sync with head.

(yamt)

- simplify ASSERT_SLEEPABLE.
- move it from proc.h to systm.h.
- add some more checks.
- make it a little more lkm friendly.

(yamt)

move schedppq from sched.h to param.h because it's used by some of
PRI_ constants.

(yamt)

Make schedintr interrupt handler per-CPU.
While there rename tickintr interrupt handler and share initialization
code with schedintr.

(nakayama)

Update the man page to note the additional cards that this driver will
match.  Most of them are experimental.

(elric)

Moved USB_PRODUCT_NOVATEL2_MERLINV620 into the UNTESTED section as the
man page indicated that it is untested.  Fixed some whitespace issues.

(elric)

Made a number of modifications, including:

1.  fixed a bug where the last bulk in/out was chosen rather
    than the first.  This got my Sierra AC 595 working,

2.  reworked some of the match logic to include flags for the
    matches,

3.  defined a flag UNTESTED which causes a warning to be output,

4.  added product/vendor IDs from other OSes (OpenBSD, FreeBSD,
    Linux) but marked them UNTESTED.

(elric)

Add a number of products which will be supported in ugensa(4).
Most of these are experimental and the product/vendor IDs are
snarfed from OpenBSD, FreeBSD and Linux.

(elric)

Make this compile after device_t/softc split.

(cube)

Fix fallout from softc split.

(cube)

Fix fallout from softc split.

(cube)

Process group and session structures are allocated so infrequently that
there is no point having seperate caches for them. Just use kmem_alloc().

(ad)

- Initialize uihashtbl in resource_init();
- Make some variables static, remove the externs from header;
- Wrap few long lines, misc;

No functional changes are intended.

(rmind)

arint_normal -> aprint_normal

(yamt)

Remove unused file.

(lukem)

we always know the bdaddr and type of a link when creating it, so
pass them to the _alloc() function and let it set them.

(plunky)

errexit() now provides the trailing \n (since fsck_ffs assumed that
from a conversion from err(3)), so "make it so".

(lukem)

insert new links at the tail of the queue so that if a create_connection
command fails to start we can find the relevant link, since it will be
the first one with the pending flag set.

(plunky)

Workaround the case, when l_cpu changes to l_target_cpu, and causes
the locking against oneself. Will be revisited. OK by <ad>.

(rmind)

sync from fsck_ffs/dir.c 1.21:
lost+found directories are mode -1700

(lukem)

Fix compilation.  Noted by yamt@.

(cube)

+CFG (context-free grammar, control-flow graph)

(dholland)

Always deviter_release().

(dyoung)

Patterned after the other ports, define _MACHINE and MACHINE before
including <arm/arm32/param.h>.  Fixes build problem for rump.

(he)

Only SH7750 and SH7750S require updates to memory-mapped data cache
arrays to be performed while running on P2.  Don't penalize other cpus
that can do it from P1.

(uwe)

Rewrite RUN_P2, RUN_P1 in asm.  gcc4 is too happy to optimize away
assigned-goto versions if try to do anything fancy with them
(discovered by trying to wrap RUN_P2 into an "if").

(uwe)

Remove unnecessary comment-outed call.

(kiyohara)

Remove TEXTADDR.  We start to 0x100000.

(kiyohara)

Adapt to device_parent() taking a device_t, not a struct device *.

(he)

Don't assume a struct device is embedded in the softc.

(cube)

fix compilation

(yamt)

- wrap bintime with !defined(_STANDALONE) as suggested by Valeriy E. Ushakov.
- add a comment.

(yamt)

Describe how to portable extend a file. From joerg. Bump date.

(wiz)

Various improvements (mdoc, wording, punctuation).

(wiz)

Add support for optional Z85C30 serial console on Cobalt Qube 2700.
Bump version.

(tsutsui)

Xref cac(4).

(xtraeme)

Mention cac(4), now that it has basic bio(4) support.

(xtraeme)

Add more wait after sending a setup packet since
21041 on Qube 2700 seems to require more setup time.

(tsutsui)

Obtain RELEASEMACHINEDIR from the make environment.
Display the path to the binary set dir upon completion of "sets".

(lukem)

show RELEASEMACHINEDIR in params

(lukem)

If we bother to report number of wired TLB entries we might as well do
it correctly.

(uwe)

Add '-q' to maketars and makesrctars, to quieten them.
Improve MAKEVERBOSE support.

(lukem)

Cleanup .tgz.tmp files on exit.

(lukem)

Split device_t and softc for the ... wait for it ... [air quotes] MI
variant of fdc(4)/fd(4).  Other cosmetic changes as appropriate.

(cube)

Split device_t and softc for isapnp(4).  Use proper types, ansify when
appropriate.

