Remove the ATF_CHECK calls I added during the conversion of this test to
atf.  They break on i386 (because the test was conceptually broken anyway);
reported by pooka@ in private mail.

Now... the current test does not actually check anything AFAICT... but this
is how it was before.

(jmmv)

... and HAVE_STRNLEN

(mrg)

xlsclients wants xcb-atom as well.

(mrg)

oops, install all the xcb-util headers.

(mrg)

xlsatoms 1.1.0 wants to link against libxcb, not all the other libs.

(mrg)

merge xlsfonts 1.0.3.

(mrg)

merge xlsclients 1.1.0

(mrg)

merge xlsatoms 1.1.0.

(mrg)

initial import of xlsfonts-1.0.3

Status:

Vendor Tag: xorg
Release Tags: xlsfonts-1-0-3

(mrg)

initial import of xlsclients-1.1.0

Status:

Vendor Tag: xorg
Release Tags: xlsclients-1-1-0

(mrg)

initial import of xlsatoms-1.1.0

Status:

Vendor Tag: xorg
Release Tags: xlsatoms-1-1-0

(mrg)

Add ACPI CPU.

(jruoho)

Note conversion of lint1 tests to atf.

(jmmv)

Remove lint1 tests; they have been converted to atf and now live in
tests/util/xlint/lint1.

(jmmv)

Add entries for lint1 tests.

(jmmv)

Add directories for lint1 tests.

(jmmv)

Convert the lint1 tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

(jmmv)

Uh, it's not 2001 any more.  Fix dates of my previous entries.

(jmmv)

add a couple of missing source files.

(mrg)

Add missing CVS identifiers.

(jruoho)

Add a Makefile for ACPI CPUs.

It is safe to load and unload the whole ACPI CPU -functionality at runtime.

(jruoho)

Merge a driver for ACPI CPUs with basic support for processor power states,
also known as C-states. The code is modular and provides an easy way to add
the remaining functionality later (namely throttling and P-states).

Remarks:

  1. Commented out in the GENERICs; more testing exposure is needed.

  2. The C3-state is disabled for the time being because it turns off
timers, among them the local APIC timer. This may not be universally
true on all x86 processors; define ACPICPU_ENABLE_C3 to test.

  3. The algorithm used to choose a power state may need tuning. When
evaluating the appropriate state, the implementation uses the
previous sleep time as an indicator. Additional hints would include
for example the system load.

Also bus master activity is evaluated when choosing a state. The
usb(4) stack is notorious for such activity even when unused.
Typically it must be disabled in order to reach the C3-state,
but it may also prevent the use of C2.

  4. While no extensive empirical measurements have been carried out, the
power savings are somewhere between 1-2 W with C1 and C2, depending
on the processor, firmware, and load. With C3 even up to 4 W can be
saved.  The less something ticks, the more power is saved.

ok jmcneill@, joerg@, and discussed with various people.

(jruoho)

build and install libxcb and xcb-util libraries.

(mrg)

add some more transforms to the hacky ugly sed script for pkg-config files,
and note a few more that are not yet handled but should be.

(mrg)

add generated libxcb config.h

(mrg)

add build infrastructure for xcb-utils 0.3.6.

(mrg)

don't install xprint.h.

(mrg)

Shorten some of the descriptions.

(jruoho)

Use standard MII register definitions from mii.h.  No binary change.

(jakllsch)

install the xcb-proto files; not used yet.

(mrg)

build support for libxcb 1.6.  not yet used.

XXX: i couldn't get the python/xml generator to run outside of
XXX: the libxcb build tree, so i just copied these from there,
XXX: but ideally files/Makefile could know how to do this.

(mrg)

- add X11SHAREDIR definition
- X11SRCDIR's for libxcb, xcb-proto and xcb-utils

(mrg)

Put the original NULL pmf handler establishment back where it started, so
as not to call it in some error paths.

Instead, if there is a watchdog present, deregister the NULL handler before
establishing the one for the watchdog.

(pgoyette)

If this itesio(4) has a watchdog, then set up a real suspend handler to
make sure we don't suspend when the wdog time is active.

(pgoyette)

Register ipmi(4) with power management subsystem so we might have a chance
of suspending.  Suspending will still be denied if the watchdog is active.

