tests/make: test edge cases in pattern matching

(rillig)

make: condense Str_Match

The test for '\\' followed by '\0' was redundant since at that point,
*str is guaranteed to be not '\0', which takes the next 'return false'.

No functional change.

(rillig)

tests/make: demonstrate another pathological pattern match

(rillig)

Specify -fno-unwind-tables to shrink binaries.

(tsutsui)

make: rename loadfile to LoadFile

It was the only function in parse.c that had a lowercase name.

No functional change.

(rillig)

make: split IncludeFile into separate functions

No functional change.

(rillig)

make: extract half of IncludeFile into a separate function

No functional change.

(rillig)

make: condense code for searching a file in the paths

No functional change.

(rillig)

Mention about luna68k psgpam(4) audio.

(tsutsui)

luna68k: add psgpam(4) audio for LUNA, using YM2149 PSG/SSG sound chip.

"PSG PAM" is a audio driver using "Pulse Amplitude Modulation" with the
YM2149 PSG/SSG chip connected to the I/O processor HD647180X on LUNA.
The driver plays up to 30 kHz PCM audio by manipulating PSG channel volumes
at extremely high speed with 6.144 MHz HD647180X, a Z80 compatible sub CPU.
https://www.youtube.com/watch?v=fR7PqaLpijQ
https://www.youtube.com/watch?v=kCDZCgkwgBA

Compatibilities with the existing XPIOCDOWNLD ioctl(2) to download and
exec user programs (PSG sound driver player ported from PC-6001 etc.)
on the HD647180X is still provided.
https://twitter.com/tsutsuii/status/1533247362965315584

This psgpam(4) driver for NetBSD/luna68k was written by @moveccr
during isaki-audio2 branch developments, and it was also demonstrated
at Open Source Conference 2018 Kyoto:
https://github.com/isaki68k/netbsd-src/tree/audio2
https://mail-index.netbsd.org/netbsd-advocacy/2018/08/07/msg000781.html
https://twitter.com/tsutsuii/status/1025504875625365505

Note specifications of interrupts from the HD647180X to the host LUNA
was confirmed by Kenji Aoyama on his LUNA-88K2 running OpenBSD/luna88k.

I would like to thank all people still working on ancient LUNA machines!

(tsutsui)

tests/make: reword comment in Makefile to be more specific

(rillig)

tests/make: demonstrate spaces and '#' in command line arguments

(rillig)

tests/lint: fix test for loss of accuracy on ILP32 platforms

The test had been wrong since msg_132.c 1.14 from 2022-05-30.

Using 'unsigned long' in a test that was intended to behave the same on
ILP32 and LP64 was an accident.  Use 'unsigned long long' instead, which
is 64-bits wide on all platforms supported by lint.

Move the test about conversion from 'long' to 'int' to the
platform-specific test files.

Noticed by martin@ on powerpc.

(rillig)

Enable crypto, swcrypto and cgd.

(martin)

Tickets #1466 - #1469

(martin)

Pull up following revision(s) (requested by skrll in ticket #1469):

libexec/ld.elf_so/arch/hppa/hppa_reloc.c: revision 1.49
libexec/ld.elf_so/arch/hppa/rtld_start.S: revision 1.14

Set DP early so that any binary functions that override others get the
right value if they're called before _start.  This is true of bash where
it provides its own getenv.

Part of port-hppa/56118: sporadic app crashes in HPPA -current

(martin)

Pull up following revision(s) (requested by skrll in ticket #1468):

sys/arch/hppa/hppa/trap.c: revision 1.120
sys/arch/hppa/include/ptrace.h: revision 1.11-1.12 (via patch)

Define a PTRACE_ILLEGAL_ASM

Match up PTRACE_BREAKPOINT_ASM with PTRACE_BREAKPOINT which is the
gdb breakpoint instruction.

Only report the SSBREAKPOINT break instruction as SIGTRAP/TRAP_TRACE. All
other break instructions will be reported as SIGTRAP/TRAP_BRKPT
This fixes a mistake I made back in 2008.

