Create a new header lwp_private.h to contain _lwp_getprivate_fast,
_lwp_gettcb_fast, _lwp_settcb and remove them from mcontext.h, so that:
1. we don't need special hacks to hide them
2. we can include <lwp.h> where needed to get the necessary prototypes
without redefining them locally.

branches: 1.28.36;
Remove the kernel PMC code. Sent yesterday on tech-kern@.

This change:

* Removes "options PERFCTRS", the associated includes, and the associated
ifdefs. In doing so, it removes several XXXSMPs in the MI code, which is
good.

* Removes the PMC code of ARM XSCALE.

* Removes all the pmc.h files. They were all empty, except for ARM XSCALE.

* Reorders the x86 PMC code not to rely on the legacy pmc.h file. The
definitions are put in sysarch.h.

* Removes the kern/sys_pmc.c file, and along with it, the sys_pmc_control
and sys_pmc_get_info syscalls. They are marked as OBSOL in kern,
netbsd32 and rump.

* Removes the pmc_evid_t and pmc_ctr_t types.

* Removes all the associated man pages. The sets are marked as obsolete.

branches: 1.27.16; 1.27.18;
fenv for m68k

branches: 1.26.12; 1.26.30;
Retire varargs.h support. Move machine/stdarg.h logic into MI
sys/stdarg.h and expect compiler to provide proper builtins, defaulting
to the GCC interface. lint still has a special fallback.
Reduce abuse of _BSD_VA_LIST_ by defining __va_list by default and
derive va_list as required by standards.

Stop installing <machine/intr.h> on hp300. Suggested by simonb.

branches: 1.24.2;
Merge newlock2 to head.

branches: 1.23.4;
don't install <machine/db_machdep.h>, this is kernel only

No need to install autoconf.h.

branches: 1.21.4; 1.21.8; 1.21.16;
merge ktrace-lwp.

branches: 1.20.12;
Factor out W{CHAR,INT}_{MAX,MIN} into their own header file.

branches: 1.19.2;
Merge the nathanw_sa branch.

Remove KDIR=, since SYS_INCLUDE=symlinks and KDIR are not supported any more.

Implement pmc(9) -- An interface to hardware performance monitoring
counters. These counters do not exist on all CPUs, but where they
do exist, can be used for counting events such as dcache misses that
would otherwise be difficult or impossible to instrument by code
inspection or hardware simulation.

pmc(9) is meant to be a general interface. Initially, the Intel XScale
counters are the only ones supported.

branches: 1.16.8;
No point in installing svr4_machdep.h.

branches: 1.15.2; 1.15.8;
Add definitions of C99 integer format conversion macros.
XXX Fastest minimum-width integer types haven't been decided upon yet.

Add definitions of C99 specified-width integer type limits.
XXX Fastest minimum-width integer types haven't been decided upon yet.

Add definitions of C99 integer constant macros.
Tidy Makefiles up a little.

Add definitions of C99 minimum-width and greatest-width integer types.
XXX Fastest minimum-width integer types haven't been decided upon yet.

branches: 1.11.2;
Add <machine/int_types.h>, which provides namespace-pure definitions
of exact-width integer types.

Require that each each MACHINE/MACHINE_ARCH supply a lock.h. This file
contains the values __SIMPLELOCK_LOCKED and __SIMPLELOCK_UNLOCKED, which
replace the old SIMPLELOCK_LOCKED and SIMPLELOCK_UNLOCKED. These files
are also required to supply inline functions __cpu_simple_lock(),
__cpu_simple_lock_try(), and __cpu_simple_unlock() if locking is to be
supported on that platform (i.e. if MULTIPROCESSOR is defined in the
_KERNEL case). Change these functions to take an int * (&alp->lock_data)
rather than the struct simplelock * itself.

These changes make it possible for userland to use the locking primitives
by including <machine/lock.h>.

Install "machineendian_machdep.h".

C99: Define a NAN macro in <math.h> which evaulates to a constant expression of
a single-precision quiet NaN; only to be defined on platforms that do support
this value.

As per Chris G. Demetriou <cgd@netbsd.org> don't install
loadfile_machdep.h into /usr/include/hp300/, instead create a
link for machine and hp300 in ${.OBJDIR} like
sys/arch/alpha/stand/Makefile.buildboot does for alpha.

Add loadfile_machdep.h to includes to be installed - this makes
sys/arch/hp300/stand/* compile again.

branches: 1.5.2; 1.5.8;
Add COMPAT_SVR4 for m68k.

branches: 1.4.4;
Pullin m68k/ieee.h.

Move the bswap functions from libutil to libc (this bups the
minor of libc and the major of libutil). For little-endian architectures
merge the bnswap() assembly versions with nto* and hton* using symbols
aliasing. Use symbol renaming for the bswap function in this case to avoid
namespace pollution.
Declare bswap* in machine/bswap.h, not machine/endian.h. For little-endian
machines, common code for inline macros go in machine/byte_swap.h
Sync libkern with libc.
Adjust #include in kernel sources for machine/bswap.h.

Add elf_machdep.h to the INCS list.

Rework the way kernel include files are installed. In the new method,
as with user-land programs, include files are installed by each directory
in the tree that has includes to install. (This allows more flexibility
as to what gets installed, makes 'partial installs' easier, and gives us
more options as to which machines' includes get installed at any given
time.) The old SYS_INCLUDES={symlinks,copies} behaviours are _both_
still supported, though at least one bug in the 'symlinks' case is
fixed by this change. Include files can't be build before installation,
so directories that have includes as targets (e.g. dev/pci) have to move
those targets into a different Makefile.

Switch all users (except ia64) of custom machine/ansi.h to common_ansi.h

Deduplicate the code among ports and poll definitions of types
directly from a compiler.

This fixes miscompilation of certain programs that instruct compilers
to generate code for different types. This bug has been detected with
-fshort-wchar in EFI firmware.

Proposed and discussed on a mailing list (twice).

Itanium uses custom !ELF fallback switch, temporarily leave it as it is.

branches: 1.4.256;
new RCS ID format.

stubs which pint to m68k include

add rcsids to everything and clean up headers

add from net2. were symlinks to ../../vax/include/*

Remove multiple inclusion protection from files which just include
MI m68k headers and don't have any own declarations.

branches: 1.13.56; 1.13.72;
Don't set __LDPGSZ to 4096; we use 8192 like the rest of m68k-land, now.

repoint at shared aout_machdep.h, not exec.h

Use <m68k/exec.h>.

Garbage collect some stuff that isn't used. Rename the relocation info
structure to be consistent with the other m68k ports. XXX that struct needs
to move to an m68k common file.

KERNEL -> _KERNEL

new RCS ID format.

HPUXCOMPAT -> COMPAT_HPUX

Fix spelling error in Copyright notice

Put definition of __LDPGSZ and `struct relocation_info' where they belong:
in <machine/exec.h>.

Remove cpu_exec_makecmds() macro and declare it as a function. Also, wrap
everything in `#ifdef KERNEL'.

fix stupid missing ||

add support for a macro to be called from execve to check a_mid type

machine-dependent exec stuff

new RCS ID format.

stub

extent(9) -> vmem(9)

Add proper consinit(9) support for sti(4) at sgc framebuffer on hp300.

The cnattach functions for sti(4) and service switch check method
for 425e in com_frodo.c are taken from OpenBSD.
The strategy how to choose the console device in hp300_cninit() is
quite diverged from 4.4BSD and OpenBSD so it's tweaked by me.

Also put several changes in sti_sgc.c to reduce diffs from OpenBSD/hp300.

Tested on 425e and 362 (which still uses gendiofb(4), not sti(4)).

XXX: sti(4) requires uvm_km_alloc(9) and uvm_map_protect(9)
to copy and call ROM functions on the executable memory region, so
it can be called before UVM and related initializations are complete.
Probably it's time to consider about MI "deferred consinit()" API
in init_main.c (or elsewhere) for modern complicated VM system...

branches: 1.12.4; 1.12.10; 1.12.14; 1.12.18; 1.12.28;
Switch NetBSD/hp300 to wscons with rasops. Simply ported from OpenBSD/hp300.

- Only A1416 Kathmandu (topcat) framebuffer on 425t is tested, but
all other variants (TigerShark, Hyperion, DaVinci, GatorBox, Renaissance)
should also work if they are working on OpenBSD/hp300.
- sti(4) and SGC bus support are not pulled because I don't have 425e
and I can't confirm that 362 and 382 actually have SGC bus.
(I'll commit a DIO based dumb driver for 362 and 382 framebuffers later)
- Xorg server with wsfb driver will also be integrated soon
once after keycode with NoSymbol problem is addressed.
(We have to re-think what code should be used on WSDISPLAY_COMPAT_RAWKBD)
- MI HIL keyboard and mouse drivers are working fine though
cngetc via hilkbd has some problem (still we can input commands).
- No old HP-UX like HIL ioctl compatibility (we removed COMPAT_HPUX anyway).
grfinfo(8) and hilinfo(8) will be removed shortly.

