System utilities, boot programs and kernel modules are machine
(port) specific not CPU, so use MACHINE not MACHINE_ARCH.

Convert more MACHINE tests to MACHINE_ARCH

branches: 1.20.78; 1.20.86; 1.20.92;
merge ktrace-lwp.

merge ultrasparc ofwboot / ofwboot.net.

branches: 1.18.2;
Don't traverse here in case of sparc64 32-bit environment.

decommission old installboot; it doesn't work on bootxx anymore anyway.

branches: 1.16.4; 1.16.8;
build sparc64 bootblk, ofwboot, & ofwboot.net on sparc

branches: 1.15.22;
Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Don't use _SUBDIRUSE outside of bsd.subdir.mk.

Adapt to sys/lib/lib{sa,kern}/Makefile.inc changes

Use ${INSTALL}.

update for kern/sa lib Makefile.inc changes.

Brr.. libkern.o is way too big for our bootxx..

Script to assist installing second-stage boot programs.

Make this work with OBJ dirs.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Adapt to new lib{sa,kern} regime.

support obj directory

copyright/Id cleanup

Make it work for V0 proms.
First cut at two stage bootstrap.

Adapt to current libsa.

Sparc bootblocks (a tight fit)

deal with PIE and SSP

branches: 1.31.2;
Add a few includes of bsd.own.mk to avoid broken test conditions
on ACTIVE_CC - this might not be the correct fix for PR 50930,
but it avoids the issue.

When building with clang, optimise for minimal size and skip frame
pointers when possible.

branches: 1.29.6;
Build boot loaders without unwind tables.

branches: 1.28.8; 1.28.18; 1.28.22;
factor out the floppy detection code.

Introduce .WAIT into the cleandir rule for blasting the libsa/libkern/etc.
trees, because it can race with the libsa/libkern/etc. makefiles' own
cleandir rules. I think I've found all of the uses of the offending idiom...
Closes PR 43360.

branches: 1.26.2; 1.26.4;
Add src/common/include to -Ipath for proplib(3) so that build works
even without installed DESTDIR.

Use <bsd.klinks.mk> to create machine and ${MACHINE_ARCH} symlinks.

- use -Os to shrink bianries
- remove local shrinked memcpy() from bootxx.c bacause
now it's identical with __OPTIMIZE_SIZE__ libsa one
derived from src/common/lib/libc/string/bcopy.c.

Tested netboot from le0 and bootxx + boot from sd0 on SS1+.

branches: 1.23.78; 1.23.86; 1.23.92;
merge ktrace-lwp.

Ignore errors on some rm -rf's for platforms that aren't quite
POSIX compliant
and return errors on r/o source (FreeBSD) for -rf. Fixes PR#25022

branches: 1.21.2;
sprinkle some: libkern.h, some -ffreestanding and some cast-y goodness,
all for the pleasure of the gcc 3 :-)

branches: 1.20.2;
Make sure to remove lib directories.

Add support for large kernels by searching for a physical memory segment
to fit it in. As a bonus, the kernel is now also mapped to the virtual
address (i.e. KERNBASE) it is linked at.

make sparc64 friendly.

branches: 1.17.8;
MKfoo=no -> NOfoo

Don't make the symlinks on clean rules as well as obj rules. Breaks r/o source
trees that haven't gotten to the make obj recursion part of make build
otherwise.

branches: 1.15.4; 1.15.8;
Change the relocation address of `boot' once again to steer clear of
memory used by the PROM on sun4 and early sun4c machines.

Change to relocation addresses that currently work on all machines.
boot programs at alternative addresses no longer built and installed
by default.

Pull down from 1.4 branch: default relocation address = 3E0000

branches: 1.12.4;
SUN_BOOTPARAM compile-time option is no longer used.

bootxx links at 0x340000
2nd-stage boot programs choices: 0x200000,0x300000,0x400000,0x700000 (default: 0x300000)

Build several versions of the second-stage boot program, each linked
to run at a different base address. Boot image build procedures can
pick one which is least likely to fail..

The default boot program is again build to run at 0x340000.

Clean up the machine symlink stuff ever so slightly. Needs to be
standardized between files.

Suppress the `machine' symlink when making target `obj'.
Also add it to the clean target list.

branches: 1.7.2; 1.7.8;
Initialize the heap (for alloc()) explicitly, since the `end' symbol
is not necessarily double-word aligned.

Checkpoint of elf boot stuff:
- Use loadfile for both boot and installboot
- Fix problems with linking in bootxx
- Make installboot work with both a.out and ELF bootxx
- *WILL NOT BOOT YET ELF KERNELS* (a.out kernels work) I am working on it.

branches: 1.5.2; 1.5.4;
Install the `machine' link again; it went missing a while ago.
Add `-I.' to CPPFLAGS to use it.

* Add promlib.c to common sources.
* Compile with prototype warnings turned on.

convert from NOxxx= to MKxxx=no.
include <bsd.own.mk> if testing a MKxxx variable.

Define _STANDALONE instead of STANDALONE.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

deal with PIE and SSP

branches: 1.13.2;
We don't want PIE!

Add a few includes of bsd.own.mk to avoid broken test conditions
on ACTIVE_CC - this might not be the correct fix for PR 50930,
but it avoids the issue.

Ignore unused arguments for clang as we pass assembler flags and the
like to the preprocessor-only mode. Conditionalize GCC flags that are
not known to clang accordingly.

branches: 1.10.122; 1.10.142;
merge ktrace-lwp.

sparc needs mostly the same compilation flags for standalone build as does
sparc64. Move them out of the sparc64 conditional.

Okayed by martin@

branches: 1.8.16; 1.8.18;
sprinkle some: libkern.h, some -ffreestanding and some cast-y goodness,
all for the pleasure of the gcc 3 :-)

branches: 1.7.24;
add/merge sparc64 support.

branches: 1.6.22; 1.6.32;
Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

LOCORE -> _LOCORE

Brr.. libkern.o is way too big for our bootxx..

Make this work with OBJ dirs.

Fixed RELOC's to be hex numbers (as expected) rather than decimal.
(e.g. s/340000/0x340000/). Fixes DVMA panic on my sun-4/280 when
booted diskless.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Refer to installboot(8)

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

copyright/Id cleanup

Make it work for V0 proms.
First cut at two stage bootstrap.

bootblocks must be stripped.

A one liner for now.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

ffs bootblock install can be simplified since installboot(8) handles that now.
To install a TFTPable boot image, we have to play the header tricks here.

Fix typo that caused sun4 machines to be misidentified as sun4c-class.
Fixes reported problem of binstall not stripping the a.out header from
the boot block on sun4 machines.

Sync filesystem before running installboot.

branches: 1.1.2;
Script to assist installing second-stage boot programs.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Correct some prototypes.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Reserve arguments 2-4 for SunOS loaders; pass DDB stuff in args 5 and 6.

Use the twiddle() from libsa.

Follow prototype changes in libsa.

cosmetic change.

round text segment to __LDPGSZ

copyright/Id cleanup

Use PROM node properties to determine device type.

Make it work for V0 proms.
First cut at two stage bootstrap.

Adapt to current libsa.

DDB support

Sparc bootblocks (a tight fit)

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Avoid reading the filesystem superblock in bootxx; all info now prepared
by installboot (per Gordon Ross).

Adapt to libsa changes & some other minor nits.

read superblock into &sblock rather than LOADADDR. we can't read
into LOADADDR because it can't be mapped into DVMA space. this fixes
a dvma panic on a sun4 boot from an xd disk.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Use the twiddle() from libsa.

copyright/Id cleanup

Make it work for V0 proms.
First cut at two stage bootstrap.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Adapt to libsa changes & some other minor nits.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Do away with ROM printf.

copyright/Id cleanup

Make it work for V0 proms.
First cut at two stage bootstrap.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Correct some prototypes.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

copyright/Id cleanup

Adapt to current libsa.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

copyright/Id cleanup

Sparc bootblocks (a tight fit)

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Document `-a karch' option.
Fix a formatting nit.

