Pull up following revision(s) (requested by rmind in ticket #1241):
sys/kern/kern_softint.c: revision 1.30
softint_execute: release/re-acquire kernel-lock depending on SOFTINT_MPSAFE
flag.  Keeping it held for MP-safe cases break the lock order assumptions.
Per discussion with <martin>.

(bouyer)

Pull up following revision(s) (requested by rmind in ticket #1240):
sys/kern/kern_runq.c: revision 1.28
sched_catchlwp: fix the case when other CPU might see curlwp->l_cpu != curcpu()
while LWP is finishing context switch.  Should fix PR/42539, tested by martin@.

(bouyer)

Pull up following revision(s) (requested by rmind in ticket #1240):
sys/kern/kern_runq.c: revision 1.28
sched_catchlwp: fix the case when other CPU might see curlwp->l_cpu != curcpu()
while LWP is finishing context switch.  Should fix PR/42539, tested by martin@.

(bouyer)

Pull up following revision(s) (requested by mrg in ticket #1239):
external/mit/xorg/lib/dri/i965/Makefile: revisions 1.7, 1.8
external/mit/xorg/lib/dri/i915/Makefile: revisions 1.5, 1.6
replace LIBS with LIBDPLIBS.  makes these actually link against drm_intel.
should fix a problem reported by snj@.
link in libOSMesa as well.

(bouyer)

Pull up following revision(s) (requested by rmind in ticket #1238):
sys/kern/kern_sig.c: revision 1.302
sigactsunshare(): set reference count in a case of new sigacts allocation.
Bug (e.g. memory leak) can happen when using clone(2) call.

(bouyer)

Pull up following revision(s) (requested by rmind in ticket #1238):
sys/kern/kern_sig.c: revision 1.302
sigactsunshare(): set reference count in a case of new sigacts allocation.
Bug (e.g. memory leak) can happen when using clone(2) call.

(bouyer)

Apply patch, requested by snj in ticket #1215:
        gnu/usr.bin/xcvs/cvsbug/Makefile: patch
Properly replace @MKTEMP_SH_FUNCTION@, @MKTEMP@, and @SENDMAIL@.
Fixes PR bin/39629.

(bouyer)

Revert previous, wrong branch

(bouyer)

Apply patch, requested by snj in ticket #1215:
gnu/usr.bin/xcvs/cvsbug/Makefile: patch
Properly replace @MKTEMP_SH_FUNCTION@, @MKTEMP@, and @SENDMAIL@.
Fixes PR bin/39629.

(bouyer)

Fix leak of a usbd_xfer_handle when a interrupt pipe is aborted then
closed:
usbd_open_pipe_intr() allocates a usbd_xfer_handle for pipe->intrxfer.
Most usb device drivers using interrupt pipes call usbd_abort_pipe()
then usbd_close_pipe(), usbd_close_pipe() is supposed to free pipe->intrxfer.
But usbd_abort_pipe() calls [uoe]hci_device_intr_abort() which,
if the xfer aborted is pipe's intrxfer, sets pipe->intrxfer to NULL.
So usbd_close_pipe() can't free it and the usbd_xfer_handle is lost.

To fix this, in usbd_abort_pipe() remember the pipe->intrxfer's value
on entrie, and if it's different after usbd_ar_pipe(), call
usbd_free_xfer with the original value.
Confirmed to fix the memory leak on close() with umodem(4) and
uplcom(4).

(bouyer)

Typo

(roy)

Remove tn3270 from sets

(roy)

Document tn3270 removal.

(roy)

Remove tn3270 from base system.
If anyone really needs it then it can be put into pkgsrc.
Importing c3270 into pkgsrc is probably a better option though.

(roy)

Add a couple of comments and some whitespace.

(skrll)

Tweak comment. From OpenBSD.

(skrll)

Simplify

(skrll)

Tweak astpending handling.

(skrll)

Remove cargo cult dlinfo entry point.

(skrll)

Use the description in the ACD for T_DBREAK.

(skrll)

fix wording:
instead of the a file -> instead of a file

(mbalmer)

Check cpl level has returned to entry value when leaving trap/syscall.

From OpenBSD.

(skrll)

Add a comment.

(skrll)

Count interrupts in uvmexp.

(skrll)

Rename type_raw to trapnum and use in a couple more places.