Demonstrated on NetBSD booth at Open Source Conference 2011 Kagawa.

branches: 1.11.22; 1.11.28; 1.11.30;
Remove clause 3 and 4 from TNF licenses

branches: 1.10.40; 1.10.42; 1.10.44;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.9.26;
Protect against multiple inclusion.

branches: 1.8.6; 1.8.22;
Replace resource map usage with extent maps.

Update the console attach interface now that bus_space is available.
Basically means that hardware drivers with console attachments are
now largely machine independent.

branches: 1.6.26; 1.6.30;
configure() prototype is in <sys/device.h>

Fix tipo inherited from old version of TNF copyright template.

branches: 1.4.4;
Add prototypes as appropriate.

Fix logic errors in new console probe code that caused framebuffers
outside of "internal i/o" space to not even have a chance at console-hood.
Fix tested by Thorsten Frueauf <frueauf@ira.uka.de>, and offered with
my sincerest apologies that the bug ever existed in the first place.

Copyright assigned to The NetBSD Foundation.

Completely rework how the console is probed. Console probing no longer
requires pre-autoconfigured devices. Fix up some prototypes. Part of the
long journey towards new config. (GETTING THERE!)

merge ktrace-lwp.

De-__P

remove license clauses 3 and 4 from my cpoyright notices

branches: 1.4.2; 1.4.8; 1.4.32;
Nuke unused btinfo_symtab field.

Add console type for APCI.

Add a type field to btinfo_bootpath.

Add a bootinfo facility, as is used on i386 and pmax, ports. We expect
the bootinfo to be located in the first page of kernel text (which has
been zero-filled for a Very Long Time to allow us to unmap KVA 0); the
boot loader will place it there after loading the kernel image.

Include <sys/bswap.h> for function prototypes. i386, pc532 and vax
still include <machine/byte_swap.h> and define macros for some of
the bswap*() functions.

Move the bswap functions from libutil to libc (this bups the
minor of libc and the major of libutil). For little-endian architectures
merge the bnswap() assembly versions with nto* and hton* using symbols
aliasing. Use symbol renaming for the bswap function in this case to avoid
namespace pollution.
Declare bswap* in machine/bswap.h, not machine/endian.h. For little-endian
machines, common code for inline macros go in machine/byte_swap.h
Sync libkern with libc.
Adjust #include in kernel sources for machine/bswap.h.

hp300: Specify proper constraints for bus_space_read region and multi ops.

Synk with next68k.
Also use consistent register name constrains, i.e. no '%' prefix.

TAB/space cleanup.

branches: 1.22.18;
Provide a bus_space_mmap that always fails. Nothing uses it yet, but it
will be very soon.

OK tsutsui@

Remove cargo-culted '#if 0' code that was designed to produce a
compile-time error if any of the bus_space_*_8 functions was used,
but was documented that it produces a link-time error.

branches: 1.20.4;
s/impliment/implement/ in comment.

Provide PRIxBUSADDR, PRIxBUSSIZE, PRIuBUSSIZE, and PRIxBSH for all arches
to follow arm and (generic) mips.

Reviewed by christos.

branches: 1.18.40;
Add dummy bus_dma(9) typedefs (enabled by __HAVE_NO_BUS_DMA) in <sys/bus.h>
for ports which don't bother to have unnecessary bus_dma(9) implementation
to appease MI driver modules that require bus_dma_tag_t.

For amiga, move <m68k/bus_dma.h> inclusion from <machine/pci_machdep.h>
to <machine/bus.h> since amiga will want actual bus_dma(9) in future
for PCI devices.

Fixes builds on sys/modules/if_axe etc on ancient ports.

"Sounds good" from martin@ on port-m68k@.

branches: 1.17.4; 1.17.8;
Add definitions for stream methods.

Pull basic SGC bus attachment from OpenBSD to test devices.
Disabled by default since there is no SGC drivers yet.

branches: 1.15.22; 1.15.28; 1.15.30;
Remove clause 3 and 4 from TNF licenses

branches: 1.14.52; 1.14.54; 1.14.56;
__unused removal on arguments; approved by core.

branches: 1.13.14; 1.13.16;
Change "inline" back to "__inline" in .h files -- C99 is still too
new, and some apps compile things in C89 mode. C89 keywords stay.

As per core@.

branches: 1.12.2; 1.12.4; 1.12.6;
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.

merge ktrace-lwp.

branches: 1.10.10;
u_intNN_t -> uintNN_t

De-__P

TAB/space cleanup.

Modify hp300 bus_space(9) structures/functions to add hooks to override
access functions. This is required by some DIO devices to use MI drivers.

branches: 1.6.16;
Type cast for bus_space_vaddr()

branches: 1.5.2;
Implement bus_space_vaddr()

branches: 1.4.8; 1.4.10;
define a "BUS_SPACE_MAP_PREFETCHABLE" flag which basically means that
device accesses are idempotent (but should not be cached by the CPU)

Add register prefixes.

branches: 1.2.8; 1.2.10; 1.2.12;
Add a macro to check for sufficient pointer alignment in bus.h context,
"BUS_SPACE_ALIGNED_POINTER()".
Equal to the param.h "ALIGNED_POINTER()" normally, but obeys additional
requirements of the bus_space_xxx_n() macros. (BUS_SPACE_DEBUG)

bus.h for the hp300 port. Only supports bus_space for now.

Re-factor the 680x0 Function Code definitions into their own
separate file (as was done on the sun2/sun3 ports ages ago)
and switch everyone to the common header.

Change "inline" back to "__inline" in .h files -- C99 is still too
new, and some apps compile things in C89 mode. C89 keywords stay.

As per core@.

branches: 1.6.2; 1.6.4; 1.6.6;
Remove leading __ from __(const|inline|signed|volatile) -- it is obsolete.

merge ktrace-lwp.

branches: 1.4.16;
add register storage class for gcc3

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.2.2; 1.2.8;
use "static __inline" instead of "extern __inline" so that we can
build without optimization.

finish conversion to the common m68k cacheops:
add support for machine-specific flavors of the ops for off-chip caches.
many thanks to Izumi Tsutsui for his help with this.

Remove multiple inclusion protection from files which just include
MI m68k headers and don't have any own declarations.

branches: 1.1.66; 1.1.82;
This file, which will be included by <sys/cdefs.h>, will contain macros
such as __warn_references() and __weak_reference() which are actually
machine dependant. This will make it easier for ports that are being
bootstraped with ELF and ECOFF based toolchains.

This change also introduces a new macro, _C_LABEL(x). _C_LABEL expands
its argument, an identifier, to a character string of the identifier
name as it is represented in an object file.

For most ports, _C_LABEL(x) will expand to "_x", for ELF based ports
_C_LABEL(x) will expand to "x".

Remove intiolimit and define and use INTIOSIZE instead.

Largely unify the <machine/cpu.h> headers on the m68k platforms.

Consistently use "intr_depth" as the name of the interrupt depth counter
on m68k platforms.

Switch hp300 over to the common interrupt dispatch code.

XXX There are still some things to fix up here, but it's no worse
than it was before (the problems date back to when we flattened
the device interrupt levels into IPL_VM).

Switch hp300 over to the common m68k vector table.

Remove duplicated / slightly-tweaked loadustp() (load user segment table)
routines from all of the m68k ports using the shared pmap. Instead, in
pmap_init(), set up a function pointer to the appropriate mmu_load_urp*()
function in mmu_subr.s.

TAB/space cleanup.

branches: 1.72.26;
cpu_need_resched():

- Remove all code that should be MI, leaving the bare minimum under arch/.
- Make the required actions very explicit.
- Pass in LWP pointer for convenience.
- When a trap is required on another CPU, have the IPI set it locally.
- Expunge cpu_did_resched().

Refactor the 68040 writeback() from each m68k port's trap handler into a
shared copy: m68040_writeback(). It is essentially a copy of the Atari
version, with some minor cosmetic tweaks and one small performance optimization
from the mvme68k port.

Tested by rjs@ on a Quadra 950. (Thanks!)

- Cleanup for dynamic sysctl:
- Remove unused *_NAMES macros for sysctl.
- Remove unused *_MAXID for sysctls.
- Move CTL_MACHDEP sysctl definitions for m68k into m68k/include/cpu.h and
use them on all m68k machines.

branches: 1.69.14; 1.69.16;
Fix typo "one the" and architecture where appropriate.

branches: 1.68.6; 1.68.10;
make cpu_need_resched() macros consistent; __USE flags

branches: 1.67.6; 1.67.10;
Always include <m68k/cpu.h> (i.e. outside #if defined(_KERNEL))
for crash(8) as amiga and sun3. There are _KERNEL protections in it.

Briefly tested by "build.sh -m news68k -U build".
Further possible botch will be fixed later.

branches: 1.66.4; 1.66.8;
- merge and move pasted m68k MD setreg() functions into m68k/m68k_machdep.c
- move m68881_save() and m68881_restore() declarations into <m68k/m68k.h>

Briefly tested and no obvious breakage on atari, sun3, and x68k.

Remove clause 3 (UCB advertising clause) from the University of Utah
copyright. Confirmed by Mike Hibler, mike at cs.utah.edu - thanks!
Also, merge UCB and Utah copyright texts back into one, as they
originally were.

