remove the recently added -O0 for vax & gcc 14, the compiler bug is fixed.

thanks to Kalvis.

vax-gcc14 needs -O0 for vsbus.c and ncr53c9x.c.

Switch from -std=gnu99 to -std=gnu11 by default.

This is needed for newer language syntax like u"foo" for UTF-16
string literals in UEFI.

As proposed on tech-kern:

https://mail-index.netbsd.org/tech-kern/2025/01/05/msg029919.html

Some commentators requested -std=c11 instead of -std=gnu11, but that
requires additional work -- some of it is a matter of providing asm
and typeof as aliases for __asm__ and __typeof__, but there's other
issues that need to be resolved too like PR toolchain/58969: use of
alloca is warning, not error, with -std=c11. This is a smaller
incremental change.

PR toolchain/58962: clang build broken by C11 features with -std=gnu99

branches: 1.301.2;
use objcopy's --compress-debug-sections when creating debug files.

this reduces the size of the installed files by over half in most cases,
though the debug set size doesn't really change much (which looks like
close to 1GB of space on amd64 with xdebug installed, similar on arm64,
and about 600MB without xdebug.)

tested by running GDB on a few things, seems just as functional, on amd64,
arm64, and slightly on riscv64.

(first attempt for this feature used "gcc -gz=zlib", but that ends up
making CTF unhappy, but fortunately this works in binutils to create
the .debug files separate to any ctf usage of the main file.)

Makefile.kern.inc: fix grammar in a comment

Simplify fix for PR toolchain/57146

Introduce ARCH_STRIP_SYMBOLS variable to centralize logic for debug
symbols from MD Makefile's to Makefile.kern.inc.

Fix kernel size inflation for arm and aarch64 (PR toolchain/57146)

For some conditions, SYSTEM_LD_TAIL is set for arm and aarch64.
Then, ctfmerge(1) in default SYSTEM_LD_TAIL is unintentionally
skipped, which results in the catastrophic kernel size inflation,
as reported in the PR.

Also, introduce and use OBJCOPY_STRIPFLAGS variable instead of
STRIPFLAGS, as strip(1) is replaced by objcopy(1) during MI
kernel build procedure.

XXX
For Makefile.{arm,aarch64}, weird logic is used to determine how
to handle debug symbols; MKDEBUG{,KERNEL} are taken into account
later in sys/conf/Makefile.kern.inc.

adapt to ${CC_WNO_ADDRESS_OF_PACKED_MEMBER}

Simplify CWARNFLAGS to use ${CC_WNO_ADDRESS_OF_PACKED_MEMBER}
which works for both clang and gcc, and remove compiler-specific
equivalents.

bsd.own.mk: rename to CC_WNO_ADDRESS_OF_PACKED_MEMBER

Provide a single variable
CC_WNO_ADDRESS_OF_PACKED_MEMBER
with options for both clang and gcc, to replace
CLANG_NO_ADDR_OF_PACKED_MEMBER
CC_NO_ADDR_OF_PACKED_MEMBER
GCC_NO_ADDR_OF_PACKED_MEMBER

Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.

branches: 1.295.4;
sys: document how to run lint on the kernel code (experimental)

sys: Build kernel with -Wno-type-limits.

The type-limits warning is actively harmful because it discourages
writing safe portable overflow detection logic which happens, on some
architectures, to be dead code.

Fix for systems that override SYSTEM_LD_TAIL

use a function "runit" to echo and execute avoiding set -x.

- With DEBUG=-g (which is also set when we build CTF) we build netbsd.gdb *and*
netbsd-${KERNEL_CONFIG}.debug because it is not worth having two sets of
rules (one that builds just netbsd.gdb and one that builds both netbsd.gdb
and netbsd-${KERNEL_CONFIG}.debug. This maintains compatibility building
netbsd.gdb when DEBUG=-g
- When either MKDEBUGKERNEL=yes or MKDEBUG=tes we set DEBUG=-g if DEBUG was
not set.
- We only make a debuginstall rule to install netbsd-${KERNEL_CONFIG}.debug
if MKDEBUG=yes

forgot one KERNEL_BUILD (Tobias Nygren)

Use the tail of KERNEL_BUILD since it can be a full path.

Add ${_MKTARGET_INSTALL}

use ${INSTALL_FILE} so that unpriv builds work.

Enable split-debug files for kernels. Enabled by default with MKDEBUG=yes.

Rename:
MKKDEBUG -> MKDEBUGKERNEL
MKTOOLSDEBUG -> MKDEBUGTOOLS
while keeping compatibility with the old names. Add missing documentation.
Now all debugging tunables are prefixed with MKDEBUG.

revert previous: http://mail-index.netbsd.org/source-changes/2021/10/25/msg133295.html

going to add __always_inline to the functions called from _mcount()
discussed on http://mail-index.netbsd.org/source-changes-d/2021/10/25/msg013480.html

In some arch, _mcount() would be called recursively when built with COPTS=-O0.

Normally, functions called from mcount.c are expected to be expanded inline,
so _mcount() will never be called recursively. But when build with COPTS=-O0,
`static inline' functions aren't inlined, and _mcount() will be called
recursively.

Even if _mcount() has `__attribute__((__no_ instrument_function__))',
it has no effect on the calling external (no-inlined) function.

To avoid this, PROF.<fn> is added can be set the profiling flag of any file.
"PROF.mcount.c" is set to blank by default, mcount.c itself is compiled
without -pg.

Add missing double p and d for stopped and overriden accordingly.
Fix few more typos along the way, mainly in copy-pasted comments.

drm2: build with -fwrapv, since the Linux kernel does and we don't
want to spend too much time fixing their code. silences a lot of
KUBSAN warnings.

we need to use COPTS to have per-option compiler flags. help from
Riastradh.

branches: 1.280.2; 1.280.6;
Revert rev 1.278 which tried to call ctfmerge only once - this has
caused kernels with embedded symbol tables to explode in size.

Don't use /usr/bin/time (it is not portable)

Move the ctfmerge call from SYSTEM_LD_TAIL (used by both STAGE1 and STAGE2)
to SYSTEM_LD_TAIL_STAGE2 only. ctfmerge is slow enough that just calling
it once is enough.

branches: 1.277.2;
sys/conf: remove useless dollars from Makefile.kern.inc

When parsing a variable modifier such as :C or :S, all '$' but the very
last are silently swallowed.

See usr.bin/make/unit-tests/varmod-subst.mk for details.

Since August 2020, make has a strict lint mode that is activated by
passing it the option -dL. In that mode, these redundant '$' are not
accepted anymore.

branches: 1.276.2;
add support for new GCC 9 warnings that may be too much to fix
right now. new address-of-packed-member and format-overflow
warnings have new GCC_NO_ADDR_OF_PACKED_MEMBER amd
GCC_NO_FORMAT_OVERFLOW variables to remove these warnings.

apply to a bunch of the tree. mostly, these are real bugs that
should be fixed, but in many cases, only by removing the 'packed'
attribute from some structure that doesn't really need it. (i
looked at many different ones, and while perhaps 60-80% were
already properly aligned, it wasn't clear to me that the uses
were always coming from sane data vs network alignment, so it
doesn't seem safe to remove packed without careful research for
each affect struct.) clang already warned (and was not erroring)
for many of these cases, but gcc picked up dozens more.

enable -Walloca, warn if something uses alloca()

enable gcc stack usage limit for kernel functions, set to 3.5 KiB for now
as that seems to be enough to accomodate the current biggest stack usages

there are about six functions which use over 3KiB local stack, and
about a dozen between 2-3 KiB, so pushing this further needs more work
if desired

compile tested on amd64, i386, sparc64, sparc, powerpc (evbppc - BookE),
m68k (mac68k)

Use -fstack-usage only if we are using gcc.

Pointed out by maxv@

Discard -fstack-usage for vax, for which GCC does not support that option
at the moment.

Add -fstack-usage to kernel builds. Produces .su files that can be
analysed to find large stack users.

Do the chmod last so that the kernels get consistent permissions for MKREPRO
Reported by wiz@.

Infrastructure for putting kernel+modules in /netbsd/kernel and
/netbsd/modules respectively instead of /netbsd and
/stand/<arch>/<version>/modules. This is only supported for x86,
and is turned off by default. To try it, add KERNEL_DIR=yes in your
/mk.conf and install a system from that build.

pass -Wno-error=implicit-fallthrough for zlib.c.

XXX: consider using copts.mk for various warning/copt flags passed
in kernel builds currently set via 'makeoptions' in files.* files.
this is suboptimal, as those all get embedded into the kernel with
config_file.h.

Merge the [pgoyette-compat] branch

- Determine KERN_AS automatically depending on whether OPT_MODULAR is
set or not, in the same way as libcompat.

- Specify OPT_MODULAR in the port Makefile instead of KERN_AS.

Now, KERN_AS=library is used for kernels without module(7) for all ports.

OK christos

Add CFLAGS, CPPFLAGS, and CWARNFLAGS from options to .c files.

branches: 1.264.2; 1.264.4;
Restrict -fno-delete-null-pointer-checks to GCC.

use -fno-delete-null-pointer-checks for kernels.

- Coverity does not like -std=gnu99
- Another lose is that you need to manually edit the compiler XML to add
all the kernel options because cov-configure barfs because it tries to
test the compilation environment thinking that we are userland even
when options like -ffreestanding are present.

We use .BEGIN for our own cleanup and won't ever need DTS support - so
avoid inclusion of dts.mk by defining _SKIP_DTS

branches: 1.260.2;
Move the dts rule to a separate file and add dependencies handling.

Add support for dts files outside of external/gpl2.

Assign DTB files to a variable so we can make -V DTB

Add support for building DTB files during kernel build, from christos.

branches: 1.256.8;
Disable a couple of warnings until further investigation.

Move to ssp.c suggested by uebayasi@

Add COPTS for xhci.c, since the do_command_locked causes a spurious stack
proector warning.

Collect a list of kernel swap files to be removed (Yorick Hardy)

match the M* and N* targets.

PR/51772: Yorick Hardy: Don't hard-code "netbsd", use ${.TARGET} consistently.

branches: 1.250.2;
Use && to fail early if subcommands of kernel link fail.

(Remind me why we string together shell commands instead of just
using make targets?)

use commands() instead of target() because assym.mk creates all the targets
unconditionally.

PR/50483: David A. Holland: fix building of assym.d:

1. use += to add to DEPS
2. move inclusion of rules mk. files before the rules are defined
so that variables are set.

Now we get on amd64:

mkdep: ignoring acpi_wakeup_low.d: No such file or directory
mkdep: ignoring amd64func.d: No such file or directory
mkdep: ignoring busfunc.d: No such file or directory
mkdep: ignoring copy.d: No such file or directory
mkdep: ignoring cpu_in_cksum.d: No such file or directory
mkdep: ignoring cpufunc.d: No such file or directory

because there are no rules to generate those, but it is harmless.

Clean up splash image build. Print messages when creating files.

Indent.

Fix dependency of *.d -> assym.h in `-S' mode.

Preserve the order of files in ${ALLFILES} -> ${OBJS} conversion.

More build ordering.

Don't pass ``-DCOPY_SYMTAB'' to ${CC}; it's replaced with opt_copy_symtab.h.

Move ldscript related code out of Makefile.kern.inc.

Clean up. Move assym.h/vers.c related code out of Makefile.kern.inc.

Clean up. Move SSP-specific adjustment out of Makefile.kern.inc.

Allow MD to link libraries. Build rules are written in MD makefiles until
config(1) will support library properly.

Make config(1) output only ${ALLFILES}, selected input files, and pass it to
Makefile.kern.inc, where files are filtered by suffix.

