Shut up lint.

(joerg)

Add nl_langinfo_l, catopen_l and wcsftime_l.

(joerg)

fix gethostbyaddr prototype per TOG.

(christos)

Whitespace

(skrll)

G/C arm COMPILERCRTSTUFF

(skrll)

Add END()

(matt)

fix cfi_register -> cfi_offset

(matt)

Instead of trying to pass two kernels to beagleboard.conf, pass one and
derive the path to the second from the first.

(jmcneill)

Rework to allow thumb armv7 compilation.
Add atomic_simplelock.c for thumb

(matt)

Thumbify (and use .cfi ops)

(matt)

thumbify
add .cfi ops (for thumb)

(matt)

This is ARM only

(matt)

Add thumb version
Use STRONG_ALIAS

(matt)

Use STRONG_ALIAS
Add thumb variation

(matt)

Add .cfi ops
Thumbify

(matt)

Add cfi ops.
Thumbify

(matt)

make sure u_int is defined

(matt)

Add END() and clarify thumb/arm

(matt)

Thumbify

(matt)

Add END()

(matt)

Thumbify

(matt)

ip -> r2
teq -> cmp

(matt)

Thumbify

(matt)

ip -> r2
teq -> cmp

(matt)

Thumbify

(matt)

Use ip as a temporary

(matt)

Change previous use of r2 to r3

(matt)

teq -> cmp
ip -> r2
add/sub -> adds/subs
(thumbify part 1)

(matt)

For EABI, add .cfi ops

(matt)

Add .cfi for __ARM_EABI__
Thumbify

(matt)

Add END(memcpy)

(matt)

For Thumb, use naive version

(matt)

Thumbify

(matt)

Thumbify

(matt)

Add .cfi ops if EABI.
Thumbify.

(matt)

Thumbify

(matt)

Fix CLKF_INTR for EABI to (void) its argument so it isn't mark unused by the
optimizer.

(matt)

Fix <arm/locore.h> lossage

(matt)

Fix <arm/locore.h> lossage

(matt)

Fix <arm/locore.h> lossage

(matt)

Fix more <arm/locore.h> lossage

(matt)

Support __format_arg.

(joerg)

Rename fixup to update_global to better reflect the purpose.
Remove the various Current pointers, they are implementation internals
that shouldn't be leaked. Remove _CATEGORY_DEFAULT related handling and
directly pick the implemenation pointer from the C locale.
Merge the C locale data into global_locale.c with the exception of
_DefaultRuneLocale. Mark that one hidden. Move _RUNE_LOCALE and
_CITRUS_CTYPE into multibyte.h and pick up the correct rl_citrus_ctype
for non-default locales.

(joerg)

make sure the correct header files are picked up

(agc)

Revert previous.

(skrll)

Fix typo

(matt)

Fix <arm/locore.h> lossage.

(matt)

Include <arm/cpufunc.h> if !_LOCORE

(matt)

Don't include <arm/locore.h>, just <arm/armreg.h>

(matt)

<arm/locore.h> fallout (fixes some include ordering errors)

(matt)

Update comment.

(wiz)

getnameinfo is now thread-safe so remove bugs section.

(christos)

appease lint.

(christos)

Remove ancient spl* variants no longer used

(martin)

Properly recognize Thumb2 instructions.
(a thumb2 FP programs correctly runs now).

(matt)

Only add in __aeabi_read_tp.S when building for earm.

(skrll)

Add a CPU_IS_ARMV7_P() macro (nonv7 evals to false, v7 only true, otherwise
cpu_armv7_p is checked (and is set by cpuconf for cortex and pj4b).

(matt)

include <arm/locore.h>

(matt)

If _KMEMUSER, include <sys/intr.h>

(matt)

Expose SOFTINT_COUNT for _KMEMUSER

(matt)

Don't use cfi ops in non EABI MCOUNT

(matt)

Sort options. Consistency. Serial comma. Describe -F instead of -S twice :)

(wiz)

Use <arm/locore.h>

(matt)

Include <arm/locore.h>

(matt)

There are only 16 faults.

