Reimport Heimdal dist converted into NetBSD dist format.  The prior
import did not catch all of the $Id$ tags because many of them are
unexpanded and the RE was only looking for expanded keywords.

Status:

Vendor Tag: HEIMDAL
Release Tags: head-20110412

(elric)

Fix this so that it builds (missing "struct" keyword).

(he)

add sparc64/intr.h

(mrg)

use HOSTPROG to build the mkext program (which creates external.c and
external.h for ccom) rather than HOST_CC directly.

because PROG and HOSTPROG don't play nicely together, put mkext in its
own SUBDIR

(plunky)

Use .set arch=xlr to access RMI specific instructions.

(matt)

xz-5.0.2 out.

(wiz)

port to userspace.

(mrg)

enable crash(8) for sparc64.  it's still sort of minimal and segvs
on some commands, but it sort of works and isn't useless.

(mrg)

Add L_PRIVATE

(matt)

Support LWP_PC

(matt)

Remove old mfcr/mtcr definitions

(matt)

Add bfd_mach_mips_xlr

(matt)

Pull in MIPS XLR (XLS) support from patch in
http://sourceware.org/ml/binutils/2009-01/msg00383.html
This replaces local changes to support some xlr instructions.

(matt)

Simplify a little. No functional change.

(jruoho)

No need to use \*[q], use plain double quotes instead.

(wiz)

cpuwatch_* stuff is #if (MIPS32 + MIPS32R2 + MIPS64 + MIPS64R2) > 0
now cobalt can build

(cliff)

Use the common power resource code. Also add a comment explaining
what can happen if we do not power on the resource upon resume from S3.

(jruoho)

Simplify by using a static array for the reference counting.

(jruoho)

- fix comment at cpucore_rmixl_run()

(cliff)

rearrange some common code in lockdebug_unlocked.
When printing a lockdebug error, add a '*' to indicate whether lock or
unlock was performed.

(matt)

- no need to check lsw_cpu_run != NULL before calling, nullop is default

(cliff)

Add a KASSERT

(matt)

- add loocoresw slot for lsw_cpu_run
- fix comments for locoresw fields

(cliff)

Xref tm(3).

(jruoho)

- add loocoresw slot for lsw_cpu_run
- fix comments for locoresw entries

(cliff)

Improve as per discussion on source-changes.

(jruoho)

Add an assert to make sure no unexpected spinlocks are held in mi_switch

(matt)

- in rmixl_pcix_intr_string() convert irq to vector when calling
rmixl_intr_string()
- in rmixl_pcix_intr_establish(), initialize dispatch data 'counts' pointer
- in rmixl_pcix_pip_add_1(), zero out pip_new after allocated.

(cliff)

- in rmixl_pcie_intr_string(), convert irq to vector when calling
rmixl_intr_string()

(cliff)

- add prototype for rmixl_eirr_ack()

(cliff)

- report of BIST result is information only; don't fail attach because of it

(cliff)

- fix RCSID
- add rmixl_eirr_ack() to ack the EIRR, using COP0_SYNC & JR_HB_RA as needed
- in rmixl_cpu_trampoline, remove old KSEG0 address reconstruction
of trampoline args pointer, and comment/explain the new way
- also in rmixl_cpu_trampoline, remove old watchpoint hack used for debugging

(cliff)

- remove most of the PARANOIA code
- add COP0_SYNC and JR_HB_RA following CP0 after all
writes to STATUS or EIMR that change interrupt control.
- all interrupt control now done w/ EIMR, except for
initial set of IE in STATUS.

(cliff)

- add compile time check in case NIPIS ever exceeds number of available vectors

(cliff)

- in evbmips_iointr(), call assembly function rmixl_eirr_ack()
to ack the EIRR, instead of using a bunch of asm() here.
- in rmixl_ipi_intr(), remove overly paranoid assert that the given
IPI request is pending; if the request is clear, it was previously processed.
- in rmixl_vecnames_common[], give ipi vectors unique (numbered) names

(cliff)

- add sc_running flag to allow cpucore_rmixl_run() to do once-per-core
initializations once we know what threads are configured
- add sc_threads_enb and sc_threads_dis to track what threads are
enabled by firmware and configured and attach, or not.
- add prototype for cpucore_rmixl_run()

(cliff)

- in cpucore_rmixl_attach(), keep track of which threads are
enabled by firmware and configured and attach, or not.
- add cpucore_rmixl_run() to do the post-running initialization:
disable unused threads in RMIXL_PCR_THREADEN, and
set Round Robin thread scheduling mode.

