Include compiler-rt in libc, libm and libkern.

Some assembly files need to compiled -marm since they are Thumb compatible.

Add EHABI unwind stubs to libkern so prevent errors in linking if unwind
tables are present.

branches: 1.18.2;
Add back {,u}modsi3 for libkern since if we are building standalone tools
we might be using the old ABI and might need them.

modsi3/umodsi3 are only needed !EABI arm

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

Modify earm rule to match any machine_arch starting with earm.

Add strnlen.c to SRCS (which will automatically use the .S version if it
exists).

pull in strnlen.S

add new divsi3 related files.

add split files.

branches: 1.10.2;
If earm*, compile new aeabi runtime routines.

branches: 1.9.12; 1.9.18; 1.9.20;
Move that majority of the 'SRCS+= foo.c' into the main Makefile.libkern.
Any .S files added by the arch/*/Makefile.inc cause the .c file to
be excluded.
Specific exclusions added using NO_SRCS to match previous files.
At least sparc, sparc64, i386, amd64 and vax GENERIC still build.
(There is a fubar with the naming of the byte-swap files ...)

Split the lines that add .c to SRCS from those that add .S
No changes (apart from the order) intended.

Add a _clzsi2 for armv4 platforms.

branches: 1.6.104; 1.6.108; 1.6.114;
Sync arm asm libkern files with libc's asm files.

Add an assembler version of strcmp, based on example code from the ARM
ARM. As an example of the performance difference that this provides
a Dhrystone score on my Shark goes from 213k to 261k.

pull in ffs.S from libc for arm.
The main benefit is that ffs always runs in constant time.

branches: 1.3.8;
Tidy up some memory copying cruft:
bcopy.S is no longer needed
memmove and memcpy were both stacking r0 and unstacking it to keep the return value, so push this down into _memcpy.
rename _memcpy.S to memcpy.S.
memmove.S is now just a placeholder otherwise the make system automagically adds a memmove.c file to libkern.
memmove is just another entry point for memcpy.

branches: 1.2.2;
Remove bcmp, bcopy and bzero. We don't need them as they aren't
called from asm code, and we shouldn't have them if we don't need them.

branches: 1.1.2; 1.1.4;
Create a shared set of ARM sources for libkern and move arm26 over to using
them.

These are identical to the current arm32 sources with the following exceptions:
- References to C labels are wrapped in _C_LABEL().
- Function returns have an alternate version inside #ifdef __APCS_26__.

Tidy up some memory copying cruft:
bcopy.S is no longer needed
memmove and memcpy were both stacking r0 and unstacking it to keep the return value, so push this down into _memcpy.
rename _memcpy.S to memcpy.S.
memmove.S is now just a placeholder otherwise the make system automagically adds a memmove.c file to libkern.
memmove is just another entry point for memcpy.

branches: 1.2.2; 1.2.4;
Sync with libc for licence update.

branches: 1.1.2;
Create a shared set of ARM sources for libkern and move arm26 over to using
them.

These are identical to the current arm32 sources with the following exceptions:
- References to C labels are wrapped in _C_LABEL().
- Function returns have an alternate version inside #ifdef __APCS_26__.

Tidy up some memory copying cruft:
bcopy.S is no longer needed
memmove and memcpy were both stacking r0 and unstacking it to keep the return value, so push this down into _memcpy.
rename _memcpy.S to memcpy.S.
memmove.S is now just a placeholder otherwise the make system automagically adds a memmove.c file to libkern.
memmove is just another entry point for memcpy.

branches: 1.1.2; 1.1.4; 1.1.6;
Create a shared set of ARM sources for libkern and move arm26 over to using
them.

These are identical to the current arm32 sources with the following exceptions:
- References to C labels are wrapped in _C_LABEL().
- Function returns have an alternate version inside #ifdef __APCS_26__.

Use common sources with userland.

merge ktrace-lwp.

branches: 1.4.4;
nuke trailing whitespace

branches: 1.3.2; 1.3.10; 1.3.12;
NetBSD/acorn26 has used APCS-32 for years, so unifdef -U__APCS_26__.

Avoid leaving junk in the top half of R0 on return.
This fixes port-arm/17440.

branches: 1.1.2; 1.1.4; 1.1.6; 1.1.18; 1.1.20;
Create a shared set of ARM sources for libkern and move arm26 over to using
them.

These are identical to the current arm32 sources with the following exceptions:
- References to C labels are wrapped in _C_LABEL().
- Function returns have an alternate version inside #ifdef __APCS_26__.

Use common sources with userland.

branches: 1.2.18;
NetBSD/acorn26 has used APCS-32 for years, so unifdef -U__APCS_26__.

branches: 1.1.2;
Create a shared set of ARM sources for libkern and move arm26 over to using
them.

These are identical to the current arm32 sources with the following exceptions:
- References to C labels are wrapped in _C_LABEL().
- Function returns have an alternate version inside #ifdef __APCS_26__.

bzero is no longer used in the kernel.

branches: 1.1.2; 1.1.26;
Create a shared set of ARM sources for libkern and move arm26 over to using
them.