(matt)

Use blx ip if _ARM_ARCH_5
prepare to get of switchframe

(matt)

Add END(beagle_start)

(matt)

Move parts of cpu.h that are not needed by MI code in <arm/locore.h>
Don't include <machine/cpu.h> or <machine/frame.h>, use <arm/locore.h>
Use <arm/asm.h> instead of <machine/arm.h>

(matt)

Move parts of cpu.h that are not needed by MI code in <arm/locore.h>
Don't include <machine/cpu.h> or <machine/frame.h>, use <arm/locore.h>
Use <arm/asm.h> instead of <machine/arm.h>

(matt)

Move parts of cpu.h that are not needed by MI code in <arm/locore.h>
Don't include <machine/cpu.h> or <machine/frame.h>, use <arm/locore.h>
Use <arm/asm.h> instead of <machine/arm.h>

(matt)

Also add an output format argument to multigest(1).

This can take one of three values:

+ digest (meaning print output in digest(1) format)
i.e. "ALG (file) = digest"

+ openssl (meaning print output in openssl(1) format), or
i.e. "ALG(file)= digest"

+ multigest (meaning output is in multigest format)
i.e. "ALG (file) (substition) = digest"

The default is multigest format.

(agc)

Fix local symbol to be local
use END/ASEND

(matt)

Add options MODULAR

(matt)

fix comment.

(matt)

remove __thumb__ PIC specifics.  no different from ARM except being pc+4 not
pc+8

(matt)

struct pcb; for a function prototype.

(matt)

Add more TF_* symbols

(matt)

introduce an output separator to multigest(1), suggested by Thomas Klausner.

Use the output separator to separate the individual digests in the output.
The output separator defaults to the null string, i.e. no output separator.

(agc)

Move the MD parts of <arm/cpu.h> to here.

(matt)

Support thumb1 indirectly and thumb2 directly

(matt)

Nuke atomic.S since we have MI defined atomic ops

(matt)

#define __HAVE_FAST_SOFTINT

(matt)

Since nothing in <sys/intr.h> depends on <machine/intr.h>, include the
latter last so SOFTINT_COUNT and friends are defined before including it.
Since <machine/intr.h> may need to include <machine/cpu.h> which might need
SOFTINT_COUNT.

(matt)

document the changes to multigest_data() and multigest_file() signatures

(agc)

adapt the signature of the multigest_file() and multigests_data()
functions to return the output digest (albeit in raw format), and to
take the algorithm list as the first argument.  this makes it mirror
the normal digest *Data() and *File() functions signatures.

(agc)

Use  -fno-unwind-tables when creating a.out kernels.

(skrll)

Drop trailing whitespace.

(wiz)

Explain command line options. Thanks to wiz for the nudge!

(agc)

Remove empty file.

(joerg)

+ fix a bug whereby large multiple digests overflowed the context space
  because they weren't taking into account the current algorithms context
  space needs
+ add my own copyright where substantial work has been done
+ move to a portable, lightweight RFC1738-inspired encoding for VCS identifiers
  (we can't be assured that any of the vis(3) routines are going to be available)
+ re-do all the regression test results to avoid the RCS Ids being expanded
  in checked-in regression tests
+ make keccak sources thread-safe by moving the initialisation space into the
  keccak context space
+ print details of expected and calculated digests during check if the
  digests do NOT match (from Thomas Klausner, the original behaviour was
  from debugging tests - thanks!)

