This change arm, armeb, earm, earmeb, earmhf, earmhfeb so all builds that
share a MACHINE_ARCH for userland so that except for etc, will produce
identical sets.

usr/include/machine now points to usr/include/arm

There are other little endian arm machine_arches other than "arm"
Use MACHINE_CPU and TARGET_ENDIANNESS instead of MACHINE_ARCH

branches: 1.10.2;
Whitespace.

branches: 1.9.10; 1.9.26;
fix, use += when adint to SUBDIR list

add gzboot support for gemini

GEMINI_dram_0x01600000 defines inflation from concatenated
gzboot & netbsd.bin.gz at sdram 0x01600000 to 0x00200000

note that ns16550.c INB() and OUTB() macros now have
"mult 4" type 4 byte regsiter offset stride
if NS16550_AX4 is defined. this is needed for gemini,
and was OKed by gimpy.

branches: 1.7.74; 1.7.78; 1.7.84; 1.7.86;
merge ktrace-lwp.

descend into TS7200_flash_0x60660000

Whoops, reenable some stuff.

Add gzboot support for SMDK2410.

gzboot for SMDK2800.

branches: 1.2.2;
Don't build a big-endian image on platforms which don't run big-endian.

Descend into the gzboot directories.

Allow a configuration to specify an arbitrary address where the
compressed kernel image is located, in the event it can't be appended
to or embedded within the gzboot binary itself.

branches: 1.37.2;
1. All Makefiles that use ARM_APCS_FLAGS add -marm unconditionally and
bsd.own.mk also adds it for gcc only (so we end up with 2 -marm for gcc).
Just add -marm unconditionally in bsd.own.mk in ARM_APCS_FLAGS and remove
it from everywhere else.
2. gcc-8 now warns about thumb internetwork in APCS mode. This does not make
sense because the documentation says that this flag does not make a
difference in apcs mode, but might produce larger code in non-APCS mode.
So we just add it to silence the warnings.

branches: 1.36.14;
centralize vers.c building for standalone programs.

GC more logic that was intended for GCC <4.8 (or !LLVM)

Same as before (-marm implies -mno-thumb), but now the condition was flipped

branches: 1.34.2; 1.34.4;
Don't reuse LDFLAGS for arguments passed directly to ${LD}, share/mk
expects it to be passed to ${CC} or ${CXX}, so they will generally
prefix flags with -Wl.

remove the MKPIE and MKCTF no hack, and move it to the .mk files

Disable PIE and CTF for standalone code
XXX: Must be done centrally, and there must be a better way.

Clear STRIPFLAGS (stripping in general is dangerous with the -BE8 object
format, and should never be done before the final linking has happened).
From Rin Okuyama.

branches: 1.30.6;
Add empty LIBCRTI= as LIBCRT0 to build sa programs without installed DESTDIR.

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

Move flags for ARM's APCS ABI into a common variable. Add some forces to
deal with limitations in the Clang <-> LLVM backend interaction to pick
the correct ABI and target assembler.

Use -marm for clang builds, too.

explicitly add -marm

-mno-thumb-interwork is the default, and there is no -mthumb in gcc 4.8

Use <bsd.klinks.mk>

Add CPPFLAGS+= -mno-thumb -mno-thumb-interwork

branches: 1.23.2;
Add -fno-unwind-tables

gcc will supply -Wa,-eabi=gnu so don't include. Use -mabi=apcs-base, not atpcs

Make sure these are always compiled with -mabi=atpcs -mfloat-abi=soft
-Wa,-meabi=gnu so they use the old compilation methods.

branches: 1.20.2;
Fixed build with USE_SSP=yes

ok christos@

Fix to allow EABI compilation.
Make sure to DISCARD new .ARM.attributes and .ARM.ebxidx sections.