Add generated *.c files (devsw.c and ioconf.c) to ${CFILES} by config(1)
internally. ${MI_CFILES} in Makefile.kern.inc is no longer needed.

Prepare for relative ${MD_OBJS}.

Move `all' target definition to Makefile.kern.inc.

Kernel depends on swap<kernel>.o too.

Sort dependendy to keep alphabetical build order.

Define kernel dependency in Makefile.kern.inc.

Rename ${COMPILE.c} to ${KCOMPILE.c} to avoid potential conflicts with those
in sys.mk. Define .go/.po suffix rules.

Refactor.

Simplify. Wrap long lines.

Stop ordering objects alphabetically now that I am sure I can fix fallouts.

Build conf/param.c normally.

Fix link of kernels whose name don't start with "netbsd*".

Build param.c under conf/ using suffix rules in `-S' mode.

Clean up `swap<kernel>.o' handling.

Define ${OBJS} in `Makefile.kern.inc'. Relative paths are not truncated
(a/b/c/x.c -> x.c) in `-S' mode. Don't provide some explicit rules for `-S'
mode for safety.

Output relative path for ${CFILES} and ${SFILES} in generated `Makefile'.
`Makefile.kern.inc' knows how to deal with $S prefix, so don't bother it in
config(1).

Add swap<kernel>.o to common ${SYSTEM_LD} as it is always linked.

Clean up some rules and make them less error prone.

Define suffix rules at the end to override system definitions. Provide them
only if ___USE_SUFFIX_RULES___ is defined (config -S).

Experimental suffix-rules build (disabled by default).

Fix "Define compile rules in sys/conf/Makefile.kern.inc ...".

Use `.if !commands(xxx)' to check if `xxx' has a defined, overriden rule,
instead of `.if !targets(xxx)'. The latter evaluates as true even when `xxx'
has an empty rule to define a dependency.

A makeoptions `DEFWARNINGS', forcibly disabling all ${CC} warnings, is only
used by hpcmips's NULLCONF, which doesn't even build for other causes.

If some source doen't build and needs a workaround, that should be deal with
by modifying per-file ${CC} options (e.g. COPTS.foo.c+=-fno-xxx). Let's not
use DEFWARNINGS any longer.

Move splash logic, keep Makefile.kern.inc clean.

Put back MEMORY_DISK_IMAGE logic, but use generated opt_memory_root_image.h
and don't polute Makefile.kern.inc.

Define compile rules in sys/conf/Makefile.kern.inc except those overriden by
"file ... compile-with ...". As rules are still explicit, be careful to not
override by checking .if !target(xxx.o).

Use makeoptions_COPY_SYMTAB via opt_copy_symtab.h in sources. Remove a hack
to set -DCOPY_SYMTAB from sys/conf/Makefile.kern.inc. Remove unnecessary
dependencies too.

Move less important part out of Makefile.kern.inc.

Define two suffix rules, .c -> .o and .S -> .o, using ${NORMAL_C} and
${NORMAL_S} respectively. Use the .c rule to build devsw.c, ioconf.c, and
param.c. Other .c/.S files have explicit rules in the generated `Makefile',
and unaffected.

Now that MAXUSERS is in opt_param.h, param.o doesn't depend on Makefile.

According to nxr.netbsd.org, nothing uses MEMORY_DISK_IMAGE. Retire it.
Premature design and its build rule bloats Makefile.kern.inc. There are
other ways like MEMORY_DISK_DYNAMIC.

(When kernel will be built as relocatable, embedding binary data will be much
easier, and md(4), splash(4), and ksyms(4) will all benefit.)

genassym.cf and machdep.c are _KERNEL_OPT'ed, don't depend on Makefile.

conf.c doesn't exist.

autoconf.c exists, but it's just a C file, no need to depend on Makefile.

I don't see how uipc_proto.c can change its content by Makefile, don't depend
on Makefile. vfs_conf.c doesn't even exist, remove it too.

Convert MAXUSERS, take 2.

Define MAXUSERS in opt_param.h and include it from sys/param.h, for more
accurate dependency. Don't define -DMAXUSERS and add it to ${CPPFLAGS}.
config(5)'s "maxusers" syntax doesn't change.

(The version just before revert was OK, but I had something bad in my obj
dirs.)

Revert MAXUSERS changes. Need fresh air.

Define MAXUXERS in opt_param.h. Bump config(1) version.

Comment.

Define some suffix rules. Not used and harmless, because all rules are
explicit now.

Specify explicit output (-o) to ${LINT} too.

Specifiy explicit output (-o) to ${CC}.

netbsd depends on $S/kern/kern_ksyms_buf.c; rarely updated but better safe.

Tweak previous.

Echo exact objects (swap*.o etc.) when linking kernel.

Have MI genassym.cf.

Don't allow MD to override the assym.h generation rule.

(Nothing defines it except i386/bioscall/Makefile, but it is not part of
kernel build.)

Make <kernel>.ldscript ready to include assym.h.

Redo Rev. 1.188:
o Explicitly add dependency of `<kernel>' on `<kernel>.ldscript', instead of
relying on ${SYSTEM_DEP}, which is evaluated much earlier than
`sys/conf/Makefile.kern.inc' is included, to avoid evaluation ordering
problems.
o Print a message when creating a `<kernel>.ldscript'.
o Trim empty lines in `<kernel>.ldscript'.

Fix the build: provide compatibility with ports that define KERNLDSCRIPT to
a file?

If a kernel linker script is provided, preprocess it by cpp(1).

Thus some constants (PAGE_SIZE, COHERENCY_UNIT, ...) can be replaced in linker
scripts without setting unnecessary temporary symbols, or doing ugly sed(1)
hacks. No headers are included yet.

Fix vers.o dependency:
o Depend on ${SYSTEM_OBJ} instead of ${SYSTEM_DEP}. The latter is obviously
meant for kernels.
o Depend on Makefile instead of ${SYSTEM_SWAP_DEP}, which has existed since
Rev. 1.1, but seems not used now. Makefile is a file that must be updated
after every config(1) run. This is simpler than dependency on swap<kernel>.c.

Define ${LINKSCRIPT} in one place.

If MD ldscript is generated, give it dependency on ${SYSTEM_OBJ}.

Include splash with genfb.
splash.o is only included with options SPLASHSCREEN.
splash_image.o target moved to Makefile.kern.inc.

This commit was approved by matt@

branches: 1.183.2;
Stop ordering dependencies alphabetically.

Don't sort objects for modular build.

Separate linker script argument (${LINKSCRIPT}) from ${LINKFORMAT}

${LINKFORMAT} specifies ABI options, and used for ld -r. ${LINKSCRIPT} (-T
ldscript) is used only for final link.

`ld -r' needs ${LINKFORMAT}. This fixes the experimental config(1) modular
build for MIPS64 ports.

Trim " " at the beginning of version string to follow the change
in rev.1.60 of newvers.sh.

-const char sccs[] = "@(#)${fullversion}";
+const char sccs[] = "@(#)" ${fullversion_source};

Leave .map when linking .ko in modular build.

More clean files.

Fix link echo indent.

Define a command to generate relocatable objects using ld -r.

Also copy the symtab to the .gdb image. If we are producing a .gdb, grab
the symtab from that since it will be a bit larger than the one from the
stripped kernel and thus can't be inserted into the .gdb kernel.

Add a commented out CWARNFLAGS to enable -Wc++-compat but don't treat them
as errors.

More tidyup after joerg's copy&paste disaster. I have ddb symbols again.

Drop stray }

Reorganize symbol table embedding. The existing option SYMTAB_SPACE is
replaced by the make option COPY_SYMTAB set to any value. The copy of
the symbol table is no longer put into a buffer in kern_ksyms.o, but a
small helper object. This object is build first with a dummy size, then
the kernel is linked to compute the real dimension of the symbol table
buffer. After that, the helper object is rebuild and the kernel linked
again.

branches: 1.169.2;
The bpendtsleep label has been gone for a long time, so cleanup the
associated hack here.

As discussed on tech-toolchain, offer a new make variable MKKDEBUG - if
set to "yes", kernel builds will gain debug info and a netbsd.gdb will
be created (same as adding "makeoptions DEBUG=-g" to the config file).

branches: 1.167.2;
Do COMPILE_CTFCONVERT on vers.o as same as other targets

Otherwise, we will get the error when building a kernel with -V MKDTRACE:
ERROR: nbctfmerge: Input file vers.o was partially built from C sources,
but no CTF data was present

Use CTFMFLAGS, not CTFFLAGS, for CTFMERGE

Introduce GENASSYM_CPPFLAGS for options during genassym processing.
Consistently drop assembler flags.

revert previous

add a device-major for lua(4)

Update LLVM/Clang snapshot to r191105. This brings in a better detection
of unused static variables and functions. Disable this for some external
code and for ioconf.c in the kernel.

Fix debug kernel builds when the automatic symbol table resizing kicks in.
Previously, the TAIL/HEAD link steps were being executed twice and the
(relinked/stripped) kernel would overwrite the debug version.

XXX this is all ugly.

Use += for SYSTEM_DEP not ?=

branches: 1.159.2;
Add references to ${_NETBSD_VERSION_DEPENDS} for files that
need to be re-created when the NetBSD version changes. They
will also be re-created when any build settings are changed.

branches: 1.158.2;
Deal with optional HAVE_GCC.

Enable -fno-common for GCC compiled kernels

Redo rev 1.154 by using an explicit MKSHNOECHO that does not output anything.

Don't pass assembler flags to genassym, it doesn't compile.

Don't print blank lines when !CTFCONVERT.

adjust the size of db_symtab space automatically.
proposed on tech-toolchain@, and got no objections.

branches: 1.152.4;
Change old-style function defintions to C89 prototypes.

Approved by releng.

-Wold-style-definitions for i386 and amd64

Move Clang-warning flags for ah_regdomain.c into sys/conf, they are
platform independent. Use the selective -Wno-* flags instead of
-Wno-error.

branches: 1.149.2; 1.149.6;
Allow analyze to work by including bsd.clang-analyze.mk for kernels

for multi-cpu machines, cpu_hatch() straddles the init of
__stack_chk_guard, so ensure stack protection is disabled
on x86 architectures (I don't know about other archs)

let kernel makefile override *.d targets

Remove gcc-4.5 hacks

let md makefile add newvers.sh flags by setting NVFLAGS

Restrict -Wno-array-bounds to tulip.c with gcc.

Define SYMTAB_SPACE in opt_ksyms.h, not opt_ddbparam.h. Ensure
that ddb(4) code is independent of SYMTAB_SPACE value.

adb_kbd.c and adb_ms.c need -Wno-cast-qual due to sysctl const issue.

apply some -Wno-foo to 3 files that are non-trivial to solve, for GCC 4.5.

remove all the code that supported HAVE_GCC=3. mostly from chuq.

If dbsym fails, not fail the command, remove the target too so that the
next invokation will try to remake it.

branches: 1.138.2;
Don't duplicate -Werror here, bsd.sys.mk already provides it

Default to -Wno-sign-compare -Wno-pointer-sign for clang.
Push -Wno-array-bounds down to the cases that depend on it.
Selectively disable warnings for 3rd party software or non-trivial
issues to be reviewed later to get clang -Werror to build most of the
tree.

Reorder flag order to not depend on -Wextra being applied before
-Wno-sign-compare

Inherit SSP settings from bsd.sys.mk

Spell --fatal-warnings with two hyphens

Add support for source specific AFLAGS and COPTS.

Make ${DBSYM} failure fatal.

correct comment

If generating a gdb kernel and using dbsym, insert the symbols into it.
(This is useful when using the .gdb with an emulator so that ddb in the
emulator will have symbols).

branches: 1.129.2; 1.129.4;
If using a kernel ldscript, make sure it's added to the list of kernel
dependencies.