(agc)

multigest changes:

+ fix a bug whereby large multiple digests overflowed the context space
  because they weren't taking into account the current algorithm's context
  space needs
+ add my own copyright where substantial work has been done
+ move to a portable, lightweight RFC1738-inspired encoding for VCS identifiers
  (we can't be assured that any of the vis(3) routines are going to be available)
+ re-do all the regression test results to avoid the RCS Ids being expanded
  in checked-in regression tests
+ make keccak sources thread-safe by moving the initialisation space into the
  keccak context space
+ print details of expected and calculated digests during check if the
  digests do NOT match (from Thomas Klausner, the original behaviour was
  from debugging tests - thanks!)

(agc)

Revert previous; the SiI3114 is not even BAR/register compatible with
siisata(4).

(jakllsch)

Try to pick up the right location for BEAGLEBONE and BEAGLEBOARD kernels.

(jmcneill)

Enable sdiv/udiv instructions for both Thumb2 and ARM if hwdiv is supported.

(matt)

If hardware divide is supported on the CPU, predefine __ARM_ARCH_EXT_IDIV__

(matt)

Enforce that this must be not compiled using Thumb1

(matt)

Thumbify & add END

(matt)

inject a PMFE_SPEED_CHANGED when changing the speed.

(matt)

Add _ARM_ARCH_5T (needed for blx <reg> support)

(matt)

Add END

(matt)

Save r3 instead of r4 when calling CERROR in thumb mode.

(matt)

thumbify and add END

(matt)

Thumbify & optimize a little

(matt)

Use _INVOKE_CERROR
use stmia instead of two str
add END

(matt)

Rework SYSTRAP for thumb.
Add END to PSEUDO_NOERROR and PSEUDO

(matt)

Fix botch.

(matt)

__mcount is always arm code (even if compiled -mthumb).
Add .cfi for unwind tables

(matt)

Teach gcc to recognize a7 & a15 as a9 but with hwdiv.

(matt)

Don't define RETc if Thumb1
Add REL_SYM(a, b)
Make GOT_INIT* thumb aware.

(matt)

Make _libc_arm_fpu_present hidden

(matt)

new hostent related tests.

(christos)

test harness for gethostbyname()/gethostbyaddr() and their internal bits.
XXX[1]: How can we avoid using hard-coded hosts for DNS
XXX[2]: How do we test NIS?

(christos)

Add not advertised reentrant functions: {get,set,end}hostent_r,
gethostbyname{,2}_r, gethostbyaddr_r. Make getnameinfo(3) use
gethostbyaddr_r(3) so it is re-entrant (ahem __ypdomain). These
are not being advertised because there is a bunch of different
implementation of them that have a variety of type signatures.

If people want to follow someone's implementation, it is now easy.

(christos)

Make FPU_EMULATE work on NetBSD/hp300.  Commented out by default in GENERIC.

Tested on HP9000/362 without optional MC68882 FPU in its socket.

(tsutsui)

Prevent inlining of the main testcase body function.
Do not mark the inner functions static, as they might get cloned then
(ending up being called "myfunc3.clone.3" instead of "myfunc").

(martin)

Use makeinfo built from src/tools/texinfo.  Fixes PR toolchain/48123.
ok apb@

(gsutre)

Replace standard sentence by standard macro in standard section.

(wiz)

New sentence, new line.

(wiz)

Hardwire ld0 -> sdmmc0 and ld1 -> sdmmc1. Support local configuration in
BEAGLEBONE.local

(jmcneill)

Add a hidden version for libpthread.

(matt)

support beaglebone and beagleboard with the same image

(jmcneill)

When compiling for Thumb1, the swp instruction is not availble.  So this
stub is added to provide __cpu_simple_lock and __cpu_simple_lock_try via
thumb interwork support.  It is compiled with -mno-thumb.

(matt)

Add an EABI ARM version of MCOUNT that doesn't do ARM26 compat.
Add an EABI Thumb version of MCOUNT (doesn't do ARM26 compat either).
Add .size to all MCOUNT

(matt)

Add support for an optional ARCHSUBDIR/Makefile.inc

(matt)

Use the thumb1 versions of sev/wfe for thumb && armv6+.
if using armv5t don't do anything for thumb.