Add missing ${_MKTARGET_LINK}

branches: 1.17.4; 1.17.10;
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: 1.16.2; 1.16.4;
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: 1.15.4; 1.15.6;
- WARNSfy
- add src/common/dist/zlib to incpath for build without DESTDIR

branches: 1.14.72; 1.14.76; 1.14.84;
The new zlib uses HEAP_SIZE, so instead use BOARD_HEAP_SIZE as the
symbol to use here. Also adapt to the new zlib with the changed name
for the header file.

branches: 1.13.2;
merge ktrace-lwp.

branches: 1.12.2;
dependencies for .BEGIN is not accepted by the latest nbmake.

include <bsd.obj.mk> to teach make the rules for OBJDIR names.

This is to fix a problem on compiling gzboot with BUILDID defined
(by build.sh -B BUILDID).

branches: 1.10.4; 1.10.6; 1.10.10;
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: 1.9.2;
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.

branches: 1.8.2;
* Add ${PROGSYM} to CLEANFILES.
* Instsall either the .bin or the .sym, based on whether or not
the resulting gzboot needs to be mdsetimage'd.

Also build a flat binary version of the gzboot program.

Garbage-collect the DYNAMIC_CRC_TABLE stuff. The table it computed
was incompatible with the new CRC code, and it is largely unnecessary
now, since the static table is so much smaller.

Fixes PR kern/20935.

branches: 1.5.6;
Give gzboot a 1M heap, so that loading gzip -9'd kernels works
reliably.

branches: 1.4.2;
Add support for relocating gzboot's .text out of flash and into
RAM (while still decompressing the image directly from flash).

This makes gzboot run a LOT faster.

Fix the -DRELOC=... CPP flag. Add -DDYNAMIC_CRC_TABLE to shrink
the text (and, thus, the space taken up in flash) a bit.

Add support for running directly out of flash:
* Require that the builder Makefile provide a linker script.
* After making sure the MMU is disabled, check to see if
_etext == __data_start. If not, then copy the .data contents
into RAM.
* Put the stack in .bss.

gzboot -- boot a gzip'd kernel image. This is useful for booting
compressed kernel images from flash.

This needs more work, but is a good first-pass.

Descend into the gzboot directories.

Allow a configuration to specify an arbitrary address where the
compressed kernel image is located, in the event it can't be appended
to or embedded within the gzboot binary itself.

branches: 1.15.56; 1.15.62;
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: 1.14.4; 1.14.6; 1.14.8;
kill extra whitespaces
reviewed by tsutsui@

bcopy -> memcpy

bzero -> memset

branches: 1.11.76; 1.11.84; 1.11.90;
free -> dealloc
unsigned -> size_t for alloc/dealloc

The new zlib uses HEAP_SIZE, so instead use BOARD_HEAP_SIZE as the
symbol to use here. Also adapt to the new zlib with the changed name
for the header file.

branches: 1.9.2;
merge ktrace-lwp.

branches: 1.8.16;
remove license clauses 3 and 4 from my cpoyright notices

branches: 1.7.2;
Allow MAXIMAGESIZE to be set to 0. For gzboot's that are linked
for flash, this allows us to basically do:

cat gzboot.bin netbsd.bin.gz > gzimg

to produce the bootable image.

branches: 1.6.6;
* Call board_init() rather than mem_init() (board_init() will take
care of that for us).
* Call board_fini() right before jumping to the kernel entry point.

branches: 1.5.2;
Improve zlib error reporting.

Use bcopy() and bzero(), since the rest of libsa does. This means
we don't have to pull in memset() and memcpy() from libsa.

Don't need to provide zcalloc() and zcfree(); there are such routines
in libz.

Fix a silly bug in the propellor spinning code.

gzboot -- boot a gzip'd kernel image. This is useful for booting
compressed kernel images from flash.

This needs more work, but is a good first-pass.

branches: 1.1.2; 1.1.14;
gzboot -- boot a gzip'd kernel image. This is useful for booting
compressed kernel images from flash.

This needs more work, but is a good first-pass.

merge ktrace-lwp.

align md_root_loadaddr to word boundary.

branches: 1.2.2; 1.2.10; 1.2.12;
Allow MAXIMAGESIZE to be set to 0. For gzboot's that are linked
for flash, this allows us to basically do:

cat gzboot.bin netbsd.bin.gz > gzimg

to produce the bootable image.

branches: 1.1.2; 1.1.14;
Add support for relocating gzboot's .text out of flash and into
RAM (while still decompressing the image directly from flash).