Extra verification by snj@.

branches: 1.64.2; 1.64.4;
Make sure MAXADDR is unsigned otherwise it could be mangled in shift ops
and causes silent hang on bootstrap. Analyzed on HP9000/362 I get today,
but I'm not sure how 040 lkptpa code has worked on my HP382...

Collect cpu_info and friends and move to m68k/include/cpu.h
Add a define __HAVE_CPU_DATA_FIRST which means that cpu_data is the first
member in struct cpu_info.

branches: 1.62.4;
MAXADDR used for PA == VA mapping should be page size dependent.

Remove uarea swap-out functionality:

- Addresses the issue described in PR/38828.
- Some simplification in threading and sleepq subsystems.
- Eliminates pmap_collect() and, as a side note, allows pmap optimisations.
- Eliminates XS_CTL_DATA_ONSTACK in scsipi code.
- Avoids few scans on LWP list and thus potentially long holds of proc_lock.
- Cuts ~1.5k lines of code. Reduces amd64 kernel size by ~4k.
- Removes __SWAP_BROKEN cases.

Tested on x86, mips, acorn32 (thanks <mpumford>) and partly tested on
acorn26 (thanks to <bjh21>).

Discussed on <tech-kern>, reviewed by <ad>.

Fix few more l_pflag and LP_OWEUPC confusions.

branches: 1.59.4; 1.59.14; 1.59.20; 1.59.22;
Remove CTL_MACHDEP_NAMES, it's not used anywhere.

Ok by martin@.

branches: 1.58.2; 1.58.6;
- account idepth in a clock interrupt handler too
- use idepth for CLKF_INTR()

branches: 1.57.2;
Merge the ppcoea-renovation branch to HEAD.

This branch was a major cleanup and rototill of many of the various OEA
cpu based PPC ports that focused on sharing as much code as possible
between the various ports to eliminate near-identical copies of files in
every tree. Additionally there is a new PIC system that unifies the
interface to interrupt code for all different OEA ppc arches. The work
for this branch was done by a variety of people, too long to list here.

TODO:
bebox still needs work to complete the transition to -renovation.
ofppc still needs a bunch of work, which I will be looking at.
ev64260 still needs to be renovated
amigappc was not attempted.

NOTES:
pmppc was removed as an arch, and moved to a evbppc target.

Use unsigned values for device addresses.

branches: 1.55.2; 1.55.4;
Add ci_cpuid where it's missing.

branches: 1.54.2; 1.54.4; 1.54.8;
Move declarations of m68k common functions
from <machine/cpu.h> to <m68k/m68k.h>.
While here, remove some obsolete function decls.

Add missing call to lwp_startup() in lwp_trampoline() (which was renamed
from proc_trampoline to match the other ports).

A DIAGNOSTIC kernel will now boot and run. LOCKDEBUG still doesn't work yet.
Also, my amiga no longer loses time.

Adapt hp300 to yamt-idlelwp. Tested on 9000/382.

branches: 1.51.2; 1.51.4; 1.51.10;
- add missed '*'
- move extiobase decl into <machine/cpu.h>

Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.49.2;
Remove spllowersoftclock() and CLKF_BASEPRI(), and always dispatch callouts
via a soft interrupt. In the near future, softclock will be run from process
context.

Apply changes for newslock2, taken from other m68k ports.

merge ktrace-lwp.

branches: 1.46.12;
don't expose cpu_info to userland.

move some per-cpu data definitions to MI place so that they can be modified
without touching all ports. discussed on tech-kern@.

De-__P

Rearrange process exit path to avoid need to free resources from different
process context ('reaper').

From within the exiting process context:
* deactivate pmap and free vmspace while we can still block
* introduce MD cpu_lwp_free() - this cleans all MD-specific context (such
as FPU state), and is the last potentially blocking operation;
all of cpu_wait(), and most of cpu_exit(), is now folded into cpu_lwp_free()
* process is now immediatelly marked as zombie and made available for pickup
by parent; the remaining last lwp continues the exit as fully detached
* MI (rather than MD) code bumps uvmexp.swtch, cpu_exit() is now same
for both 'process' and 'lwp' exit

uvm_lwp_exit() is modified to never block; the u-area memory is now
always just linked to the list of available u-areas. Introduce (blocking)
uvm_uarea_drain(), which is called to release the excessive u-area memory;
this is called by parent within wait4(), or by pagedaemon on memory shortage.
uvm_uarea_free() is now private function within uvm_glue.c.

MD process/lwp exit code now always calls lwp_exit2() immediatelly after
switching away from the exiting lwp.

g/c now unneeded routines and variables, including the reaper kernel thread

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.41.2;
Merge the nathanw_sa branch.

finish conversion to the common m68k cacheops:
add support for machine-specific flavors of the ops for off-chip caches.
many thanks to Izumi Tsutsui for his help with this.

merge the 12 copies of vm_machdep.c on the m68k platforms.
clean up some other stuff along the way, including:
- use m68k/cacheops.*, remove duplicates from cpu.h.
- centralize a few declarations in (all the copies of) cpu.h.
- define M68K_VAC on platforms which have a VAC.
- switch the sun platforms to the (now common) proc_trampoline().
- do the phys_map thang on the sun platforms too, no reason not to.

branches: 1.38.8;
use _KERNEL_OPT

branches: 1.37.2;
Unification of the m68k syscall() function.

Make need_resched() take a "struct cpu_info *" argument. This
causes gives a primitive form of processor affinity. Its use in
roundrobin() still needs some work.

Add missing semicolon.

branches: 1.34.2;
First sweep at scheduler state cleanup. Collect MI scheduler
state into global and per-CPU scheduler state:

- Global state: sched_qs (run queues), sched_whichqs (bitmap
of non-empty run queues), sched_slpque (sleep queues).
NOTE: These may collectively move into a struct schedstate
at some point in the future.

- Per-CPU state, struct schedstate_percpu: spc_runtime
(time process on this CPU started running), spc_flags
(replaces struct proc's p_schedflags), and
spc_curpriority (usrpri of processes on this CPU).

- Every platform must now supply a struct cpu_info and
a curcpu() macro. Simplify existing cpu_info declarations
where appropriate.

- All references to per-CPU scheduler state now made through
curcpu(). NOTE: this will likely be adjusted in the future
after further changes to struct proc are made.

Tested on i386 and Alpha. Changes are mostly mechanical, but apologies
in advance if it doesn't compile on a particular platform.

branches: 1.33.2;
Define cpu_number() as discussed on tech-smp.

synchronize types, and s/curproc/p/ in one forgotten place

mac68k part of pr 6152 fix, and missing bits for the others

Changes to support fork_kthread():
- cpu_set_kpc() now takes void *arg third argument, passed to the
entry point.
- cpu_fork() allows parent to be non-curproc iff parent is proc0.
When forking non-curproc, assume its state has already been saved.
- Adjust various pieces of machine-dependent code to account of all of this.

vm_offset_t -> {paddr_t,vaddr_t}, vm_size_t -> vsize_t

Avoid declaring the "astpending" and "want_resched" globals in cpu.h,
as this breaks C++ code that happens to indirectly include this header.
Both Matthias Scheler and I noticed this, independently.

This problem notably does not affect the atari and sun3/sun3x ports,
which have already implemented a similar solution.

Make these compile after m68k/m68k.h 1.4->1.5.

Move the prototype of hp300_calibrate_delay() to cpu.h

branches: 1.25.4;
- Move a bunch of hp300 SPU-specific stuff out of cpu.h, and place it
in new hp300spu.h file. This includes the code that associates
model number with cpu/mmu/cache combinations.
- Define "mmuid" constants for several models, derived from code
provided by Mike Smith.

Rearrange the interrupt code slightly:
- place all interrupt-related macros, prototypes, etc. into <machine/intr.h>,
like other ports.
- rename isrlink/isrunlink/isrinit, etc. intr_establish/intr_disestablish/
intr_init, etc., and move to intr.c for naming consistency.
- don't call intr_init() twice.

regdump() is prototyped in m68k/cpu.h now.

Add prototypes as appropriate.

Prototypes for functions in locore.s and machdep.c (oops, forgot to
check this in the other day...)

Use <m68k/cpu.h>.

remove the now unused MHz constants, add a declaration for vectab[],
and document some globals.

doboot() takes no arguments (from Thorsten Frueauf)

Prototype doboot() and give it the __noreturn__ attribute, so that
machdep.c will compile.

Remove deprecated cpu_setstack().

Rather than using individual HP300 model names, map those model names to
properties such as "M68040" and "M68K_MMU_MOTOROLA", used to conditionally
compile code. This is more in-line with now the other m68k ports do it,
and makes it much easier to find "common ground" among them.

In locore.s, use the MMU-type constants defined my <machine/cpu.h> rather
than "expanded" values.

Because HP300-model names are mapped to CPU/MMU combinations in
<machine/cpu.h>, no change to kernel configuration will be necessary.

remove unused cpu_exec() definitions. moved "broken swap" markers, for
ports that still need it, to types.h.

Eliminate the double mapping of the kernel stack. Enable swapping.

define BROKEN_SWAP and/or cpu_swapout as appropriate.