Kernels modules are c99 (gnu99) just like the kernel.
Add(move) -std=gnu99 to CPPFLAGS (from COPTS).

branches: 1.127.2;
Use 'a && b && ...' contructs for running multiple commands on a single line
in Makefiles, to properly detect/report failures.

DTrace: Add an SDT (Statically Defined Tracing) provider framework, and
implement most of the proc provider. Adds proc:::create, exec,
exec_success, exec_faillure, signal_send, signal_discard, signal_handle,
lwp_create, lwp_start, lwp_exit.

branches: 1.125.2;
"cgd.c" doesn't use variable size stack allocations anymore.

Nothing has used config_time.h since 1/8/2008.

Add a MKREPRO variable (defaults to "no") that invokes newvers.sh with
the -r flag if it is set to "yes", resulting in a reproducible kernel build.

fix cscope target

Philip Paeps, PR kern/41012

Unbreak ssp kernels. The issue here that when the ssp_init() call was deferred,
it caused the return from the enclosing function to break, as well as the
ssp return on i386. To fix both issues, split configure in two pieces
the one before calling ssp_init and the one after, and move the ssp_init()
call back in main. Put ssp_init() in its own file, and compile this new file
with -fno-stack-protector. Tested on amd64.
XXX: If we want to have ssp kernels working on 5.0, this change needs to
be pulled up.

branches: 1.120.2;
Not every grep knows -q. Ok by apb.

Sort the kernel objects.

branches: 1.118.2; 1.118.4;
Use ${TOOL_SED} instead if plain sed in Makefiles.

Use ${TOOL_AWK} instead of ${AWK} or plain "awk" in make commands.
Pass AWK=${TOOL_AWK:Q} to shell scripts that use awk.

Compile NetBSD/amd64 kernels with "-Wextra". Patches contributed by
Juan RP in PR port-amd64/39266.

Don't add "${CPUFLAGS}" and "${CWARNFLAGS}" to "AFLAGS" and "CFLAGS" twice.
Patch submitted by Juan Romero Pardines in PR kern/39265.

Wrap compiler-specific flags with HAVE_GCC and HAVE_PCC as necessary. Add a few flags for PCC.

branches: 1.113.2;
pcc needs -Wno-unreachable-code.

branches: 1.112.2;
Print kernel version string along with the size output, to avoid having
to dig it out manually if installing by version number... and also to
make it somewhat easier to notice up front if one accidentally boots
the wrong test kernel. not like I've ever done that. ;-)

PR kern/38563.

branches: 1.111.2; 1.111.4; 1.111.6;
Fix flags for kern_synch.c: it is COPTS.<cfile>, not CFLAGS.<cfile>.

Since we say the kernel is C99, add -std=gnu99 to COPTS

Add assym.d to DEPS so it gets properly cleaned.

branches: 1.108.2; 1.108.6;
Add a syntax-only target which attempts to compile the entire kernel as one
entity. Think of it as super-lint. Due to bugs in the GCC C frontend it
doesn't quite work right but it does provide useful feedback.

Add a stub file that includes a bunch of kernel headers and always get
compiled with -g. For the initial set, netbsd on amd64 grows by around
80KB. This allows much easier use of GDB for post-mortem debugging as
it can understand the layout of data structures. The additional data can
be strip(1)ped off normally for size constraint environments.

Remove unnecessary references to config_time.h.

Use BUILDSYMLINKS from <bsd.files.mk> instead of custom copy rules
for config_time.h and athhal_options.h.

Note: we still copy param.c because I'm told that we should still support
people editing that on a per-compile basis.

branches: 1.104.6;
add dev/ofw/ofw_subr.c to VARSTACK

branches: 1.103.4;
Fix accidental USE_SSP -> USE_FORT: now USE_SSP works as intended for
kernel builds. Thanks to xtraeme for pointing this out.

ufs is ssp clean

branches: 1.101.6; 1.101.8; 1.101.10;
Add new Makefile knob, USE_FORT, which extends USE_SSP by turning on the
FORTIFY_SOURCE feature of libssp, thus checking the size of arguments to
various string and memory copy and set functions (as well as a few system
calls and other miscellany) where known at function entry. RedHat has
evidently built all "core system packages" with this option for some time.

This option should be used at the top of Makefiles (or Makefile.inc where
this is used for subdirectories) but after any setting of LIB.

This is only useful for userland code, and cannot be used in libc or in
any code which includes the libc internals, because it overrides certain
libc functions with macros. Some effort has been made to make USE_FORT=yes
work correctly for a full-system build by having the bsd.sys.mk logic
disable the feature where it should not be used (libc, libssp iteself,
the kernel) but no attempt has been made to build the entire system with
USE_FORT and doing so will doubtless expose numerous bugs and misfeatures.

Adjust the system build so that all programs and libraries that are setuid,
directly handle network data (including serial comm data), perform
authentication, or appear likely to have (or have a history of having)
data-driven bugs (e.g. file(1)) are built with USE_FORT=yes by default,
with the exception of libc, which cannot use USE_FORT and thus uses
only USE_SSP by default. Tested on i386 with no ill results; USE_FORT=no
per-directory or in a system build will disable if desired.

branches: 1.100.2;
Add the ability for platform-specific Makefiles to indicate which kernel
images to install into / (which might be more than just "netbsd").

Addresses PR port-shark/17569.

branches: 1.99.2; 1.99.6;
We don't need to include kern_exit43.c to the list of files that we cannot
use SSP on now that zero sized arrays are treated properly in gcc.

Add --param ssp-buffer-size=1 if USE_SSP as for userspace.
Add kern_exit_43.c to VARSTACK -- it isn't, really, but it causes an error
because it has a 0-element array on the stack and SSP always emits the
error that it is not protecting such a small array (ssp-buffer-size=0 is
not supported, so, really, it should not emit this error!).

use -fstack-protector instead of -fstack-protector-all since this breaks
on amd64 (it works on sparc64).

there is no stack-protector for gcc3

Add arch/xen/i386/gdt.c to list of kernel files that do variable-size
allocations on the stack. This allocation could potentially be quite
large -- I am not sure how to best fix that.

Fixes USE_SSP i386 build.sh failure.

branches: 1.94.2;
add aic97xx.c to the varstack list. from shannonjr.

add another file to the VARSTACK list

Add SSP support.
XXX: This is broken for me right now, because my kernel resets after fxp0
is probed, but it could be some bug in the driver/compiler.

Add -Wextra -Wno-unused for prep.

Insert a .WAIT between "depend" and "all" in the "dependall" target, so
that "make -j4 dependall" has the intended effect.

Sparc64 is -Wextra ready - but we are not sure where to go with __unused,
so add -Wno-unused-parameter for now.
Ok'd by christos.

- sprinkle __unused on function decls.
- fix a couple of unused bugs
- no more -Wno-unused for i386

Add -Wextra -Wno-unused for i386. Portmasters, please check your ports
and make sure they compile with these flags.

branches: 1.86.2;
Add AFLAGS+= ${AOPTS.${.IMPSRC:T}}

make the bpendtsleep: label only active if KERN_SYNCH_BPENDTSLEEP_LABEL
is defined. if this option is present in the Makefile CFLAGS and we are
using GCC4, build kern_synch.c with -fno-reorder-blocks, so that this
actually works.

XXX be nice if KERN_SYNCH_BPENDTSLEEP_LABEL was a normal 'defflag' option
XXX but for now take the easy way out and make it checkable in CFLAGS.

branches: 1.84.2;
Add -Wa,-fatal-warnings to AFLAGS. We want assembler warnings to be
fatal as well as compiler warnings.
Vague nods of agreement from Christos and Gimpy.

Create a linker map file.
Very useful when trying to find out where recent 'bloat' has come from,
as well as giving exact details of what is actually in the kernel.

remove support for building (with) GCC 2.95. also:
- always install <stdbool.h>
- don't generate a fake one for vax / gettext.

it is attributes no attribute

Put back -Wno-attribute for now since not all archs have been fixed.

Don't use -Wno-attributes for gcc4. All the code has been fixed and it will
just hide bugs.

branches: 1.78.2;
if GCC>3:
- add -Wno-attributes -Wno-pointer-sign to CWARNFLAGS.
- add -fno-strict-aliasing to CFLAGS [*]

our kernel again needs a bunch of work for this to be enabled.

Also search for headers in common/include

Order & uniq DEPS, a la <bsd.dep.mk>.

Sort SYSTEM_OBJ before adding to SYSTEM_DEP.
Kernel builds are now more consistent with userland
(and IMHO building in order is more asthetically pleasing.)

Don't add CPUFLAGS to CFLAGS twice.

branches: 1.73.4; 1.73.6; 1.73.8; 1.73.10; 1.73.12;
If we are going to run mkdep separately for each file, we might as well
get this makefile to execute the mkdep commands - no need for a submake
and xargs at that point.
However we do need to do something to stop the 'mkdep -d' and CLEANDEPEND
command lines being to long.
Note that 'echo ${xxx} | cmd' is ok because echo is a shell builtin, and
isn't (usually) subject to the kernel's command line limits.

Remove the .d files during __CLEANDEPEND

use .d mechanism for kernel as well.

. Have dbsym explicitly suggest increasing SYMTAB_SPACE when
an overflow occurs.
. Make this error a fatal build time error
. Move the support for dbsym into the MI Makefile.kern.inc,
conditional upon the SYMTAB_SPACE option being defined in
the kernel config file.

branches: 1.69.2;
don't pass -t option to xargs unless MAKEVERBOSE is 2.

branches: 1.68.2;
Do not use -Wreturn-type with gcc 2.95.x

Sometimes .BEGIN target is not wanted - eg in sub-make's.

Don't assume infinite command line length.

Add CWARNFLAGS.<fn>

- move COPTS.${.IMPSRC:T} and friends after other options so that
they can override global options.
- put CPPFLAGS.${.IMPSRC:T} into CPPFLAGS, rather than CFLAGS.

Add overrides for tools we use (genassym and hexdump). Now kernel compiles
work again (hi jason!)

Most of the kernel source now works with "-Wcast-qual -Wshadow
-Wreturn-type -Wwrite-strings -Wswitch" so I am turning them on. There
is going to be a little lossage on some platforms but you can always
compile with "make DEFWARNINGS=no". Tested on sparc64 and i386.

Use TOOL_GENASSYM. Part of PR toolchain/30350

tweak reference of DEBUG variable to ignore options that don't start with -g

nuke trailing whitespace

branches: 1.58.2; 1.58.4;
Adjust the cscope target so that each source file on a separate line to avoid
overrunning cscope's read buffer and prefix files within lib/{kern,compat}
with the appropriate path.

work around a binutils bug where converting ELF kernels to a.out with objcopy
produces corrupted binaries when the link_set_* sections extend into another
page after the end of the .text section by using a generated an ldscript that
puts all the link_set_* data into the .text section in the first place.

Rename LIBCOMPAT* -> SYSLIBCOMPAT*, now that <bsd.own.mk> provides LIBCOMPAT.
Fixes kernel link problem noted by Sebastien Erard.

Make uudecode into a tool that can be used to build a kernel and redo
the athhal-elf.o handling to take advantage of it. This means that
bsd.files.mk is no longer included by Makefile.i386 so CFLAGS won't be
duplicated.

Narrow a sed(1) RE handling a file/directory name so that build under a
directory named like 'xxx.cxxx/' succeeds.