Manual page for installboot(8).

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

installboot(8) now handles required a.out header munging.
Also, the munging is extended by inserting an instruction that branches
past the header, so the same boot image can be used on all supported
sun platforms.

Add `-a <arch>' switch to force the target architecture. Retain `-h' for
compatibility.

UFS => FFS conversion.

Correct test for sun4 machines (thanks Chuck).
Since uname currently does not allow to discriminate different sparc
models, we use `sysctl.hw.cpumodel' to do this somewhat heuristicly:
if the returned strings starts with "SUN4-" a sun4 architecture is
assumed. This information is used to determine whether or not to strip
the a.out header off the 1st-stage bootblocks.

branches: 1.9.2;
Avoid reading the filesystem superblock in bootxx; all info now prepared
by installboot (per Gordon Ross).

sync filesystems just prior to writing the bootblocks to the raw device.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Include exec header in length computation.

don't assume the f_fsnamelen is nul-truncated or longer than MFSNAMELEN

Prototype internal function.

copyright/Id cleanup

Remove bogus fsstat().

Make it work for V0 proms.
First cut at two stage bootstrap.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

New ARP system, supports IPv4 over any hardware link.

Some of the stuff (e.g., rarpd, bootpd, dhcpd etc., libsa) still will
only support Ethernet. Tcpdump itself should be ok, but libpcap needs
lot of work.

For the detailed change history, look at the commit log entries for
the is-newarp branch.

branches: 1.3.8;
mountroot: return error code to caller, not -1.

Adapt to libsa changes & some other minor nits.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

New ARP system, supports IPv4 over any hardware link.

Some of the stuff (e.g., rarpd, bootpd, dhcpd etc., libsa) still will
only support Ethernet. Tcpdump itself should be ok, but libpcap needs
lot of work.

For the detailed change history, look at the commit log entries for
the is-newarp branch.

branches: 1.2.10;
Adapt to libsa changes & some other minor nits.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Not all v2 proms understand the `mac-address' word (from Theo).

Check input channel in gettick(); for `L1-A' processing.

branches: 1.14.2;
make sure devopen() sets pd->devtype properly when BOOTXX is defined
(should be DT_BLOCK). otherwise the bootxx boot block will fail if
you are on a system with a v2 openprom [because it will not do the
v2_seek in obp_strategy()]. fixes the problem recently noted on
the port-sparc list by matthew green <mrg@eterna.com.au>. i tested
fix on a ss2 with a v2 prom.

Adapt to libsa changes & some other minor nits.

Correct some prototypes.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Adapt declaration to stand.h; PR #1036

Use the twiddle() from libsa.

Follow prototype changes in libsa.

putchar is void

copyright/Id cleanup

Keep those pointers moving.

Use PROM node properties to determine device type.

Make it work for V0 proms.
First cut at two stage bootstrap.

Adapt to current libsa.

Sparc bootblocks (a tight fit)

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Adapt to libsa changes & some other minor nits.

Correct some prototypes.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Clarify the comment about enabling interrupts above level 11, suggested
by Andrew Gillham.

Enable just enough interrupts to get "L1-A" through. Patch from Jason to
avoid spurious floppy interrupts on Jason's machine...

branches: 1.5.4;
Enable interrupts; some ROM versions apparently turn them off.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

copyright/Id cleanup

Make it work for V0 proms.
First cut at two stage bootstrap.

Sparc bootblocks (a tight fit)

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Bump version number to celebrate the fact that the bootblocks can be used
unmodified across all platforms.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

copyright/Id cleanup

Make it work for V0 proms.
First cut at two stage bootstrap.

Sparc bootblocks (a tight fit)

MKfoo=no -> NOfoo

branches: 1.2.22; 1.2.26;
convert from NOxxx= to MKxxx=no.
include <bsd.own.mk> if testing a MKxxx variable.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Convert from getopt to getopts (script is slightly smaller as a
result, but not enough to save a block ... but if /usr/bin/getopt is
not used elsewhere, and it should not be, it will no longer be required.)

General cleanups...
Better quote usage
No more need for sed
(should be no need for getopt either, but did not do that one)

PR install/53610

Remove use of -a in test(1) - deprecated in standard, and no
longer supported in SMALL (install media) versions of test in /bin/sh

branches: 1.15.148; 1.15.150;
Omit specifying the secondary bootloader on the installboot command line
for sparc64 systems. This should fix PR#33987.

Also adapt to the way sparc64 32-bit kernels now report its CPU
architecture -- these still need the sparc64 boot loader, but now
report hw.machine = sparc. Adapt by testing machdep.cpu_arch
instead.

Hints from martin and mrg.

branches: 1.14.4; 1.14.8; 1.14.16;
merge ktrace-lwp.

branches: 1.13.16;
ofwboot.net is gone. from hannken@

shell coding tweaks

branches: 1.11.10;
Patch up to work with MI installboot(8).

- add a -i path-to-installboot option
- also recognise $INSTALLBOOT from the environment
- update Usage & share the option list with Help.

clarify usage. idea from david obrien.

branches: 1.8.4; 1.8.8;
Add -f option. If used it will perform the low level block operations on the
file rather than requiring a raw device. Useful when dealing with vnd's
images and building snapshots. Also allows snapshots to be built in
securelevel > 0.

fix ultrasparc support.

remove some ".sparc64" file suffixes that are not necessary.

add sparc64 support.

branches: 1.4.4;
Allow second-stage boot program to be specified on the command line.

branches: 1.3.2; 1.3.8;
Try not to run when kernel is in secure mode.

Header magic is now done in stand/boot/Makefile.

branches: 1.1.8;
Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

adapt to ${CC_WNO_ADDRESS_OF_PACKED_MEMBER}

Simplify CWARNFLAGS to use ${CC_WNO_ADDRESS_OF_PACKED_MEMBER}
which works for both clang and gcc, and remove compiler-specific
equivalents.

bsd.own.mk: rename to CC_WNO_ADDRESS_OF_PACKED_MEMBER

Provide a single variable
CC_WNO_ADDRESS_OF_PACKED_MEMBER
with options for both clang and gcc, to replace
CLANG_NO_ADDR_OF_PACKED_MEMBER
CC_NO_ADDR_OF_PACKED_MEMBER
GCC_NO_ADDR_OF_PACKED_MEMBER

Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.

add support for new GCC 9 warnings that may be too much to fix
right now. new address-of-packed-member and format-overflow
warnings have new GCC_NO_ADDR_OF_PACKED_MEMBER amd
GCC_NO_FORMAT_OVERFLOW variables to remove these warnings.

apply to a bunch of the tree. mostly, these are real bugs that
should be fixed, but in many cases, only by removing the 'packed'
attribute from some structure that doesn't really need it. (i
looked at many different ones, and while perhaps 60-80% were
already properly aligned, it wasn't clear to me that the uses
were always coming from sane data vs network alignment, so it
doesn't seem safe to remove packed without careful research for
each affect struct.) clang already warned (and was not erroring)
for many of these cases, but gcc picked up dozens more.

Appease clang -Waddress-of-packed-member for <ufs/lfs/lfs_accessors.h>,
in the same manner as done for sparc64; pointer substitution is just for
sanity check and harmless there.

Also move bootjs.net after we've evaluated $OBJS (thanks gson@)

branches: 1.41.4;
move .for generated rules after OBJS have been set.

centralize vers.c building for standalone programs.

branches: 1.39.2; 1.39.4;
Strip debug symbols from boot code

branches: 1.38.12; 1.38.30;
emit some linker -Map files so that we can figure out what symbols
in the boot loaders are.

