Kill empty .TP

(joerg)

kdtrace_{pro,thread}c_ctor: use kmem_zalloc(), instead of manual memset().

(rmind)

proc_alloc: move kdtrace_proc_ctor() out of the proc_lock.

(rmind)

New sentence, new line. Bump date for previous.

(wiz)

Make sure that .TH actually gets a section number.

(joerg)

Drop trailing non-sense arguments for .TH. Too many of those have been
provided anyway.

(joerg)

fix some cosmetics.

(chs)

Add definitions for TOOL_MANDOC_* for standalone users. Requested by
dyoung.

(joerg)

Collect garbage: sc_rbus_{iot,memt} are not used.

(dyoung)

print human readable exit code.

(christos)

Add machine class manual paths. No manual pages have moved yet.

(christos)

Regen for pathconf/fpathconf rumpification.

(njoly)

Rumpify pathconf(2)/fpathconf(2)

ok from pooka@

(njoly)

atf 0.9 includes a backwards incompatible change that requires rebuilding
the sh tests.  Add a note to tell users to clean up the old files, as
otherwise the rebuilds will not be triggered.

(jmmv)

Added support for spdupdate command in setkey

(vanhu)

A suggestion from veego to fix it (taken from ../../include/fontsproto).

(ghen)

Misc KNF.

(skrll)

Import dhcpcd-5.2.4

(roy)

Import dhcpcd-5.2.4 with the following changes:
* Fix crash when using clientid and the interface re-configures
* log the pid of dhcpcd
* Indicate server IP received message from even if server ID not   
  present
* Fix crashes on IPv4LL failure and add more logging

Status:

Vendor Tag: roy
Release Tags: dhcpcd-5-2-4

(roy)

Note import of atf 0.9.

(jmmv)

atf-0.9 introduces an use.fs test-case property to allow tests to write to
their work directory.  The purpose is to be able to know which tests intend
to touch the file system and to allow a minor optimization in atf-run.

Define use.fs=true for all those tests requiring it.  (This highlights that
some tests currently require modifying the file system but conceptually
they shouldn't be... which leaves room for further improvements/cleanups
later :-)

(jmmv)

Adjust file lists for atf-0.9.  Most notably, some files are gone for good!

(jmmv)

Remove directories not required by atf-0.9 any more.

(jmmv)

Simplify build of shell-based test cases: atf-compile is gone.

(jmmv)

Bye bye atf-compile, you won't be missed.

(jmmv)

Adjust reachover makefiles for atf 0.9.

(jmmv)

Fix import conflicts for atf 0.9.

(jmmv)

Import atf 0.9:

* Added atf-sh, an interpreter to process test programs written using
  the shell API. This is not really a shell interpreter by itself
  though: it is just a wrapper around the system shell that eases the
  loading of the necessary ATF libraries.

* Removed atf-compile in favour of atf-sh.

* Added the use.fs metadata property to test case, which is used to
  specify which test cases require file system access. This is to
  highlight dependencies on external resources more clearly and to speed
  up the execution of test suites by skipping the creation of many
  unnecessary work directories.

* Fixed test programs to get a sane default value for their source
  directory. This means that it should not be necessary any more to pass
  -s when running test programs that do not live in the current
  directory.

* Defining test case headers became optional. This is trivial to achieve
  in shell-based tests but a bit ugly in C and C++. In C, use the new
  ATF_TC_WITHOUT_HEAD macro to define the test case, and in C++ use
  ATF_TEST_CASE_WITHOUT_HEAD.

Status:

Vendor Tag: TNF
Release Tags: atf-0-9

(jmmv)

Remove unncessary cast.

(skrll)

Change struct cpu_info layout to improve cache footprint.  From OpenBSD.

(skrll)

Fix sume bugs.
  - Why +1 ?
  - mvidmac_dump_xoredesc() -> gtidmac_dump_xoredesc().
  - bitmask_snprintf() -> snprintb().
Support big endian machines with GTIDMAC_CCHR_ENDIAN_BE.

(kiyohara)

make sure we pass a valid consdev pointer to the console keyboard so kbd
knows it's console and passes the right flags to wskbd, now the wskbd will
not drop off the mux when leaving event mode.
Symptoms: now the keyboard should no longer be dead when leaving X and there
          should be no more panics caused by it
I tested this on an SS5 and an SS20, may need some more attention and zs.c
sure could use some cleanup - there are a few differences to sparc64 that
really don't make much sense to me, like not passing console flags to zstty
which sparc64 does but sparc avoids, which caused this problem in the first
place.

