Move to a flat space register convention. %sr[4-7] are all now the space
number allocated to the process.  gcc produces (slightly) better code
with this convention.

Retain backwards compatiblity.

Welcome to 6.99.36

(skrll)

Whitespace

(skrll)

Use %t2 instead of %ret1 for style sake only.

(skrll)

Remove unnecessary %sr2 instructions in spstrcpy

(skrll)

remove unused variable to avoid warning from gcc 4.8.

(nonaka)

`Controller' takes two l's.

(riastradh)

Update

(skrll)

fix to pass collect memory address to xrstor.

(nonaka)

Repeat after me... don't delete entries mark them obsolete.

(skrll)

Note hppa switch to gcc 4.8

(skrll)

note sparc* switch to gcc 4.8.

(mrg)

switch sparc* to GCC 4.8.

(mrg)

Include new/missing files for vax

(martin)

Provide sqrtl alias

(martin)

Simple tanf() for vax

(martin)

Simplistic tanhf() for vax

(martin)

Simplistic log1pf() for vax

(martin)

log10f() for vax

(martin)

log2() for vax

(martin)

Add brute-force adaptions of the ieee version of exp2() and exp2f for
vax. Needs some polishing (and table adaption), but should get the build
going for now.

(martin)

Enable building ld.elf_so for powerpc
Produce a link map for ld.elf_so

(matt)

Make ELF64 aware.

(matt)

If marking a shared image TEXTREL, print out a warning about the section
that caused it.

(matt)

Welcome to 6.99.35

(matt)

add ep_entryoffset to exec_package so one can calculate the relocabase
of an ET_DYN image.

(matt)

When building a DEBUG kernel, mdsetimage into the .gdb as well (so if it's
used with gxemul, it has the md image in it.)

(matt)

Adapt to new elf_machdep.h
Add limited ELF64 support.

(matt)

Switch hppa to gcc 4.8

(skrll)

More comments.
rest of powerpc64 relocs

(matt)

powerpc64 changes

(matt)

Startup file for powerpc64.  (too many difference with ppc32 to do
#ifdef dance)

(matt)

a couple minor adjustments, sitting around since last july

(dholland)

bid the libtool.m4 lossage.

(christos)

Don't include corelow.o and solib.o the {powerpc,sh3}*-netbsd cases since
those are provided by Makefile.in.  This allows MKCROSSGDB to build again
for powerpc.

(matt)

Remove automatically generated file "libgcc-std.ver" while cleaning.

(tron)

Get GCC 4.8 closer to build with "USE_SSP" set to "yes".

(tron)

Drop tautologic check.

(joerg)

Fix abs usage.

(joerg)

move it to something that does not confuse cvs

(christos)

A couple of the atan tests are randomly failing.
Print the incorrect value.

(dsl)

Fix some typos.
Make the infinity and nan constants 'double' not 'long double'.

(dsl)

Fix build of DTrace with GCC 4.8.

(tron)

Move the #defines that simplified the test definitions and checks
  into a separate header than can be used by the other libm tests.
Make the subtest index 'unsigned int' so that simple constants (eg 0)
  print correctly.

(dsl)

apcs-gnu only passes one register on the stack.
ldrd always loads little endian (low address, low register).

(matt)

Define all the instruction conditions.
Use __SHIFTOUT to get the condition.
Add bool arm_cond_ok_p(uint32_t insn, uint32_t psr) inline
Add tlbtr inline

(matt)

Fix build with "USE_SSP" set to "yes".

Problem reported by B Harder on "current-users" mailing list.

(tron)

Delete the word "buggy" in the comment of NULLFS.
This has been added in 1999, but it's not a buggy now.
Adviced from tsutsui@, msaitoh@.

(isaki)

Enabled SYSV* options, which all ports have (and which is needed by 3rd
party software like SANE, e.g.).
Added ugen*.

(phx)

Current support for iterating over mnt_vnodelist is rudimentary.  Every
caller has to care about list and vnode mutexes, reference count being zero,
intermediate vnode states like VI_CLEAN, VI_XLOCK, VI_MARKER and so on.

Add an interface to iterate over a vnode list:

void vfs_vnode_iterator_init(struct mount *mp, struct vnode_iterator **marker)
void vfs_vnode_iterator_destroy(struct vnode_iterator *marker)
bool vfs_vnode_iterator_next(struct vnode_iterator *marker, struct vnode **vpp)

vfs_vnode_iterator_next() returns either "false / *vpp == NULL" when done
or "true / *vpp != NULL" to return the next referenced vnode from the list.

