Add a match quirk to prevent matching any interface on SiPEED FPGA
development boards (e.g. Tang Nano 9K).  The FT2232s on these boards
are wired up only for JTAG.

(thorpej)

Remove a superflouous printf().

(thorpej)

b3_2706_map_vme(): Use VM_BESTFIT.

(thorpej)

_vme_space_get(): Use VM_BESTFIT.

(thorpej)

vmem_init(): Ensure that the quantum is a power of 2, and that if private
tags are being used, they are added to the arena before the first span is
added.

(thorpej)

Add a missing barrier in irongate_conf_read0().

(thorpej)

In dec_6600_device_register(): If we're running on a DS10, set the
"pciide-disable-dma" property on the on-board "aceride" to true.  This
is intended to be a temporary measure until the issue that makes it not
work correctly is found and fixed.

(thorpej)

Always call platform.device_register(), even if bootdev_data is NULL.

(thorpej)

As the final step in configuraing bus-master DMA, consult the boolean device
property "pciide-disable-dma" and disable bus-master DMA if that property is
present and true.

(thorpej)

Add machdep.rpb_type and machdep.rpb_variation sysctl nodes.

(thorpej)

Add an example devpubd hook that looks for Minipro-compatible EEPROM
programmers and sets the access permissions to 0660.

(thorpej)

Add an example devpubd hook that detects the generic USB interface
portion of a Tigard debug board and changes the permissions of the
appropriate /dev/ugenN.* nodes to allow access without superuser
permissions, suitable for using e.g. openocd with the device.

This example can be easily modified to support other generic USB devices
that have user-space drivers where running as the superuser is not desired.

(thorpej)

Descend into share/examples/devpubd/hooks

(thorpej)

Add an example devpubd hook that makes stable path names, using symlinks,
for USB serial interfaces, regardless of where the interface is connected
or the order of enumeration.  This requires the USB device to have a
"serialnumber" to function.  Examples:

- uftdi serial adapter with two ports:

/dev/tty-uftdi-FT64S4YP-1 -> /dev/ttyU0
/dev/tty-uftdi-FT64S4YP-2 -> /dev/ttyU1

- uslsa serial adapter:

/dev/tty-uslsa-01E7ABCC -> /dev/ttyU4

This allows something like the following in /etc/remote:

sun3:dv=/dev/tty-uftdi-FT64S4YP-1:br#9600:pa=none:dc:

That path will always be stable regardless of which /dev/ttyU* node is
actually assigned when the serial adapter is plugged in.

(thorpej)

Add an examples directory for devpubd hooks.

(thorpej)

ugen and ugenif share the same /dev/ugenN.xx namespace in such a way
that the device unit number does not necessarily match the /dev/ugenN.xx
unit number (N).  If you ONLY have ugen devices, it happens to work out
and devpubd scripts can be extremely naive.  If you ONLY have ugenif
devices, it also happens to work out, but your devpubd scripts have to
slightly more informed.  If you have a mix of ugen AND ugenif devices,
though, you're pretty much out of luck.

So, this change adds a "ugen-unit" device property which devpubd scripts
can query to determine which /dev/ugenN.xx nodes a given ugen or ugenif
device is using.

(thorpej)

Add a "match quirk" mechanism to the uftdi driver that allows it to
selectively reject individual interfaces based on the combination of
- Vendor ID
- Product ID
- Interface number
- Vendor string
- Product string

This is necessary[*] to allow some devices that would otherwise match
uftdi (and thus instantiate a ucom) to be matched by ugenif instead,
which is required to make the device available to libusb1.

[*] ...due to a deficiency in the USB stack that does not provide a
mechanism for a user-space driver to claim a device from a kernel driver
and then return it back at a later time.

Use this new match quirk mechanism to reject "interface 1" of the
FTDI 2232C-based Tigard debug board; On this board, "interface 0"
is brought out to regular TTL-level UART pins, but "interface 1" is
brought out to SWD and JTAG headers, and is really only useful when
used with something like openocd.  Because the FTDI 2232C on this board
just uses the standard FTDI vendor and product IDs, it can only be
distinguished by the strings, which cannot be specified usbdevices.config,
thus necessitating the match quirk entry (that works in combination
with the ugenif entry added in usbdevices.config).

