makefiles

(tsutsui)

Chop init_sysctl into base nodes (init_sysctl_base.c) and the
kitchen sink (init_sysctl.c).  Further surgery may be needed down
the line.

(pooka)

Add comments that Makefile.inc should have shared stuff with
src/distrib/utils/x_ifconfig for install media.

(tsutsui)

sync with head

(yamt)

create interfaces only after ifnef is initialized

(pooka)

work around tap bug: if /dev/tap<n> was previously non-blocking, newly
opened fd's will also be non-blocking.
(yeayea, i'll fix the kernel some day, but I don't want to reboot
my host OS now)

(pooka)

Make cgthree(4) match prior to genfb(4). Required by cgthree(4) on SS1+.

(tsutsui)

Rewrite code which looks for an unallocated lebuffer and pairs it
with possible orphaned this le at sbus one on old PROMs not using
homegrown struct sbusdev but common cfdriver_t and device_t structures.
Now nothing refers struct sbusdev registered by sbus_establish().

Tested an SBus le/esp combo card on SS1+.

(tsutsui)

Allow for 'options ACPI_DEBUG' by providing module declarations
and using memory allocation macros instead of calling AcpiOs* stubs
directly.

(mlelstv)

The ramdisk includes dmesg, so use it in preference to /kern/msgbuf

(abs)

Remove bogus -L entry.

(haad)

Distinguish between UFS1 and UFS2 inodes by reading the ufsmount structure,
the previous heuristic of comparing the size fields of inode and dinode
failed.

(mlelstv)

switch port-sparc over to Xorg by default.

(mrg)

Tickets 940, 945, 947-949, and 952-957.

(snj)

Tickets 945, 948, 953, and 955.

(snj)

Pull up following revision(s) (requested by mhitch in ticket #957):
sys/arch/alpha/conf/GENERIC: revision 1.329
Add the I2O iop controller, as used in the ES40 and ES45.

(snj)

Pull up following revision(s) (requested by mhitch in ticket #956):
sys/arch/amiga/conf/files.amiga: revision 1.141
Switch the amiga to the common m68k/m68k/pmap_motorola.c (Finally!!).

(snj)

Pull up following revision(s) (requested by mhitch in ticket #955):
sys/arch/vax/vax/clock.c: revision 1.51
Not understanding what vax_mfpr_get_counter() was doing, my fix for
backwards time was incorrect, and actually disabled the use of mfpr for
timecounters.  The intent was to emulate a 32 bit counter using the
hardclock_ticks from the clock interrupt and the contents of the
Interval Counter register.  The problem with that was when the ICR
wrapped, but the clock interrupt was blocked resulted in an incorrect
count.  Work around this by keeping track of the previous ICR value
and hardclock_ticks to ensure the 32 bit counter doesn't go backwards.
Also, the ICR runs from -10000 to -1, so adjust the value when reading it.
Now mfpr works quit nicely on my 4000/90.

(snj)

Pull up following revision(s) (requested by mhitch in ticket #955):
sys/arch/vax/vax/clock.c: revision 1.51 via patch
Not understanding what vax_mfpr_get_counter() was doing, my fix for
backwards time was incorrect, and actually disabled the use of mfpr for
timecounters.  The intent was to emulate a 32 bit counter using the
hardclock_ticks from the clock interrupt and the contents of the
Interval Counter register.  The problem with that was when the ICR
wrapped, but the clock interrupt was blocked resulted in an incorrect
count.  Work around this by keeping track of the previous ICR value
and hardclock_ticks to ensure the 32 bit counter doesn't go backwards.
Also, the ICR runs from -10000 to -1, so adjust the value when reading it.
Now mfpr works quit nicely on my 4000/90.

(snj)

Pull up following revision(s) (requested by mhitch in ticket #954):
sys/arch/alpha/alpha/pmap.c: revision 1.242
Alpha will size the mutex structure in pmap_tlb_shootdown_q to 16 bytes.
Change the command and pad accordingly;  now pmap_tlb_shootdown_q entries
really are cache-aligned.

(snj)

Pull up following revision(s) (requested by mhitch in ticket #953):
sys/arch/alpha/alpha/pmap.c: revision 1.241
Now that secondary cpus are added the cpu_info list earlier, attempt to
send pmap tlb shootdowns to them cause the shootdown job queue to fill up,
but since the cpus aren't running yet, no IPIs get sent.  When the job
queue is full, the bit mask of cpus to send the IPI to is not set and no
shootdown IPI ever gets sent after the cpu is marked running.  Always
set the cpumask even when the queue is full.  Now I get shootdown ipis
on all the secondary cpus.

(snj)

Pull up following revision(s) (requested by mhitch in ticket #953):
sys/arch/alpha/alpha/pmap.c: revision 1.241
Now that secondary cpus are added the cpu_info list earlier, attempt to
send pmap tlb shootdowns to them cause the shootdown job queue to fill up,
but since the cpus aren't running yet, no IPIs get sent.  When the job
queue is full, the bit mask of cpus to send the IPI to is not set and no
shootdown IPI ever gets sent after the cpu is marked running.  Always
set the cpumask even when the queue is full.  Now I get shootdown ipis
on all the secondary cpus.

(snj)

Set early clobber on status.  Make sure to preserve KX.

