Change octal and hex parsing to not use strtoul so that they don't handle
'-'. From Martijn van Duren.
Also add a warning if the conversion fails (like the gnu printf does)

(christos)

go back to right-aligning the titles for CPU and WCPU.

(christos)

Limit buffer size for device capabilities requests as a work-around for PR
kern/56109.

(reinoud)

fix typo

(ryo)

when the packet size was just n*MCLBYTES, there was an inconsistency in the length of mbuf.

(ryo)

- Remove obsolete cltom(9) and mtocl(9).
- Add missing manual page's link of m_adj.

(msaitoh)

spello in comment

(skrll)

Restore our local change to remove GNU_STACK program header:

http://cvsweb.netbsd.org/bsdweb.cgi/src/external/gpl3/gcc/dist/gcc/config/rs6000/rs6000.c#rev1.13

This was removed accidentally when merging GCC10 into our tree.

Note that some firmwares, e.g., Explora, refuse to execute ELF
kernel image if this program header is present.

Now, GCC10 becomes just fine for all three powerpc sub-archs, i.e.,
oea, booke, and ibm4xx, as far as I can see. No regressions are
observed for ATF.

OK mrg

(rin)

Fix regression introduced to GCC10, where it wrongly recognizes 32-bit
processors as POWER9 if -misel flag is specified.

rs6000_machine_from_flags() assumes ISEL instructions are supported only
for POWER9 and successors. However, ISEL is also implemented for 32-bit
booke processors.

Since our kernel for booke is compiled with -misel, this regression
completely breaks it.

As a fix, check whether CPU is 64-bit capable or not, before checking
-misel flag.

The problem has been reported as 100108 to upstream:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100108

OK mrg

(rin)

Added missing locking order between sppp and IFNET_LOCK

(yamaguchi)

Stop and destroy timeout after sppp_detach and if_detach
for safety

The functions may use resources of pppoe(4) while detaching,
so the release should move after it.

(yamaguchi)

Fix regression introduced to GCC10, by which inline assembler codes for
403/405 are miscompiled.

Redundant .machine directive introduced by upstream commit
2d94f7dea9c73ef3c116a0ddc722724578a860fe:

https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=2d94f7dea9c73ef3c116a0ddc722724578a860fe

clobbers CPU flags passed to assembler. This results in miscompile for
inline assembler codes specific to 403/405 processors, at least.

Therefore, revert this commit locally.

The problem has been reported as 100107 to upstream:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100107

OK mrg

(rin)

Remove unnecessaly lock holdings to avoid dead lock

The locks were held while callout_halt() and workqueue_wait()
without reason.
And the locks also were held at callout and workqueue handler
so that the handler kicked by those function couldn't acquire
the lock.

The reasons why those are unneccesary are:
- Items of callout_t are protected by callout_lock
- Items of struct workqueue and struct work are protected
  by q_mutex in struct workqueue
- Items of struct sppp_work protected by atomic_cas(3)
- struct pppoe_softc does not free before workqueue_wait() and
  callout_halt() even if the locks are not held

(yamaguchi)

Fix not to put the wrong error message

(yamaguchi)

Stop ppp layer at first of destroying pppoe interface

(yamaguchi)

Sort initialization sequence in pppoe_clone_create() out
for refactoring

It has no functionality impact

(yamaguchi)

Use kmem_zalloc to allocate pppoe_softc

(yamaguchi)

Move initialization of sc_lock in pppoe_softc to first

The lock may be held in callbacks for ppp layer or other
components so that it should be initialized early.

(yamaguchi)

commonize error handling in pppoe_clone_create()

(yamaguchi)

restore gcc-8 subdirs and mark as obsolete.  oops.

noted by pgoyette.

(mrg)

check the result value of pmap_extract() and panic if it fails.
this is a "should never fail" case, and GCC 10 noticed that it
allows an uninitialised variable use.

tested by rin@ on amiga and mac68k.

(mrg)

make: remove type name for the abort status in job handling

(rillig)

tests/make: demonstrate handling of null bytes

(rillig)

make: rename PrintOutput to PrintFilteredOutput to avoid confusion

In the default configuration, the function PrintOutput did nothing.
Only if the shell has defined an output filter, something happens at
all.

(rillig)

make: document two previously unknown bugs in job mode

(rillig)

fix reversed comment

(christos)

fix reversed comment

(christos)

- Define NEED_ISYSTEM here
- include <bsd.init.mk> now that we can do it, because we need Makefile.rump
  to be included first, so that NOLINT gets defined, so that we don't end up
  building lint modules just for this.

(christos)

- Don't define NEED_ISYSTEM. This is a rump thing and should be only exposed
  there.