-fno-zero-initialized-in-bss only to gcc3, not gcc2.95

Force zero-initialised variables into the .data segment.

branches: 1.51.2;
fix the MAKEVERBOSE messages to be consistent with the standard ones

prefix ${.TARGET} with "${.CURDIR:T}/" in various _MKSHMSG macros;
per similar change to <bsd.own.mk>

Change default verbosity level to `traditional'.

NORMAL_C,NOPROF_C,NORMAL_S: put `@' in front of these to get saner output.
Also we can't use _MKTARGET_COMPILE here. For now, just hard-code the message.

SYSTEM_LD: don't display every object file on the planet thrice.

Provide appropriate no-op defaults for _MK* macros, in case <bsd.own.mk>
is too old.

Add support for MAKEVERBOSE (a first pass, at least).
Use COPTS.<sourcefile> instead of <sourcefile>_G.
Use HOST_SH (which defaults to `sh' if not set).

Remove the .if !target(dependall) so that 'make dependall' works (again)
for the kernel.

Prevent mkdep from parsing compiler flags like -pipe.

force -Wno-uninitialized if the compiler is gcc 2.95

Back out -Wno-uninitialized for m68k -- the compiler bug is fixed.

Add COPTS.<fn>, CPUFLAGS.<fn>, CPPFLAGS.<fn> to CFLAGS to mirror the
file specific flags support of bsd.{lib,prog}.mk.

From Christos.

sys/dev/ata/wd.c causes spurious warning about blkno not
being initialized before used. This appears to be a bug
specific to (at least) the m68k target, and will be reported
to the GCC maintainers.

Delete -Wno-uninitialized. Tested on sparc64 and i386. Rationale:
Gcc-3.3 produces valid warnings in most cases and uncovers bugs. In the
cases where a variable is known to be initialized, we initialize it and
add: /* XXX: gcc */. If a $MACHINE_ARCH's compiler is busted, it can add
-Wno-uninitialized in the MD Makefile.

Make the kernel makefiles pay attention to NOGCCERROR

allow DEBUG to be set to "-gstabs" or "-gstabs+", to enable building
netbsd.gdb. (you need "-gstabs" for gcc3.3 and old gdb.)

Rename a large chunk of the make(1) variables which refer to a
program/tool from "FOO" to "TOOL_FOO". The new variables are:
TOOL_ASN1_COMPILE TOOL_CAP_MKDB TOOL_CAT TOOL_CKSUM TOOL_COMPILE_ET
TOOL_CONFIG TOOL_CRUNCHGEN TOOL_CTAGS TOOL_DB TOOL_EQN TOOL_FGEN
TOOL_GENCAT TOOL_GROFF TOOL_HEXDUMP TOOL_INDXBIB TOOL_INSTALLBOOT
TOOL_INSTALL_INFO TOOL_M4 TOOL_MAKEFS TOOL_MAKEINFO TOOL_MAKEWHATIS
TOOL_MDSETIMAGE TOOL_MENUC TOOL_MKCSMAPPER TOOL_MKESDB
TOOL_MKLOCALE TOOL_MKMAGIC TOOL_MKTEMP TOOL_MSGC TOOL_MTREE
TOOL_PAX TOOL_PIC TOOL_PREPMKBOOTIMAGE TOOL_PWD_MKDB TOOL_REFER
TOOL_ROFF_ASCII TOOL_ROFF_DVI TOOL_ROFF_HTML TOOL_ROFF_PS
TOOL_ROFF_RAW TOOL_RPCGEN TOOL_SOELIM TOOL_SUNLABEL TOOL_TBL
TOOL_UUDECODE TOOL_VGRIND TOOL_ZIC

For each, provide default in <bsd.sys.mk> of the form:
TOOL_FOO?= foo
and for the ${USETOOLS}=="yes" case in <bsd.own.mk>, provide override:
TOOL_FOO= ${TOOLDIR}/bin/${_TOOL_PREFIX}foo

Document all of these in bsd.README.

This cleans up a chunk of potential (and actual) namespace collision
within our build infrastructure, as well as improves consistency in
the share/mk documentation and provision of appropriate defaults for
each of these variables.

branches: 1.34.2;
Add a CPUFLAGS variable which can be set by the user's mk.conf or in
the environment:

CPUFLAGS Additional flags to the compiler/assembler to select
CPU instruction set options, CPU tuning options, etc.

Since CPUFLAGS is not implicitly set by any part of the make infrastructure,
it is safe to set in mk.conf, unlike COPTS or DBG.

Allow a memory disk image to be statically included in the kernel
to be specified in the kernel configuration file, like so:

makeoptions MEMORY_DISK_IMAGE="/path/to/image.fs"

Add ${EXTRA_OBJ} so that pre-compiled .o's can be linked into the kernel.

Remove an antiquated comment and make .depend depend on config_time.h
so that the depend target can properly complete now that perry has the
time to use the time.

Add -Wno-sign-compare; that flag was made implicit with -Wall in
GCC 3.3, and our sources aren't really up to the task, yet.

Alter config so that it emits a config_time.src file (the source file
for config_time.h) that contains, for example:

/* Sun Nov 17 05:37:51 2002 GMT */
#define CONFIG_TIME 1037511471
#define CONFIG_YEAR 2002
#define CONFIG_MONTH 11
#define CONFIG_DATE 17
#define CONFIG_HOUR 5
#define CONFIG_MINS 37
#define CONFIG_SECS 51

These values represent the current time as of when config was last
run, so that functions (eg, inittodr()) can use these values instead
of being updated once every year or so with the "current" time.

The associated modification to Makefile.kern.inc makes config_time.h
depend on this depend on this and the kernel Makefile, so that
granularity of kernel builds is not reduced.

Remove an extra }

lint infrastructure overhaul to finally lint the kernel libs right.

XXX the dependencies for the kernel .ln files are just plain wrong. We
need some sort of extended make depend hack to do it right. We will
suffer with it for now.

use -X to ignore static function unused warnings, and echo the command.

Back out revision 1.23, it breaks all kernel configs that do

makeoptions COPTS="foo"

There are quite a few of those in the tree, as well.

all ports are now fully prototyped and pointer-arith clean.
turn on these warnings all the time.

Replace COPTS?= with COPTS+= so that a setting of COPTS in mk.conf
(common on sparc systems) does not cause the -O2 in DEFCOPTS to
be lost - which results in an unloadable kernel on sparc.

Reviewed by: christos

Merge the gehenna-devsw branch into the trunk.

This merge changes the device switch tables from static array to
dynamically generated by config(8).

- All device switches is defined as a constant structure in device drivers.

- The new grammer ``device-major'' is introduced to ``files''.

device-major <prefix> char <num> [block <num>] [<rules>]

- All device major numbers must be listed up in port dependent majors.<arch>
by using this grammer.

- Added the new naming convention.
The name of the device switch must be <prefix>_[bc]devsw for auto-generation
of device switch tables.

- The backward compatibility of loading block/character device
switch by LKM framework is broken. This is necessary to convert
from block/character device major to device name in runtime and vice versa.

- The restriction to assign device major by LKM is completely removed.
We don't need to reserve LKM entries for dynamic loading of device switch.

- In compile time, device major numbers list is packed into the kernel and
the LKM framework will refer it to assign device major number dynamically.

Build a .gdbinit in the kernel build directory. The new .gdbinit
will source all of the system-provided gdb scrips in sys/gdbscripts,
as well as any files specified by the user in the GDBINIT make var.

vers.o should also depend upon $S/conf/newvers.sh and $S/conf/osrelease.sh

cscope.out target needs depend too, zap old comments

Fix cscope and mkid targets so that they work even when the kernel
compile directory is not under /usr/src/sys (i.e. when 'S' is not
'../../../..'). Pointed out by Robert Elz in PR 17384.

Thanks again to Andrew Brown for figuring out how to rip .depend apart.

Updated version of cscope/mkid support. Check libkern and compat lib
for source files. Also include header files in mkid run.

Add cscope and mkid (and cscope.out and ID) targets to kernel compile
Makefiles. The main feature added by these targets is that they cover
ONLY the source files used for a given kernel and no other. Thus when
examining MD routines provided by all machines, you will see only
those applicable for your kernel.

Make sure DBSYM is defined if not overridden by USETOOLS=yes.

Add -Wno-format-zero-length, for forward compatibility with gcc 3.2.
Our in-tree gcc 2.95.3 accepts the flag as well.

branches: 1.13.2; 1.13.4;
lint target depends on ${LINTSTUBS}, not on (nonexistent) ${KERNLINTSTUBS}.

Make sure USETOOLS value is propagated (for USETOOLS?=no case in each
kernel Makefile). Perhaps we should .include this file earlier....

make .s and .S files depend on assym.h

branches: 1.10.2;
Add -ffreestanding to kernel builds. It's not always safe to
use GCC's built-ins.

don't strip any symbols from profiling kernels,
that makes it harder to figure out what's going on.

Make the kernel depend on `vers.o' instead of `newvers'. This
eliminates the need to unconditionally rebuild vers.o and subsequently
to relink the kernel every time you type make, even if you just
finished one.

Add the rest of the infrastructure needed to support the rest of the
ports. This includes cleaning out DBG, cleaning up the `clean'
target, and tweaking the warnings flags (cesfic, amigappc, and the arm
ports are a little less warning resistant).

Oh, and let's `install' the kernel into ${DESTDIR} if someone says
`make install'. We have to think about cross-compilers here.

Misplaced colon.

Alter the way that LINKTEXT and LINKDATA (and LINKENTRY) are set so
that empty values for TEXTADDR and DATAADDR (and ENTRYPOINT) will not
screw things up. Add support for SYSTEM_LD_TAIL_EXTRA which some
ports (not yet converted) are using. Add support for GENASSYM_EXTRAS
which has just been added to (some of) the arm ports.

Instead of checking for previous usage, just append to some make
variables. Tweak some variables to provide for more uniform defaults.
Minor adaption to makefiles for already converted ports.

branches: 1.3.2;
The four tools mkdep, strip, objcopy, and objdump are not part of the
defined toolchain if we're not using the new toolchain. We need them,
so we add our own definitions here.

Add a MKDEP?=mkdep line so that depend targets will work again.

Abstract out a large chunk of the kernel build machinery so that new
build features (such as ross's DEBUGLIST) can easily be applied to all
ports. This should reduce the complexity of each port's kernel
Makefile considerably. Line counts:

227 arch/i386/conf/Makefile.i386.orig
98 arch/i386/conf/Makefile.i386
227 arch/alpha/conf/Makefile.alpha.orig
99 arch/alpha/conf/Makefile.alpha
219 arch/sparc/conf/Makefile.sparc.orig
102 arch/sparc/conf/Makefile.sparc
215 arch/vax/conf/Makefile.vax.orig
102 arch/vax/conf/Makefile.vax
253 conf/Makefile.kern.inc

Roll i386, alpha, sparc, and vax over to the new build machinery.

assym.mk: Fix assym.d generation

It seems that the genassym/mkdep output has been changed.
This commit adapts the assym.d generation rule to the new output.

factor out common code.

Drop -fstack-usage* from CFLAGS passed genassym.
We don't want it to create a "-.su" file.

branches: 1.5.2; 1.5.18;
Fix dependency of *.d -> assym.h in `-S' mode.

Simplify the dependency line further.

Improve previous.

Restore *.[sS] -> assym.h dependency in `-S' mode.

Clean up. Move assym.h/vers.c related code out of Makefile.kern.inc.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

Create compat_100 module infrastructure now that we have branched
for NetBSD-10

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

version dup3

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

note that COMPAT_X implies all COMPAT_X+N.

branches: 1.1.2;
clean up COMPAT_* options for native netbsd code:
- new series of files that are useful for saying "i want everything
since netbsd 1.4", etc.
- use the fact COMPAT_* options have future dependancies to remove
many redundant options.

removes about 3000 lines total across kernel configuration files.
tested about 30 random kernels in the changed list.