(macallan)

Fix .Xr; dirent(5) -> dirent(3).

(jruoho)

backout the xfree86 4.5 change to GetTimeInMillis() that returns
relative times instead of absolute times.  it causes problems in
openmotif at the very least, and is not what xorg has done.

from Cliff Wright in PR#39209.

(mrg)

Add a comment and check off in mmmmap.

(skrll)

Revert previous.

(ghen)

Don't use rumpuser_malloc() directly.

(pooka)

apply build framework and config changes for pcc-100602, plus add a
prepare-import.sh script for future updates

(plunky)

merge changes from pcc-090902 branch

(plunky)

Update pcc to current [0.9.9 20100603] version. There have been many
bugfixes and feature additions, full changelog is below. The major
feature I think was the GCC compatibility which is now enabled by
default, with many of the commonly used GCC attributes and builtin
functions now supported, and pcc defines __GNUC__ as version 4.3

This version can build all of bin/, sbin/ and games/ except for dab(5),
with only a few outstanding issues in usr.bin/ and usr.sbin/. There are
still issues with PIC code generation (affects building libraries, on
i386 at least) and I have not tried applying pcc to kernel builds.

ChangeLog, in most-recent first order, screen-scraped from
    http://pcc.ludd.ltu.se/fisheye/changelog/pcc

Each paragraph is a separate commit, the [xxx] refers to the module
that the following commits were in:

[ccom]
    Fix designated initializer bug as reported by Iain Hibbert on pcc-list.

    Put in prototype for dainit() to avoid warnings when using yacc.

    Discard parameter attributes, they ended up to be handled as arguments
    otherwise.  Suppose to fix Jira#PCC-132 by Iain Hibbert.

    Rewrite __builtin_offsetof() to handle non-constant array offset values.
    This actually made the offsetof code much simpler.
    Requested by Iain Hibbert on pcc-list.

    Op DOT will be BINODE.

    Support gcc ugly construction a?:b, as discussed on pcc-list.

    Add gcc compat range initializers, reported by Iain Hibbert.

[amd64]
    Do not put eax as hidden argument if not needed.

    Better support for the Intel addressing modes.  Now ByteBench only takes
    77% of the earlier time to complete.

[mip]
    Do not base hash functions on memory addresses.  These addresses may
    change between different compilations and we want predictability.
    Also check if an address of a node is lower than expected to avoid
    short-range temporaries in the Freeze() code.

[i386]
    Clear flags field on TEMP nodes so that function pointers will get their
    stack adjusted afterwards, otherwise there might be junk left.

[ccom]
    Aet alignment to at least ALCHAR for structs when passing them over to
    pass2.  Failure to do so caused bug spotted by Iain Hibbert.

[sparc64]
    Fix string init bug, from Lars Zetterlund.

[mip]
    Declare printip2() as static function for whole file, as pointed out by
    Iain Hibbert.

[ccom]
    Make the canary value on stack volatile, so that the compiler won't
    try to put it into a register.  Solves stack overwrite error reported
    by Iain Hibbert on pcc-list.

[i386]
    Fix uninitialized member bug in picsymtab breaking -k -g.  Suppose to fix
    Jira#PCC-138 by Nicolas Joly, patch provided by Iain Hibbert.

[ccom]
    Tell scan to ignore typedefs of the same names goto labels. Fixes
    Jira#PCC-141 by Andreas Kaiser with recommended fix from him.

[mip]
    Run deljumps again after the ssa pass to cleanup the code before doing
    register allocation.

[mip]
    Do liveness analysis before turning a function into SSA code, and later
    use that to to find out whether we need to insert a phi function or not.
    This is usually called "Pruned SSA".
    Now the SSA code tend to be significantly better quality.

    Fix deljumps label updating bug.
    Remove statements that are moves to itself in deljumps.

[libpcc]
    Define _VA_LIST_DEFINED for compatibility eglibc headers.
    From tg@mirbsd.de.

[ccom]
    Ensure CM towers are built to the left, fixes parsing bug pointed put
    by Jonathan Gray. Also add "used" attribute.