PR/56866: hppa: kernel gets confused between actual breakpoints and
single-step breakpoints

(martin)

Pull up following revision(s) (requested by skrll in ticket #1467):

sys/arch/hppa/include/pmap.h: revision 1.41
sys/arch/hppa/hppa/pmap.c: revision 1.117

port-hppa/56849: Wacko kernel memory accounting in current/hppa

Two fixes:
- Don't include direct mapped memory in pmap statistics
- Decrement pmap statistics counts in pmap_kremove

(martin)

Pull up following revision(s) (requested by skrll in ticket #1466):

sys/arch/hppa/hppa/trap.c: revision 1.119

Handle 'NA' (non-access) traps for the lpa and probe instructions.  The
change is inspired by OpenBSD with a bunch of my own, mainly stylistic,
changes.

Thanks to Tom Lane for the analysis.

PR/56118: sporadic app crashes in HPPA -current

(martin)

"inp_hash" is not used now.

(knakahara)

Use LIST_FOREACH macro.

(knakahara)

PR 56874: avoid using uninitialized memory, patch from Tom Lane.

(martin)

PR 56875: Bump timeout for the bignumber test to 2h, as suggested
by Tom Lane to give slower architectures a chance to complete it.

(martin)

Add support for explicitly selecting "EFI system partition" as a partition
type.

(martin)

There are two hard problems in computer science cache invalidation, naming
and off-by-one errors.

Fix a naming and off-by-one bug... Thanks to Tom Lane for spotting my typo.

(skrll)

Only report the SSBREAKPOINT break instruction as SIGTRAP/TRAP_TRACE. All
other break instructions will be reported as SIGTRAP/TRAP_BRKPT

This fixes a mistake I made back in 2008.

PR/56866: hppa: kernel gets confused between actual breakpoints and single-step breakpoints

(skrll)

Oops... revert commit mistake

(skrll)

Handle 'NA' (non-access) traps for the lpa and probe instructions.  The
change is inspired by OpenBSD with a bunch of my own, mainly stylistic,
changes.

Thanks to Tom Lane for the analysis.

PR/56118: sporadic app crashes in HPPA -current

(skrll)

fix typo in comment: white spae -> whitespace.

(andvar)

refactor: use TAILQ_FOREACH instead of TAILQ_FOREACH_SAFE about inpt_queue.

They don't use "ninph" pointer and don't remove elements.

(knakahara)

s/spcae/space/ in comment and error messages.

(andvar)

wsfontload(8): Mention .wsf font files.  Fix formatting.

(uwe)

Remove now unused bootdefault() function.

Part of PR 56862.

(wiz)

Do not use default entry's parameters for for plain "boot" command

Go back to the "menu" instead of you want that.

Patch from RVP in PR 56862, ok uwe@

(wiz)

wsfont.c: use __arraycount instead of defining it locally

(uwe)

Terminus Font: Import latin1 encoded versions from 4.49.1

(uwe)

bdfload: Teach it to write wsf font files.

wsfontload(8) now groks simple "wsf" file format - that is basically a
header of kinda struct wsdisplay_font but with string fields embedded
and numeric fields in little-endian and font data afterwards.

Teach bdfload to write out such files if -o file.wsf argument is
specified.

While here fix a couple of bugs too (sorry, should have been separate
commits).  Fix an off by one in numchars calculations.  Fix the type
of buflen.

(uwe)

initialize a variable to appease clang

(macallan)

This kernel has no midi, so comment out pseudo-device sequencer

(martin)

Terminus Font: Regen after fixing an off-by-one in bdfload.

A bug in bdfload caused numchars parameter to be one less than
reality, so the last glyph was lost.  It was not immediately
noticeable b/c in IBM (cp437) encoded fonts the last glyph is a
(non-breaking) space.

While here, change the font names to have -CP437 suffix.

(uwe)

Do not warn for ELF note for FDO packaging metadata
(https://systemd.io/COREDUMP_PACKAGE_METADATA/),
found in userland binaries of Fedora 36.

(rin)

Functionalize frame pointer backtrace.

(ryo)

This kernel has no midi, so comment out pseudo-device sequencer

(martin)

This kernels have no midi configured, so comment out pseudo-device sequencer

(martin)

This kernel has no midi configured, so comment out pseudo-device sequencer

(martin)

Add method to initialize errinfo so that npfctl does not print random strings.

(christos)

On aarch64, ddb backtrace can be performed without framepointer by specifying
the /s modifier to the ddb trace command (trace/s, bt/s).
The default is trace with framepointer (same as before).