Drop bootprog_maker (formerly enabled by -M) and bootprog_date (formerly
disabled by -D) from the output of newvers_stand.sh. Change -D to the
inverted logic, so that it adds the date to bootprog_rev in ().

Change all platforms accordingly. -D is added if MKREPRO is not yes and
wasn't present before. Platforms that didn't use -D don't depend on
MKREPRO now either.

branches: 1.36.4; 1.36.6; 1.36.8;
System utilities, boot programs and kernel modules are machine
(port) specific not CPU, so use MACHINE not MACHINE_ARCH.

Convert more MACHINE tests to MACHINE_ARCH

branches: 1.34.78; 1.34.86; 1.34.92;
merge ktrace-lwp.

Add SUPPORT_DHCP - otherwise bootp does not work (packet is too short for
root-path). From Christos.

branches: 1.32.14; 1.32.16;
Rename the JavaStation network boot loader: bootjs.elf -> bootjs.net.

Add support to build bootjs.elf - the secondary bootstrap used to netboot
JavaStations. Based on patches from Valeriy E. Ushakov.

Use ${HOST_SH} instead of `sh'.

If necessary, pull in <bsd.sys.mk> to get the definition of HOST_SH;
Makefiles that pull in one of (most of) <bsd.*.mk> will get this anyway.

sprinkle some: libkern.h, some -ffreestanding and some cast-y goodness,
all for the pleasure of the gcc 3 :-)

branches: 1.28.2;
Add support for large kernels by searching for a physical memory segment
to fit it in. As a bonus, the kernel is now also mapped to the virtual
address (i.e. KERNBASE) it is linked at.

Don't declare which exec formats to support in loadfile_machdep.h.
Instead, give each user of loadfile() control over which exec formats
it wants to support.

branches: 1.26.4;
objcopy -> ${OBJCOPY}; remove redundant definitions of OBJCOPY?= (it is
now in bsd.own.mk).

use the 32-bit compiler on sparc64 (but not yet enabled.)

branches: 1.24.2; 1.24.4;
Add PROM patches (forth code) as a table indexed with root node name
and prom version. Little tweaks to prom during secondary boot are
much easier then fragile work-arounds in the kernel. Currently this
is used for JavaStation 1 with OBP3.

Dispose of `boot.net.big'.

switch to sys/conf/newvers_stand.sh for generation of version file
sync the message in boot/version with other ports using newvers_stand.sh

Build a boot_big.net (based at 700000) in addition to boot.net,
allowing you to netboot a kernel with a ramdisk inside on systems
where memory is laid out to allow this.

branches: 1.20.4;
Make dependall work.

Revert to using <bsd.prog.mk> again, so we support all targets that may
be presented when building the system (noticed by Juergen Hannken-Illjes).

We need <bsd.obj.mk>.

Build several versions of the second-stage boot program, each linked
to run at a different base address. Boot image build procedures can
pick one which is least likely to fail..

The default boot program is again build to run at 0x340000.

branches: 1.16.2;
Move the boot block load address from 3MB to 7MB. The INSTALL kernel is
already too large, and GENERIC won't be far behind.

branches: 1.15.2; 1.15.4; 1.15.6;
Use ${SIZE}, ${OBJCOPY} rather than size and objcopy.

Better pad the artificial header to exactly 32 bytes.

Add bootinfo struct to pass symtab arguments.
XXX: This does not work with ELF yet, but I don't understand why...

Checkpoint of elf boot stuff:
- Use loadfile for both boot and installboot
- Fix problems with linking in bootxx
- Make installboot work with both a.out and ELF bootxx
- *WILL NOT BOOT YET ELF KERNELS* (a.out kernels work) I am working on it.

branches: 1.11.2; 1.11.4;
Use `-Ttext' switch & collect ld options in LINKFLAGS.

Use PROMLIB to streamline this code. This also provides the bulk of
Open Firmware boot code.

Clean vers.c & vers.o.

Use DESTDIR consistently when constructing `boot.net'.

Change load address to avoid cache-aliasing on machines with Harvard-style
cache implementations (notably the Hypersparcs).

Install a version of the secondary boot block in /usr/mdec that is
suitable for use on a server for diskless booting. Done mostly to
simplify the installation notes and avoid questions ...

branches: 1.5.8;
Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

remove header munging hack.

update for kern/sa lib Makefile.inc changes.

Make this work with OBJ dirs.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

PR 51200 gets in libsa considered harmful: use kgets

Don't use dynamically sized unions, they are a horrible GNUism. Just use
the existing upper limit all the time.

branches: 1.28.2; 1.28.16;
Pass the "boothowto" from the bootloader to the kernel.
If (an old) bootloader did not pass the flags, check explicitly if we
missed RB_USERCONF and invoke userconf_prompt() in MD code, as MI code
already missed it.
This fixes PR 46466.

branches: 1.27.4; 1.27.8; 1.27.10;
Drop bootprog_maker (formerly enabled by -M) and bootprog_date (formerly
disabled by -D) from the output of newvers_stand.sh. Change -D to the
inverted logic, so that it adds the date to bootprog_rev in ().

Change all platforms accordingly. -D is added if MKREPRO is not yes and
wasn't present before. Platforms that didn't use -D don't depend on
MKREPRO now either.

branches: 1.26.2; 1.26.4;
factor out the floppy detection code.

Don't LOAD_BACKWARDS for floppies.

branches: 1.24.44; 1.24.64; 1.24.66;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.23.10;
De-__P. ANSIify.

Same object code is generated (except in prf.c, where putchar argument
type was changed to int to match the prototype and one instrcution
changes accordingly).

strategy() in ofdev.c was not ANSIified in this sweep, b/c that
changes code generated at call sites!

branches: 1.22.4; 1.22.8;
merge ktrace-lwp.

branches: 1.21.2;
constify

Do not rely on strings to which pointers are returned by promlib functions
to persist accross further promlib calls.
This fixes (at least) bootjs.net.

branches: 1.19.4; 1.19.10;
compat mode: skip double map at VA 0 if the physical load address is 0.
This allows a much wider range of historic kernels to be booted (w/ `-C').

branches: 1.18.2;
Move UCB-licensed code from 4-clause to 3-clause licence.

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

sprinkle some: libkern.h, some -ffreestanding and some cast-y goodness,
all for the pleasure of the gcc 3 :-)

compat mode:
* print warning if kernel image is too large.
* adjust `required space' computation to be more compatible to
what it was before.
* since there's still no code to recognise `old kernels', provide
a way to turn on `compatmode' manually, using the `-C' boot option.

branches: 1.15.2;
Add support for large kernels by searching for a physical memory segment
to fit it in. As a bonus, the kernel is now also mapped to the virtual
address (i.e. KERNBASE) it is linked at.

* Use fdloadfile() to avoid opening the the underlying device twice.
* Cleanup old stuff.

* loadfile() return a file descriptor that must be closed.
* check the kernel size before loading

Use BOOT_FLAG() macro to process flags.

Pass actual loaded kernel file in bootinfo[].

branches: 1.10.2; 1.10.8;
Add PROM patches (forth code) as a table indexed with root node name
and prom version. Little tweaks to prom during secondary boot are
much easier then fragile work-arounds in the kernel. Currently this
is used for JavaStation 1 with OBP3.

Include <sys/exec.h> instead of <a.out.h>/

branches: 1.8.2; 1.8.4; 1.8.8;
Initialize the heap (for alloc()) explicitly, since the `end' symbol
is not necessarily double-word aligned.

Casts for PROM_LOADADDR which is not casted anymore.

Add bootinfo struct to pass symtab arguments.
XXX: This does not work with ELF yet, but I don't understand why...