- added sunos_machdep.c for sun3, atari, amiga and mac68k.
- changed machdep.c and trap.c to use struct emul.
- remove ep_setup references.
- added struct emul to all emulations.

KERNEL -> _KERNEL

new RCS ID format.

Merge with 4.4-Lite.

cpu_swapin and need_proftick

Add sysctl() stuff.

Most of the changes needed to make this continue to run.

Reorder clockframe.

Add RCS identifiers (this time on the correct side of the branch), and
incorporate recent changes in netbsd-0-9 branch.

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

obsolete DB_AOUT_SYMBOLS. however, we need to leave most of the code
in db_sym.[ch] as it is used by the elf version of crash(8).

i will be cleaning up the db_sym.c code in a follow up commit to avoid
having dead code compiled.

branches: 1.4.100; 1.4.106;
Remove defines DB_ELF_SYMBOLS and DB_ELFSIZE, which are defined
in <m68k/db_machdep.h>.

XXX: should we remove DB_AOUT_SYMBOLS from <m68k/db_machdep.h>, or
XXX: should all m68k ports still have DB_AOUT_SYMBOLS?

branches: 1.3.24; 1.3.40;
Add ELF support.

Thanks to Steve Woodford for his help!

branches: 1.2.32;
new RCS ID format.

stub

Add getlabelusesmbr(), as proposed in
http://mail-index.netbsd.org/tech-userlevel/2011/08/25/msg005404.html
This is used by disk tools such as disklabel(8) to dynamically decide is
the undelyling platform uses a disklabel-in-mbr-partition or not
(instead of using a compile-time list of ports).
getlabelusesmbr() reads the sysctl kern.labelusesmbr, takes its value from the
machdep #define LABELUSESMBR.
For evbmips, make LABELUSESMBR 1 if the platform uses pmon
as bootloader, and 0 (the previous value) otherwise.

Don't use DEV_BSIZE for LABELSECTOR.
We have never supported 256byte/sector media, and
current definition won't work on >512byte/sector media.

branches: 1.2.78;
Use _HP300_FOO_H_ instead of _MACHINE_FOO_H_.

branches: 1.1.66; 1.1.82;
move arch dependence to archs. clean up a bit. deal with weird MAXPARTITONS.

Remove multiple inclusion protection from files which just include
MI m68k headers and don't have any own declarations.

branches: 1.1.48; 1.1.64;
Basic elf_machdep.h for m68k; doesn't include relocations, yet.

In the `MY THAT'S GROSS' department...
Eliminate the recursive include of machine/endian.h from sys/endian.h.

branches: 1.7.8;
Format consistency.

new RCS ID format.

stubs which pint to m68k include

Add quad stuff, etc.

get rid of #ifndef KERNEL around inclusion of cdefs.h

add rcsids to everything and clean up headers

add from net/2 sources. was symlink to ../../tahoe/include/endian.h

branches: 1.1.6;
In the `MY THAT'S GROSS' department...
Eliminate the recursive include of machine/endian.h from sys/endian.h.

moved to aout_machdep.h (via repository copy)

Use <m68k/exec.h>.

Garbage collect some stuff that isn't used. Rename the relocation info
structure to be consistent with the other m68k ports. XXX that struct needs
to move to an m68k common file.

KERNEL -> _KERNEL

new RCS ID format.

HPUXCOMPAT -> COMPAT_HPUX

Fix spelling error in Copyright notice

Put definition of __LDPGSZ and `struct relocation_info' where they belong:
in <machine/exec.h>.

Remove cpu_exec_makecmds() macro and declare it as a function. Also, wrap
everything in `#ifdef KERNEL'.

fix stupid missing ||

add support for a macro to be called from execve to check a_mid type

machine-dependent exec stuff

Re-factor the 680x0 Function Code definitions into their own
separate file (as was done on the sun2/sun3 ports ages ago)
and switch everyone to the common header.

branches: 1.1.2; 1.1.18;
fenv for m68k

Prototype __flt_rounds() consistently.

branches: 1.7.18;
Wrap with #ifndef _XXX_FLOAT_H_/#define _XXX_FLOAT_H_/ ... /#endif.

new RCS ID format.

stubs which pint to m68k include

Adjust DBL_MIN so it doesn't underflow.

Adjust DBL_MAX so it doesn't overflow.

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

new RCS ID format.

Merge with 4.4-Lite.

stubs which pint to m68k include

Add RCS identifiers (this time on the correct side of the branch), and
incorporate recent changes in netbsd-0-9 branch.

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

branches: 1.2.2; 1.2.6;
Remove clause 3 (UCB advertising clause) from the University of Utah
copyright. Confirmed by Mike Hibler, mike at cs.utah.edu - thanks!
Also, merge UCB and Utah copyright texts back into one, as they
originally were.

Extra verification by snj@.

branches: 1.1.2;
Switch NetBSD/hp300 to wscons with rasops. Simply ported from OpenBSD/hp300.

- Only A1416 Kathmandu (topcat) framebuffer on 425t is tested, but
all other variants (TigerShark, Hyperion, DaVinci, GatorBox, Renaissance)
should also work if they are working on OpenBSD/hp300.
- sti(4) and SGC bus support are not pulled because I don't have 425e
and I can't confirm that 362 and 382 actually have SGC bus.
(I'll commit a DIO based dumb driver for 362 and 382 framebuffers later)
- Xorg server with wsfb driver will also be integrated soon
once after keycode with NoSymbol problem is addressed.
(We have to re-think what code should be used on WSDISPLAY_COMPAT_RAWKBD)
- MI HIL keyboard and mouse drivers are working fine though
cngetc via hilkbd has some problem (still we can input commands).
- No old HP-UX like HIL ioctl compatibility (we removed COMPAT_HPUX anyway).
grfinfo(8) and hilinfo(8) will be removed shortly.

Demonstrated on NetBSD booth at Open Source Conference 2011 Kagawa.

s/combindations/combinations/ in comment.

Make the M68K_MMU_* defines real kernel options, available in opt_m68k_arch.h.
For 68030, 68040, and 68060, infer M68K_MMU_MOTOROLA. For 68010 and 68020, it
must be an explicit option (set in std.${MACHINE} or, in the case of hp300,
inferred from the specific model option).

Rather than testing for and defining the M680x0 CPU option in
hp300spu.h, make the individual model options (e.g. HP320, HP425,
etc.) declare a dependency on the correponding M680x0 options.
This makes them appear correctly in opt_m68k_arch.h rather than
having to pull in all of <machine/cpu.h>.

fix PR 6724 - convert m68k options to defflag's. this means that
M680[12346] are now available from opt_m68k_arch.h. FPSP meantioned
in the PR has already been fixed, and i could not find any more.

i built these kernels to ensure i did not break their builds:

amiga: GENERIC DRACO
atari: HADES FALCON MILAN-PCIIDE
mac68k: GENERIC
sun2: GENERIC
sun3: GENERIC GENERIC3X
cesfic: attempted GENERIC, does not build due to lack of machine/bus.h
hp300: GENERIC
luna68k: GENERIC
mvme68k: GENERIC
news68k: GENERIC
next68k: GENERIC
x68k: GENERIC

branches: 1.13.20; 1.13.22;
Remove clause 3 and 4 from TNF licenses

branches: 1.12.58; 1.12.60; 1.12.62;
Identify models HP362 and HP382. From OpenBSD, tested on my HP382.

Note now options HP362 and options HP382 are required for these models
in your kernel config file.

branches: 1.11.4; 1.11.8;
merge ktrace-lwp.

branches: 1.10.12;
fix typo.

branches: 1.9.6;
HP360 does not have L2 cache, so no need to define CACHE_HAVE_PAC for it.

merge the 12 copies of vm_machdep.c on the m68k platforms.
clean up some other stuff along the way, including:
- use m68k/cacheops.*, remove duplicates from cpu.h.
- centralize a few declarations in (all the copies of) cpu.h.
- define M68K_VAC on platforms which have a VAC.
- switch the sun platforms to the (now common) proc_trampoline().
- do the phys_map thang on the sun platforms too, no reason not to.

branches: 1.7.8;
use _KERNEL_OPT

branches: 1.6.2;
LKMify

branches: 1.5.2;
Add MMUIDs for the 385 and 425e, and a machine ID for the 385. From OpenBSD.

branches: 1.4.10;
Update for changes to config.

Fix tipo inherited from old version of TNF copyright template.

branches: 1.2.4;
Oops, the numbers for "t" and "s" got swapped around. No functional
difference, just correct now.

- Move a bunch of hp300 SPU-specific stuff out of cpu.h, and place it
in new hp300spu.h file. This includes the code that associates
model number with cpu/mmu/cache combinations.
- Define "mmuid" constants for several models, derived from code
provided by Mike Smith.

Change u_long to vaddr_t/vsize_t in exec code where appropriate (mostly
involves setregs and vmcmds). Should result in no code differences.

Remove clause 3 and 4 from TNF licenses

branches: 1.22.74; 1.22.76; 1.22.78;
merge ktrace-lwp.

