GCC sanitizers: only build the sanitizers if $MKGCCCMDS != "no".

there's a minor use-case for the configuration of "building sanitizers
for distribution, but not the compiler", but it's pretty uncommon, and
the latest sanitizers don't build at all on m68000 so enhance the
option to not build /usr/bin/gcc (which simply can't modern GCC on a
sun2) to not build the sanitizers either.

(if we were to split these, MKGCCSANITIZER would be good variable.)

update set lists for far too many files, including fixing MKDEBUG for
a couple of builds.

don't mark the gcc-10 + sanitizer directories obsolete yet.

they're still created by the build, and they're owned by the base set.
i'll clean the up when i delete gcc-10 entirely.

branches: 1.300.2;
distrib/lists/sets/comp: Add gcc=12 obsolete entries for gcc-10/*.h.

mark all previously gcc=9 files as obsolete.

also mark a bunch of obsolete gcc headers as "comp-obsolete" "set".

should fix a 'postinstall fix obsolete' issue reported by uwe@.

more updates for sets and GCC 12

- renable sanitizer, add new shlib versions, enable tsan for arm64,
including moving some stuff into the per-cpu/arch dir from 'mi'.
- fix mips64, 68k, and ia64
- mark many any-gcc-version things just as "gcc" and remove dups
- mark many gcc=9 only things as obsolete
- mark many gcc=10 only things as obsolete with gcc=12 (more could
be done here.)

various updates for GCC 12

- bump some shlibs
- add new header files
- add compiler-version header files
- disable sanitizer libraries for gcc12 for now

Unbreak creation of i386 PIEs on amd64.

Functional patch by RVP; set lists updated accordingly.

This fixes PR toolchain/57380.

Add the .xsce scripts

add some files that are obsolete with binutils 2.39

sort

Turn on -z separate-code for x86, like linux has done. This splits the text
and data segment load sections into two pieces, for at total of 4 PT_LOAD
sections: two for text (r--, r-x) and two for data (r--, rw-). For x86_64
this changes the default max page size from 2M to 4K.

annotate the 234 linker scripts

distrib/sets: remove more extra keywords from obsolete entries

Found by fmt-list.

Move tsan entries marked machine=amd64 from .mi files into md.amd64

Move the sanitizer libraries from 9.0.0 -> 13.0.0 to match our current clang
version.

Update LLVM build system for 249b40b558955afe5ac2b549edcf2d7f859c8cc9

This enables the use of modules for a significant build performance gain
when building with clang as host compiler or when using HAVE_LLVM=yes.
Switch libc++ to using the copy from the mono-repo.

branches: 1.286.2;
Add <x86/gdt.h> (we could gc now the <machine/gdt.h> files if we fix the
sources that use them.

various updates for GCC 10 support. remove dead gcc7/8 support
and properly mark obsolete files in gcc9/gcc10.

testing on amd64 and shark for GCC 9, and many platforms for GCC 10,
(though not arm yet.)

Add missing lint library

Add HAVE_NVMM and use it to control the build of NVMM related components.
Defined to "yes" on amd64, "no" everywhere else.

mostly support GCC 9 builds. some what tested with GCC 9 but
tested to not break GCC 8 builds for x86, arm*, vax, and sparc*.

Add back <xen/xenio3.h>, it's needed by xentools413

Sort and deduplicate set lists.

<xen/xenio.h> is in fact used by the newly imported xentools 4.13,
reinstantiate it

header is for use by xentools only, which define domid_t themselves,
so it should not be a problem the header doesn't define it per PR port-xen/52874

don't install xen/xenio.h and xen/xenio3.h anymore, xentools don't use it

PR port-xen/52874

Add /usr/include/xen/intrdefs.h to sets
Add xen to KLINKFILES for amd64/i386

Install LLVM sanitizers to /usr/lib/clang/9.0.0

This is the proper location where clang 9.0.0 looks for them.

Update LLVM to 10.0.0git (01f3a59fb3e2542fce74c768718f594d0debd0da)

gcc 8 set lists updates. i don't appear to have broken gcc 7 builds
but i only tested a couple of them.

Add new multiboot2.h header.

Correct the build rules of LLVM sanitizers

Fix generation of .a.syms files in LLVM sanitizers. The old rules weren't
functional and sometimes they were fatal in the build.

Add the .a.syms files to distribution sets.
These files are generated only in certain .a files in 64bit version of
sanitizers.

Correct the location of files with LLVM sanitizers. Move the LLVM sanitizer
entries from base to comp files.

branches: 1.271.2;
Revert MKLLVMCMDS changes.

split MKLLVM into MKLLVM and MKLLVMCMDS, like MKGCC and MKGCCCMDS:

- only build clang or clang-tblgen if MKLLVMCMDS != "no"
- move HAVE_MESA_VER earlier so we can detect mesa 18 and enable
MKLLVM on platforms that need it
- move the clang commands and headers into llvmcmds group
- only build the clang specific LLVM libraries if MKLLVMCMDS != "no"

much of this work came from maya@'s github commit
f90685c11d4460d3098fa35f48b58d1893e974e0, reworked for HAVE_MESA_VER
and llvmcmds vs MKCLANG.


build tested on several platforms and on amd64 and i386 with several
different build options (clang build, gcc build, gcc build with llvm
libs).

get rid of binutils=227-specific files.

Sort, and add a couple obsolete files for binutils=231.

- add new entries of gcc=7 and libgomp, libasan, liblsan, libstdc++, libubsan
- add missing header files fpr gcc=7
- add some missing gcc=5 obsolete entries
- create ./usr/include/dev/bluetooth in mtree not make, and move it into the
base set where all other directories are.

add basic gcc-7 support in sets lists and filesystem. mips gained
a header. assumes it's otherwise mostly the same as gcc-7.

- mark gcc=5 files obsolete
- make includes always available, not only when gcccmds
- mark files obsolete consistently

new binutils is 231

add nvmm lint library

Fix the libnvmm sets, do the same as libx86_64.

Add libnvmm, NetBSD's new virtualization API. It provides a way for VMM
software to effortlessly create and manage virtual machines via NVMM.

It is mostly complete, only nvmm_assist_mem needs to be filled -- I have
a draft for that, but it needs some more care. This Mem Assist should
not be needed when emulating a system in x2apic mode, so theoretically
the current form of libnvmm is sufficient to emulate a whole class of
systems.

Generally speaking, there are so many modes in x86 that it is difficult
to handle each corner case without introducing a ton of checks that just
slow down the common-case execution. Currently we check a limited number
of things; we may add more checks in the future if they turn out to be
needed, but that's rather low priority.

Libnvmm is compiled and installed only on amd64. A man page (reviewed by
wiz@) is provided.

Add NVMM - for NetBSD Virtual Machine Monitor -, a kernel driver that
provides support for hardware-accelerated virtualization on NetBSD.

It is made of an MI frontend, to which MD backends can be plugged. One
MD backend is implemented, x86-SVM, for x86 AMD CPUs.

We install

/usr/include/dev/nvmm/nvmm.h
/usr/include/dev/nvmm/nvmm_ioctl.h
/usr/include/dev/nvmm/{arch}/nvmm_{arch}.h

And the kernel module. For now, the only architecture where we do that
is amd64 (arch=x86).

NVMM is not enabled by default in amd64-GENERIC, but is instead easily
modloadable.

Sent to tech-kern@ a month ago. Validated with kASan, and optimized
with tprof.

Update LLVM/Clang build system to r337282.

Remove the kernel PMC code. Sent yesterday on tech-kern@.

This change:

* Removes "options PERFCTRS", the associated includes, and the associated
ifdefs. In doing so, it removes several XXXSMPs in the MI code, which is
good.