This makes gzboot run a LOT faster.

This file is obsolete.

gzboot -- boot a gzip'd kernel image. This is useful for booting
compressed kernel images from flash.

This needs more work, but is a good first-pass.

merge ktrace-lwp.

Deal with case where __data_store == _edata.

branches: 1.5.6;
Continue the " - . - 8" purge. Specifically:

add rd, pc, #foo - . - 8 -> adr rd, foo
ldr rd, [pc, #foo - . - 8] -> ldr rd, foo

Also, when saving the return address for a function pointer call, use
"mov lr, pc" just before the call unless the return address is somewhere
other than just after the call site.

Finally, a few obvious little micro-optimisations like using LDR directly
rather than ADR followed by LDR, and loading directly into PC rather than
bouncing via R0.

branches: 1.4.2; 1.4.14;
Fix a couple of comments.

Add support for relocating gzboot's .text out of flash and into
RAM (while still decompressing the image directly from flash).

This makes gzboot run a LOT faster.

Add support for running directly out of flash:
* Require that the builder Makefile provide a linker script.
* After making sure the MMU is disabled, check to see if
_etext == __data_start. If not, then copy the .data contents
into RAM.
* Put the stack in .bss.

gzboot -- boot a gzip'd kernel image. This is useful for booting
compressed kernel images from flash.

This needs more work, but is a good first-pass.

Bump version to 1.2 for the concatenation feature.

branches: 1.2.2; 1.2.14;
Add support for relocating gzboot's .text out of flash and into
RAM (while still decompressing the image directly from flash).

This makes gzboot run a LOT faster.

gzboot -- boot a gzip'd kernel image. This is useful for booting
compressed kernel images from flash.

This needs more work, but is a good first-pass.

Fix CPUFLAGS to the correct value so a passed down CPUFLAGS doesn't result
in broken images.

branches: 1.2.142; 1.2.150; 1.2.152;
Make these images concatenating images.

Gzboot support for the ADI BRH eval board. This flash address allows
the kernel to be decompressed and started at board reset if the BRH's
rotary switch is set to position 5.

Fix to allow EABI compilation.
Make sure to DISCARD new .ARM.attributes and .ARM.ebxidx sections.

branches: 1.5.12;
Deal with the following change to ld:

* The default output section LMA has changed for allocatable sections from
being equal to VMA, to keeping the difference between LMA and VMA the same as
the previous output section in the same region. This is a more useful
default when using overlays and other cases where you specify an LMA
differing from the VMA for some sections.

branches: 1.4.78;
merge ktrace-lwp.

Pad .sdata section to a 4-byte boundary so that the .image section
is always aligned. This makes md_root_loadaddr always 4 byte aligned.
Without this, may get an un-aligned access trap before we even print
anything on the console which was a pain in the neck to debug since so
early in the bootstrap the CPU usually just halts on exceptions.

branches: 1.2.2;
Don't set the output format explicitly; let the linker use its defaults.

Gzboot support for the ADI BRH eval board. This flash address allows
the kernel to be decompressed and started at board reset if the BRH's
rotary switch is set to position 5.

Fix CPUFLAGS to the correct value so a passed down CPUFLAGS doesn't result
in broken images.

branches: 1.3.2;
Fix to allow EABI compilation.
Make sure to DISCARD new .ARM.attributes and .ARM.ebxidx sections.

branches: 1.2.4; 1.2.6; 1.2.10; 1.2.22; 1.2.30;
- nullcom.c provides nulldev-like console device for silent gzboot
- GEMINI_dram_0x01600000/Makefile uses nullcom, shhh

add gzboot support for gemini

GEMINI_dram_0x01600000 defines inflation from concatenated
gzboot & netbsd.bin.gz at sdram 0x01600000 to 0x00200000

note that ns16550.c INB() and OUTB() macros now have
"mult 4" type 4 byte regsiter offset stride
if NS16550_AX4 is defined. this is needed for gemini,
and was OKed by gimpy.