As discussed on tech-kern@

XXX The pmf handlers for this and all other watchdogs should be factored
XXX out into a common handler for a generic wdog(4) pseudo-device, but
XXX that's left for the future.

(pgoyette)

initial import of xcb-proto-1.6

Status:

Vendor Tag: xorg
Release Tags: xcb-proto-1-6

(mrg)

initial import of libxcb-1.6

Status:

Vendor Tag: xorg
Release Tags: libxcb-1-6

(mrg)

initial import of xcb-util-0.3.6

Status:

Vendor Tag: xorg
Release Tags: xcb-util-0-3-6

(mrg)

Note the conversion of the include and csu tests to atf.

(jmmv)

Remove the lib/csu/ and include/ tests; they have been converted to atf.

(jmmv)

Add tests for lib/csu/ and include/.

(jmmv)

Add directories for the lib/csu/ and include/ tests.

(jmmv)

Convert the include tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

(jmmv)

Convert the csu tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

(jmmv)

Remove obsolete __HAVE_GENERIC_SOFT_INTERRUPTS,
which is mandatory since 2007/07:
http://mail-index.NetBSD.org/source-changes/2007/07/14/0039.html

(tsutsui)

Bump MAXPARTITIONS from 8 to 16 using __HAVE_OLD_DISKLABEL compat method
as i386 and hpcmips. PR port-hpcarm/18256

(tsutsui)

Code to set PCI cfg xfer mode registers of 0680A IDE, to make sure the
chip is in PIO mode.
Patch submitted by Toru Nishimura.

(phx)

Fix argument handling of the "-b" option on 64bit platforms. Using
"-b 1024k" would previously fail with this error message:

progress: buffer size 1048576 is greater than -1.

(tron)

this never got bumped properly for libXrender 0.9.6; do so now.

(mrg)

(a) - don't enable KMS support
(b) - fix compile error when XF86_DRMMODE is not set

(mrg)

regenerate these with the right thing.

(mrg)

fix the previous; it wasn't the right one.

(mrg)

remove now-gone neo_bank.c

(mrg)

update for xf86-video-sis 0.10.3

(mrg)

update this for xf86-video-tsent 1.2.4

(mrg)

update for xf86-video-ati 6.13.1

(mrg)

merge xf86-video-sis 0.10.3

(mrg)

merge xf86-video-rendition 4.2.4

(mrg)

merge xf86-video-neomagic 1.2.5

(mrg)

merge xf86-video-ati 6.13.1

(mrg)

initial import of xf86-video-tseng-1.2.4

Status:

Vendor Tag: xorg
Release Tags: xf86-video-tseng-1-2-4

(mrg)

initial import of xf86-video-trident-1.3.4

Status:

Vendor Tag: xorg
Release Tags: xf86-video-trident-1-3-4

(mrg)

initial import of xf86-video-sis-0.10.3

Status:

Vendor Tag: xorg
Release Tags: xf86-video-sis-0-10-3

(mrg)

initial import of xf86-video-rendition-4.2.4

Status:

Vendor Tag: xorg
Release Tags: xf86-video-rendition-4-2-4

(mrg)

initial import of xf86-video-neomagic-1.2.5

Status:

Vendor Tag: xorg
Release Tags: xf86-video-neomagic-1-2-5

(mrg)

initial import of xf86-video-ati-6.13.1

Status:

Vendor Tag: xorg
Release Tags: xf86-video-ati-6-13-1

(mrg)

Add a missing newline in BOOT_DEBUG mode.

(isaki)

Fix build with SYSVBFS_VNOPS_DEBUG.

(njoly)

Small typo in comment.

(njoly)

Ticket 1397.

(riz)

Pull up following revision(s) (requested by dyoung in ticket #1397):
sys/netinet6/udp6_output.c: revision 1.41
Under some circumstances, udp6_output() would call ip6_clearpktopts()
with an uninitialized struct ip6_pktopts on the stack, opt.
ip6_clearpktopts(&opt, ...) could dereference dangling pointers,
leading to memory corruption or a crash.  Now, udp6_output() calls
ip6_clearpktopts(&opt, ...) only if opt was initialized. Thanks to
Clement LECIGNE for reporting this bug.
Fix a potential memory leak: it is udp6_output()'s responsibility
to free its mbuf arguments on error.  In the unlikely event that
sa6_embedscope() failed, udp6_output() would not free its mbuf
arguments.
I will ask for this to be pulled up to -4, -5, and -5-0.

