Support the brightness slider on PowerBook 1xx series.
Update copyright.

(nat)

Use the right location for powering off the PowerBook.

The previous location would only power off the computer if the AC adaptor was
connected.

(nat)

PR toolchain/58699: fix nbtool_config.h filename in tools/README

(martin)

Note about sti(4) support for 362 and 382 on hp300.

(tsutsui)

Add sti(4) support for 9000/362 and 9000/382 on-board framebuffers.

Mostly pulled from OpenBSD/hp300.

This supersedes old and dumb gendiofb.  Currently there is no mmap
or cmap (palette) support, but I'll investigate how Bt474 RAMDAC
should be handled.

Tested on 382 with 640x480 and 1024x768 variants, and 362 with 640x480.
No visible regression on 425e (EVRX) and 425t (CRX).

(tsutsui)

Avoid redundant and incorrect hypervisor detection in second
identify_hypervisor() call

The second call to identify_hypervisor() during autoconf needlessly repeats
hypervisor detection logic and could incorrectly override vm_guest values
(e.g., replacing VM_GUEST_KVM with VM_GUEST_VM). This change ensures that
if vm_guest is already set to a known hypervisor, the second pass skips
detection, preserving correct identification from the early call.

Tested on NetBSD/amd64, NetBSD/i386 with QEMU/KVM and QEMU/NVMM (BIOS and PVH),
Xen domU and Xen PVH.

(imil)

Use uint32_t instead of u_int32_t.

(tsutsui)

Introduce cpu_max_hypervisor_cpuid to cache hypervisor CPUID leaf

This variable stores the maximum supported hypervisor CPUID leaf so that
future checks can avoid repeated calls to x86_cpuid().

(imil)

enable apc(4) by default.

it seems that the original commit left it disabled as the linux
driver has a comment about it hanging on prototype systems (that
i don't recall ever hearing about with netbsd), but it's enabled
by default there, so let's enable it by default.

the real benefit here is that qemu-system-sparc detects this and
it idles on the host properly.

(mrg)

t_ptrace_*.c: Make debug output a little less verbose.

Maybe it will stop overflowing the releng testbed file systems this
way.

(riastradh)

Convert endianness of superblock on load and save.

(perseant)

Convert endianness when loading superblock and FATs.  There is likely more
that needs to be done in this regard.

(perseant)

Do not check boot block checksum by default.  Add -x flag to turn on this check, if desired by the user.
Convert the endianness of the superblock and other on-disk data structures.
Do not attempt to write the superblock if passed -n.

(perseant)

Expose exatfs_upcase()

(perseant)

Declare struct dkwedge_info

(perseant)

sljit: avoid compiler warning due to switch fallthru on mips

Unsure why, but this is only a problem when building sljit in
userspace via rump, which we don't do by default on mips,
although it's included in the OCTEON kernel.

Not enough to allow setting MKSLJIT=yes on mips but gets us
closer.

(nia)

Fix build for XEN not defined

(imil)

mention KERNEL_DIR

(pgoyette)

From FreeBSD D50078 by jrtc27:

Linux's struct stat does not have an st_flags member, but we still want
to be able to create FreeBSD disk images with non-zero st_flags. Add a
replacement member to struct fsinode to store it on Linux.

(christos)

Fix build on i386

(imil)

Remove conditionals on KERNEL_DIR, and reorder the potential boot
locations to put the legacy ones first.  This allows us to use a
single version of the bootstrap code regardless of the use of the
KERNEL_DIR build option, and should avoid spurious error messages
disrupting the automated testbeds.

Tested on i386 (non-KERNEL_DIR) and amd64 (both with and without
KERNEL_DIR).  In all cases the anita installations succeeded and
the atf tests were initiated.  Any remaining issues are most likely
due to subtle changes in text interaction and will probably need
updates to "expect" processing in anita.

If necessary this commit can be reverted, but please try to avoid
more churn, and update the testbed drivers as appropriate.
cvs: ----------------------------------------------------------------------

(pgoyette)

make: fix order of error messages in the ":?" modifier

While here: In Var_Parse, EvalStack_Pop needs to be called before
freeing the short-lived variable, as its name is part of the evaluation
stack.

(rillig)

tests/make: demonstrate wrong order of error messages

(rillig)

Introduce pvh_boot boolean to identify the real hypervisor when booting in PVH
mode.

As of now, sys/arch/x86/x86/identcpu.c / identify_hypervisor() returns in the
case of vm_guest being VM_GUEST_GENPVH, yet this VM type is not an actual
hypervisor but an information recorded in locore.S to drive boot method.
We need to investigate what type of hypervisor is really running the VM in
order to apply specifics, so instead of relying on vm_guest_is_pvh() which only
checks for VM_GUEST_XENPVH || VM_GUEST_GENPVH, pvh_boot informs on the boot
method while allowing to identify the real hypervisor.

