make: backport to C90

In the past few months I had accidentally used C99 features in the make
code.  According to tools/README, tools that are used in the build
system should restrict themselves to C90.

This allows make to build with GCC's options "-pedantic
-Wno-system-headers -Dinline= -Wno-error=cast-qual".

I didn't notice anyone actively complaining though, I just wanted to see
how much work this backporting would be.  The identifier __func__ is
still used, as in other tools.

No functional change.

(rillig)

make: rename function parameters to match boolean constants

No functional change.

(rillig)

Add a -L option to follow all symbolic links.  Useful if you have symlinks
in a makefs directory tree but want to refer to the actual file.

(simonb)

Deal with the case when a device gets recreated (eg when mtree is called by
MAKEDEV) and the stat buffer contains information about the original device
before checking/setting the owner/group/mode of the new device file.

(simonb)

Centralise the setitimer() timer type validation in dosetitimer() as is
done with dogetitimer().

(simonb)

Print the "preposterous TOD clock time" message only if the RTC is
before very late 2020 (1/1/2021 minus a few leap days) instead of
late 1994.  Someone(TM) should remember to adjust this in somewhat
less than 26 years.

(simonb)

make: use standard %H in strftime instead of %k

The conversion specifier %k is neither in C99 nor in C90.

These timestamps are only used in the debug output.

(rillig)

CTASSERT that NSIG <= 128.  There are many hard-coded assumptions that
there are <= 4 x 32bit signal mask bits.

(simonb)

make: use C99 bool type instead of defining its own

No functional change.

(rillig)

For MIPS N32, register_t needs to be uint64_t as "long" isn't
big enough.

(simonb)

config_attach_loc() -> config_attach() with CFARG_LOCATORS argument.

(thorpej)

Revert parsedate(3) XXX as per discussion with kre@.

(simonb)

Rework about the rest note in speaker(4).
- Obsolete the sc_rest callback.  The rest note operation can be done by
  the common spkr layer.  This also fixes PR kern/56060.
  This work-in-progress patch was left in my local tree for years. :(
- Improve calculations of tone and rest length.

(isaki)

Improve SPKRDEBUG code.
- Replace wrong aprint_debug_dev() with device_printf().
  By this, it no longer need to print dev_t.
- Improve some messages.

(isaki)

Give config_attach() the tagged variadic argument treatment and
mechanically convert all call sites.

(thorpej)

ismt_config_interrupts(): remove a now-unused variable.

(thorpej)

lint: reword message about constant argument to '!'

In the other messages, operators are typically written in their literal
form as well.  Using single quotes disambiguates them from normal
punctuation.

(rillig)

tests/lint: add test for constant argument to '!'

(rillig)

config_found_ia() -> config_found() w/ CFARG_IATTR.

(thorpej)

lint: reduce memory usage

No functional change.

(rillig)

tests/lint: prevent typo from tree.c 1.264 from happening again

(rillig)

lint: fix typo from previous commit in merge_qualifiers

(rillig)

lint: clean up merge_qualifiers

No functional change.

(rillig)

clarify comment in previous
( this is a hack, should go away when the root cause is fixed etc. )
no functional change

(macallan)

lint: rename local variable in build_colon

The name 'rtp' was confusing since it did not correspond to 'rn'.  The
'r' meant 'result', not 'right-hand side'.

No functional change.

(rillig)

lint: replace a new modtab with op_name

This reduces the types of variables that are passed around.

No functional change.

(rillig)

lint: rename getopname to op_name

There are several functions called 'get' that allocate memory.  Remove
this possible confusion.

No functional change.

(rillig)

lint: reorder struct members to be in comprehension order

No functional change.

(rillig)

Fix style to appease changes2html.

(rin)

lint: rename members of brace_level

The terms 'member' and 'subscript' are distinctive enough.

No functional change.

