ddb/db_active.h: New home for extern db_active.

This can be included unconditionally, and db_active can then be
queried unconditionally; if DDB is not in the kernel, then db_active
is a constant zero. Reduces need for #include opt_ddb.h, #ifdef DDB.

introduce DDB_END_CMD and replace more than 20 copies of the same
list of NULLs and 0. idea from rillig@.

all touched ports built, several booted.

branches: 1.65.2;
SH3_P2SEG_FUNC(f) - get a P2 address of a function.


Just SH3_P1SEG_TO_P2SEG + cast, but we now define SH3_P1SEG_TO_P2SEG
and SH3_P2SEG_TO_P1SEG to use arithmetic, not bitwise ops. That gives
the same result for the correct P1/P2 input values, but addition can
be done at link time with addends. Thus SH3_P2SEG_FUNC compiles to a
constant P2 address.

db_tlbdump_cmd - get rid of RUN_P2/P1, use functions instead.

Move the bits of code to read TLB entries into separate functions and
call them via P2 pointers instead of doing P2/P1 switch inline.

Declare __db_cachedump_sh3/sh4 __noinline too.

Previous commit only did that for the definitions. Forgot to apply
the attribute to the declarations too.

db_cachedump_cmd - call the dump function via P2 directly.

Instead of calling RUN_P2 at the very beginning of both cpu-specific
cache dump functions just call them via P2 pointers. This effectively
just makes db_cachedump_cmd into a shared RUN_P2 code for both of
them. Drop RUN_P1 at the end as we are getting back to P1 anyway
because that's where the return address is. Any necessary padding is
provided by the sh_icache_sync_all() call.

Implement PT_STEP.

BRCR and BAMRA have different format in sh3 and sh4, so there's no
point in using SH_() wrapper to refer to them in code that is already
model-specific.

branches: 1.59.4; 1.59.6; 1.59.8;
Use lwp_getpcb() on mips, powerpc and sh3, clean from struct user usage.

Ansify function definitions w/o arguments. Generated with sed.

branches: 1.57.6; 1.57.12;
Match PC address for single-stepping, program BARA = tf->tf_spc before
returning from exception.

Makes single-stepping work on sh3 - without matching the address sh3
gets stuck at the instruction we are at, getting EXPEVT_BREAK before
executing the instruction.

Cosmetics: add blank lines to make db_trap() call more prominent.
s/kdb/ddb/ is a comment (does anyone really still calls it kdb?)

We never pass type = -1 to kdb_trap() so g/c that cargo-culted case.

Use _SH_TRA_BREAK instead of hardcoded constant under ifdef KGDB too.

Support single-stepping in DDB for sh3 too, it has different format of
UBC registers.

Add GBR register to switchframe.

While here, reorder switchframe so that we can shave off a few
instructions when restoring it (group stack registers at the end, make
sure that last and first registers are safe to load/store in delay
slots).

Add GBR register to trapframe.

branches: 1.50.2; 1.50.4; 1.50.6;
Revert interrupt stack changes. There are some issues and, at least
in the current form, they complicate implementation of fast software
interrupts. I'd rather get fast softints in now and may be revisit
interrupt stack later.

Oops, delete commented out line forgotten in previous.

On SH4 we must be running on P2 to access memory-mapped TLB arrays.
Print DA2 again - that messenger was shot in 1.34

Add "mach reset" command that resets machine by taking a trap with
exceptions disabled. Fix db_stackcheck_cmd (have_addr is now bool).
Supply help strings for mach commands.

branches: 1.46.2; 1.46.6;
Use dedicated interrupt stack. Tested on hpcsh and landisk. This has
been sitting in my tree for quite a while. While here s/top/bottom/
in a few places related to stacks to confirm to conventional
terminology.

branches: 1.45.4; 1.45.8;
Merge the ppcoea-renovation branch to HEAD.

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

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

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

Add a new option DDB_VERBOSE_HELP that adds online help to ddb.
From Adam Hamsik.
Minor modifications by me, all bugs are probably mine.

branches: 1.43.8; 1.43.10;
Print pc/pr when faulting in ddb, it comes in handy.

branches: 1.42.4; 1.42.6; 1.42.12;
Sprinkle static so that gcc can eliminate simple helper functions that
get inlined.

