Adapt t_ptrace_wait* tests for recent kernel changes

Adapt AUXV reading loop to handle io.piod_len == 0 as EOF.

Temporarily disable trace_thread_lwpcreate until it will be verified that
it is stable.

(kamil)

Introduce fixes for ptrace(2)

Stop disabling LWP create and exit events for PT_SYSCALL tracing.
PT_SYSCALL disabled EXEC reporting for legacy reasons, there is no need
to repeat it for LWP and CHLD events.

Pass full siginfo from trapsignal events (SEGV, BUS, ILL, TRAP, FPE).
This adds missing information about signals like fault address.

Set ps_lwp always.

Before passing siginfo to userland through p_sigctx.ps_info, make sure
that it was zeroed for unused bytes. LWP and CHLD events do not set si_addr
and si_trap, these pieces of information are passed for crashes (like
software breakpoint).

LLDB crash reporting works now correctly:

(lldb) r
Process 552 launched: '/tmp/a.out' (x86_64)
Process 552 stopped
* thread #1, stop reason = signal SIGSEGV: invalid address (fault address: 0x123456)

(kamil)

add NetMos NM9912 entries

PR kern/54149 by Mouse

(jdolecek)

fix for clang.

(christos)

Add more code from FreeBSD, still needs work

(scole)

Restore UC_MCONTEXT, UC_MCONTEXT, and clean up some whitespace

(scole)

Remove some unused assembly and other variables to reduce compile warnings

(scole)

It's the Normal Memory Remap Register (not Normal Region Remap Register)

(skrll)

Explicitly indicate that nexttowardl() is identical to nextafterl().

Requested by kamil.

(mgorny)

Fix one more writable directory check to "directory" in options seciton
when named is reloaded case.

(taca)

Fix Normal Region Remap Register bitmask names

(skrll)

take interrupt lock when closing backend.

(mlelstv)

on netbsd when defaulting to the native system, also derived MACHINE_ARCH
from the uname -p value already fetched.

this means eg, build.sh on evbarm will build the native platform rather
than "earm" (armv5) -- which was the default even on arm64 hosts.

(mrg)

Rename and partially enable trace_thread ATF ptrace(2) tests

Rename trace_thrad[1234] to more meaningful names:

- trace_thread_nolwpevents
- trace_thread_lwpexit
- trace_thread_lwpcreate
- trace_thread_lwpcreate_and_exit

In my local tests LWP CREATE events work as expected.
LWP EXIT ones are still racy and keep them disabled racy.

PR kern/51995

(kamil)

Allow cancelled byte transfer in ptrace(2) again

Current GDB depends on it on NetBSD (which is probably a GDB support bug).

In future GDB will be examined for this and this code revisited.

(kamil)

ATF ptrace(2) tests suspend1 and resume1 now pass

Verified on bare metal and in qemu.

PR kern/51995

(kamil)

Call MD code in mi_startlwp() before MI check for debugger

This allows to get initialized mcontext.

(kamil)

ATF ptrace(2) tests signal9 and signal10 now pass

PR kern/51918

(kamil)

Correct passing debugger related events for LWP create and exit

Add MI toplevel startlwp function.

Switch all userland LWPs to go through lwp_create using a shared
mi_startlwp() function between all MD ABIs.

Add debugger related event handling in mi_startlwp() and continue with
standard p->p_emul->e_startlwp at the end of this routine.

Use eventswitch() to notify the event of LWP exit in lwp_exit().

ATF ptrace(2) tests signal9 and signal10 now pass.

(kamil)

Assert that debugger event is triggered only for userland LWP

All passing ATF ptrace(2) tests still pass.

(kamil)

Add new fork/vfork ATF trests in t_ptrace_wait*

Add new tests:

- fork_detach_forker
- vfork_detach_vforker
- vfork_detach_vforkerdone
- fork_kill_forker
- vfork_kill_vforker
- vfork_kill_vforkerdone

All tests pass with recent kernel changes.