These are identical to the current arm32 sources with the following exceptions:
- References to C labels are wrapped in _C_LABEL().
- Function returns have an alternate version inside #ifdef __APCS_26__.

Include compiler-rt in libc, libm and libkern.

Thumbify

Support thumb

branches: 1.4.2; 1.4.4;
Add ARMv5 support the clz instruction

branches: 1.3.2; 1.3.10; 1.3.16; 1.3.34;
Use RET/RETc as appropriate

Need to be __clz[sd]i2

Add a _clzsi2 for armv4 platforms.

Use common sources with userland.

merge ktrace-lwp.

branches: 1.5.4;
nuke trailing whitespace

branches: 1.4.2; 1.4.10; 1.4.12;
NetBSD/acorn26 has used APCS-32 for years, so unifdef -U__APCS_26__.

Use .L prefix for all local labels.

branches: 1.2.8;
Sync to libc version of divsi3.S. This is to fix up the stack for 8 byte alignment.

branches: 1.1.2; 1.1.4; 1.1.6;
Create a shared set of ARM sources for libkern and move arm26 over to using
them.

These are identical to the current arm32 sources with the following exceptions:
- References to C labels are wrapped in _C_LABEL().
- Function returns have an alternate version inside #ifdef __APCS_26__.

Use common sources with userland.

merge ktrace-lwp.

branches: 1.4.4;
nuke trailing whitespace

branches: 1.3.2; 1.3.10; 1.3.12;
NetBSD/acorn26 has used APCS-32 for years, so unifdef -U__APCS_26__.

branches: 1.2.2; 1.2.4; 1.2.6;
Local label fixup.

pull in ffs.S from libc for arm.
The main benefit is that ffs always runs in constant time.

Use common sources with userland.

merge ktrace-lwp.

branches: 1.4.16;
Duh, fix the align-to-32-bits code.

Xscale-optimised mem* functions, contributed by Wasabi Systems.
(Note: memcmp/memset improvements also benefit non-Xscale).

memcmp() - Compare 32-bits at a time if possible. Special-case 6-byte
comparisons, for the benefit of the network stack.

memset() - More loop unrolling, plus use of 'strd' instruction,
results in > 100% speedup on Xscale.

memcpy() - Big-endian support, unrolled loops, 'strd/ldrd/pld', plus
special-cases for very common length/alignment combinations
(at least in the kernel). Benchmarks show ~50% improvment on
Xscale.

memmove() - Big-endian support. Use fast memcpy(), above, if the regions
don't overlap. Otherwise unchanged.

branches: 1.2.2;
NetBSD/acorn26 has used APCS-32 for years, so unifdef -U__APCS_26__.

branches: 1.1.2;
Sync arm asm libkern files with libc's asm files.

Use common sources with userland.

merge ktrace-lwp.

branches: 1.6.16;
Xscale-optimised mem* functions, contributed by Wasabi Systems.
(Note: memcmp/memset improvements also benefit non-Xscale).

memcmp() - Compare 32-bits at a time if possible. Special-case 6-byte
comparisons, for the benefit of the network stack.

memset() - More loop unrolling, plus use of 'strd' instruction,
results in > 100% speedup on Xscale.

memcpy() - Big-endian support, unrolled loops, 'strd/ldrd/pld', plus
special-cases for very common length/alignment combinations
(at least in the kernel). Benchmarks show ~50% improvment on
Xscale.

memmove() - Big-endian support. Use fast memcpy(), above, if the regions
don't overlap. Otherwise unchanged.

big endian support

thanks to smi@sm.sony.co.jp

branches: 1.4.2;
NetBSD/acorn26 has used APCS-32 for years, so unifdef -U__APCS_26__.

Use .L prefix for all local labels.

branches: 1.2.8;
Tidy up some memory copying cruft:
bcopy.S is no longer needed
memmove and memcpy were both stacking r0 and unstacking it to keep the return value, so push this down into _memcpy.
rename _memcpy.S to memcpy.S.
memmove.S is now just a placeholder otherwise the make system automagically adds a memmove.c file to libkern.
memmove is just another entry point for memcpy.

branches: 1.1.2; 1.1.4; 1.1.6;
Create a shared set of ARM sources for libkern and move arm26 over to using
them.

These are identical to the current arm32 sources with the following exceptions:
- References to C labels are wrapped in _C_LABEL().
- Function returns have an alternate version inside #ifdef __APCS_26__.

