Force -fno-stack-protector on "boot block" that absolutely can't have
a stack protector (probably not even a -fstack-protector-strong) because
the bloat would render them unuseable. This also means the system
compiler can now take on any more it wants, and all the pieces which
can't use the stack protector are properly marked.
ok kettenis

Allow strip and objcopy to be overrridden, for easier cross-buildability. NFC

__remq.S will be needed soon.

Use arc4 to bit-spread the 512-byte random buffer over the .openbsd.randomdata
section, which has grown a fair bit with the introduction of retguard.
Mortimer discovered the repeated 512-byte sequence as retguard keys, and
this resolves the issue. (Chacha does not fit on the media, so 1.5K early
drop RC4 is hopefully sufficient in our KARL link universe)
Version crank the bootblocks. sysupgrade -s will install new bootblocks.
ok djm mortimer

Remove -S from install commands

As of usr.bin/xinstall/install.c revision 1.68, -S is a no-op and
install(1) will always create files safely, thus clean the option usage
from the tree.

Diff from Lauri Tirkkonen <lotheac at iki dot fi>, thanks.

destroy lint remnants.
okay millert@ deraadt@

Prep for relro: make sure it's off for any non-PIE stand/ program

ok millert@ kettenis@

Create a MOP version of the network boot loader; verified to boot on DEC 3000.

Fix alpha bootblocks: Compile with -fno-pie, link with -nopie.

tested by naddy@; ok deraadt@

Add NOPIE= bits for sys/arch/*/stand to ensure that bootblocks will always be
built with -fno-pie. This gets the hairiest part of PIE out of the way ...

ok deraadt@

Update alpha bootblocks to use MI loadfile code instead of rolling our own;
also print a meaningful message if the kernel load address conflicts with the
SRM memory areas, so that we can know which kernel load address to use.

Requires `make includes' before attempting to build bootblocks.

Tweaks mk@, `nice' deraadt@

we want strlen.c

libkern, begone. Move to a new mechanism where config(8)'s "file"
directive can select between MI and MD versions of these files. At
the same time, adjust the boot programs to pick exactly what they need,
instead of the 7 or 8 mechanisms previously used.

There will be some fallout from this, but testing it all by myself is a
ridiculously slow process; it will be finished in-tree.

Various developers were very nice and avoided making fun of me when I
was gibbering in the corner..

enable -DSUN_BOOTPARAMS; henning@ martin@ miod@ ok

suggested by art, go for it deraadt, ok miod@
remove ECOFF support

Use objcopy to build the bootblocks instead of a hairy dd.
Bootblock built with elf utils work now.

Enable loading of ELF kernels.

branches: 1.3.14;
Updates from NetBSD (cgd):
- seriously clean up makefiles. use libsa/libkern/libz as appropriate,
and don't build the various sources into objs locally by specifying
them directly in the Makefiles.
- move the secondary bootstrap down (to 0x20004000), and add a HEAP_LIMIT
so that we don't exhaust our bootstrap address space (when using the new,
not yet checked in, memory allocator).
- clean up installboot, deal with secondary bootstrap programs not on
'a' partition or 'a' partition not starting at 0.
- add padding to structures in bbinfo.h for future expansion. Add a
netbbinfo structure to allow netboot information to be hard-coded in
network boot blocks, so that they can be made to work even on systems
with firmware which doesn't support the new "ethernet address in boot
device" convention.
- move the sync() calls to the correct place in installboot.c
- remove a kludge in disk.c which was breaking multiple opens/closes
of the disk
- allow netboot ethernet address to be hard-coded into binary so that
machines with old firmware which doesn't pass it in the boot device
can work. Assume that if the ethernet address isn't passed in, it
uses the old (dain-bramaged) 'read' interface works on my 3000/300LX.
- Add setnetbootinfo, a program to hard code an ethernet address into
a network boot.
- move an 'int debug;' into boot.c (it was the only diff between boot.c and
netboot.c), and nuke netboot.c (using boot.c instead for net boot blocks).
- add cd9660 file system ops to the file system ops table in filesystem.c
- if a file name is given (i.e. BOOTED_FILE) is set, boot only that file,
but if not then try to boot "bsd", "bsd.bak", "bsd.old", and
"obsd" (in that order) until one is found or until the list of names
is exhausted.
- add support for reading gzipped kernels.
- use strerror() to print errors, rather than just printing error numbers
- if no disk label exists, fake one up
- slightly relax the block size checks in bootxx.c; they were a bit to paranoid

And local changes:
- don't build a copy of libsa/libkern/libz for each boot prog,
just build a single copy and use it for everything.