(rillig)

lint: remove unused brace_level.bl_array_of_unknown_size

Since init.c 1.177 from 2021-03-29, the type of the object to be
initialized is no longer modified in the middle of the initialization,
as required by C99 6.7.8p22.  Therefore it is no longer necessary to
keep this redundant information around.

No functional change.

(rillig)

tests/lint: add test for enum type mismatch

(rillig)

lint: add parentheses after sizeof, as required by share/misc/style

No functional change.

(rillig)

For ports with __HAVE_LEGACY_INTRCNT, turn intrcnt[] and derived
variables into u_int, to match with kern/subr_evcnt.c.

(rin)

lint: rename functions for duplicating types

No functional change.

(rillig)

Make extern declaration of intrcnt a u_int instead of a long.  Matches
some declartions in some ports.  Fixes build on news68k (and probably
newsmips).

(simonb)

lint: rename members of struct memory_block

No functional change.

(rillig)

lint: name memory allocation functions consistently

No functional change.

(rillig)

named/filter-aaaa.so is unrelated to DYNAMICROOT

(martin)

lint: rename getnode to expr_zalloc_tnode

The new name highlights that the returned memory is only valid in the
scope of the current expression.  This was misleading before since the
other related functions all have a 't' (probably for 'temporary') in
their names.

Also encode in the function name that the returned memory is zeroed out
as that could not be inferred from the old name.

No functional change.

(rillig)

lint: extract filename ID counter into separate function

No functional change.

(rillig)

lint: rename mbl to memory_block

No functional change.

(rillig)

Call usbnet_attach before usbnet_{lock_core,busy}

kern/56090: panic in udav_attach

(skrll)

Add gpt wedge support

(martin)

lint: make debug logging compatible with C90

The previous code used the GCC-style varargs macros, which did not even
conform to C99.

No functional change.

(rillig)

rndctl: make -l's "Flags" field reflect current reality

don't print "legacy options" RND_FLAG_ESTIMATE_TIME and
"RND_FLAG_ESTIMATE_VALUE"

only print "estimate" if we have actually counted any bits from
something, since it's no longer really possible to "enable estimation".

ideally, there should also be a "samples" field so it's clear
collected bits are not being counted.

(nia)

Document DHT kernel for evbppc.

(rin)

- Hook DHT kernel
- Add bin and ub to KERNEL_SUFFIXES
- Group and sort kernels

(rin)

rndctl.8: reflect current reality

(nia)

evbppc: Add support for DHT Walnut 405GP evaluation board.
(Digital Home Technologies PCB 01070201 Rev. 1.1)

Official web page seems gone, but analyses by Linux people are at:

- https://elinux.org/DHT-Walnut
- http://web.archive.org/web/20070225171826/http://supernova.stanford.edu/dingdong/

The board features:

- 266MHz PowerPC 405GP processor
- one PC133 SDRAM slot
- two 32-bit 5V PCI slot
- on-chip ethernet controller with DP83843 PHY
- on-chip serial port / GPIO controller
- on-board PDC20265 IDE controller

Hardware limitations:

- no MAC address assigned
- no RTC present

Known problem:

- atabus(4) channels cannot be attached to pdcide(4) for cold boot

dmesg:

- https://dmesgd.nycbug.org/index.cgi?do=view&id=5997

Have fun!

(rin)

Add bpf

(martin)

Add an XXX reminder to convert at(1) to use parsedate(3) in <util.h>.

(simonb)

Drop setgid kmem, simplify nlist setup.

(simonb)

Add bit-field definitions for DCR_SDRAM0_B[0-3]CR registers.

(rin)

Set f_namemax during mount time like all the other filesystems so that
it does gets the right data in copy_statvfs_info(). Otherwise f_namemax
can end up being 0. To reproduce: unmount the remote filesystem, remount
it, and kill -HUP mountd to refresh exports.

(christos)

regen

(rin)

PDC20265 is Ultra/100, not 66.