To make vrecycle() work in this environment change it to

bool vrecycle(struct vnode *vp)

where "vp" is a referenced vnode to be destroyed if this is the last reference.

Discussed on tech-kern.

Welcome to 6.99.34

(hannken)

Don't probe beyond chan_ntargets in atapi_probe_bus in the "all" case.

PR/48626: ahci_atapi_probe_device panic with kmemguard

(skrll)

It is md.${MACHINE}.${MACHINE_ARCH}

(martin)

Tidy up Makefile of cyclic and dtrace modules to support ARM

(ozaki-r)

Add cyclic_machdep.c for ARM

The code is still almost same as i386's one.

(ozaki-r)

Now that maphist is statically initialized, remove initialize tests

(matt)

Add defined(__arm__)

(ozaki-r)

Use native atomic function instead of DTrace's own one

So we don't need to implement it for each architecture.

This change is from FreeBSD (not as-is though).

(ozaki-r)

No functional change:
- Rename function for consistency.
- Add prototypes.
- KNF.

(msaitoh)

Use UVMHIST_INITIALIZER (KERNHIST_INITIALIZER) to statically initialize
maphist.  This allows maphist to used very very early in boot well before
uvm has been initialized.

(matt)

Add KERNHIST_INITIALIZER(NAME,BUF) macro to be used with KERNHIST_DEFINE to
static initialize a kern_history.  After boot, it still needs to linked
to list of kern_history's and a KERNHIST_LINK_STATIC(NAME) macro was added
to do that.

(matt)

Changes to transit-20140304

+ add 2 accessor functions for lists/dicts, one to get a field by its index,
  and the other for dicts to get the field value by name

+ bump version number to 20140304

(agc)

merkletree - minor changes for packaging infrastructure

no functional change

(agc)

multigest - minor cosmetic changes

bump version number to 20140304

(agc)

Add yet another TLA.

(ginsbach)

Build net/zlib.c for DTrace

This is needed for building DTrace on ARM.

(ozaki-r)

Fix dtrace_cas32 logic

- compare with cmp, not new
- return old value

(ozaki-r)

frame -> tf
printf out the insn, faulting pc, lr, sp when we get an undefined insn
int the kernel.

(matt)

rearrange some printfs.
when using the extended mmu, make sure to disable ttbr0 until there is an
activated pmap.

(matt)

push/pop unsaved registers before using printf.

(matt)

Only do UVMHIST_LOG is maphist have its entry buffer initialized

(matt)

Use uint32_t instead of u_int32_t

(ozaki-r)

mknative-gcc for GCC 4.8 and coldfire.

(mrg)

coldfire does not want m68k/t-floatlib.
enable powerpc*-netbsd* -- ie, match ppc64.

(mrg)

Disable abs() checking, a lot of computations get implicit truncations
or use unsigned values.

(joerg)

partial mknative-gcc for gcc 4.8 and ppc64.

(mrg)

Make pointer NULL check non-fatal for a bogus condition.

(joerg)

Make the abs() use check non-fatal for a bogus use in the LFS cleanerd.

(joerg)

When using Clang, disable integrated assembler for genassym. It creates
syntactically invalid assembler, which is normally not a problem as it
is also using -S.

(joerg)

Update build glue for LLVM/Clang r202566. News:
- match ARM CPU default of GCC
- better support for SPARC64
- type checkings for abs/fabs
- null pointer checks against struct member
- various other improvements

(joerg)

Add GENASSYM_CPPFLAGS in two more places.

(joerg)

check the correct field in the structure in dsa_verify - pointed out by
joerg - thanks!

(agc)

on 601 CPUs skip to normal memcpy if both source and destination are 32bit
aligned

(macallan)

Don't export __sync* if _KERNEL || _STANDALONE are defined.
(except if _RUMPKERNEL is defined)

(matt)

Remove duplicate END(y) macro

(matt)

Don't byte swap the data, assume the bus_space will do it.

(matt)

Introduce GENASSYM_CPPFLAGS for options during genassym processing.
Consistently drop assembler flags.

(joerg)

forgotten mknative-gcc 4.8 for mips64el.

(mrg)

Replace xalloc(), xrealloc() and xfree() with malloc(), realloc() and free().

Per following Xorg changes:
> Replace X-allocation functions with their C89 counterparts
http://cgit.freedesktop.org/xorg/xserver/commit/?id=3f3ff971ecff9936cebafc813af9193b97bba89c

(tsutsui)

Rename md.evbppc to md.evbppc-powerpc, to separate it from the evbppc64
one (which is nonexistant for now).