Checkpoint of elf boot stuff:
- Use loadfile for both boot and installboot
- Fix problems with linking in bootxx
- Make installboot work with both a.out and ELF bootxx
- *WILL NOT BOOT YET ELF KERNELS* (a.out kernels work) I am working on it.

branches: 1.4.4;
Use PROMLIB to streamline this code. This also provides the bulk of
Open Firmware boot code.

add support for finding kernels other than `netbsd'. the list is:
netbsd, netbsd.gz, netbsd.old, netbsd.old.gz, onetbsd, onetbsd.gz
and vmunix, in that order. this is used by the bootable `bootfs'
image (suitable for floppies and cdroms), which contains a netbsd.gz
by default.
while i'm here, fix up `boot [device] -s' which got broke when
booting kernels with a - in their name was fixed..

branches: 1.2.2;
Do not use KERNBASE to calculate the `end of symbols' location. Instead
use the `a_entry' field from the kernel's header as a base.

Pass on a new DDB magic number to the kernel, so kernels linked for a
different KERNBASE value can apply a correction when booted from old
bootblocks.

branches: 1.1.4;
Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Change about 4500 of the K&R function definitions to ANSI ones.
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)

branches: 1.2.8; 1.2.14;
Remove clause 3 and 4 from TNF licenses

branches: 1.1.2; 1.1.140; 1.1.142; 1.1.144;
Add bootinfo struct to pass symtab arguments.
XXX: This does not work with ELF yet, but I don't understand why...

Remove clause 3 and 4 from TNF licenses

branches: 1.2.58; 1.2.60; 1.2.62;
De-__P. ANSIify.

Same object code is generated (except in prf.c, where putchar argument
type was changed to int to match the prototype and one instrcution
changes accordingly).

strategy() in ofdev.c was not ANSIified in this sweep, b/c that
changes code generated at call sites!

branches: 1.1.2; 1.1.54; 1.1.68; 1.1.72;
Add bootinfo struct to pass symtab arguments.
XXX: This does not work with ELF yet, but I don't understand why...

merge ktrace-lwp.

Constify. Make (the only) submatch function static.
Prettify some comments.

add a prom patch that creates a device node for the Krups' audio chip
in case it's missing ( like with OF 3.11.10 ) - makes audiocs at ebus work
on these machines.

branches: 1.9.14; 1.9.16;
Rename PROM_getpropstring* => prom_getpropstring*

Uppercase CPU, plural is CPUs.

sprinkle some: libkern.h, some -ffreestanding and some cast-y goodness,
all for the pleasure of the gcc 3 :-)

branches: 1.6.6;
Extend the matching routine to take a function pointer, so that additional
(arbitrary) matching can be done.
Add match function and patch for Cycle 5 IP (Sparc 5 clone).

Reviewed by Uwe.

OBP 2.30 in JS1 is romvec version 3, so g/c PROM_OBP_V2 entry.
Add bare-bones "8042" (pckbc) node and rename bogus zs nodes out of
the way to prevent them from matching.

add support for OBP 3.x as seen on some JavaStation 1's

Add patches for Mr.Coffee with OBP2.
Miscellaneous cosmetic changes.

branches: 1.2.4;
getprop* -> PROM_getprop*

branches: 1.1.2; 1.1.4;
Add PROM patches (forth code) as a table indexed with root node name
and prom version. Little tweaks to prom during secondary boot are
much easier then fragile work-arounds in the kernel. Currently this
is used for JavaStation 1 with OBP3.

merge ktrace-lwp.

add a prom patch that creates a device node for the Krups' audio chip
in case it's missing ( like with OF 3.11.10 ) - makes audiocs at ebus work
on these machines.

branches: 1.17.2; 1.17.16; 1.17.18;
Add support for large kernels by searching for a physical memory segment
to fit it in. As a bonus, the kernel is now also mapped to the virtual
address (i.e. KERNBASE) it is linked at.

Increment version number for match function and Cycle 5 IP changes.

PROM patches for JavaStation-1.

Bump the version numbers of all bootloaders which use loadfile_elfXX() now
that it tries hard to avoid backwards seeks.

branches: 1.13.4; 1.13.6;
Bump version numbers of all bootloaders that use loadfile/ELF, to account for
my changes to symbol loading. I should probably have done this at the time,
but it's better late than never.

branches: 1.12.4;
switch to sys/conf/newvers_stand.sh for generation of version file
sync the message in boot/version with other ports using newvers_stand.sh

branches: 1.11.2;
Add bootinfo struct to pass symtab arguments.
XXX: This does not work with ELF yet, but I don't understand why...

Checkpoint of elf boot stuff:
- Use loadfile for both boot and installboot
- Fix problems with linking in bootxx
- Make installboot work with both a.out and ELF bootxx
- *WILL NOT BOOT YET ELF KERNELS* (a.out kernels work) I am working on it.

branches: 1.9.4;
look for /^Version:/ to find the boot block version, rather than using
the "version" file's RCS id (which is useless for branches).

add support for finding kernels other than `netbsd'. the list is:
netbsd, netbsd.gz, netbsd.old, netbsd.old.gz, onetbsd, onetbsd.gz
and vmunix, in that order. this is used by the bootable `bootfs'
image (suitable for floppies and cdroms), which contains a netbsd.gz
by default.
while i'm here, fix up `boot [device] -s' which got broke when
booting kernels with a - in their name was fixed..

branches: 1.7.8;
need to bump this back to version 1.7 (last one).

need to bump this back to version 1.7.

need to bump this back to version 1.7.

need to bump this back to version 1.7.

need to bump this back to version 1.7.

need to bump this back to version 1.7.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

no more _LKM -> _MODULE

branches: 1.16.16;
deal with PIE and SSP

branches: 1.15.6; 1.15.10;
Add GENASSYM_CPPFLAGS in two more places.

Use <bsd.klinks.mk>

branches: 1.13.10; 1.13.20; 1.13.24;
Major overhaul:

1) Add support for multiple filesystems including ffsv1, ffsv2, lfsv1 and lfsv2.
(ffsv1 and lfsv2 are known to work. ffv2 support needs more work. lfsv1 is
probably untested.)

2) Only 32-bit Fcodes are used. All 64-bit math uses two cells. This means it
should work on 32-bit machines.

Additional work to consider:

1) Rename genfth.cf to genffs.cf, or keep it and move the ffs symbols to genffs.cf

2) Move the ffs code and the lfs code to different source files so you can
selectively fload the just the filesystems you want.

3) Add code to load ELF files directly instead of relying on the host firmware
ELF support. But, since the binary is now 7250 bytes there may not be room.

branches: 1.12.2;
assym.fth.h also requires sparc symlink. Fix build without DESTDIR.

branches: 1.11.2; 1.11.8;
Use ${TOOL_AWK} instead of ${AWK} or plain "awk" in make commands.
Pass AWK=${TOOL_AWK:Q} to shell scripts that use awk.

branches: 1.10.56; 1.10.60; 1.10.66;
Add src/common/include to include header search path since
<sys/device.h> includes <prop/proplib.h>.

branches: 1.9.4; 1.9.8;
merge ktrace-lwp.

branches: 1.8.2;
Use TOOL_GENASSYM. Part of PR toolchain/30350

Use ${HOST_SH} instead of `sh'.