(cliff)

- add cpu_rmixl_run(), and set in mips_locoresw.lsw_cpu_run
to be called from cpu_hatch() once cpus are running,
so we can determine what threads are configured
and running, and can finish initialization of per-core registers
depending on that.
- in cpu_rmixl_db_watch_init() clear IEU_DEFEATURE[DBE],
and init all COP0 watchpoint regs
- option MIPS_DDB_WATCH is deprecated, removed; use of cpu watchpoints
is longer depends on that or DDB

(cliff)

- add mips_watchpoint_init() to discover number of CPU watchpoints,
and call that from {mips32,mips32r2,mips64,mips64r2}_vector_init()

(cliff)

- option MIPS_DDB_WATCH is dedprecated, removed; now using
(MIPS32 + MIPS32R2 + MIPS64 + MIPS64R2) > 0 for conditional
compile of db_mach_watch stuff
- MIPS CPU (COP0) watchpoint support moved from db_interface.c
to cpu_subr.c, cpu.h; ddb_mach_watch &etc now use those cpu functions.
- ddb_cpu now volatile
- 'struct db_mach_watch' definition &etc moved to mips/include/db_machdep.h
- db_mach_watch_tab is replaced by curcpu()->ci_watch_tab
to allow per-cpu watchpoint control
- improve MP function in kdb_trap()
- fix conditions for printing cp0 regs

(cliff)

- MIPS CPU (COP0) watchpoint support moved from db_interface.c to cpu_subr.c
see the cpuwatch_* functions
- ci_cpuwatch_count in allocated cpu_info is inherited from cpu_info_store
- cpu_hatch() calls lsw_cpu_run (if not NULL)
to optionally allow running MIPS chip-specific code
after subordinate cpus have been set running
- #ifdef DDB cpu_debug_dump() to allow compile when DDB not defined

(cliff)

- add lsw_cpu_run function pointer to struct locoresw

(cliff)

- remove include <mips/proc.h>, unused
- db_mach_watch_set_all() is deprecated, removed,
superceded by cpuwatch_set_all()

(cliff)

- MIPS CPU (COP0) watchpoint support moved from db_machdep.h to cpu.h
- CPU watchpoints are per-cpu; add ci_cpuwatch_count, ci_watchpoint_tab[]
to struct cpuinfo

(cliff)

- option MIPS_DDB_WATCH is deprecated, removed

(cliff)

- remove cpucore and cpu reconfiguration stuff
cares and threads can be specified here,
but it's easier to do at boot time from firmmware
e.g. using 'userapp_mask_cpus <mask>' command

(cliff)

- 32 bit kernel must MEMLIMIT to 512MB

(cliff)

- remove options LOCKDEBUG; is set in XLSATX64.MP
- option MIPS_DDB_WATCH is deprecated, removed

(cliff)

Initial import of the libhmac(3) library and hmac(1) utility to
calculate an HMAC value.

Some tests based on the IETF SHA2 HMAC tests are included, although
not installed by default.

Examples of usage:

    % hmac -k '  ' -a sha256 Makefile
    sha256 (Makefile) = 9612652f4e2371dc41e75f7d1b80e305a7be54b171b6a6002ba5caa072f3e825
    % hmac -a sha256 Makefile
    hmac key:
    sha256 (Makefile) = 9612652f4e2371dc41e75f7d1b80e305a7be54b171b6a6002ba5caa072f3e825
    % hmac -a sha256 < Makefile
    hmac key:
    9612652f4e2371dc41e75f7d1b80e305a7be54b171b6a6002ba5caa072f3e825
    %

Hashing algorithms supported are:  md5, rmd160, sha1, sha224, sha256,
sha384 and sha512 (using code from libc).  Default algorithm is sha512.