* Removes the PMC code of ARM XSCALE.

* Removes all the pmc.h files. They were all empty, except for ARM XSCALE.

* Reorders the x86 PMC code not to rely on the legacy pmc.h file. The
definitions are put in sysarch.h.

* Removes the kern/sys_pmc.c file, and along with it, the sys_pmc_control
and sys_pmc_get_info syscalls. They are marked as OBSOL in kern,
netbsd32 and rump.

* Removes the pmc_evid_t and pmc_ctr_t types.

* Removes all the associated man pages. The sets are marked as obsolete.

branches: 1.257.2; 1.257.4;
obsolete some GCC 4.8 files.

it's called avx512bwintrin.h.

more updates for GCC 6. with other uncommited changes, i386 seems
to work well enough for hello.c, hello.cc and hello.m.

updates for GCC 6.4: add gcc-6 specific headers, combine some
all-gcc files, bump the shlib versions on new libs.

convert HAVE_GCC handling to modern GCC release numbering:
- HAVE_GCC=5 is now the default (vs. HAVE_GCC=53 we've been using for
GCC 5.4 and GCC 5.5.)
- remove some more GCC 4.8 code. we don't support GCC 4 here.
- adjust set lists to gcc=5 from gcc=53.

add some basic HAVE_GCC=6 handling (totally unused so far.)

pcommitintrin.h has been removed from gcc 5.5

Remove vm86.

Pass 4.

Should be comp-obsolete.

Remove ibcs2_machdep.h on i386, and don't install it.

Move freebsd_machdep.h into sys/compat/freebsd, and don't install it. Now,
the compat_freebsd files are all contained in sys/compat/freebsd.

Update LLVM build glue for 5.0RC1.

Restore marking ./usr/include/i386/svr4_machdep.h obsolete
instead of removing

(so it can be removed by postinstall)

Remove svr4_machdep.h right away, no one should include it.

If i386/svr4_machdep.h is no longer to be installed in /usr/include
it should no longer be listed in the sets lists as installed.
Mark it obsolete (hopefully unbreak builds.)

branches: 1.243.4;
Introduce PT_GETDBREGS and PT_SETDBREGS in ptrace(2) on i386 and amd64

This interface is modeled after FreeBSD API with the usage.

This replaced previous watchpoint API. The previous one was introduced
recently in NetBSD-current and remove its spurs without any
backward-compatibility.

Design choices for Debug Register accessors:
- exec() (TRAP_EXEC event) must remove debug registers from LWP
- debug registers are only per-LWP, not per-process globally
- debug registers must not be inherited after (v)forking a process
- debug registers must not be inherited after forking a thread
- a debugger is responsible to set global watchpoints/breakpoints with the
debug registers, to achieve this PTRACE_LWP_CREATE/PTRACE_LWP_EXIT event
monitoring function is designed to be used
- debug register traps must generate SIGTRAP with si_code TRAP_DBREG
- debugger is responsible to retrieve debug register state to distinguish
the exact debug register trap (DR6 is Status Register on x86)
- kernel must not remove debug register traps after triggering a trap event
a debugger is responsible to detach this trap with appropriate PT_SETDBREGS
call (DR7 is Control Register on x86)
- debug registers must not be exposed in mcontext
- userland must not be allowed to set a trap on the kernel

Implementation notes on i386 and amd64:
- the initial state of debug register is retrieved on boot and this value is
stored in a local copy (initdbregs), this value is used to initialize dbreg
context after PT_GETDBREGS
- struct dbregs is stored in pcb as a pointer and by default not initialized
- reserved registers (DR4-DR5, DR9-DR15) are ignored

Further ideas:
- restrict this interface with securelevel

Tested on real hardware i386 (Intel Pentium IV) and amd64 (Intel i7).

This commit enables 390 debug register ATF tests in kernel/arch/x86.
All tests are passing.

This commit does not cover netbsd32 compat code. Currently other interface
PT_GET_SIGINFO/PT_SET_SIGINFO is required in netbsd32 compat code in order to
validate reliably PT_GETDBREGS/PT_SETDBREGS.

This implementation does not cover FreeBSD specific defines in their
<x86/reg.h>: DBREG_DR7_LOCAL_ENABLE, DBREG_DR7_GLOBAL_ENABLE, DBREG_DR7_LEN_1
etc. These values tend to be reinvented by each tracer on its own. GNU
Debugger (GDB) works with NetBSD debug registers after adding this patch:

--- gdb/amd64bsd-nat.c.orig 2016-02-10 03:19:39.000000000 +0000
+++ gdb/amd64bsd-nat.c
@@ -167,6 +167,10 @@ amd64bsd_target (void)

#ifdef HAVE_PT_GETDBREGS

+#ifndef DBREG_DRX
+#define DBREG_DRX(d,x) ((d)->dr[(x)])
+#endif
+
static unsigned long
amd64bsd_dr_get (ptid_t ptid, int regnum)
{


Another reason to stop introducing unpopular defines covering machine
specific register macros is that these value varies across generations of
the same CPU family.

GDB demo:
(gdb) c
Continuing.

Watchpoint 2: traceme

Old value = 0
New value = 16
main (argc=1, argv=0x7f7fff79fe30) at test.c:8
8 printf("traceme=%d\n", traceme);

(Currently the GDB interface is not reliable due to NetBSD support bugs)

Sponsored by <The NetBSD Foundation>

branches: 1.242.2;
Update LLVM build glue for pre-4.0.0 r291444.

Tag cleanup: if the last column contains 'obsolete', the one before should too.

Add support for hardware assisted watchpoints/breakpoints API in ptrace(2)

Add new ptrace(2) calls:
- PT_COUNT_WATCHPOINTS - count the number of available hardware watchpoints
- PT_READ_WATCHPOINT - read struct ptrace_watchpoint from the kernel state
- PT_WRITE_WATCHPOINT - write new struct ptrace_watchpoint state, this
includes enabling and disabling watchpoints

The ptrace_watchpoint structure contains MI and MD parts:

typedef struct ptrace_watchpoint {
int pw_index; /* HW Watchpoint ID (count from 0) */
lwpid_t pw_lwpid; /* LWP described */
struct mdpw pw_md; /* MD fields */
} ptrace_watchpoint_t;

For example amd64 defines MD as follows:
struct mdpw {
void *md_address;
int md_condition;
int md_length;
};

These calls are protected with the __HAVE_PTRACE_WATCHPOINTS guard.

Tested on amd64, initial support added for i386 and XEN.

Sponsored by <The NetBSD Foundation>

Sort the unsorted sets lists (they should all be correctly sorted now.)

everyone is now on binutils-2.26 or more

branches: 1.237.2;
GCC 5.3 sets for amd64.

Update build system for LLVM 3.8.0rc3.

add missing cpu_rng.h includes

handle both binutils 2.23 and 2.26

now that we only support gcc 4.8, make all the gcc tags just "gcc".
this should help with next-gcc, when we can add back gcc=48 and
gcc=5x tags, as necessary.

obsolete and clean up gcc=3 (3.3), gcc=4 (4.1), and gcc=45 (4.5) sets.

Update some of the set lists to use the compatdir/compatfile
compattestsdir/compattestsfile as well the kmod auto-entry creation.
Remove entries that are now instantiated by the above.

Update set lists for LLVM 3.6RC1.

stuff added by blacklist

Install cpuid.h on X86 and arm_neon.h on ARM.

Consistently tag libasan with gcc=48

Add libasan

sort

branches: 1.224.2;
Update build glue for LLVM/Clang 3.6svn r215315. Changes include various
performance improvements across all backends and a much improved PPC32
backend.