Fix to allow EABI compilation.
Make sure to DISCARD new .ARM.attributes and .ARM.ebxidx sections.

branches: 1.2.12;
Deal with the following change to ld:

* The default output section LMA has changed for allocatable sections from
being equal to VMA, to keeping the difference between LMA and VMA the same as
the previous output section in the same region. This is a more useful
default when using overlays and other cases where you specify an LMA
differing from the VMA for some sections.

branches: 1.1.6; 1.1.8; 1.1.14;
add gzboot support for gemini

GEMINI_dram_0x01600000 defines inflation from concatenated
gzboot & netbsd.bin.gz at sdram 0x01600000 to 0x00200000

note that ns16550.c INB() and OUTB() macros now have
"mult 4" type 4 byte regsiter offset stride
if NS16550_AX4 is defined. this is needed for gemini,
and was OKed by gimpy.

Fix CPUFLAGS to the correct value so a passed down CPUFLAGS doesn't result
in broken images.

branches: 1.3.142; 1.3.150; 1.3.152;
Fix the name of the image.

Make these images concatenating images.

Rename the IQ80310 directory to have all nibbles of the address in the name.

Fix to allow EABI compilation.
Make sure to DISCARD new .ARM.attributes and .ARM.ebxidx sections.

branches: 1.4.12;
Deal with the following change to ld:

* The default output section LMA has changed for allocatable sections from
being equal to VMA, to keeping the difference between LMA and VMA the same as
the previous output section in the same region. This is a more useful
default when using overlays and other cases where you specify an LMA
differing from the VMA for some sections.

branches: 1.3.78;
merge ktrace-lwp.

Pad .sdata section to a 4-byte boundary so that the .image section
is always aligned. This makes md_root_loadaddr always 4 byte aligned.
Without this, may get an un-aligned access trap before we even print
anything on the console which was a pain in the neck to debug since so
early in the bootstrap the CPU usually just halts on exceptions.

branches: 1.1.2;
Rename the IQ80310 directory to have all nibbles of the address in the name.

Fix CPUFLAGS to the correct value so a passed down CPUFLAGS doesn't result
in broken images.

branches: 1.3.142; 1.3.150; 1.3.152;
Make these images concatenating images.

branches: 1.2.6; 1.2.8;
Add iq80321.c

Example gzboot config for the IQ80321.

Fix to allow EABI compilation.
Make sure to DISCARD new .ARM.attributes and .ARM.ebxidx sections.

branches: 1.7.12;
Deal with the following change to ld:

* The default output section LMA has changed for allocatable sections from
being equal to VMA, to keeping the difference between LMA and VMA the same as
the previous output section in the same region. This is a more useful
default when using overlays and other cases where you specify an LMA
differing from the VMA for some sections.

branches: 1.6.78;
merge ktrace-lwp.

Pad .sdata section to a 4-byte boundary so that the .image section
is always aligned. This makes md_root_loadaddr always 4 byte aligned.
Without this, may get an un-aligned access trap before we even print
anything on the console which was a pain in the neck to debug since so
early in the bootstrap the CPU usually just halts on exceptions.

branches: 1.4.6; 1.4.8; 1.4.14;
Use LOADADDR() where appropriate.

Lose -nbsd

Add -nbsd to OUTPUT_FORMAT BFD target names.

Example gzboot config for the IQ80321.

branches: 1.1.6;
Add gzboot support for NSLU2. This is a work-in-progress. This configuration
follows the existing NSLU2 RedBoot default bootstrap procedure. (Other
configurations are possible, but not yet implemented.)

branches: 1.2.6;
Because everything goes into the same memory region at run-time in this
configuration *and* image data concatenated onto the end, the BSS segment
needs to be allocated in the file.

Add gzboot support for NSLU2. This is a work-in-progress. This configuration
follows the existing NSLU2 RedBoot default bootstrap procedure. (Other
configurations are possible, but not yet implemented.)

merge ktrace-lwp.

branches: 1.3.2;
Assign to CPUFLAGS to add -march=armv4 instead of adding to COPTS.

add -DCPU_S3C{2800,2410} for board/sscom.c

Add gzboot support for SMDK2410.