Make compilable in our env. Involved disabling ELF support,
reordering includes (due to -Wstrict-prototypes) and other -Wall goo

Merge to NetBSD 961020. Retained our kernel APIs where NetBSD has changed.
-Wall -Wstrict-prototypes -Wmissing-prototypes too.

Allow strip and objcopy to be overrridden, for easier cross-buildability. NFC

__remq.S will be needed soon.

Use arc4 to bit-spread the 512-byte random buffer over the .openbsd.randomdata
section, which has grown a fair bit with the introduction of retguard.
Mortimer discovered the repeated 512-byte sequence as retguard keys, and
this resolves the issue. (Chacha does not fit on the media, so 1.5K early
drop RC4 is hopefully sufficient in our KARL link universe)
Version crank the bootblocks. sysupgrade -s will install new bootblocks.
ok djm mortimer

Remove -S from install commands

As of usr.bin/xinstall/install.c revision 1.68, -S is a no-op and
install(1) will always create files safely, thus clean the option usage
from the tree.

Diff from Lauri Tirkkonen <lotheac at iki dot fi>, thanks.

destroy lint remnants.
okay millert@ deraadt@

Prep for relro: make sure it's off for any non-PIE stand/ program

ok millert@ kettenis@

Create a MOP version of the network boot loader; verified to boot on DEC 3000.

Fix alpha bootblocks: Compile with -fno-pie, link with -nopie.

tested by naddy@; ok deraadt@

Add NOPIE= bits for sys/arch/*/stand to ensure that bootblocks will always be
built with -fno-pie. This gets the hairiest part of PIE out of the way ...

ok deraadt@

Update alpha bootblocks to use MI loadfile code instead of rolling our own;
also print a meaningful message if the kernel load address conflicts with the
SRM memory areas, so that we can know which kernel load address to use.

Requires `make includes' before attempting to build bootblocks.

Tweaks mk@, `nice' deraadt@

we want strlen.c

libkern, begone. Move to a new mechanism where config(8)'s "file"
directive can select between MI and MD versions of these files. At
the same time, adjust the boot programs to pick exactly what they need,
instead of the 7 or 8 mechanisms previously used.

There will be some fallout from this, but testing it all by myself is a
ridiculously slow process; it will be finished in-tree.

Various developers were very nice and avoided making fun of me when I
was gibbering in the corner..

enable -DSUN_BOOTPARAMS; henning@ martin@ miod@ ok

suggested by art, go for it deraadt, ok miod@
remove ECOFF support

Use objcopy to build the bootblocks instead of a hairy dd.
Bootblock built with elf utils work now.

Enable loading of ELF kernels.

branches: 1.3.14;
Updates from NetBSD (cgd):
- seriously clean up makefiles. use libsa/libkern/libz as appropriate,
and don't build the various sources into objs locally by specifying
them directly in the Makefiles.
- move the secondary bootstrap down (to 0x20004000), and add a HEAP_LIMIT
so that we don't exhaust our bootstrap address space (when using the new,
not yet checked in, memory allocator).
- clean up installboot, deal with secondary bootstrap programs not on
'a' partition or 'a' partition not starting at 0.
- add padding to structures in bbinfo.h for future expansion. Add a
netbbinfo structure to allow netboot information to be hard-coded in
network boot blocks, so that they can be made to work even on systems
with firmware which doesn't support the new "ethernet address in boot
device" convention.
- move the sync() calls to the correct place in installboot.c
- remove a kludge in disk.c which was breaking multiple opens/closes
of the disk
- allow netboot ethernet address to be hard-coded into binary so that
machines with old firmware which doesn't pass it in the boot device
can work. Assume that if the ethernet address isn't passed in, it
uses the old (dain-bramaged) 'read' interface works on my 3000/300LX.
- Add setnetbootinfo, a program to hard code an ethernet address into
a network boot.
- move an 'int debug;' into boot.c (it was the only diff between boot.c and
netboot.c), and nuke netboot.c (using boot.c instead for net boot blocks).
- add cd9660 file system ops to the file system ops table in filesystem.c
- if a file name is given (i.e. BOOTED_FILE) is set, boot only that file,
but if not then try to boot "bsd", "bsd.bak", "bsd.old", and
"obsd" (in that order) until one is found or until the list of names
is exhausted.
- add support for reading gzipped kernels.
- use strerror() to print errors, rather than just printing error numbers
- if no disk label exists, fake one up
- slightly relax the block size checks in bootxx.c; they were a bit to paranoid

And local changes:
- don't build a copy of libsa/libkern/libz for each boot prog,
just build a single copy and use it for everything.