(cube)

Split device_t and softc for opl(4).  That should have been done along with
midi(4), because opl_softc embed a midi_softc.
Fixes PR#38242 by Onno van der Linden.

(cube)

defflag cache related options.

(uwe)

Split device_t and softc for sb(4) and all its attachments.

(cube)

s/prototyoe/prototype/

(bjs)

Use macros to specify a interrupt level.

(nakayama)

Update alignment comment.

(christos)

Use proper types.

(cube)

Split device_t and softc for pckbd(4) and pms(4).

(cube)

After a process in swapped in, refresh the pcb's paddr.

(matt)

Add support for the new GENERIC.MP kernel set

(martin)

Update supported hardware, add new kernel sets.

(martin)

Add -momit-leaf-frame-pointer, saves a few bytes.

(dsl)

Split device_t and softc for isa(4).

(cube)

Set scope on IPv6 multicast address to give carp a chance to work for IPv6, too.
>From FreeBSD.

(ws)

Swap red and blue for 8bpp paletted mode.

(jmcneill)

Add a 'maxcpus' global.

(ad)

- Special case mounting a devfs from init(8) so that a devfs cannot be
  unmounted unless it is forced. This is a (possibly temporary) workaround
  to the problem that the /etc/rc.d/root script umounts all file systems
  before mounting /.

- Clean up devfs_create_node() based on feedback from wrstuden@.

- Add devfs_remove_node() (counterpart to devfs_create_node).

- Add function to create a console and dctl(4) device node for every
  devfs mount.

- Support for VOP_MKNOD() on a devfs file system has been dropped, so
  that devfsd(8) can fully control device nodes that are created in devfs.

- Allow devfs to be mounted -o union.

(mjf)

Split device_t and softc for pckbc(4) and its attachments.

(cube)

strcmp doesn't taken a length argument.

(skrll)

whitespace '\t' -> ' '

(rtr)

remove my old (invalid) email address

(plunky)

remove old email address

(plunky)

ARM/Thumb interworking support

(rearnsha)

Allow detection of VFP devices on ARM11 cores.

(rearnsha)

Allow use of the VFP.

(rearnsha)

VFP support.

(rearnsha)

VFP support.

(rearnsha)

VFP support.

(rearnsha)

VFP support.

(rearnsha)

VFP support.

(rearnsha)

Wrap mips3_cp0_count_read() stuff (for DECstation 5000/150) with
#if defined(MIPS3) so that kernel config without options MIPS3 compiles.
Problem reported by Erik Bertelsen on port-pmax.

(tsutsui)

Make sure M_PROMISC isn't already set, before we need to see if we are going
to see if we need to set M_PROMISC.
Assume the interface is not CARP'ed.

(matt)

Add "mach reset" command that resets machine by taking a trap with
exceptions disabled.  Fix db_stackcheck_cmd (have_addr is now bool).
Supply help strings for mach commands.

(uwe)

Switch dz driver to PRIVALLOC and device_* accessors.

(matt)

This is not (has never?) being used.

(matt)

Switch to PRIVALLOC and device_* accessors

(matt)

Use 0 sized softc and switch to PRIVALLOC

(matt)

Fix device access.  Use a 0 sized softc and switch to PRIVALLOC for mem_cmi.

(matt)

Use device_*

(matt)

Use a 0 sized softc / switch to PRIVALLOC.

(matt)

Cosmetic: make rtw_hw_keysupp() return a bool instead of int.  In
rtw_wep_setkeys(), add a debug printf, and move the initialization
of keylen from the for-loop to the declaration.

(dyoung)

Comment out xmi and nmi devices.

(matt)

Use device_* accessors.

(jkunz)

Extract subroutine rtw_tx_kick() from rtw_start(), and reuse it in
rtw_io_enable() to restart the transmitter after resetting it.

(dyoung)

In vga_is_console(), use deviter_first/_next() and device_t accessors.
Ok cube@.

(dyoung)

Change tests for "impossible" keys to KASSERT()s.  Get rid of a
temporary variable.  In rtw_key_delete(), do not invalidate the
hardware keys unless a WEP key is deleted.

(dyoung)

Fix compilation under options RND_COM and COM_DEBUG.
Reported by Kurt Schreiner on current-users@.

(cube)

Add a strchr() - for dosfs in particular.
The i386 asm version in src/common is rather larger than this one.

(dsl)

Add a local strcasecmp() instead of pulling the bloat from libkern.

(dsl)

Split device_t and softc for all attachments of vga(4).

(cube)

Fix falout from recent VAX code rototill:
MicroVAX 4000/[3-7]00 could not get HW Ethernet address.

(jkunz)

Split device_t and softc for irframe(4).

(cube)

if_ze.c

(jkunz)

Re-run mknative.