If necessary, pull in <bsd.sys.mk> to get the definition of HOST_SH;
Makefiles that pull in one of (most of) <bsd.*.mk> will get this anyway.

sprinkle some: libkern.h, some -ffreestanding and some cast-y goodness,
all for the pleasure of the gcc 3 :-)

Rename a large chunk of the make(1) variables which refer to a
program/tool from "FOO" to "TOOL_FOO". The new variables are:
TOOL_ASN1_COMPILE TOOL_CAP_MKDB TOOL_CAT TOOL_CKSUM TOOL_COMPILE_ET
TOOL_CONFIG TOOL_CRUNCHGEN TOOL_CTAGS TOOL_DB TOOL_EQN TOOL_FGEN
TOOL_GENCAT TOOL_GROFF TOOL_HEXDUMP TOOL_INDXBIB TOOL_INSTALLBOOT
TOOL_INSTALL_INFO TOOL_M4 TOOL_MAKEFS TOOL_MAKEINFO TOOL_MAKEWHATIS
TOOL_MDSETIMAGE TOOL_MENUC TOOL_MKCSMAPPER TOOL_MKESDB
TOOL_MKLOCALE TOOL_MKMAGIC TOOL_MKTEMP TOOL_MSGC TOOL_MTREE
TOOL_PAX TOOL_PIC TOOL_PREPMKBOOTIMAGE TOOL_PWD_MKDB TOOL_REFER
TOOL_ROFF_ASCII TOOL_ROFF_DVI TOOL_ROFF_HTML TOOL_ROFF_PS
TOOL_ROFF_RAW TOOL_RPCGEN TOOL_SOELIM TOOL_SUNLABEL TOOL_TBL
TOOL_UUDECODE TOOL_VGRIND TOOL_ZIC

For each, provide default in <bsd.sys.mk> of the form:
TOOL_FOO?= foo
and for the ${USETOOLS}=="yes" case in <bsd.own.mk>, provide override:
TOOL_FOO= ${TOOLDIR}/bin/${_TOOL_PREFIX}foo

Document all of these in bsd.README.

This cleans up a chunk of potential (and actual) namespace collision
within our build infrastructure, as well as improves consistency in
the share/mk documentation and provision of appropriate defaults for
each of these variables.

branches: 1.4.6;
Use ${FGEN} rather than fgen.

branches: 1.3.8;
Fix problem introduced with the MKMAN->NOMAN conversion.

MKfoo=no -> NOfoo

branches: 1.1.2; 1.1.4; 1.1.8; 1.1.12;
move ofwboot & bootblk here; add a Makefile to build ofwboot.net

Fix various typos in comments.

branches: 1.17.8;
Restore backward compatibility of UFS2 with previous NetBSD releases by
disabling support in UFS2 for extended attributes (including ACLs).
Add a new variant of UFS2 called "UFS2ea" that does support extended attributes.
Add new fsck_ffs operations "-c ea" and "-c no-ea" to convert file systems
from UFS2 to UFS2ea and vice-versa (both of which delete all existing extended
attributes in the process).

Fix all remaining typos, mainly in comments but also in few definitions and log messages, reported by me in PR kern/54889.
Also fixed some additional typos in comments, found on review of same files or typos.

branches: 1.15.34;
Use lfs32_dinode_SIZEOF instead of ufs1_dinode_SIZEOF for lfs. These
are the same so it doesn't make any difference, but technically one
should use lfs parts with lfs.

Make these compile again after changes to LFS.

branches: 1.13.18; 1.13.36;
indir-block needs to be the size of a filesystem block or bad things can happen.

branches: 1.12.2;
Fix shifting by zero.

Fix ffsv2 support.

Major overhaul:

1) Add support for multiple filesystems including ffsv1, ffsv2, lfsv1 and lfsv2.
(ffsv1 and lfsv2 are known to work. ffv2 support needs more work. lfsv1 is
probably untested.)

2) Only 32-bit Fcodes are used. All 64-bit math uses two cells. This means it
should work on 32-bit machines.

Additional work to consider:

1) Rename genfth.cf to genffs.cf, or keep it and move the ffs symbols to genffs.cf

2) Move the ffs code and the lfs code to different source files so you can
selectively fload the just the filesystems you want.

3) Add code to load ELF files directly instead of relying on the host firmware
ELF support. But, since the binary is now 7250 bytes there may not be room.

branches: 1.9.2;
Move Eduardo Horvath's license to 2 clause. OK eeh@.

Instead of bailing out with "Inode not directory", assume that the disk is
a RAIDframe mirror, apply the RAID boot offset, and try to boot again.

Fixes the problem where disks would be unbootable if partitioned normally
and then converted to RAID 1 without being zeroed first.

branches: 1.7.24; 1.7.68; 1.7.74; 1.7.78; 1.7.80; 1.7.82;
merge ktrace-lwp.

branches: 1.6.10; 1.6.20; 1.6.22;
Use RF_PROTECTED_SECTORS instead of a (wrong) magic value.
The bootblock now is able to load ofwboot from a raid1 root.

branches: 1.5.2; 1.5.4;
Track changes introduced when UFS2 was merged, so that this builds again.

Dynamically allocate the buffer we use for filesystem deblocking based on
the filesystem blocksize in the super block. Allows us to read filesystems
with block sizes > 8KB.

branches: 1.3.6; 1.3.14; 1.3.16;
Support for raidframe roots. I hope. (Someone needs to test this.)

branches: 1.2.2;
withough -> without

branches: 1.1.2; 1.1.4; 1.1.6;
move ofwboot & bootblk here; add a Makefile to build ofwboot.net

Use TOOL_GENASSYM. Part of PR toolchain/30350

Clean up tmp files on exit. Fixes PR#23723

branches: 1.1.2; 1.1.4; 1.1.28;
move ofwboot & bootblk here; add a Makefile to build ofwboot.net

Restore backward compatibility of UFS2 with previous NetBSD releases by
disabling support in UFS2 for extended attributes (including ACLs).
Add a new variant of UFS2 called "UFS2ea" that does support extended attributes.
Add new fsck_ffs operations "-c ea" and "-c no-ea" to convert file systems
from UFS2 to UFS2ea and vice-versa (both of which delete all existing extended
attributes in the process).

Move member di_inumber from ufs1_dinode to ulfs1_dinode.

No functional change intended. Tested on sparc64.

Stuff UFS_ in front of a few of ufs's symbols to reduce namespace
pollution. Specifically:
ROOTINO -> UFS_ROOTINO
WINO -> UFS_WINO
NXADDR -> UFS_NXADDR
NDADDR -> UFS_NDADDR
NIADDR -> UFS_NIADDR
MAXSYMLINKLEN -> UFS_MAXSYMLINKLEN
MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency)

Sort out ext2fs's misuse of NDADDR and NIADDR; fortunately, these have
the same values in ext2fs and ffs.

No functional change intended.

branches: 1.7.10; 1.7.20;
Major overhaul:

1) Add support for multiple filesystems including ffsv1, ffsv2, lfsv1 and lfsv2.
(ffsv1 and lfsv2 are known to work. ffv2 support needs more work. lfsv1 is
probably untested.)

2) Only 32-bit Fcodes are used. All 64-bit math uses two cells. This means it
should work on 32-bit machines.

Additional work to consider:

1) Rename genfth.cf to genffs.cf, or keep it and move the ffs symbols to genffs.cf

2) Move the ffs code and the lfs code to different source files so you can
selectively fload the just the filesystems you want.

3) Add code to load ELF files directly instead of relying on the host firmware
ELF support. But, since the binary is now 7250 bytes there may not be room.

branches: 1.6.20;
Move TNF licenses to 2 clause form

branches: 1.5.74; 1.5.76; 1.5.78;
merge ktrace-lwp.

Use RF_PROTECTED_SECTORS instead of a (wrong) magic value.
The bootblock now is able to load ofwboot from a raid1 root.

branches: 1.3.2; 1.3.4;
Track changes introduced when UFS2 was merged, so that this builds again.

try to make this work with ufs1 only.

branches: 1.1.2; 1.1.4;
move ofwboot & bootblk here; add a Makefile to build ofwboot.net

adjust comments slightly

Make these compile again after changes to LFS.

branches: 1.2.10;
Move member di_inumber from ufs1_dinode to ulfs1_dinode.

No functional change intended. Tested on sparc64.

branches: 1.1.2; 1.1.6; 1.1.14; 1.1.24;
Major overhaul:

1) Add support for multiple filesystems including ffsv1, ffsv2, lfsv1 and lfsv2.
(ffsv1 and lfsv2 are known to work. ffv2 support needs more work. lfsv1 is
probably untested.)