This allows backtracing even on kernels compiled with -fomit-frame-pointer.

(ryo)

Whitespace

(skrll)

PR/56865: hppa: t_ptrace_core_wait test doesn't know it must advance PC after breakpoint

Add __hppa__ to the list of architectures that need to advance to the instruction
after the breakpoint

Thanks to Tom Lane for the fix.

(skrll)

use stp if possible.

(ryo)

revert bits and pieces I accidentally included in my last commit

from hannken

(nia)

Tickets #1464 and #1465

(martin)

Pull up following revision(s) (requested by skrll in ticket #1465):

usr.sbin/makemandb/makemandb.c: revision 1.63

Don't index outside the mdocs array of function pointers.  Analysis and
suggested fixes from Tom Lane.  I played it safe and went with (my
variation of) the minimal fix.

PR port-hppa/56118: sporadic app crashes in HPPA -current

(martin)

Pull up following revision(s) (requested by skrll in ticket #1464):

lib/libc/arch/hppa/sys/ptrace.S: revision 1.8

Save and restore %r19 the "linkage table pointer register" across the call
to __cerror so if the ptrace syscall fails we can call __cerror again with
the correct %r19 value.

Do this even though the call of __cerror doesn't go via the PLT because
__cerror calls __errno which does.

Analysis and fix from Tom Lane in
PR port-hppa/56864: hppa: ptrace(2) dumps core when returning an error

I changed the location of where %r19 is stored on the stack to follow the
ABI.

(martin)

Ticket #1748

(martin)

Apply patch, requested by msaitoh in ticket #1746:

sys/dev/pci/ixgbe/ixgbe.c (apply patch)

Fix the location of "break" statement. No functional change intended.

(martin)

Remove accidentally committed merge conflict line.

(nia)

build system: Revert all the recent additions of MK[...] knobs that
allow conditionally disabling the building of certain user space
programs in the 'base' set.

There is not enough consensus that this is the right way and a few
people had strong objections, see source-changes-d@.

(nia)

Don't index outside the mdocs array of function pointers.  Analysis and
suggested fixes from Tom Lane.  I played it safe and went with (my
variation of) the minimal fix.

port-hppa/56118: sporadic app crashes in HPPA -current

(skrll)

Match up PTRACE_BREAKPOINT_ASM with PTRACE_BREAKPOINT which is the
gdb breakpoint instruction.

(skrll)

PR kern/51254
uvm_map_findspace(): Output current value of "entry" when KASSERT fires.

(rin)

Define a PTRACE_ILLEGAL_ASM

(skrll)

Save and restore %r19 the "linkage table pointer register" across the call
to __cerror so if the ptrace syscall fails we can call __cerror again with
the correct %r19 value.

Do this even though the call of __cerror doesn't go via the PLT because
__cerror calls __errno which does.

Analysis and fix from Tom Lane in

port-hppa/56864: hppa: ptrace(2) dumps core when returning an error

I changed the location of where %r19 is stored on the stack to follow the
ABI.

(skrll)

Ticket #1463.

(msaitoh)

Pull up following revision(s) (requested by martin in ticket #1463):

crypto/external/bsd/openssh/dist/sshkey-xmss.c patch
crypto/external/bsd/openssh/dist/version.h patch

Apply upstream fix for CVE-2019-16905.
Call this NetBSD_Secure_Shell-20220604.

(msaitoh)

s/filer/filter/ in comment. No functional change.

(msaitoh)

uvm(9): Sprinkle assertions into uvm_map_findspace.

May help to diagnose PR kern/51254.

(riastradh)

bwfm(4) uses firmload.

(rjs)

Match explicit section names and not .{text,data}.* for example.

This test now passes on hppa.

(skrll)

gcc 9 is no longer relevant here.

(may be in gcc.old, but soon after netbsd-10 branch we'll both
obsolete it on the branch and likely upgrade -current gcc.)

(mrg)

uvm(9): Don't duplicate vm_map_min/max in `show map' output.

Didn't notice these were already there, oops!

(riastradh)

uvm(9): Sprinkle more info into hint/orig_hint assertions.

May help to diagnose PR kern/51254.

(riastradh)

uvm(9): Print min/max address and first_free entry in ddb `show map'.

May help to diagnose PR kern/51254.

(riastradh)

tests/lib/libc: Test mmap(2) with bad hints.

