Sync with OpenBSD r1.98. Lots of improvements including checksum offload,
hardware vlan tagging, and support for multiple receive queues.

(jmcneill)

make(1): inline Buf_Len

(rillig)

regen

(jmcneill)

Add some more product IDs for mcx(4)

(jmcneill)

Add symmetric toeplitz implementation with integration for NICs, from OpenBSD.

(jmcneill)

make(1): remove __predict_false

The effect (at least on x86_64) is so minimal that it is not worth
cluttering the code.

(rillig)

make(1): inline Buf_GetAll

(rillig)

make(1): only clean up the Buffer data in CLEANUP mode

Cleaning up the members is only useful during debugging but not during
use in production.

(rillig)

make(1): split Buf_Destroy into Buf_Done and Buf_DoneData

In all cases except one, the boolean argument to Buf_Destroy was
constant.  Removing that argument by splitting the function into two
separate functions makes the intention clearer on the call site.  It
also removes the possibility for using the return value of Buf_Done,
which would have made no sense.

The function Buf_Done now pairs with Buf_Init, just as in HashTable and
Lst.

Even though Buf_Done is essentially a no-op, it is kept as a function,
both for symmetry with Buf_Init and for clearing the Buffer members
after use (this will be done only in CLEANUP mode, in a follow-up
commit).

(rillig)

Add a few network interface offload feature initialisms

(jakllsch)

lint: rename incompl to is_incomplete

No functional change.

(rillig)

lint: fix type conversion for very large data types

Data types that are 4 GB or larger are an edge case.  Nevertheless,
compute their size correctly.

(rillig)

lint: remove redundant parentheses

No functional change.

(rillig)

lint: remove message 189, add tests for a few other messages

Message 189 would have applied to traditional C and was supposed to
detect assignments between struct and union types.  The corresponding
check had never been implemented though.

Traditional C has been superseded for 30 years now, therefore there is no
point in adding this check retroactively.

(rillig)

Add the standard compat string from the Device Tree bindings.

(thorpej)

Add standard compat strings from the Device Tree bindings.

(thorpej)

Add the standard compat string from the Device Tree bindings.

(thorpej)

lint: add tests that trigger a few more messages

(rillig)

make(1): run lint with strict bool mode

The make code is not supposed to use constructs such as "if (strcmp(s1,
s2))" or "if (p && *p)", instead all boolean expressions have a form
that would be accepted by a C#, Go or Java compiler as well.  This also
ensures that pre-C99 compilers generate equivalent code as C99
compilers, at least regarding boolean expressions.

(rillig)

make(1): fix lint warning about empty declarations

(rillig)

make(1): reduce boilerplate for printing bit sets in debug mode

No functional change.

(rillig)

make(1): update documentation of test lint.mk

(rillig)

make(1): clean up JobWriteShellCommands

(rillig)

make(1): clean up comments in JobSaveCommands and JobStart

(rillig)

make(1): add test for combining the options -j, -n, -t

This is a preparation for refactoring the complicated condition in
JobStart.

(rillig)

make(1): explain test opt-touch-jobs

(rillig)

Print the path that we failed to open on error

(jmcneill)

Fix "Cannot allocate memory" failure on amd64.

BIOCGBLEN ioctl of bpf(4) requires u_int, not size_t.
The problem is reported by Kenji Aoyama.

Should be pulled up to netbsd-9 and netbsd-8.

(tsutsui)

Don't print undefined interrupt value

(skrll)

Pretty print

(skrll)

Improve handling of receive overflows.  When we get an overflow, don't
reset the chip if we are still receiving packets.  Real overflows can
happen when network and CPU(s) are loaded, so the reset is not needed
in this case.
Add a counter to track receive overflows (when GEM_COUNTERS is defined).

This reverts some changes (always reset on overflow) from
PR port-sparc64/46260.

(jdc)

Add a proper compat string for this device, following the standard
conventions.

(thorpej)

If we're going to keep a reference on the "props" dictionary from
the i2c_attach_args, we should retain it.

(thorpej)

make(1): explain seemingly redundant condition in jobs mode

(rillig)

make(1): remove unnecessary fflush for empty shell files

If nothing is done to any file, there is no need to flush stdout.  Move
the call to each of the remaining branches, except for -t mode, where
Job_Touch already takes care of everything.