(thorpej)

Define a "flags 1" config directive for ugenif, which is similar to ugen's,
but rather forces the ugenif to match at the *lowest* match priority rather
than the highest.  This allows ugenif to claim only otherwise unclaimed
interfaces.

(thorpej)

Tidy up TLSB autoconfiguration just a bit.

(thorpej)

Alpha systems based on the "Laser" and "TurboLaser" system architecture
have an interesting feature: the RTC and console UART are present on each
CPU module, but only those peripherals on the "primary" CPU module matter,
because each CPU's module's periperals are mapped to the same physical
address, but are only accessible by that CPU module.  The firmware selects
a primary CPU to boot the system, and that CPU's RTC and UART are the
system RTC and console, respectively.

To handle this, on systems where it's needed, we wrap the RTC gettime/settime
calls and, if not running on the primary CPU already, cross-call to the primary
to perform the RTC access.

(thorpej)

Statically initialize the PRIMARY and RUNNING flags in cpu_info_primary.

(thorpej)

Wrap the mcclock driver's mc146818_softc inside a new mcclock_softc.
This commit represents no functional change, but forms the basis for
a future functional change.

(thorpej)

bus_space-ify.

(thorpej)

Add a bus space implementation for the Gbus, the general 8-bit bus present
on Laser / TurboLaser CPU modules.

(thorpej)

Expose mc146818_{get,set}time_ymdhms() and allow a front-end to override
these function pointers in the TODR handle, allowing the front-end to
wrap mc146818_{get,set}time_ymdhms() with special handling, if needed.

(thorpej)

Revert previous until I can diagnose a failure reported by gson.

(thorpej)

Early in main(), assert that curcpu() evaluates as the primary CPU and
stash away a pointer to it as the boot CPU for quick reference later.

(thorpej)

Rename the local "boot_cpu" variable to "uvm_boot_cpu".

(thorpej)

Move the at-shutdown call to resettodr() from cpu_reboot() to kern_reboot().

It's a small step, but it's a step.

(thorpej)

prom_uses_prom_console() needs to consider ST_DEC_7000, as well.

(thorpej)

tlsb -> gbus.  No functional change.

(thorpej)

Add the basic framework for the DEC KN7AA ("Ruby") systems -- DEC 7000
and DEC 10000.

This is a work-in-progress, but this should be sufficient for the system
to boot, using the PROM console routines (and then proceed to not find any
devices because we don't yet support the "Laser System Bus").

(thorpej)

Shuffle GBus-related files into their own directory in preparation for
other GBus attachments.  No functional change here; just minimal editing
to reflect path differences after moving the files.

(thorpej)

Revert the CLOCK_FORMAT0 work-around.

(thorpej)

Fix up the mac68k interrupt handlers to work with the new style common
clockframe layout and remove the CLOCK_FORMAT0 work-around.  As a nice
side-effect, this also eliminates the super-sketchy stack unwinding used
by rtclock_intr to get at the interrupt stack frame.

(thorpej)

Define mode macros for CPOL and CPHA, and define SPI_MODE_* in terms
of CPOL and CPHA.  Add SPI_FREQ_*() helper macros to make setting the
transfer frequency less error prone.

(thorpej)

Add the Sun3 obio flavor's DMA controller access registers.

(thorpej)

Move sireg.h from vme/ to sun/ -- the same (or nearly so) register defns
are used by multiple bus attachments.

(thorpej)

Rather than having different "builtin" tables (3xx, 362, 382, 4xx),
just have a single "builtin" table with a mask indicating which of
those groups of machines have a particular built-in device.

(thorpej)

Catch up with revision 1.5 of atomic_init_cas.c.  Fixes building libc
on sun2, and probably others.

(thorpej)

What good are comments, really, if you can't have a little color?  Add
a note above CLKF_INTR() about why NetBSD will likely never switch to
using the "master" stack for the kernel, and thus why the not-enabled
version of CLKF_INTR() will never be enabled.

(Adding the comment here, rather than just deleting the code, has value,
I think, for future retrocomputing archeologists.)