(riz)

Ticket 1397

(riz)

Pull up following revision(s) (requested by dyoung in ticket #1397):
sys/netinet6/udp6_output.c: revision 1.41
Under some circumstances, udp6_output() would call ip6_clearpktopts()
with an uninitialized struct ip6_pktopts on the stack, opt.
ip6_clearpktopts(&opt, ...) could dereference dangling pointers,
leading to memory corruption or a crash.  Now, udp6_output() calls
ip6_clearpktopts(&opt, ...) only if opt was initialized. Thanks to
Clement LECIGNE for reporting this bug.
Fix a potential memory leak: it is udp6_output()'s responsibility
to free its mbuf arguments on error.  In the unlikely event that
sa6_embedscope() failed, udp6_output() would not free its mbuf
arguments.
I will ask for this to be pulled up to -4, -5, and -5-0.

(riz)

Some of the msdosfs tests are killed by SSP due to stack limit
being exceeded.  I cannot figure out what is going on by code
reading, nor repeat this either on my desktop or in qemu, so skip
those tests for msdosfs until I can get to the bottom of it.

(pooka)

Tickets 1412, 1417, 1423, 1424, 1426-1428.

(riz)

Pull up following revision(s) (requested by dyoung in ticket #1428):
sys/netinet6/udp6_output.c: revision 1.41
Under some circumstances, udp6_output() would call ip6_clearpktopts()
with an uninitialized struct ip6_pktopts on the stack, opt.
ip6_clearpktopts(&opt, ...) could dereference dangling pointers,
leading to memory corruption or a crash.  Now, udp6_output() calls
ip6_clearpktopts(&opt, ...) only if opt was initialized. Thanks to
Clement LECIGNE for reporting this bug.
Fix a potential memory leak: it is udp6_output()'s responsibility
to free its mbuf arguments on error.  In the unlikely event that
sa6_embedscope() failed, udp6_output() would not free its mbuf
arguments.
I will ask for this to be pulled up to -4, -5, and -5-0.

(riz)

Tickets 1424, 1426, 1428.

(riz)

Pull up following revision(s) (requested by dyoung in ticket #1428):
sys/netinet6/udp6_output.c: revision 1.41
Under some circumstances, udp6_output() would call ip6_clearpktopts()
with an uninitialized struct ip6_pktopts on the stack, opt.
ip6_clearpktopts(&opt, ...) could dereference dangling pointers,
leading to memory corruption or a crash.  Now, udp6_output() calls
ip6_clearpktopts(&opt, ...) only if opt was initialized. Thanks to
Clement LECIGNE for reporting this bug.
Fix a potential memory leak: it is udp6_output()'s responsibility
to free its mbuf arguments on error.  In the unlikely event that
sa6_embedscope() failed, udp6_output() would not free its mbuf
arguments.
I will ask for this to be pulled up to -4, -5, and -5-0.

(riz)

Pull up following revision(s) (requested by jklos in ticket #1427):
gnu/dist/binutils/gas/config/atof-vax.c
gnu/dist/binutils/gas/atof-generic.c
From Matt Thomas, suggested upstream:
* atof-generic.c (atof_generic): recognize snan and qnan in
addition to nan and inf.
* atof-vax.c (flonum_gen2vax): deal properly with nan, snan, qnan,
and +Inf and -Inf codes.
This fixes PR toolchain/43303.

(riz)

Pull up following revision(s) (requested by christos in ticket #1426):
sys/netsmb/subr_mchain.c: revision 1.19
sys/netsmb/smb_subr.c: revision 1.35
sys/netsmb/smb_subr.h: revision 1.19
sys/netsmb/mchain.h: revision 1.9
sys/netsmb/smb_dev.h: revision 1.7
Convert sizes/lengths to unsigned (size_t) or uint32_t
(for binary compatibility).