[mip]
    Teach dce() to remove ip structs without destroying the basic block
    structure.

[ccom]
    Check if type is an array before checking its dimension.  Fixes bug
    reported by Iain Hibbert on pcc-list.

[i386]
    Tell table to prefer 2-op insns before 3-op.

    Do not try to match args on builtins.

[mip]
    Do not favorize 3-op instruction in the ssa case, simple statements
    like i++ have been rewritten to t1 = t2 + 1 and we want the register
    allocator to try to put t1 and t2 in the same register if there are
    2-op insns.

    Help coloring algorithm by at the time of assigning registers see if
    there are any frozen moves to nodes that have already got their color
    and we can give this node the same register.
    This will avoid unneccessary moves.

    Move list init into the function that initializes it.

    Put two more variables into p2env, to have a cleaner way of passing
    information in the optimizer phase.

    Fix broken indentation.

    Do not put a move on the move list more than once.
    Count precolored nodes as infinite K, not 0, otherwise it may make
    colorable graphs uncolorable when there are many moves.

[ccom]
    Fix reporting on bogus error of struct parameters, as reported by Iain
    Hibbert on pcc-list.

[amd64]
    Remove cast of pointer converted ints, not needed on amd64.  Fixes
    Jira#PCC-136 by Nicholas Joly.

[ccom]
    Add and handle gcc attribute __warn_unused_result__.  Requested by Iain
    Hibbert but with a different implementation.

    Do not allow initialization of multi-dimension incomplete arrays, fixes
    bug reported by Jesus Sanchez on pcc-list.
    Also fix another bug spotted, which had as a result that dynamic-sized
    arrays initialized by strings did not become null-terminated.

[amd64]
    amd64 cannot handle imemdiate constants for 64-bit simple ops.
    Fixes Jira#PCC-134 by Nicolas Joly.

[ccom]
    Allow any expression inside typeof.  Supposes to fix Jira#PCC-135
    by Iain Hibbert.

[mip]
    Replace deljumps() with the stuff used in the 2BSD c2.
    Still some parts to implement though.

[i386]
    Yech, yesterdays worst thinko about where to put the return check for
    floats.  Now awk works again.

    Use SSTDCALL, not FSTDCALL for symtab entries.  Spotted by Iain Hibbert.

    Reorganize late mangle code a little.  Now stdcall should work.

[mip]
    Remove some #if 0'ed code that is phased out.

[i386]
    The n_flags fields should use NLOCALx defines, not hardcoded numbers.
    Found by Iain Hibbert.

[ccom]
    Check for NaN string is only valid for ICON nodes.  Bug found by
    Iain Hibbert.

[i386]
    Detect the situation where a function returns a float number but the
    return value is ignored; the float stack must be popped anyway.
    FIxes a bug in awk where the float stack overflows.

[ccom]
    Clear out symtab pointers for NAME and ICON nodes in functions to
    be inlined, since accessing them will result in crashes.  Fixes
    stdcall/ccall bugs reported by Iain Hibbert.

[ccom, i386]
    Add support for stdcall and cdecl attributes.  Patch from iain Hibbert.

[ccom]
    Remove redundant GETSUE() define, from Iain Hibbert.

    Fix yet another situation where attributes causes warnings.
    Reported and slightly modified patch from Iain Hibbert.

    Handling of argument to scalinit() was incorrect causing wrong type to
    be returned.  Fixes Jira#PCC-131 by Iain Hibbert.

    Fix problem with offsetof of zero-length arrays and a debug segfault.
    Fixes Jira#PCC-129 by Iain Hibbert.

    Fix checking of struct sizes.  Solves Jira#PCC-128 by Iain Hibbert.

    Remove redundant check that caused function names beginning with L to
    be treated as wide strings in __func__.  Patch from Iain Hibbert.

[amd64]
    Add last memcpy for function returning structs.  Should be complete now.

[configure]
    Set default wchar type to int instead of unsigned.

[cc]
    Fix wchar defines as requested by Iain Hibbert.

[cpp]
    One more comment-handling bug found by Iain Hibbert.

    Forgot to null-terminate yytext.  Bug pointed out by Iain Hibbert.

[ccom]
    No reason not to allow pointers larger than LONG, as pointed out by
    Peter Kuschnerus.