Status:

Vendor Tag: CROOKS
Release Tags: hmac-base

(agc)

use ioctl(FBIOSVIDEO) in CG14SaveScreen(), now the Xserver can actually turn
the monitor off
TODO: hw cursor

(macallan)

add the "special" pipe2.

(christos)

move dup3 to a more appropriate place because pipe is "special". Gotta love
linux.

(christos)

Add DTYPE_SEM.

(rmind)

Minor comment fix.  Use fd_close() in sys__ksem_destroy(), it is cleaner.

(rmind)

use the same function whenever we change colour depth
also, when opening the fb device switch to 32bit and back to 8 on close, not
the other way around

(macallan)

regen (+ ATI Radeon HD4650)

(wiz)

+ ATI Radeon HD4650.

(wiz)

no more amd, factor out more stuff.

(christos)

initialize mutex in attach

(cliff)

initialize mutex in attach

(cliff)

Build framework for Heimdal.

(elric)

Autogenerated headers for heimdal head-20110412.

(elric)

Autogenerated files for HEIMDAL head-20110412.

(elric)

Script to convert a git clone of Heimdal into our dist format.

(elric)

Autogenerated headers for heimdal head-20110412.

(elric)

ExtClk for QNAP should be the same as for Synology's 266MHz systems.
This seems logical, as both boards are very similar, and the clock precision
is ok now.

(phx)

_gss_DES3_get_mic_compat() requires that ctx->target has been defined, and,
well, it hasn't yet.  Move the call down to after it is defined and things
are better.

(elric)

Conditionalise the sqlite3 HDB backend.

(elric)

Import latest Heimdal from the head of their git repository into the
new location for externally maintained software.

Status:

Vendor Tag: HEIMDAL
Release Tags: head-20110412

(elric)

default to wskbd
tested on amd64, sparc(64), shark
Now X without a config file should work on most hardware

(macallan)

use PRIx64 for printing 64-bit values

(plunky)

pass ${MACHINE} as -D__<whatever> so we can have arch-specific defaults

(macallan)

- get platform #ifdefs in a slightly more sane order
- look for SBus and UPA devices on sparc(64)
- add default drivers for shark and sgimips
- only try wsfb if we can't find any other usable device
With this Xorg without config file works fine on shark, sparc(64) and amd64,
should work on other archs as well.

(macallan)

xf86LoaderReqSymLists() is no more

(macallan)

Move the forward declaration of _sys_readlink() outside of the #if,
so that the build succeeds even if _FORTIFY_SOURCE isn't > 0.

(he)

Convert to using -nostdinc, and set include paths and create symlinks
so that we include the source tree's sys/, machine/ and m68k/ headers,
and only as a last resort use DESTDIR's /usr/include.  For some reason
the host's bootinfo.h got included ahead of the x68k variant, causing
a build failure in my case.

(he)

Explicitly list -lterminfo after -lcurses, to support static builds.

(he)

expose the KSTACK_LOWEST_ADDR and KSTACK_SIZE to _KMEMUSER as well,
like the x86 versions do.  for crash(8).

(mrg)

Collect also the division functions to single place, div(3).

(jruoho)

Collect abs(3), labs(3), llabs(3), and imaxabs(3) to a single small page.

(jruoho)

Fix another bug (no 'imaxint_t' type).

(jruoho)

Fix obvious bug.

(jruoho)

move the include sys/types.h xor stdbool.h to the top of the file,
so that "bool" will be present when used later in the file.

(mrg)

fix sbusDeviceTable to actually include driver names

(macallan)

Split out from ieee(3) also the remainder(3) and copysign(3) families.

(jruoho)

re-include sparc64/sparc64/intreg.h in kernel mode.  it's necessary.

(mrg)

export a bunch of things to _KMEMUSER as well.

XXX: should consolidate the curcpu/curlwp definitions.

(mrg)

port to userspace.  this needs db_user.h, stddef.h, stdbool.h and not
promlib.h.  put db_read/write_bytes() under _KERNEL.

XXX: for now, provide an empty db_machine_command_table[].