Use _HP300_FOO_H_ instead of _MACHINE_FOO_H_.

branches: 1.20.12;
De-__P

TAB/space cleanup.

catch up with const ksiginfo_t and const sigset_t

SA_SIGINFO support for m68k (compat hpux)

branches: 1.16.2;
Back out the lwp/ktrace changes. They contained a lot of colateral damage,
and need to be examined and discussed more.

Adapt to ktrace/lwp changes.

Pass lwp pointers throughtout the kernel, as required, so that the lwpid can
be inserted into ktrace records. The general change has been to replace
"struct proc *" with "struct lwp *" in various function prototypes, pass
the lwp through and use l_proc to get the process pointer when needed.

Bump the kernel rev up to 1.6V

Merge the nathanw_sa branch.

Add kernel support for having userland provide the signal trampoline:

* struct sigacts gets a new sigact_sigdesc structure, which has the
sigaction and the trampoline/version. Version 0 means "legacy kernel
provided trampoline". Other versions are coordinated with machine-
dependent code in libc.
* sigaction1() grows two more arguments -- the trampoline pointer and
the trampoline version.
* A new __sigaction_sigtramp() system call is provided to register a
trampoline along with a signal handler.
* The handler is no longer passed to sensig() functions. Instead,
sendsig() looks up the handler by peeking in the sigacts for the
process getting the signal (since it has to look in there for the
trampoline anyway).
* Native sendsig() functions now select the appropriate trampoline and
its arguments based on the trampoline version in the sigacts.

Changes to libc to use the new facility will be checked in later. Kernel
version not bumped; we will ride the 1.6C bump made recently.

branches: 1.11.26; 1.11.30; 1.11.38;
Update for signal changes.

Fix tipo inherited from old version of TNF copyright template.

Fix execve(2) and *setregs() interfaces so emulations can set registers in a
more correct way. (See tech-kern.)

branches: 1.8.4;
Garbage-collect a bunch of _very_ hp300-specific code, most of which
doesn't work anyhow (due to current and future fundamental differences
between NetBSD and HP-UX device semantics). This can probably be shared
with other m68k ports now.

Add forward declarations for struct exec_package and struct exec_vmcmd,
to minimize the number of unnecessary #includes.

Garbage-collect hpux_dumpu().

Prototype hpux_setregs().

Define the HP-UX sigcontext structure here, and prototype hpux_sendsig().

Copyright assigned to The NetBSD Foundation.

Completely rework how the console is probed. Console probing no longer
requires pre-autoconfigured devices. Fix up some prototypes. Part of the
long journey towards new config. (GETTING THERE!)

Move the obviously machine-dependent HP-UX compat code to hpux_machdep.{c,h}.
A fair bit of this, the m68k core dump and exec goo, can probably be
made into a generic m68k hpux module, eventually.

More to be placed in hpux_machdep.c - keep your eyes peeled...

A foolish consistency.

Pullin m68k/ieee.h.

Fix typo: ieefp.h -> ieeefp.h.

Move fp_except and fp_rnd definitions to machine/include.

branches: 1.1.2; 1.1.4; 1.1.12;
Add definitions of C99 integer constant macros.

branches: 1.1.2; 1.1.4; 1.1.12;
Add definitions of C99 integer format conversion macros.
XXX Fastest minimum-width integer types haven't been decided upon yet.

branches: 1.1.2; 1.1.4; 1.1.12;
Add definitions of C99 specified-width integer type limits.
XXX Fastest minimum-width integer types haven't been decided upon yet.

branches: 1.1.2; 1.1.4; 1.1.12;
Add definitions of C99 minimum-width and greatest-width integer types.
XXX Fastest minimum-width integer types haven't been decided upon yet.

branches: 1.1.2;
Add <machine/int_types.h>, which provides namespace-pure definitions
of exact-width integer types.

Add _KERNEL guards around most of the definitions in these files.

Fix computation of the appropriate auto-vector interrupt level to use
for the DMA controller by tracking all of the DIO devices that have
"ISRPRI_BIO" interrupt handlers, and finding the highest auto-vector
level among them and informing the DMA controller.

Previously, in a post-flattening world, the DMA controller was always
told to interrupt at lev5, which is fine I guess although definitely not
ideal, but would not work on aaaancient HP320 Rev A DMA boards. To that
effect, print a warning if a Rev B DMA board ends up at ipl != 3 (we can't
differentiate between a Rev A and a Rev B board).

Switch hp300 over to the common interrupt dispatch code.

XXX There are still some things to fix up here, but it's no worse
than it was before (the problems date back to when we flattened
the device interrupt levels into IPL_VM).

hp300/intr.h: Put most of this under #ifdef _KERNEL.

Only ipl_cookie_t really needs to be exposed now, for crash(8).

hp300/intr.h: Fix missing includes.

branches: 1.34.86;
Remove trailing spaces.

branches: 1.33.4; 1.33.6;
Include <sys/evcnt.h> instead of <sys/device.h> for struct evcnt.

<sys/device.h> now requires <sys/mutex.h> for struct device_lock, and
<machine/mutex.h> requires <machine/intr.h> for ipl_cookie_t, so
including <sys/device.h> from <machine/intr.h> causes recursive dependency.

branches: 1.32.2;
Remove obsolete code for old software interrupt emulation.
The MI softint(9) handles it in the kernel threads nowadays.

Note "Emulation of VAX REI instruction" is still there to handle ASTs.

branches: 1.31.4; 1.31.6;
Make cpu_intr_p() inline.

Don't recompute an ipl level for IPL_VM during or after cpu_configure(9)
and use a static and possible highest ipl value (ipl 5) for IPL_VM
so that we can put ipl2psl_table[] conversion into makeiplcookie(9),
which could be called before cpu_configure(9).

Fixes "panic: sbappendaddr" on NFS load.

Unify splraiseipl(9) implementation among m68k ports as per
suggestion from isaki@ on port-m68k.

For hp300:
- use ipl2psl_table[] for makeiplcookie() conversion
- use uint16_t for ipl2psl_table[]

branches: 1.28.2; 1.28.4;
Remove clause 3 and 4 from TNF licenses

branches: 1.27.6; 1.27.8; 1.27.10;
- account idepth in a clock interrupt handler too
- use idepth for CLKF_INTR()

Interrupt handling changes, in discussion since February:

- Reduce available SPL levels for hardware devices to none, vm, sched, high.
- Acquire kernel_lock only for interrupts at IPL_VM.
- Implement threaded soft interrupts.

branches: 1.25.2;
Merge the ppcoea-renovation branch to HEAD.

This branch was a major cleanup and rototill of many of the various OEA
cpu based PPC ports that focused on sharing as much code as possible
between the various ports to eliminate near-identical copies of files in
every tree. Additionally there is a new PIC system that unifies the
interface to interrupt code for all different OEA ppc arches. The work
for this branch was done by a variety of people, too long to list here.

TODO:
bebox still needs work to complete the transition to -renovation.
ofppc still needs a bunch of work, which I will be looking at.
ev64260 still needs to be renovated
amigappc was not attempted.

NOTES:
pmppc was removed as an arch, and moved to a evbppc target.

Add IPL_SERIAL definition since MI com(4) now requires it for mutex(9).

branches: 1.23.8; 1.23.16; 1.23.18; 1.23.20;
- Shrink ipl_cookie_t from 32-bits to 16-bits (large enough to hold
an SR value or an IPL_* constant).
- Take advange of the smaller ipl_cookie_t to shrink kmutex_t from
16 bytes to 8 bytes by overlapping storage where possible.
- Implement a RAS-based _lock_cas() for mc68010 systems (Sun2). See
sun68k/sun68k/isr.c.

Tested on various m68k platforms, but NOT Sun2. In any case, at least
Sun2 compiles now.

branches: 1.22.2;
Remove #ifdef _KERNEL since this file is no longer exported.

Hide #include <m68k/softintr.h> inside #ifdef _KERNEL in <machine/intr.h>
to avoid exporting unnecessary files to userland.
Should fix build.sh failure, which was pointed out by isaki@.

XXX: which userland program would require contents of <machine/intr.h>?

Make hp300 use common m68k/softintr.c.

branches: 1.19.2;
Remove spllowersoftclock() and CLKF_BASEPRI(), and always dispatch callouts
via a soft interrupt. In the near future, softclock will be run from process
context.

Fix an indent in the previous.

Rename _IPL_SOFT* macro to SI_SOFT* to sync with other m68k ports.

merge yamt-splraiseipl branch.

- finish implementing splraiseipl (and makeiplcookie).
http://mail-index.NetBSD.org/tech-kern/2006/07/01/0000.html
- complete workqueue(9) and fix its ipl problem, which is reported
to cause audio skipping.
- fix netbt (at least compilation problems) for some ports.
- fix PR/33218.

branches: 1.15.20; 1.15.22;
merge ktrace-lwp.

branches: 1.14.10;
u_intNN_t -> uintNN_t

Switch DCA and APCI serial devices to use MI com(4) driver.

DCA is tested on HP362 and HP382 with serial console,
but APCI on HP4xx is not tested yet.

branches: 1.12.16;
softintr(9) interface for hp300

branches: 1.11.2; 1.11.8;
Remove the use of splimp() from the NetBSD kernel. splnet()
and only splnet() is allowed for the protection of data structures
used by network devices.

branches: 1.10.2;
Make sure everybody has an splvm() and equate it with splimp() (splimp()
is the historical name for this interrupt level, and the historical name
is going to go away in the near future).