(riz)

Pull up following revision(s) (requested by christos in ticket #1426):
sys/netsmb/subr_mchain.c: revision 1.19
sys/netsmb/smb_subr.c: revision 1.35
sys/netsmb/smb_subr.h: revision 1.19
sys/netsmb/mchain.h: revision 1.9
sys/netsmb/smb_dev.h: revision 1.7
Convert sizes/lengths to unsigned (size_t) or uint32_t
(for binary compatibility).

(riz)

Apply patches (requested by chs in ticket #1424):
sys/arch/amd64/amd64/netbsd32_machdep.c:        patch
sys/arch/i386/i386/trap.c:                      patch
Fix several panics that can be caused by applications using
bad segment register values with setcontext() or sigreturn().

(riz)

Apply patches (requested by chs in ticket #1424):
sys/arch/amd64/amd64/netbsd32_machdep.c:        patch
sys/arch/i386/i386/trap.c:                      patch
Fix several panics that can be caused by applications using
bad segment register values with setcontext() or sigreturn().

(riz)

Pull up following revision(s) (requested by joerg in ticket #1423):
lib/libc/stdlib/exit.c: revision 1.12
lib/libc/misc/initfini.c: revision 1.6
Ensure that initfini.c is referenced by exit.c. The start up code has to
reference the latter as a return of main() results in a call to exit(3),
so this ensures that the libc constructors are run for statically linked
programs. Fixes PR 37454.

(riz)

Pull up following revision(s) (requested by jakllsch in ticket #1417):
sys/nfs/nfs_bio.c: revision 1.185
Fix memory leak during some NFS writes.

(riz)

Pull up following revision(s) (requested by phx in ticket #1412):
sys/arch/amiga/dev/clock.c: revision 1.51
sys/arch/amiga/amiga/locore.s: revision 1.149 via patch
sys/arch/amiga/dev/kbd.c: revision 1.53
sys/arch/amiga/amiga/amiga_init.c: revision 1.118
Reverted the CIA-timer based delay() to the pre-5.0 method of a calibrated
delay loop.
This fixes keyboard handshaking problems with some A1200 models since 5.0
and restores the precision for short delays on DraCo systems (the QuickLogic
timer has only a seventh of the CIA precision).
Changed the keyboard handshaking delay from 2000 back to 200ms, although
even the recommended 85ms were successfully tested on the most problematic
A1200 keyboards.
All those changes were tested on an A3000 and A1200 with 68060/50 CPU, and
previously discussed on the port-amiga ML.

(riz)

Add missing pthread_mutex_init call in mutex3 testcase.

(njoly)

Add some sanity checks for statvfs values.

(njoly)

Note that the libpthread tests have been converted to atf.

(jmmv)

The libpthread tests have been converted to atf; remove the old versions.

(jmmv)

Add libpthread tests.

(jmmv)

Add dirs for libpthread tests.

(jmmv)

Convert the libpthread tests to atf.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

I think that this, together with the previous conversion of librt, obsoletes
the tests in the semaphore/ directory.  Will investigate later.

(jmmv)

Fix a wrong arg for callout_reset(9) in txcom_txintr(). PR port-hpcmips/43474

(tsutsui)

Add missing callout_init(9) calls. PR port-hpcmips/43472

(tsutsui)

Add some missing .debug suffix.

(njoly)

skip directory test on sysvbfs

(pooka)

Note conversion of some extra tests to atf.

(jmmv)

Remove tests converted to ATF.

(jmmv)

Add tests for libposix, libprop, librt, libskey and libutil.

(jmmv)

Add new tests dirs.

(jmmv)

Convert the libposix, libprop, librt, libskey and libutil tests to ATF.
Initial work from the GSoC 2008 project by Lukasz Strzygowski.

(jmmv)

Fix typo in comment.  comment tested by wizd.

(pooka)

Fill in PR kern/43626 now that it exists.

(pooka)

Do the famous renamerace test using directories.  Uh oh, bad idea.
PR coming soon.

(pooka)

This test does not always fail for LFS, so apply same kludge as
elsewhere while waiting for atf to grow support for these cases.

(pooka)

Use a kmutex to protect the hash chains and always take this mutex
before removing a node from the hash chain.

Release the hash list lock before calling getnewvnode() and check the
hash list again like other file systems do.

Take v_interlock before calling vget().

(hannken)

Always take the hash list lock before removing a node from the hash chain.

Release the hash list lock before calling getnewvnode() and check the
hash list again like other file systems do.