(riastradh)

- Use register prefix (%r).
- Remove workaround for old gas(1) bug.

No binary changes.

(rin)

uvm(9): Fix mmap optimization for topdown case.

PR kern/51393

(riastradh)

uvm(9): Fix 19-year-old bug in assertion about mmap hint.

Previously this would _first_ remember the original hint, and _then_
clamp the hint to the VM map's range:

orig_hint = hint;
if (hint < vm_map_min(map)) { /* check ranges ... */
if (flags & UVM_FLAG_FIXED) {
UVMHIST_LOG(maphist,"<- VA below map range",0,0,0,0);
return (NULL);
}
hint = vm_map_min(map);
...
KASSERTMSG(!topdown || hint <= orig_hint, "hint: %#jx, orig_hint: %#jx",
    (uintmax_t)hint, (uintmax_t)orig_hint);

Even if nothing else happens in the ellipsis, taking the branch
guarantees the assertion will fail in the topdown case.

(riastradh)

mmap(2): If we fail with a hint, try again without it.

`Hint' here means nonzero addr, but no MAP_FIXED or MAP_TRYFIXED.

This is suboptimal -- we could teach uvm_mmap to do a fancier search
using the address as a hint.  But this should do for now.

Candidate fix for PR kern/55533.

(riastradh)

Fix syntax error

(pgoyette)

Include sequencer code conditionally.  Fixes build for sparc and sparc64
(and anything else that has a midi but no sequencer).

(pgoyette)

Ticket #1747

(martin)

Pull up following revision(s) (requested by tsutsui in ticket #1747):

sys/arch/atari/atari/stalloc.c: revision 1.17

Restore NULL pointer checks lost in rev 1.16.

Fixes PR port-atari/56859, ok'ed mlelstv@, and confirmed on TT030.
Should be pulled up to netbsd-9 and netbsd-8.

(martin)

Ticket #1462

(martin)

Pull up following revision(s) (requested by tsutsui in ticket #1462):

sys/arch/atari/atari/stalloc.c: revision 1.17

Restore NULL pointer checks lost in rev 1.16.

Fixes PR port-atari/56859, ok'ed mlelstv@, and confirmed on TT030.
Should be pulled up to netbsd-9 and netbsd-8.

(martin)

Add sets-lists entries for the new combined midi_seq module

(pgoyette)

Combine the midi and sequencer modules into a single midi_seq module
to avoid a circular dependency as noted in kern/56772.  Retain minimal
modules of the original names to accomodate auto-loading upon access
to the /dev/xxx nodes.

(pgoyette)

ipf(8): document all -l options

References to "none", "state", and "nat" were missing in places.
Addresses PR bin/36763 from Wolfgang Stukenbrock.

(gutteridge)

optimize. reduce 2 instructions.

(ryo)

remove reference to gatekeeper.dec.com ftp.

(andvar)

some srm firmware (including legacy platforms) still available at hp ftp.
not sure if HPE selling firmware CDs, but it is definitely not Compaq anymore.

(andvar)

fix link to pkgsrc.tar.gz file in postinstall section.

(andvar)

fix the title of the boot tape creation section.

(andvar)

Ticket #1746

(martin)

Pull up the following revisions, requestes by msaitoh in ticket #1746:

sys/dev/pci/ixgbe/ix_txrx.c 1.95
sys/dev/pci/ixgbe/ixgbe.c 1.261,1.263,
1.265-1.268,1.273,
1.275-1.277,1.305,
1.312,
1.316-1.321 via patch
sys/dev/pci/ixgbe/ixgbe.h 1.85 via patch
sys/dev/pci/ixgbe/ixgbe_mbx.c 1.19
sys/dev/pci/ixgbe/ixgbe_netbsd.h 1.15-1.16
sys/dev/pci/ixgbe/ixgbe_type.h 1.46-1.47
sys/dev/pci/ixgbe/ixgbe_x550.c 1.26
sys/dev/pci/ixgbe/ixv.c 1.178,1.182 via patch

- Reduce code duplication between ixgbe_msix_admin() and
  ixgbe_legacy_irq().
  - Add missing code which was not in ixgbe_msix_admin() from
    ixgbe_legacy_irq() and vice versa.
  - Reorder code.
- Disable/enable the OTHER interrupts correctly.
- Don't return in the middle of ixgbe_msix_admin() when an flow
  director reinit failed. NetBSD currently doesn't support flow
  director, so this is not a real bug.
- Print ECC, PHY and temp error log using with ratecheck().
- Correctly re-enable queue interrupt in ixgbe_legacy_irq().
- Correctly enter the recovery mode.
- No functional change:
  - Add some debug printf()s.
  - Don't use "more" flag for simplify.
  - Use macro.
  - Fix typos in comment.
  - KNF.