(thorpej)

Hygiene.

(thorpej)

Support the hardware-assisted AST on news68k in the common interrupt
stubs.

(thorpej)

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

(thorpej)

Now that we've agreed on the name "intr_depth", let's agree on the type, too.

(thorpej)

Switch the x68k port to use the same clockframe layout as the other
m68k ports.

(thorpej)

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

(thorpej)

Switch the amiga port to use the same clockframe layout as the other
m68k ports.

(thorpej)

Revert unintended commit.

(thorpej)

Adjust the clockframe to be the same as what the common m68k interrupt
dispatch code expects.

(thorpej)

Ensuire VECO_TO_VECI() returns an 8-bit index.

(thorpej)

Make sure to include opt_mvmeconf.h.

(thorpej)

opt_mvmetype.h -> opt_mvmeconf.h to match mvme68k (which wins out because
it supports more boards).

(thorpej)

Move physaccess(), physunaccess(), and kvtop() from vm_machdep.c to
pmap_motorola.c.  They're legacy functions tied to that pmap implementation.

(thorpej)

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

(thorpej)

Remove a comment that is no longer applicable.

(thorpej)

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).

(thorpej)

Switch next68k over to common interrupt dispatch and G/C __HAVE_LEGACY_INTRCNT.
Also included is G/C of the old ssir stuff that's no longer used.

(thorpej)

Bump required version (mkflagvar).

(thorpej)

Bump version.

(thorpej)

Add "options M68010" for completeness.

(thorpej)

Don't put the MVME1xx model defines into IDENT, make them defflag options and
put them in opt_mvmeconf.h.  Make these options depend on the appropriate
M680x0 option (MVME147 -> M68030, MVME16[27] -> M68040, MVME17[27] -> M68060)
so that the CPU option appears correctly in opt_m68k_arch.h.

(thorpej)

mkflagvar M68010 M68020 M68030 M68040 M68060

(thorpej)

With config(1) as it exists today, a kernel Makefile is able to implement
logic on kernel options so long as those options are not defflag'd or
defparam'd.  This works because such options are automatally added to the
IDENT var in the kernel Makefile as a preprocessor define, and the Makefile
can then do an operation like:

.if !empty(IDENT:M-DSOMECOOLCPUOPTION)
CFLAGS+= -mcpu=somecoolcpu
.endif

Unfortunately, this precludes making it possible to generate a compile-time
dependency on SOMECOOLCPUOPTION, or having SOMECOOLCPUOPTION imply another
kernel config option using the normal config(1) option dependency function.

Fix this by introducing a new option description keyword: mkflagvar.  This
keyword marks an already defflag'd option as wanting a kernel Makefile var
defined if that option is selected in the kernel config file.  So:

defflag opt_coolcpu.h SOMECOOLCPUOPTION ANOTHERCOOLCPUOPTION
mkflagvar SOMECOOLCPUOPTION ANOTHERCOOLCPUOPTION

will cause:

KERNEL_OPT_SOMECOOLCPUOPTION="1"
KERNEL_OPT_ANOTHERCOOLCPUOPTION="1"

...to be emitted into the kernel Makefile if those options are, in fact,
selected with "options ..." in the kernel config file, thus allowing for
a compile-time dependency on the option to be generated in addition to
Makefile logic, which now looks like:

.if !empty(KERNEL_OPT_SOMECOOLCPUOPTION)
CFLAGS+= -mcpu=somecoolcpu
.endif

(thorpej)

Don't put the news1[27]00 model defines into INDENT, make them
defflag options and put them in opt_newsconf.h.  Make these
options depend on the appropriate M680x0 option (M68030 for each)
so that the CPU option appears correctly in opt_m68k_arch.h.

(thorpej)

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>.

(thorpej)

Make sigcode.s and sunos_sigcode.s build as their own stand-alone files.

(thorpej)

Make compat_13_sigreturn13.s and compat_16_sigreturn14.s build as their
own stand-alone files and G/C the now-empty sigreturn.s.

(thorpej)

cpu_switchto(): No need for the pmap_activate() dance here; it's
already done for us in mi_switch().

(thorpej)

G/C intrcnt[].