Take v_interlock before calling vget().

(hannken)

Under some circumstances, udp6_output() would call ip6_clearpktopts()
with an uninitialized struct ip6_pktopts on the stack, opt.
ip6_clearpktopts(&opt, ...) could dereference dangling pointers,
leading to memory corruption or a crash.  Now, udp6_output() calls
ip6_clearpktopts(&opt, ...) only if opt was initialized. Thanks to
Clement LECIGNE for reporting this bug.

Fix a potential memory leak: it is udp6_output()'s responsibility
to free its mbuf arguments on error.  In the unlikely event that
sa6_embedscope() failed, udp6_output() would not free its mbuf
arguments.

I will ask for this to be pulled up to -4, -5, and -5-0.

(dyoung)

With Xen, PDPpaddr should contain a guest physical address (== PFN).

(jym)

pte0 is used to map the page required by bioscall() trampoline code, so
add VM_PROT_EXECUTE protection for it.

Currently, this does not change much, as page execute permission is not
enforced for i386 (except for Xen, where BIOS calls are not used anyway).

(jym)

Document the format of the system crontab file, since cron(8) says that we do.

(christos)

f_namemax is one of the static fields overridden by copy_statvfs_info(),
so be sure to set it to the value coming from the file server as
part of mount args.

exposed, like so many other problems, by njoly's tests

(pooka)

Default f_namemax to MAXNAMLEN, mostly for initial statvfs struct
passed to kernel in mount.

XXX: there's currently no way for a file server to override the
one passed in mount.

(pooka)

Check the virtual address 'va' for each PDIR_SLOT_PTE entry. PDP_SIZE
is 4 with PAE (Xen only currently), 1 otherwise: loop should be unrolled
when PDP_SIZE is 1.

pmap_alloc_level() is used by pmap_growkernel(), the PDE is a kernel
mapping: mark it so with PG_k. While here, use pmap_pa2pte() for physical
address 'pa'.

No functional change.

(jym)

PR/43623: Takahiro Kambe: Allow /etc/crontab processing if it is not writable,
don't insist that it is mode 0600.

(christos)

remove bad cast.

(christos)

To help find the cause of kernel complaints such as "/netbsd:
nd6_storelladdr: sdl_alen == 0, dst=... if=wm1", add printfs for some
"impossible" conditions, and make the nd6_storelladdr() printf more
informative by printing the value of sdl_alen.

(dyoung)

Make the comment about PDPpaddr more thorough.

(jym)

Purely cosmetic: move L2_SLOT_KERNBASE out of the PAE #ifdef/#endif block.

(jym)

In Xen PAE case, fix argument size passed to init386(), by pushing the
upper bits onto stack (we do not expect first_avail to be above 4GiB, so
assume their value is 0).

Remove macros (PROC0PDIR and PROC0STACK) that were never used.

(jym)

some recent changes

(pooka)

s/PG_DIRECT/PG_XIP/

(uebayasi)

Rename PG_DIRECT to PG_XIP.  PG_XIP is marked to XIP vnode pages.

(uebayasi)

Regen for libarchive 2.8.4

(joerg)

Merge libarchive 2.8.4

(joerg)

Import libarchive 2.8.4:
Symlink dereference fix for Linux broke the build there; corrected.
Fix issues on ancient FreeBSD, QNX, ancient NetBSD and Minix.
Improved reliability of hash function detection.

Status:

Vendor Tag: KIENTZLE
Release Tags: libarchive-2-8-4

(joerg)

Add the new kernel/t_sigaction test program.

(jmmv)

Add a couple of simple tests for sigaction(2): one for the case where flags
is set to 0 and the other for the case where we set SA_RESETHAND.

The latter test is intended to expose PR port-macppc/43619 and is marked as
an expected failure on such platform.

(jmmv)

xfail test on lfs.  It goes badaboom faster than you can find your
multipass.  Borrow PR kern/43582 used earlier for rmdirrace, as it
looks pretty much like the same problem.

(pooka)

start converting renameraces to vfs

(pooka)

Convert "The Original" rename race test from to vfs and retire the
ffs/tmpfs versions.  The only difference is that the origamical
one mounted ffs with MNT_LOG (and therein actually lay the bug).

(pooka)