Catch up with bool rototill.

TRUE -> true, FALSE -> false

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.38.2;
Remove duplicate #includes, patch contributed in private mail
by Slava Semushin <slava.semushin@gmail.com>.

To verify that no nasty side effects of duplicate includes (or their
removal) have an effect here, I've compiled an i386/ALL kernel with
and without the patch, and the only difference in the resulting .o
files was in shifted line numbers in some assert() calls.
The comparison of the .o files was based on the output of "objdump -D".

Thanks to martin@ for the input on testing.

branches: 1.37.2;
__db_print_symbol: don't try to be too smart with db_print_loc_and_inst.

We can have in a register an address that points to/into some variable
in the data segment, but db_print_loc_and_inst only looks for
functions, so it will misprint it as something unrelated from libkern
+ huge offset. E.g. instead of netbsd:cpu_info_store it would print
netbsd:prop_string_create_cstring+0xdeadbeef

Worse, if the address happens to be odd (char field in a struct, an
element of char array), attempt at printing the "instruction" at that
address will cause a fault and will abort "mach frame".

Disassemly is not really that useful in "mach frame" listing anyway
and more often just clutters things by overflowing 80 columns.

Cosmetics - space auxiliary defines so that they are not visually
hidden in surrounding code.

s/0/NULL/ for pointers.

Make columns align with the header when dumping sh4 tlb.
Ifdef out dump of data array 2, it causes weird effects on landisk.
Minor whitespace cosmetic.

exp_type[] was constified in exception.c but an extern declaration in
db_interface.c was not updated. Declare it publicly. While here,
constify the array itself too.

branches: 1.32.4; 1.32.8;
__asm__ -> __asm
__const__ -> const
__inline__ -> inline
__volatile__ -> volatile

Declare ddb_regs extern in db_machdep.h and provide the defintion in
db_interface.c.

merge ktrace-lwp.

branches: 1.29.2;
Catch up with constification.

Adapt to ddb constification.

adapt to const changes.

db_stackcheck_cmd: catch up with proc => lwp changes.
This is just a quick fix to make KSTACK_DEBUG kernels compilable.
Formatting needs improvment (group lwps from one process together, etc).
PR port-sh3/29621

KNF: use "continue" as explicit empty loop body.

Use uint<N>_t instead of u_int<N>_t.

branches: 1.23.8; 1.23.10;
__KERNEL_RCSID()

branches: 1.22.2;
Use PAGE_SIZE rather than NBPG.

Add _reg_bset_[124]() and _reg_bclr_[124]() macro for
SH embeded device registers.

Fix asm usage. "tf" is an output operand, not an input.

branches: 1.19.6;
Nuke local extern label_t *db_recover; it's now in <ddb/db_extern.h>

cosmetic changes.

adapted to pmap and exception code changes.

Use "#define<tab>", white space nits.

make these compile with gcc -O0

Rework interrupt code.
+ Fully utilize SH SR.I[0:3] interrupt level.
+ software interrupt is emulated by TMU1, 2 one shot interrupt.
+ implement generic soft interrupts.
+ implement clockframe correctly.

add kernel stack and trapframe debug command.

SR related parts moved to psl.h. cpufunc.h segments.h are removed.
kernel mode checking is only SR.MD. no check stack pointer.

Initial support for KGDB on the sh3.

Cleanup register definition

Clean up function declaration.

add cache dump

ANSIfy KNF.

machine tlb command print PFN with AREA number.

TLB dump command added to DDB

s/BYTE_ORDER/_BYTE_ORDER/g

branches: 1.3.4;
DDB single step and stack trace support.

remove include of <vm/vm.h>. <vm/vm.h> -> <uvm/uvm_extern.h>

branches: 1.1.2;
Merge in NetBSD/sh3 from cvs.kame.net repository.

Tree structure:
- sys/arch/sh3: sh3 generic code
As commented, in-chip device drivers are put into sys/arch/sh3/dev.
- sys/arch/evbsh3: sh3 evaluation boards (pure sh3 CPU, no fancy external HW)
- sys/arch/mmeye: Brains mmEye, www.brains.co.jp
MI source code includes couple of #ifdef for sh3-coff support.
(sh3 uses coff or elf)

Needs some more improvements, especialy in sys/arch/sh3/conf/files.sh3,
to compile the tree (due to last minute tree structure change).

introduce DDB_END_CMD and replace more than 20 copies of the same
list of NULLs and 0. idea from rillig@.

all touched ports built, several booted.