Reduces diff to OpenBSD.

(skrll)

Count traps in uvmexp. Reduces diff to OpenBSD.

(skrll)

UltraSPARC-IIIi CPUs select the BUSY/NACK pair in the interrupt
vector dispatch status register based on the lower two bits of the
ITID.

From OpenBSD and details are described in the section 10.4.3 of
UltraSPARC IIIi Processor User's Manual.

(nakayama)

Add processor implementation codes, from OpenBSD.

(nakayama)

Use __arraycount().

(dyoung)

* remove just-for-kicks locking
* KNF
* remove outdated comment (quite a funny one to read in 2010, though)

(pooka)

Note sparc apc(4).

(bouyer)

Fixed crash, when /rtas was not found: OF_finddevice() returns -1 when
the device is not found.

(phx)

Add apc(4), a driver for the Aurora Personality Chip (APC) found
on SPARCstation-4/5, and emulated by qemu to idle the simulator
when the CPU is idle. Infos about the registers from the linux driver.
Not enabled by default because it can cause some Sparc systems to
hang (so says the linux driver).
Only the CPU idle part implemented at this time; fan speed and
Convenience power outlet management to be added.
Tested on qemu.

(bouyer)

Accomplish messy stack with slightly less messy code.

(pooka)

Amazingly, we've been freeing a handle and then using that which it referred
to for years. Bad.

(mjacob)

vclean() actually sets v_tag to VT_NON but doesn't touch v_type.
getcleanvnode() sets v_type to VNON after releasing v_interlock.
So the thread doing quotaon(), quotaoff() or qsync() could vget()
a vnode which is being recycled in getcleanvnode(), after is has
been cleaned and v_interlock released, but before v_type has been
reset, leading to KASSERT(vp->v_usecount == 1) firing in
getnewvnode(), or qsync() dereferending a NULL pointer as in
PR kern/42205.
Fix by using the same tests as other ffs function traversing the mount
list: also check for VTOI(vp) == NULL, and VI_XLOCK in addition
to VI_CLEAN.

(bouyer)

Use .%U instead of .%O for URLs.

(joerg)

Use .%U instead of .%O for URLs.

(joerg)

Revert previous. The KASSERT() is right and my analysis is wrong,
as pointed out by pooka@.

(bouyer)

Use .%U instead of .%O for URLs.

(joerg)

Use .%U. Give full URL for draft-ietf-ipsec-ecn-02.

(joerg)

Use .%U for URLs instead of .%O.

(joerg)

Use .%U instead of .%O for URLs.

(joerg)

Use subr_percpu.c instead of homegrown implementation.  ...except
when using malloc(3)-relegated allocators (happens in production
e.g. on Linux), since subr_percpu.c uses vmem and i don't want to
reimplement vmem.

(pooka)

print mbuf statistics a few times.  incidentally, can be used to
test percpu.

(pooka)

Add function names defined by macro in comment for reference.

(nakayama)

Fix typo in previous.

(nakayama)

Regenerate.

(asau)

Test for <sys/endian.h> and provide HAVE_SYS_ENDIAN_H.
Fixes FreeBSD hosted builds (PR 42440 and PR 42470).

Approved by <dholland>.

(asau)

Put the dlopen,and friends entry points back.

They're needed by pre-2.0 binaries. Backward compatibiliity for these
was broken by another commit recently which I'll fix shortly.

(skrll)

Get rid of most of the studly caps.
First to MULTIPROCESSOR support for pmap: move some pmap globals to cpu_info.
Define per-cpu ASID number spaces.
Remove some mips1/!mips1 difference in db_interface.c
Add mips32/64 knowledge to stacktrace.

(matt)

Revert ticket 1218.

(snj)

Fix reference counting for vfsops in mount.  Otherwise it's possible
(for an unprivileged user) to force vfs modules to remain loaded
forever.  Also, it's possible for an admin with fat fingers to have
to curse out loud (a lot) and reboot.

.. or at least fix things as much as seems to be possible without
involving 1000 zorkmids.  do_sys_mount() takes either struct vfsops
(which hopefully came properly referenced) or a userspace string
for file system type.  The standard in-kernel calling convention
of "do_sys_mount(l, vfs_getopsbyname("nfs"), NULL," is not to be
considered healthy, kosher, or even tasty (although if vfs_getopsbyname()
fails the whole thing *currently* fails without the program counter
pointing to hyperspace).