Add spllock(). See spl(9) for details.

Make sure we provide splsched() as described in spl(9).

branches: 1.7.2;
Change the semantics of splsoftclock() to be like other spl*() functions,
that is priority is rasied. Add a new spllowersoftclock() to provide the
atomic drop-to-softclock semantics that the old splsoftclock() provided,
and update calls accordingly.

This fixes a problem with using the "rnd" pseudo-device from within
interrupt context to extract random data (e.g. from within the softnet
interrupt) where doing so would incorrectly unblock interrupts (causing
all sorts of lossage).

XXX 4 platforms do not have priority-raising capability: newsmips, sparc,
XXX sparc64, and VAX. This platforms still have this bug until their
XXX spl*() functions are fixed.

Store computed PSL values in an array indexed by abstract interrupt
level, rather than individual shortwords.

Slightly rearrange the SSIR inline assembly implementation in order to avoid
making assumptions on C symbol name prefixes.

branches: 1.4.10;
Spell `interrupt' correctly.

Fix tipo inherited from old version of TNF copyright template.

branches: 1.2.2;
Use optimized _splraise() implementation, from mac68k/intr.h

Rearrange the interrupt code slightly:
- place all interrupt-related macros, prototypes, etc. into <machine/intr.h>,
like other ports.
- rename isrlink/isrunlink/isrinit, etc. intr_establish/intr_disestablish/
intr_init, etc., and move to intr.c for naming consistency.
- don't call intr_init() twice.

Remove multiple inclusion protection from files which just include
MI m68k headers and don't have any own declarations.

branches: 1.1.56; 1.1.72;
kcore.h for hp300 - just pulls in the common m68k kcore.h

- Normalize inclusion protection (remove)
- Move CHAR_{MIN,MAX} to a common file.
- Fix broken comments

branches: 1.9.112; 1.9.116;
Remove multiple inclusion protection from files which just include
MI m68k headers and don't have any own declarations.

branches: 1.8.48; 1.8.64;
multiple include protect machine/limits.h, fixes pr 4473 (from Mika Nystrom)

new RCS ID format.

stubs which pint to m68k include

Make sure min values are correct when used at a larger size.

Add quad support.

Removed CLK_TCK, it's value is not machine dependant.
It's supposed to be defined in <time.h>, not <limits.h> anyway.

add rcsids to everything and clean up headers

add from net2. were symlinks to ../../vax/include/*

Consistently define WARN in a way that passes format string checks, i.e.
always uses the same number of arguments as given in the format string.

branches: 1.7.44; 1.7.60;
Remove clause 3 and 4 from TNF licenses

branches: 1.6.72; 1.6.74; 1.6.76;
free -> dealloc
unsigned -> size_t for alloc/dealloc

branches: 1.5.2;
merge ktrace-lwp.

branches: 1.4.12;
De-__P

branches: 1.3.2; 1.3.18;
Split elf32 vs. elf64 handling in loadfile(), including support for
each separately and explicitly. BOOT_ELF is now BOOT_ELF32 and
BOOT_ELF64, and ELFSIZE should no longer be defined in loadfile_machdep.h.

branches: 1.2.4; 1.2.6;
Convert both bootblocks and mkboot to use libsa's loadfile(), and make
both ELF aware.

branches: 1.1.2; 1.1.8;
First cut at loadfile() support for the hp300 boot program. Doesn't
work yet (kernel fails before MMU can be enabled).

branches: 1.1.6;
Require that each each MACHINE/MACHINE_ARCH supply a lock.h. This file
contains the values __SIMPLELOCK_LOCKED and __SIMPLELOCK_UNLOCKED, which
replace the old SIMPLELOCK_LOCKED and SIMPLELOCK_UNLOCKED. These files
are also required to supply inline functions __cpu_simple_lock(),
__cpu_simple_lock_try(), and __cpu_simple_unlock() if locking is to be
supported on that platform (i.e. if MULTIPROCESSOR is defined in the
_KERNEL case). Change these functions to take an int * (&alp->lock_data)
rather than the struct simplelock * itself.

These changes make it possible for userland to use the locking primitives
by including <machine/lock.h>.

branches: 1.1.4;
Create a new header lwp_private.h to contain _lwp_getprivate_fast,
_lwp_gettcb_fast, _lwp_settcb and remove them from mcontext.h, so that:
1. we don't need special hacks to hide them
2. we can include <lwp.h> where needed to get the necessary prototypes
without redefining them locally.

branches: 1.1.6;
C99: Define a NAN macro in <math.h> which evaulates to a constant expression of
a single-precision quiet NaN; only to be defined on platforms that do support
this value.

Merge the nathanw_sa branch.

branches: 1.1.2;
file mcontext.h was initially added on branch nathanw_sa.

GC this file. It has been unused for quite a while.

new RCS ID format.

Add RCS identifiers (this time on the correct side of the branch), and
incorporate recent changes in netbsd-0-9 branch.

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

branches: 1.2.4;
Merge newlock2 to head.

branches: 1.1.2;
file mutex.h was initially added on branch newlock2.

proper sizing of kmem_arena on different ports

PR port-i386/45946: Kernel locks up in VMEM system

branches: 1.52.4; 1.52.8;
Remove clause 3 (UCB advertising clause) from the University of Utah
copyright. Confirmed by Mike Hibler, mike at cs.utah.edu - thanks!
Also, merge UCB and Utah copyright texts back into one, as they
originally were.

Extra verification by snj@.

branches: 1.51.4; 1.51.6; 1.51.8;
Refactoring MD segment related definitions on m68k ports:
- move macro SEGSHIFT, NBSEG, and SEGOFSET from <machine/param.h>
to <m68k/pmap_motorola.h> where they should belong
- also simplify NBSEG macro for both 020/030 and 040/060 cases
- also move m68k_trunc_seg(), m68k_round_seg(), and m68k_seg_offset()
(that use SEGOFSET) from <m68k/param.h> to <m68k/pmap_motorola.h>
- add comments about each segment size value used on pmap_motorola
implementation
- add TIA_SIZE, TIA_SHIFT, TIB_SIZE, and TIB_SHIFT macro which can be
used for ste/pte index sizes
- use proper vaddr_t cast instead of unsigned for SEGOFSET/PGOFSET macro

Tested on atari, hp300, and news68k.

Fix wrong NBSEG values for all hp300 pmap derived m68k ports.
They were incorrect since 1997 on amiga and atari, and since 2002
on other ports, but fortunately they don't look so fatal.

Anyway, these values will be moved into <m68k/pmap_motorola.h> soon
since they are quite pmap implementation dependent.

branches: 1.49.10;
Remove COMPAT_HPUX.

branches: 1.48.20; 1.48.26; 1.48.32;
Merge #ifdef _KERNEL stuff into one place.

branches: 1.47.26;
Use _HP300_FOO_H_ instead of _MACHINE_FOO_H_.

Patterned after next68k, omit inclusion of <sys/device.h> if we
are not in _KERNEL. This should work around our lint problems
related to "unspecified-size array at end of struct" constructs.

branches: 1.45.10; 1.45.12; 1.45.14;
De-__P

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22364, verified by myself.

branches: 1.43.2;
Merge the nathanw_sa branch.

merge pmap.h for all 4k-page motorola-MMU (or compatible) m68k platforms.

branches: 1.41.8;
use _KERNEL_OPT

branches: 1.40.6;
Update for the NKMEMPAGES changes.

CL* discarding.

branches: 1.38.14; 1.38.20;
defopt COMPAT_HPUX

Move the definition of MSGBUFSIZE up to the machine-arch level if
possible. Pointed out by Bernd Ernesti.

Implement the kernel part of pr-1891. This allows for a more flexible sized
msgbuf. Note that old 'dmesg' and 'syslogd' binaries will continue running,
though old 'dmesg' binaries will output a few bytes of junk at the start of
the buffer, and will miss a few bytes at the end of the buffer.

branches: 1.35.2;
Restore the defines of UPAGES and NPTEPG.

The 'Mach derived conversion macros' are now in <m68k/param.h>

Use the MI <m68k/param.h> include.

Rearrange the interrupt code slightly:
- place all interrupt-related macros, prototypes, etc. into <machine/intr.h>,
like other ports.
- rename isrlink/isrunlink/isrinit, etc. intr_establish/intr_disestablish/
intr_init, etc., and move to intr.c for naming consistency.
- don't call intr_init() twice.

Add prototypes as appropriate.

Change MID_MACHINE to MID_M68K; we now use the same executable format as
the rest of the m68k ports.

Define ALIGNED_POINTER

branches: 1.28.4;
- Implement _splraise(), like _spl(), but only modifies the SR if the
new ipl is higher than the ipl currently set in SR. (Thanks to
scottr@netbsd.org for implementing it while I was busy with other stuff!)
- Define spl{bio,net,tty,imp} in terms of _splraise().
- Move definition of splhil() here, and define splkbd() as an alias for it.

Dynamically compute the PSL value for spl{bio,net,tty,imp}() calls.