SH3_P2SEG_FUNC(f) - get a P2 address of a function.


Just SH3_P1SEG_TO_P2SEG + cast, but we now define SH3_P1SEG_TO_P2SEG
and SH3_P2SEG_TO_P1SEG to use arithmetic, not bitwise ops. That gives
the same result for the correct P1/P2 input values, but addition can
be done at link time with addends. Thus SH3_P2SEG_FUNC compiles to a
constant P2 address.

db_tlbdump_cmd - get rid of RUN_P2/P1, use functions instead.

Move the bits of code to read TLB entries into separate functions and
call them via P2 pointers instead of doing P2/P1 switch inline.

Declare __db_cachedump_sh3/sh4 __noinline too.

Previous commit only did that for the definitions. Forgot to apply
the attribute to the declarations too.

db_cachedump_cmd - call the dump function via P2 directly.

Instead of calling RUN_P2 at the very beginning of both cpu-specific
cache dump functions just call them via P2 pointers. This effectively
just makes db_cachedump_cmd into a shared RUN_P2 code for both of
them. Drop RUN_P1 at the end as we are getting back to P1 anyway
because that's where the return address is. Any necessary padding is
provided by the sh_icache_sync_all() call.

Implement PT_STEP.

BRCR and BAMRA have different format in sh3 and sh4, so there's no
point in using SH_() wrapper to refer to them in code that is already
model-specific.

branches: 1.59.4; 1.59.6; 1.59.8;
Use lwp_getpcb() on mips, powerpc and sh3, clean from struct user usage.

Ansify function definitions w/o arguments. Generated with sed.

branches: 1.57.6; 1.57.12;
Match PC address for single-stepping, program BARA = tf->tf_spc before
returning from exception.

Makes single-stepping work on sh3 - without matching the address sh3
gets stuck at the instruction we are at, getting EXPEVT_BREAK before
executing the instruction.

Cosmetics: add blank lines to make db_trap() call more prominent.
s/kdb/ddb/ is a comment (does anyone really still calls it kdb?)

We never pass type = -1 to kdb_trap() so g/c that cargo-culted case.

Use _SH_TRA_BREAK instead of hardcoded constant under ifdef KGDB too.

Support single-stepping in DDB for sh3 too, it has different format of
UBC registers.

Add GBR register to switchframe.

While here, reorder switchframe so that we can shave off a few
instructions when restoring it (group stack registers at the end, make
sure that last and first registers are safe to load/store in delay
slots).

Add GBR register to trapframe.

branches: 1.50.2; 1.50.4; 1.50.6;
Revert interrupt stack changes. There are some issues and, at least
in the current form, they complicate implementation of fast software
interrupts. I'd rather get fast softints in now and may be revisit
interrupt stack later.

Oops, delete commented out line forgotten in previous.

On SH4 we must be running on P2 to access memory-mapped TLB arrays.
Print DA2 again - that messenger was shot in 1.34

Add "mach reset" command that resets machine by taking a trap with
exceptions disabled. Fix db_stackcheck_cmd (have_addr is now bool).
Supply help strings for mach commands.

branches: 1.46.2; 1.46.6;
Use dedicated interrupt stack. Tested on hpcsh and landisk. This has
been sitting in my tree for quite a while. While here s/top/bottom/
in a few places related to stacks to confirm to conventional
terminology.

branches: 1.45.4; 1.45.8;
Merge the ppcoea-renovation branch to HEAD.

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

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

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

Add a new option DDB_VERBOSE_HELP that adds online help to ddb.
From Adam Hamsik.
Minor modifications by me, all bugs are probably mine.

branches: 1.43.8; 1.43.10;
Print pc/pr when faulting in ddb, it comes in handy.

branches: 1.42.4; 1.42.6; 1.42.12;
Sprinkle static so that gcc can eliminate simple helper functions that
get inlined.

Catch up with bool rototill.