Make compilable in our env. Involved disabling ELF support,
reordering includes (due to -Wstrict-prototypes) and other -Wall goo

Merge to NetBSD 961020. Retained our kernel APIs where NetBSD has changed.
-Wall -Wstrict-prototypes -Wmissing-prototypes too.

__remq.S will be needed soon.

Use arc4 to bit-spread the 512-byte random buffer over the .openbsd.randomdata
section, which has grown a fair bit with the introduction of retguard.
Mortimer discovered the repeated 512-byte sequence as retguard keys, and
this resolves the issue. (Chacha does not fit on the media, so 1.5K early
drop RC4 is hopefully sufficient in our KARL link universe)
Version crank the bootblocks. sysupgrade -s will install new bootblocks.
ok djm mortimer

Remove -S from install commands

As of usr.bin/xinstall/install.c revision 1.68, -S is a no-op and
install(1) will always create files safely, thus clean the option usage
from the tree.

Diff from Lauri Tirkkonen <lotheac at iki dot fi>, thanks.

destroy lint remnants.
okay millert@ deraadt@

Prep for relro: make sure it's off for any non-PIE stand/ program

ok millert@ kettenis@

Create a MOP version of the network boot loader; verified to boot on DEC 3000.

Fix alpha bootblocks: Compile with -fno-pie, link with -nopie.

tested by naddy@; ok deraadt@

Add NOPIE= bits for sys/arch/*/stand to ensure that bootblocks will always be
built with -fno-pie. This gets the hairiest part of PIE out of the way ...

ok deraadt@

Update alpha bootblocks to use MI loadfile code instead of rolling our own;
also print a meaningful message if the kernel load address conflicts with the
SRM memory areas, so that we can know which kernel load address to use.

Requires `make includes' before attempting to build bootblocks.

Tweaks mk@, `nice' deraadt@

we want strlen.c

libkern, begone. Move to a new mechanism where config(8)'s "file"
directive can select between MI and MD versions of these files. At
the same time, adjust the boot programs to pick exactly what they need,
instead of the 7 or 8 mechanisms previously used.

There will be some fallout from this, but testing it all by myself is a
ridiculously slow process; it will be finished in-tree.

Various developers were very nice and avoided making fun of me when I
was gibbering in the corner..

enable -DSUN_BOOTPARAMS; henning@ martin@ miod@ ok

suggested by art, go for it deraadt, ok miod@
remove ECOFF support

Use objcopy to build the bootblocks instead of a hairy dd.
Bootblock built with elf utils work now.

Enable loading of ELF kernels.

branches: 1.3.14;
Updates from NetBSD (cgd):
- seriously clean up makefiles. use libsa/libkern/libz as appropriate,
and don't build the various sources into objs locally by specifying
them directly in the Makefiles.
- move the secondary bootstrap down (to 0x20004000), and add a HEAP_LIMIT
so that we don't exhaust our bootstrap address space (when using the new,
not yet checked in, memory allocator).
- clean up installboot, deal with secondary bootstrap programs not on
'a' partition or 'a' partition not starting at 0.
- add padding to structures in bbinfo.h for future expansion. Add a
netbbinfo structure to allow netboot information to be hard-coded in
network boot blocks, so that they can be made to work even on systems
with firmware which doesn't support the new "ethernet address in boot
device" convention.
- move the sync() calls to the correct place in installboot.c
- remove a kludge in disk.c which was breaking multiple opens/closes
of the disk
- allow netboot ethernet address to be hard-coded into binary so that
machines with old firmware which doesn't pass it in the boot device
can work. Assume that if the ethernet address isn't passed in, it
uses the old (dain-bramaged) 'read' interface works on my 3000/300LX.
- Add setnetbootinfo, a program to hard code an ethernet address into
a network boot.
- move an 'int debug;' into boot.c (it was the only diff between boot.c and
netboot.c), and nuke netboot.c (using boot.c instead for net boot blocks).
- add cd9660 file system ops to the file system ops table in filesystem.c
- if a file name is given (i.e. BOOTED_FILE) is set, boot only that file,
but if not then try to boot "bsd", "bsd.bak", "bsd.old", and
"obsd" (in that order) until one is found or until the list of names
is exhausted.
- add support for reading gzipped kernels.
- use strerror() to print errors, rather than just printing error numbers
- if no disk label exists, fake one up
- slightly relax the block size checks in bootxx.c; they were a bit to paranoid

And local changes:
- don't build a copy of libsa/libkern/libz for each boot prog,
just build a single copy and use it for everything.