(pooka)

Remove KASSERT(vp->v_usecount == 1) in getnewvnode() and ungetnewvnode().
Another process could be vget()ing the vnode and bump v_usecount while
getcleanvnode() is vclean()ing it (as vclean drops the interlock).
vget() will then wait for VI_XLOCK or VI_FREEING to clear; and we could test
this assertion while the other process is still slepping. We could even
end up in ungetnewvnode() before this other process got a chance to run.

(bouyer)

from main with return 0;
(Why did I manage to succesfully build n times, including a full
distribution build, without tripping this before ?!?)

(pooka)

Remove the entry points for dlopen() and friends. They haven't been used
since Aug 2003.

(skrll)

Remove puffs| from vfstype before making comparison to determine if
fs is mounted.

Fixes slightly-miscategorized kern/37626.

(pooka)

undef and redefine instead of checking if it is define, from joerg

(christos)

add mount_puffs

(pooka)

If getargs and vfs has PUFFS_TYPEPREFIX, use mount_puffs to query args.

(pooka)

Descend into mount_puffs.

(pooka)

provide __arraycount for those that don't have it.

(christos)

Add a generic puffs kernel utility for querying mount arguments.
E.g.:
golem> dtfs -r 'chr 15 25' dtfs /puffs
golem> mount_puffs -o getargs dtfs /puffs
version=26, flags=0x10, root cookie=0xbb90c0c0, root type=VCHR, root size=0, root rdev=0xf19

(pooka)

In case the operations thread has exited, do not queue any more
operations.  This prevents kernel memory leaks (one of which happened
every time the file system was unmounted via PUFFSOP_UNMOUNT ...
and incidentally would've been trivially caught with the old
malloc(9) interface.  I wonder if the message is to use a ton of
pools instead of regression-attractive kmem interface).

(pooka)

Fixes the rx stall problem on 82578 by MANY workaround code.
We need more work for 82577.

(msaitoh)

On i386/amd64, define two flag bits in the boot parameters which control
whether modules are loaded and whether boot.cfg is evaluated, and
set both to "off" per default in the PXE bootloader.
Extend "installboot" to toggle the bits.
This way, pxeboot works with existing dhcp server setups (and as
described in the manpage) out of the box. Also, boot.cfg reading
involves a stat() call which is horribly inefficient with the
TFTP pseudo file system.

(drochner)

Simplify mips_page_physload.

(matt)

Add a list of functions where we usually don't care about their return
code. Can be bypassed by -hh

(christos)

Remove if defined(__RCSID) around __RCSID() macro.
#include "nbtool_config.h" is enough for tools.

(tsutsui)

Since VOP_GETATTR() does not require a locked vnode, resolve and
reference the puffs_node before sending the request to the file
server.  This diminishes the window where the inode can be reclaimed
and be invalidated before it is accessed (but does not completely
eliminate the race, as that is a caller problem which we cannot
fix here).

(pooka)

More consistency with obsolete tag:
it should be only be accompanied by gcc={3,4} or {xorg,x11}, and only
if they are necessary, i.e. if the same line exists for the other setting
of gcc or xorg/x11.
Make it so.

(wiz)

In addition to setting module=no if ${MKKMOD} is "no", also
set kmod=no.  This should get evbppc building correctly again.

(he)

Make DEBUG build compile.

(skrll)

Make DEBUG build compile.

(skrll)

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

(wiz)

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

(wiz)

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

(wiz)

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

(wiz)

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

(wiz)

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

(wiz)

Wrap a few long lines.

(skrll)

Shut gcc up.

(skrll)

Regen

(sborrill)

Fix typo in (unused) vendor name

(sborrill)

s/PCI_PRODUCT_ALI_M1543/PCI_PRODUCT_ALI_M1533/ now that dsl@ has fixed
pcidevs.

Hi dsl!

(skrll)

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

(wiz)

Make ums.c work with a "Microsoft Natural Ergonomic Desktop 7000" mouse.
XXX The tilt function still doesn't work for me.

(matthias)

Add cas(4).

(jdc)

in BS(unmap), instead of assuming KSEG2 bsh address came from uvm_km_alloc,
check that the address is NOT (KSEG0 or KSEG1) address.