(martin)

Ticket 1461

(martin)

Pull up following revision(s) (requested by gutteridge in ticket #1461):

usr.sbin/makemandb/apropos.1: revision 1.19
usr.sbin/makemandb/apropos.c: revision 1.25
usr.sbin/makemandb/apropos.c: revision 1.26
usr.sbin/makemandb/apropos.1: revision 1.20
usr.sbin/makemandb/apropos.1: revision 1.21
usr.sbin/makemandb/apropos.1: revision 1.22
usr.sbin/makemandb/apropos.1: revision 1.23
usr.sbin/makemandb/apropos-utils.c: revision 1.46
usr.sbin/makemandb/apropos-utils.c: revision 1.47
usr.sbin/makemandb/apropos-utils.c: revision 1.49

PR/54343: Prevent NULL pointers in callback strings; use "*?*" for now to
identify them.

PR bin/54343:  We want the callback_args.machine to be NULL if it is not
present in the DB.

The previous commit fixed the problem of allowing apropos to not crash and
produce output even if the database is missing values for certain mandatory
fields, such as name, section etc. Normally we don't expect those values
to be missing in the database but in case of parsing errors it can happen.

However, the machine architecture is an optional field since not all man pages
are hardware specific so that should be allowed to be set to NULL if not
present in the database.

apropos.c: fix pager functionality

Issue reported by Rocky Hotas on NetBSD-Users, patch input from RVP on
same, adjustments by me.

apropos.1: document the PAGER environment variable

apropos(1): use proper -width

apropos(1): use proper -width for the list of options too

apropos(1): Tweak the description of -1, ... -9, and -s

-s is not for compatibility only, because section names can be
anything.  E.g. we have 3lua and 9lua in base.  We have rudiments of
3f (for FORTRAN libs).  Some packages in pkgsrc also use suffixed 1
and 3 sections.

apropos(1): Use the official spelling for "SQLite".
While here, use .Bx to refer to 3BSD.

apropos(1): improve error handling in edge cases
Patch from RVP on NetBSD-Users, with an additional comment tweak by me.

Summary from RVP:
1. Ignore SIGPIPE so that we're not killed in the middle of some
  DB operation by a botched $PAGER:
$ env PAGER=3D/non-existent apropos -p ...
2. Return proper exit status in case of write errors:
$ apropos ... >/dev/full || echo fail

(martin)

fix folloing->following typos in more files.
also s/begginning/beginning/.

(andvar)

Restore (again) support for grant tables v1: pvshim doesn't support
v2 (it doesn't even support GNTTABOP_set_version) some setups will
likely require pvshim in the future.

(bouyer)

1460

(snj)

Pull up following revision(s) (requested by msaitoh in ticket #1460):
sys/dev/pci/ixgbe/ixgbe.c: patch
Fix previous to re-enable legacy interrupt correctly. It's a part of
ixgbe.c rev. 1.273.

(snj)

typo in comment

(dholland)

Modify comment. ixgbe_enable_queue() can be used on both MSI-X and legacy intr.

(msaitoh)

KNF. Modify comment. No functional change.

(msaitoh)

check platform-do-*-mute properties to find out the resp. GPIO's polarity
now this works properly on pmac7,3

(macallan)

If we get the disk description via device properties, the strings may
have arbitrary length - make sure the "available disks" menu fits
on the current screen.