Rename sljitarch.h to sljit_machdep.h.

infrastructure for bind-9.10.0-P2:
adding the tools delv and dnssec-importkey
adding libirs to support delv

Drop entries for LLVM/Clang older than 3.4.

Update build system for LLVM/Clang snapshot r209886. This brings in the
merged AArch64 backend and the move of eARM to DWARF exception handling
for NetBSD. C++11 is required for building.

sort

/usr/include/g++/bits/${CONFIG} directories are created independent of
MKGCC, so fix file lists.

branches: 1.217.2;
Fix for /usr/include/g++/bits directory, which is only for MKGCC=yes

Unobsolete fpu.h

Mark i386/npx.h obsolete.
I'd forgotten it was released here as well.

Move sys/arch/amd64/amd64/fpu.c and sys/arch/amd64/include/fpu.h
into sys/arch/x86 in preparation for using the same code for i386.

Add cpu_extended_state.h

Update LLVM family to 3.5svn r198450. Interesting changes include:
- Improvements in the SPARC64 support
- ARM backend can almost build a full NetBSD/EABI system with the
integrated assembler
- PowerPC grows disassembler support
- LLD supports copy relocation, beginning of MIPS backend
- New diagnostic for use of bool expression as size_t argument for
strncmp and friends

Drop superflous "picinstall" flag on "obsolete" items

Reported by apb@, thanks!

Mark all _pic.a archive files with the 'picinstall' flag.

This fixes checkflist with MKPICINSTALL=no.

ok christos@.

Make /usr/include/g++/bits/i386 and x86_64 not depend on gcccmds.

Directories are created even when gcc=48 is not true, so remove
that flag as well.

Fix the build for amd64.

Even though the files are installed only if "compat" is set, the
/usr/include/g++/bits directories are unconditionally created via
the mtree files.

update set lists for gcc 4.8.

_g libraries don't belong here.

more gcc-4.8 files

add support for GCC 4.8 sets.

sort

Update LLVM/Clang snapshot to r192551 for various bug fixes.

Fix amd64 build sets with new architectures.

Update LLVM/Clang snapshot to r191105. This brings in a better detection
of unused static variables and functions. Disable this for some external
code and for ioconf.c in the kernel.

Sort entries in strcmp order.

Update LLVM/Clang snapshot to r182734. Brings initial support for
integrated PowerPC assembler.

Introduce MKLIBSTDCXX, defaulting to yes.

Add build glue for libc++.

Update LLVM/Clang snapshot to r180004. This brings in the last pieces of
C++11 support in Clang as well as initial SPARC64 support.

more libevent set fixes

libevent sets

Hook Lutok into the build. This is all protected by the MKKYUA guard.

- fix the debug sets which were broken because of _LIB_PREFIX lossage.
- add missing tpm files.

Update LLVM/Clang snapshot to r172366. Catch up with the move past 3.2
release.

Add a new "debug" set that gets built when ${MKDEBUG} is set. On evbarm:
-rw-r--r-- 1 root src 35806068 Jan 16 10:47 base.tgz
-rw-r--r-- 1 root src 57374069 Jan 16 10:48 comp.tgz
-rw-r--r-- 1 root src 101286193 Jan 16 10:49 debug.tgz
...
This allows always building the debugging code and libraries but not
installing it (although now it is placed in the default sets if ${MKDEBUG}
is set

bump libmagic

add entries built for MKDEBUG and MKDEBUGLIB to the set lists
Thanks to Hisashi T Fujinaka <htodd@twofifty.com> for pointing out the gap.

bump libssh

Adjust libraries debug entries following recent bind update.

Fix set lists after the merge of netpgpverify(1) and libnetpgpverify(3)
from the agc-netpgp-standalone branch.

Update LLVM/Clang snapshot to r168187 for various bug fixes.

bump resolver library.

Include sljitarch.h even when MKSLJIT=no.

Build sljit test when MKSLJIT != no and set MKSLJIT to yes on amd64 and i386.

Update LLVM/Clang snapshot to r166772. Now builds without RTTI and
exceptions. Also features the first round of a loop vectorizer.

*Sigh* provide, by default, the dumb "allocate 64K on the stack" variant
of the lzf_compress API by default. This is because there is evidently
code floating around out there in the world with broken autoconf scripts that
don't check for LZF_STATE_ARG in the lzf header files and just assume the
dumb version of the API.

We rename the sane API to lzf_compress_r and implement the dumb API in
terms of it. Consequently, bump liblzf major version number. This is
fine because nothing used our liblzf with the old API.

Add lzf(1) utility from LZF distribution. Add manual page for lzf(1).

branches: 1.176.2;
Extend the CPU microcode update framework to support Intel x86 CPUs.
Contrary to the AMD implementation, it doesn't use xcalls to distribute
the update to all CPUs but relies on cpuctl(8) to bind itself to the
right CPU -- to keep it simple and avoid possible problems with
hyperthreading.
Also, it doesn't parse the vendor supplied file to pick the right
part for the present CPU model but relies on userland to prepare
files with specific filenames. I'll commit a pkg for this in a minute
(pkgsrc/sysutils/intel-microcode).
The ioctl interface changed; compatibility is provided (should be
limited to COMPAT_NETBSD6 as soon as this is available).

Make GCC entries optional.

on x86, <machine/cpufunc.h> only pulls in <x86/cpufunc.h>. The latter
is not installed to userland and noone missed it, so the former ones
can not be useful either. Don't install them.

adjust for MKRUMP = no

new openssl

update lists after version bump of libc.

Update file sets for atf 0.16.

In particular, add entries for the new libatf-c and libatf-c++ shared
libraries. I have followed existing entries to come up with the right
lines to add... but this is quite tricky, so please forgive any fallout
that may follow (and let me know if there is any!).

mark idea and mdc2 libraries obsolete.

add crypto shared libraries

Fix the lists.

Update LLVM/Clang snapshot to r158533. This brings in support for
stack-realignment in functions using alloca on x86. Move to
/usr/include/clang-3.2 to reflect the version.

bump for bind-9.9.1-P1

libexecinfo sets

bump libssh

- add getdisk{raw,cooked}name
- bump libutil

bump libutil and new getfsspecname

adjust for expat-2.1.0

Implement sem_timedwait.

Move "libexpat" from X11 into base which fixes PR lib/40346.
This change was approved by the Core Group.

Merge XZ 5.0.3.

Add _dlauxinfo helper function to provide access to the ELF auxilary
vector. This can normally be found via __ps_strings, but libc is
initialised too early when linked dynamically and doesn't have access to
it yet, so provide an alternative mechanism via ld.elf_so.
Bump libc minor.

branches: 1.155.2;
Adjust debug objects for recent libc minor bump.

Set list updates for libquota bump.

Set list updates for libquota minor bump.

(the 32-bit lib and debug lib lists really ought to be automatically
generated)

Bump libsqlite3 minor version for new functions in 3.7.10.

From tnozaki@: make fpos_t a complex object that keeps track of the parse
state of the stream. Change argument of the seek function to funopen() from
fpos_t to off_t. Make f{g,s}etpos() use the new fpos_t struct, while providing
backwards compatible entry points. Approved by releng@

Setlist fix for MKDEBUGLIB=yes on amd64. Probably needs a similar
fix on sparc64.

Update LLVM/Clang snapshot to r147888. Most importantly, this fixes a
number of ABI issues for i386 and provides the correct
__FLT_EVAL_METHOD__ when using i387.

Bump minor version of libquota.so.

undo the pam bump. requested by joerg.

openpam changes

+libtre.so.0.8.debug

add tre to the sets.

branches: 1.143.2;
do not build profiling versions of libgcc or libgcc_eh.
they are not used, and they don't build on arm with gcc 4.5.