    Cast 0x80 to char to avoid a warning.  Fixes compiler warning as
    reported by Peter Kuschnerus.

[amd64]
    Add arg count index to builtin switch table.

[ccom]
    Fix segfault bug, found and patch provided by Iain Hibbert.
    While here, make the switch table const.

    Do not segfault if the string given to __builtin_nan is a NULL pointer.
    Bug reported by Iain Hibbert.

[ccom]
    Discard comments in yylex() so that -C works even with comments in
    #if directives.  Supposes to fix Jira#PCC-126 by Iain Hibbert.

    Add huge/inf/nan builtins as requested by Iain Hibbert.
    Also add a generic builtin function, based on a slightly modified
    version of a patch by Iain Hibbert.

    Add another cast routine which will cast without adding a "cast" and
    therefore it will complain if it's a bad cast (like int to pointer).

[amd64]
    Fix bugs where int should be long.  Fixes Jira#PCC-125 by Nicholas Joly.

[ccom]
    Break out buintins to their own file.  Common argument count check
    for builtins. Should have argument cast checks here also.

[i386]
    Add support for attribute "visibility".

[amd64]
    Do not emit visibility type "default" since it's ``default''.

    Print out visibility directives for as.

[ccom]
    Add attribute support for visibility per patch from Iain Hibbert.

[i386]
    Add support for constructor/destructor attribute.
    XXX defloc() should be possible elf-common code.

[amd64]
    Add con/destructor support. XXX pragma init/fini should be MI.

[ccom]
    Add con/destructor as attributes per patch from Iain Hibbert.
    Different target implementations though.

[amd64]
    Fix struct argument/return bugs.  Add PIC support.

[ccom]
    Change some forgotten code to find the bottom type size; since type
    sizes are not directly inherited anymore.  Fixes coredump bug by
    Iain Hibbert.

[cc, cpp, ccom]
    Fix arrows from joerg@netbsd via plunky on pcc-list.

    Escape some dashes and use an em-dash (noted by Thorsten).

    Fix broken markup in manpages (from plunky on pcc-list).

[ccom]
    Add __builtin_memset, per patch from Iain Hibbert.

[i386]
    Remove unused variable, as Iain Hibbert kindly informed.

    Remove unused (and irrelevant) statement.

[ccom]
    Add cast() function that is an easier-to-use wrapper around buildtree().


[ccom, mip]
    kill some commons; ragge@ ok

[cpp]
    Allow # alone on lines, as requested by Mickey.

[ccom]
    Accept gcc attributes after a gcc asm() rename directive. Bugfix
    provided by Iain Hibbert.  Supposed to fix Jira#PCC-104.

    Implement __builtin_memcpy().  Now pcc can compile itself again on OS X.

    Add support for variable renaming when assingning; i.e.
    int foo asm("bar") = 2;
    and not only
    int foo asm("bar");
    Should fix Jira#PCC-116 by Altamiranus.

    Add support for attribute modes SF, DF, byte, word, pointer.
    Should fix Jira#PCC-114 by Altamiranus.

    Understand QI mode attribute as CHAR.  Fixes Jira#PCC-113 by Altamiranus.

    Understand function attributes pure and const.  Fixes Jira#PCC-110 and
    Jira#PCC-111 by Altamiranus.

[cpp]
    Fix const usage.  From Iain Hibbert.

[mip]
    License (and small code) change in the code repository where the
    function basename() came from.  From Iain Hibbert.

[sparc64]
    Correct format strings and STASG return register.
    Addresses PCC-107 and PCC-108 from Dmitry Alenichev.

[i386]
    Remove unused locals on non-openbsd builds.

[cc]
    Don't define __GNUC__ on win32 until problems with header files are
    resolved.

[ccom]
    Add my local changes for building with the Microsoft compiler.

    Add support for attribute mode and some mode types. Should solve
    Jira#PCC-101 by Detlef Riekenberg.

[cpp]
    Print out an extra \n before emitting a new line-file directive.
    Solves Jira#PCC-97 by Gregory McGarry.

[ccom]
    Add gcc attribute nothrow as patch in Jira#PCC-102 by Detlef Riekenberg.

[cpp]
    Understand gcc directive #include_next. Solves Jira#PCC-106 by Altamiranus.