2) Only 32-bit Fcodes are used. All 64-bit math uses two cells. This means it
should work on 32-bit machines.

Additional work to consider:

1) Rename genfth.cf to genffs.cf, or keep it and move the ffs symbols to genffs.cf

2) Move the ffs code and the lfs code to different source files so you can
selectively fload the just the filesystems you want.

3) Add code to load ELF files directly instead of relying on the host firmware
ELF support. But, since the binary is now 7250 bytes there may not be room.

adapt to ${CC_WNO_ADDRESS_OF_PACKED_MEMBER}

Simplify CWARNFLAGS to use ${CC_WNO_ADDRESS_OF_PACKED_MEMBER}
which works for both clang and gcc, and remove compiler-specific
equivalents.

bsd.own.mk: rename to CC_WNO_ADDRESS_OF_PACKED_MEMBER

Provide a single variable
CC_WNO_ADDRESS_OF_PACKED_MEMBER
with options for both clang and gcc, to replace
CLANG_NO_ADDR_OF_PACKED_MEMBER
CC_NO_ADDR_OF_PACKED_MEMBER
GCC_NO_ADDR_OF_PACKED_MEMBER

Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.

add support for new GCC 9 warnings that may be too much to fix
right now. new address-of-packed-member and format-overflow
warnings have new GCC_NO_ADDR_OF_PACKED_MEMBER amd
GCC_NO_FORMAT_OVERFLOW variables to remove these warnings.

apply to a bunch of the tree. mostly, these are real bugs that
should be fixed, but in many cases, only by removing the 'packed'
attribute from some structure that doesn't really need it. (i
looked at many different ones, and while perhaps 60-80% were
already properly aligned, it wasn't clear to me that the uses
were always coming from sane data vs network alignment, so it
doesn't seem safe to remove packed without careful research for
each affect struct.) clang already warned (and was not erroring)
for many of these cases, but gcc picked up dozens more.

Uses LFS boot code

branches: 1.17.40;
Add empty LIBCRTI= as LIBCRT0 to build sa programs without installed DESTDIR.

XXX: probabry we should have bsd.saprog.mk or something.

branches: 1.16.112; 1.16.122; 1.16.128;
merge ktrace-lwp.

Define LIB{CRT0,C,CRTBEGIN,CRTEND} as nothing, we don't need to link
against them.

branches: 1.14.10;
Convert `bootxx' to raw binary format which is what the MI installboot(8)
program needs.

branches: 1.13.8; 1.13.12;
bootxx links at 0x340000
2nd-stage boot programs choices: 0x200000,0x300000,0x400000,0x700000 (default: 0x300000)

RELOC -> RELOC_DEFAULT

branches: 1.11.2;
Use ${SIZE}, ${OBJCOPY} rather than size and objcopy.

Add -DLIBSA_NO_CLOSE

branches: 1.9.2; 1.9.4;
Use `-Ttext' switch & collect ld options in LINKFLAGS.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

STRIP -> STRIPFLAG

remove header munging hack.

We don't need `rem.o'.

branches: 1.4.2;
Brr.. libkern.o is way too big for our bootxx..

Make this work with OBJ dirs.

Don't strip bootxx on installation.

Network booting; "oldmon" support; combine all functionality in single images.
Many thanks to Gordon Ross and Chuck Cranor for figuring out and testing
the old monitor (sun4) stuff.

Don't check if an array is (not) NULL, check if the string is empty.

branches: 1.25.22; 1.25.40;
kill extra whitespaces
reviewed by tsutsui@

We need an explicit include of <sys/exec_aout.h> here as well.

- use -Os to shrink bianries
- remove local shrinked memcpy() from bootxx.c bacause
now it's identical with __OPTIMIZE_SIZE__ libsa one
derived from src/common/lib/libc/string/bcopy.c.

Tested netboot from le0 and bootxx + boot from sd0 on SS1+.

Now that the stand/ code was converted to use memcpy(), the private
implementation of bcopy() needs to become memcpy() as well. The
simpler local implementation is needed so that the boot code doesn't
overflow the available space.

bcopy -> memcpy

branches: 1.20.8; 1.20.14;
Remove clause 3 and 4 from TNF licenses

branches: 1.19.40; 1.19.42; 1.19.44;
Make the loadboot() declaration consistent, use char* because it
does pointer arithmetic on the argument.

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

branches: 1.17.10;
De-__P. ANSIify.

Same object code is generated (except in prf.c, where putchar argument
type was changed to int to match the prototype and one instrcution
changes accordingly).

strategy() in ofdev.c was not ANSIified in this sweep, b/c that
changes code generated at call sites!

Nothing in this file needs <sys/time.h>

Get strcpy prototype from libkern.h, not string.h

branches: 1.14.4; 1.14.8; 1.14.14;
merge ktrace-lwp.

branches: 1.13.2;
constify

Ooops, missed this file when changing the bootdevice signature.
Pointed out by Juergen Hannken-Illjes.

branches: 1.11.10; 1.11.16; 1.11.18; 1.11.24;
bootblock.h:
- unify sparc_bbinfo (1064 bytes, with 256 block entries)
and sun68k_bbinfo (296 byte, with 64 block entries)
into shared_bbinfo (512 bytes, with 118 block entries),
which will be also shared by future bbinfo-using platforms
(including macppc)
- add datestamp to *_BBINFO_MAGIC strings, to prevent installboot vs
bootxx version skew.
- add macppc support

*/bootxx.c:
- migrate to new shared_bbinfo structure

installboot:
- add macppc support (still needs applepartmap support and testing)
- improve and add some more warnings & errors to installboot
- implement shared_bbinfo_clearboot() and shared_bbinfo_setboot(), which
perform the majority of the work for bbinfo-using back-ends
(rather than replicating that across multiple back-ends).

convert to <sys/bootblock.h>

Rather than using nlist(3), use a magic structure with a magic
number that installboot(8) can search for. Suggested by cgd@netbsd.org.

branches: 1.8.8; 1.8.10; 1.8.12;
Fix doubled 'the's in comments.

Include <sys/exec.h> instead of <a.out.h>/

branches: 1.6.2; 1.6.4; 1.6.8;
Initialize the heap (for alloc()) explicitly, since the `end' symbol
is not necessarily double-word aligned.

Casts for PROM_LOADADDR which is not casted anymore.

branches: 1.4.4;
Use PROMLIB to streamline this code. This also provides the bulk of
Open Firmware boot code.

Assign my copyrights to TNF.

Call the device close routine when we're done reading.

branches: 1.1.4;
Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Fix typo of "following" in comment.

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.

Remove clause 3 and 4 from TNF licenses

branches: 1.9.40; 1.9.42; 1.9.44;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.8.10;
De-__P. ANSIify.

Same object code is generated (except in prf.c, where putchar argument
type was changed to int to match the prototype and one instrcution
changes accordingly).

strategy() in ofdev.c was not ANSIified in this sweep, b/c that
changes code generated at call sites!

branches: 1.7.2;
s/u_intN_t/uintN_t/

branches: 1.6.4; 1.6.6;
__asm__ -> __asm
__const__ -> const
__inline__ -> inline
__volatile__ -> volatile

merge ktrace-lwp.

branches: 1.4.2;
constify

branches: 1.3.2;
Move all `ID prom' stuff into the prom library.

Casts for PROM_LOADADDR which is not casted anymore.

branches: 1.1.4;
Use PROMLIB to streamline this code. This also provides the bulk of
Open Firmware boot code.

decommission old installboot; it doesn't work on bootxx anymore anyway.

branches: 1.1.2; 1.1.4; 1.1.6;
Rather than using nlist(3), use a magic structure with a magic
number that installboot(8) can search for. Suggested by cgd@netbsd.org.