XXX2: this file puts a lot of it under #ifdef DDB, but i'm not sure
why this is done.  for now it makes it easier to port to userspace :)

(mrg)

port to userspace.  needs <stdbool.h>, <sys/cpu.h>.  delete unused
ULOAD().  implement KLOAD() in terms of db_read_bytes() in userspace.
can't use proc_find_raw() in userspace; disable this functionality
for now.  use db_printf().

(mrg)

apply s/TRUE/true/; s/FALSE/false/;.  needed for userland.

(mrg)

this needs db_user.h and stdlib.h in userland.

(mrg)

install <machine/intr.h> for crash(8).

(we could avoid this, but it would require a lot of contortions that
make the header files ugly.  maybe someone else can figure out a less
ugly way of doing this later.)

(mrg)

this needs sys/types.h in userspace.

(mrg)

Add prototypes for mipsNN_cp0_watch{lo,hi}_{read,write}

(matt)

Add mipsNN_cp0_watch{lo,hi}_{read,write}

(matt)

do_ksem_open: do not leak the file descriptor on error path.
sys__ksem_unlink: fix comment, release one lock ealier.

(rmind)

Fix SSP builds (Vladimir Kirillov)

(christos)

Fix SSP builds (Vladimir Kirillov)

(christos)

Get rid of an errant KASSERT

(matt)

Dont bother taking out TLBINFO lock in pmap_tlb_asid_deactivate since the
atomic ops will DTRT with it.  Remove spl calls too.

(matt)

Add __lwp_getprivate_fast

(matt)

avoid a crash when the display size is reported as 1x1
XXX this is hackish, it should be properly fixed instead

(ahoka)

add NAND_BBT

(ahoka)

Add a CHMU handler to implement __lwp_getprivate_fast.

(matt)

Fix GENERIC_SUN4U build.

(nakayama)

Fix MKDEBUG set lists.

(matt)

Add missing ./usr/libdata/debug/usr/tests/lib/libcurses

(matt)

Add support for TLS relocations.
From Joerg Sonnenberger

(matt)

Add RCSIDs

(matt)

Make x68k bootloader to use MI libsa, libkern, and libz Makefiles in
src/sys/lib dirs for maintainability.  Tested on X68030, ok from isaki@.

(tsutsui)

Put asctime(3) explicitly to SEE ALSO.

(jruoho)

Split out FUNCTIONS to make this tolerable to read. Also improve readability
with lists. Describe the functions in the order of appearance in SYNOPSIS.
Use CAVEATS instead of NOTES. Remove "man page spam". Etc.

(jruoho)

include <sys/cpu.h> when using cpu specific info.  fixes building
these as part of sparc64 crash(8).

(mrg)

Xref tm(3).

(jruoho)

Move some of the discussion from the messy ctime(3) to tm(3).

(jruoho)

Add a small summary parge for struct tm from <time.h>. Cf. timeval(3).

(jruoho)

make sparc/frame.h idempotent.

(mrg)

Split the nextafter(3) family of functions to their own page and properly
document these.

(jruoho)

.Dv NULL.

(jruoho)

previous fix does not work if there is exactly only one entry where continue
exits the loop.
Apply fix from Konrad Wilke on port-xen@
That makes NetBSD DomU boot on Linux Dom0 with xl.

(cegger)

Use RETURN VALUES rather than DIAGNOSTIC for... return values.

(jruoho)

Small rewording.

(jruoho)

rm -f ${.TARGET} before trying to > write to it; some how mine ended up
being mode 444 and the build failed.

(mrg)

PR lib/41931 should be fixed; remove expected failure.

(jruoho)

Better to still leave the old PR number as a comment.

(jruoho)

PR lib/33262 should be fixed; remove expected failure.

(jruoho)

replace a non-working make construct with a working one.
now crash only builds unsupported.c on non-x86 systems again.

(mrg)

Make sure we init MIPS COP0 OSSCRATCH for UP as well as for MP.

(matt)

Add code to fixup the text segment to replace loads
from L_CPU(MIPS_CURLWP) or curlwp->l_cpu with
loads from COP0 register OSSCRATCH.

(matt)

Adjust --sysroot related usage:
-Wl,-rpath does not expand =, so just drop it.

Drop -Wl,-rpath-link entries that duplicate the -Wl,-rpath entries, this
is done implicitly now that ld is built with sysroot support.