(martin)

Add missing END for each function

(ozaki-r)

Get rid of unused variable definition za from mount.h

There is no user for it. Even worse, it breaks dtrace build for arm.

(ozaki-r)

bump local pkg version.

(mrg)

don't look for po/Makefile.

(mrg)

mknative-gcc 4.8 for most platforms.  missing:  ppc64, m68010, ia64, coldfire.

(mrg)

include dbxelf.h elfos.h freebsd-spec.h rs6000/sysv4.h netbsd.h
and netbsd-elf.h on powerpc64-netbsd.

port netbsd64.h forward from GCC changes.  (it's a pity that
a vast portion of linux64.h isn't in some common header that
netbsd64.h and freebsd64.h could also used.  there's 100s of
lines of copypasta here, oh well.)

(mrg)

Add a different version vfp_fpscr_changable if FPU_VFP was not defined.
If no FPU was found, reinit vfp_fpscr_changeable/default to values appropriate
for softfloat.

(matt)

compat16 -> compat_16

(matt)

Fix #if/#endif nesting

(matt)

#include <m68k/m68k.h>

(matt)

fix typo.

(matt)

The fixuns* come from compiler_rt so softfloat doesn't need to provide them.

(matt)

Avoid a warning and just do the trap #1 instead of calling sigreturn inline.
If it fails, it'll fall into the botch case so don't bother calling CERROR
on failure.

(matt)

Make this work with coldfire.

(matt)

use _KERNEL_OPT around #include

(matt)

Make this coldfire compatible

(matt)

Fix non-EABI loading of argument.  Deal with endian issues.
Fixes PR/48635

(matt)

Add atomic_sub_64.S

(matt)

Load new value from correct stack location in _atomic_cas_64_up

(matt)

Fetch value from correct stack location.  Push an even number of registers
so ldrd won't fail.

(matt)

Reformat a little.  Add m68k quirk

(matt)

Fix wrong conditional statement in dtrace_debug_output

Reported in PR kern/48547 by ozaki-r.

(ozaki-r)

Add almost-pass-through configure and Makefile.in scripts for the
package distribution. No functional change

(agc)

Minor changes to multigest

Introduce a new multigest_format_raw() function to the multigest
library, which can be used to format a raw multigest value into a
formatted hexadecimal number.

Add a new multigest_algs_rawsize() function, which will return the
space needed (in bytes) to hold a raw multigest for the algorithms
specified. The multigest_get_rawsize() function continues to work
on a multigest_t structure which has already been initialised.

Make blake2 compile without a c99 compiler.

Add faux-autoconf infrastructure for packaging purposes.

(agc)

Remove unnecessary files.i2c inclusion

files.i2c is already included in conf/files. So we don't need to include
it in this file.

This fix shuts up many warnings on building sys/modules/iic.

(ozaki-r)

Use the global pci_mode to avoid 'set but not used' warnings from gcc 4.8.3.

(dsl)

apply netbsd not rebuilding to mips-tables.opt; fixes r/o src builds.

(mrg)

Avoid running a helper program for atf-sh as a test.

Should fix failures observed in PR bin/48624.

(jmmv)

Pass down TOOL_STAT here as well.

(joerg)

also include bsd.own.mk, so value for RUMP_NPF_TESTING can come from mk.conf

(pooka)

Spell it TOOL_STAT.

(joerg)

Use _KERNEL_OPT around #include "opt_ppcarch.h"

(matt)

add man page for rumpfs

(pooka)

Use nbstat when cross-compiling, unbreaks build on Linux.

(joerg)

Remove the print of the rounding mode.
It was added in case it was non-zero - which it isn't.
It still isn't clear why acos(-1) gives the wrong result on some amd64
  systems.

(dsl)

Fix the memory limiter to work for non-pagedaemon threads again

reported by njoly

(pooka)

mention ppc601 support

(macallan)

config goop for ppc601 support
GENERIC_601 more or less resembles a Power Macintosh 7200 with a few drivers
thrown in for convenience while keeping the kernel itself smaller than 4MB

(macallan)

support ppc601
from scole_mail, ok matt@

(macallan)

on ppc601 do byte-wise copies when in _KERNEL
from scole_mail, ok matt@

(macallan)

More control bits.
SYNCHPRIM defines.

(matt)

If we are going to map all of physical memory, we don't need to allocate
l2 page tables for the kernel.