TRUE -> true, FALSE -> false

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.38.2;
Remove duplicate #includes, patch contributed in private mail
by Slava Semushin <slava.semushin@gmail.com>.

To verify that no nasty side effects of duplicate includes (or their
removal) have an effect here, I've compiled an i386/ALL kernel with
and without the patch, and the only difference in the resulting .o
files was in shifted line numbers in some assert() calls.
The comparison of the .o files was based on the output of "objdump -D".

Thanks to martin@ for the input on testing.

branches: 1.37.2;
__db_print_symbol: don't try to be too smart with db_print_loc_and_inst.

We can have in a register an address that points to/into some variable
in the data segment, but db_print_loc_and_inst only looks for
functions, so it will misprint it as something unrelated from libkern
+ huge offset. E.g. instead of netbsd:cpu_info_store it would print
netbsd:prop_string_create_cstring+0xdeadbeef

Worse, if the address happens to be odd (char field in a struct, an
element of char array), attempt at printing the "instruction" at that
address will cause a fault and will abort "mach frame".

Disassemly is not really that useful in "mach frame" listing anyway
and more often just clutters things by overflowing 80 columns.

Cosmetics - space auxiliary defines so that they are not visually
hidden in surrounding code.

s/0/NULL/ for pointers.

Make columns align with the header when dumping sh4 tlb.
Ifdef out dump of data array 2, it causes weird effects on landisk.
Minor whitespace cosmetic.

exp_type[] was constified in exception.c but an extern declaration in
db_interface.c was not updated. Declare it publicly. While here,
constify the array itself too.

branches: 1.32.4; 1.32.8;
__asm__ -> __asm
__const__ -> const
__inline__ -> inline
__volatile__ -> volatile

Declare ddb_regs extern in db_machdep.h and provide the defintion in
db_interface.c.

merge ktrace-lwp.

branches: 1.29.2;
Catch up with constification.

Adapt to ddb constification.

adapt to const changes.

db_stackcheck_cmd: catch up with proc => lwp changes.
This is just a quick fix to make KSTACK_DEBUG kernels compilable.
Formatting needs improvment (group lwps from one process together, etc).
PR port-sh3/29621

KNF: use "continue" as explicit empty loop body.

Use uint<N>_t instead of u_int<N>_t.

branches: 1.23.8; 1.23.10;
__KERNEL_RCSID()

branches: 1.22.2;
Use PAGE_SIZE rather than NBPG.

Add _reg_bset_[124]() and _reg_bclr_[124]() macro for
SH embeded device registers.

Fix asm usage. "tf" is an output operand, not an input.

branches: 1.19.6;
Nuke local extern label_t *db_recover; it's now in <ddb/db_extern.h>

cosmetic changes.

adapted to pmap and exception code changes.

Use "#define<tab>", white space nits.

make these compile with gcc -O0

Rework interrupt code.
+ Fully utilize SH SR.I[0:3] interrupt level.
+ software interrupt is emulated by TMU1, 2 one shot interrupt.
+ implement generic soft interrupts.
+ implement clockframe correctly.

add kernel stack and trapframe debug command.

SR related parts moved to psl.h. cpufunc.h segments.h are removed.
kernel mode checking is only SR.MD. no check stack pointer.

Initial support for KGDB on the sh3.

Cleanup register definition

Clean up function declaration.

add cache dump

ANSIfy KNF.

machine tlb command print PFN with AREA number.

TLB dump command added to DDB

s/BYTE_ORDER/_BYTE_ORDER/g

branches: 1.3.4;
DDB single step and stack trace support.

remove include of <vm/vm.h>. <vm/vm.h> -> <uvm/uvm_extern.h>

branches: 1.1.2;
Merge in NetBSD/sh3 from cvs.kame.net repository.

Tree structure:
- sys/arch/sh3: sh3 generic code
As commented, in-chip device drivers are put into sys/arch/sh3/dev.
- sys/arch/evbsh3: sh3 evaluation boards (pure sh3 CPU, no fancy external HW)
- sys/arch/mmeye: Brains mmEye, www.brains.co.jp
MI source code includes couple of #ifdef for sh3-coff support.
(sh3 uses coff or elf)

Needs some more improvements, especialy in sys/arch/sh3/conf/files.sh3,
to compile the tree (due to last minute tree structure change).