[ccom]
    Handle more cases of attribute.  Adds another shift/reduce conflict.

    Understand malloc attribute and ignore it, we do not do aliasing yet.

    Handle sizes of pointers to forward declarated strust/unions better.

[i386]
    Print out constants in the same radix as gcc, to simplify debugging.

    Correct return values of STASG.

[mip]
    STASG must be treaded as ASSIGN in sanity checks.

[i386, amd64]
    generate symbol types and section attributes for elf; ragge@ ok

[ccom]
    Ueck, forgot some debug printouts :-(

    Handling of attribute strings was broken; they were treated as names.
    Reported by Michael Shalayeff who provided a diff, but fixed in a more
    common way.

[i386]
    Fix bug causing comparision of two long-long values to sometimes fail
    (low-order values must be compared unsigned). Reported by Jonathan Gray.
    Also fix a bug in struct assignment causing broken code to be generated.

[cc]
    Get GCC_COMPAT from configure.

    At last - defined __GNUC__ etc.  and also __GNUC_STDC_INLINE__.

[mip]
    Only try to shortcut the conditional code generation if the operands are
    of types +-*/%&|^>><<, otherwise side effects may be lost.  Bug found
    by Jonathan Gray.

[ccom]
    Add the g77 types, for compatibility.

[amd64]
    Give 32-bit register to movl instruction.  Fixes Jira#PCC-100 by
    Dara Hazeghi.

    Allow moving of TEMPs to regs. Fixes Jira#PCC-99 by Dara Hazeghi.

[sparc64]
    Sparc64 patches from Lars Zetterlund.  Are supposed to fix Jira#PCC-20.

[ccom]
    Add __builtin_offset(), which is becoming more and more common.
    Suggedted by Jonathan Gray.

[csu]
    Use 'test -f' rather than 'test -e' for improved portability.
    From Adam Hoka.

[powerpc]
    Improve alignment.

[linux]
    Improve building of shared libraries on linux.  Tested on Ubuntu 8.04.

[libpcc]
    Remove definition of off_t, since it is frequently wrong.  Confirmed OK on
    Ubuntu 8.04, OSX and Win32.

[i386]
    Fix alignments on OSX.

[tests]
    Get test to work on win32.

    Don't use __attribute__ in this test, since most systems will define it
    away if not gcc.

[mip]
    Change static array to allocated linked list for temporaries not allowed
    to be spilled.  libssl seems to make a heavy use of extended assembler.
    Bug reported by Jonathan Gray.

[i386]
    dd support for gcc xasm arg types I, J, K, L, M, N.  Fixes Jira#PCC-96
    by Jonathan Gray.

[ccom]
    kill unused variables; ragge@ ok

[hppa]
    fix merge error

[ccom]
    Put back nocon_e, creates better yacc code.

[hppa]
    fixx fallout form walkf() change and other gcc bitchings

    add pragma support (from amd64 ;)

[powerpc, i386, arm, amd64, mips]
    Return to the text segment after putting reference in constructor table.
    Should fix PCC-65.

[ccom]
    Rearrange yacc code slightly to avoid a shift/reduce conflict, that
    could cause unwanted syntax errors.

    Giving multiple declarations of the same prototype failed if there
    were dynamic arrays as arguments.

    Fix bug in debug printout of prototype parameters.

    Accept [*] as [] in prototypes.  Fixes Jira#PCC-94 by Jonathan Gray.

[mip]
    Must add moves before and after a +r xasm directive even if the node
    is in a temporary; the temporary may need to be spilled.

Status:

Vendor Tag: ragge
Release Tags: pcc-100603

(plunky)

Use proper signal names. Prefixing them with "SIG" is an extension which
is not supported by posix shells.

(christos)

Don't quote strings to be expanded as we quote the content already.

(joerg)

need errno for the debug build.

(christos)

set -e is supposed to work inside eval; skip EV_TESTED.

(christos)

Pass all the set -e tests.

(christos)

We have required sigaction() for quite a while.
Use bmake_signal() - a wrapper around sigaction()
rather than signal() to ensure that signals are handled
consistently.

(sjg)

perhaps this fixes the build issue?

(christos)

Use MAP_FILE|MAP_SHARED instead of MAP_FILE for the flags parameter of mmap(2).
Patch is from martin.