(rillig)

make(1): rename JobOpenTmpFile to JobWriteShellCommands

The old name didn't reflect that the function not only opens the
temporary file for writing but also actually writes the shell commands
to that file.

(rillig)

make(1): do not create empty shell files in jobs mode

In a chain of dependencies such as "all: build; build: do-build", the
intermediate targets do not have any commands.  In jobs mode,
nevertheless, an empty file was created and fed to the shell.  This was
unnecessary.  See jobs-empty-commands.mk.

The case of the special command line "...", which was suggested on
current-users, is not optimized since it doesn't occur in practice.

Suggested by Mateusz Guzik on current-users:
https://mail-index.netbsd.org/current-users/2021/01/26/msg040215.html

(rillig)

make(1): demonstrate unnecessary creation of empty files in jobs mode

(rillig)

Tickets #1189 and #1190

(martin)

Pull up following revision(s) (requested by lukem in ticket #1190):

usr.bin/ftp/progressbar.c: revision 1.24
usr.bin/ftp/ssl.c: revision 1.9
usr.bin/ftp/progressbar.h: revision 1.9
usr.bin/ftp/ftp.c: revision 1.171
usr.bin/ftp/version.h: revision 1.92

ftp: don't use restartable signals

Refactor to not rely upon restartable signals (SA_RESTART),
possibly fixing intermittent failures with -q QUITTIME.
ftp transfers: handle EINTR/EAGAIN in copy_bytes(),
instead of relying upon restartable signals.

http/https transfers: Explicitly print an error similar to
progressmeter() when timing-out for -Q QUITTIME in fetch_wait(),
and set errno to ETIMEDOUT so that the warn() in fetch_url()
prints a more accurate error message.

PR/55857

(martin)

Pull up following revision(s) (requested by lukem in ticket #1189):

usr.bin/progress/progress.c: revision 1.23

progress: handle EINTR in writes. PR/55914

(martin)

Ticket #1647

(martin)

Pull up following revision(s) (requested by nonaka in ticket #1647):

sys/dev/hyperv/if_hvn.c: revision 1.20 (via patch)
sys/dev/hyperv/hvkbd.c: revision 1.7 (via patch)
sys/dev/hyperv/vmbus.c: revision 1.12 (via patch)

hvkbd(4): Don't wait forever.
vmbus(4): Don't wait forever.
hvn(4): Don't wait forever.

(martin)

Use acpi_compatible_match().

(thorpej)

Use acpi_compatible_match() / acpi_compatible_lookup().

(thorpej)

Use acpi_compatible_match().

(thorpej)

fdtbus_intr_establish -> fdtbus_intr_establish_xname

(skrll)

fdtbus_intr_establish -> fdtbus_intr_establish_xname

(skrll)

Fix build without MULTIPROCESSOR.

(rin)

Sort headers. No binary changes.

(rin)

the last_fault_code DEBUG variable is not MP safe, so don't compile it in.

(skrll)

More debug

(skrll)

Fix previous; sunxidrm(4) couldn't be probed on Allwinner A64 due to
trailing whitespace in compat string.

(rin)

Typo in comment

(skrll)

Trailing whitespace

(skrll)

hvn(4): Don't wait forever.

(nonaka)

vmbus(4): Don't wait forever.

(nonaka)

hvkbd(4): Don't wait forever.

(nonaka)

Use acpi_compatible_match().

(thorpej)

Don't try print an error message when there is none.

(christos)

PR/55962: Andreas Gustafsson: Some tar(1) error messages lack file name and
newline

(christos)

Fix sanity check to match the fixed off-by-one logic in the sizing
logic.

(joerg)

fix logic botch from previous - if we're asked to draw the cursor somewhere
else and it's still visible we need to clear it

(macallan)

In device_is_a(), handle dev or dev->dv_cfdriver being NULL.  This
makes the calling pattern:

device_is_a(device_parent(dev), "whatever")

safe.

(thorpej)

Rewrite and streamline virtio device config read/write and explicitly cater
for the Aarch64-eb bus problem with Qemu. This removes lots of bus_space
`magic' and cleans up the code.

(reinoud)

No need to consult ia->ia_ncompat before calling iic_compatible_lookup().

(thorpej)

Use iic_compatible_lookup() in ssdfb_i2c_attach().

(thorpej)

Use iic_compatible_lookup() in admtemp_setflags().