Idea ok'd by bouyer@, tested on Xen domU, Xen dom0 with GENERIC PVH and
qemu GENERIC PVH boot.

(imil)

Load and check the upcase table when checking the filesystem.

(perseant)

Initialize fs structure, to make -w option work correctly.

(perseant)

t_ptrace_wait, t_ptrace_sigchld: Enable debug output.

Otherwise failures lead only to nearly useless output like:

FAILED: /work/src/tests/lib/libc/sys/t_ptrace_signal_wait.h:852: info.psi_siginfo.si_signo != sig

What _was_ si_signo, you ask?  Sorry, you have to guess or run the
test yourself!

(riastradh)

doc: mips enablement

(nia)

KNF - blank linebetween data and code

Also make sure we leave space in the path[] for a possible '.gz' to
be appended.

Both suggestions from kre@ - thanks

(pgoyette)

Bump a max number of bootable file (actually MAC address lines) to 16.

There are >10 HP9000 machines with Ethernet here.

(tsutsui)

Update to pathnames for booting under KERNEL_DIR

Tested on i386 (wwithout KERNEL_DIR) and amd64 (both with and without)

(pgoyette)

PR 59307: fix longjmp(9) for sparc64 too (there is only a single arg,
return value from setjmp is always 1).

(martin)

Fix typo.

Noticed by Miod Vallat.

(tsutsui)

Fix indent.

(tsutsui)

PR port-powerpc/59306 has been fixed, remove expected failure.

(martin)

PR port-powerpc/59306: kernel longjmp(9) only takes one argument,
always make setjmp(9) return 1.
Change as suggested by Taylor.

(martin)

PR port-sparc/59307 has been fixed, remove expected failure

(martin)

PR port-sparc/59307: kernel longjmp(9) only gets a single argument,
make setjmp(9) always return 1 via longjmp(9).

(martin)

Backout /filename/kernel changes, this breaks default installs
and needs more testing.

(martin)

PR port-sparc/59307 does not apply to sparc64

(martin)

If only a filename is given on the boot command, try /filename/kernel
and /filename/kernel.gz, and then /filename to minimize differences
between old andnew recovery procedures.  No change to ``naked'' boot
commands, nor to boot commands with pathname containing `;'

(pgoyette)

mk: Enable SSP on MIPS, since it works and defends properly.

Also carry over RELRO enablement to n64 variants.

PR pkg/59368

(nia)

PR 59370: compat 1.3 setjmp has been fixed

(martin)

PR 59370: compat_13: since we do not initialize the %o0 slot in jmpbuf
in setjmp() we need to store the full 64bit argument from longjmp()
there, otherwise the upper 32bit will be random garbage.

(martin)

Skip the test if we are not allowed to load modules.

(martin)

Make main() actually return zero. From Riastradh.
PR kern/59327: user stack pointer is not aligned properly

(martin)

Allow out of source builds (hrosenfeld)

(christos)

tests/lib/libc/setjmp: Mark compat 1.3 setjmp busted on sparc64.

PR port-sparc64/59370: compat 1.3 setjmp is busted

(riastradh)

xen: Stop-gap FPU PCB fix; disable Intel AMX for now.

Since the custom cpu_uarea_alloc/free are disabled under XENPV,
nothing would initialize struct pcb::pcb_savefpu to point either to
struct pcb::pcb_savefpusmall, or to a separately allocated large area
on machines with Intel AMX TILECFG/TILEDATA requiring it.  So the
memset in fpu_lwp_fork would crash on null pointer dereference:

[  1.0000030] uvm_fault(0xffffffff8094a300, 0x0, 2) -> e
[  1.0000030] fatal page fault in supervisor mode
[  1.0000030] trap type 6 code 0x2 rip 0xffffffff8062795c cs 0xe030 rflags 0x10202 cr2 0 ilevel 0 rsp 0xffffffff80adad38
[  1.0000030] curlwp 0xffffffff8078f880 pid 0.0 lowest kstack 0xffffffff80ad62c0
kernel: page fault trap, code=0
Stopped in pid 0.0 (system) at  netbsd:memset+0x2c:    repe stosq      %es:(%rdi)
memset() at netbsd:memset+0x2c
lwp_create() at netbsd:lwp_create+0x2f1
fork1() at netbsd:fork1+0x42c
main() at netbsd:main+0x44f

In order to support Intel AMX TILECFG/TILEDATA, or any other CPU
extensions that increase the XSAVE area beyond what fits in a single
page after struct pcb, we would need to enable the the custom
cpu_uarea_alloc/free.  Currently that would imply allocating stack
guard pages (`redzone') under XENPV; if there's some reason the stack
guard pages don't work, we could also push #ifdef XENPV conditionals
into cpu_uarea_alloc/free to cover the guard pages -- to be
considered.