Hook up SQLite into the build

obsolete libgcc_pic and friends. all libgcc's are supposed to be pic
for platforms with shlibs already (and are.)

this obsoletes our hacks for the libgcc specs to use libgcc_pic, and
fixes a couple of other issues reported to me directly.

add new files for binutils 2.21.

- bump libbfd and libopcodes for new binutils.
- new elfedit program

build and install libgomp, as needed by "cc -fopenmp", and its
suplimentary files like libgomp.spec and omp.h.

bump libc

bump libisc too.

sets for bind.

Fix up amd64 set lists for MKDEBUGLIB=yes build. There are still
issues for other arches.

bump libssh

Add getlabelusesmbr(), as proposed in
http://mail-index.netbsd.org/tech-userlevel/2011/08/25/msg005404.html
This is used by disk tools such as disklabel(8) to dynamically decide is
the undelyling platform uses a disklabel-in-mbr-partition or not
(instead of using a compile-time list of ports).
getlabelusesmbr() reads the sysctl kern.labelusesmbr, takes its value from the
machdep #define LABELUSESMBR.
For evbmips, make LABELUSESMBR 1 if the platform uses pmon
as bootloader, and 0 (the previous value) otherwise.

bump libutil for strpct

Build and install ppath(3), property-list paths library.

Mark a few gcc-4.1 entries as obsolete with gcc-4.5.

Add scalbnl to libm.

bump libssh to 17.

build GMP, MPFR and MPC as private libraries just for GCC. don't
install the headers or librarys into the system.

someone who really cares could make them shlibs again and have them
installed into eg, /usr/lib/gcc. not convinced it is worth it.

Retire varargs.h support. Move machine/stdarg.h logic into MI
sys/stdarg.h and expect compiler to provide proper builtins, defaulting
to the GCC interface. lint still has a special fallback.
Reduce abuse of _BSD_VA_LIST_ by defining __va_list by default and
derive va_list as required by standards.

We no longer install <machine/bus.h> on amd64 and i386, so mark it
obsolete in the set lists.

bump libdns.so minor forgotten for amd64/sparc64 compat debug libraries.

- un-obsolete a pair of files we want to remain installed (XXX i wonder
if a way to have "clean builds" not run "postinstall fix obsolete" in
the $DESTDIR so that these sorts of errors will appear as set list
errors instead of being skipped
- fix the list of headers to install for GCC 4.5, based upon the mknative
${G_EXTRA_HEADERS}. XXX move the location of G_EXTRA_HEADERS.

bumping the major of libcrypto and libssl as suggested by tls@

various updates for GCC 4.5. finally obsolete some very old gcc stuff.

* Arrange for interfaces that create new file descriptors to be able to
set close-on-exec on creation (http://udrepper.livejournal.com/20407.html).

- Add F_DUPFD_CLOEXEC to fcntl(2).
- Add MSG_CMSG_CLOEXEC to recvmsg(2) for unix file descriptor passing.
- Add dup3(2) syscall with a flags argument for O_CLOEXEC, O_NONBLOCK.
- Add pipe2(2) syscall with a flags argument for O_CLOEXEC, O_NONBLOCK.
- Add flags SOCK_CLOEXEC, SOCK_NONBLOCK to the socket type parameter
for socket(2) and socketpair(2).
- Add new paccept(2) syscall that takes an additional sigset_t to alter
the sigmask temporarily and a flags argument to set SOCK_CLOEXEC,
SOCK_NONBLOCK.
- Add new mode character 'e' to fopen(3) and popen(3) to open pipes
and file descriptors for close on exec.
- Add new kqueue1(2) syscall with a new flags argument to open the
kqueue file descriptor with O_CLOEXEC, O_NONBLOCK.

* Fix the system calls that take socklen_t arguments to actually do so.

* Don't include userland header files (signal.h) from system header files
(rump_syscallargs.h).

* Bump libc version for the new syscalls.

Bump major for libmj.so debug libraries, forgotten in previous.

Update libcrypto.so and libssl.so minor forgotten in previous update,
for compat and debug libraries.

branches: 1.116.2;
fix a lot of different issues (doesn't anyone build anymore)
- libdns had the wrong version
- zfs attributes missing
- missing debug libs
- typos in library names
- typos in library versions
- _g libraries missing for new .so files (since we don't install .a files
we should not be installing _g.a files either, why are we?)
Hopefully this will fix the sparc64 build.

bump libmagic

Update LLVM and Clang snapshot

Attempt to fix build after perfuse addition (there is no MKPERFUSE)

Enable the build of perfused and libperfuse

Add missed MKDEBUG entry.

Fix MKDEBUG build for all 64bit architectures which have compat lib installed,
after kerberos import.

hi elric@!

various updates to deal with heimdal update.

Install clang's MD headers into /usr/include/clang-3.0 similar to GCC.

Upgrade Heimdal to 1.5pre1 by switching the build from crypto/dist/heimdal
to crypto/external/bsd/heimdal. The latter was just imported as the head
of the Heimdal tree as of a few days ago.

Move the headers for compiler and platform specific intrinsincs from
/usr/include to /usr/include/gcc-4.1.

Fix libquota_pic.a

Add a new libquota library, which contains some blocks to build and/or
parse quota plists; as well as a getfsquota() function to retrieve quotas
for a single id from a single filesystem (whatever filesystem this is:
a local quota-enabled fs or NFS). This is build on functions getufsquota()
(for local filesystems with UFS-like quotas) and getnfsquota();
which are also available to userland programs.
move functions from quota2_subr.c to libquota or libprop as appropriate,
and ajust in-tree quota tools.
move some declarations from kernel headers to either sys/quota.h or
quota/quota.h as appropriate. ufs/ufs/quota.h still installed because
it's needed by other installed ufs headers.
ufs/ufs/quota1.h still installed as a quick&dirty way to get a code
using the old quotactl() to compile (just include ufs/ufs/quota1.h instead of
ufs/ufs/quota.h - old code won't compile without this change and this is
on purpose).
Discussed on tech-kern@ and tech-net@ (long thread, but not much about
libquota itself ...)

Fix various vis/unvis issues:
- no need for all the weak symbols
- define a new _VIS_END flag for UNVIS_END so that there are no collisions
between and vis and unvis flags.
- add bound versions of the vis and unvis functions that take the length of
the destination buffer. Unlike the OpenBSD ones they return -1 or NULL if
the buffer is not large enough, instead of silently truncating.

bump for bind-9.8.0rc1

missing lvm keyword.

Add entries for libsaslc
Add some missings entries for amd64

Fix typo.

Add compat entries for libdevmapper, libdm, and liblzf

Fix build after libdm import sorry for breaking build.

remquo{,f} from FreeBSD via Stathis Kamperis

bump libc for arc4random_{buf,uniform}

NPF checkpoint:
- Add libnpf(3) - a library to control NPF (configuration, ruleset, etc).
- Add NPF support for ftp-proxy(8).
- Add rc.d script for NPF.
- Convert npfctl(8) to use libnpf(3) and thus make it less depressive.
Note: next clean-up step should be a parser, once dholland@ will finish it.
- Add more documentation.
- Various fixes.

Add missing .debug extension.

branches: 1.92.2;
missed librumphijack_pic.a and .ln in previous

hook librumphijack to the build

Hook libisns into the build, and modify the set lists accordingly.

Note the addition of iSNS protocol support in doc/CHANGES

With thanks to Wasabi Systems for contributing the code.