Use common sources with userland.

merge ktrace-lwp.

branches: 1.3.2;
Re-sync with libc version of memcpy_arm.S, which has deals correctly
with big-endian ARM (non-XScale) systems.

nuke trailing whitespace

branches: 1.1.4; 1.1.10; 1.1.12;
Xscale-optimised mem* functions, contributed by Wasabi Systems.
(Note: memcmp/memset improvements also benefit non-Xscale).

memcmp() - Compare 32-bits at a time if possible. Special-case 6-byte
comparisons, for the benefit of the network stack.

memset() - More loop unrolling, plus use of 'strd' instruction,
results in > 100% speedup on Xscale.

memcpy() - Big-endian support, unrolled loops, 'strd/ldrd/pld', plus
special-cases for very common length/alignment combinations
(at least in the kernel). Benchmarks show ~50% improvment on
Xscale.

memmove() - Big-endian support. Use fast memcpy(), above, if the regions
don't overlap. Otherwise unchanged.

Use common sources with userland.

merge ktrace-lwp.

branches: 1.2.2; 1.2.4;
nuke trailing whitespace

branches: 1.1.4; 1.1.10; 1.1.12;
Xscale-optimised mem* functions, contributed by Wasabi Systems.
(Note: memcmp/memset improvements also benefit non-Xscale).

memcmp() - Compare 32-bits at a time if possible. Special-case 6-byte
comparisons, for the benefit of the network stack.

memset() - More loop unrolling, plus use of 'strd' instruction,
results in > 100% speedup on Xscale.

memcpy() - Big-endian support, unrolled loops, 'strd/ldrd/pld', plus
special-cases for very common length/alignment combinations
(at least in the kernel). Benchmarks show ~50% improvment on
Xscale.

memmove() - Big-endian support. Use fast memcpy(), above, if the regions
don't overlap. Otherwise unchanged.

Use common sources with userland.

merge ktrace-lwp.

branches: 1.4.2; 1.4.4;
nuke trailing whitespace

branches: 1.3.8; 1.3.10;
Xscale-optimised mem* functions, contributed by Wasabi Systems.
(Note: memcmp/memset improvements also benefit non-Xscale).

memcmp() - Compare 32-bits at a time if possible. Special-case 6-byte
comparisons, for the benefit of the network stack.

memset() - More loop unrolling, plus use of 'strd' instruction,
results in > 100% speedup on Xscale.

memcpy() - Big-endian support, unrolled loops, 'strd/ldrd/pld', plus
special-cases for very common length/alignment combinations
(at least in the kernel). Benchmarks show ~50% improvment on
Xscale.

memmove() - Big-endian support. Use fast memcpy(), above, if the regions
don't overlap. Otherwise unchanged.

branches: 1.2.16;
Tidy up some memory copying cruft:
bcopy.S is no longer needed
memmove and memcpy were both stacking r0 and unstacking it to keep the return value, so push this down into _memcpy.
rename _memcpy.S to memcpy.S.
memmove.S is now just a placeholder otherwise the make system automagically adds a memmove.c file to libkern.
memmove is just another entry point for memcpy.

branches: 1.1.2; 1.1.4; 1.1.6;
Create a shared set of ARM sources for libkern and move arm26 over to using
them.

These are identical to the current arm32 sources with the following exceptions:
- References to C labels are wrapped in _C_LABEL().
- Function returns have an alternate version inside #ifdef __APCS_26__.

Use common sources with userland.

merge ktrace-lwp.

branches: 1.4.16;
Xscale-optimised mem* functions, contributed by Wasabi Systems.
(Note: memcmp/memset improvements also benefit non-Xscale).

memcmp() - Compare 32-bits at a time if possible. Special-case 6-byte
comparisons, for the benefit of the network stack.

memset() - More loop unrolling, plus use of 'strd' instruction,
results in > 100% speedup on Xscale.

memcpy() - Big-endian support, unrolled loops, 'strd/ldrd/pld', plus
special-cases for very common length/alignment combinations
(at least in the kernel). Benchmarks show ~50% improvment on
Xscale.

memmove() - Big-endian support. Use fast memcpy(), above, if the regions
don't overlap. Otherwise unchanged.

branches: 1.3.2;
NetBSD/acorn26 has used APCS-32 for years, so unifdef -U__APCS_26__.

Use .L prefix for all local labels.

branches: 1.1.2; 1.1.4; 1.1.6; 1.1.18;
Create a shared set of ARM sources for libkern and move arm26 over to using
them.

These are identical to the current arm32 sources with the following exceptions:
- References to C labels are wrapped in _C_LABEL().
- Function returns have an alternate version inside #ifdef __APCS_26__.

Use common sources with userland.

branches: 1.3.18;
NetBSD/acorn26 has used APCS-32 for years, so unifdef -U__APCS_26__.

branches: 1.2.2;
Sync arm asm libkern files with libc's asm files.

Add an assembler version of strcmp, based on example code from the ARM
ARM. As an example of the performance difference that this provides
a Dhrystone score on my Shark goes from 213k to 261k.

Use common sources with userland.

branches: 1.2.18;
NetBSD/acorn26 has used APCS-32 for years, so unifdef -U__APCS_26__.

branches: 1.1.2;
Sync arm asm libkern files with libc's asm files.

branches: 1.2.2; 1.2.6; 1.2.10;
Fix typos (prX not cpX)

Add EHABI unwind stubs to libkern so prevent errors in linking if unwind
tables are present.