PR kern/59371: Xen domU uvm_fault since FPU state allocation patch

PR port-amd64/57661: Crash when booting on Xeon Silver 4416+ in
KVM/Qemu

(riastradh)

regen.

(andvar)

fix typo in previous commit s/FHC/FCH/ for AMD RAID mode SATA controller.

(andvar)

clear whitespace in IPV6 neighbor solicitation

(joe)

regen.

(andvar)

Add AMD X370/X399/A320/A520/600 Series SATA/xHCI devices, 600 Series and Navi
PCIe Switch Upstream/Downstream Ports, Raven Ridge Vega GPUs and GeForce 6200A.

(andvar)

allow to specify the font's height in WSDISPLAYIO_SFONT.
now you can:
/home/ml# wsfontload -l
Gallant 12x22
Droid Sans Mono 10x20
Droid Sans Mono 12x22
Terminal 8x13
HP ROM 10x20
/home/ml# wsconsctl -dw font="Droid Sans Mono@20"
font -> "Droid Sans Mono@20"
/home/ml# wsconsctl -dw font="Droid Sans Mono@22"
font -> "Droid Sans Mono@22"
/home/ml# wsconsctl -dw font="Droid Sans Mono@15"
wsconsctl: WSDISPLAYIO_SFONT: No such file or directory

(macallan)

remove code to stash the console font in video memory - this was an experiment
based on guesswork which probably won't work if the card runs in anything
other than 1280x1024, and wasn't significantly faster than drawing characters
line by line using BINC writes.

(macallan)

t_sigstack: Don't do atf_tc_expect_fail then atf_tc_expect_signal.

atf interprets this sequence, with no intervening atf_tc_fail, to
mean the xfail failed to fail so it fails.  (Got that?)  That is, it
is interpreted somewhat like

atf_tc_expect_fail(...);
... there is supposed to be a failure here ...
atf_tc_expect_pass();
... remaining checks are expected to pass ...

where a lack of failures between the xfail and the xpass means the
expectation was not met.

But in this case, there are two compounding bugs:

1. the compat 1.3 setjmp has botched its registers, as uwe@
  discovered, so it is expected to crash on SIGSEGV (PR 59351); and

2. the compat 1.3 longjmp restores the stack poiter and signal mask
  in the wrong order, so it is expected to fail the test (PR 57946).

Since the crash in setjmp triggers before the failure after longjmp,
only do atf_tc_expect_signal; once that is fixed we can re-enable the
atf_tc_expect_fail for longjmp.

PR lib/57946: longjmp fails to restore stack first before restoring
signal mask on most architectures

PR port-arm/59351: compat_setjmp is busted

(riastradh)

Correct a couple of typos.

(kre)

Mention new KERNEL_DIR stuff

(pgoyette)

Actually include a bootable kernel in the KERNEL_DIR boot images.

This enables us to actually boot the image into qemu from /netbsd/kernel
and load modules from /netbsd/modules/... , keeping the modules and
kernel together, and removing kernel version numbers from the modules'
path.

From christos@

LGTM, commit it...

(pgoyette)

tests/kernel/setjmp_tester: Apparently kmods have .debug data too!

(Fix set list.)

PR port-riscv/59304: kernel longjmp(9) fails to make setjmp(9) return 1
PR port-hppa/59305: kernel longjmp(9) fails to make setjmp(9) return 1
PR port-powerpc/59306: kernel longjmp(9) fails to make setjmp(9) return 1
PR port-sparc/59307: kernel longjmp(9) fails to make setjmp(9) return 1
PR port-vax/59308: kernel longjmp(9) fails to make setjmp(9) return 1

(riastradh)

posix_spawn(2): Allocate a new vmspace at process creation time.

This allocates a new vmspace for the process at the time the new
process is created, rather than sharing some other vmspace temporarily.
This eliminates any risk of anything bad happening due to temporary
sharing, since there isn't any sharing.

Resolves a race to where:

1. we set up the child to share proc0.p_vmspace at first,

2. another process tries to read the new child's psstrings via
  kern.proc_args.<childpid>.argv or similar with the child's
  p_reflock held and gets stuck in a uvm fault loop because
  proc0.p_vmspace doesn't have the child's psstrings address
  (inherited from the parent) mapped,

3. the child is waiting for p_reflock before it can replace its
  p_vmspace or psstrings.

By allocating the vmspace up front, with no mappings in it, we avoid
exposing the child in this scenario.  Minor possible downside is that
sysctl kern.proc_args.<childpid>.argv might spuriously fail with
EFAULT during this time (rather than fail with EBUSY as it does if
p_reflock is held concurrently) but that's not a particularly big
deal.