Fix version numbers in comments

branches: 1.2.4; 1.2.8;
Merge the [pgoyette-compat] branch

branches: 1.1.2;
file compat_netbsd80.config was initially added on branch pgoyette-compat.

branches: 1.2.6;
Fix version numbers in comments

branches: 1.1.2;
Add a new member to struct vfsstat and grow the unused members
The new member is caled f_mntfromlabel and it is the dkw_wname
of the corresponding wedge. This is now used by df -W to display
the mountpoint name as NAME=

Fix kernel builds.

aes(9): Rewrite x86 SSE2 implementation.

This computes eight AES_k instances simultaneously, using the
bitsliced 32-bit aes_ct logic which computes two blocks at a time in
uint32_t arithmetic, vectorized four ways.

Previously, the SSE2 code was a very naive adaptation of aes_ct64,
which computes four blocks at a time in uint64_t arithmetic, without
any 2x vectorization -- I did it at the time because:

(a) it was easier to get working,
(b) it only affects really old hardware with neither AES-NI nor SSSE3
which are both much much faster.

But it was bugging me that this was a kind of dumb use of SSE2.

Substantially reduces stack usage (from ~1200 bytes to ~800 bytes)
and should approximately double throughput for CBC decryption and for
XTS encryption/decryption.

I also tried a 2x64 version but cursory performance measurements
didn't reveal much benefit over 4x32. (If anyone is interested in
doing more serious performance measurements, on ancient hardware for
which it might matter, I also have the 2x64 code around.)

Prompted by:

PR kern/59774: bearssl 32-bit AES is too slow, want 64-bit optimized
version in kernel

aes(9): New 64-bit bitsliced implementation.

Derived from BearSSL's aes_ct64 code. Compared to the aes_ct code,
on machines with native 64-bit integer arithmetic, aes_ct64 should
have approximately:

- the same throughput for:
. CBC encryption,
. CCM encryption/decryption, and
. CBC-MAC;

- double the throughput for:
. CBC decryption,
. XTS encryption/decryption.

(aes_ct computes AES on two blocks at a time; aes_ct64 computes it on
four blocks at a time, with roughly the same number of instructions.
CBC encryption and CBC-MAC are inherently sequential; CCM, being a
combination of CTR and CBC-MAC, can only really be parallelized two
ways, so having four ways available doesn't help; and CBC decryption
and XTS admit parallelism limited only by the size of the inputs.)

Enable with `options AES_BEAR64'. Should be a reasonable default on
all platforms with 64-bit integer registers. Caveat: uses about 1200
bytes of stack space. (Could approximately halve that, like the
BearSSL aes_ct code, at some speed cost which I haven't measured --
by moving the br_aes_ct64_skey_expand logic into add_round_key in
aes_ct64_{enc,dec}.c.)

PR kern/59774: bearssl 32-bit AES is too slow, want 64-bit optimized
version in kernel

sys/conf/copts.mk: Drop fallthrough hack for zlib.c

branches: 1.11.6;
introduce new GCC 12 warning disables and use them in a few places

this introduces 4 new warning disable flags:

CC_WNO_MISSING_TEMPLATE_KEYWORD
CC_WNO_REGISTER
CC_WNO_STRINGOP_OVERREAD
CC_WNO_ARRAY_BOUNDS

and documents them in README.warnings. of these, the string op
and array bounds are both problematic (real bugs) and also spurious
(not real bugs), and the other 2 are mostly temporary for older
3rd party code.

add some new uses of CC_WNO_STRINGOP_OVERFLOW.

fix m68k build for gallium and GCC 12.

adapt to ${CC_WNO_IMPLICIT_FALLTHROUGH}

Use ${CC_WNO_IMPLICIT_FALLTHROUGH} instead of
the older style more complex expressions.

bsd.own.mk: rename GCC_NO_* to CC_WNO_*

Rename compiler-warning-disable variables from
GCC_NO_warning
to
CC_WNO_warning
where warning is the full warning name as used by the compiler.

GCC_NO_IMPLICIT_FALLTHRU is CC_WNO_IMPLICIT_FALLTHROUGH

Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.

bsd.own.mk: rename to CC_WNO_ADDRESS_OF_PACKED_MEMBER

Provide a single variable
CC_WNO_ADDRESS_OF_PACKED_MEMBER
with options for both clang and gcc, to replace
CLANG_NO_ADDR_OF_PACKED_MEMBER
CC_NO_ADDR_OF_PACKED_MEMBER
GCC_NO_ADDR_OF_PACKED_MEMBER

Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.

branches: 1.7.4;
sys/conf: fix typo in comment

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.

branches: 1.5.6;
XXX:

turn off -Wunused-but-set-variable for mips' linux_machdep.c and GCC 8.

convert HAVE_GCC == 7 to HAVE_GCC >= 7.

branches: 1.3.4;
mark pf.c, radeon_cs.c and via_dmablit.c for no fall thru error.

this was already done in the various modules, but not in the main
kernel.

be sure to only apply zlib.c's -Wno-error=implicit-fallthrough to GCC 7.
push the setting into the rump and module version too.

pass -Wno-error=implicit-fallthrough for zlib.c.

XXX: consider using copts.mk for various warning/copt flags passed
in kernel builds currently set via 'makeoptions' in files.* files.
this is suboptimal, as those all get embedded into the kernel with
config_file.h.

Welcome to 2025!

branches: 1.22.2;
Update copyright notice for 2024.

Welcome to 2023. Wrap lines so the years fit in to 80 columns with
a leading kernel log timestamp.

branches: 1.20.4;
Welcome to 2022!

welcome 2021. YES.

branches: 1.18.8;
Welcome to 2020.

branches: 1.17.4;
Welcome to 2019!

branches: 1.16.2; 1.16.4;
Update for 2018 new year

branches: 1.15.8;
goodbye 2016, hello 2017!

branches: 1.14.2;
WELCOME 2016!

Welcome to 2015!

branches: 1.12.4; 1.12.6;
Welcome to 2014!

branches: 1.11.2;
Happy 2013 everybody

branches: 1.10.6;
Welcome to 2012, Happy New Year!

Don't forget to adjust copyright date in new code.

branches: 1.9.8; 1.9.12;
Revert previous. Add 2011 to the list in old format.
Let board@ decide if it's ok to switch to the range format.

Welcome to 2011! Happy New Year!

Do not forget to update copyright notices when you add new code.

branches: 1.7.4;
Welcome to 2010.

Don't forget to update copyright notices when you add new code.

MUHAHAHA, FIRST POST!!!!111!1!!ONE!!1


(also, "welcome to 2009 and don't forget to update copyright notices
when you add new code" and all that stuff)

branches: 1.5.6; 1.5.10; 1.5.18; 1.5.20; 1.5.24;
Welcome to 2008.

Don't forget to update copyright notices when you add new code.

branches: 1.4.20; 1.4.26; 1.4.32;
Welcome to 2007.

Don't forget to update copyright notices when you add new code.

branches: 1.3.18; 1.3.22; 1.3.24;
Welcome to 2006.

Don't forget to update copyright notices when you add new code.

(On behalf of an absent friend. Maybe he can do this himself next year.)

branches: 1.2.10; 1.2.20;
Welcome to 2005.

Don't forget to update copyright notices when you add new code.

branches: 1.1.2; 1.1.4; 1.1.8;
Store the copyright text in conf/copyright, and use conf/newvers.sh
to generate the appropriate const char copyright[] = "...";
statement instead of hard coding it into kern/init_main.c.
Idea from Simon Burge.

branches: 1.1.2; 1.1.18;
Move less important part out of Makefile.kern.inc.

Expose struct kauth_cred for the benefit of the debugger. I can't convince gcc
to produce debug info for the structure if it does not appear in more than
one source file.

branches: 1.3.22; 1.3.40;
Remove some unecessary includes sys/user.h header.

Put a dummy function in here that uses a couple of the types.
Otherwise, gcc won't emit any debug info.

branches: 1.1.2; 1.1.4; 1.1.6; 1.1.12; 1.1.14; 1.1.16; 1.1.18; 1.1.20;
Add a stub file that includes a bunch of kernel headers and always get
compiled with -g. For the initial set, netbsd on amd64 grows by around
80KB. This allows much easier use of GDB for post-mortem debugging as
it can understand the layout of data structures. The additional data can
be strip(1)ped off normally for size constraint environments.

Obsoleted by @@@'s kernel makefile work.

branches: 1.2.42; 1.2.44;
Remove some junk, and add _KERNEL and _LOCORE.

branches: 1.1.1;
Initial revision

change -h to -n because it is more "portable"

Add -h to the link command so that we replace the target file each time; this
should fix:

*** Failed target: .BEGIN
*** Failed command: ln -sf /usr/src/sys/external/gpl2/dts/dist/arch/arm/boot/dts dts/arm
*** Error code 1 (ignored)
ln: dts/arm/dts: Permission denied

Add support for building overlay DT blobs specified in the kernel
config file like so:

makeoptions DTS_OVERLAYDIR="${THISARM}/conf"
makeoptions DTS_OVERLAYS="
my-cool-overlay.dts
"

The result is my-cool-overlay.dtbo in the kernel compile directory.

branches: 1.9.2;
Compile dtbs with -@ so that fdtoverlay(1) et.al. can symbolically
reference phandles.

branches: 1.8.2;
- Allow multiple .BEGIN targets
- Make their protection consistent

use -@ like bsd.klinks.mk

Use similar restrictions to bsd.klinks.mk for .BEGIN rules.

handle absolute $S paths; we want to add ../ if they are relative only.

Add more complexity because of the path mess in the dts files. Requested
by jmcneill@

branches: 1.3.2;
Add support for building dts files from multiple paths, idea from christos

branches: 1.2.2; 1.2.4; 1.2.6;
do dts depend processing in two passes, one for /include/ and one for #include

Move the dts rule to a separate file and add dependencies handling.

A driver for the NXP SC16IS7xx family of UART chips. This family is
mostly a 16C450 with features from the 16C550, 16C650 and 16C750 and
makes use of src/sys/dev/ic/com.c for much of the heavy lifting.

A BIG difference is that the chip family is attached via a I2C or SPI
bus and is in no way, shape or form, attached to the computer via a
computer bus.

Since spin locks are not allowed to be held while calling to the I2C
or SPI framework com.c grew a new hw feature flag where it could be
run at something other than a hard IP level and doesn't hold spin lock
(for very long). In particular, IPL_SOFTSERIAL. This method is
enabled via a hw feature flag and should not effect any other use of
com.c. Other changes with this new method of use include the use of
workqueues and such.

A further feature added to com.c was the abilty to take advantage of
the MCR prescaler that is present in the SC16IS7xx family and likely
present in the 16C650. This prescaler is tried if the baud rate
requested can not be met with the frequency oscillator that the chip is
built with.

The chip family does not use any particular frequency oscillator for
the baud rate generator. In order to deal with this the frequency can
be set via sysctl, a kernel option or with a FDT overlay if that is
supported.

Most of the features of the chip family are supported including the
gpio pins via gpiobus(4) that are present with some of the family
members.

The chip can attach via I2C or SPI and can make sure of FDT if that is
present in the system.

This chip family has been around for quite some time and there are a
number of breakout boards on Amazon. It is also reasonably priced and
you can roll your own board if you can perform simple SMD soldering as
very few additional components are required.

Put the pcfiic driver declaration in the global location, and adapt
the sparc64 instance to be a proper separate attachment.

Allow MODDULAR_AUTOUNLOAD_UNSAFE as a kernel build option. Defaults to
false, so no change in current behaviour if not selected.

scmi: Add Arm SCMI performance protocol support.