(rin)

avoid mapping 0xfffff000 - my beige G3 DSIs on it
with this my the machine boots again
tested on a variety of G4 and G5 models with no problems

(macallan)

Put a copy of our existing data first in the non-error case (noticed by RVP).

(christos)

lint: remove wrong assumption from comment

The size in bits of a struct or union is not measured at all at this
point since portable_size_in_bits only takes the broad type
classification (tspec_t), not the precise type information (type_t).

No functional change.

(rillig)

lint: replace empty macro arguments with non-empty arguments

Empty arguments are a new feature of C99.  Since lint is one of the
build tools, it is supposed to only use C90 features.  C99 6.10.3 "Macro
replacement" explicitly allows empty macro arguments while C90 3.8.3
"Macro replacement" last sentence defines an empty macro argument as
undefined behavior.

This change makes the '1's stick out less from the table.  I tried to
use '.' instead of '-' as well, but that made it too hard to see the
commas.

No functional change.

(rillig)

lint: do not error out of a struct is initialized without braces

This allows to process lib/libc/gen/sysctl.c 1.38 from 2021-03-30, as
well as its precedessor 1.37, which had a workaround just for lint.

While unusual, C99 allows these.

(rillig)

Move the named filter-aaaa plugin to the shared list

(martin)

Add a sysctl hashstat collector for ubchash.

(simonb)

Add a sysctl hashstat collector for vcache.

(simonb)

Add a sysctl hashstat collector for bufhash.

(simonb)

Add a sysctl hashstat collector for uihash.

(simonb)

Use kernel sysctl hashstat collection instead of kmem grovelling
directly.  Also GC a few old hash nlist entries that no longer exist.

(simonb)

Add support for kernel hash statistics generation for vmstat -h/-H.  As
well as not needing any kmem grovelling, also much faster as it doesn't
need a kmem read for hash bucket and for each item in the hash bucket
chains.

(simonb)

Now that ports that use legacy intrcnt interrupt accounting have event
counters to show that data, for the live kernel case just show INTR
events for "vmstat -i".

(simonb)

Define __HAVE_LEGACY_INTRCNT to indicate this port uses legacy intrcnt
interrupt accounting.

(simonb)

Expose olde style intrcnt interrupt accounting via event counters.
This code will be garbage collected once our last legacy intrcnt
user is update to native evcnts.

(simonb)

Whitespace: #define<tab>

(simonb)

New vax documentation, vax/covid(4)

(riastradh)

Tickets #1238 - #1240

(martin)

Pull up following revision(s) (requested by christos in ticket #1240):

sys/arch/sparc/sparc/trap.c: revision 1.199

PR port-sparc/55573: remove kernel message about disabled coprocessor
instructions - it is triggered by userland trying to detect availability
of sparcv9 VIS instructions.

(martin)

Apply patch, requested by tsutsui in ticket #1239:

sys/arch/sparc/dev/audioamd.c (apply patch)

Fix for PR 56078 (different solution applied to -current): fix wrong
lock usage.

(martin)

Pull up following revision(s) (requested by skrll in ticket #1238):

sys/arch/arm/cortex/gic.c: revision 1.47

Only target the boot cpu for real with SPI interrupts.  I tried to do
this back in 2014, but somehow I missed a spot.

This is a quick-and-dirty fix for the USB stack which expects transfer
completions to be in-order.  If interrupts happen across the CPUs then
this isn't guaranteed (yet).

kern/55243 panic at usb_transfer_complete() on raspberry pi 4

(martin)

Adjust the number of entries in the intrcnt array to match the number
of names in the intrnames list.

(simonb)

KNF a bit. No functional change.

(msaitoh)

Modify error message to sync with ixgbe.c

(msaitoh)

Reenable kernel modules for MIPS now that they work.

(simonb)

Use compat/exec.mk - zfs isn't compat32 friendly (yet?).