Patch and first paragraph of commit message written by chs@; minor
tweaks to comments -- and any mistakes in the analysis -- by me.

PR kern/59037: deadlock in posix_spawn
PR kern/59175: posix_spawn hang, hanging other process too

(riastradh)

t_signal_and_sp: Try to fix mips o32 tests.

Evidently, for o32, PIC_PROLOGUE (which expands to .cpload) must be
the very first instruction of the function, or else it will get the
$gp calculation wrong.

Since it seems that the mips PIC_PROLOGUE does not, after all, mess
with the stack, let's just dispense with the temporary copy of sp in
t0 which I had added in paranoia over what magic might happen inside
PIC_PROLOGUE.  (Such paranoia is not entirely unjustified -- for
example, the sh3 PIC_PROLOGUE _does_ push data on the stack, though
this can be bypassed with PIC_PROLOGUE_NOSAVE.)

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

tests/kernel: New test for setjmp(9)/longjmp(9).

PR port-riscv/59304: kernel longjmp(9) fails to make setjmp(9) return 1
PR port-hppa/59305: kernel longjmp(9) fails to make setjmp(9) return 1
PR port-powerpc/59306: kernel longjmp(9) fails to make setjmp(9) return 1
PR port-sparc/59307: kernel longjmp(9) fails to make setjmp(9) return 1
PR port-vax/59308: kernel longjmp(9) fails to make setjmp(9) return 1

(riastradh)

lib/libc/arch/alpha: Say STACK_ALIGNBYTES, not 0xf, where relevant.

No functional change intended.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

lib/libc/arch/alpha/gen/_lwp.c, makecontext.c: Sort includes.

No functional change intended.

Preparation for tidying for:

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

Replace broken link to the GEM Gigabit Ethernet ASIC Specification
with a working archive.org link.

(gson)

t_signal_and_sp arm/execsp.S: Try using registers that aren't in use.

The first draft of this was instead using all the registers that
_are_ in use on entry to the elf entry point, which understandably
confused __start when we jumped to it after trashing all its inputs.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

t_signal_and_sp arm/signalsphandler.S: Fix pasto.

Copied & pasted this from contextspfunc.S and forgot to change one
reference to contextsp to signalsp instead, oops.

Should resolve:

*** Check failed: /work/src/tests/kernel/t_signal_and_sp.c:440: signalsp != NULL not met

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

t_ptrace_wait*: Deal with quirk of softfloat trap SIGFPE delivery.

PR misc/56820: Many FPE related tests fail on softfloat machines

(riastradh)

libc softfloat: Make SIGFPE for trapped fp exceptions non-ignorable.

In hardfloat, when the kernel delivers SIGFPE for a floating-point
exception trap, if SIGFPE is masked or ignored, it is handled _as if_
SIGFPE were unmasked and had the default signal disposition:

    964 /*
    965 * If the signal is masked or ignored, then unmask it and
    966 * reset it to the default action so that the process or
    967 * its tracer will be notified.
    968 */
    969 const bool ignored = action == SIG_IGN;
    970 if (masked || ignored) {
    971 mutex_enter(&ps->sa_mutex);
    972 sigdelset(mask, signo);
    973 sigdelset(&p->p_sigctx.ps_sigcatch, signo);
    974 sigdelset(&p->p_sigctx.ps_sigignore, signo);
    975 sigdelset(&SIGACTION_PS(ps, signo).sa_mask, signo);
    976 SIGACTION_PS(ps, signo).sa_handler = SIG_DFL;
    977 mutex_exit(&ps->sa_mutex);
    978 }

https://nxr.netbsd.org/xref/src/sys/kern/kern_sig.c?r=1.410#964

In other words, if you have asked for floating-point exception traps
via fpsetmask(3) or feeenableexcept(3), then you can't simply defer
or ignore them at the signal level any more than you can defer or
ignore SIGBUS or SIGSEGV.  And if the signal handler returns, it will
restart the instruction and -- if nothing has changed in the set of
trapped floating-point exceptions -- trap again, delivering the same
signal again.

Since we can't lock access to signal dispositions, I don't think we
can guarantee the same effect from userland atomically, but we can
get a close approximation (should be good enough for single-threaded
programs, at least) by mimicking this logic and calling sigqueueinfo
in a loop.

PR misc/56820: Many FPE related tests fail on softfloat machines

(riastradh)

tests/kernel/Makefile: Back out DPSRCS for t_execregs arch helpers.