Add support booting from UFS2 partition for sparc.
Tested on tme emulating SS2.

Closes PR port-sparc/36068.

branches: 1.1.70; 1.1.108; 1.1.124; 1.1.158;
Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Drop 3rd and 4th clauses. Approved by gwr@ (copyright holder).

branches: 1.14.62;
De-__P. ANSIify.

Same object code is generated (except in prf.c, where putchar argument
type was changed to int to match the prototype and one instrcution
changes accordingly).

strategy() in ofdev.c was not ANSIified in this sweep, b/c that
changes code generated at call sites!

branches: 1.13.2; 1.13.6;
It's free -> dealloc, not free -> size_t (hi, Christos :).

free -> dealloc
unsigned -> size_t for alloc/dealloc

branches: 1.11.2;
merge ktrace-lwp.

branches: 1.10.16;
Spell address with two d's. Inspired by similar changes in OpenBSD,
originating from Jonathon Gray and forwarded by jmc@openbsd.

branches: 1.9.24;
Align the DVMA mapping on NBPSG boundaries.

branches: 1.8.2;
Rearrange previous slightly to avoid getting values that are too low;
the only goal here is to leave some space for the stack.

Account for the stack when computing the base va. This makes the
bootblocks work again on the 4/110.

branches: 1.6.2;
dvma_init: avoid integer variable overflow

Determine base address at run time.

Nuke unused code.

branches: 1.3.2; 1.3.8;
Use PROMLIB to streamline this code. This also provides the bulk of
Open Firmware boot code.

Move around a few PTE-manipulating macros.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

branches: 1.1.2; 1.1.4; 1.1.10;
factor out the floppy detection code.

branches: 1.1.2; 1.1.4; 1.1.10;
factor out the floppy detection code.

kill sprintf

branches: 1.8.34; 1.8.44; 1.8.50;
Remove clause 3 and 4 from TNF licenses

branches: 1.7.74; 1.7.76; 1.7.78;
merge ktrace-lwp.

Move common Sun ID PROM definitions into <dev/sun/idprom.h>.

Make pmap_map() work on sun4 machines.
Thanks to Matthew and Julian for finding this.

branches: 1.4.2; 1.4.6; 1.4.8;
pmap_map_srmmu: use the `map-pages' Forth word to let prom map all pages
at once.

pmap_map(): round up size to compute # of pages.

branches: 1.2.2;
Drop redundant argument to printf("\n"). Obvious pasto.

Add support for large kernels by searching for a physical memory segment
to fit it in. As a bonus, the kernel is now also mapped to the virtual
address (i.e. KERNBASE) it is linked at.

remove shadowed (common) declarations of various libsa variables
from various boot programs. for macppc and zaurus, avoid building
with -fcommon any more.

branches: 1.10.4;
rootpath is owned by libsa

branches: 1.9.74;
Drop 3rd and 4th clauses. Approved by gwr@ (copyright holder).

branches: 1.8.62;
De-__P. ANSIify.

Same object code is generated (except in prf.c, where putchar argument
type was changed to int to match the prototype and one instrcution
changes accordingly).

strategy() in ofdev.c was not ANSIified in this sweep, b/c that
changes code generated at call sites!

branches: 1.7.18; 1.7.32; 1.7.36;
Note which network boot protocol we're trying.

- Enable NFS_BOOT_DHCP (as well as the existing NFS_BOOT_BOOTPARAM).
DHCP/Bootp is tried before bootparams in the MI nfs mountroot code.
- Change the order in boot.net from "bootparams then bootp" to
"bootp then bootparams", to be consistent with the MI nfs mountroot code.

(Consistency with other NetBSD ports is good, and things still work
for sites that run bootparams but not dhcp/bootp, although I'd argue that
dhcp/bootp is much easier to setup and manage and is easier to debug
than our rpc.bootparamd...)

Per discussion with Paul Kranenburg and Matt Green.

branches: 1.5.8; 1.5.12;
Try both the BOOTPARAMS and BOOTP protocol (in that order). This also
takes care of PR#8738.

branches: 1.4.2;
include <lib/libkern/libkern.h> for intoa()/inet_ntoa()

branches: 1.3.4;
Use PROMLIB to streamline this code. This also provides the bulk of
Open Firmware boot code.

Don't use <lib/libsa/if_ether.h> anymore. This was a copy of the
old (before ARP changes) <netinet/if_ether.h>, intended for
temporary use.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Drop 3rd and 4th clauses. Approved by gwr@ (copyright holder).

bzero -> memset

branches: 1.10.2;
Replace time_t values in libsa sources with the following two types
to avoid unnecessary 64 bit ops which would make binaries larger:

satime_t (currently unsigned int):
numbers in seconds returned by the machine dependent getsecs() function
which are used to measure relative time

saseconds_t (currently int):
numbers in seconds used to specify timeout to network drivers

Per discussion on current-users.

branches: 1.9.58; 1.9.62; 1.9.70;
De-__P. ANSIify.

Same object code is generated (except in prf.c, where putchar argument
type was changed to int to match the prototype and one instrcution
changes accordingly).

strategy() in ofdev.c was not ANSIified in this sweep, b/c that
changes code generated at call sites!

branches: 1.8.4; 1.8.8;
merge ktrace-lwp.

branches: 1.7.16;
Replace myetheraddr() by prom_getether(). Convert the file descriptor to
deivce tree node (if possible) to support the "local-mac-address' property.

branches: 1.6.2;
Explicit cast for netif.

Approved by: Matthias Drochner <drochner@netbsd.org>

Move all `ID prom' stuff into the prom library.

Remove inclusion of unused headers.

branches: 1.3.8; 1.3.10; 1.3.14;
Use PROMLIB to streamline this code. This also provides the bulk of
Open Firmware boot code.

Don't use <lib/libsa/if_ether.h> anymore. This was a copy of the
old (before ARP changes) <netinet/if_ether.h>, intended for
temporary use.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

switch to sys/conf/newvers_stand.sh for generation of version file
sync the message in boot/version with other ports using newvers_stand.sh

branches: 1.2.8; 1.2.18;
look for /^Version:/ to find the boot block version, rather than using
the "version" file's RCS id (which is useless for branches).

branches: 1.1.8;
Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

sparc promdev.c: Nix trailing whitespace. Fix intertoken space.

No functional change intended.

Move the check for RAID boot into the strategy routines.
This avoids a Data Access Exception at boot time on older PROM versions.

branches: 1.29.36;
don't compare string pointers against \0

branches: 1.28.14;
Virtio block devices may have a "scsi" device_type instead of "block",
so accept both.

branches: 1.27.6; 1.27.10; 1.27.14;
Avoid buffer overruns (no practical difference for any sane firmware
properties, but better be safe).

branches: 1.26.8; 1.26.18; 1.26.22;
Fix printf formats in -DDEBUG_PROM to make it compile.

XXX: print daddr_t after casting it to int since libsa is not built
with LIBSA_PRINTF_LONGLONG_SUPPORT by default.

factor out the floppy detection code.

Don't LOAD_BACKWARDS for floppies.

branches: 1.23.2; 1.23.4;
Don't try to read disklabel to check FS_RAID on floppy boot since
reopening floppy could cause Data Access Exception later.
Fixes PR port-sparc/42186, ok'ed by mrg@.

Replace time_t values in libsa sources with the following two types
to avoid unnecessary 64 bit ops which would make binaries larger:

satime_t (currently unsigned int):
numbers in seconds returned by the machine dependent getsecs() function
which are used to measure relative time

saseconds_t (currently int):
numbers in seconds used to specify timeout to network drivers

Per discussion on current-users.

branches: 1.21.4; 1.21.12; 1.21.14; 1.21.20;
Add support booting from UFS2 partition for sparc.
Tested on tme emulating SS2.

Closes PR port-sparc/36068.

branches: 1.20.8; 1.20.24; 1.20.58;
De-__P. ANSIify.