Changes for the new delay() function.

add _MACHINE and _MACHINE_ARCH, which are like MACHINE and MACHINE_ARCH,
execpt without quotes. meant to be __CONCAT()ted for easy #includes
of machine-dependent headers for MI code (e.g. for the MI ISA/EISA/PCI/TC
bus code).

LOCORE -> _LOCORE

Replace splnet() with splsoftnet(). Add splnet().

kill user-land DELAY macro, as suggested by gwr.

clean up definitions of ctod(), dtoc(), ctob(), btoc(), btodb(), and
dbtob(). remove unused definitions of ctos(), stoc(), and dtob().

KERNEL -> _KERNEL

don't be so quick to cast macro arguments to unsigned; if the
arguments are really off_t's (e.g. to btodb), then you can lose
information. This was the "> 4G file systems don't work" bug; physio
uses btodb, which was broken.

Make these consisteny with i386. Mostly cosmetic.

new RCS ID format.

move arch dependence to archs. clean up a bit. deal with weird MAXPARTITONS.

asm -> __asm (PR #463).

Nuke struct ste, struct ste40, struct pte, and HP_PAGE_SIZE. Use my
pv_entry allocator. Change pa_index() to pmap_page_index() for consistency
with other ports.

Add USPACE.

Merge with 4.4-Lite.

Kill stupid #ifndef.

splstatclock

HPUXCOMPAT -> COMPAT_HPUX

define disklabel location.

add MACHINE_ARCH

Set MID_MACHINE to MID_M68K4K.

Add ALIGNBYTES.

Add RCS identifiers (this time on the correct side of the branch), and
incorporate recent changes in netbsd-0-9 branch.

add MID_MACHINE #defined to be MID_HP300

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

Need a common <m68k/pcb.h> for gdb.

Eliminate the double mapping of the kernel stack. Enable swapping.

new RCS ID format.

Merge with 4.4-Lite.

struct md_coredump

Update to match proc.h.

Add RCS identifiers (this time on the correct side of the branch), and
incorporate recent changes in netbsd-0-9 branch.

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

Since we're batting 1.000 on "platforms that adopt new-pmap_68k that also
use machine_bootmap[]", just remove the option (it looks like it would
be opted into on every platform).

No need to explicitly include mmu_{30,40}.h here; hp300 does not use the
TT register for anything.

Enable pmap_prefer() on hp300 systems to handle the 320/350 VAC.

The hp300 reserves the last page of RAM to use for the MMU trampoline.
Since this page would otherwise be wasted, put it to good use as the
NULL segment table using a new NULL_SEGTAB_PA hook.

The Hibler/Utah pmap put the virtual kernel PTE array near the top of
the address space because it needed to use the very last page of kernel
virtual space to map the last page of RAM VA==PA to faciliate MMU on/off
transitions, and doing so made for more efficient use of kernel PT pages.
We use the SYSMAP_VA hook to do the same for the same reason.

(The *2 is because we can't use the very top of the address space
for the virtual kernel PTE array because MAXADDR is otherwise occupied.)

First step to make hp300 port the new m68k pmap ready:

- Add new pmap files to config(5) file definitions
- Add commented out options __HAVE_NEW_PMAP_68K in conf/std.hp300
- Adjust <machine/pmap.h> and <machine/pte.h> for new pmap
initialization ops etc.
- Update <machine/vmparam.h> for the new pmap specific address layout
- Allocate intio, extio, and bootinfo_va space using the pmap_bootmap
structure and initialize va variables in hp300_init() in
__HAVE_NEW_PMAP_68K case

On hp300 the kernel is not mapped VA==PA so more work is necessary
to turn on/off the MMU (on-going).

merge pmap.h for all 4k-page motorola-MMU (or compatible) m68k platforms.

it really helps to get the stub right before cutting + pasting it 27 times.
alas, I did not. doh.

add pmap_remove_all() hook (empty on most platforms so far).

branches: 1.32.4;
Update pmap_update to now take the updated pmap as an argument.
This will allow improvements to the pmaps so that they can more easily defer expensive operations, eg tlb/cache flush, til the last possible moment.

Currently this is a no-op on most platforms, so they should see no difference.

Reviewed by Jason.

branches: 1.31.2; 1.31.4;
replace vm_page_t with struct vm_page *.

replace {simple_,}lock{_data,}_t with struct {simple,}lock {,*}.

Garbage-collect pmap_page_index().

Undo a misguided previous change to the pmap_update() API.

Give pmap_update() an argument (a pmap_t) so that it knows which
pmap it should be updating.

#define away pmap_update() in <machine/pmap.h> so that no function
call overhead is incurred as we start sprinkling pmap_update() calls
throughout the source tree (no pmaps currently defer operations, but
we are adding the infrastructure to allow them to do so).

branches: 1.25.8; 1.25.20;
hp300 part of the fix for pr 6152

Implement pmap_prefer() for systems w/ an HP MMU (which has a virtually-tagged
cache). This helps to avoid cache alias problems, and can improve performance
in the case where physical pages have multiple mappings (since the pages will
not have to be marked cache-inhibited).

vm_offset_t -> {paddr_t,vaddr_t}, vm_size_t -> vsize_t

GC the old MACHINE_NONCONTIG code.

Move pmap_map() function definition to MD headers, as appropriate. It's
an internal function, and the VM system shouldn't try to prototype it.
(Note that some ports _don't_ prototype it.)

pmap_page_index() is not used in the MACHINE_NEW_NONCONTIG case.

Implement MACHINE_NEW_NONCONTIG and switch the hp300 port to use it.

Garbage-collect pm_stchanged; it's not used by anything.

Garbage-collect use of the PCB's copy of the user segment table pointer.

Now that all ports have pmap_activate(), and it has an identical interface,
prototype it in <vm/pmap.h>

Change an argument name to PMAP_ACTIVATE() to more accurately describe the
semantics of the argument.

Make pmap_activate() and pmap_deactivate() take a struct proc *.

s/hp300_btop/m68k_btop/

Make active_user_pmap() DTRT while a process is exiting.

Add an active_user_pmap() macro, and use it as appropriate.

Rather than using individual HP300 model names, map those model names to
properties such as "M68040" and "M68K_MMU_MOTOROLA", used to conditionally
compile code. This is more in-line with now the other m68k ports do it,
and makes it much easier to find "common ground" among them.

In locore.s, use the MMU-type constants defined my <machine/cpu.h> rather
than "expanded" values.

Because HP300-model names are mapped to CPU/MMU combinations in
<machine/cpu.h>, no change to kernel configuration will be necessary.

KERNEL -> _KERNEL

Bring back pmap_kernel(), for now always inlined as a pointer to
kernel_pmap_store.

KERNEL -> _KERNEL

new RCS ID format.

Nuke struct ste, struct ste40, struct pte, and HP_PAGE_SIZE. Use my
pv_entry allocator. Change pa_index() to pmap_page_index() for consistency
with other ports.

Merge with 4.4-Lite.

cast

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

Remove the kernel PMC code. Sent yesterday on tech-kern@.

This change:

* Removes "options PERFCTRS", the associated includes, and the associated
ifdefs. In doing so, it removes several XXXSMPs in the MI code, which is
good.

* Removes the PMC code of ARM XSCALE.

* Removes all the pmc.h files. They were all empty, except for ARM XSCALE.

* Reorders the x86 PMC code not to rely on the legacy pmc.h file. The
definitions are put in sysarch.h.

* Removes the kern/sys_pmc.c file, and along with it, the sys_pmc_control
and sys_pmc_get_info syscalls. They are marked as OBSOL in kern,
netbsd32 and rump.

* Removes the pmc_evid_t and pmc_ctr_t types.

* Removes all the associated man pages. The sets are marked as obsolete.

branches: 1.1.2; 1.1.4; 1.1.6; 1.1.202; 1.1.204;
Implement pmc(9) -- An interface to hardware performance monitoring
counters. These counters do not exist on all CPUs, but where they
do exist, can be used for counting events such as dcache misses that
would otherwise be difficult or impossible to instrument by code
inspection or hardware simulation.

pmc(9) is meant to be a general interface. Initially, the Intel XScale
counters are the only ones supported.

m68k syscall rototill:

- Switch all m68k-based ports over to __HAVE_SYSCALL_INTERN.
- Add systrace glue.
- Define struct mdproc in <m68k/proc.h> instead of <machine/proc.h>.
(They were all defined exactly the same anyway, other than a couple
of the MDP_* flags.)

branches: 1.8.30; 1.8.34; 1.8.42;
Add missing definition needed for COMPAT_SUNOS

branches: 1.7.4;
Nuke MDP_HPUXTRACE.

new RCS ID format.

MDP_AST is deprecated.

branches: 1.4.2;
Merge with 4.4-Lite.

Most of the changes needed to make this continue to run.

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

new RCS ID format.

For 4.4 profiling stuff.

new RCS ID format.