(skrll)

Fix spl(9) botch in cpu_intr() on ews4800mips:
Don't enable unhandled interrupts before all interrupts are processed.

(tsutsui)

Fix spl(9) botch in cpu_intr() on arc:
Don't enable unhandled interrupts before all interrupts are processed,
and also change all interrupt handlers to return processed MIPS_INT_MASK
values, rather than masked values of them.

(tsutsui)

SMP support for sparc64 now works, so make GENERIC.MP kernel build
by default.

(nakayama)

SMP support for sparc64 now works on some models.

(nakayama)

Pull in some updates for NetBSD/sh gdb from wrstuden-fixsa.

- support siginfo trampoline detection/unwind
- modernise wrt core file regsets
- use inf-ptrace

(skrll)

#include "opt_multiprocessor.h"

(nakayama)

Complete per-CPU TSBs and MMU contexts changes.

(nakayama)

Provide a to_pid_to_exec_file

(skrll)

Improve FPU state save/clear like x86, idea from OpenBSD.

(nakayama)

On SMP kernel,
- use "tick" interrupt for per-CPU statclock.
- disable "counter-timer" #1 interrupt and use it as timecounter.

(nakayama)

Enable ETHERCAP_VLAN_MTU.

(tsutsui)

Improve IPI handling:
- make IPI takes two arguments.
- add IPI event counters per-CPU.
- implement IPI functions which were missing or broken.
- insert DELAY while halting primary CPU in IPI handler.

(nakayama)

Split device_t and softc for all com(4) devices (well, everything that
uses a com_softc backend).  Use proper types and ansify where appropriate.

(cube)

Don't fetch DMA descriptor after BUS_DMASYNC_PREREAD op.

(tsutsui)

Split device_t and softc, catching up with changes of all other ne(4)
variants.

(cube)

If a cd(4) is attaching on ata instead of scsi, the controller is
the grandparent, not the parent.

(martin)

Remove a tty_lock for now - I know that this is wrong, but the whole
tty subsystem locking is not consistent and Andrew didn't have
time to come up (or guide to) the right fix.
This prevents machines with serial console on zs(4) from deadlocking on
first input.

(martin)

Initial bio(4) support; only volume status is handled.  The driver doesn't
yet know about physical drives.  Derived from the OpenBSD ciss(4) bio(4)
support.

(mhitch)

make compile
  - add xenbus_read_ull prototype into include/xenbus.h

(ichiro)

Add swedge support to xbd backed, and make the virtual block device handle
sizes larger than 4TB. Tested by Jukka Marin.
Should fix kern/37370.

(bouyer)

Implement a C version of splx() in evtchn_do_event(). using plain splx()
here will reenable interrupts as a side effect, and we don't want it here.

It could cause some event handlers to run twice (which should be harmless),
and trap() to be called on the wrong LWP in doreti_checkast (which can
probably cause some damage).

(bouyer)

Modernise NetBSD/alpha gdb - taken from gdb-current.

Add support for siginfo trampolines as well as sigcontext.

(skrll)

Modernise NetBSD/alpha gdb - taken from gdb-current.

Add support for siginfo trampolines as well as sigcontext.

(skrll)

if "ddb_vgapost" is called but its data structures were not initialized,
print a message, suggested by joerg
(the message might be invisible if the console is in X mode)

(drochner)

Reflect change of PMF_FN_ARGS.

(taca)

Since _file is a short, check that the fd fits in it, otherwise bail with
EMFILE. We treat _file as an unsigned short to double our range, with a
special case for -1 (closed). Make a note of what we should do about stdio
if we ever bump libc. We could change _file in the future compatibly to an
int by putting it in the extension space but for now we don't bother.

(christos)

Cache referenced rootvp in struct ukfs.  Makes it possible to run
pre/postcall wrapper around the exported namei, since postcall no
longer needs to execute VFS_ROOT (which is a bad thing to do while
potentially holding on to a locked vnode, mmkay).

(pooka)

Fix download-vulnerability-list's -c option. From Jim Bernard in PR 38231.

(wiz)

Don't include i810 hack if DRM_NO_AGP is defined, as it results in
an undefined symbol (agp_i810_borrow) at kernel link time.

(bjs)

-add a function to vga_post which can be called from DDB to get a
usable VGA console ("call ddb_vgapost")
-allow to switch from/to screens occupied by an X server if the graphics
device is console and in polling mode (ie DDB)
This together allows to get a DDB session on a VGA console if the
system crashed while X11 was running.
As long as the protocol to tell X servers about virtual screen switches
is as primitive as it is, it is unsafe to restart an X session afterwards.
So this is basically for crash analysis.