(matt)

For now, make it possible to not specify -D_NPF_TESTING by compiling
with RUMP_NPF_TESTING_=no.  This in turn makes it possible to use npf
in a rump kernel for real packet filtering instead of just testing.

no change to default

(pooka)

Pull in the changes from the dpdk, netmap and snabb switch repos.
There are two major changes:

1) All thread context policy is pushed down to the hypercalls.
  This is meant to help performance-mongering implementations be
  able to control packet scheduling better (e.g. pin down packet
  reception to certain physical cores).
2) Generalize linkstr, meaning that the interface can now take an
  arbitrary string which is passed to the create() hypercall.  This can
  be used to map backend device characteristics to the rump kernel
  interface instance.

(pooka)

Restore the hp700 file entries and mark them obsolete.

Don't mark any hppa files obsolete - delete them.

(skrll)

Define HAVE_FENV_H on architectures that support it.
What a mess, we need a central place for this!

(martin)

Correct usage of clri.

(blymn)

Remove all cargo-cult #ifndef __vax__ from the tests, that just paper
over bugs in the vax libm.

(martin)

Fix build for platforms w/o fenv.h.
Remove some bogus #ifdef __vax__ (and add a few, hopefully non-bogus,
new ones).

(martin)

Prefix link to cpsw with 'evbarm'.

(wiz)

New sentence, new line.

(wiz)

Further cpu_info inits in cpu_hatch

(matt)

Add more instructions including ldrex/strex variants

(matt)

Add a mpsafe flag to the intrsource

(matt)

Query the media and vfp feature registers to determine what our default
mode should be and what bits in the fpscr can be changed.
Print what features are supported:
vfp0 at cpu0: NEON MPE (VFP 3.0+), rounding, NaN propogation, denormals

(matt)

Add most of the cpu_* variables to <arm/locore.h>
Add and initialize cpu_synchprim_present

(matt)

Drop in a fabs() after the fsqrt().
While it may seem pointless, it the rouding mode is set to round towards
minus infinity then acos(-1) calculates atan2(sqrt(1.0 - 1.0), -1) the subtract
generates -0.0 which sqrt() preserves.
atan2(-0, -1) is -pi, but acos(-1) is expected to be +pi.
This might 'fix' the test failures seen in some environments, but they
are not failing due to an obvously incorrent x87 rounding mode.

(dsl)

sun4v: SUN4V_TLB_TSB_LOCK define should have the same LL (instead of UL) as the other defines

(palle)

merge GCC r208254.

(mrg)

Move IO KVA so it doesn't share the same TLB index as the kernel.

(matt)

Add micphy

(ozaki-r)

Add micphy.4

(ozaki-r)

Mention micphy(4)

(ozaki-r)

Add micphy(4) manpage

(ozaki-r)

Don't core-dump if a type cannot be resolved. Still gives an error.

(christos)

Fix bundling of the atf version into pkgconfig files and atf-version.

Sigh; one more attempt.  This time I'm sure I've verified that the
.pc files contain the right number and that atf-version also outputs
the right stuff...  Both with a clean and non-clean obj directory.

Should fix part of the problems reported in PR bin/48624.

(jmmv)

Include the subtest number in any error output.
Also temporarily print the rounding mode.
I think that acos(-1) is ending up with the wrong sign because the test
  is being run with 'round towards -ve infinity' set.
I think it getting set somewhere and causing this test to fail.
The acos() code probably needs fixing - it shouldn't depend on the
round mode like this. But first I want to know if this if the error.

(dsl)

If we are compiling with -g and using gcc4.8 and dtrace downgrade
dwarf generation to version 2 for now.

(christos)

Make cleanup routines actually work.

The only way to pass global state from the body to the cleanup is via the
file system.

Fixes leaks of global system resources (in all cases, given that the body
does not by itself clean things up).

(jmmv)

Remove unnecessary and broken cleanup routine.

(jmmv)

Don't overwrite MACHINE_ARCH if it is already set.

(joerg)

Initialize variables used to setup a dummy element.
XXX bogus, needs review

(joerg)

Mark cpu_reset as dead and add an explicit __builtin_unreachable for a
terminal function call for clang's noreturn diagnostic.

(joerg)

Be explicit about the double register use.

(joerg)

Bump SYMTAB_SPACE for clang.

(joerg)

Bump sshramdisk size to 3800k, reserve 4MB in the install kernel.

(joerg)

Clang needs a slightly larger MD image, so bump the size to 2300k.

(joerg)

GC ipending.

(joerg)

Skip a tautologic check if we know it is one.

(joerg)

GC ipending.

(joerg)

Make the empty loop body explicit.

(joerg)

Explicitly enable VFP support for vfp_init.c. Fixes build of
INTEGRATOR_CP with clang.

(joerg)

Use assembler code for 32-bit sparc compat library on sparc64.

(nakayama)