Solves my own PR kern/43346

(veego)

Implement a sort-of pagedaemon: adjust all memory allocators to go
through an in-rumpkernel hypermemory allocator which knows it should
kick the pagedaemon and block in case ``waitok'' memory allocation
fails.

This allows us to recover from some out-of-memory situations.
Realworld'istically speaking (as opposed to whatever "should be"
theory), these OOM situations will happen extremely rarely if ever
when our hypervisor is a regular process.  Speculatively, this
should be useful for other types of hosts.

issues remaining:
* the hypervisor does not know how to reclaim kernel memory (and
  for the reason I stated above, I'm not sure if it makes sense
  to teach the current implementation about that)
* vfs memory (buffers, vm object pages etc.) is not reclaimed

(pooka)

Applied a patch by Toru Nishimura:
Fixed kmutex usage error and replaced tsleep by cv_waitsig.
Now "cat /dev/satmgr" no longer panics when pressing a button.

(phx)

Report result of pool_reclaim() from pool_drain_end().

(pooka)

.include <bsd.lib.mk> to handle the pkgconfig file correctly.

(ghen)

Needs additional single quotes to keep the double qoutes. Magic...

(joerg)

* Add missing .WAIT before libdrm_intel since it now depends on libdrm
* Move libchromeXvMC further down since it depends on some other libraries

(veego)

Link against libXext, libX11, libXv, libXvMC and libdrm to fix a hidden
dependency

(veego)

Link against libdrm to fix a hidden dependency

(veego)

Link against libz to fix a hidden dependency

(veego)

It is __NetBSD_Version__ and not __NetBSD_Version
Fixed a hidden issue with __progname

The i910 and i965 dri modules still have a problem with the __progname usage
in MesaLib/dist/src/mesa/drivers/dri/intel/intel_context.c

(veego)

disable the DNS tests which rely on external hosts, as
they are not useful with automated testing systems without
internet connectivity.

(plunky)

Note that set -o tabcomplete requires either set -o emacs or set -o vi
to work.

(dholland)

Fix DMA setting for read/write on conventional PCI bus devices.
This bug was added in rev. 1.166. Sorry...

(msaitoh)

Prevent if_detach() from crashing while it walks the routing table
to find and unlink routes that reference the detached ifnet: make
if_rt_walktree() return ERESTART whenever it has deleted a route.
Whenever rt_walktree() returns ERESTART, if_detach() restarts it.

I believe that this fix resembles one by Jonathan Kollasch or by someone
else, which has languished in a PR for too long.  Sorry!

Tested by me and by Jeff Rizzo.

XXX It's supposed to be safe for rn_walktree() to apply to the routing
XXX table a routine that may delete routes.  Why isn't it safe in
XXX practice?

(dyoung)

Avoid .TS H and .TH for now as it doesn't alter the output in this case and
conflicts with the header check from mandoc. Having somewhat working
HTML output is more useful.

(joerg)

Move all miiverbose/module related functions from ukphy.c to mii_physubr.c -
ukphy is not included in all kernels, but mii_physubr.c should be in every
kernel that has a phy.

(martin)

XORGVERSION should be a single quoted expression as it represents a
single argument for .TH

(joerg)

Improve EHCI isochronous code, fix scheduling.

- Create and use symbolic constants.

- Convert some switch statements to functionally-similar caculations.

- Correct scheduling interval of high-speed isochronous transactions.
  Previous calculation produced half the intended rate.

(jakllsch)

Check return value of posix_memalign ... always helpful to not return
garbage memory in case of failure.

(pooka)

Remove orphaned heading "Mice"

(dholland)

COMPAT_43 should not be sorted between COMPAT_40 and COMPAT_50.

(dholland)

COMPAT_43 should not be sorted between COMPAT_40 and COMPAT_50.

(dholland)

Add symbolic constant for number of USB 2.0 microframes per frame.

(jakllsch)

uint8_t -> unsigned char, int16_t -> short.

(tnozaki)

it's a wonderful static

(pooka)

remove duplicated towctrans internal function.

(tnozaki)

Drop empty .TP. Kill some useless quotes while here.

(joerg)

Drop empty .TP.

(joerg)

In aiodone, call uvm_pageout_done() with number of PG_PAGEOUT pages
processed.