(christos)

- Don't include <bsd.own.mk> It should be done by the enclosing Makefile
- Don't define NEED_ISYSTEM. This is a rump thing and should be only exposed
  there.

(christos)

Fix two problems on old topcat(4) framebuffers found on HP332/340.

- Fix panic on monochrome framebuffers. They don't have palette registers.
  The problem was reported from Anders Gustafsson and also Andrew Gillham
  back in 2013:
  https://mail-index.netbsd.org/port-hp300/2013/09/27/msg000086.html
  https://mail-index.netbsd.org/port-hp300/2013/09/28/msg000087.html
- Fix incorrect framebuffer width of 98542/98544 framebuffers on HP332.
  Reported from Andrew Gillham (98542) as above and Chris Hanson (98543):
  https://mail-index.netbsd.org/port-hp300/2013/09/27/msg000083.html
  https://mail-index.netbsd.org/port-hp300/2020/05/01/msg000164.html

Worth to pullup to netbsd-9.

(tsutsui)

mtu for revision B0 and B1 is supported up to 16338 bytes

(ryo)

fixed the problem of instability when changing mtu in IFF_UP state

(ryo)

add support hardware tcp/udp rx checksum offloading

(ryo)

Fix fallout from alpha/interrupt.c rev 1.93:

http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/arch/alpha/alpha/interrupt.c#rev1.93

by which interrupts *seem* to occupy 100% CPU time.

Now, we bump ci_intrdepth for clock interrupt. Therefore, if ci_intrdepth > 1
is observed in statclock(), CPU is actually occupied by interrupts.

Thanks jklos and thorpej for notice!

(rin)

avoid second typedef for boolean_t in kernel code originally defined in
<sys/types.h>

(christos)

remove extra typedef for vnode_t done in <sys/vnode.h>

(christos)

document various gcc warnings that occur frequently enough to
inspire the GCC_NO_* series of variables, and explain some
about what the issues may be for each, hopefully providing some
direction on how to fix new warnings in this space.

christos asked me for this last year, and it's finally complete
enough to commit.  :-)

(mrg)

feedback from rin@: alpha and m68k work well, powerpc not so much
but fixes are known.

thanks!

(mrg)

i've test-built all platforms.  everything except arm64, x68k and sun2
builds.  many of them work as well.  most mips are (minimally) tested.

(mrg)

convert 'extern phys_ram_seg_t mem_clusters[0];' to the '[]' form
so that when this code accesses mem_clustes[0] and mem_clusters[1]
it does't reach out of bounds.

XXX: should be in a header, but it's ugly to fix and i stopped
    when bus_dma.c wouldn't build.

(mrg)

vax vs gcc:  fsdb.c loses.  also build this with -O0.

(mrg)

one port complains about string overflow here that i can't see.

(mrg)

more -fcommon for boot programs.

(mrg)

intrcnt[] is changed from long to u_int. Use sizeof(var) instead of
sizeof(type) to catch up with this change.

No binary changes as all ports with __HAVE_LEGACY_INTRCNT are ILP32, IIUC.

(rin)

dk_start(): retry device-dependent start() routine later, also when it
returns ENOMEM in addition to EAGAIN.

Device-dependent start() routine may allocate buffer directly, or via
bus_dma(9) API (some implementations for bus_dma(9) like alpha allocate
memory internally).

If these attempts fail with ENOMEM, this is not a disk error, therefore
we must retry later, as already done for EAGAIN.

(rin)

Contrary to comment here, nowadays hardclock() and statclock() should be
called with cpu_intr_p() is turning on, if used from the interrupt context.

Otherwise, entropy_enter() is used instead of entropy_enter_intr(), which
results in KASSERT() failure.

Therefore, bump ci_intrdepth for clock interrupt. Remove stale comment also.

(rin)

Fix inverted logic!!

vmem_xalloc(9) should be used with VM_NOSLEEP for BUS_DMA_NOWAIT.

(rin)

sandpoint has its own module_init_md(). So, do not compile in
that function in powerpc_machdep.c.

(rin)

Use uvm_km_alloc(9) with UVM_KMF_NOWAIT flag in pte_enter(), in order not to
sleep in pmap_enter(9) and pmap_kenter_pa(9), which can result in dead lock.

In most cases, pmap_enter(9) is used with PMAP_CANFAIL flag. In this case,
if pte_enter() fails due to uvm_km_alloc(9), we can safely return ENOMEM.
UVM layer will take care of it.

uvm_km_alloc(9) fails for pmap_enter(9) without PMAP_CANFAIL or for
pmap_kenter_pa(9), we have no choice but to panic there.