(martin)

Ticket #1459

(martin)

Pull up the following revisions, all via patch, requested by msaitoh
in ticket #1459:

sys/dev/pci/ixgbe/ixgbe.c 1.261,1.265-1.268,1.273,1.275-1.277,
1.312,1.316-1.319 via patch
sys/dev/pci/ixgbe/ixgbe.h 1.85 via patch
sys/dev/pci/ixgbe/ixgbe_type.h 1.46-1.47
sys/dev/pci/ixgbe/ixgbe_x550.c 1.26
sys/dev/pci/ixgbe/ixv.c 1.182

- Reduce code duplication between ixgbe_msix_admin() and
  ixgbe_legacy_irq().
  - Add missing code which was not in ixgbe_msix_admin() from
    ixgbe_legacy_irq() and vice versa.
  - Reorder code.
- Disable/enable the OTHER interrupts correctly.
- Don't return in the middle of ixgbe_msix_admin() when an flow
  director reinit failed. NetBSD currently doesn't support flow
  director, so this is not a real bug.
- Print ECC, PHY and temp error log using with ratecheck().
- Correctly re-enable queue interrupt in ixgbe_legacy_irq().
- Correctly enter the recovery mode.
- No functional change:
  - Add some debug printf()s.
  - Don't use "more" flag for simplify.
  - Fix typos in comment.
  - KNF.

(martin)

Increase the resource limit so these tests pass on hppa.

LGTM from Roland

(skrll)

Trailing whitespace

(skrll)

add timeout period to timeout message.

seen on radeon 5450 on an asrock m/b with ryzen 5600g cpu.

(mrg)

regen.

(mrg)

add some AMD zen2/3 devices.

(mrg)

tidy up backtrace from crash(9) on aarch64

- fix to dump trapframe when backtracing from crash(8).
- use db_read_bytes() when reading kernel memory.

(ryo)

KNF. No functional change.

(msaitoh)

As a tiny optimization, rearrange cases as follows:

(1) load/store of format D (base + disp)
(2) load/store of format X (base + index)
(3) lmw and stmw

For clang-compiled userland (*), their frequencies are roughly,
(1) > (2) >> (3) ~ 0.

Improvement should be minor; we are already trapped in the alignment
fault handler.

(*) clang unconditionally emits unaligned memory access for powerpc.
    Undocumented -disable-ppc-unaligned option does not work...

(rin)

Localize one more variable to NPCI > 0 block. Fix ia64 build.

(rin)

tmux 3.3 out.

(wiz)

cprng(9): cprng_fast is no longer used from interrupt context.

Rip out logic to defer reseeding to softint.

(riastradh)

mvxpsec(4): Fix missing change to freesession return type.

Not sure how this didn't get committed before -- it was supposed to be
part of this change:

https://mail-index.netbsd.org/source-changes/2022/05/22/msg138764.html

(riastradh)

tmpfs_read: respect MNT_NOATIME.

(hannken)

delay slot indentation

(skrll)

Revert previous, the mask is intented to be like sc_output_mask, a bitmask
of active ports.

(martin)

In SNAPPER_OUTPUT_SELECT the "mask" is a bitmap of muted outputs, not
active outputs.

(martin)

Recognize the OF labels used for line output and headphone output on a
G5 powermac 7,3.

(martin)

Correctly enter the recovery mode. Not tested.

(msaitoh)

Correctly re-enable queue interrupt in ixgbe_legacy_irq().

- Don't enable queue 1-15 interrupt.
- Don't enable queue 0 interrupt when the queue work is being scheduled.
- OK'd by knakahara.

(msaitoh)

Simplify ixgbe_msix_que(). No functional change.

(msaitoh)

add now required pci.h / NPCI.  (the ia64 version builds with PCI
enabled, but i guess it won't be usable.)

XXX:  enable this on evbarm*?  it is x86/ia64 only currently.

(mrg)

reorganise most of the NetBSD portion of nvkm_mem_dtor().

when nvkm_mem_new_host() is called via the in-kernel ioctl method,
we copy the supplied dmamap, use it's dm_nsegs value for allocation
of "mem->dma", and assume it remains valid until we're done.

when this path is taken "mem->mem" remains NULL so all the code in
nvkm_mem_dtor() is ignored, and the "mem->dma" is leaked.  this is
one leak seen in PR#56826.  as "dmamap->dm_nsegs" can become invalid
before the dtor call, store the value in "mem->nseg" for use in the
dtor, and convert the dtor to free "mem->dma" if "mem->dma" is set.
additionally, "mem->pages" should end up being the same value as
"nseg" here, ASSERT() this.

while here properly mark NetBSD specific code in nvkm_mem_new_host().

additionally, destroy the dmamap created in the non-ioctl path of
nvkm_mem_new_host().  this is another leak seen in PR#56826.

with both of these fixes my "kmem-04096" pool does not grow rapidly
while using "mpv -vo gpu".  infact, once i loaded the relevant file
into memory, this pool remains stable after at least one minute of
video playback.

ok riastradh@

(mrg)

allow ACPI to compile without PCI support.  fixes ia64 build.