(matt)

Pull up following revision(s) (requested by mhitch in ticket #949):
sys/arch/alpha/alpha/cpu.c: revision 1.86
There's now some per-cpu initialization that occurs before the secondary
cpus are told to begin running.  Since the seconedary cpus weren't being
added to the cpu_info list until then, that initialization wasn't being
done and resulted in crashes on the secondary cpus.  Add the secondary
cpus to the cpu_info_list after they've been started (but waiting to be
told to start running).  This fixes the problem specifically stated in
PR port-alpha/41106.  MP alphas will now at least boot and begin running,
but will eventually crash in various ways later.

(snj)

Pull up following revision(s) (requested by mhitch in ticket #949):
sys/arch/alpha/alpha/cpu.c: revision 1.86
There's now some per-cpu initialization that occurs before the secondary
cpus are told to begin running.  Since the seconedary cpus weren't being
added to the cpu_info list until then, that initialization wasn't being
done and resulted in crashes on the secondary cpus.  Add the secondary
cpus to the cpu_info_list after they've been started (but waiting to be
told to start running).  This fixes the problem specifically stated in
PR port-alpha/41106.  MP alphas will now at least boot and begin running,
but will eventually crash in various ways later.

(snj)

Pull up following revision(s) (requested by pooka in ticket #948):
sys/dev/rnd.c: revision 1.74
rnd_wakeup_readers() uses rndpool_mtx for its own consistency
management, so it cannot be called with the mutex held.  There is
no consistency requirement to synchronize over the whole add-and-wakeup
operation, as if data is consumed in the window the mutex is locked,
wakeups will simply be skipped.
fix from Juho Salminen in PR kern/42020.

(snj)

Pull up following revision(s) (requested by pooka in ticket #948):
sys/dev/rnd.c: revision 1.74
rnd_wakeup_readers() uses rndpool_mtx for its own consistency
management, so it cannot be called with the mutex held.  There is
no consistency requirement to synchronize over the whole add-and-wakeup
operation, as if data is consumed in the window the mutex is locked,
wakeups will simply be skipped.
fix from Juho Salminen in PR kern/42020.

(snj)

Pull up following revision(s) (requested by jmcneill in ticket #947):
sys/dev/pad/files.pad: revision 1.4
PR# kern/39899: audio attach with pad(4) segfaults
Make sure pad pulls in auconv mulaw support.

(snj)

Pull up following revision(s) (requested by jmcneill in ticket #945):
sys/dev/pad/padvol.c: revision 1.4
Catch up to audio(4) device_t/softc split.

(snj)

Pull up following revision(s) (requested by jmcneill in ticket #945):
sys/dev/pad/padvol.c: revision 1.4
Catch up to audio(4) device_t/softc split.

(snj)

Pull up following revision(s) (requested by jmcneill in ticket #940):
sys/arch/x86/x86/bus_space.c: revision 1.24
Fix a bug where mapping the very end of iomem accidentally returns an
address in the ISA hole (because addr+size calculations wrap to 0). Fixes
ohci on VirtualPC 7 for Mac, which places OHCI at base address 0xfffff000
size 0x1000.

(snj)

Deal with argv on LP64

(matt)

Pull up following revision(s) (requested by mrg in ticket #952):
sys/arch/dreamcast/dev/g2/g2bus_bus_mem.c: revision 1.15
sys/arch/dreamcast/include/bus.h: revision 1.15
Add an empty bus_space_mmap(9) function as a workaround
for MI pci(4) mmap function in sys/dev/pci/pci_usrreq.c.
XXX: no bus_space_subregion(9) for g2bus.

(snj)

Pull up following revision(s) (requested by mrg in ticket #952):
sys/arch/cobalt/cobalt/bus.c: revision 1.36
sys/arch/cobalt/include/bus.h: revision 1.23
Add an empty bus_space_mmap(9) function as a workaround
for MI pci(4) mmap function in sys/dev/pci/pci_usrreq.c.

(snj)

Add ELF64 config for P5064 and do the LP64 dance in mach_init.  Deal with
envp being a list of 32bit pointers.  XXX still need to do argv.

(matt)

When making a ELF64 kernel, use objcopy to make a ELF32 N32 version as well.
The loader can't tell and this avoid requiring 64bit aware bootloaders.

(matt)

When working with address preferences, sockaddr_externalize() both
addresses before comparing them.

This allows IPv6 link-local addresses (which have an embedded scope id)
to have a preference set on them.

ok dyoung

(jakllsch)

Update the NetBSD roadmaps to reflect recent changes.

Please contact core if you have other major projects and requirements
that you would like to see included.

(agc)

Simplify activation routines.  Don't block interrupts, if_deactivate()
does it.  Compile-tested, only.

(dyoung)

fix undefined return values of Read/WriteMemory, found by clang
static analyzer

(drochner)

fix undefined result of stat(), found by clang static analyzer

(drochner)

Simplify activation routines: do not call mii_activate(), it is a
no-op.  Do not block interrupts, if_deactivate() will take care of
that for us.

(dyoung)

Simplify activation routines: don't call mii_activate(), it's a
no-op.  Don't block interrupts, if_deactivate() does it.
Compile-tested, only.

(dyoung)

To ease a future device_t/softc split, se device_private() instead
of casting device_t to softc.

(dyoung)

unifdef -D__NetBSD__ -U__FreeBSD__.

No functional change intended.

(dyoung)

Mention pfsync(4) import

(degroote)

fix typo assigning bus_space_tag in rmixl_com_cnattach
set sc_dev in rmixl_com_attach

(cliff)