(thorpej)

Build m68k/support.s on its own; don't include it from locore.s

(thorpej)

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).

(thorpej)

Several years ago, the interrupt priority levels for devices were "flattened"
such that IPL_BIO, IPL_NET, and IPL_TTY (logical interrupt priority levels)
became aliases of IPL_VM (the logical interrupt priority level above which
memory allocation is not allowed).  Unfortuantely, this meant that any
use of these logical interrupt priority levels to differentiate between
different interrupt service routines at the same auto-vectored interrupt
level was pointless... "when everyone is special, no one is".

Before this flattening happened, these distinct logical interrupt priority
levels levels were dynamically mapped to physical auto-vectored interrupt
levels based on which types of devices registered themselves at which auto-
vectored level.  Now, splbio(), splnet(), etc. are all equivalent to splvm(),
which effectively blocks all device interrupts.

Switch to using the ISRPRI_* values provided by the common m68k interrupt
dispatch code, which allows interrupt handlers for devices that are more
latency-sensitive to be sorted earlier in the list of handlers at a given
auto-vectored interrupt level, as was originally intended.

(thorpej)

Provide PSLTOIPL() and IPLTOPSL() macros.

(thorpej)

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).

(thorpej)

Add a hook to allow a platform to suppress reporting stray auto-vectored
interrupts.

(thorpej)

Increment and decrement idepth in the assembly stubs, not C code.  This
provides two advantages:
- Greater coverage for detecting "interrupt time".
- More flexibility for platform-specific interrupt stubs that might need
  to do special processing before calling the common dispatcher.

(thorpej)

Switch mvme68k over to the common interrupt dispatch code and G/C
__HAVE_LEGACY_INTRCNT.

(thorpej)

Use m68k_intrvec_intrhand() to implement isrunlink_vectored().

(thorpej)

- Declare idepth as volatile.
- Provide a m68k_intrvec_intrhand() routine that returns the interrupt
  handle for a given vectored interrupt.  XXX This is gross and should
  burn to the ground, but is needed to support legacy ISR interfaces.

(thorpej)

Switch next68k over to common interrupt dispatch and G/C __HAVE_LEGACY_INTRCNT.
Also included is G/C of the old ssir stuff that's no longer used.

(thorpej)

G/C unused isrlink_evcnt().

(thorpej)

G/C __HAVE_LEGACY_INTRCNT from news68k.

(thorpej)

G/C __HAVE_LEGACY_INTRCNT from luna68k.

(thorpej)

No need for our own spurious interrupt handler now that m68k_intr.c
handles them for us.

(thorpej)

No need for our own spurious interrupt handler now that m68k_intr.c
handles them for us.

(thorpej)

No need for our own spurious interrupt handler now that m68k_intr.c
handles them for us.

(thorpej)

Insert a spurious interrupt handler at auto-vector IPL 0 that simply
abosrbs them.  They'll get counted during the normal course of auto-vector
interrupt handling.

(thorpej)

G/C __HAVE_LEGACY_INTRCNT from cesfic.

(thorpej)

Provide a m68k_count_intr() macro that hides the details of how interrupts
are counted, and also ensures that ci_data.cpu_nintr is incremented.

(thorpej)

Revert unintended commit (didn't hit CTRL-C fast enough I guess).

(thorpej)

Provide an ev_count32 field for situations where a 32-bit counter is
sufficient (and, notably, might be desirable to avoid 64-bit math on
an older 32-bit platform).  This is overlaid on the 64-bit counter
field, and simply references the correct half based on byte order.

(thorpej)

G/C __HAVE_LEGACY_INTRCNT from the sun68k platforms.

(thorpej)

Use evcnt to count interrupts for the non-__HAVE_LEGACY_INTRCNT case.

(thorpej)

Guard references to protorp with M68030.

(thorpej)

Switch cesfic over to the common interrupt dispatch code.

(thorpej)

Switch cesfic over to the common vector table.

(thorpej)

Switch luna68k over to the common interrupt dispatch code.