To use x86/cpu.h struct cpu_info from userland, we need via_padlock.h installed.

sets for tzcode enhancements

setlists updated for OpenLDAP 2.4.23

add obsolete entries for libamu to make MKDEBUGLIB=yes work on amd64 again

add missing liblua_g.a debuglib entries needed for MKDEBUGLIB=yes

rumpcrypto should never have been its own faction, so finally make
it a component under kern, i.e. rumpcrypto -> rumpkern_crypto.

sets for new libpcap and tcpdump

Remove the do-external-lib and do-gnu-lib targets, along with
external/lib/Makefile and crypto/external/lib/Makefile, replacing
them all with SUBDIRs directly from lib/Makefile.

compat/compatsubdirs.mk becomes simpler now, as everything is built
from lib/Makefile, meaning all the libraries will now be built under
compat so update the set lists to account for that.

fix sets for bind 9.7.2-P3

bump libssh to 16

setlists updated for OpenSSH 5.6

Add rumpclient to setlists.

You know, I used to think setlists were ridiculous back when ~20
entries per lib were required. It seems we're up to 47 and I've
long since run out of superlatives to describe the situation.

Hook up xz into the build system.

Add lua debug files.

Fix file lists for amd64 and sparc64 since on these arches 32bit versions of
the libraries are built as well.

libc minor was bumped.

merge conflicts

this isn't how this was planned, but needed for releases to build. will
be revisited later

Add support for fenv.h interface for i386 and amd64.

Submitted by Stathis Kamperis as part of GSoC 2010 and ported from
FreeBSD.

ldap is configured to depend upon MKCRYPTO no, so force MKLAP=no if
MKCRYPTO=no.
don't build pkg_install, libcrypto tests or rump_smbfs if MKCRYPTO=no.
mark librumpcrypto, rump_smb, pkg_*

fix set lists as appropriate.

Install x86/pte.h

mark some missing "libiscsi" files as "iscsi" so that MKISCSI=no works.
reported by director <director@ktinos.info> on port-amd64.

1. MB_LEN_MAX switch MD to MI.
2. unfortunately hppa's MB_LEN_MAX is defined incorrectly 6 instead of 32
so we have to add more setlocale(3) __RENAME func, __setlocale50.
3. move setlocale1.c and setlocale32.c to lib/libc/compat/locale/*
prepareing for next libc major crunk.
4. bump libc minor version.

Bump major of libpuffs, librefuse and libp2k.

Hmm, 60 lines changed (!!), 3 medium camels sacrificed and half
the total working time spent here. I picked the wrong day to not
quit editing setlists.

Add missing debuglib libraries to fix MKDEBUGLIB=yes builds.

Fix MKDEBUG=yes library builds for amd64

Userland now builds and uses terminfo instead of termcap.

OK: core@, jdc@

libmagic compat lib is installed again, remove obsolete marker.

Convert remaining base-obsolete to comp-obsolete.

Consistency: set for obsolete files should be "foo-obsolete".

Add missing entries for libiscsi compat library. Hi mrg@!

update mips64* sets for reality, and fix a couple of newly introduced
issues with sparc64 and amd64.

MKDEBUGLIB=yes fixes.

obsolete a bunch of files that should never have been installed, found
by the new src/compat framework that builds libraries more like they
are normally built.

Obsolete compat libbfd*.a and libbfd.so, following native behaviour.

Rename MKBFD to MKBINUTILS and support HAVE_BINUTILS for set lists. This
makes {MK,HAVE_}BINUTILS consistent with {MK,HAVE_}{GCC,GDB}.

Allow MKBFD to defines MKBINUTILS as a backwards compatibility hook.

Update the sets lists and add conditionals for lib{bfd,opcodes}.

Build & install rumpdev, a device component for rump. It currently
supports autoconf and is required for example by pseudo-device rump
components.

install librumpcrypto component: kernel crypto routines

+iasl.debug

/usr/lib/i386/libbfd.a is built and installed, do not mark it as obsolete.
"obsolete" added in 1.46 without a particular reason, file is not obsolete in
md.sparc64.

Remove premature addition of compat32 libs to arch-specific sets.

fix netpgp library issues.

remove libsdp as improved functionality has been enabled in libbluetooth.

Fix keywords: /usr/lib/i386/libiberty* need 'bfd', not 'debuglib'.
/usr/lib/i386/libvers* need 'kerberos', not 'debuglib'.

Apply the same keywords to the 32-bit compat files as apply to the
non-compat equivalent files, to stop amd64/sparc64 builds with MKxxx=no
options from failing because files in the DESTDIR are missing from the
METALOG.

amd64 builds, now, with several MKxxx=no options set. sparc64 probably
needs some more work.

FWIW, join(1) saved me some manual labor as I copied keywords from
non-compat to compat files:

1 Collect the files that are missing from the DESTDIR into
compat-i386.files

2 Replicate the first column of compat-i386.files. Remove i386/ from
the first column.

3 sort -k 1,1 distrib/sets/lists/comp/{md.amd64,mi,shl.*}
distrib/sets/lists/base/{md.amd64,mi,module.amd64,module.mi,shl.elf,sh
l.mi} | join -1 1 -2 1 -o '2.2 1.3' - compat-i386.files >
compat-i386.keywords

4 Process distrib/lists/{base,comp}/md.amd64 with join -1 1 -2 1 -o '1.1
1.2 1.3,2.2' -a1 - compat-i386.keywords

Flag ldscripts with 'bfd'. Sort by first column.

branches: 1.44.2;
Split vfs out of rumpkern into rumpvfs. Non-fs rumps no longer
include the file system code. File system rumps explicitly need
to include rumpvfs from now on.

Update compat library setlists for MKPUFFS removal.

Move libx86_64.so to the base set, for consistency (.so.0 and .so.0.0
are already there).

Add Makefiles and file lists for various optional crypto libraries.
Fixes PR lib/39846 and needs to be pulled up to netbsd-5.

fix MKPIC=no builds.

branches: 1.39.2;
Fix amd64 (and sparc64) build with MKPUFFS=no.

fix a bunch of MKPROFILE=no and MKSHARE=no errors introduced by myself
and a few others.

add the compat librefuse / libp2k entries.

move the compat shlibs into "base-compat-shlib", not "base-i18n-shlib".

add files built with MKCOMPAT != "no".

- install i386 includes on amd64
- sort

- install ldscripts on sparc64
- move ldscripts to /usr/libdata/ldscripts

Share cpu.h between the x86 ports.

branches: 1.32.2; 1.32.4;
install pmmintrin.h for SSE3

branches: 1.31.2; 1.31.4; 1.31.6;
Install xenio.h and xenio3.h in /usr/include/xen, for the benefit
of xentools3.

XXX ignore those in Makefile.ioctl-c, they don't compile properly outside
of the Xen context and the ioctls from xenio.h conflicts with
soundcard.h

add x86/pmap_pv.h

Add /usr/include/amd64/frame_regs.h

- Make __cpu_simple_lock and similar real functions and patch at runtime.
- Remove old x86 atomic ops.
- Drop text alignment back to 16 on i386 (really, this time).
- Minor cleanup.

Install the linker scripts on amd64, which allows to link x86 32bit objects
using the native ld(1). This helps building pkgsrc/sysutils/xentools3-hvm
on amd64; is also allows to build the x86 boot blocks with the
native toolchain.

sync with kernel after the merge of yamt-x86pmap branch.

branches: 1.25.4;
Add amd64 specific debug files.
Fix build.sh release for amd64 with MKDEBUG=yes

+ x86/sysarch.h

Sync with changes introduced by merging the newlock2 branch.

add ./usr/include/x86/cputypes.h

Move the single remaining line from md.x86_64 to md.amd64, and put
md.x86_64 back to deleted state. It is a long time since the latter
file was inspected when putting together set lists for amd64.

Add linux32_machdep.h .

Revert yesterday's change that attempted to fix the detection of the
boot device when using a Multiboot boot loader. It couldn't work because
these boot loaders do not pass a checksum of the disk so matchbiosdisk()
cannot really find any matches. I should have gone to sleep before
commiting...