(simonb)

Simplify the CTF float encoding type logic: all floating point types of
4 bytes or less are "float", 8 bytes or less are "double" and 16 bytes
or less are "long double".  Make ctfconvert much happier with N32 ABI
where we have 4 byte pointers and 16 byte long doubles.  Will also help
ctfconvert if ever used on m68k or 32-bit RISC-V.

@christos: LGTM

(simonb)

add named plugin dir

(christos)

handle copying of plugins to the chroot

(christos)

new named plugin (simon@)

(christos)

Add plugin glue (simon@)

(christos)

There are two interface attriubtues on this device, so we need to be
explcit about it in BOTH config_found() calls.

(thorpej)

mention regex changes

(christos)

sun4v: update TODO after pcb spill improvements and status when running on a S7 host

(palle)

lint: clean up duplicate code in handling of initialization

No functional change.

(rillig)

sun4v: handle spill trap properly when trap level is 1 and otherwin is non-zero - ensure that all windows are spilled to the pcb like the sun4u winfixsave code path

(palle)

tests/lint: test 'scalar type cannot use designator'

(rillig)

tests/lint: remove unnecessary empty lines from test

In a separate commit to not cause an unnecessarily large diff in the
previous commit.

(rillig)

lint: add error about out-of-bounds array subscripts

This check is not strictly necessary since any C99 compiler must
diagnose them as well, it is rather meant for demonstrating how to do
the check in lint, and for symmetry with the 'unknown member' error
message.  These provide insight into how the data structures in init.c
are meant to be accessed.

(rillig)

libc/gen: fix hack for previously unsupported lint initializers

Supported since init.c 1.182 from 2021-03-30.

(rillig)

Make this format a little bit nicely.

(rin)

lint: add type information for message about unknown member name

(rillig)

lint: reword message for very unlikely .member in array initialization

(rillig)

lint: add type information to message 175

(rillig)

tests/lint: remove outdated comments

(rillig)

make: remove workaround for bug in lint

The bug has been fixed in lint.

No functional change.

(rillig)

Use ``for (;;)'' for infinite loop, as required by style.
No binary changes.

(rin)

- Write-back msgbuf and update bootstr for reboot from halt.
- Remove dead code.

(rin)

lint: rewrite handling of initializations, fixing several bugs

The previous implementation had a wrong model of how initialization
happens in C99, its assertions failed in all kind of edge cases and it
was not possible to fix the remaining bugs one at a time without running
into even more obscure assertion failures.

The debug logging was detailed but did not help to clarify the
situation.  After about 20 failed attempts at fixing the small details I
decided to start all over and rewrite the initialization code from
scratch.  I left the low-level parts of handling designators, the code
that is independent of brace_level and the high-level parts of how the
parser calls into this module.  Everything else is completely new.

The concept of a brace level stays since that is how C99 describes
initialization.  The previous code could not handle multi-level
designations (see d_init_pop_member.c).  There are no more assertion
failures in the initialization code.

Some TODO comments have been left in the tests to keep the line numbers
the same in this commit.  These will be cleaned up in a follow-up
commit.

The new implementation does not handle initialization with "missing"
braces.  This is an edge case that both GCC and Clang warn about, so it
is not widely used.  If necessary, it may be added later.

The new implementation does not use any global variables in the vast
majority of the functions, to make all dependencies and possible
modifications obvious.

(rillig)

Fix nearly 20 year old type - cache sizes are 2048 bytes, not 2848 bytes.

(simonb)

- G/C unused.
- Sort headers.

No binary changes.

(rin)

- Explicitly include <powerpc/spr.h>.
- Drop unused headers.
- Group headers.

No binary changes.

(rin)

- Include tlb.h directly, instead of uvm_extern.h.
- Drop unused headers.
- Sort and group headers.

No binary changes.

(rin)

- Include tlb.h directly, instead of uvm_extern.h.
- Explicitly include param.h.
- Drop unused headers.
- Sort and group headers.