(thorpej)

Add the "ISR priority" notion that's used on some m68k platforms:

      /*
        * Some devices are particularly sensitive to interrupt
        * handling latency.  Unbuffered serial ports, for example,
        * can lose data if their interrupts aren't handled with
        * reasonable speed.  For this reason, we sort interrupt
        * handlers by an abstract "ISR" priority, inserting higher-
        * priority interrupts before lower-priority interrupts.
        */

(...within the same shared auto-vectored interrupt list.)

(thorpej)

Wrap references to intrcnt[] in __HAVE_LEGACY_INTRCNT.

(thorpej)

Switch news68k over to the common interrupt dispatch code.

(thorpej)

Declare intrcnt as a u_int array, and add a missing m68k_ih_free() call
in m68k_intr_disestablish().

(thorpej)

G/C isrlink_custom(); just use vec_set_entry() directly.

(thorpej)

Switch the sun68k platforms over to the common m68k interrupt dispatch code.

(thorpej)

Add a common m68k interrupt dispatch implementation.

(thorpej)

Provide a _splx() inline.

(thorpej)

Merge common sun68k types defns into <sun68k/types.h>.

(thorpej)

Don't bother with isr_add_custom().  It's just a thin wrapper around
vec_set_entry(), so just use vec_set_entry() directly.

(thorpej)

Switch luna68k over to the common m68k vector table.

(thorpej)

Use NAUTOVECTORS and NUSERVECTORS constants from <m68k/vectors.h>

(thorpej)

Provide NAUTOVECTORS and NUSERVECTORS constants.

(thorpej)

Switch news68k over to the common m68k vector table.

(thorpej)

Switch next68k over to the common m68k vector table.

(thorpej)

Switch mvme68k to the common m68k vector table.

(thorpej)

Expose saved_vbr directly.

(thorpej)

Switch hp300 over to the common m68k vector table.

(thorpej)

Switch the sun68k platforms over to the common m68k vector table.

(thorpej)

Provide some vector table override hooks and interfaces needed by the
sun68k platforms.

(thorpej)

Add buserr2030 and addrerr2030 aliases to busaddrerr2030.

(thorpej)

Switch to common m68k vector table.

(thorpej)

Add a common vector table implementation for m68k platforms.

(thorpej)

G/C some unused definitions.

(thorpej)

Make getsfc() and getdfc() inlines; it's slightly smaller.

(thorpej)

Fix a comment.

(thorpej)

G/C unreferenced copies of _spl() and _splraise().

(thorpej)

We only need one copy of getvbr() and setvbr().

(thorpej)

Make getsr() an inline in <m68k/psl.h> and G/C all of the now-redudant
copies.

(thorpej)

Remove getsp(); nothing uses it.

(thorpej)

Properly implement mm_md_physacc() and garbage-collect the now unused
"lowram" variable (hold-over from hp300 lineage).

(thorpej)

Statically-initialize the '851 / '030 protorp with MMU51_SRP_BITS.  In
pmap_init(), re-initialize protorp with MMU51_CRP_BITS.  Remove the now-
redundant code that does this in each of the m68k platforms.

(thorpej)

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.

(thorpej)

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.

(thorpej)

This platform does not have a loadustp() function.

(thorpej)

Add PROCFS and MSDOSFS.

(thorpej)

Use aprint_error() rather than a bare printf() in a few spots, and
an aprint_normal() in some others.

(thorpej)

Accept "root=xxx" for the root device and "flags=xxx" as the boothowto
flags from the kernel command line arguments.

(thorpej)

Add bootinfo_getarg(), which gets the name kernel command line argument
and returns its value (sort of like getenv()).

(thorpej)

Don't to go IPL0 right before calling main(), instead call spl0() at the
end of cpu_configure().  It seems that Qemu doesn't reset pending interrupts
correctly after a RESET request, which was causing an explosion when an
interrupt was delivered to the new kernel instance.  This change makes
reboot work.

Thx to mlelstv@ for figuring out what the problem was.

(thorpej)

Add virt68k etc stuff.

(thorpej)

Add installation ramdisk for virt68k.  With the GENERIC kernel, you can
just specify this file with "-initrd /path/to/inst-ramdisk.fs" on the
Qemu command line and it will boot right into the installer.