From OpenBSD, adapted for the NetBSD cpufreq sysctl interface.

branches: 1.1313.2;
version dup3

dk(4): Add support for discovering Atari TOS partitions as wedges

Any partitioning scheme which conforms to the Atari AHDI 3.00 spec should be
recognized by the new DKWEDGE_METHOD_TOS.

fix spelling mistakes, mainly in comments and log messages.

Re-factor the Goldfish RTC driver into attach-front-end and generic
back-end; Goldfish virtual devices can be found on virtual platforms
that don't use FDT.

Rework AHCISATA_EXTRA_DELAY for kern/56737

- Remove AHCI_QUIRK_EXTRA_DELAY as issue appears to be drive and
not controller related
- Replace AHCISATA_EXTRA_DELAY with AHCISATA_REMOVE_EXTRA_DELAY,
so defaulting to enabling the extra delay, as the downside of
slower probing on systems which do not need it is less than having
other systems intermittently fail to probe and attach drives
- Also allow disabling extra delay with AHCISATA_EXTRA_DELAY_MS = 0

We should return to this code to work out which of the extra delays
are needed, and how long they need to be. It may be that faster
systems are more likely to trigger the issue (I've only seen it on
a 13th gen i7-13700, though only tested on a limited set)

XXX pullup -10

config(5): enable NTP option implicitly when PPS_SYNC option is enabled.

PPS_SYNC requires NTP, but some code blocks are ifdef'ed without it,
thus allowing an attempt to build kernel with PPS_SYNC, and without NTP
(and it almost builds with one line change).

Update options(4) man page to reflect the change.

Solution suggested by riastradh.

Separate DesignWare watchdog driver and FDT glue.

remove lingering strip(4) remnants

Create compat_100 module infrastructure now that we have branched
for NetBSD-10

branches: 1.1304.2;


Split the BMP280 / BME280 driver into common code and create I2C and
SPI attachments.

Add driver for Cadence I2C controller.

MI PMAP EFI_RUNTIME support

use defparam; noted by tnn@.

make MAXLWP a real option that triggers rebuilds properly.

drm: Use CPPFLAGS.drmkms in all local drm drivers too.

This way we don't pollute the NetBSD kernel namespace with all the
Linux compat shim definitions needed to build drm, except for the
local drm drivers that need the API.

bwfm(4) uses firmload.

Add MODULAR_DEFAULT_VERBOSE option.

Add initial support for Intel GPIO chips

sys: Include files.acpica unconditionally.

This way acpica.h always exists unconditionally, so MI code can be
conditional on NACPICA.

It is not great to have such conditionals, but it's better to have

#include "acpica.h"

#ifdef NACPICA > 0

than to have

#if defined(__i386__) || defined(__x86_64__) || defined(__arm__)
#include "acpica.h"
#endif

#ifdef NACPICA > 0

and we can still grep for NACPICA to find places that could be
factored better.

Re-factor and overhaul the "mcp23s17gpio" driver as "mcpgpio", and
add support for 8-bit and I2C variants of the chip:
- MCP23008 / MCP23S08: 8-bit (I2C / SPI)
- MCP23017 / MCP23S17: 16-bit (I2C / SPI)
- MCP23018 / MCP23S18: 16-bit (I2C / SPI), open-drain outputs

The MCP23x17 and MCP23x18 are essentially identical, software-wise; we
merely report different GPIO pin capabilities (no push-pull output for
MCP23x18). Also, remove the tri-state capability that was previously
advertised by the old version of this driver; these chips have no way
to put the pin into a HI-Z mode.

All 3 I2C versions are supported, but the SPI front-end still only
supports the MCP23S17 for now (SPI autoconfiguration needs an overhaul).

mcp23s17gpio(4) remains present as a link to the new mcpgpio(4) man page.

XXX Still to-do: FDT integration, interrupt suppoort.

(File missed in prior commit.)

Add driver for DesignWare Ethernet Quality-of-Service controller.

s/sytem/system/

A driver and user land utility for the Sparkfun Serial Controlled Motor
Driver module as illustrated here:

https://www.sparkfun.com/products/13911

A SCMD module is a ARM SOC simular to a Arduino in front of a motor
driver chip. The single SCMD module can control two motors and up to
16 additional modules can be chained together using an internal I2C
bus. One can interface with the SCMD using tty uart commands, SPI or
I2C. The driver in this commit adds a kernel driver for the I2C and
SPI interfaces. The command line utility provides a set of
convenience commands that support most of the functions of the SCMD
and is able to use the tty uart mode, SPI user land or the included
kernel driver in a uniform manor.

The use of the SCMD module is mostly for small robots and the like,
but it can control anything that is controllable by voltage.

ahcisata(4): Introduce AHCI_QUIRK_EXTRA_DELAY quirk for devices that
need extra delays as done by AHCISATA_EXTRA_DELAY option.

Enable this quirk for "C600/X79 AHCI". Also add commented out quirk
entries for "Bay Trail SATA (AHCI)" and "Mobile AHCI SATA Controller",
for which non-reproducible failures worked around by extra delays have
been reported.

500 ms of delays inserted by these option/quirk may be too much. Add
AHCISATA_EXTRA_DELAY_MS option to adjust number of delays in ms, like:

----
options AHCISATA_EXTRA_DELAY_MS=200
----

Thanks prlw1@ and jun@ for testing!

add defflag for AHCISATA_EXTRA_DELAY

defparam NMBCLUSTERS_MAX.

Separate MI smbios interface from MD specific code.

Fix not to do if_down() before reconnect

Almost network interface do not use if_down() even when there is no
connectivity. So, pppoe(4) is also made be not used it.
This behavior can be rollbacked by SPPP_IFDOWN_RECONNECT option.

Put the MSGBUFSIZE option in opt_param.h so that changes of its setting
in kernel config files will be picked up automatically without needing
a "make clean". Added to opt_param.h since ports define their default
MSGBUFSIZE settings in their <arch/include/param.h> (or equivalent).

Added a kernel option to change the number of processing packets
at one pppoeintr()

Add a new link-aggregation pseudo interface named lagg(4)

- FreeBSD's lagg(4) based implementation
- MP-safe and MP-scalable

Add a parameter to change keepalive interval in each PPPoE I/F

Added SPPP_NORECV_TIME option to change pp_max_noreceive

Introduce SPPP_KEEPALIVE_INTERVAL option
to change the interval between LCP echo requests

branches: 1.1279.4; 1.1279.6;
defflag USBHIST_PRINT

remove unused "_DIAGNOSTIC" option and opt_diagnostic.h.
note that this is unrelated to the widely used "DIAGNOSTIC" option.

branches: 1.1277.2;
move vmt(4) from MD to MI, and add support vmt on aarch64. tested on ESXi-Arm Fling

- move from sys/arch/x86/x86/{vmt.c,vmtreg.h,vmtvar.h} to sys/dev/vmt/{vmt_subr.c,vmtreg.h,vmtvar.h},
and split the attach part of the cpufeaturebus and fdt
- add aarch64 vmware backdoor op
- add include guard to vmt{reg,var}.h
- Yet there is still some little-endian dependency. it needs to be fixed in order to work properly on aarch64eb

Wups, forgot to commit this with vether.

[ozaki-r] Changes to the kernel core for wireguard

Remove references to BRIDGE_IPF, it is now compiled in by default.

New ChaCha API in kernel.

This will enable us to adopt MD vectorized implementations of ChaCha.

Convert ieee80211_crypto_ccmp.c to new aes_ccm API.

This will make it easier to provide better hardware acceleration
without fpu enabling/disabling overhead for each block of data.

New cgd cipher adiantum.

Adiantum is a wide-block cipher, built out of AES, XChaCha12,
Poly1305, and NH, defined in

Paul Crowley and Eric Biggers, `Adiantum: length-preserving
encryption for entry-level processors', IACR Transactions on
Symmetric Cryptology 2018(4), pp. 39--61.

Adiantum provides better security than a narrow-block cipher with CBC
or XTS, because every bit of each sector affects every other bit,
whereas with CBC each block of plaintext only affects the following
blocks of ciphertext in the disk sector, and with XTS each block of
plaintext only affects its own block of ciphertext and nothing else.

Adiantum generally provides much better performance than
constant-time AES-CBC or AES-XTS software do without hardware
support, and performance comparable to or better than the
variable-time (i.e., leaky) AES-CBC and AES-XTS software we had
before. (Note: Adiantum also uses AES as a subroutine, but only once
per disk sector. It takes only a small fraction of the time spent by
Adiantum, so there's relatively little performance impact to using
constant-time AES software over using variable-time AES software for
it.)

Adiantum naturally scales to essentially arbitrary disk sector sizes;
sizes >=1024-bytes take the most advantage of Adiantum's design for
performance, so 4096-byte sectors would be a natural choice if we
taught cgd to change the disk sector size. (However, it's a
different cipher for each disk sector size, so it _must_ be a cgd
parameter.)

The paper presents a similar construction HPolyC. The salient
difference is that HPolyC uses Poly1305 directly, whereas Adiantum
uses Poly1395(NH(...)). NH is annoying because it requires a
1072-byte key, which means the test vectors are ginormous, and
changing keys is costly; HPolyC avoids these shortcomings by using
Poly1305 directly, but HPolyC is measurably slower, costing about
1.5x what Adiantum costs on 4096-byte sectors.

For the purposes of cgd, we will reuse each key for many messages,
and there will be very few keys in total (one per cgd volume) so --
except for the annoying verbosity of test vectors -- the tradeoff
weighs in the favour of Adiantum, especially if we teach cgd to do
>>512-byte sectors.

For now, everything that Adiantum needs beyond what's already in the
kernel is gathered into a single file, including NH, Poly1305, and
XChaCha12. We can split those out -- and reuse them, and provide MD
tuned implementations, and so on -- as needed; this is just a first
pass to get Adiantum implemented for experimentation.

cgd(4): Switch from legacy rijndael API to new aes API.

Rework AES in kernel to finally address CVE-2005-1797.

1. Rip out old variable-time reference implementation.
2. Replace it by BearSSL's constant-time 32-bit logic.
=> Obtained from commit dda1f8a0c46e15b4a235163470ff700b2f13dcc5.
=> We could conditionally adopt the 64-bit logic too, which would
likely give a modest performance boost on 64-bit platforms
without AES-NI, but that's a bit more trouble.
3. Select the AES implementation at boot-time; allow an MD override.
=> Use self-tests to verify basic correctness at boot.
=> The implementation selection policy is rather rudimentary at
the moment but it is isolated to one place so it's easy to
change later on.

This (a) plugs a host of timing attacks on, e.g., cgd, and (b) paves
the way to take advantage of CPU support for AES -- both things we
should've done a decade ago. Downside: Computing AES takes 2-3x the
CPU time. But that's what hardware support will be coming for.

Rudimentary measurement of performance impact done by:

mount -t tmpfs tmpfs /tmp
dd if=/dev/zero of=/tmp/disk bs=1m count=512
vnconfig -cv vnd0 /tmp/disk
cgdconfig -s cgd0 /dev/vnd0 aes-cbc 256 < /dev/zero
dd if=/dev/rcgd0d of=/dev/null bs=64k
dd if=/dev/zero of=/dev/rcgd0d bs=64k

The AES-CBC encryption performance impact is closer to 3x because it
is inherently sequential; the AES-CBC decryption impact is closer to
2x because the bitsliced AES logic can process two blocks at once.

Discussed on tech-kern:

https://mail-index.NetBSD.org/tech-kern/2020/06/18/msg026505.html

Add fault(4).

i.MX support merged into GENERIC kernel.