(cliff)

set a delta value on the master channel and move it to the outputs class so
volume control via PMF works

(macallan)

set a delta value for the master channel so volume control via PMF works

(macallan)

actually assign audio control command codes to the audio control keys

(macallan)

Move AMD64's bswap64 implementation from libc to src/common and share it
with the kernel.

(joerg)

Provide inline assembly version of bswap64.

(joerg)

more avail_end/avail_start changes

(matt)

More avail_end/avail_start fallout.

(matt)

More fixes for the CFATTAL_DECL_NEW changes and rmixl cpucore/cpu changes.

(matt)

Deal with various changes with avail_end and avail_start, etc.

(matt)

device_t, CFATTACH_DECL_NEW, ansify, ...

(matt)

Revert code I added in 1.133 that touches the PCI_COMMAND_IO_ENABLE and
PCI_COMMAND_MEM_ENABLE bits in PCI_COMMAND_STATUS_REG.  This causes
problems where the BAR is 0.  'options PCI_ADDR_FIXUP' is more
appropriate when these bits are not set correctly by the BIOS.

(jakllsch)

Shut gcc up.

(skrll)

Deal with new cpu_identity()

(matt)

use RTLD_NOW now that it works.

(christos)

PR/39240: Satoshi Suetake: Don't fail when attempting to resolve weak symbols
when we are doing immediate binding, leave them alone and they will be dealt
with later during lazy binding. From skrll@

(christos)

Use __arraycount instead of locally defined macro.

(njoly)

Fix tab/space for www changes page.

(tsutsui)

Bump the ramdisk size so that the contents fits again.
OK'ed by reinoud@

(he)

recognize SVM PauseFilter

(cegger)

Explicitly clear AUXIO4M_FTC bit in FTC_FLIP macro used on
pseudo-dma for floppy, as well as AUXIO4C_FTC bit for sun4c.
A comment in the macro says AUXIO4M_FTC bit is auto-clear,
but my two SS20s (including compatible) with 150MHz hyperSPARCs
get data_overrun without it, and no bad side effect on SS5
(works with and without this change).

Closes PR port-sparc/42516, which is the last one of a bunch of
floppy issue on NetBSD/sparc since NetBSD 2.0 days. See
http://mail-index.NetBSD.org/port-sparc/2009/12/20/msg000484.html
for details.

(tsutsui)

fix some bitmask in PCI Express capabilities decoding
being here, fix some typos and add a comma for clarity

(drochner)

RFC783 says the ErrMsg sent with an ERROR packet needs to be
NUL-terminated

(drochner)

Add dhcp support to NetBSD/hp700 install stuff.

(skrll)

- clean up formatting in print functions

(cliff)

- don't match if attach arg name doesn;t match cf_name

(cliff)

- nuke mainbussearch()
- use config_found (vs. config_search_ia) for cpucore's and obio
- print L2 cache attributes

(cliff)

- spec CIDFL's for RMI L2, cores, threads attributes
- add cputab[] entry for RMI XLS404LITE
- cpu_identify() now gets device_t arg; use it instead of static 'label'

(cliff)

- cpu_identify() now gets device_t arg
- add CIDFL's for RMI L2, cores, threads attributes

(cliff)

- attach at cpucore
- match multiple cpus (RMI "threads")
- print of cpu freq and call to cpu_identify() moved to parent

(cliff)

- mainbus gets 'core' locator, loses (unused) 'addr' locator
- attach cpucore at mainbus
- attach cpu at cpucore

(cliff)

- attach cpucore at mainbus
- attach cpu at cpucore

(cliff)

- add driver to represent processor core(s) in device tree

(cliff)

- add autoconf.h, need mainbus_atach_args

(cliff)

Avoid adding -Os to CFLAGS if DBG is set and contains -g.
This helps gdb a lot when compiling local debug versions of sysinst.

(martin)

Use the proper structs for ioctls to fetch network configuration.
Might fix PR 42436.

(martin)

revert brainfart

(macallan)

need object directory for etc/mtree to run distrib-dirs

(pooka)

In rumptest, create obj dirs before distrib dirs.  Fail if distrib dir
creation fails.

per clue-by-4 from mrg

(pooka)

replace local ci_cpuname[] and use new cpu_name().