(kamil)

Correct handling of corner cases in fork1(9) code under a debugger

Correct detaching and SIGKILLing forker and vforker in the middle of its
operation.

(kamil)

Add eventswitch() in signal code

Route all crash and debugger related signal through eventswitch(), that
calls sigswitch() with preprocessed arguments.

This code avoids code duplication and allows to introduce changes that
will affect all callers of sigswitch() in debugger-related events.

No functional change intended.

(kamil)

Disallow resuming program with PC=0x0 in ptrace(2)

If the address parameter is 0, report error.
It's a popular mistake to set Program Counter to 0x0.
In certain kernels this is allowable parameter and causes
portability issue.

Disallow explicitly zeroed PC, instead of triggering
a harder to debug crash later.

(kamil)

Start converting the x86 CPU functions to inlined ASM. Matters for NVMM,
where some are invoked millions of times.

(maxv)

Pull up following revision(s) (requested by mlelstv in ticket #1253):

sys/kern/sys_pipe.c: revision 1.147
sys/kern/sys_pipe.c: revision 1.148

Clean up pipe structure before recycling it.

Handle half-closed pipes in FIONWRITE and FIONSPACE.

(martin)

Remove unused functions and reorder a little.

(maxv)

Add ether_ioctl() call for SIOCSIFADDR and SIOCSIFFLAGS to make behavior
the same as before.

XXX need cleanup.

(msaitoh)

set_params -> set_format in comment.

(isaki)

Don't release sc_lock and sc_intr_lock in trigger_{input,output}.
In the past, sc_lock was IPL_SCHED and (probably) it had conflicted
with usb subroutines.  But at some point, sc_lock has changed to use
IPL_SOFTUSB so such problems should been gone.

(isaki)

Adapt to audio2.
- Remove obsoleted or empty methods.
- Remove AUFMT_VALIDATE()/INVALIDATE().  At first, this flag is
  set/cleared but no one in the driver seems to refer at least now.
  And second, don't use such flag for other purpose.  If you need
  to do such thing, use .driver_data instead.

(isaki)

Fix a wrong and extra ether_ioctl() call in alc_ioctl(). _IOWR types call
broke ifreq by the first ether_ioct() and the second ether_ioctl() caused
unexpected behavior. One of bug example is that SIOCGIFMEDIA always failed.
This bug was added in if_alc.c rev. 1.16.

XXX pullup-8

(msaitoh)

Adapt dbri to audio2.
- Add some mutex_enter/exit.
- Remove DBRI_BIG_BUFFER option.  Such big buffer will not necessary
  in audio2.
Thank you, macallan@.

(isaki)

Enable pwm1 and add a pwm-fan node to enable the +FAN- header on ROCKPro64

(jmcneill)

Add rkpwm and pwmfan

(jmcneill)

Add support for RK3399 PWM controller.

(jmcneill)

Tickets #1691 and #1692

(martin)

Pull up following revision(s) (requested by mlelstv in ticket #1692):

sys/kern/sys_pipe.c: revision 1.147
sys/kern/sys_pipe.c: revision 1.148

Clean up pipe structure before recycling it.

Handle half-closed pipes in FIONWRITE and FIONSPACE.

(martin)

Pull up following revision(s), via patch as options
differ on the branch (requested by sevan in ticket #1691):

sys/dev/veriexec.config: revision 1.1

Add a configuration file for enabling Veriexec in kernel config(5) files.

(martin)

Ticket #1690

(martin)

Ticket #1690

(martin)

Use __nothing macro.

(isaki)

Apply patch, requested by roy in ticket #1690:

external/bsd/dhcpcd/dist/configure
external/bsd/dhcpcd/dist/src/auth.c
external/bsd/dhcpcd/dist/src/dhcp.c
external/bsd/dhcpcd/dist/src/dhcp6.c
external/bsd/dhcpcd/dist/compat/consttime_memequal.h

Security fixes for dhcpcd:
Fix a potential 1 byte read overflow with DHO_OPTSOVERLOADED.
Use consttime_memequal(3) to compare hashes.