However, my testbeds for 403 and 405 survive more than a week at least
without hitting this panic.

(rin)

Style fixes:

- Add "static inline" to pte_enter(), to match with its declaration.
- Remove parentheses from return.
- Use NULL instead of 0 for pointer initialization.

No binary changes.

(rin)

lint: remove stray semicolon from grammar, make empty rule visible

No functional change.

(rillig)

tests/lint: adjust test for C11 to the change in grammar and options

(rillig)

use an enum instead of constant variables so that they work in CTASSERT.

(christos)

there is also an optional type qualifier list in the array size

(christos)

gnu enables c11 extensions

(christos)

lint: un-abbreviate code for passing options to cpp, lint1 and lint2

No functional change.

(rillig)

lint: add option to accept C11 features

The list of available letters for the command line options gets shorter
and shorter.  Most of the interesting letters are already used for some
warning categories.  Curiously, -A, -W and -E were all still available.

The option -A nicely matches the intention of the option, which is to
allow a certain set of language features.  To keep the option available
for further extensions, define -Ac11 as the currently only valid option
of that kind.  This allows straight-forward extension for C17 and future
language standards, as well as independent feature-sets.  The options -W
and -E may someday complement the -A option, using the allow/warn/error
categories.

(rillig)

lint: arrange options in the usage messages

* lowercase the word 'usage'
* group the preprocessor flags on a line of their own
* add missing closing ']' for '-X'
* properly indent the lines

(rillig)

lint: remove redundant condition in update_location

(rillig)

lint: add support for C11-isms such as int[static 3]

(rillig)

lint: add test for newly added message about static array size

(rillig)

make: let the compiler decide whether to inline string functions

On x86_64, this reduces the binary size by 2 kB.

(rillig)

make: turn run-time string concatenation into compile-time

No functional change.

(rillig)

make: remove unnecessary modifier ':U' for certain fixed expressions

No functional change, since the expression is evaluated using
VARE_WANTRES, not using VARE_UNDEFERR.

(rillig)

make: reduce memory allocations in the modifiers ':D' and ':U'

(rillig)

Add DIOCGDISKINFO compat.

(mlelstv)

make: rename members of ModifyWord_LoopArgs

No functional change.

(rillig)

make: clean up pattern flags for the modifiers ':S' and ':C'

No special handling is necessary for C90 since none of the struct
members is a const_member.

The prefix 'Var' is not necessary since this type does not apply to a
variable but only to a modifier.

No functional change.

(rillig)

support parsing c99's static class for array size expressions.

(christos)

lto-common.c needs the stack protector treatment

(christos)

more -fcommon.

(mrg)

build lfs_node.c with -O3 on ia64 to avoid assembler overflow issues.

(mrg)

for GCC, built memset.c with -fno-builtin.  this avoids GCC 10 compiling
memset() and emitting calls to memset() where it seems code that looks
like what memset() does, which ends up recursing and blowing the stack.

this makes mips userland with GCC 10 work.

(mrg)

OpenSSL man pages have moved

(nia)

intro.3: add a few

(nia)

Add a ddb disassembler for riscv.

builds, but not really tested yet.

(dholland)

Fix the problem "pcictl pci0 list" causes "panic: trap_el1h_error" on rockpro64.

The panic occures in bus_space_barrier() in rk3399_pcie.c:rkpcie_conf_read().
We expected bus_space_peek_4() to trap and recover in the path
trap_el1h_sync() -> data_abort_handler(), but In fact, the read is delayed
until bus_space_barrier(), and we get an SError interrupt (trap_el1h_error)
instead of a Synchronous Exception (trap_el1h_sync).

To catch this correctly, An implicit barrier in bus_space_peek have been added,
and trap the SError interrupt to recover from.

(ryo)

regen

(christos)

regen

(christos)

use the first element of the array instead

(christos)

dtrace systrace.c issue fixed by christos (thanks!)

mips issue is related to memset.c miscompiling and does not
appear with -ffreestanding (thanks joerg & simonb.)

(mrg)

regen

(christos)

regen

(christos)

Elide [NUM] in arguments for systrace, use the first address of the array
and cast to intptr_t instead.

(christos)

note amd64 dtrace issue is only warning for now.
note that i386 has problems in -current.

(mrg)

Pass -isystem and -imacro ccp flags to lint

(christos)

Allow type attributes after function pointer parameters

(christos)

add a test for a type attribute after a param function pointer

(christos)

apply -Wno-error=array-bounds for GCC 10.

these trigger "pointer to end of array" issues, and i'm not
entirely sure what is happening in the code generated for the
two "INDIR" system calls from syscalls.master.