(thorpej)

port-arm/55957: Odroid C2 can not access eMMC card

A few changes to clock setup:
- Disable clock while changing dividers
- Set / clear DDR flag before changing clock
- Adjust TX/RX/core phases

(jmcneill)

Modify the generated make wrapper to avoid complaints about
non-writable OBJDIR from make (which happen a lot when cross-building
kernels re-using a makewrapper build for some other user).

(martin)

Remove x86_genfb_mtrr_init. PATs have been available since the Pentium III
and this code has been #if notyet'd shortly after being introduced.

(jmcneill)

add optional glyph cache for genfb
enable with options GENFB_GLYPHCACHE=n
with n being the desired size of the cache in MB. Should be enough to cache
at least 900 glyphs in whatever video mode used in order to be effective
in 32bit per pixel that's about 1MB

(macallan)

Don't silently ignore the errors from npfctl_run_op. We end up returning
packets to userland that are missing required fields (like in rule_add the
id of the rule) and npfctl aborts.

(christos)

Add verbose output options commented out.

(skrll)

binutils-2.36 out.

(wiz)

if_rge.c: avoid aprint_error*() abuse

(jakllsch)

Trailing whitespace... heh

(skrll)

vcaudio: Do not report recording mode in hardware format

(nia)

Style.  NFCI.

(skrll)

Fix non-DIAGNOSTIC build

(skrll)

Note support for QEMU "mipssim" emulator.

(simonb)

Add support for the QEMU MIPS "mipssim" simulator.
TODO- mipsnet network driver (root on md(4) only for now).

(simonb)

Introduce weighted matching for ACPI autoconfiguration, and provide
acpi_compatible_match() based around device_compatible_entry.  Matches
against _HID score big, matches against _CID are weighted in the
standard most-to-least-specific ordering, less than _HID.

Also provide a maching value for _CLS, that's always less than _HID
and _CID matches, and use that in acpi_match_class().

Also provide acpi_compatible_lookup(), that returing the matching
entry based on the same criteria.

(thorpej)

Introduce autoconfiguration helpers based around device_compatible_entry:
- pci_compatible_match(): matches against the PCI ID.
- pci_compatible_match_subsys(): matches against PCI SUBSYS ID.
- pci_compatible_lookup(): look up entry by PCI ID.
- pci_compatible_lookup_subsys(): look up entry by PCI SUBSYS ID.
- pci_compatible_lookup_id(): look up entry by an arbitrary ID using the
  PCI ID code conventions.

- Define PCI_COMPAT_EOL as a compat data array sentinel.

(thorpej)

G/C of_match_compatible().

(thorpej)

Add device_compatible_match_id() and device_compatible_lookup_id(), which
are like device_compatible_match() and device_compatible_lookup(), but
take a single scalar value as an ID.

Add a "uintptr_t id" field to device_compatible_entry, in an anonymous
union with the existing "const char *compat" field.

(thorpej)

Use eisa_compatible_{match,lookup}().

(thorpej)

Add eisa_compatible_match() and eisa_compatible_lookup().

(thorpej)

There is not much point in of_compatible() returning -1 for "no match"
and >= 0 for "match".  Just make it return 0 for "no match" and >0 for
"match" so it can be treated like a boolean expression.

(thorpej)

Rename of_match_compat_data() to of_compatible_match().  Similarly,
rename of_search_compatible() to of_compatible_lookup().

Standardize on of_compatible_match() for driver matching, and adapt
all call sites.

(thorpej)

Rename of_match_compat_data() to of_compatible_match().  Similarly,
rename of_search_compatible() to of_compatible_lookup().

Standardize on of_compatible_match() for driver matching, and adapt
all call sites.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Replace 2 uses of of_match_compatible().  Also use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Use DEVICE_COMPAT_EOL.

(thorpej)

Define a macro to terminate the common usage of device_compatible_entry
arrays, in order to placate the angry hordes objecting to use of a GCC
extension.

(thorpej)

make(1): fix irrelevant message in -W mode

(rillig)

make(1): demonstrate irrelevant message about warnings as errors

(rillig)

make(1): in -dp mode, print stack trace with each diagnostic

This makes it easier to track down where a warning or an error
originated from.  This information could be further enriched for .for
loops, to also include the variable names and their values.  For now,
it's good enough to replace the large comment describing how a stack
trace _could_ be generated with actual code that implements that idea.