(drochner)

Implement all ukfs interfaces (except getdents, that's a bit more
tricky) with the help of rump system calls instead of calling namei
by themselves.  Trust me when I say it was non-trivial to get the
namei calls right.

(pooka)

start dealing with RGB vs. BGR textures, also handle repetitititive textures
in some cases

(macallan)

fix previous so firefox doesn't trigger a crash anymore
tested by Jukka Salmi and myself

(macallan)

Use device_t and accessors.

(dyoung)

Make some cosmetic changes:

        Use fewer 'error = ...; break;' statements and more 'return
        ...;'

        Make the SIOCSIFFLAGS case more clear by using a switch
        statement instead of an if-else if-else chain.

        Shorten a staircase, and remove two unnecessary curly
        braces.

(dyoung)

epic_stop: clear IFF_RUNNING before disabling.

(dyoung)

sonic_stop: clear IFF_RUNNING before disabling.

(dyoung)

Make console drivers support PCI_IOC_CFGREAD/WRITE ioctls. This is useful for
XFree86, which can alternatively find the graphics card through /dev/ttyE0.
Approved by macallan.

(phx)

sf_stop: clear IFF_RUNNING before disabling.

(dyoung)

Do not use sc_invalid, it is no more.

(dyoung)

Use device_t and its accessors throughout.  Use aprint_*_dev().

Improve PMF-ability.

        Add a 'flags' argument to suspend/resume handlers and
        callers such as pmf_system_suspend().

        Define a flag, PMF_F_SELF, which indicates to PMF that a
        device is suspending/resuming itself.  Add helper routines,
        pmf_device_suspend_self(dev) and pmf_device_resume_self(dev),
        that call pmf_device_suspend(dev, PMF_F_SELF) and
        pmf_device_resume(dev, PMF_F_SELF), respectively.  Use
        PMF_F_SELF to suspend/resume self in ath(4), audio(4),
        rtw(4), and sip(4).

        In ath(4) and in rtw(4), replace the icky sc_enable/sc_disable
        callbacks, provided by the bus front-end, with
        self-suspension/resumption.  Also, clean up the bus
        front-ends.  Make sure that the interrupt handler is
        disestablished during suspension.  Get rid of driver-private
        flags (e.g., RTW_F_ENABLED, ath_softc->sc_invalid); use
        device_is_active()/device_has_power() calls, instead.

        In the network-class suspend handler, call if_stop(, 0)
        instead of if_stop(, 1), because the latter is superfluous
        (bus- and driver-suspension hooks will 'disable' the NIC),
        and it may cause recursion.

        In the network-class resume handler, prevent infinite
        recursion through if_init() by getting out early if we are
        self-suspending (PMF_F_SELF).

rtw(4) improvements:

        Destroy rtw(4) callouts when we detach it.  Make rtw at
        pci detachable.  Print some more information with the "rx
        frame too long" warning.

Remove activate() methods:

        Get rid of rtw_activate() and ath_activate().  The device
        activate() methods are not good for much these days.

Make ath at cardbus resume with crypto functions intact:

        Introduce a boolean device property, "pmf-powerdown".  If
        pmf-powerdown is present and false, it indicates that a
        bus back-end should not remove power from a device.

        Honor this property in cardbus_child_suspend().

        Set this property to 'false' in ath_attach(), since removing
        power from an ath at cardbus seems to lobotomize the WPA
        crypto engine.  XXX Should the pmf-powerdown property
        propagate toward the root of the device tree?

Miscellaneous ath(4) changes:

        Warn if ath(4) tries to write crypto keys to suspended
        hardware.

        Reduce differences between FreeBSD and NetBSD in ath(4)
        multicast filter setup.

        Make ath_printrxbuf() print an rx descriptor's status &
        key index, to help debug crypto errors.

        Shorten a staircase in ath_ioctl().  Don't check for
        ieee80211_ioctl() return code ERESTART, it never happens.

(dyoung)

In smap_stop(), clear IFF_RUNNING before disabling the device.

(dyoung)

Cosmetic: use device_t and accessors.  Change to ANSI-style
prototypes.  Delete some dead code.

(dyoung)

Print the TPPoll register when the Tx DMA engine doesn't shut down
in the time we allow.

(dyoung)

Sync with MI NE2000 changes:
> Split device_t and softc for the NE2000 Ethernet chip and all its variants
> and attachments.  Use device_t accessors, correct types, and ANSIfy when
> appropriate.

(tsutsui)

Support chdir.  This is useful for example when testing and not
wanting to stress namei() so much as test individual ops.
XXX: it should be implemented per calling thread, not per fs.

(pooka)

Split device_t and softc for the NE2000 Ethernet chip and all its variants
and attachments.  Use device_t accessors, correct types, and ANSIfy when
appropriate.

(cube)