Use ${DESTDIR} explicitly for the remaining -Wl,-rpath-link entries.

(joerg)

Extend, not reassign CPPFLAGS.

(joerg)

Don't fail on platforms with "char" == "unsigned char"

(joerg)

Highly rewrite POSIX semaphore.  Use file descriptor subsystem, greatly
simplify reference counting, and fix various bugs which existed before,
e.g. some incorrect locking or mq_unlink() not checking for permissions.

Also, fixes PR/43452.

(rmind)

Collect all of var/chroot/tcpdump in one set - so we have a chance to
win the checkflist game.

(martin)

Ooops, fix source path for the "protocols" file

(martin)

Clean up some compile warnings

(roy)

POSIX says that term.h should define tgetstr and friends
Fixes PR lib/43941

(roy)

GCC 4.1 doesn't have the support code for -I =... yet, so don't use it
for now.

(joerg)

Add a new check to populate /var/chroot/tcpdump/etc

(martin)

Should no longer fail - if it still does, be sure to have a proper
/var/chroot/tcpdump/etc/protocols file installed (i.e. by running
postinstall)

(martin)

When run as root, tcpdump will chroot to /var/run/tcpdump - but it can
not look up /etc/protcols in there. So install a copy of /etc/protocols
into the chroot area.
Fixes PR bin/44721.

(martin)

softint_execute: add assert which could catch locking bugs in softint handlers.

(rmind)

G/C unused MUTEX_COUNT_BIAS (it was for VAX)

(rmind)

Partially revert part of -r1.167; it was wrong. From dholland.

(jakllsch)

Enable strtc(4). Should theoretically work.
It works indeed for my DSM-G600 after I removed the battery. But there
must be something which confuses the chip later on...

(phx)

Fix build with stack smash protection enabled.

(tron)

Use bsd.klinks.mk to create machine and ${MACHINE_ARCH} symlinks.

(tsutsui)

let log(<0) return NaN rather than -inf in POSIX/XOPEN modes, as
requested in PR lib/41931 by Havard Eidnes (the PR refers to POSIX,
the OSF/1 manpage suggests that XOPEN should behave that way too)
being here, do the same to log10 and log2

(drochner)

Add a test case for sysV modifier subst

(sjg)

Continue scanning for other frontends when initialization
of one frontend failed. Bug reported by Konrad Wilk on port-xen@.
Fix this for all error pathes within the loop.

(cegger)

If a "long" modifier name does not match, goto default case
so SysV modifier can be tried.

(sjg)

Some KNF. No binary change.

(tsutsui)

Bump version to denote serial console fix.

(tsutsui)

Add missing breaks.  No need to output to SERIAL on ITE console.

(tsutsui)

Fix pasto.  Now serial console output on bootloader works properly
after ~decade.  ok from minoura@

(tsutsui)

Add libm/t_infinity

(martin)

Try to generate infinite values in all available floating point formats
on the FPU and check isinf() and fpclassify() results. This is
different from the libc internal consistency checks testing PR
lib/33262.

(martin)

remove DB_AOUT_SYMBOLS.

(mrg)

Apply patch provided by mrg to fix bad calls to vw_printw and vw_scanw

(blymn)

Fix NetBSD tag field.

(blymn)

Fix incorrect call example.

(blymn)

Incorrect licence was applied - fix this.

(blymn)

build xen kernels again after db_trace merge

(cegger)

Remove trailing whitespace.

(wiz)

Markup improvements.

(jruoho)

don't take pmap_lock in a few places.  there's a bunch more work
we can do in here, but i've run with this change for ages.

(mrg)

avoid compiling dead code into crash.

(mrg)

obsolete DB_AOUT_SYMBOLS.  however, we need to leave most of the code
in db_sym.[ch] as it is used by the elf version of crash(8).

i will be cleaning up the db_sym.c code in a follow up commit to avoid
having dead code compiled.

(mrg)

description:
Update comments.

(dholland)

Add comment warning about case with LOCKPARENT but not LOCKLEAF. bleh.

(dholland)

description:
Remove dead assignment of "error" and simplify some uses of it.

(dholland)