SH3_P2SEG_FUNC(f) - get a P2 address of a function.


Just SH3_P1SEG_TO_P2SEG + cast, but we now define SH3_P1SEG_TO_P2SEG
and SH3_P2SEG_TO_P1SEG to use arithmetic, not bitwise ops. That gives
the same result for the correct P1/P2 input values, but addition can
be done at link time with addends. Thus SH3_P2SEG_FUNC compiles to a
constant P2 address.

db_tlbdump_cmd - get rid of RUN_P2/P1, use functions instead.

Move the bits of code to read TLB entries into separate functions and
call them via P2 pointers instead of doing P2/P1 switch inline.

Declare __db_cachedump_sh3/sh4 __noinline too.

Previous commit only did that for the definitions. Forgot to apply
the attribute to the declarations too.

db_cachedump_cmd - call the dump function via P2 directly.

Instead of calling RUN_P2 at the very beginning of both cpu-specific
cache dump functions just call them via P2 pointers. This effectively
just makes db_cachedump_cmd into a shared RUN_P2 code for both of
them. Drop RUN_P1 at the end as we are getting back to P1 anyway
because that's where the return address is. Any necessary padding is
provided by the sh_icache_sync_all() call.

Implement PT_STEP.

BRCR and BAMRA have different format in sh3 and sh4, so there's no
point in using SH_() wrapper to refer to them in code that is already
model-specific.

branches: 1.59.4; 1.59.6; 1.59.8;
Use lwp_getpcb() on mips, powerpc and sh3, clean from struct user usage.

Ansify function definitions w/o arguments. Generated with sed.

branches: 1.57.6; 1.57.12;
Match PC address for single-stepping, program BARA = tf->tf_spc before
returning from exception.

Makes single-stepping work on sh3 - without matching the address sh3
gets stuck at the instruction we are at, getting EXPEVT_BREAK before
executing the instruction.

Cosmetics: add blank lines to make db_trap() call more prominent.
s/kdb/ddb/ is a comment (does anyone really still calls it kdb?)

We never pass type = -1 to kdb_trap() so g/c that cargo-culted case.

Use _SH_TRA_BREAK instead of hardcoded constant under ifdef KGDB too.

Support single-stepping in DDB for sh3 too, it has different format of
UBC registers.

Add GBR register to switchframe.

While here, reorder switchframe so that we can shave off a few
instructions when restoring it (group stack registers at the end, make
sure that last and first registers are safe to load/store in delay
slots).

Add GBR register to trapframe.

branches: 1.50.2; 1.50.4; 1.50.6;
Revert interrupt stack changes. There are some issues and, at least
in the current form, they complicate implementation of fast software
interrupts. I'd rather get fast softints in now and may be revisit
interrupt stack later.

Oops, delete commented out line forgotten in previous.

On SH4 we must be running on P2 to access memory-mapped TLB arrays.
Print DA2 again - that messenger was shot in 1.34

Add "mach reset" command that resets machine by taking a trap with
exceptions disabled. Fix db_stackcheck_cmd (have_addr is now bool).
Supply help strings for mach commands.

branches: 1.46.2; 1.46.6;
Use dedicated interrupt stack. Tested on hpcsh and landisk. This has
been sitting in my tree for quite a while. While here s/top/bottom/
in a few places related to stacks to confirm to conventional
terminology.

branches: 1.45.4; 1.45.8;
Merge the ppcoea-renovation branch to HEAD.

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

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

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

Add a new option DDB_VERBOSE_HELP that adds online help to ddb.
From Adam Hamsik.
Minor modifications by me, all bugs are probably mine.

branches: 1.43.8; 1.43.10;
Print pc/pr when faulting in ddb, it comes in handy.

branches: 1.42.4; 1.42.6; 1.42.12;
Sprinkle static so that gcc can eliminate simple helper functions that
get inlined.

Catch up with bool rototill.

TRUE -> true, FALSE -> false

Replace the Mach-derived boolean_t type with the C99 bool type. A
future commit will replace use of TRUE and FALSE with true and false.

branches: 1.38.2;
Remove duplicate #includes, patch contributed in private mail
by Slava Semushin <slava.semushin@gmail.com>.