Found by xtraeme@.

When booting an i386 kernel with Multiboot, properly detect the boot device
by looking it up in the x86_alldisks table (instead of trying to match it
to 'wd*' manually).

In order to do this, move the cpu_rootconf function from x86 common code
to amd64 and i386 specific one. This way, i386 can do an extra step (call
the appropriate Multiboot code) in the appropriate place (after
x86_matchbiosdisks and before findroot()).

Exclude additional files from the set lists if MKGCCCMDS=no.

Move the lib<foo>.so entry (link) from comp to base, so that
applications on systems without the `comp' set installed can
dlopen() a library without supplying the major number.

add a <machine/joystick.h> which just includes the new common one

Add <machine/wchar_limits.h>.

Add emmintrin.h.

branches: 1.12.2;
MKLINT=no support

reformat whitespace

Support MKPROFILE=no

Significant overhaul of how the sets are used:

* Add an optional third field to the sets file which is a list of
comma separated keywords that control if the line is printed.
Currently supported keywords
kerberos4 ${MKKERBEROS4} != no
kerberos ${MKKERBEROS} != no
lint ${MKLINT} != no
obsolete ${obsolete} != 0.
In this case, non obsolete files are not printed.
(This will allow future support for builds with variables such as
MKHESIOD and MKYP set to "no".)

* Use sh(1)'s getopts where appropriate, and otherwise cleanup the
various scripts.

* Move defaults for sets.subr from sets.defaults into sets.subr.
Move replicated code for determining stuff such as shlibs type
from various scripts into sets.subr.

* Merge the obsolete.*, krb.*, krb4.* and lint.* into the appropriate
main lists with the relevant third field keyword(s).

Add MMX/SSE (i386,amd64) and Altivec (powerpc) include files.

Add <x86/math.h>.

update for amd64 build.

XXX: still missing ./usr/bin/gdb & friends, which are in comp/mi, etc.

Fix syspkg names on some libraries (these are under comp-, not base-).

Add atomic.h

Add some obsolete files (after x86_64 -> amd64).

x86_64 -> amd64

x86_64 -> amd64

don't mark the gcc-10 + sanitizer directories obsolete yet.

they're still created by the build, and they're owned by the base set.
i'll clean the up when i delete gcc-10 entirely.

distrib/lists/sets/comp: Add gcc=12 obsolete entries for gcc-10/*.h.

mark all previously gcc=9 files as obsolete.

also mark a bunch of obsolete gcc headers as "comp-obsolete" "set".

should fix a 'postinstall fix obsolete' issue reported by uwe@.

more updates for sets and GCC 12

- renable sanitizer, add new shlib versions, enable tsan for arm64,
including moving some stuff into the per-cpu/arch dir from 'mi'.
- fix mips64, 68k, and ia64
- mark many any-gcc-version things just as "gcc" and remove dups
- mark many gcc=9 only things as obsolete
- mark many gcc=10 only things as obsolete with gcc=12 (more could
be done here.)

various updates for GCC 12

- bump some shlibs
- add new header files
- add compiler-version header files
- disable sanitizer libraries for gcc12 for now

Unbreak creation of i386 PIEs on amd64.

Functional patch by RVP; set lists updated accordingly.

This fixes PR toolchain/57380.

Add the .xsce scripts

add some files that are obsolete with binutils 2.39

sort

Turn on -z separate-code for x86, like linux has done. This splits the text
and data segment load sections into two pieces, for at total of 4 PT_LOAD
sections: two for text (r--, r-x) and two for data (r--, rw-). For x86_64
this changes the default max page size from 2M to 4K.

annotate the 234 linker scripts

distrib/sets: remove more extra keywords from obsolete entries

Found by fmt-list.

Move tsan entries marked machine=amd64 from .mi files into md.amd64

Move the sanitizer libraries from 9.0.0 -> 13.0.0 to match our current clang
version.

Update LLVM build system for 249b40b558955afe5ac2b549edcf2d7f859c8cc9

This enables the use of modules for a significant build performance gain
when building with clang as host compiler or when using HAVE_LLVM=yes.
Switch libc++ to using the copy from the mono-repo.

branches: 1.286.2;
Add <x86/gdt.h> (we could gc now the <machine/gdt.h> files if we fix the
sources that use them.

various updates for GCC 10 support. remove dead gcc7/8 support
and properly mark obsolete files in gcc9/gcc10.

testing on amd64 and shark for GCC 9, and many platforms for GCC 10,
(though not arm yet.)

Add missing lint library

Add HAVE_NVMM and use it to control the build of NVMM related components.
Defined to "yes" on amd64, "no" everywhere else.

mostly support GCC 9 builds. some what tested with GCC 9 but
tested to not break GCC 8 builds for x86, arm*, vax, and sparc*.

Add back <xen/xenio3.h>, it's needed by xentools413

Sort and deduplicate set lists.

<xen/xenio.h> is in fact used by the newly imported xentools 4.13,
reinstantiate it

header is for use by xentools only, which define domid_t themselves,
so it should not be a problem the header doesn't define it per PR port-xen/52874

don't install xen/xenio.h and xen/xenio3.h anymore, xentools don't use it

PR port-xen/52874

Add /usr/include/xen/intrdefs.h to sets
Add xen to KLINKFILES for amd64/i386

Install LLVM sanitizers to /usr/lib/clang/9.0.0

This is the proper location where clang 9.0.0 looks for them.

Update LLVM to 10.0.0git (01f3a59fb3e2542fce74c768718f594d0debd0da)

gcc 8 set lists updates. i don't appear to have broken gcc 7 builds
but i only tested a couple of them.

Add new multiboot2.h header.

Correct the build rules of LLVM sanitizers

Fix generation of .a.syms files in LLVM sanitizers. The old rules weren't
functional and sometimes they were fatal in the build.

Add the .a.syms files to distribution sets.
These files are generated only in certain .a files in 64bit version of
sanitizers.

Correct the location of files with LLVM sanitizers. Move the LLVM sanitizer
entries from base to comp files.

branches: 1.271.2;
Revert MKLLVMCMDS changes.

split MKLLVM into MKLLVM and MKLLVMCMDS, like MKGCC and MKGCCCMDS:

- only build clang or clang-tblgen if MKLLVMCMDS != "no"
- move HAVE_MESA_VER earlier so we can detect mesa 18 and enable
MKLLVM on platforms that need it
- move the clang commands and headers into llvmcmds group
- only build the clang specific LLVM libraries if MKLLVMCMDS != "no"

much of this work came from maya@'s github commit
f90685c11d4460d3098fa35f48b58d1893e974e0, reworked for HAVE_MESA_VER
and llvmcmds vs MKCLANG.


build tested on several platforms and on amd64 and i386 with several
different build options (clang build, gcc build, gcc build with llvm
libs).

get rid of binutils=227-specific files.

Sort, and add a couple obsolete files for binutils=231.

- add new entries of gcc=7 and libgomp, libasan, liblsan, libstdc++, libubsan
- add missing header files fpr gcc=7
- add some missing gcc=5 obsolete entries
- create ./usr/include/dev/bluetooth in mtree not make, and move it into the
base set where all other directories are.

add basic gcc-7 support in sets lists and filesystem. mips gained
a header. assumes it's otherwise mostly the same as gcc-7.

- mark gcc=5 files obsolete
- make includes always available, not only when gcccmds
- mark files obsolete consistently

new binutils is 231

add nvmm lint library

Fix the libnvmm sets, do the same as libx86_64.

Add libnvmm, NetBSD's new virtualization API. It provides a way for VMM
software to effortlessly create and manage virtual machines via NVMM.