(mrg)

Make LCD BrightnessUp/BrightnessDown work on C7x0/860.

Patch from steleto:
https://gist.github.com/steleto/10f62a074bff0c188fcc10c14ef40b5a
and also confirmed by me on SL-C700.

Worth to pullup to netbsd-9.

(tsutsui)

- fix memory leak
- xxx questionable allocation
- remove casts
- use sizeof(*var)
- bcopy -> memcpy/memmove

(christos)

with dtrace enabled, there is at least one remaining issue for amd64.

(mrg)

ignore a "should be impossible" uninitialised variable error.

(mrg)

XXX: ignore warning about overlapping buffers for sprintf().

(mrg)

sparc and amd64 successfully finished their atf runs.

(mrg)

more -fcommon for sources not ready yet

(mrg)

update the status of GCC.  summary:  mostly looking good.

- all targets build or can be attempted
- arm64 needs __aarch64_swp1_acq in both libgcc and libkern
- earmv4 works, earmv7hf works, all arm32 builds, rest untested
- hppa, most/all m68k, alpha, and most ppc builds, none tested
- ia64 kernel runs as well as before, userland fails to build
- i386 runs, seems mostly fine except lib/libc/sys/t_ptrace_wait
  resume test is hung, and cpu-spinning in a zombie
- sun2 grew again, too large but builds besides RAMDISK not fitting
- mips seems problematic.  some kernels fail some work, userland
  has problems with both static and dynamic n32 and n64 programs
- sh3el seems ok in gxemul/landisk, sh3eb not tested
- vax seems OK in simh (same c++ exception issues)

all ports kernels have been tested to build (though perhaps with
some uncommited changes only needed for some, while most changes
are commited, about 15 are not but many ports build without them.)

(mrg)

avoid duplicating symbols in libterminfo.

(mrg)

make this match the builtin prototypes.

(mrg)

apply -fcommon here as well.

(mrg)

some platforms trip -fcommon here too.  we should fix this.

(mrg)

more -O0 for vax

(mrg)

more GCC 10 fixes.

mDNSResponder: another wrong return local address

dhcp: ignore a seemingly impossible stringop overflow

hpacel: avoid maybe uninitialised error that is wrong.

rsh: avoid impossible malloc(0)

udf: cast pointers through (uintptr_t) to fool invalid boundary checks

(mrg)

avoid potential uninitialised variable use.

(mrg)

GCC 10 does not like casting from one identical anonymous enum
to another enum.  supply some quietening cast.

(mrg)

Reschedule softint to process packets enqueued to ppoediscinq
while doing pppoe_data_input

And added a empty check for ppoeinq, for safety

(yamaguchi)

Added missing counter clear when a pppoe state changes to PADI_SENT

(yamaguchi)

apply some -Wno- to ignore several new warnings with GCC 10.

ntp: ignore truncation beyond api sizes, and ignore wrongly
guessed underflow

tmux: ignore maybe uninitialised warning for impossible case

libbfd: signed/unsigned variables assigned in the same statement

dri/gallium: -Wno-builtin-declaration-mismatch for u_atomic.c as
it implements backend functions with different in-C-machine but
same-in-real-machine types

libXfont/libXfont2: signed/unsigned variables assigned in the same
statement

i915drm: ignore impossible maybe uninitialised warnings

sysinst: ignore an invalid string truncation issue

(mrg)

Added a NULL check for parent interface of pppoe

(yamaguchi)

Hold the lock for pppoe while referencing sc_id
that is an item of struct pppoe_softc

(yamaguchi)

use -fcommon here.

(mrg)

use -fcommon here too.

(mrg)

remove extra line that clearly isn't meant to be there and may trigger
uninitialised reads of variables.

(mrg)

use correct enum types for various functions.

(mrg)

ensure that pointer is filled in when used upon return.

(mrg)

no longer need -Wno-error=implicit-fallthrough.
don't provide common symbols that are in libterminfo now days.

(mrg)

increase the size of a buffer by 1 byte.

i think GCC 10 is right here.

(mrg)

properly terminate after using strncpy().

(mrg)

properly terminate a string after strncpy().

(mrg)

use -fcommon.

(mrg)

apply -fcommon to sources that still rely upon it.

(mrg)

in struct evcnt_sysctl, change ev_strings[0] to ev_strings[].

GCC 10 does not like accesses to zero length arrays.

(mrg)

fake-use alloca()'s return value to quieten -Werror=unused-result

(mrg)

allocate the potential trailing nul as well.

(mrg)

pass dest buffer size to strncpy() and ensure nul termination.

(mrg)

make: spell SysV in mixed case

No functional change.