Same object code is generated (except in prf.c, where putchar argument
type was changed to int to match the prototype and one instrcution
changes accordingly).

strategy() in ofdev.c was not ANSIified in this sweep, b/c that
changes code generated at call sites!

Add support to the 2nd stage boot loader for booting from RAID
partitions. We do this by checking the NetBSD label to see if the
boot partition is of type RAID. If so, we offset reads from the
disk so that the kernel image can be read.

Note that with this code, sun4 machines (with PROM monitor) will only
boot from RAID if the RAID partition is the first one on the disk.

Tested on a SPARCstation 20, a SPARCstation 2 and a 4/330.

branches: 1.18.2;
struct filesystem is now struct fs_ops, since i don't know when.

branches: 1.17.4; 1.17.6; 1.17.8; 1.17.12; 1.17.14;
merge ktrace-lwp.

branches: 1.16.2;
Do not rely on strings to which pointers are returned by promlib functions
to persist accross further promlib calls.
This fixes (at least) bootjs.net.

branches: 1.15.6; 1.15.8; 1.15.14;
sprinkle some: libkern.h, some -ffreestanding and some cast-y goodness,
all for the pleasure of the gcc 3 :-)

branches: 1.14.2;
Move all `ID prom' stuff into the prom library.

remove trailing \n in panic(). approved perry.

branches: 1.12.4;
getprop* -> PROM_getprop*

branches: 1.11.2;
avoid an unused variable warning

branches: 1.10.2;
* Missing PROM_OPENFIRM: case in getdevtype() and prom_getether().
* In getdevtype(), use prom_instance_to_package() to get to the device node.

From Valeriy E. Ushakov on port-sparc@netbsd.org.

branches: 1.9.2;
On V0 proms, read the `idprom' property at the root node to get to the
station's ethernet address.

branches: 1.8.2; 1.8.12;
Include <sys/systm.h> to get prototype for memset.

branches: 1.7.4;
Use PROMLIB to streamline this code. This also provides the bulk of
Open Firmware boot code.

Move around a few PTE-manipulating macros.

rename getprop routine to avoid conflict with the one in bsd_openprom.h

Can't afford a 1K buffer in the data segment.

add support for finding kernels other than `netbsd'. the list is:
netbsd, netbsd.gz, netbsd.old, netbsd.old.gz, onetbsd, onetbsd.gz
and vmunix, in that order. this is used by the bootable `bootfs'
image (suitable for floppies and cdroms), which contains a netbsd.gz
by default.
while i'm here, fix up `boot [device] -s' which got broke when
booting kernels with a - in their name was fixed..

Properly split boot spec into file name and arguments.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

factor out the floppy detection code.

Don't LOAD_BACKWARDS for floppies.

Replace time_t values in libsa sources with the following two types
to avoid unnecessary 64 bit ops which would make binaries larger:

satime_t (currently unsigned int):
numbers in seconds returned by the machine dependent getsecs() function
which are used to measure relative time

saseconds_t (currently int):
numbers in seconds used to specify timeout to network drivers

Per discussion on current-users.

branches: 1.13.8;
Remove clause 3 and 4 from TNF licenses

branches: 1.12.58; 1.12.60; 1.12.62;
De-__P. ANSIify.

Same object code is generated (except in prf.c, where putchar argument
type was changed to int to match the prototype and one instrcution
changes accordingly).

strategy() in ofdev.c was not ANSIified in this sweep, b/c that
changes code generated at call sites!

branches: 1.11.4; 1.11.8;
merge ktrace-lwp.

branches: 1.10.2;
Do not rely on strings to which pointers are returned by promlib functions
to persist accross further promlib calls.
This fixes (at least) bootjs.net.

branches: 1.9.2; 1.9.8; 1.9.10; 1.9.16;
Add support for large kernels by searching for a physical memory segment
to fit it in. As a bonus, the kernel is now also mapped to the virtual
address (i.e. KERNBASE) it is linked at.

Move all `ID prom' stuff into the prom library.

`xmit' & `recv' return `ssize_t'

Add bootinfo struct to pass symtab arguments.
XXX: This does not work with ELF yet, but I don't understand why...

branches: 1.5.4;
Use PROMLIB to streamline this code. This also provides the bulk of
Open Firmware boot code.

Move around a few PTE-manipulating macros.

Assign my copyrights to TNF.

Define a new DDB magic number to pass to the kernel.

branches: 1.1.4;
Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

Deal with _edata not being 4-byte aligned

branches: 1.11.44; 1.11.64;
Remove clause 3 and 4 from TNF licenses

branches: 1.10.74; 1.10.76; 1.10.78;
merge ktrace-lwp.

Add openfirmware() entry point.

branches: 1.8.2;
Only JavaStations have OFW, so set cpu type to CPU_SUN4M in is_openfirm.

Skip relocation if OpenFirmware already loaded us at the right address.

branches: 1.6.2; 1.6.8;
Save OpenFirmware client interface handler in romp, where it is
expected to be. GC unused opf_romp variable.

branches: 1.5.2;
When relocating, there's no need to copy the BSS area.

branches: 1.4.8;
Use <machine/asm.h> to construct C identifiers.

Transfer copyright to TNF.

Use PROMLIB to streamline this code. This also provides the bulk of
Open Firmware boot code.

Completely rearrange SPARC boot code, modeling it after alpha and hp300
ports' boot programs. Also added support for loading gzip'd kernels,
useful for booting from floppy. From Jason R. Thorpe <thorpej@nas.nasa.gov>,
tested by Matthew Green <mrg@eterna.com.au> and Erik Fair <fair@clock.org>.

PR 56829: Fix "map" call to Open Firmware and no longer mess up OFW stack.
This unbreaks miniroot.fs.

Thanks to Björn Johannesson for originally pointing this out and to
mrg@ for pointing out the actual issue and suggesting a fix.

Add more details about the boot device as a new bootinfo record type.
From within the bootloader, when we have readily accessible instance
handles of the boot device, it is easy to query more details like SCSI
LUN, target, and FC-AL wwn from the firmware.
The kernel later would have a hard time getting theses, but can make good
use to match the boot device.

Add compile time asserts to make sure we have properly picked up types
for a 64bit cpu and 64bit openfirmware.

Fix kernel loading failures from partitions started from over first
4GB of disks on sparc64.

branches: 1.13.4; 1.13.6;
The "close" call returns no arguments

branches: 1.12.4; 1.12.10; 1.12.14; 1.12.16; 1.12.18; 1.12.24;
WARNSfy and fix dumb bugs noticed by warnings.

branches: 1.11.4; 1.11.6;
Make ofwboot can handle over 4GB physical memory by using OpenFirmware
calls properly, and some cosmetic changes. Idea from OpenBSD.

branches: 1.10.20; 1.10.30; 1.10.40; 1.10.44;
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.

Some aliases (like "cdrom") expand to a path that includes a partition -
deal with them.

branches: 1.8.14; 1.8.22; 1.8.32; 1.8.34;
De-__P. ANSIify.

Same object code is generated (except in prf.c, where putchar argument
type was changed to int to match the prototype and one instrcution
changes accordingly).

strategy() in ofdev.c was not ANSIified in this sweep, b/c that
changes code generated at call sites!

branches: 1.7.2; 1.7.6;
Alter sparc64 bootstrap:

- Map kernel with permanent 4MB pages to ease kernel's life
- Pass mappings down to the kernel as part of bootinfo structure
- Try to detect whether kernel is aware of the new boot scheme or not and
in the former case map the kernel using OpenFirmware and do not pass
bootinfo structure
- Provide -C switch to force compatibility mode if autodetection fails
- While on this, fix PR #25763: If load of default kernel fails, try one by
one those specified in compiled-in list

Much inspired by the bootloader of FreeBSD, with some imported code from there.