Make crypto/rijindael optional again as cprng_strong does no longer
depend on it. Dependency is explicitly declared in files.foo if a
component requires it.

Introduce POOL_NOCACHE, simple option to cancel pool_caches and go directly
to the pool layer. It is taken out of POOL_QUARANTINE.

Advertise POOL_NOCACHE for kMSan rather than POOL_QUARANTINE. With kMSan
we are only interested in the no-caching effect, not the quarantine. This
reduces memory pressure on kMSan kernels.

branches: 1.1264.2;
remove also files.netsmb from here

remove SMBFS and nsmb(4) - kernel part

it's unmaintained and supports only obsolete SMB1

dme needs mii

undo previous since config has been fixed

Fix MIPS builds by working around an apparent bug in config(8)'s elif
handling.

Also exclude EXEC_AOUT stuff for aarch64

XXX Pullup-9 and -8

Fix comment - thanks Ryo Shimizu

mips64 has compat_netbsd32 but cannot have exec_aout; all other users
of compat_netbsd32 need exec_aout

Addresses PR kern/55037.

XXX pullup-9

Add support for Broadcom GENET v5 ethernet controller as found on the
Raspberry Pi 4 (BCM2711).

Retire KLEAK.

KLEAK was a nice feature and served its purpose; it allowed us to detect
dozens of info leaks on the kernel->userland boundary, and thanks to it we
tackled a good part of the infoleak problem 1.5 years ago.

Nowadays however, we have kMSan, which can detect uninitialized memory in
the kernel. kMSan supersedes KLEAK: it can detect what KLEAK was able to
detect, but in addition, (1) it operates in all of the kernel and not just
the kernel->userland boundary, (2) it requires no user interaction, and (3)
it is deterministic and not statistical.

That makes kMSan the feature of choice to detect info leaks nowadays;
people interested in detecting info leaks should boot a kMSan kernel and
just wait for the magic to happen.

KLEAK was a good ride, and a fun project, but now is time for it to go.

Discussed with several people, including Thomas Barabosch.

COMPAT_43 requires COMPAT_60

Remove FDDI support.

Remove filemon(4).

Discussed on tech-kern:
https://mail-index.NetBSD.org/tech-kern/2020/01/13/msg025938.html

This was never (intentionally) enabled by default, and the design has
some shortcomings. You can get mostly the same results with ktrace,
as in usr.bin/make/filemon/filemon_ktrace.c which is now used instead
of filemon for make's meta mode.

If applications require higher performance than ktrace, or nesting
that ktrace doesn't support, we might consider adding something back
into the vfs system calls themselves, without hijacking the syscall
table. (Might want a more reliable output format too, e.g. one that
can handle newlines in file names.)

Remove Token Ring support.

Remove HIPPI support and the esh(4) driver that uses it. There have not
been any users of HIPPI for some time, and it is unlikely to be resurrected.

- options NAMECACHE_ENTER_REVERSE is no more.

- Partially sort the list of per-vnode namecache entries by using a TAILQ.
Put the real name to the head, and put dot and dotdot to the tail so that
cache_lookup_reverse() doesn't have to consider them.

branches: 1.1248.2;
Revert the removal of filemon.

add Analogix DisplayPort core driver

Retire filemon, discussed on tech-kern@.

Nuke crypto/arc4. Has not been used since 2003. Will not be missed.

Add support for Kernel Memory Sanitizer (kMSan). It detects uninitialized
memory used by the kernel at run time, and just like kASan and kCSan, it
is an excellent feature. It has already detected 38 uninitialized variables
in the kernel during my testing, which I have since discreetly fixed.

We use two shadows:
- "shad", to track uninitialized memory with a bit granularity (1:1).
Each bit set to 1 in the shad corresponds to one uninitialized bit of
real kernel memory.
- "orig", to track the origin of the memory with a 4-byte granularity
(1:1). Each uint32_t cell in the orig indicates the origin of the
associated uint32_t of real kernel memory.

The memory consumption of these shadows is consequent, so at least 4GB of
RAM is recommended to run kMSan.

The compiler inserts calls to specific __msan_* functions on each memory
access, to manage both the shad and the orig and detect uninitialized
memory accesses that change the execution flow (like an "if" on an
uninitialized variable).

We mark as uninit several types of memory buffers (stack, pools, kmem,
malloc, uvm_km), and check each buffer passed to copyout, copyoutstr,
bwrite, if_transmit_lock and DMA operations, to detect uninitialized memory
that leaves the system. This allows us to detect kernel info leaks in a way
that is more efficient and also more user-friendly than KLEAK.

Contrary to kASan, kMSan requires comprehensive coverage, ie we cannot
tolerate having one non-instrumented function, because this could cause
false positives. kMSan cannot instrument ASM functions, so I converted
most of them to __asm__ inlines, which kMSan is able to instrument. Those
that remain receive special treatment.

Contrary to kASan again, kMSan uses a TLS, so we must context-switch this
TLS during interrupts. We use different contexts depending on the interrupt
level.

The orig tracks precisely the origin of a buffer. We use a special encoding
for the orig values, and pack together in each uint32_t cell of the orig:
- a code designating the type of memory (Stack, Pool, etc), and
- a compressed pointer, which points either (1) to a string containing
the name of the variable associated with the cell, or (2) to an area
in the kernel .text section which we resolve to a symbol name + offset.

This encoding allows us not to consume extra memory for associating
information with each cell, and produces a precise output, that can tell
for example the name of an uninitialized variable on the stack, the
function in which it was pushed on the stack, and the function where we
accessed this uninitialized variable.

kMSan is available with LLVM, but not with GCC.

The code is organized in a way that is similar to kASan and kCSan, so it
means that other architectures than amd64 can be supported.

Remove acorn26's upc(4) devices.

Add support for internal DesignWare HDMI PHYs

Add Kernel Concurrency Sanitizer (kCSan) support. This sanitizer allows us
to detect race conditions at runtime. It is a variation of TSan that is
easy to implement and more suited to kernel internals, albeit theoretically
less precise than TSan's happens-before.

We do basically two things:

- On every KCSAN_NACCESSES (=2000) memory accesses, we create a cell
describing the access, and delay the calling CPU (10ms).

- On all memory accesses, we verify if the memory we're reading/writing
is referenced in a cell already.

The combination of the two means that, if for example cpu0 does a read that
is selected and cpu1 does a write at the same address, kCSan will fire,
because cpu1's write collides with cpu0's read cell.

The coverage of the instrumentation is the same as that of kASan. Also, the
code is organized in a way similar to kASan, so it is easy to add support
for more architectures than amd64. kCSan is compatible with KCOV.

Reviewed by Kamil.

Add a new member to struct vfsstat and grow the unused members
The new member is caled f_mntfromlabel and it is the dkw_wname
of the corresponding wedge. This is now used by df -W to display
the mountpoint name as NAME=

Add vHCI, a driver which allows to send and receive USB packets directly
from userland via /dev/vhci. Using this, it becomes possible to test and
fuzz the USB stack and all the USB drivers without having the associated
hardware.

The vHCI device has four ports independently addressable.

For each xfer on each port, we create two packets: a setup packet (which
indicates mostly the type of request) and a data packet (which contains
the raw data). These packets are processed by read and write operations
on /dev/vhci: userland poll-reads it to fetch usb_device_request_t
structures, and dispatches the requests depending on bRequest and
bmRequestType.

A few ioctls are available:

VHCI_IOC_GET_INFO - Get the current status
VHCI_IOC_SET_PORT - Choose a vHCI port
VHCI_IOC_USB_ATTACH - Attach a USB device on the current port
VHCI_IOC_USB_DETACH - Detach the USB device on the current port

vHCI has already allowed me to automatically find several bugs in the USB
stack and its drivers.

Switch from NIST CTR_DRBG with AES to NIST Hash_DRBG with SHA-256.

Benefits:

- larger seeds -- a 128-bit key alone is not enough for `128-bit security'
- better resistance to timing side channels than AES
- a better-understood security story (https://eprint.iacr.org/2018/349)
- no loss in compliance with US government standards that nobody ever
got fired for choosing, at least in the US-dominated western world
- no dirty endianness tricks
- self-tests

Drawbacks:

- performance hit: throughput is reduced to about 1/3 in naive measurements
=> possible to mitigate by using hardware SHA-256 instructions
=> all you really need is 32 bytes to seed a userland PRNG anyway
=> if we just used ChaCha this would go away...

XXX pullup-7
XXX pullup-8
XXX pullup-9

branches: 1.1237.2;
Add KASAN_PANIC, an option to turn KASAN warning into kernel panics,
requested by Siddharth. While here clarify a little.

Implement an aggressive psref leak detector

It is yet another psref leak detector that enables to tell where a leak occurs
while a simpler version that is already committed just tells an occurrence of a
leak.

Investigating of psref leaks is hard because once a leak occurs a percpu list of
psref that tracks references can be corrupted. A reference to a tracking object
is memorized in the list via an intermediate object (struct psref) that is
normally allocated on a stack of a thread. Thus, the intermediate object can be
overwritten on a leak resulting in corruption of the list.

The tracker makes a shadow entry to an intermediate object and stores some hints
into it (currently it's a caller address of psref_acquire). We can detect a
leak by checking the entries on certain points where any references should be
released such as the return point of syscalls and the end of each softint
handler.

The feature is expensive and enabled only if the kernel is built with
PSREF_DEBUG.

Proposed on tech-kern

Remove Designware timer code. This was used by the old Rockchip port and
is no longer required.

Introduce POOL_QUARANTINE, a feature that creates a window during which a
freed buffer cannot be reallocated. This greatly helps detecting
use-after-frees, because they are not short-lived anymore.

We maintain a per-pool fifo of 128 buffers. On each pool_put, we do a real
free of the oldest buffer, and insert the new buffer. Before insertion, we
mark the buffer as invalid with KASAN. On each pool_cache_put, we destruct
the object, so it lands in pool_put, and the quarantine is handled there.

POOL_QUARANTINE can be used in conjunction with KASAN to detect more
use-after-free bugs.

defparam all of the config variables associated with SYSV IPC stuff.
The variables were removed from sys/conf/param.c and moved into the
SYSV IPC code, but config options were never propagated via any opt_*
file.

This should fix an issue reported on netbsd-users list from Dima Veselov.

Note that this does not address other parameters included in that report,
including CHILD_MAX and NOFILE; this commit only affects items related to
the SYSV IPC code. Also note that this does not affect non-built-in
sysv_ipc modules, for which you need to update the Makefile to use any
non-standard config values - just like any other non-built-in modules
which have config params.

XXX Pull-up to -8 and -8-0

XXX Note that there are a couple of panic() calls in msginit() which
XXX really should be changed to simple printf() and then result in
XXX msginit failure. Unfortunately msginit() currently doesn't return
XXX a value so we cannot indicate failure to the caller. I will fix
XXX this is a future commit.

Remove compat_osf1, discussed on tech-kern@.

config glue for ssdfb

Add KCOV - kernel code coverage tracing device

The KCOV driver implements collection of code coverage inside the kernel.
It can be enabled on a per process basis from userland, allowing the kernel
program counter to be collected during syscalls triggered by the same
process.

The device is oriented towards kernel fuzzers, in particular syzkaller.

Currently the only supported coverage type is -fsanitize-coverage=trace-pc.

The KCOV driver was initially developed in Linux. A driver based on the
same concept was then implemented in FreeBSD and OpenBSD.

Documentation is borrowed from OpenBSD and ATF tests from FreeBSD.

This patch has been prepared by Siddharth Muralee, improved by <maxv>
and polished by myself before importing into the mainline tree.