(pooka)

rumpvm_init -> uvm_init to get rid of local prototype.
no functional change

(pooka)

Initialize the initial snap block list's count.

From Antti Kantee <pooka@netbsd.org>.

(hannken)

Restore PHYSMEM_MAX* options (hi cegger!)

(joerg)

Move memory protection parameters to MI part.
Gtpci believes the initialization of the firmware(boot-loader) if protection
information is not obtained.

(kiyohara)

Set NULL to tag when the property doesn't get it.

(kiyohara)

Remove unused gpio.h.

(kiyohara)

Add Marvell Gigabit Ethernet Controller on SC/SoC.
  Will support on ofppc soon.
  Tested on G4+PegasosII (kiyohara@)
    G3+PegasosII (phx@)

(kiyohara)

Use kmem(9) instead of malloc(9) in gtidmac_attach().

(kiyohara)

Erase the old definition macro that forgot to be erased.

(kiyohara)

Remove unused header-file.

(kiyohara)

Check for a null keyring before listing keys - with thanks to Jeff Rizzo
for the problem report.

(agc)

commit version number change

(agc)

Add code, dev_mem_getva() and dev_mem_relva(), to deal with cache-aliasing
issues by allocating an appropriate KVA from physical address, according to
the colour.  Used by architectures, which have such requirement.  For now,
enable only for MIPS, others will follow.  This renames previously invented
mm_md_getva() and mm_md_relva(), since we do this in MI way, instead of MD.
Architectures just need to define __HAVE_MM_MD_CACHE_ALIASING as indicator.

Reviewed by Matt Thomas.

(rmind)

ACPICA version 20100528 has been released.

(jruoho)

Install mdocml's default CSS as /usr/share/man/style.css.  Switch HTML
man pages to use mandoc unconditional as it gives reasonable output for
all man pages, not only a subset of mdoc(7).  Use the newly installed
style.css for formatting and produce hyperlinks for .Xr.

(joerg)

Add __cacheline_aligned and __read_mostly annotations.

These annotations help to mitigate false sharing on multiprocessor
systems.

Variables annotated with __cacheline_aligned are placed into the
.data.cacheline_aligned section in the kernel. Each item in this
section is aligned on a cachline boundary - this avoids false
sharing. Highly contended global locks are a good candidate for
__cacheline_aligned annotation.

Variables annotated with __read_mostly are packed together tightly
into a .data.read_mostly section in the kernel. The idea here is that
we can pack infrequently modified data items into a cacheline and
avoid having to purge the cache, which would happen if read mostly
data and write mostly data shared a cachline. Initialisation variables
are a prime candiate for __read_mostly annotations.

(mjf)

Update to mdocml-1.10.0

(joerg)

Merge mdocml-1.10.0

(joerg)

Import mdocml-1.10.0:
- basic support for the low-level roff macros
- various cleanups and improvements

Status:

Vendor Tag: KRISTAPS
Release Tags: mdocml-1-10-0

(joerg)

Fix typo in sdpd, reported by Jukka Salmi in PR 43401.
This also affected bthcid, so I also fixed it there.

(wiz)

Don't pass "canfail" down to rumpuser_malloc -- there's quite little
we can do with that info way down there.  Instead, pass alignment.
Implement rumpuser_malloc() with posix_memalign().

(pooka)

adjust autoconfig output: this item is part of a verbose line

(plunky)

Always use rumpuser_malloc() for allocating both poolpage and
poolpage_cache -- its bootstrap cost is slightly higher than
anonmmap, but it's faster in the long run.

(pooka)

From Manuel Bouyer:

In the matt-nb5-mips64 branch, in mips_emul.c:MachEmulateBranch(), the
BRANCHTARGET macro was changed to take the value instead of a pointer to
instruction. But the effect of this change is that now, the instruction's word
value is used to compute the new PC, instead of the instruction's address. Of
course this can't give good results, and in my case this gave an unaligned PC.

(matt)

Fix bad initialization spotted by Manuel Bouyer.

(matt)

tidy up memory allocation and don't unnecessarily print "./" before names.

(christos)

refactoring old locale-db(BSDCTYPE) loading method with mmap(2).

(tnozaki)

more split ctype.h -> sys/ctype_inline.h, sys/ctype_bits.h

(tnozaki)

Use .An.