(martin)

Apply patch, requested by roy in ticket #1690:

external/bsd/dhcpcd/dist/configure
external/bsd/dhcpcd/dist/src/auth.c
external/bsd/dhcpcd/dist/src/dhcp.c
external/bsd/dhcpcd/dist/src/dhcp6.c
external/bsd/dhcpcd/dist/compat/consttime_memequal.h

Security fixes for dhcpcd:
Fix a potential 1 byte read overflow with DHO_OPTSOVERLOADED.
Use consttime_memequal(3) to compare hashes.

(martin)

Use the comm page to inject events, rather than ioctls, and commit them in
vcpu_run. This saves a few syscalls and copyins.

For example on Windows 10, moving the mouse from the left to right sides of
the screen generates ~500 events, which now don't result in syscalls.

The error handling is done in vcpu_run and it is less precise, but this
doesn't matter a lot, and will be solved with future NVMM error codes.

(maxv)

Tickets #1251 - #1253

(martin)

Pull up following revision(s) (requested by alnsn in ticket #1252):

distrib/amd64/ramdisks/ramdisk-cgdroot/list: revision 1.2
distrib/i386/ramdisks/ramdisk-cgdroot/list: revision 1.2

Add a symlink to /altroot/stand to help the kernel find modules.

(martin)

Pull up following revision(s) via patch, as options differ between
current and this branch (requested by sevan in ticket #1251):

sys/dev/veriexec.config: revision 1.1

Add a configuration file for enabling Veriexec in kernel config(5) files.

(martin)

No need to define COMPAT_50 any longer.  All the compat decisions are
made with HOOKs.

(pgoyette)

Fix casing in Dt. Fix typo.

(wiz)

Fix casing in Dt. Fix SYNOPSIS. Fix typo.

(wiz)

Fix casing in Dt.

(wiz)

Handle ISA/EISA interrupts like isa_machdep.c.

(mlelstv)

slotdata might not be attached, don't use it then.

(mlelstv)

allow NONE build

(mlelstv)

Add man pages for Amiga clock chips

(mlelstv)

Reimplement emuxki driver.
- Use single voice per playback and per recording.
- Use fixed format, 2ch/48kHz, to simplify.
- Fix several problems in previous driver.
  And now it works even on alpha!
The driver is written by Y.Sugahara.  Thanks!

(isaki)

Adapt to audio2.
- XXX New set_format is not implemented at this point.
  This implementation can not adapt to audio2 (and it also has not
  worked since netbsd-8).
  The driver supports multiple hardware stream which is named 'voice'.
  This voice was allocated at open() and was deallocated at close().
  However, in order to support mixing, some interfaces such as
  set_params() (or set_format in audio2) has to be called before open().

(isaki)

fix a locking botch in ucomhwiflow():

tty.c always calls t_hwiflow() with tty_lock held, and the caller
of this for ucom always holds sc->sc_lock when calling down into
the tty layer.

don't try to re-take the sc_lock in ucomhwiflow() (locking against
myself is triggered here currently), but instead assert that the
lock is already held _and_ that tty_lock is held.

in ucom_detach(), when closing pipes set sc_bulkin_pipe and
sc_bulkout_pipe to NULL.  fixes bug noticed by code inspection:
a failed detach would attempt to close them a second time.

(mrg)

cvs.1 & cvs.texinfo: add various small corrections

Add various corrections mostly sourced from upstream's cvs.texinfo,
e.g., note NetBSD's local change to log admin commands in history
(subsequently added on upstream's HEAD, since it was noted in
r. 1.689), add some references to rannotate and rlog (from r. 1.686),
some amendments concerning rdiff (from r. 1.687), and miscellaneous
typo fixes.

(gutteridge)

Use mainline dts for rockpro64 now.

(jmcneill)

Use PRWORD for printing -m/-mW totals so fields don't run in to each other.

While here, update field widths for 'vmstat -mW' for modern machines.

(simonb)

Simplify card detect

(jmcneill)

Use correct size of softc in CFATTACH_DECL_NEW

(jmcneill)

Fix mux register offset

(jmcneill)

Add new ATF ptrace(2) tests: bytes_transfer_eof

Assert that error is returned after attempt to read or write post EOF.