(rillig)

make: reduce memory allocation and strlen calls in modifier ':from=to'

Previously, SysVMatch was quite verbose and felt like hand-optimized
assembler code, which made it difficult to discover the underlying idea
of the code.

All this code was replaced with two simple calls to Substring_HasPrefix
and Substring_HasSuffix.  Now that the operands of that modifier are no
longer passed as C strings, there is no need to collect all information
in a single scan through the word and the pattern.

It was not necessary to call Var_Subst unconditionally.  Calling it only
when the string contains a '$' saves another memory allocation and two
string copies (because of the Buf_DoneDataCompact).

No functional change.

(rillig)

tests/make: fix test for modifier ':from=to'

I had forgotten to remove the single quotes, to produce a really empty
word.  Plus I had forgotten the modifier ':Q' for the result, which
meant that the single quotes did not show up in the output.

(rillig)

tests/make: add table test for the modifier ':from=to'

Just to prevent any accidental change in the upcoming refactoring to
reduce memory allocation in ApplyModifier_SysV.

(rillig)

Add attribute fallthrough

(christos)

Add an example of attribute fallthrough

(christos)

make: in the ':Q' modifier, only allocate memory if necessary

(rillig)

hdaudioctl.8: add missing arguments

pointed out by RVP

(nia)

Print target addresses similar to aarch64 as "address <sym+off>"
instead of "<sym+off> [addr:address]".  Uses less columns, a bit
easier on the eyes.

(simonb)

mknative-gcc for GCC 10 and earm ports (now including earmv7hf) after
the arm/netbsd-elf.h fix.  shark appears to work at least now.

(mrg)

avoid attempting to print NULL.

(mrg)

use strlcpy() to ensure desired nul termination.
trailing nul bytes are not needed here.

(mrg)

avoid uninitialised variable access.  make this case like the other 2 around it.

(mrg)

handle cd1400_compute_baud() failure and avoid reading uninitialised
variables.

(mrg)

be sure to only access vcpu if it was initialised.

(mrg)

convert union member from [0] array to [3] array, so that attempts
to access these members later will  work.

does not affect size of the structure or layout in any other way.

(mrg)

use libsa/saerrno.h instead of our own errno definitions.

(mrg)

avoid common symbols.

(mrg)

properly terminate a buffer after strncpy().

(mrg)

pass the right enum to krb5_auth_con_setkeytype().

(mrg)

avoid an unlikely buffer truncation.

(mrg)

add a default: case to __kernel_standard(), failing with EDOM

(mrg)

avoid leaving off the trailing nul in a literal string.

(mrg)

arm/netbsd-elf.h is necessary for eabi targets.

(mrg)

properly size a buffer to include the trailing nul.

(mrg)

avoid common symbols.

(mrg)

updates:
- arm64 doesn't build (missing __aarch64_swp1_acq in kernels)
- mips cpu targets all build (maybe not all ports)
- powerpc builds (maybe not all ports)

(mrg)

mknative-gcc for GCC 10.3 and powerpc ports.

(mrg)

merge our changes into GCC 10.  parts of the rs6000.c file were
split into separate files, and an old change to rs6000.c now
belongs in rs6000-logue.c.

(mrg)

Reapply font calculation fix.

(mlelstv)

new GCC_NO_* uses for warning issues.  most of the users of the new
GCC_NO_RETURN_LOCAL_ADDR are bugs in GCC itself, not the code.

(mrg)

fix a buffer size (off by one), and fix all the comments for the
3 other correctly sized buffers with the same comment/context.

(mrg)

include bsd.own.mk to make sure $S is set.

(mrg)

include bsd.own.mk to be sure NETBSDSRCDIR is set.

(mrg)

add GCC_NO_IMPLICIT_FALLTHRU, GCC_NO_MAYBE_UNINITIALIZED, and
GCC_NO_RETURN_LOCAL_ADDR.

(mrg)

avoid common symbols.

(mrg)

use -fcommon where necessary.

(mrg)

remove shadowed (common) declarations of various libsa variables
from various boot programs.  for macppc and zaurus, avoid building
with -fcommon any more.

(mrg)

add some new uses of existing GCC_NO_* variables for warning issues.
remove an no longer relevant for gcc7 workaround (works fine in both
gcc9 and gcc 10.)

(mrg)

avoid assigning both signed and unsigned variables in the same statement.

(mrg)

avoid duplicate "ddb_regs" in crash with GCC 10 and -fcommon default.

(mrg)

make a numeric literal unsigned as it is bit-negated.

(mrg)

create the gcc-10 subdirs, and don't create the gcc-8 ones.

(mrg)