(mrg)

document gomoku

(christos)

Elide vax breakage

(christos)

For some reason on my system, the guest starts with a pending event in
the shared structure. On i386, this cause hypervisor_callback to be
entered before cpu_info_primary is fully initialised, especially on i386
ci_intrstack is still NULL, which cause a crash when we try to use it.
Work around by recycling the boot's tmp stack for this until cpu_attach()
is called.

(bouyer)

When we have pending events in stipending(), evt_set_pending() has to set
the ih_pending flag for each handler too. Xen/i386 should be stable again.

(bouyer)

fix lint build on vax

(christos)

Restore NULL pointer checks lost in rev 1.16.

Fixes PR port-atari/56859, ok'ed mlelstv@, and confirmed on TT030.
Should be pulled up to netbsd-9 and netbsd-8.

(tsutsui)

libpthread(3): Fix a marvellous interaction with rtld.

Patch from chs@.  Comment explaining the story by me.  This patch may
not be optimal -- maybe it would be better in pthread__init, or
better for rtld to call _lwp_unpark after _lwp_park in the contened
case -- but we've tested this version and it's annoying to reproduce,
so let's take this version and worry about testing improvements
later.

(riastradh)

Revert previous; evt_set_pending() will set ret to 1 if needed to this was
not our bug.

(bouyer)

vnd(4): Work around deadlock in VNDIOCCLR.

Since the changes this year to eliminate a host of races and
deadlocks in open, close, revoke, attach, and detach, closing the
last instance of a device special node has the side effect of waiting
for all concurrent I/O operations (read, write, ioctl, strategy, &c.)
on the device to complete.

Unfortunately, while this works for physical devices which revoke
open device nodes in their autoconf detach functions, as invoked by
some hardware interrupt indicating that the device is no longer
present, pseudo-devices like vnd(4) work differently -- or, work by
luck, or don't work any more.

VNDIOCCLR acts kind of like an autoconf detach function in that it
revokes open device nodes, which closes the last instance.  But
VNDIOCCLR is itself called via ioctl, which is an I/O operation that
close waits for.  So we end up with a deadlock, spec_io_drain waiting
for spec_close lower down in the call stack:

> spec_io_drain() at netbsd:spec_io_drain+0x84
> spec_close() at netbsd:spec_close+0x1c6
> VOP_CLOSE() at netbsd:VOP_CLOSE+0x38
> spec_node_revoke() at netbsd:spec_node_revoke+0x14d
> vcache_reclaim() at netbsd:vcache_reclaim+0x4e7
> vgone() at netbsd:vgone+0xcd
> vrevoke() at netbsd:vrevoke+0xfa
> genfs_revoke() at netbsd:genfs_revoke+0x13
> VOP_REVOKE() at netbsd:VOP_REVOKE+0x35
> vdevgone() at netbsd:vdevgone+0x64
> vnddoclear.part.0() at netbsd:vnddoclear.part.0+0xaa
> vndioctl() at netbsd:vndioctl+0x78c
> bdev_ioctl() at netbsd:bdev_ioctl+0x91
> spec_ioctl() at netbsd:spec_ioctl+0xa5
> VOP_IOCTL() at netbsd:VOP_IOCTL+0x41
> vn_ioctl() at netbsd:vn_ioctl+0xb3
> sys_ioctl() at netbsd:sys_ioctl+0x555

In the past, there was a workaround for what was presumably a crash
instead of a deadlock here: don't issue revoke (vdevgone) on the open
character devices for the minor number in use by the ioctl.  If you
use, e.g., `vnconfig -u vnd0', and vnconfig(8) picks /dev/rvnd0c or
/dev/rvnd0d, that special case kicks in.  But if you use `vnconfig -u
/dev/vnd0d', the ioctl will be issued on the block device instead, so
the special case doesn't kick in, so the operation deadlocks.

It is actually probably safe not to revoke the block device if what
the ioctl caller holds open is that, because specfs(9) forbids more
than one open of a block device, so nothing else can have it open
anyway.