This time for real -- accidentally backed out the wrong commit last
time:

https://mail-index.netbsd.org/source-changes/2025/04/26/msg156620.html

Doesn't work the way I'd hoped for update builds -- I had tested it
only with a .h file, not with the .S files:

#    create  kernel//work/src/tests/kernel/arch/arm/execsp.d
CC=/work/tools/bin/arm--netbsdelf-eabi-gcc /work/tools/bin/nbmkdep -f /work/src/tests/kernel/arch/arm/execsp.d.tmp  -- --sysroot=/branches/hosts-current/evbearmv5 -D_KERNTYPES -I/work/src/tests/kernel/..    -x assembler-with-cpp /work/src/tests/kernel/arch/arm/execsp.S &&  mv -f /work/src/tests/kernel/arch/arm/execsp.d.tmp /work/src/tests/kernel/arch/arm/execsp.d
nbmkdep: unable to write to file /work/src/tests/kernel/arch/arm/execsp.d.tmp: Permission denied

PR kern/59084: exec/spawn leaks register content

(riastradh)

t_execregs: Redo better diagnostics for child.

Accidentally backed out the wrong commit here:

https://mail-index.netbsd.org/source-changes/2025/04/26/msg156620.html

Meant to back out this one:

https://mail-index.netbsd.org/source-changes/2025/04/26/msg156607.html

Instead backed out this one (which I am now restoring):

https://mail-index.netbsd.org/source-changes/2025/04/25/msg156596.html

PR kern/59084: exec/spawn leaks register content

(riastradh)

distrib/amd64/cdroms/spec.in: Add buffer line after $NetBSD$.

Reduces chance of merge conflicts when changing content we might
plausibly edit.  No functional change intended.

(riastradh)

tests/lib/csu: Build h_hello.o with -fPIE, not -fpie.

Static libraries like libc may overflow the bounds assumed by -fpie.
Resolves:

h_hello.o: in function `main':
h_hello.c:(.text.startup+0x30): relocation truncated to fit: R_SPARC_GOT13 against symbol `__sF' defined in .data.rel section in /tmp/build/2025.04.27.10.22.23-sparc64/destdir/usr/lib/libc.a(findfp.o)

PR lib/59359: static pies are broken

(riastradh)

Add pie as set lists attribute

(martin)

Add new paths for kernel code when using KERNEL_DIR

(pgoyette)

Fix names of new CSU tests

(martin)

Remove a trailing tab.

(tsutsui)

Use proper uintNN_t integer types.

(tsutsui)

lib/csu: Add tests of {static, dynamic} x {pie, non-pie}.

Main goal here is to test static pies, but let's take advantage of
this to test the same code several ways, since the csu logic has some
differences for static vs dynamic builds too like the ordering of
_libc_init calls.

PR lib/59359: static pies are broken

(riastradh)

bsd.own.mk: If NOPIE is defined, set MKPIE=no.

This way, on sun2, we don't wind up with the bizarre setting of
MKPIC=no MKPIE=yes by default.

Preparation for conditionalizing automatic tests for:

PR lib/59359: static pies are broken

(riastradh)

Fix builds on NetBSD hosts.

Do not #include prop_object_impl.h in tools or _STANDALONE
builds, it (directly or indirectly) brings in far too much
that is inappropriate in those environments.

If its sole reason for being added was for the fallback
definition of _PROP_EXPORT then in those cases, simply
supply that here.

This is most probably not the most elegant of fixes, and a
better way of supplying that fallback definition than dumping
it in the middle of prop_object_impl.h should be found.

Someone else's problem - but DO test any alternative version
of this with a complete clean build on a NetBSD host before
committing.  And note, that since prop_object_impl.h includes
nbtool_config.h, when appropriate, it MUST be included before
any other system include file (but that change alone is not
sufficient to fix things, so do it this way instead).

This should unbreak all (well, all the previously unbroken)
builds on NetBSD hosts.

(kre)

t_execregs: implement sh3 h_execregs helper

Just dumps mcontext __gregset_t for now.
Not yet hooked into the build.

(uwe)

sh3/mcontext.h: make asm-friendly, respect _LOCORE

(uwe)

amd64/param.h: Fix KASAN/KMSAN build (and ALL build by extension).

Make UPAGES match what it was before my recent changes, as verified
by the __CTASSERT below, justifying the existence of the __CTASSERT.

As I recall, SVS is incompatible with KASAN/KMSAN, so it doesn't
contribute to the sum; presumably KASAN/KMSAN requires three pages,
though I'm not sure where this is documented.  If it turns out this
accounting is wrong, we should fix it and cross-reference any relevant
constraints affecting the accounting.  (But for now I'm just making
sure to restore the status quo of these definitions, which was my
intent all along with adding the __CTASSERT.)

Issue noted by hannken@.

PR port-amd64/57661: Crash when booting on Xeon Silver 4416+ in
KVM/Qemu

(riastradh)

tests/libexec/ld.elf_so/Makefile: Sort.

Preparation for:

PR bin/59360: ld.elf_so(8): missing RELR support

(riastradh)

sys/exec_elf.h: Add some missing numbers from the ELF gABI.