(thorpej)

Fix dump paste-o.  Halt works properly.  Reboot does reboot, but the
new kernel instance crashes early.

(thorpej)

Fix stupid typo in cpu_set_reset_func().

(thorpej)

Implement bootinfo_setup_rndseed().  Thx Riastradh@ for the suggestion.

(thorpej)

Put some meat on the bones of the Goldfish TTY driver.  Works well enough
for sysinst.

(thorpej)

The RAM disk specified by bootinfo is not necessarily loaded along with
the static kernel image.  Map is separately and account for it in the
available memory segments.

(thorpej)

First crack at virt68k sysinst support, more-or-less cloned from mvme68k
with the names changed to protect the guilty and boot block installation
removed.

(thorpej)

First cut at set list for virt68k.

(thorpej)

Perform VirtIO 1.0 feature negotation in the MMIO transport.  This is
required for HV's that have MMIO-v2 and don't support transitional devices.

(thorpej)

G/C some bit defs that are redundant with mmu_{51,40}.h

(thorpej)

get_physical(): Use bit defs from <m68k/mmu_40.h>

(thorpej)

Add gftty at char 12.

(thorpej)

Add virt68k to a list of machines in a comment.

(thorpej)

Add virt68k headers.

(Why on earth is this a manual step!@#??)

(thorpej)

Add mvme68k to MACHINES.m68k.

(thorpej)

Add virt68k to valid_MACHINE_ARCH.

(thorpej)

Add virt68k to MACHINES.m68k.

(thorpej)

Add virt68k.

(thorpej)

Add a driver for the Qemu Virtual System Controller that can be used
to halt or reset the system.

(thorpej)

Provide a mechanism for a system reset controller to register to be used
to reset/halt the system.

(thorpej)

Collect some garbage.

(thorpej)

Detect a RND seed passed in by the loader.
XXX We don't actually *do* anything with it yet.

(thorpej)

Move the initrd setup into bootinfo.c

(thorpej)

If the loader passed us a RAM disk, use it as the root file system.

(thorpej)

Remove non-applicable comments.

(thorpej)

Bump date.

(thorpej)

Comment out LOCKDEBUG.

(thorpej)

Remove yet another remnant of debugging.

(thorpej)

Collect more garbage.

(thorpej)

G/C some scratch-pad notes.

(thorpej)

Add a port to the Qemu m68k "virt" platform, a 68040-based virtual machine
with VirtIO I/O devices.

This is a work-in-progress, but "successfully mounts root file system
and then complains about missing /sbin/init that is, in fact, not there"
seems like a good checkpoint.

(thorpej)

Adapt to changes in the generic gfrtc code.

(thorpej)

Fix file-copy-o.

(thorpej)

Add support for the Goldfish virtual TTY.  This is only enough for
early console; the rest will come later.

(thorpej)

Add support for the Goldfish virtual programmable interrupt controller.

(thorpej)

Add support for the timer portion of the Goldfish RTC device.

(thorpej)

- The VirtIO 1.0 spec says that the MMIO interface uses little-endian
  registers.  For some VMs, at least, this appears to be independent of
  the config/struct byte-order.  Detect this and handle it.
- Add support for the "v2" MMIO personality.

(thorpej)

Define PTE used in the pmap module int terms of the bit definitions
in mmu_{51,40}.h.

(thorpej)

gfrtc_settime(): Write the low half of the time valueinto the RTC_TIME_LOW
register, not the RTC_TIME_HIGH register.

(thorpej)

Provide a vtophys() function like other platforms do, and implement
kvtop() (which will be g/c'd shortly) in terms of it.

(thorpej)

pmap_bootstrap(): When initializing the L2 PTEs, don't use l2pte_index(),
as it is designed to wrap from 1023->0 during the normal course of use.
However, when setting up the initial kernel page tables, if we have a
large enough config that we end up with more than L2 PT page, we rely
on being able to index > 1023 into the adjacent PT pages to initialize
those entries.

Fixes a MM fault panic during early boot on larger memory configs
(reported by Dave McGuire with an 8GB ES40 and John Klos with a 12GB
DS25).  I've successfully booted at 16GB config in Qemu with this fix.

(thorpej)