Make compilable in our env. Involved disabling ELF support,
reordering includes (due to -Wstrict-prototypes) and other -Wall goo

Merge to NetBSD 961020. Retained our kernel APIs where NetBSD has changed.
-Wall -Wstrict-prototypes -Wmissing-prototypes too.

Use arc4 to bit-spread the 512-byte random buffer over the .openbsd.randomdata
section, which has grown a fair bit with the introduction of retguard.
Mortimer discovered the repeated 512-byte sequence as retguard keys, and
this resolves the issue. (Chacha does not fit on the media, so 1.5K early
drop RC4 is hopefully sufficient in our KARL link universe)
Version crank the bootblocks. sysupgrade -s will install new bootblocks.
ok djm mortimer

Remove -S from install commands

As of usr.bin/xinstall/install.c revision 1.68, -S is a no-op and
install(1) will always create files safely, thus clean the option usage
from the tree.

Diff from Lauri Tirkkonen <lotheac at iki dot fi>, thanks.

destroy lint remnants.
okay millert@ deraadt@

Prep for relro: make sure it's off for any non-PIE stand/ program

ok millert@ kettenis@

Create a MOP version of the network boot loader; verified to boot on DEC 3000.

Fix alpha bootblocks: Compile with -fno-pie, link with -nopie.

tested by naddy@; ok deraadt@

Add NOPIE= bits for sys/arch/*/stand to ensure that bootblocks will always be
built with -fno-pie. This gets the hairiest part of PIE out of the way ...

ok deraadt@

Update alpha bootblocks to use MI loadfile code instead of rolling our own;
also print a meaningful message if the kernel load address conflicts with the
SRM memory areas, so that we can know which kernel load address to use.

Requires `make includes' before attempting to build bootblocks.

Tweaks mk@, `nice' deraadt@

we want strlen.c

libkern, begone. Move to a new mechanism where config(8)'s "file"
directive can select between MI and MD versions of these files. At
the same time, adjust the boot programs to pick exactly what they need,
instead of the 7 or 8 mechanisms previously used.

There will be some fallout from this, but testing it all by myself is a
ridiculously slow process; it will be finished in-tree.

Various developers were very nice and avoided making fun of me when I
was gibbering in the corner..

enable -DSUN_BOOTPARAMS; henning@ martin@ miod@ ok

suggested by art, go for it deraadt, ok miod@
remove ECOFF support

Use objcopy to build the bootblocks instead of a hairy dd.
Bootblock built with elf utils work now.

Enable loading of ELF kernels.

branches: 1.3.14;
Updates from NetBSD (cgd):
- seriously clean up makefiles. use libsa/libkern/libz as appropriate,
and don't build the various sources into objs locally by specifying
them directly in the Makefiles.
- move the secondary bootstrap down (to 0x20004000), and add a HEAP_LIMIT
so that we don't exhaust our bootstrap address space (when using the new,
not yet checked in, memory allocator).
- clean up installboot, deal with secondary bootstrap programs not on
'a' partition or 'a' partition not starting at 0.
- add padding to structures in bbinfo.h for future expansion. Add a
netbbinfo structure to allow netboot information to be hard-coded in
network boot blocks, so that they can be made to work even on systems
with firmware which doesn't support the new "ethernet address in boot
device" convention.
- move the sync() calls to the correct place in installboot.c
- remove a kludge in disk.c which was breaking multiple opens/closes
of the disk
- allow netboot ethernet address to be hard-coded into binary so that
machines with old firmware which doesn't pass it in the boot device
can work. Assume that if the ethernet address isn't passed in, it
uses the old (dain-bramaged) 'read' interface works on my 3000/300LX.
- Add setnetbootinfo, a program to hard code an ethernet address into
a network boot.
- move an 'int debug;' into boot.c (it was the only diff between boot.c and
netboot.c), and nuke netboot.c (using boot.c instead for net boot blocks).
- add cd9660 file system ops to the file system ops table in filesystem.c
- if a file name is given (i.e. BOOTED_FILE) is set, boot only that file,
but if not then try to boot "bsd", "bsd.bak", "bsd.old", and
"obsd" (in that order) until one is found or until the list of names
is exhausted.
- add support for reading gzipped kernels.
- use strerror() to print errors, rather than just printing error numbers
- if no disk label exists, fake one up
- slightly relax the block size checks in bootxx.c; they were a bit to paranoid