Fix the meta-number DT_ENCODING while here: it is defined to be 32,
not 31; it is a name for the smallest DT_* tag that follows the rule
of only even-numbered tags using d_ptr, and odd-numbered tags using
either nothing or d_val.  Nothing uses DT_ENCODING directly, though,
so this shouldn't affect anything.

https://www.sco.com/developers/gabi/latest/ch5.dynamic.html

Preparation for:

PR bin/59360: ld.elf_so(8): missing RELR support

(riastradh)

sys/exec_elf.h: Nix trailing whitespace.

No functional change intended.

Preparation for:

PR bin/59360: ld.elf_so(8): missing RELR support

(riastradh)

t_signal_and_sp: Fix RCSID in sh3/execsp.S.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

t_signal_and_sp: sh3 - mark expected failures

misaligned_sp_and_signal and signalsp_sigaltstack currently fail on
sh3 b/c the stack is not force-aligned for signal handlers.  Make
signalsphandler more robust by not touching the stack - we can save
r12 (GOT) in a register.

PR kern/59327: user stack pointer is not aligned properly

(uwe)

t_signal_and_sp: add sh3 support

PR kern/59327: user stack pointer is not aligned properly

(uwe)

mk: Enable partial relro on mips64.

(nia)

add an entry for ctftools

(christos)

avoid possible use of uninitilaized error

reviewed by christos@

(joe)

Instead of faking the kernel path in KERNEL_DIR, use proper entries in
the array. This has the advantage of producing correct error messages.

(christos)

t_signal_and_sp: Define __EXPOSE_STACK to get STACK_ALIGNBYTES.

This is a bit silly: sys/param.h unconditionally defines
STACK_ALIGNBYTES if machine/param.h does, but if machine/param.h
doesn't, then it only provides the default

#define STACK_ALIGNBYTES __ALIGNBYTES

under the condition that __EXPOSE_STACK is defined.

But using __EXPOSE_STACK here is a lower-risk change than touching
sys/param.h.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

t_signal_and_sp: Get STACK_ALIGNBYTES from sys/param.h.

Fills in a default of __ALIGNBYTES if machine/param.h doesn't define
it.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

Hide the private symbols.

(thorpej)

tests/kernel/Makefile: Back out DPSRCS for t_signal_and_sp helpers.

Doesn't work the way I'd hoped for update builds -- I had tested it
only with a .h file, not with the .S files:

#    create  kernel//work/src/tests/kernel/arch/arm/execsp.d
CC=/work/tools/bin/arm--netbsdelf-eabi-gcc /work/tools/bin/nbmkdep -f /work/src/tests/kernel/arch/arm/execsp.d.tmp  -- --sysroot=/branches/hosts-current/evbearmv5 -D_KERNTYPES -I/work/src/tests/kernel/..    -x assembler-with-cpp /work/src/tests/kernel/arch/arm/execsp.S &&  mv -f /work/src/tests/kernel/arch/arm/execsp.d.tmp /work/src/tests/kernel/arch/arm/execsp.d

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

tests/kernel/Makefile: Back out DPSRCS for t_execregs arch helpers.

Doesn't work the way I'd hoped for update builds -- I had tested it
only with a .h file, not with the .S files:

#    create  kernel//work/src/tests/kernel/arch/arm/execsp.d
CC=/work/tools/bin/arm--netbsdelf-eabi-gcc /work/tools/bin/nbmkdep -f /work/src/tests/kernel/arch/arm/execsp.d.tmp  -- --sysroot=/branches/hosts-current/evbearmv5 -D_KERNTYPES -I/work/src/tests/kernel/..    -x assembler-with-cpp /work/src/tests/kernel/arch/arm/execsp.S &&  mv -f /work/src/tests/kernel/arch/arm/execsp.d.tmp /work/src/tests/kernel/arch/arm/execsp.d
nbmkdep: unable to write to file /work/src/tests/kernel/arch/arm/execsp.d.tmp: Permission denied

PR kern/59084: exec/spawn leaks register content

(riastradh)

s/LIP Rset/LIP Reset/ in comment.

(andvar)

doc: two weeks of changes

(nia)

Unbreak bus_dmamem_map for mipsco.

Patch is sent from Miod Vallat:
> The switch to common code in 2009 did not account for private fields in
> bus_dma_segment_t, which are now left uninitialized but were relied upon
> by bus_dmamem_map().
>
> Remove these fields as their values were trivial to compute from the MI
> fields.
>
> This gives Ethernet and SCSI a chance to work again on this platform.

"switch to common code in 2009" was the following commit:
https://mail-index.netbsd.org/source-changes/2009/08/21/msg224405.html

We should eventually switch to the common mips bus_dma and it still
requires private _ds_vaddr for VIPT cache in MIPS3 case, but for now
his priority is to unbreak the code before attempting to improve it.