To verify that no nasty side effects of duplicate includes (or their
removal) have an effect here, I've compiled an i386/ALL kernel with
and without the patch, and the only difference in the resulting .o
files was in shifted line numbers in some assert() calls.
The comparison of the .o files was based on the output of "objdump -D".

Thanks to martin@ for the input on testing.

branches: 1.37.2;
__db_print_symbol: don't try to be too smart with db_print_loc_and_inst.

We can have in a register an address that points to/into some variable
in the data segment, but db_print_loc_and_inst only looks for
functions, so it will misprint it as something unrelated from libkern
+ huge offset. E.g. instead of netbsd:cpu_info_store it would print
netbsd:prop_string_create_cstring+0xdeadbeef

Worse, if the address happens to be odd (char field in a struct, an
element of char array), attempt at printing the "instruction" at that
address will cause a fault and will abort "mach frame".

Disassemly is not really that useful in "mach frame" listing anyway
and more often just clutters things by overflowing 80 columns.

Cosmetics - space auxiliary defines so that they are not visually
hidden in surrounding code.

s/0/NULL/ for pointers.

Make columns align with the header when dumping sh4 tlb.
Ifdef out dump of data array 2, it causes weird effects on landisk.
Minor whitespace cosmetic.

exp_type[] was constified in exception.c but an extern declaration in
db_interface.c was not updated. Declare it publicly. While here,
constify the array itself too.

branches: 1.32.4; 1.32.8;
__asm__ -> __asm
__const__ -> const
__inline__ -> inline
__volatile__ -> volatile

Declare ddb_regs extern in db_machdep.h and provide the defintion in
db_interface.c.

merge ktrace-lwp.

branches: 1.29.2;
Catch up with constification.

Adapt to ddb constification.

adapt to const changes.

db_stackcheck_cmd: catch up with proc => lwp changes.
This is just a quick fix to make KSTACK_DEBUG kernels compilable.
Formatting needs improvment (group lwps from one process together, etc).
PR port-sh3/29621

KNF: use "continue" as explicit empty loop body.

Use uint<N>_t instead of u_int<N>_t.

branches: 1.23.8; 1.23.10;
__KERNEL_RCSID()

branches: 1.22.2;
Use PAGE_SIZE rather than NBPG.

Add _reg_bset_[124]() and _reg_bclr_[124]() macro for
SH embeded device registers.

Fix asm usage. "tf" is an output operand, not an input.

branches: 1.19.6;
Nuke local extern label_t *db_recover; it's now in <ddb/db_extern.h>

cosmetic changes.

adapted to pmap and exception code changes.

Use "#define<tab>", white space nits.

make these compile with gcc -O0

Rework interrupt code.
+ Fully utilize SH SR.I[0:3] interrupt level.
+ software interrupt is emulated by TMU1, 2 one shot interrupt.
+ implement generic soft interrupts.
+ implement clockframe correctly.

add kernel stack and trapframe debug command.

SR related parts moved to psl.h. cpufunc.h segments.h are removed.
kernel mode checking is only SR.MD. no check stack pointer.

Initial support for KGDB on the sh3.

Cleanup register definition

Clean up function declaration.

add cache dump

ANSIfy KNF.

machine tlb command print PFN with AREA number.

TLB dump command added to DDB

s/BYTE_ORDER/_BYTE_ORDER/g

branches: 1.3.4;
DDB single step and stack trace support.

remove include of <vm/vm.h>. <vm/vm.h> -> <uvm/uvm_extern.h>

branches: 1.1.2;
Merge in NetBSD/sh3 from cvs.kame.net repository.

Tree structure:
- sys/arch/sh3: sh3 generic code
As commented, in-chip device drivers are put into sys/arch/sh3/dev.
- sys/arch/evbsh3: sh3 evaluation boards (pure sh3 CPU, no fancy external HW)
- sys/arch/mmeye: Brains mmEye, www.brains.co.jp
MI source code includes couple of #ifdef for sh3-coff support.
(sh3 uses coff or elf)

Needs some more improvements, especialy in sys/arch/sh3/conf/files.sh3,
to compile the tree (due to last minute tree structure change).