It is mostly complete, only nvmm_assist_mem needs to be filled -- I have
a draft for that, but it needs some more care. This Mem Assist should
not be needed when emulating a system in x2apic mode, so theoretically
the current form of libnvmm is sufficient to emulate a whole class of
systems.

Generally speaking, there are so many modes in x86 that it is difficult
to handle each corner case without introducing a ton of checks that just
slow down the common-case execution. Currently we check a limited number
of things; we may add more checks in the future if they turn out to be
needed, but that's rather low priority.

Libnvmm is compiled and installed only on amd64. A man page (reviewed by
wiz@) is provided.

Add NVMM - for NetBSD Virtual Machine Monitor -, a kernel driver that
provides support for hardware-accelerated virtualization on NetBSD.

It is made of an MI frontend, to which MD backends can be plugged. One
MD backend is implemented, x86-SVM, for x86 AMD CPUs.

We install

/usr/include/dev/nvmm/nvmm.h
/usr/include/dev/nvmm/nvmm_ioctl.h
/usr/include/dev/nvmm/{arch}/nvmm_{arch}.h

And the kernel module. For now, the only architecture where we do that
is amd64 (arch=x86).

NVMM is not enabled by default in amd64-GENERIC, but is instead easily
modloadable.

Sent to tech-kern@ a month ago. Validated with kASan, and optimized
with tprof.

Update LLVM/Clang build system to r337282.

Remove the kernel PMC code. Sent yesterday on tech-kern@.

This change:

* Removes "options PERFCTRS", the associated includes, and the associated
ifdefs. In doing so, it removes several XXXSMPs in the MI code, which is
good.

* Removes the PMC code of ARM XSCALE.

* Removes all the pmc.h files. They were all empty, except for ARM XSCALE.

* Reorders the x86 PMC code not to rely on the legacy pmc.h file. The
definitions are put in sysarch.h.

* Removes the kern/sys_pmc.c file, and along with it, the sys_pmc_control
and sys_pmc_get_info syscalls. They are marked as OBSOL in kern,
netbsd32 and rump.

* Removes the pmc_evid_t and pmc_ctr_t types.

* Removes all the associated man pages. The sets are marked as obsolete.

branches: 1.257.2; 1.257.4;
obsolete some GCC 4.8 files.

it's called avx512bwintrin.h.

more updates for GCC 6. with other uncommited changes, i386 seems
to work well enough for hello.c, hello.cc and hello.m.

updates for GCC 6.4: add gcc-6 specific headers, combine some
all-gcc files, bump the shlib versions on new libs.

convert HAVE_GCC handling to modern GCC release numbering:
- HAVE_GCC=5 is now the default (vs. HAVE_GCC=53 we've been using for
GCC 5.4 and GCC 5.5.)
- remove some more GCC 4.8 code. we don't support GCC 4 here.
- adjust set lists to gcc=5 from gcc=53.

add some basic HAVE_GCC=6 handling (totally unused so far.)

pcommitintrin.h has been removed from gcc 5.5

Remove vm86.

Pass 4.

Should be comp-obsolete.

Remove ibcs2_machdep.h on i386, and don't install it.

Move freebsd_machdep.h into sys/compat/freebsd, and don't install it. Now,
the compat_freebsd files are all contained in sys/compat/freebsd.

Update LLVM build glue for 5.0RC1.

Restore marking ./usr/include/i386/svr4_machdep.h obsolete
instead of removing

(so it can be removed by postinstall)

Remove svr4_machdep.h right away, no one should include it.

If i386/svr4_machdep.h is no longer to be installed in /usr/include
it should no longer be listed in the sets lists as installed.
Mark it obsolete (hopefully unbreak builds.)

branches: 1.243.4;
Introduce PT_GETDBREGS and PT_SETDBREGS in ptrace(2) on i386 and amd64

This interface is modeled after FreeBSD API with the usage.

This replaced previous watchpoint API. The previous one was introduced
recently in NetBSD-current and remove its spurs without any
backward-compatibility.

Design choices for Debug Register accessors:
- exec() (TRAP_EXEC event) must remove debug registers from LWP
- debug registers are only per-LWP, not per-process globally
- debug registers must not be inherited after (v)forking a process
- debug registers must not be inherited after forking a thread
- a debugger is responsible to set global watchpoints/breakpoints with the
debug registers, to achieve this PTRACE_LWP_CREATE/PTRACE_LWP_EXIT event
monitoring function is designed to be used
- debug register traps must generate SIGTRAP with si_code TRAP_DBREG
- debugger is responsible to retrieve debug register state to distinguish
the exact debug register trap (DR6 is Status Register on x86)
- kernel must not remove debug register traps after triggering a trap event
a debugger is responsible to detach this trap with appropriate PT_SETDBREGS
call (DR7 is Control Register on x86)
- debug registers must not be exposed in mcontext
- userland must not be allowed to set a trap on the kernel

Implementation notes on i386 and amd64:
- the initial state of debug register is retrieved on boot and this value is
stored in a local copy (initdbregs), this value is used to initialize dbreg
context after PT_GETDBREGS
- struct dbregs is stored in pcb as a pointer and by default not initialized
- reserved registers (DR4-DR5, DR9-DR15) are ignored

Further ideas:
- restrict this interface with securelevel

Tested on real hardware i386 (Intel Pentium IV) and amd64 (Intel i7).

This commit enables 390 debug register ATF tests in kernel/arch/x86.
All tests are passing.

This commit does not cover netbsd32 compat code. Currently other interface
PT_GET_SIGINFO/PT_SET_SIGINFO is required in netbsd32 compat code in order to
validate reliably PT_GETDBREGS/PT_SETDBREGS.

This implementation does not cover FreeBSD specific defines in their
<x86/reg.h>: DBREG_DR7_LOCAL_ENABLE, DBREG_DR7_GLOBAL_ENABLE, DBREG_DR7_LEN_1
etc. These values tend to be reinvented by each tracer on its own. GNU
Debugger (GDB) works with NetBSD debug registers after adding this patch:

--- gdb/amd64bsd-nat.c.orig 2016-02-10 03:19:39.000000000 +0000
+++ gdb/amd64bsd-nat.c
@@ -167,6 +167,10 @@ amd64bsd_target (void)

#ifdef HAVE_PT_GETDBREGS

+#ifndef DBREG_DRX
+#define DBREG_DRX(d,x) ((d)->dr[(x)])
+#endif
+
static unsigned long
amd64bsd_dr_get (ptid_t ptid, int regnum)
{


Another reason to stop introducing unpopular defines covering machine
specific register macros is that these value varies across generations of
the same CPU family.

GDB demo:
(gdb) c
Continuing.

Watchpoint 2: traceme

Old value = 0
New value = 16
main (argc=1, argv=0x7f7fff79fe30) at test.c:8
8 printf("traceme=%d\n", traceme);

(Currently the GDB interface is not reliable due to NetBSD support bugs)

Sponsored by <The NetBSD Foundation>

branches: 1.242.2;
Update LLVM build glue for pre-4.0.0 r291444.

Tag cleanup: if the last column contains 'obsolete', the one before should too.

Add support for hardware assisted watchpoints/breakpoints API in ptrace(2)

Add new ptrace(2) calls:
- PT_COUNT_WATCHPOINTS - count the number of available hardware watchpoints
- PT_READ_WATCHPOINT - read struct ptrace_watchpoint from the kernel state
- PT_WRITE_WATCHPOINT - write new struct ptrace_watchpoint state, this
includes enabling and disabling watchpoints

The ptrace_watchpoint structure contains MI and MD parts:

typedef struct ptrace_watchpoint {
int pw_index; /* HW Watchpoint ID (count from 0) */
lwpid_t pw_lwpid; /* LWP described */
struct mdpw pw_md; /* MD fields */
} ptrace_watchpoint_t;

For example amd64 defines MD as follows:
struct mdpw {
void *md_address;
int md_condition;
int md_length;
};

These calls are protected with the __HAVE_PTRACE_WATCHPOINTS guard.