(kamil)

Reject ptrace(2) PIOD_READ/WRITE operations with length 0

Return EINVAL for such invalid requests.

(kamil)

Pull type value mappings are different for each pin, and not based on the bank number.

(jmcneill)

Add the missing add.    (Return to the earlier state, done differently.)

When dl_print() was converted to use lla_snprintf() the offset to
the LLA in dl_addr.dl_data was forgotten (dl_data contains both
the interface name and the LL addr, we want the latter, not the former).

When there is no data (src_len == 0), still null terminate the output buffer
(provided there is space in it for the \0).

(kre)

Whitespace consistency.  NFC.

(kre)

Return EIO for empty memory transfer from ptrace(2)

Certain operations of PT_READ/PT_WRITE and PIOD_READ/PIOD_WRITE can result
in 0 byte transfer and the ptrace(2) call still returned success.

GDB had a special handling of this case for PT_IO checking piod_len != 0,
but in LLDB this corner case caused infinite loop and breakage. The LLDB
case has been enhanced.

Unfortunately the status of operation of PT_READ/PT_WRITE is not
distinguishable between successful operation and empty opeartion. This
renders this call into a questionable one.

Change the behavior and return error with EIO in scenarios of
truncated/empty byte transfers by PT_READ/PT_WRITE and empty byte transfers
from PT_IO.

No code changed is needed in GDB and LLDB.

(kamil)

CARP, Veriexec, bufq_priocscan(9)

(sevan)

turn on the "sun50i-a64-unstable-timer" workaround on "allwinner,sun8i-a83t"
compatible CPUs.  avoids triggering the KASSERT() on cubietruck plus.

idea from jmcneill.

(mrg)

Format MAP_FMT so that it's more humanly readable - missed one bit in
previous.

(uwe)

More adapt to audio2.
- Use set_format.
- Remove an empty method.
- XXX DAC1 and secondary audio device support should be removed.
  Now mixing two (or more) sources is done by audio layer.

(isaki)

Format MAP_FMT so that it's more humanly readable.
Same object code is generated for kdump.

(uwe)

Avoid incompatible function pointer casts in thrd_create(3)

Use an intermediate function trampoline to workaround different function
pointer prototypes.

While there, correct scenario returning thrd_nomem from thrd_create(3).
In practice ENOMEM is rarely returned from pthread(3).

Older code worked on tested ports, but was depending on unneeded UB.

(kamil)

match definition of hexdigits[] to the declaration in <sys/systm.h>

(christos)

sync with reality

(maxv)

Stop taking care of the INT/NMI windows in the kernel, the emulator is
supposed to do that itself.

(maxv)

Remove useless calls to nvmm_init().

(maxv)

Rename ACPI_NAME_SIZE -> ACPI_NAMESEG_SIZE with acpica update, see
src/sys/external/bsd/acpica/dist/changes.txt

(scole)

Introduce rt_addrmsg_src which adds RTA_AUTHOR to the message.
Use this when we notify userland of a duplicate address
and set RTA_AUTHOR to the hardware address of the sender.

While here, match the logging diagnostic of INET6 to the simpler one
of INET so it's consistent.

(roy)

Move lla_snprintf from if_arp.c to dl_print.c

(roy)

Catch up with constant name change.

(christos)

rtsock: Route address message simplification

Rename rt_newaddrmsg to rt_addrmsg_rt.
Add rt_addrmsg which drops the error and route arguments which are only
needed by one caller.

(roy)

Adapt to audio2.

(isaki)

Adapt to audio2.
- Drop INDEPENDENT property.  Both play and rec seems to share the
  sample rate (however, they require different calculations?).
- XXX It's better to modify frequency list more strictly.
- Remove obsoleted and empty methods.