The syntax of the locations is <filename>:<lineno>, which intentionally
differs from the traditional "<filename>" line <lineno>, since the
former is more widely supported by editors and IDEs.

Having this stacktrace information is especially intended for
complicated systems built upon make, such as pkgsrc.

(rillig)

Fix potential off-by-one error when using hash fudging. It needs to
round up to 2/4 and not one less to guarantee that the adjusted hash
fits into array.

(joerg)

call f_touch with kq_lock held, and without KERNEL_LOCK() - for this
adjust EVFILT_USER, which is the only filter actually using that hook

kqueue_scan() now doesn't need to exit/enter the kq_lock when calling
f_touch, which removes another possible race

part of PR kern/50094

(jdolecek)

lint: remove __noinline attribute from string interning function

I had committed this accidentally while ensuring that the generated code
is still efficient even though the source code looks heavy with the
double pointer indirection.

(rillig)

remove outdated comment

(macallan)

Fix indexing bug in clean up code on error in virtio PCI v1.0 attach code.
Thanks go to Rin Okuyama for spotting it.

(reinoud)

- refactor vcons_cursor_noread() a bit so it works the same way as the other
  _noread() methods instead of duplicating half of vcons_cursor()
- skip clearing the cursor only if we use all _noread() methods
- clear RI_CURSOR in all copy*_noread() methods as well if we overwrite the
  cursor

(macallan)

Trailing whitespace

(skrll)

KNF

(skrll)

llvmlockdir -> lvmlockdir

(martin)

Replace use of of_match_compatible().

(thorpej)

Replace 2 uses of of_match_compatible().

(thorpej)

There is not much point in of_compatible() returning -1 for "no match"
and >= 0 for "match".  Just make it return 0 for "no match" and >0 for
"match" so it can be treated like a boolean expression.

As such of_match_compatible() (a wrapper around of_compatible()) is now
obsolete, and will be removed once all call sites are converted to an
appropriate replacement.

(thorpej)

Minor comment changes.

(thorpej)

Rename etc/rc.d/llvmlockdir to lvmlockdir - it has nothing to do with LLVM,
the name probably was a freudian slip of my fingers when I split that
fragment out of another rc.d script.

(martin)

Regen for previous commit to GENERIC.in.

(rin)

Regen for previous commit to GENERIC.in.

(rin)

Now that the ACPI layer no longer does this for us, evaluate the _DSM
that returns the HID descriptor address at attach time.

(thorpej)

Don't map ACPI _HID / _CID strings to "compatible" strings from the
Device Tree bindings.  Just pass those strings in the standard
most-specific-to-least-specific order to the drivers and let them
deal.

(thorpej)

Back out unintended change in previous.

(thorpej)

acpipchb is a bus, so claim all child ACPI device nodes

(jmcneill)

build fix

(jmcneill)

Add a device_t parameter to acpi_enter_i2c_devs. If non-NULL, all child
acpi_devnodes will be claimed by that device so we don't later try to
attach a duplicate device to that node at acpinodebus.

(jmcneill)

In the lam(1) example with a literal newline, use newish sh syntax.

Bump date (as of when I wrote the change and discussed it, in October;
forgot to commit since)

(dholland)

avoid needing <machine/promlib.h>.  fixes sparc64 build.

(mrg)

s/boolean_t/bool/.  boolean_t is obsolete (from Mach), and this
also fixes a new build issue in libkvm on sparc*.

(mrg)

put back clearing of KN_QUEUED and check for re-queue - as rev. 1.53 notes,
it's necessary for correct function

fixes PR kern/55946, thanks to Paul Goyette for testing

part of PR kern/50094 fix

(jdolecek)

make(1): split For_Eval into separate functions

(rillig)

Fix a thundering herd problem in pipes.

Wake only one waiter when data becomes available, not all of them.
Waking them all is not a usual case, but turns up with make's job
token pipes. (Probably make's job signalling scheme should also be
revised, assuming rillig hasn't already done that, but that's a
separate issue.)

This change will not do us much good for the moment because we don't
distinguish cv_signal from cv_broadcast for interruptible sleeps, but
that's also a separate problem.

Seen on FreeBSD; from mjg at freebsd a couple months ago. Patch was
mine (iirc) but the real work in this sort of thing is discovering the
problem.

(dholland)

make(1): extract ForLoop_New to separate function

(rillig)