All ATF tests pass.

compat_sunos depends upon compat_09. fixes:

[ 1.8785495] WARNING: module error: built-in module compat_sunos can't find builtin dependency `compat_09'
[ 1.8785495] WARNING: module error: built-in module compat_sunos prerequisite compat_09 failed, error 2

Add driver for Designware HDMI TX controller.

Move COMPAT_43 to preceed anything that depends on it

Spell COMPAT_43 correctly - no NETBSD in the middle

comma police.

COMPAT_FREEBSD also requires COMPAT_43

Adjust dependencies for COMPAT_FREEBSD and COMPAT_LINUX

Should address kern/53915

Merge the [pgoyette-compat] branch

Retire compat_ibcs2, as discussed on tech-kern@. FreeBSD did the same
recently.

Make ipmi driver available to other platforms.
Add ACPI attachment.

Remove compat_svr4 and compat_svr4_32, as discussed on tech-kern@ recently,
but also as discussed several times in the past.

Remove the COM_REGMAP option -- just use it all the time. While here,
garbage-collect the COM_FUNCMAP and COM_AU1X00 options, as there are
not used anywhere.

Introduce KLEAK, a new feature that can detect kernel information leaks.

It works by tainting memory sources with marker values, letting the data
travel through the kernel, and scanning the kernel<->user frontier for
these marker values. Combined with compiler instrumentation and rotation
of the markers, it is able to yield relevant results with little effort.

We taint the pools and the stack, and scan copyout/copyoutstr. KLEAK is
supported on amd64 only for now, but it is not complicated to add more
architectures (just a matter of having the address of .text, and a stack
unwinder).

A userland tool is provided, that allows to execute a command in rounds
and monitor the leaks generated all the while.

KLEAK already detected directly 12 kernel info leaks, and prompted changes
that in total fixed 25+ leaks.

Based on an idea developed jointly with Thomas Barabosch (of Fraunhofer
FKIE).

Add NVMM - for NetBSD Virtual Machine Monitor -, a kernel driver that
provides support for hardware-accelerated virtualization on NetBSD.

It is made of an MI frontend, to which MD backends can be plugged. One
MD backend is implemented, x86-SVM, for x86 AMD CPUs.

We install

/usr/include/dev/nvmm/nvmm.h
/usr/include/dev/nvmm/nvmm_ioctl.h
/usr/include/dev/nvmm/{arch}/nvmm_{arch}.h

And the kernel module. For now, the only architecture where we do that
is amd64 (arch=x86).

NVMM is not enabled by default in amd64-GENERIC, but is instead easily
modloadable.

Sent to tech-kern@ a month ago. Validated with kASan, and optimized
with tprof.

Add amdccp(4) driver for AMD Cryptographic Coprocessor, as found on the
A11xx Opterons. Driver currently provides RNG service only.

Rename pl061gpio to plgpio and split the device logic from the FDT glue.

Remove ISDN from the kernel. It has remained unmaintained for a long time,
is of poor quality, and is now an obstacle to MP-ification. It was removed
ten years ago from FreeBSD for the same reason.

This retires a big user of the mbuf API, and will ease maintenance of the
kernel.

Remove isic(4). It is part of ISDN, which we are now retiring.

Remove the "ifritz" driver (no man page). It is part of ISDN, which we are
retiring.

Remove ifpci(4). It is part of ISDN, which we are retiring.

Remove iavc(4).

Remove daic(4), it has never been functional.

ok martin@

Add libnv

Remove the network ATM code.

Retire the 'midway' driver. Discussed on tech-net@ recently and also three
years ago, part of removing the network ATM code.

Add support for kASan on amd64. Written by me, with some parts inspired
from Siddharth Muralee's initial work. This feature can detect several
kinds of memory bugs, and it's an excellent feature.

It can be enabled by uncommenting these three lines in GENERIC:

#makeoptions KASAN=1 # Kernel Address Sanitizer
#options KASAN
#no options SVS

The kernel is compiled without SVS, without DMAP and without PCPU area.
A shadow area is created at boot time, and it can cover the upper 128TB
of the address space. This area is populated gradually as we allocate
memory. With this design the memory consumption is kept at its lowest
level.

The compiler calls the __asan_* functions each time a memory access is
done. We verify whether this access is legal by looking at the shadow
area.

We declare our own special memcpy/memset/etc functions, because the
compiler's builtins don't add the __asan_* instrumentation.

Initially all the mappings are marked as valid. During dynamic
allocations, we add a redzone, which we mark as invalid. Any access on
it will trigger a kASan error message. Additionally, the compiler adds
a redzone on global variables, and we mark these redzones as invalid too.
The illegal-access detection works with a 1-byte granularity.

For now, we cover three areas:

- global variables
- kmem_alloc-ated areas
- malloc-ated areas

More will come, but that's a good start.

Retire EtherIP, we have L2TP instead.

More ipkdb removal

Remove the kernel PMC code. Sent yesterday on tech-kern@.

This change:

* Removes "options PERFCTRS", the associated includes, and the associated
ifdefs. In doing so, it removes several XXXSMPs in the MI code, which is
good.

* Removes the PMC code of ARM XSCALE.

* Removes all the pmc.h files. They were all empty, except for ARM XSCALE.

* Reorders the x86 PMC code not to rely on the legacy pmc.h file. The
definitions are put in sysarch.h.

* Removes the kern/sys_pmc.c file, and along with it, the sys_pmc_control
and sys_pmc_get_info syscalls. They are marked as OBSOL in kern,
netbsd32 and rump.

* Removes the pmc_evid_t and pmc_ctr_t types.

* Removes all the associated man pages. The sets are marked as obsolete.

branches: 1.1200.2;
shuffle mechanics of files.virtio config(5) include location,
without actually moving the contents thereof from dev/pci yet

Add kernel API for pwm channels.

Remove COM_AWIN option

Remove MOTG_ALLWINNER option

branches: 1.1196.2;
Remove the Econet code. It was part of acorn26, which was removed a
month ago.

move "define tcbus {}" into files.alpha. it isn't wanted for general
ports, and generates an unused symbol for pmax. with this both alpha
and pmax kernel work with gcc 5 and 6.

Add a new option CISS_NO_INTERRUPT_HACK for driving ciss(4) by callouts.
This is intended as workaround for certain Xen issues with dom0 kernels
and will likely want to have a high HZ value as well for decent
performance.

Add a proper defflag for GPROF, and include opt_gprof.h, otherwise we're
not gonna go very far.

implement 32-bit compat support for raidframe.

convert several raidframe ioctls to be bitsize idempotent so that
they work the same in 32 and 64 bit worlds, allowing netbsd32 to
configure and query raid properly. remove useless 'row' in a few
places. add COMPAT_80 and put the old ioctls there.

raidframeio.h:
RAIDFRAME_TEST_ACC
- remove, unused
RAIDFRAME_GET_COMPONENT_LABEL
- convert to label not pointer to label
RAIDFRAME_CHECK_RECON_STATUS_EXT
RAIDFRAME_CHECK_PARITYREWRITE_STATUS_EXT
RAIDFRAME_CHECK_COPYBACK_STATUS_EXT
- convert to progress info not pointer to info
RAIDFRAME_GET_INFO
- version entirely.
raidframevar.h:
- rf_recon_req{} has row, flags and raidPtr removed (they're
not a useful part of this interface.)
- RF_Config_s{} and RF_DeviceConfig_s{} have numRow/rows removed.
- RF_RaidDisk_s{} is re-ordered slightly to fix alignment
padding - the actual data was already OK.
- InstallSpareTable() loses row argument

rf_compat32.c has code for RF_Config_s{} in 32 bit mode, used
by RAIDFRAME_CONFIGURE and RAIDFRAME_GET_INFO32.

rf_compat80.c has code for rf_recon_req{}, RF_RaidDisk_s{} and
RF_DeviceConfig_s{} to handle RAIDFRAME_FAIL_DISK,
RAIDFRAME_GET_COMPONENT_LABEL, RAIDFRAME_CHECK_RECON_STATUS_EXT,
RAIDFRAME_CHECK_PARITYREWRITE_STATUS_EXT,
RAIDFRAME_CHECK_COPYBACK_STATUS_EXT, RAIDFRAME_GET_INFO.

move several of the per-ioctl code blocks into separate functions.

add rf_recon_req_internal{} to replace old usage of global
rf_recon_req{} that had unused void * in the structure, ruining
it's 32/64 bit ABI.

add missing case for RAIDFRAME_GET_INFO50.

adjust raid tests to use the new .conf format, and add a case to
test the old method as well.

raidctl:
deal with lack of 'row' members in a couple of places.
fail request no longer takes row.
handle "START array" sections with just "numCol numSpare", ie
no "numRow" specified. for now, generate old-style configuration
but update raidctl.8 to specify the new style (keeping reference
to the old style.)

note that: RF_ComponentLabel_s::{row,num_rows} and
RF_SingleComponent_s::row are obsolete but not removed yet.

add ipsec(4) interface, which is used for route-based VPN.

man and ATF are added later, please see man for details.

reviewed by christos@n.o, joerg@n.o and ozaki-r@n.o, thanks.
https://mail-index.netbsd.org/tech-net/2017/12/18/msg006557.html

Merge autofs support from: Tomohiro Kusumi
XXX: Does not work yet

Generalise FDT and OFW

Fix the linux dependency. It does not depend on COMPAT_16, it just wants
the compat functions (not really controlled by COMPAT_NETBSD, but for the
principle).

Makes it possible to load compat_linux.kmod from the filesystem without
any COMPAT_* option compiled (but COMPAT_NETBSD).

Add support for I2C designware controllers (as found in Intel PCH devices),
with a pci front-end.
The pci front-end is tied to ACPI and Intel-specific, so it's in arch/x86/pci
and not dev/pci.
Core driver from OpenBSD, PCI front-end by me.

Factor out bus-independant HID code so that it can be shared by USB, bluetooth
and i2c.
dev/usb/ukbdmap.c is renamed to dev/hid/hidkbdmap.c
dev/usb/hid.[ch] moved to dev/hid/
usage pages moved from dev/usb/usbhid.h moved to dev/hid/hid.h,
and updated with OpenBSD entries.
bus-independant code moved from dev/usb/ums.c to dev/hid/hidms.c
(the same should be done for keyboard and touchpad drivers)

Needed for the upcoming HID over I2C support, proposed on tech-kern@
on Dec, 1.

Add driver for QEMU Firmware Configuration device.

This interface allows the host to pass various data items and files to
the guest OS.

Catch up with the new config version.

Add driver for Broadcom 802.11a/b/g/n/ac USB wireless devices, based on
the OpenBSD bwfm(4) driver.

(untested) attempt to fix b5 evbarm-earmv7hf build - please revert if this
is incorrect, or harms anything else...

Remove support for VERIFIED_EXEC_FP_RMD160, VERIFIED_EXEC_FP_SHA1, and VERIFIED_EXEC_FP_MD5 options.
These algorithms are either broken or on their way to being broken.

Discussed on tech-security
http://mail-index.netbsd.org/tech-security/2017/08/21/msg000936.html

ok riastradh

Add a helper for exposing LED controls via sysctl.

spelling (DEC called it "TURBOchannel")

bump required config version for multiline makeoptions feature

Remove the 'needs-flag' from spkr.c - it's not needed by anything.

New device wsbell - allows for a console beep for non pckbds (usb etc).
Works for platforms without pcppi - can work with spkr at audio and spkr
at pcppi.

To use add the following to your kernel config:
wsbell* at spkr? console?

Ok pgoyette@.

Add driver for ARM PrimeCell PL031 (RTC) Real Time Clock.

Add driver for ARM PrimeCell Advanced Audio CODEC interface (PL041).

Don't expect this driver to work on real hardware, but QEMU emulates it.