(isaki)

Move the set-up of the net.oroute... sysctl tree to compat_50 since
the new AF_ROUTE was introduced in NetBSD-6.0

(pgoyette)

snprintb(3) says that, in the new(?) Torek format, all fields specs end with \0
The F spec is one of those, it should be terminated with \0 just like all
the others (irrelevant that it has no extra data to delimit).

Fix <sys/mman.h> to define the snprintb() format string correctly (include
the missing \0's).  Fix the copy of that definition included into
snprintb(3) to match the updated mman.h version (ride the date bump
from the day before yesterday .. this is the same change, just corrected).

Undo the previous snprintb.c change ("off by one" fix) which was an
attempt to make the broken mman.h usage work (and did, but not the way
it should be done).  Also, after using the new * format (instead of only
when something has already matched) skip the associated data so we don't
attempt to interpret it as more field specifiers.  This func needs lots of TLC!

Fix the ATF tests for snprintb() to not assume that F format is really
exactly like f format, and has data after the field specifier.  It doesn't.
Add several more tests (including testing the '*' field operator
recently added).

(kre)

For the rtsock compat code, make sure we create the "oroute" sysctl
tree.  Otherwise a 5.2 version of getifaddrs(2) gets errors.

This makes the 5.2 version of ifconfig(8) behave the same on both
NetBSD-8 and -current.  HOWEVER, both of them print nothing (for
``ifconfig -l'' command) so there's still a bug somewhere.

As reported originally by der Mouse.

(pgoyette)

acpica 20190405 changed some internal macros; update code to match

(dogcow)

new acpica

(christos)

merge conflicts

(christos)

Document that veriexec_file_add() also expects keep-filename and eval-on-load.

(alnsn)

Add pcihost at fdt, qemufwcfg at fdt, nvme at pci, and virtio at pci

(jmcneill)

one  more noreturn

(christos)

Modify the communication layer between the kernel NVMM driver and libnvmm:
introduce a bidirectionnal "comm page", a page of memory shared between
the kernel and userland, and used to transfer data in and out in a more
performant manner than ioctls.

The comm page contains the VCPU state, plus three flags:

- "wanted": the states the kernel must get/set when requested via ioctls
- "cached": the states that are in the comm page
- "commit": the states the kernel must set in vcpu_run

The idea is to avoid performing expensive syscalls, by using the VCPU
state cached, either explicitly or speculatively, in the comm page. For
example, if the state is cached we do a direct 1->5 with no syscall:

          +---------------------------------------------+
          |                    Qemu                    |
          +---------------------------------------------+
              |                                  ^
              | (0) nvmm_vcpu_getstate            | (6) Done
              |                                  |
              V                                  |
            +---------------------------------------+
            |                libnvmm                |
            +---------------------------------------+
                  |  ^          |              ^
        (1) State |  | (2) No  | (3) Ioctl:    | (5) Ok, state
        cached?  |  |          | "please cache | fetched
                  |  |          |  the state"  |
                  V  |          |              |
              +-----------+      |              |
              | Comm Page |------+---------------+
              +-----------+      |
                      ^        |
          (4) "Alright |        V
              babe"  |    +--------+
                      +-----| Kernel |
                            +--------+

The main changes in behavior are:

- nvmm_vcpu_getstate(): won't emit a syscall if the state is already
  cached in the comm page, will just fetch from the comm page directly
- nvmm_vcpu_setstate(): won't emit a syscall at all, will just cache
  the wanted state in the comm page
- nvmm_vcpu_run(): will commit the to-be-set state in the comm page,
  as previously requested by nvmm_vcpu_setstate()

In addition to this, the kernel NVMM driver is changed to speculatively
cache certain states known to be of interest, so that the future
nvmm_vcpu_getstate() calls libnvmm or the emulator will perform will use
the comm page rather than expensive syscalls. For example, if an I/O
VMEXIT occurs, the I/O Assist in libnvmm will want GPRS+SEGS+CRS+MSRS,
and now the kernel caches all of that in the comm page before returning
to userland.

Overall, in a normal run of Windows 10, this saves several millions of
syscalls. Eg on a 4CPU Intel with 4VCPUs, booting the Win10 install ISO
goes from taking 1min35 to taking 1min16.