(mrg)

introduce a new function that returns a unique string for each cpu:

char *cpu_name(struct cpu_info *);

and use it when setting up the runq event counters, avoiding an 8 byte
kmem(4) allocation for each cpu.  there are more places the cpuname is
used that can be converted to using this new interface, but that can
and will be done as future work.

as discussed with rmind.

(mrg)

Minimize unnecessary differences in rump.

(pooka)

Use a common cpu.c for evbmips single cpu machines

(matt)

./usr/libdata/ldscripts/kmodule will not be installed if MKKMOD=no, so
add 'kmod' to its keywords.

(dyoung)

Add a few symbols used by the tty code.

(pooka)

cgd is attached via the module interface now instead of pseudodevs.

(pooka)

If RUMP_VERBOSE is set in the env, set boothowto to AB_VERBOSE.

(pooka)

belated entry for 20100108

(lukem)

Provide a more complete modules hook, cgd_modcmd(), derived from
vnd_modcmd().

(dyoung)

- Add i82567LM-2 i82567LM-4 i82567V-3 LAN controller.
- Reset GMII interface after wm_reset() in wm_init().
- Rework for assigning mii_{read,write}reg(). Use PCI product ID to identify
  the PHY.
- Add code about LPLU(Low Power Link Up) function. Now we can linkup 1000BaseT
  on PCH. It seems that we have to do the same work for ICH9.

(msaitoh)

Revert 1.194 rev.

(rmind)

Add some macros

(msaitoh)

regen

(msaitoh)

add i82567LM-2 i82567LM-4 i82567V-3 LAN controller

(msaitoh)

rename some wm devices to know PHY type

(msaitoh)

Make cgd(4) into a detachable pseudo-device.  Thanks, Jan Danielsson,
for the patches!

I've lightly tested the basics: create cgd0 on vnd0d, initialize with
zeros, newfs /dev/cgd0a, mount, copy files on, unmount, drvctl -d
cgd0.  Works fine.  I also shutdown with a cgd0 configured: detached as
expected.

(dyoung)

Fix incorrect Acer Labs ID.  PCI_PRODUCT_ALI_M1543 had been incorrectly
identified as product code 0x1533, which is what is actually in the
CS20.  PCI_PRODUCT_ALI_M1543 was corrected recently and sio.c would fail
to match the bridge.

(mhitch)

- call the mapping function directly instead of el_wset().
- save the strings passed to the mapping function so that they don't get
  re-used. This leaks. To fix it properly we could either pass a flag to
  free particular entries before re-using, or allocate all of them.
  Allocating all of them wastes memory, allocating some of them makes
  the code more complex.
This fixes compatibility binding (shell tab completion for example)

(christos)

- in the argv conversion, handle NULL as NULL
- when printing tab/nl print them, don't handle them specially.

(christos)

Make r/w servers unmount themselves cleanly upon receiving SIGINT/TERM.

(pooka)

Add convenience routine puffs_unmountonsignal(), which does exactly that.

(pooka)

Change some devices to use CFATTACH_DECL_NEW/device_t/cfdata_t/device_*

(matt)

send PMF events for volume control keys

(macallan)

Adjust report descriptor for Logitech USB Receiver M/N C-BT44 to
more closely match what is in the report.  (This corrects the
off-by-one usages of most of the gray multimedia keys on a EX110
wireless keyboard so they can sensibly be used with libusbhid(3),
usbhidctl(1) and usbhidaction(1).)

This device also uses officially reserved usages (in the approximate
range of 0x1000 to 0x1100) in the Consumer usage page, but these
are out of the logical range the device reports as being valid.

(jakllsch)

Avoid treating the Griffin PowerMate knob as a ums(4).
Should fix PR/19445.

(jakllsch)

merge conflicting error sections.

(christos)

Split device_t/softc, some KNF.

(tsutsui)

Switch to using todr_[gs]ettime_ymdhms() format
which doesn't require struct timeval details.

(tsutsui)

regen.

(jakllsch)

Add Logitech C-BT44 Receiver.

(jakllsch)

Import support for upa64s device (/upa on SB1000/2000) from OpenBSD.
Attach ffb both at mainbus and at upa.

(martin)

error message cleanup
- 1 -> EXIT_FAILURE
- fprintf(stderr, -> warnx(
- better warning messages

(christos)

- 1 -> EXIT_FAILURE
- avoid assertion firing when hitting ^D in CC: line.

(christos)