And local changes:
- don't build a copy of libsa/libkern/libz for each boot prog,
just build a single copy and use it for everything.

Make compilable in our env. Involved disabling ELF support,
reordering includes (due to -Wstrict-prototypes) and other -Wall goo

Merge to NetBSD 961020. Retained our kernel APIs where NetBSD has changed.
-Wall -Wstrict-prototypes -Wmissing-prototypes too.

Remove -S from install commands

As of usr.bin/xinstall/install.c revision 1.68, -S is a no-op and
install(1) will always create files safely, thus clean the option usage
from the tree.

Diff from Lauri Tirkkonen <lotheac at iki dot fi>, thanks.

destroy lint remnants.
okay millert@ deraadt@

Prep for relro: make sure it's off for any non-PIE stand/ program

ok millert@ kettenis@

Create a MOP version of the network boot loader; verified to boot on DEC 3000.

Fix alpha bootblocks: Compile with -fno-pie, link with -nopie.

tested by naddy@; ok deraadt@

Add NOPIE= bits for sys/arch/*/stand to ensure that bootblocks will always be
built with -fno-pie. This gets the hairiest part of PIE out of the way ...

ok deraadt@

Update alpha bootblocks to use MI loadfile code instead of rolling our own;
also print a meaningful message if the kernel load address conflicts with the
SRM memory areas, so that we can know which kernel load address to use.

Requires `make includes' before attempting to build bootblocks.

Tweaks mk@, `nice' deraadt@

we want strlen.c

libkern, begone. Move to a new mechanism where config(8)'s "file"
directive can select between MI and MD versions of these files. At
the same time, adjust the boot programs to pick exactly what they need,
instead of the 7 or 8 mechanisms previously used.

There will be some fallout from this, but testing it all by myself is a
ridiculously slow process; it will be finished in-tree.

Various developers were very nice and avoided making fun of me when I
was gibbering in the corner..

enable -DSUN_BOOTPARAMS; henning@ martin@ miod@ ok

suggested by art, go for it deraadt, ok miod@
remove ECOFF support

Use objcopy to build the bootblocks instead of a hairy dd.
Bootblock built with elf utils work now.

Enable loading of ELF kernels.

branches: 1.3.14;
Updates from NetBSD (cgd):
- seriously clean up makefiles. use libsa/libkern/libz as appropriate,
and don't build the various sources into objs locally by specifying
them directly in the Makefiles.
- move the secondary bootstrap down (to 0x20004000), and add a HEAP_LIMIT
so that we don't exhaust our bootstrap address space (when using the new,
not yet checked in, memory allocator).
- clean up installboot, deal with secondary bootstrap programs not on
'a' partition or 'a' partition not starting at 0.
- add padding to structures in bbinfo.h for future expansion. Add a
netbbinfo structure to allow netboot information to be hard-coded in
network boot blocks, so that they can be made to work even on systems
with firmware which doesn't support the new "ethernet address in boot
device" convention.
- move the sync() calls to the correct place in installboot.c
- remove a kludge in disk.c which was breaking multiple opens/closes
of the disk
- allow netboot ethernet address to be hard-coded into binary so that
machines with old firmware which doesn't pass it in the boot device
can work. Assume that if the ethernet address isn't passed in, it
uses the old (dain-bramaged) 'read' interface works on my 3000/300LX.
- Add setnetbootinfo, a program to hard code an ethernet address into
a network boot.
- move an 'int debug;' into boot.c (it was the only diff between boot.c and
netboot.c), and nuke netboot.c (using boot.c instead for net boot blocks).
- add cd9660 file system ops to the file system ops table in filesystem.c
- if a file name is given (i.e. BOOTED_FILE) is set, boot only that file,
but if not then try to boot "bsd", "bsd.bak", "bsd.old", and
"obsd" (in that order) until one is found or until the list of names
is exhausted.
- add support for reading gzipped kernels.
- use strerror() to print errors, rather than just printing error numbers
- if no disk label exists, fake one up
- slightly relax the block size checks in bootxx.c; they were a bit to paranoid

And local changes:
- don't build a copy of libsa/libkern/libz for each boot prog,
just build a single copy and use it for everything.