The libnvmm API is not changed, but the ABI is. If we changed the API it
would be possible to save expensive memcpys on libnvmm's side. This will
be avoided in a future version. The comm page can also be extended to
implement future services.

(maxv)

Annotate ticket #1249

(martin)

Redo previous, this time actually using the patch requested in the ticket:
pull up following revision(s) (requested by skrl in ticket #1249):

sys/arch/cats/cats/cats_machdep.c: revision 1.87 (via patch)

The free block after the kernel from arm32_bootmem_init doesn't
account for bt_memavail.  Adjust for this.

(martin)

Adapt to audio2.
- XXX In audio2, number of blocks must be 3 or greater, so
  modify round_blocksize() to return bufsize / 4.
- XXX Doesn't neo_trigger_input() need to subtract ssz from
  buffer end like trigger_output()?

(isaki)

Remove undefined \*H

(wiz)

cvs.1 & cvs.texinfo: add a missing section documenting the "tag" and
"rtag" commands. This is taken mostly from the upstream project's
cvs.texinfo revisions 1.686 and 1.687. Additionally, I've reflected
NetBSD's local changes to log "tag" as well as "rtag" in history, and
to require admin privileges for destructive tagging commands. This
addresses PR bin/33877.

(gutteridge)

Adapt to audio2.

(isaki)

Adapt to audio2.
- Searching nearest frequency is unnecessary in audio2.

(isaki)

Don't release sc_lock on eso_halt_{input,output}.
halt_* is called with sc_lock && sc_intr_lock held.  This lock order
is first sc_lock and then sc_intr_lock.  So unlocking sc_lock with
sc_intr_lock held is wrong operation.  And cv_wait(sc_intr_lock) will
work even with sc_lock held.

(isaki)

Adapt to audio2.
- Select a few typical frequencies which doesn't have rounding error
  instead of whole range.

(isaki)

Adapt to audio2.
- Recording seems a bit tricky so I left it untouched.
  (It should work on audio2 as well if it works on -current/-8.)

(isaki)

Adapt to audio2.
- Fix wrong round_blocksize() calculation. ~0x20 -> -0x20.

(isaki)

Adapt to audio2.

(isaki)

Adapt to audio2.

(isaki)

Correct debug message in previous commit.

(isaki)

new bind

(christos)

bump bind

(christos)

merge conflicts, bump versions, sync includes

(christos)

Add a C99 symbol to libm: nexttowardl

It's an alias for an already existing symbol nextafterl.

Patch obtained from <mgorny>

Detected by the LLVM buildbot breakage in tests.

(kamil)

s/three/multiple/

Adjust text for consistency.  There are already more than three interval
timers available, and who knows if even more might be added in the future.

(pgoyette)

Remove trailing whitespace.

(wiz)

provide 1,2,4,8M entries since this are popular for loading text and data
in ELF.

(christos)

remove dup line

(christos)

Add more cases for the small alignments (useful because they cover popular
page-sizes), and add a default case.

(christos)

Document the '*' field and give a more complex example with F and *.

(christos)

Fix an off-by-one in the F modifier that ended up skipping the first : or =
entry, and add a * specifier for a "default case".

(christos)

Mmh, fix nvmm_vcpu_create(), the cpuid is given, and must not be chosen
from the free map. Looks like I forgot this after all my design rounds.
While here reorder the initialization.

(maxv)

Reorder the NVMM headers, to make a clear(er) distinction between MI and
MD. Also use #defines for the exit reasons rather than an union. No ABI
change, and no API change except 'cap->u.{}' renamed to 'cap->arch'.

(maxv)

Fix build on ppc64

(skrll)

simplify and explain

(christos)

Adapt to audio2.
- auvia_set_params_sub is also renamed because caller auvia_set_params
  is renamed to auvia_set_format.

(isaki)

More adapt to audio2.
- Shrink autri_formats[].  HW actually supports frequencies other
  than 48kHz but they may include (ignorable) rounding error in
  thier calculation.  So accept only 48kHz.  It's enough.
- Remove obsoleted and empty methods.