Note it looks the private _ds_paddr for mipsco was also pulled from
NetBSD/arc, where DMA address != physical address on jazzio devices.

(tsutsui)

Some USBHIST (aka KERNHIST) fixes from sc.dying on current-users.

(skrll)

Some USBHIST (aka KERNHIST) fixes from sc.dying on current-users.

(skrll)

A USBHIST (aka KERNHIST) fix from sc.dying on current-users.

(skrll)

Some USBHIST (aka KERNHIST) fixes from sc.dying on current-users. One
fixup from me.

(skrll)

Make endif comment match actual definition.

(andvar)

Remove dead code and variables from clock interrupt code.

These should have been removed on switching to using MI timecounter(9)
from MD microtime(9) impementation back in 2006:
https://mail-index.netbsd.org/source-changes/2006/09/15/0042.html
https://github.com/NetBSD/src/commit/12fed6d0

Patch from Miod Vallat.

(tsutsui)

Move ipl_sr_map to model-specific file and add IPL_DDB entry.

This will help adding support for other mipsco systems such as
the RC3240 eventually.

Patch from Miod Vallat.

(tsutsui)

Fix le(4) match routine to probe the address specified in config(5).

No functional changes with current kernel config files.

Also apply cosmetic changes:
- use proper variable name per examples in the bus_dma(9) man page
- fix an incorrect comment

Patch from Miod Vallat.

(tsutsui)

Remove obsolete MD 'PART_(ROOT|SWAP|BSD|RAW|USR|REST|FIRST_FREE)' defines.

Patch from Miod Vallat.

(tsutsui)

tests/kernel/Makefile: Add DPSRCS for t_execregs.c parts.

We build it with -DHAVE_EXECREGS_TEST if arch/.../execregs.h exists;
this way if you add it for a new architecture and do an update build,
t_execregs.c will be recompiled with the macro defined so it will
pick up the new stuff automatically.

PR kern/59084: exec/spawn leaks register content

(riastradh)

Add new file-system spec.in for KERNEL_DIR=yes

(pgoyette)

tests/kernel/Makefile: Add DPSRCS for t_signal_and_sp.c parts.

This way if you add, e.g., arch/foo/execsp.S, then t_signal_and_sp.c
will be recompiled with -DHAVE_EXECSP even if you don't make clean.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

Add $NetBSD$ keyword

(pgoyette)

librumphijack/t_tcpip: Fix clutter at /out.

PR misc/59354: librumphijack/t_tcpip leaves clutter at /out

(riastradh)

s/abadoned/abandoned/ and s/fobidden/forbidden/ in comments.

(andvar)

s/identiy/identity/ and s/chang /change/ in comments.

(andvar)

s/cahanged/changed/ in comment.

(andvar)

s/cahnge/change/ in comments.

(andvar)

Leave the kernel-dir sets in the $RELEASE directories so the
subsequent image-building targets can find them.

Use `ln -f'' to allow overwriting previous versions of links.

Addresses kern/55929

XXX More to come for building the image targets.

christos@ says ``go for it''

(pgoyette)

t_compat_cancel: Use -1 rather than SIGALRM for expected signal.

On kernels without (e.g.) COMPAT_13, we get SIGSYS rather than
SIGALRM if the cancellation point doesn't cancel the syscall.

Won't help if we try to interrupt in-progress waits, but currently
none of the tests do that; we'll burn that bridge when we come to it,

PR lib/59240: POSIX.1-2024: cancellation point audit
PR lib/59247: pthread_cancelstub.c is inadequately tested

(riastradh)

t_execregs: Make failure of child a little more obvious.

PR kern/59084: exec/spawn leaks register content

(riastradh)

t_execregs: On aarch64, make sure to align stack to 16 bytes.

Should avoid SIGBUS with strict alignment (SCTLR_EL0.A bit).

PR kern/59084: exec/spawn leaks register content

(riastradh)

t___sync_compare_and_swap: Mark tests xfail on armv5.

PR port-arm/56839: GCC emits wrong codes for compare_and_swap_1
bultins on armv5 (el & eb)

(riastradh)

Fix DES sources for sparc64

(martin)

t_signal_and_sp: Draft add arm support.

Compile-tested only, let's see how this goes.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

t_setjmp, t_sigstack: Mark compat13 setjmp tests xfail on arm.

(Does it even make sense to have compat_setjmp on arm?  Didn't we
break ABI compatibility with the oabi->eabi transition, long after
1.3?)

PR port-arm/59351: compat_setjmp is busted

(riastradh)

hppa makecontext: Use STACK_ALIGNBYTES to refer to stack alignment.

No functional change intended -- just make it clearer and easier to
search.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

hppa _lwp_makecontext: Align stack pointer.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

mips _lwp_makecontext: Align sp.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

hppa _lwp_makecontext, makecontext: Sort includes.