No binary changes.

(rin)

- Explicitly include <sys/param.h>.
- Drop unused headers.
- Sort headers.

No binary changes.

(rin)

- Include tlb.h directly instead of uvm_extern.h.
- Remove unused/duplicated header include.
- Sort and group headers.

No binary changes.

(rin)

- Explicitly include <sys/param.h>.
- G/C headers.
- Sort headers.

No binary changes.

(rin)

Include <dev/ic/comreg.h> for COM_FREQ.

(rin)

Sort & group headers. No binary changes.

(rin)

G/C unused. Use tlb.h directly instead of uvm_extern.h.
No binary changes.

(rin)

G/C <evbppc/obs405.h>.

(rin)

Sort headers. Fix style. No binary changes.

(rin)

G/C misleading <evbppc/tlb.h>. Use <powerpc/ibm4xx/tlb.h> directly.

(rin)

G/C unused. No binary changes (except for __LINE__ for KASSERT).

(rin)

Move common global variables, i.e., phys_map, machine, and machine_arch to
evbppc/evbppc_machdep.c (bad naming; used only for 4xx boards).

(rin)

According to "PPC405GP Embedded Processor User窶冱 Manual",
405GP has 16KB instruction cache, not 8KB.

(rin)

Support OPB running @ 33MHz for 405GP based boards.
No need to modify EMAC_MR1 register this case
(STACR_OPBC fields are reserved for 405GP).

(rin)

Convert walnut to use powerpc/ibm4xx/openbios.
No functional changes.

(rin)

Use mnemonic "bdneq". Add missing register prefix.

No binary changes.

(rin)

- Document IBM/AMCC Walnut is supported by this file.
- Misc style fixes.

(rin)

Remove unused "startkernel" argument for openbios_board_init().

(rin)

Centralize cpu_reboot().

(rin)

G/C bootpath.

(rin)

tests/lint: test initialization without braces

(rillig)

Remove obsoleted file; we had switched to powerpc/ibm4xx/pci/pchb.c long ago.

(rin)

tests/lint: demonstrate assertion failures for union initialization

(rillig)

tests/lint: test unknown member name in struct

I had not expected to trigger another assertion, I just wanted to make
sure my latest ongoing refactoring will not break this case.  Apparently
there is no need to worry about that.

(rillig)

tests/lint: test initialization of union in traditional C

(rillig)

tests/lint: ensure initialization does not modify shared type

In my not yet published rewrite of lint's init.c, I forgot to copy the
array type. Guard against this bug, which would have been hard to find.

Given that in C, the declaration 'int a[], b[]' creates two different
type objects anyway, it's not easy to come up with a test case that
actually triggers this possible bug.  I'm not sure whether this test
indeed catches this bug.  If not, I'll add another test.

(rillig)

lint: rename variable 'level' to 'bl'

For consistency with its type prefix and the other variables.  This
variable is used so often that it makes sense to abbreviate it.

No functional change.

(rillig)

lint: add helper functions

No functional change.

(rillig)

lint: rename struct members in init.c

No functional change.

(rillig)

lint: remove outdated comments, clean up style

No functional change.

(rillig)

tests/lint: add examples for initialization from C99

These all pass lint without any warnings.  Both Clang and GCC warn about
these in -Wall mode, so they are unlikely to be seen in practice.

(rillig)

PR/56083: RVP: tar -c hangs when reading from FUSE-ext2 filesystems since
it returns 0 for statvfs members and libarchive handles only -1. Centralize
the transfer setting code.

(christos)

Use ibm4xx_cpu_startup() and board_info_init().

(rin)

Use IBM405GP_UART0_BASE instead of hard-coded magic number 0xef000000.

No functional changes since:

IBM405GP_UART0_BASE & ~(TLB_PG_SIZE - 1) == 0xef000000

See ppc4xx_tlb_reserve() in ibm4xx/pmap.c.

(rin)