(isaki)

Adapt to audio2.
Tested on sparc64 by macallan@.  Thank you.

(isaki)

Adapt to audio2.
- Move auixp_formats' definition from header to source.
- Make compilable even with DEBUG_AUIXP.

(isaki)

Adapt to audio2.

(isaki)

Adapt to audio2.
- Correct frequency_type in modem_audio_format.
- {spdif,modem}_audio_format in my previous commit (in this branch)
  was also wrong. :(

(isaki)

More adapt to audio2.
- Replace set_params with new set_format.
- Remove sc_encodings which is no longer used.

(isaki)

Skip CARP & Veriexec support to reduce size

(sevan)

Skip CARP & Veriexec support in RAMDISK to reduce size

(sevan)

Add audio_indexof_format().  It finds index matches with specified
audio_param_t from array of audio_format_t.
Formerly, it was auconv_set_converter() that convert userland
params to the hardware native (linear) format and such conversion
is unnecessary in audio2.  But some drivers have gotten index from
this format array using auconv_set_converter().  This function is
an alternative to such usage.

(isaki)

Adapt to audio2.

(isaki)

Adapt to audio2.

(isaki)

Fix some typos, improve wording.

(wiz)

Add support for EnhancedIBRS, a more performant mitigation for SpectreV2,
available on future CPUs (or maybe they already exist now...).

(maxv)

Adapt to audio2.
- XXX Name conflicts.  audio2 replaces set_params with set_format
  and audio_dai already have had its own set_format.
  I named newly introduced one audio_dai_mi_set_format for now.

(isaki)

If guest events were being processed when a #VMEXIT occurred, reschedule
the events rather than dismissing them. This can happen for instance when a
guest wants to process an exception and an #NPF occurs on the guest IDT. In
practice it occurs only when the host swapped out specific guest pages.

(maxv)

Semicolon not needed

(pgoyette)

Oooppss - deleted too many chars.  Put them back.

Should fix the build break.

(pgoyette)

Optimize nvmm-intel, use inlined GCC assembly rather than function calls.

(maxv)

A few more empty-string --> NULL in required-modules lists

(pgoyette)

Adapt to audio2.

(isaki)

Adapt to audio2.

(isaki)

Use usbdevices.config & bluetooth.config to reduce duplication.

(sevan)

Rename imask to intr_mask as imask is defined globally in
sys/arch/powerpc/include/intr.h:104 and this clashes.

Resolves build on macppc.
ok <pgoyette>

(sevan)

fix uchcom(4) detach, like umodem(4) recently:

- use static normally in umodem_common.c
- add sc_refcnt, sc_lock and sc_detach_cv to softc.  usage is:
  - sc_dying is protected by sc_lock
  - sc_detach_cv is matched with sc_lock for cv operations
  - sc_refcnt is increased in open and decreased in close, any time
    it is decreased, it is checked for less than zero, and a broadcast
    performed on sc_detach_cv.  detach waits for sc_refcnt
- uchcom_param() and uchcom_set() check for sc_dying

this fixes pullout out an open ucom@uchcom.

XXX: pullup

(mrg)

Use filesystems.config to reduce duplication

(sevan)

Sort more

(sevan)

Sort

(sevan)

Backport improvements into GCC's sanitizer_linux.cc from more recent LLVM

Backport fixups for syscall()/__syscall() routines from LLVM compiler-rt
dated October 1st 2018. The commit beffore switching LLVM compiler-rt
sycall calls to libc calls for NetBSD.

GCC8 will get part of these changes from upstream and GCC9 will operate on
libc calls directly for the NetBSD port.

This is intended to correct misuse of parameters of syscall()/__syscall()
that could break !x86 ports in UBSan.

(kamil)

Enforce alignment also if the compiled in PAGE_SIZE is bigger than
getpagesize()

(christos)

Back to using MAX_PAGE_SHIFT

(christos)

Use veriexec.config to reduce duplication

(sevan)

typo

(sevan)