(matt)

If compiling for Thumb1, make __lwp_getprivate_fast call __aeabi_read_tp.

(matt)

Add _ARM_ARCH_T2 to indicate Thumb2 support
Don't define _ARM_ARCH_DWORD_OK disappear if compiling for Thumb1

(matt)

Fix tpyo.

(matt)

regen

(pooka)

translate ino_t to a portability-friendlier type

(pooka)

regen

(pooka)

Explicitly avoid MAXNAMELEN.  I'm not claiming it's pretty, but most
likely it's more bang for the buck than adding cpp support to the
sed script.

(pooka)

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

(matt)

Only assemble if !__ARM_EABI__

(matt)

Explicitly produce an error if the mythical kernel folk try to include
this header.

(pooka)

remove includes which haven't been necessary since the cudgel war

(pooka)

Give includers a way to say "yea I'll just deal with the type
compat on my own, thank you very many"

(pooka)

regen

(pooka)

type compat header list doesn't need to be autogenerated

(pooka)

revert pending fix: MAXNAMLEN causes trouble

(pooka)

Use more static and __dead.

(joerg)

regen

(pooka)

grab struct dirent

(pooka)

In the error paths where uaudio_chan_free_buffers() has to be called, acquire
the uaudio spinlock after the call because another spinlock is used indirectly
during the call.
Fixes a panic with LOCKDEBUG when a low-speed uaudio attached to a high-speed
hub is opened.

(aymeric)

Add .ARM.extab to the list of sections to be removed when converting to
aout format.

(martin)

Do not explicitly prevent unwind table generation

(martin)

Now that all architectures default to unwind data, remove the failure
for !amd64.

(joerg)

Save space and strip .eh_frame sections for the ramdisks as well.

(joerg)

Welcome to 6.1.1!

(snj)

Remove trailing space.

(wiz)

Remove trailing space, sort SEE ALSO.

(wiz)

Do not output unwind tables, they make no sense for the kernel.
Do not force MACHINE=arm, to allow building eabi kernels with MACHINE=earmv4

(martin)

Allow building of COMPAT_NETBSD32.

(martin)

This driver was actually pulled back to 6.1, so it first appeared in 6.1, not
7.0.

(rkujawa)

Link to X-Surf 100 docs, add info about the chip.

(rkujawa)

when building as a tool, only use struct tm tm_gmtoff if HAVE_STRUCT_TM_TM_GMTOFF is defined -- fixes cygwin build

(jmcneill)

regen

(jmcneill)

add test for struct tm tm_gmtoff member

(jmcneill)

Partially revert previous, there are still a few gotchas in
running ipv4/ipv6-only (too many dependencies on ifdefs).

(pooka)

Separate inet and inet6, allows inet6-only rump kernels.

(pooka)

Note AX88x9x MII support (in previous commit, which accidentally had no
commit log message), bump date.

(rkujawa)

(rkujawa)

Change the default value of rump kernels CPUs to 2.  It used to be
the number of host cores, but that value is overkill for most uses,
especially with massively multicore hosts.  Dozens of useless virtual
CPUs are relatively speaking expensive in terms of bootstrap time and
memory footprint.  On the other end of the spectrum, defaulting to 2
might shake out some bugs from the qemu test runs.

(pooka)

comments & whitespace police

(macallan)

pass HOST_SH to the build, to keep it from picking up a "bash"
from $PATH

(drochner)

remove some files which will be regenerated during build, avoids
writes into $BSDSRCDIR

(drochner)

add to CLEANFILES at some places, for generated files

(drochner)

Note X-Surf 100 support.

(rkujawa)

As MII support was added, add it also to synopsis.

(rkujawa)

Use mvnge AHI, #0x80000000 instead of mvnge AHI, ALO, lsr #1

(matt)

0x800000000 -> 0x80000000 (one too many zeroes)

(matt)

Fix movlt

(matt)

andlt -> movlt

(matt)