No functional change intended.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

alpha: Align signal pointer on entry to signal handler.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

mips: Align stack pointer on entry to signal handler.

Based on a patch by rin@.  The variant approach I took puts the stack
frame allocation and alignment logic in one place (getframe, used by
sendsig_siginfo for native (n64, on mips), netbsd32_sendsig_siginfo
for compat32 (n32/o32, on mips), and sendsig_sigcontext (compat 1.6))
and reduces the chance of provoking compiler exploitation of
undefined behaviour by doing arithmetic in uintptr_t rather than in
pointers to large aligned structs.  This also ensures the resulting
pointer is aligned for the object (struct siginfo_sigframe, struct
siginfo_sigframe32, struct sigcontext), not just for the ABI stack
alignment.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

mips __resumecontext: Align private stack frame size.

On mips64, UCONTEXT_SIZE is 664=0x298 (n32) or 744=0x2e8 (n64), but
sp is required to have 16-byte alignment in the ABI, so we need to
round the stack frame allocation up to 0x2a0 or 0x2f0.

No functional change intended on mips o32.  Unlikely to have a
practical change on mips n32/n64, but possible if signal delivery
wins a race -- until we fix sp alignment in signal handler calls,
coming soon.

Based on a patch by rin@.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

t_signal_and_sp: Check alignment of default thread stack too.

Not just with pthread_attr_setstack.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

lib/libc/arch/mips/gen/_lwp.c: Sort includes.

No functional change intended.

Preparation for:

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

sparc _lwp_makecontext, makecontext: Sort includes.

No functional change intended.

Tidying up for:

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

mips: Make sure stack is aligned on exec.

The only caller of setregs, sys/kern/kern_exec.c execve_runproc,
already guarantees STACK_ALIGNBYTES alignment, so let's assert it but
out of paranoia also clear the bits anyway just in case for non-
DIAGNOSTIC builds.

Based on a patch by rin@.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

mips/makecontext.c: Use STACK_ALIGNBYTES for clarity.

No functional change intended.

Tidying to make stack alignment more obvious, prompted by:

PR kern/59327: user stack pointer is not aligned properly

Part of a patch by rin@.

(riastradh)

mips/makecontext.c: Sort includes.

No functional change intended.

Preparation for:

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

x86: Make sure esp is aligned when delivering signal.

While here, use STACK_ALIGNBYTES consistently for the alignment mask
(or STACK_ALIGNBYTES32 in amd64 for the compat32 alignment mask).

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

Fix camelia source files for sparc64

(martin)

t_signal_and_sp: contextsplink test may not be broken on n64.

n32 and n64 have the same stack alignment requirement (16-byte), so
if there were a problem it would probably manifest on n32 too.  But
the only things that __resumecontext does with a misaligned sp are:

1. syscall getcontext (not even via the syscall stub, but via the
  syscall instruction directly), which is unlikely to care; and

2. tail-call setcontext, which does a syscall (again, unlikely to
  care) that will discard the caller's sp and replace it altogether.

So I'm not sure we can even test potential adverse consequences of
misalignment inside __resumecontext, short of interposing our own
setcontext symbol to artificially verify its stack alignment before
passing on to the syscall -- which, while feasible, is maybe not
worth the effort.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

t_signal_and_sp: Mark signalsp xfail on mips n32/n64.

The kernel takes an initially aligned stack pointer and subtracts
sizeof(struct sigframe_siginfo) or sizeof(struct sigframe_siginfo32)
from it -- that is, 872=0x368 or 792=0x318 bytes, respectively.  But
n32 and n64 require 16-byte stack alignment.

PR kern/59327: user stack pointer is not aligned properly

(riastradh)

Remove the now static _prop_format_default and _prop_string_type_tags.
Ride the minor bump from a few hours ago.
XXX should use -fvisibility=hidden and only export public symbols
XXX and then bump major

(martin)

tests/lib/libc/setjmp/t_sigstack: Update xfails.

Designations based on actual test run for i386 and code inspection
for arm and sh3 (and code inspection for all the other architectures
not listed because I expect the tests to pass on them -- they mostly
use sigreturn which will DTRT).

Move the ia64 xfail to just setjmp/sigsetjmp -- no compat13 on ia64.

Nix the long comment now that just about everything has been fixed.

PR lib/57946: longjmp fails to restore stack first before restoring
signal mask on most architectures

(riastradh)

pass lint

(christos)

1102

(snj)

Pull up following revision(s) (requested by martin in ticket #1102):
build.sh: revision 1.389-1.391 via patch
Fix "show-revisionid" operation to include both src and xsrc
data if the -x argument is present.
--
Minor cosmetic tweaks to previous
--
Simplify previous (from kre@)

(snj)

typo police.

(christos)