(wiz)

* remove rumpvm_makepage, just use uvm_pagealloc()
* update copyright to reflect reality a little better

(pooka)

Add a man page for the hp700 lcd(4) driver.

(skrll)

Add the lcd(4) driver from OpenBSD.  Thanks to Adam Hoka for doing most of
the work.

(skrll)

handle STDC and GNUC inline semantic differences by providing and
using an EXTERN_INLINE definition for functions that are defined as
inline but provide an externally callable reference.

(these are externally called in ipftest)

(plunky)

later versions of pcc may define __GNUC__ for compatibility reasons, so
having a system definition for alloca is counter productive as it seems
fairly common to use

    #ifdef __GNUC__
    #define alloca ..
    #endif

in user-headers (eg see usr.sbin/gspa/gspa/gspa_ass.h)

So, defang this definition as we don't otherwise provide alloca on NetBSD

(plunky)

avoid another compiler warning

(agc)

remove an unused variable

(agc)

Update netpgp to 3.99.3

Changes since 3.99.2:
+ avoid possible free() of new value passed to netpgp_setvar(),
  with thanks to Anon Ymous.
+ netpgpkeys(1):  print keys to stdout, not stderr - reported by Anon
  Ymous.
+ fix DSA signatures and verification
+ simplify and shorten the internals of packet processing by getting rid of
  the intermediate pseudo-abstraction layer, which detracted from understanding
  and had no benefit whatsoever. Rename some enums and some definitions.
+ add some checking to new key generation, and don't try to read in
  the keys after writing them - reported by Tyler Retzlaff

(agc)

USETBL=yes.  several of these manuals need it.  from PR#36600.

(mrg)

make sure we have created a directory when generating a new key.

don't try to re-read the key after writing it - that's done by a separate
function. Problem found by Tyler Retzlaff, fixed in a different way.

check that keyrings are non-NULL before attempting to free them - from a
nudge by Tyler Retzlaff.

(agc)

fix a macro definition error, as per PR#41900 from Henning Petersen.

(mrg)

fix a prototype error, as per PR#41899 from Henning Petersen.

(mrg)

when cleaning up, don't try to free the public key which is part of the
secret key - shown up by Mac OS X malloc.

(agc)

enable the drm drivers; they've been tested for ages and have been
enabled on amd64 for some months now.  as requested in PR#42936.

(mrg)

need bsd.own.mk to avoid the "NETBSDSRCDIR is nothing" syndrome

(pooka)

hmm, using unifdef for the previous might have been a good idea ...

(pooka)

Now that atf gdb/coredump conflicts have been solved (thanks jmmv!),
remove USE_ATF ifdefs.

(pooka)

new tests

(pooka)

add tests for:
* threading
* tsleep variants
* uvm page busying and wanting

(pooka)

Add some kernel namespace (i.e. _KERNEL) helpers for regression tests.

(pooka)

Support mtsleep() without a biglocked sleeper (uvm uses this in
UVM_UNLOCK_AND_WAIT())

(pooka)

Dump rump kernel bootstrap time.

(pooka)

Support KTHREAD_JOINABLE/kthread_join().  Also fixes earlier bug
where all pthreads were created non-detached.

(pooka)

The x86 kernel ABI depends on __cpu_simple_lock stuff being present.
Since they are practically never used (only when prehistoric code
uses simple_lock()), their efficiency doesn't matter that much and
we can simply adapt the versions from x86 lock.h.

(pooka)

Extract ACPIVERBOSE into a kernel module.  The module can be builtin
by defining 'options ACPIVERBOSE' in the kernel config file (no change
from current behavior), or it can be loaded at boot time on those
architectures that support the boot loader's "load" command.

ACPIVERBOSE also includes code to dump acpi/wmi data.

(pgoyette)

Fix previous.

(skrll)

Silently return if a power state transition is requested for a device that
has no support for it. For example, in Kurt Schreiner's Shuttle G61:

        Device (FAN)
        {
            Name (_HID, EisaId ("PNP0C0B"))
            Method (_INI, 0, NotSerialized)
            {
                Store (TP1H, CTOS)
                Store (TP1L, CTHY)
            }
        }

Obviously acpitz(4) has little use for a device that only contains _INI.
(The handle of this device is referenced incorrectly in _AL0.)

(jruoho)