Unclear what the consequences of failing to revoke the character
device are -- but this is what vnd(4) has done all along.  cgd(4) and
ccd(4) also don't bother to revoke.  We don't have a notion of
`revoke every file descriptor _except_ this one'; only a vnode as a
whole can be revoked, including all references to it.

This is a stop-gap measure to avoid a deadlock we are definitely
hitting on some users.  A slightly better measure would be to revoke
the block or character device according to which one is being used,
but that requires a little more work with two different d_ioctl
functions -- and wouldn't address isues with the character device.  A
proper solution requires identifying the appropriate protocol for all
of these pseudo-device disk drivers and using it uniformly for them.

Reported on current-users:
https://mail-index.netbsd.org/current-users/2022/05/27/msg042437.html

(riastradh)

Ticket #1745

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #1745):

sys/dev/pci/ixgbe/ix_txrx.c: revision 1.98
sys/dev/pci/ixgbe/ixv.c: revision 1.181
sys/dev/pci/ixgbe/ixgbe.c: revision 1.315
sys/dev/pci/ixgbe/ixgbe.h: revision 1.86

bus_dmamem_unmap() before bus_dmamem_free(), otherwise we may give back meomry
which is still (and will stay) mapped.

Fixes one instance of "panic: HYPERVISOR_mmu_update failed" on Xen.
There may be others.

Fix a bug that the legacy interrupt doesn't work when MSI-X allocation failed.
Fixes PR kern/56857.

Remove unused adapter->msix_mem.

(martin)

Ticket #1458

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #1458):

sys/dev/pci/ixgbe/ixv.c: revision 1.181
sys/dev/pci/ixgbe/ixgbe.c: revision 1.315
sys/dev/pci/ixgbe/ixgbe.h: revision 1.86

Fix a bug that the legacy interrupt doesn't work when MSI-X allocation failed.
Fixes PR kern/56857.

Remove unused adapter->msix_mem.

(martin)

libc: Reintroduce getentropy.

This was introduced two years ago when the getrandom/getentropy API
question was still open, and removed because the discussion was
ongoing.  Now getentropy is more widely adopted and soon to be in
POSIX.  So reintroduce the symbol into libc since we'll be keeping it
anyway.  Discussion of details of the semantics, as interpreted by
NetBSD, is ongoing, but the symbol needs to get in before the
netbsd-10 branch.  The draft POSIX text is
(https://www.opengroup.org/austin/docs/austin_1110.pdf):

    SYNOPSIS
        #include <unistd.h>

        int getentropy(void *buffer, size_t length);

    DESCRIPTION
        The getentropy() function shall write length bytes of data
        starting at the location pointed to by buffer. The output
        shall be unpredictable high quality random data, generated by
        a cryptographically secure pseudo-random number
        generator. The maximum permitted value for the length
        argument is given by the {GETENTROPY_MAX} symbolic constant
        defined in <limits.h>.

    RETURN VALUES
        Upon successful completion, getentropy() shall return 0;
        otherwise, -1 shall be retunred and errno set to indicate the
        error.

    ERRORS
        The getentropy() function shall fail if:

        [EINVAL]        The value of length is greater than
                        {GETENTROPY_MAX}.

        The getentropy() function may fail if:

        [ENOSYS]        The system does not provide the necessary
                        source of entropy.

    RATIONALE
        The getentropy() function is not a cancellation point.

Minor changes from the previous introduction of getentropy into libc:

- Return EINVAL, not EIO, on buflen > 256.
- Define GETENTROPY_MAX in limits.h.

The declaration of getentropy in unistd.h and definition of
GETENTROPY_MAX in limits.h are currently conditional on
_NETBSD_SOURCE.  When the next revision of POSIX is finalized, we can
expose them also under _POSIX_C_SOURCE > 20yymmL as usual -- and this
can be done as a pullup without breaking existing compiled programs.

(riastradh)

s/longeset/longest/ and s/collsion/collision/ in comments.

(andvar)

stipending(): if we're going to process some interrupts don't return 0.
Hopefully fixes random hang seen in i386 Xen PV.

The bug has been there ~forever but was masked by the fact that spllower()
did call event handlers much more often.

(bouyer)

s/disbale/disable/ and s/enbale/enable/ in comments. also one more typo fix.

(andvar)