Merge with 4.4-Lite.

stubs which pint to m68k include

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

First step to make hp300 port the new m68k pmap ready:

- Add new pmap files to config(5) file definitions
- Add commented out options __HAVE_NEW_PMAP_68K in conf/std.hp300
- Adjust <machine/pmap.h> and <machine/pte.h> for new pmap
initialization ops etc.
- Update <machine/vmparam.h> for the new pmap specific address layout
- Allocate intio, extio, and bootinfo_va space using the pmap_bootmap
structure and initialize va variables in hp300_init() in
__HAVE_NEW_PMAP_68K case

On hp300 the kernel is not mapped VA==PA so more work is necessary
to turn on/off the MMU (on-going).

merge pte.h for m68k w/ motorola or compatible MMU
(only 4k-page platforms so far).

branches: 1.4.50;
new RCS ID format.

Nuke struct ste, struct ste40, struct pte, and HP_PAGE_SIZE. Use my
pv_entry allocator. Change pa_index() to pmap_page_index() for consistency
with other ports.

Merge with 4.4-Lite.

Move pte.h to /usr/include/machine, like it is in other ports.

new RCS ID format.

stubs which pint to m68k include

Remove multiple inclusion protection from files which just include
MI m68k headers and don't have any own declarations.

branches: 1.9.66; 1.9.82;
new RCS ID format.

HP300 -> MACHINE

stubs which pint to m68k include

Still need old #defines for now.

Implement `struct reg'.

Move #define of NIPCREG outside #fidef.

Ditto.

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

branches: 1.2.4;
Merge newlock2 to head.

branches: 1.1.2;
file rwlock.h was initially added on branch newlock2.

make the definition of _JBLEN mach. dep. header-dependent.

Minor change.

new RCS ID format.

stubs which pint to m68k include

md signal context.

Retire varargs.h support. Move machine/stdarg.h logic into MI
sys/stdarg.h and expect compiler to provide proper builtins, defaulting
to the GCC interface. lint still has a special fallback.
Reduce abuse of _BSD_VA_LIST_ by defining __va_list by default and
derive va_list as required by standards.

Standardize format.

new RCS ID format.

stubs which pint to m68k include

add rcsids to everything and clean up headers

add from net2. were symlinks to ../../vax/include/*

branches: 1.2.18;
merge ktrace-lwp.

branches: 1.1.6;
new file

Remove compat_svr4 and compat_svr4_32, as discussed on tech-kern@ recently,
but also as discussed several times in the past.

branches: 1.3.40; 1.3.218;
Rip protection against multiple inclusion from wrapper headers.

branches: 1.2.2; 1.2.4; 1.2.6; 1.2.10;
Pull in the right generic m68k header. (Where was my mind?)

Add COMPAT_SVR4 for m68k.

new RCS ID format.

Merge with 4.4-Lite.

stubs which pint to m68k include

Add RCS identifiers (this time on the correct side of the branch), and
incorporate recent changes in netbsd-0-9 branch.

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

I am giving up attempting to bend the kernel build machinery to my will
so that assym.h can be safely included by libkern assembly sources. So,
instead, redundantly define __HAVE_M68K_BROKEN_RMC in multiple places so
that all the things that need to key off of it can.

Introduce __HAVE_M68K_BROKEN_RMC, which can be defined by m68k platforms
that cannot perform indivisible READ-MODIFY-WRITE bus cycles. The
define is named for the /RMC output signal on the CPU that is asserted
when such a cycle is performed. These cycles are only used by the CAS,
CAS2, and TAS instructions. The cycles are, as far as I can tell,
otherwise indistinguishable from garden-variety read followed by write
bus cycles, except for the assertion of /RMC. Note that other single-
instruction "read-modify-write" instructions (e.g. OR, AND, ADD, etc.
with a memory destination operand) do not assert /RMC, and are thus not
considered truly atomic from the perspective of the 68k architecture.

When __HAVE_M68K_BROKEN_RMC is defined, then we must make unavailable
the CAS, CAS2, and TAS instructions, because using them will typically
cause a bus error to occur. In this case, we handle _atomic_cas_32()
(and its aliases) in the kernel environment the same way as the 68010
does: a restartable atomic sequence. 8- and 16-bit CAS operations are
not available in the kernel environment on such platforms.

We define __HAVE_M68K_BROKEN_RMC for the hp300 platform because 68020-based
hp300 systems will generate a fault for /RMC-type cycles.

Based on work by Izumi Tsutsui.

XXX pullup-11?

Don't need to define __HAVE_RAS here; it's already defined by <m68k/types.h>.

Document via __HAVE_BUS_SPACE_8 platforms that implement bus_space_*_8

branches: 1.21.66;
Welcome to 5.99.53! Merge rmind-uvmplock branch:

- Reorganize locking in UVM and provide extra serialisation for pmap(9).
New lock order: [vmpage-owner-lock] -> pmap-lock.

- Simplify locking in some pmap(9) modules by removing P->V locking.

- Use lock object on vmobjlock (and thus vnode_t::v_interlock) to share
the locks amongst UVM objects where necessary (tmpfs, layerfs, unionfs).

- Rewrite and optimise x86 TLB shootdown code, make it simpler and cleaner.
Add TLBSTATS option for x86 to collect statistics about TLB shootdowns.

- Unify /dev/mem et al in MI code and provide required locking (removes
kernel-lock on some ports). Also, avoid cache-aliasing issues.

Thanks to Andrew Doran and Joerg Sonnenberger, as their initial patches
formed the core changes of this branch.

branches: 1.20.32; 1.20.42;
Now that __HAVE_TIMECOUNTER and __HAVE_GENERIC_TODR are invariants,
remove the conditionals and the code associated with the undef case.

branches: 1.19.2; 1.19.8;
Merge the ppcoea-renovation branch to HEAD.

This branch was a major cleanup and rototill of many of the various OEA
cpu based PPC ports that focused on sharing as much code as possible
between the various ports to eliminate near-identical copies of files in
every tree. Additionally there is a new PIC system that unifies the
interface to interrupt code for all different OEA ppc arches. The work
for this branch was done by a variety of people, too long to list here.

TODO:
bebox still needs work to complete the transition to -renovation.
ofppc still needs a bunch of work, which I will be looking at.
ev64260 still needs to be renovated
amigappc was not attempted.

NOTES:
pmppc was removed as an arch, and moved to a evbppc target.

branches: 1.18.10;
Generic soft interrupts are mandatory.

branches: 1.17.10; 1.17.18;
Switch hp300 to timecounter with MC6840 interval timer, which is
used for hardclcok(9). Old code is still left for reference.

XXX: possibly problematic if hardclock(9) is blocked more than 1/HZ,
XXX: but old microtime(9) implementation also had the similar problem.

branches: 1.16.2;
Switch hp300 to MI todr.

branches: 1.15.4; 1.15.8;
Use _HP300_FOO_H_ instead of _MACHINE_FOO_H_.

branches: 1.14.16;
Do not export __HAVE_RAS to userland. Applications are supposed to try
rasctl() and detect failure with EOPNOTSUPP.

branches: 1.13.6;
Use "#define\t" instead of "#define ".

RAS support for m68k.

branches: 1.11.8;
Use "#define<tab>".

softintr(9) interface for hp300

branches: 1.9.8; 1.9.12;
#define __HAVE_DEVICE_REGISTER on ports that have it, and check for
that, rather than a list of architecture defines, in config_attach

branches: 1.8.28;
Kill __BROKEN_INDIRECT_CONFIG

First step inn removing config_scan() and the hacks that gave devices
on indirect-config busses a (permanent) softc that they could share
between 'match' and 'attach' routines:

Define __BROKEN_INDIRECT_CONFIG so that old autoconfiguration
interfaces are used, until drivers are converted to use the new
interfaces (actually, converted back to use the _older_ interfaces)
which prohibit indirect configuration devices from receiving a softc
in their match routine that they can share with their attach routine.

Nuke __BDEVSW_DUMP_OLD_TYPE.

define __BDEVSW_DUMP_OLD_TYPE for ports where it's true. clean up
some m68k ports inclusion of common header.

new RCS ID format.

stubs which pint to m68k include

add rcsids to everything and clean up headers

branches: 1.1.1;
Initial revision

Retire varargs.h support. Move machine/stdarg.h logic into MI
sys/stdarg.h and expect compiler to provide proper builtins, defaulting
to the GCC interface. lint still has a special fallback.
Reduce abuse of _BSD_VA_LIST_ by defining __va_list by default and
derive va_list as required by standards.

new RCS ID format.

.

Switch hp300 over to the common interrupt dispatch code.

XXX There are still some things to fix up here, but it's no worse
than it was before (the problems date back to when we flattened
the device interrupt levels into IPL_VM).

Switch hp300 over to the common m68k vector table.

G/C unused P1PAGES.

First step to make hp300 port the new m68k pmap ready:

- Add new pmap files to config(5) file definitions
- Add commented out options __HAVE_NEW_PMAP_68K in conf/std.hp300
- Adjust <machine/pmap.h> and <machine/pte.h> for new pmap
initialization ops etc.
- Update <machine/vmparam.h> for the new pmap specific address layout
- Allocate intio, extio, and bootinfo_va space using the pmap_bootmap
structure and initialize va variables in hp300_init() in
__HAVE_NEW_PMAP_68K case

On hp300 the kernel is not mapped VA==PA so more work is necessary
to turn on/off the MMU (on-going).