Tested on amd64, initial support added for i386 and XEN.

Sponsored by <The NetBSD Foundation>

Sort the unsorted sets lists (they should all be correctly sorted now.)

everyone is now on binutils-2.26 or more

branches: 1.237.2;
GCC 5.3 sets for amd64.

Update build system for LLVM 3.8.0rc3.

add missing cpu_rng.h includes

handle both binutils 2.23 and 2.26

now that we only support gcc 4.8, make all the gcc tags just "gcc".
this should help with next-gcc, when we can add back gcc=48 and
gcc=5x tags, as necessary.

obsolete and clean up gcc=3 (3.3), gcc=4 (4.1), and gcc=45 (4.5) sets.

Update some of the set lists to use the compatdir/compatfile
compattestsdir/compattestsfile as well the kmod auto-entry creation.
Remove entries that are now instantiated by the above.

Update set lists for LLVM 3.6RC1.

stuff added by blacklist

Install cpuid.h on X86 and arm_neon.h on ARM.

Consistently tag libasan with gcc=48

Add libasan

sort

branches: 1.224.2;
Update build glue for LLVM/Clang 3.6svn r215315. Changes include various
performance improvements across all backends and a much improved PPC32
backend.

Rename sljitarch.h to sljit_machdep.h.

infrastructure for bind-9.10.0-P2:
adding the tools delv and dnssec-importkey
adding libirs to support delv

Drop entries for LLVM/Clang older than 3.4.

Update build system for LLVM/Clang snapshot r209886. This brings in the
merged AArch64 backend and the move of eARM to DWARF exception handling
for NetBSD. C++11 is required for building.

sort

/usr/include/g++/bits/${CONFIG} directories are created independent of
MKGCC, so fix file lists.

branches: 1.217.2;
Fix for /usr/include/g++/bits directory, which is only for MKGCC=yes

Unobsolete fpu.h

Mark i386/npx.h obsolete.
I'd forgotten it was released here as well.

Move sys/arch/amd64/amd64/fpu.c and sys/arch/amd64/include/fpu.h
into sys/arch/x86 in preparation for using the same code for i386.

Add cpu_extended_state.h

Update LLVM family to 3.5svn r198450. Interesting changes include:
- Improvements in the SPARC64 support
- ARM backend can almost build a full NetBSD/EABI system with the
integrated assembler
- PowerPC grows disassembler support
- LLD supports copy relocation, beginning of MIPS backend
- New diagnostic for use of bool expression as size_t argument for
strncmp and friends

Drop superflous "picinstall" flag on "obsolete" items

Reported by apb@, thanks!

Mark all _pic.a archive files with the 'picinstall' flag.

This fixes checkflist with MKPICINSTALL=no.

ok christos@.

Make /usr/include/g++/bits/i386 and x86_64 not depend on gcccmds.

Directories are created even when gcc=48 is not true, so remove
that flag as well.

Fix the build for amd64.

Even though the files are installed only if "compat" is set, the
/usr/include/g++/bits directories are unconditionally created via
the mtree files.

update set lists for gcc 4.8.

_g libraries don't belong here.

more gcc-4.8 files

add support for GCC 4.8 sets.

sort

Update LLVM/Clang snapshot to r192551 for various bug fixes.

Fix amd64 build sets with new architectures.

Update LLVM/Clang snapshot to r191105. This brings in a better detection
of unused static variables and functions. Disable this for some external
code and for ioconf.c in the kernel.

Sort entries in strcmp order.

Update LLVM/Clang snapshot to r182734. Brings initial support for
integrated PowerPC assembler.

Introduce MKLIBSTDCXX, defaulting to yes.

Add build glue for libc++.

Update LLVM/Clang snapshot to r180004. This brings in the last pieces of
C++11 support in Clang as well as initial SPARC64 support.

more libevent set fixes

libevent sets

Hook Lutok into the build. This is all protected by the MKKYUA guard.

- fix the debug sets which were broken because of _LIB_PREFIX lossage.
- add missing tpm files.

Update LLVM/Clang snapshot to r172366. Catch up with the move past 3.2
release.

Add a new "debug" set that gets built when ${MKDEBUG} is set. On evbarm:
-rw-r--r-- 1 root src 35806068 Jan 16 10:47 base.tgz
-rw-r--r-- 1 root src 57374069 Jan 16 10:48 comp.tgz
-rw-r--r-- 1 root src 101286193 Jan 16 10:49 debug.tgz
...
This allows always building the debugging code and libraries but not
installing it (although now it is placed in the default sets if ${MKDEBUG}
is set

bump libmagic

add entries built for MKDEBUG and MKDEBUGLIB to the set lists
Thanks to Hisashi T Fujinaka <htodd@twofifty.com> for pointing out the gap.

bump libssh

Adjust libraries debug entries following recent bind update.

Fix set lists after the merge of netpgpverify(1) and libnetpgpverify(3)
from the agc-netpgp-standalone branch.

Update LLVM/Clang snapshot to r168187 for various bug fixes.

bump resolver library.

Include sljitarch.h even when MKSLJIT=no.

Build sljit test when MKSLJIT != no and set MKSLJIT to yes on amd64 and i386.

Update LLVM/Clang snapshot to r166772. Now builds without RTTI and
exceptions. Also features the first round of a loop vectorizer.

*Sigh* provide, by default, the dumb "allocate 64K on the stack" variant
of the lzf_compress API by default. This is because there is evidently
code floating around out there in the world with broken autoconf scripts that
don't check for LZF_STATE_ARG in the lzf header files and just assume the
dumb version of the API.

We rename the sane API to lzf_compress_r and implement the dumb API in
terms of it. Consequently, bump liblzf major version number. This is
fine because nothing used our liblzf with the old API.

Add lzf(1) utility from LZF distribution. Add manual page for lzf(1).

branches: 1.176.2;
Extend the CPU microcode update framework to support Intel x86 CPUs.
Contrary to the AMD implementation, it doesn't use xcalls to distribute
the update to all CPUs but relies on cpuctl(8) to bind itself to the
right CPU -- to keep it simple and avoid possible problems with
hyperthreading.
Also, it doesn't parse the vendor supplied file to pick the right
part for the present CPU model but relies on userland to prepare
files with specific filenames. I'll commit a pkg for this in a minute
(pkgsrc/sysutils/intel-microcode).
The ioctl interface changed; compatibility is provided (should be
limited to COMPAT_NETBSD6 as soon as this is available).

Make GCC entries optional.

on x86, <machine/cpufunc.h> only pulls in <x86/cpufunc.h>. The latter
is not installed to userland and noone missed it, so the former ones
can not be useful either. Don't install them.

adjust for MKRUMP = no

new openssl

update lists after version bump of libc.

Update file sets for atf 0.16.

In particular, add entries for the new libatf-c and libatf-c++ shared
libraries. I have followed existing entries to come up with the right
lines to add... but this is quite tricky, so please forgive any fallout
that may follow (and let me know if there is any!).

mark idea and mdc2 libraries obsolete.

add crypto shared libraries

Fix the lists.

Update LLVM/Clang snapshot to r158533. This brings in support for
stack-realignment in functions using alloca on x86. Move to
/usr/include/clang-3.2 to reflect the version.

bump for bind-9.9.1-P1

libexecinfo sets

bump libssh

- add getdisk{raw,cooked}name
- bump libutil