Make compilable in our env. Involved disabling ELF support,
reordering includes (due to -Wstrict-prototypes) and other -Wall goo

Merge to NetBSD 961020. Retained our kernel APIs where NetBSD has changed.
-Wall -Wstrict-prototypes -Wmissing-prototypes too.

destroy lint remnants.
okay millert@ deraadt@

Prep for relro: make sure it's off for any non-PIE stand/ program

ok millert@ kettenis@

Create a MOP version of the network boot loader; verified to boot on DEC 3000.

Fix alpha bootblocks: Compile with -fno-pie, link with -nopie.

tested by naddy@; ok deraadt@

Add NOPIE= bits for sys/arch/*/stand to ensure that bootblocks will always be
built with -fno-pie. This gets the hairiest part of PIE out of the way ...

ok deraadt@

Update alpha bootblocks to use MI loadfile code instead of rolling our own;
also print a meaningful message if the kernel load address conflicts with the
SRM memory areas, so that we can know which kernel load address to use.

Requires `make includes' before attempting to build bootblocks.

Tweaks mk@, `nice' deraadt@

we want strlen.c

libkern, begone. Move to a new mechanism where config(8)'s "file"
directive can select between MI and MD versions of these files. At
the same time, adjust the boot programs to pick exactly what they need,
instead of the 7 or 8 mechanisms previously used.

There will be some fallout from this, but testing it all by myself is a
ridiculously slow process; it will be finished in-tree.

Various developers were very nice and avoided making fun of me when I
was gibbering in the corner..

enable -DSUN_BOOTPARAMS; henning@ martin@ miod@ ok

suggested by art, go for it deraadt, ok miod@
remove ECOFF support

Use objcopy to build the bootblocks instead of a hairy dd.
Bootblock built with elf utils work now.

Enable loading of ELF kernels.

branches: 1.3.14;
Updates from NetBSD (cgd):
- seriously clean up makefiles. use libsa/libkern/libz as appropriate,
and don't build the various sources into objs locally by specifying
them directly in the Makefiles.
- move the secondary bootstrap down (to 0x20004000), and add a HEAP_LIMIT
so that we don't exhaust our bootstrap address space (when using the new,
not yet checked in, memory allocator).
- clean up installboot, deal with secondary bootstrap programs not on
'a' partition or 'a' partition not starting at 0.
- add padding to structures in bbinfo.h for future expansion. Add a
netbbinfo structure to allow netboot information to be hard-coded in
network boot blocks, so that they can be made to work even on systems
with firmware which doesn't support the new "ethernet address in boot
device" convention.
- move the sync() calls to the correct place in installboot.c
- remove a kludge in disk.c which was breaking multiple opens/closes
of the disk
- allow netboot ethernet address to be hard-coded into binary so that
machines with old firmware which doesn't pass it in the boot device
can work. Assume that if the ethernet address isn't passed in, it
uses the old (dain-bramaged) 'read' interface works on my 3000/300LX.
- Add setnetbootinfo, a program to hard code an ethernet address into
a network boot.
- move an 'int debug;' into boot.c (it was the only diff between boot.c and
netboot.c), and nuke netboot.c (using boot.c instead for net boot blocks).
- add cd9660 file system ops to the file system ops table in filesystem.c
- if a file name is given (i.e. BOOTED_FILE) is set, boot only that file,
but if not then try to boot "bsd", "bsd.bak", "bsd.old", and
"obsd" (in that order) until one is found or until the list of names
is exhausted.
- add support for reading gzipped kernels.
- use strerror() to print errors, rather than just printing error numbers
- if no disk label exists, fake one up
- slightly relax the block size checks in bootxx.c; they were a bit to paranoid

And local changes:
- don't build a copy of libsa/libkern/libz for each boot prog,
just build a single copy and use it for everything.

Make compilable in our env. Involved disabling ELF support,
reordering includes (due to -Wstrict-prototypes) and other -Wall goo

Merge to NetBSD 961020. Retained our kernel APIs where NetBSD has changed.
-Wall -Wstrict-prototypes -Wmissing-prototypes too.