Fix to allow EABI compilation.
Make sure to DISCARD new .ARM.attributes and .ARM.ebxidx sections.

branches: 1.4.12;
Deal with the following change to ld:

* The default output section LMA has changed for allocatable sections from
being equal to VMA, to keeping the difference between LMA and VMA the same as
the previous output section in the same region. This is a more useful
default when using overlays and other cases where you specify an LMA
differing from the VMA for some sections.

branches: 1.3.78;
merge ktrace-lwp.

Pad .sdata section to a 4-byte boundary so that the .image section
is always aligned. This makes md_root_loadaddr always 4 byte aligned.
Without this, may get an un-aligned access trap before we even print
anything on the console which was a pain in the neck to debug since so
early in the bootstrap the CPU usually just halts on exceptions.

branches: 1.1.4;
Add gzboot support for SMDK2410.

merge ktrace-lwp.

branches: 1.3.2;
Assign to CPUFLAGS to add -march=armv4 instead of adding to COPTS.

add -DCPU_S3C{2800,2410} for board/sscom.c

gzboot for SMDK2800.

Fix to allow EABI compilation.
Make sure to DISCARD new .ARM.attributes and .ARM.ebxidx sections.

branches: 1.4.12;
Deal with the following change to ld:

* The default output section LMA has changed for allocatable sections from
being equal to VMA, to keeping the difference between LMA and VMA the same as
the previous output section in the same region. This is a more useful
default when using overlays and other cases where you specify an LMA
differing from the VMA for some sections.

branches: 1.3.78;
merge ktrace-lwp.

Pad .sdata section to a 4-byte boundary so that the .image section
is always aligned. This makes md_root_loadaddr always 4 byte aligned.
Without this, may get an un-aligned access trap before we even print
anything on the console which was a pain in the neck to debug since so
early in the bootstrap the CPU usually just halts on exceptions.

branches: 1.1.4;
gzboot for SMDK2800.

merge ktrace-lwp.

branches: 1.3.2;
Assign to CPUFLAGS to add -march=armv4 instead of adding to COPTS.

add -DCPU_S3C{2800,2410} for board/sscom.c

gzboot for SMDK2800.

Fix to allow EABI compilation.
Make sure to DISCARD new .ARM.attributes and .ARM.ebxidx sections.

branches: 1.4.12;
Deal with the following change to ld:

* The default output section LMA has changed for allocatable sections from
being equal to VMA, to keeping the difference between LMA and VMA the same as
the previous output section in the same region. This is a more useful
default when using overlays and other cases where you specify an LMA
differing from the VMA for some sections.

branches: 1.3.78;
merge ktrace-lwp.

Pad .sdata section to a 4-byte boundary so that the .image section
is always aligned. This makes md_root_loadaddr always 4 byte aligned.
Without this, may get an un-aligned access trap before we even print
anything on the console which was a pain in the neck to debug since so
early in the bootstrap the CPU usually just halts on exceptions.

branches: 1.1.4;
gzboot for SMDK2800.

Fix CPUFLAGS to the correct value so a passed down CPUFLAGS doesn't result
in broken images.

branches: 1.2.112; 1.2.120; 1.2.122;
merge ktrace-lwp.

branches: 1.1.4;
support for gzimg's in TS-7200 Redboot on-board flash

Fix to allow EABI compilation.
Make sure to DISCARD new .ARM.attributes and .ARM.ebxidx sections.

branches: 1.3.12;
Deal with the following change to ld:

* The default output section LMA has changed for allocatable sections from
being equal to VMA, to keeping the difference between LMA and VMA the same as
the previous output section in the same region. This is a more useful
default when using overlays and other cases where you specify an LMA
differing from the VMA for some sections.

branches: 1.2.78;
merge ktrace-lwp.

branches: 1.1.4;
support for gzimg's in TS-7200 Redboot on-board flash