Properly tag for mips64 so we don't look for (and not find it) on 32-bit

(pgoyette)

Remove the entry for MD compat_13 and _16 code.  It's done for amd64
and arm, and mips will be done while verifying the mips builds.

(pgoyette)

Hook up the arm compat_netbsd32_13 & _16 code

(pgoyette)

Fix some NTP-related function pointers

(pgoyette)

WIP - clean up some of the Makefiles and conf/files stuff for the
netbsd32 modules.

(pgoyette)

oopppsss - it's opt_coredump.h not opt_compat_coredump.h

(pgoyette)

We've already done the compat_80 module stuff, so remove it.

The coredump issue was a red-herring, replace it with an entry for
completing and "hooking up" the compat_netbsd32_machdep_13 and _16
stuff

(pgoyette)

(Try to) move the arm-specific compat_netbsd32_machdep stuff to the
arm directory, not in acorn32!

XXX We still need to come back and hook up the _13 and _16 code!

(pgoyette)

Make sure we pick up the definition of COMPAT_NETBSD32

(pgoyette)

typo: insert missing ')'

(pgoyette)

Fix hdaudio device configuration

When disabling unassociated devices, we have a special handling
for pins of type COP_AWCAP_TYPE_PIN_COMPLEX, but it came after code
that may disable any pins, including the ones that should be handled
as COP_AWCAP_TYPE_PIN_COMPLEX.

The result was that hdaudio could fail to detect some devices.  We
fix the situation by making sure that COP_AWCAP_TYPE_PIN_COMPLEX
pins always get their specific handling.

The change makes the built-in speaker work on Dell OptiPlex 5060

(manu)

Work around deadlock between fstchg and fstcnt

When suspending a filesystem in fstrans_setstate(), we wait on
fstcnt for threads to finish transactions. While we do this, any
thread trying to start a filesystem transaction will wait on fstchg
in fstrans_start(), a situation which can deadlock.

The wait for fstcnt in fstrans_setstate() can be interrupted by
a signal, but the wait for fstchg in fstrans_start() cannot. Once
most processes are stuck in fstchg, it is impossible to send a
signal to the thread that waits on fstcnt, because no process
respond anymore to user input.

We fix that by adding a timeout to the wait on fstcnt in
fstrans_setstate(). This means suspending a filesystem may fail,
but it was already the case when the sleep was interupted by
a signal, hence calling function must already handle a possible
failure.

Fixes kern/53624

(manu)

HACK:  if calling dirname() with the results of a previous dirname()
there is no need to copy the path into the output buffer, it is already
there....

All this has to change to become compat with a forthcoming POSIX update.

(kre)

Initialize CML1 clock

(jmcneill)

Register clocks with clk_attach

(jmcneill)

Add dwiic(4) and spdmem(4).

(jakllsch)

match spdmem_i2c on "atmel,spd" compatible string

(jakllsch)

Set "i2c-indirect-config" to false in the right place so that iic(4)
better behaves itself.

(jakllsch)

Add dwiic_fdt attachment for "snps,designware-i2c".

(jakllsch)

According to POSIX moving the cursor in curses(3) touches the window

Mark the old and new lines as dirty, so they will be refreshed upon next
call to getch(3)-like routine.

This also matches the ncurses behavior.

Reviewed by <roy>

(kamil)

Sync with OpenBSD src/sys/dev/ic/dwiic.c r1.4.

Makes split command+data write operations in the exec() function work.
cvs: ----------------------------------------------------------------------

(jakllsch)

Revert previous - will pass the additional args via the environment.

(martin)

Correctly use device_printf() rather than aprint_error_dev() for errors
that can occur at runtime.

(jakllsch)

avoid hardcode. don't depend that AARCH64_KSEG_START is 0xffff000000000000.

(ryo)

Allow for ${COMPRESS_PROGRAM} to hold the name of a compressor program
plus additional arguments (like -9).

(martin)

Correct detecting of terminal resize in curses(3) with keypad(,TRUE)

A previous change fixed only keypad(,FALSE) scenarios.

Handle catching terminal resize in INKEY_NORM and INKEY_ASSEMBLING (in the
middle of assembling a key code from passed codes) as both accept keys with
fgetc(3) and both can be in theory interrupted with a resize.

PR lib/53615

(kamil)

Revive netipsec/keysock.h, accidentally marked obsolete in r1.2230.

Reported by Robert Swindells on current-users.

(rin)

Don't call ukbd_cnattach() unconditionally, this causes a non-functionnal
keyboard if there is a framebuffer but the console is serial only.
Instead add a fdt_device_register_post_config() callback, which checks
if the just-attached wsdisplay is the console, and calls ukbd_cnattach()
in this case.
OK jmcneill@

(bouyer)

Introduce wsdisplay_isconsole(struct wsdisplay_softc *), which
returns sc->sc_isconsole

(bouyer)

Disallow 0-byte transfers and suppress timeout messages

(jmcneill)

Replace the NTP issue with one for COREDUMP.  This is a major problem
for machine-dependant part of compat_netbsd32

(pgoyette)

Additional use of NTP function pointers, since the NTP code might not
exist in the kernel.  (NTP is not (yet) a module, so we don't need to
use the MP-safe mechanism.)

(pgoyette)

Cortex A5 also uses PPI 27 for timer.

ODROID-C1 boots again.

(skrll)

Add PCB used in a networking context.
Wondered by alnsn at EuroBSDcon 2018.

(sevan)

Add CPPFLAGS+= -std=c99 so tools can build on systems where the
system compiler doesn't default to C99 (eg, NetBSD 6.1).

(riz)

return 0 on success of dwiic_i2c_acquire_bus(), not 1 (failure)

(jakllsch)

use the standard code instead of ours

(christos)

no need for our copy of memequal

(christos)

Fix spelling of HOSTPROG to not enable NLS for tools build.

(joerg)

Don't include NLS for the tools build of xz.

(joerg)

Fix wrong macro, use An.

(wiz)

New sentence, new line. Use An/Mt.

(wiz)

Add missing Makefile.inc for xz tools.

(joerg)

pad(4) mixer has only 1 channel, so return EINVAL in the case other than 1.

This fixes the following strange output of mixerctl(1):

        outputs.master=255,0
        inputs.dac=255,0

(nakayama)

Revert to rev.1.53.
I accidentally committed the netbsd-8 branch file in rev.1.54.

(nakayama)

Close double quotation mark

(ryoon)

Update xz entry for 5.2.4.

(joerg)

Update build glue for xz-5.2.4. Hook up tools version.

(joerg)

Revert last change, side effect of old config.h.

(joerg)

Don't try to generate files we are not interested in.

(joerg)

Revert to vendor state.

(joerg)

Update xz import script to reflect inclusion of configure.

(joerg)

Include list.h so prototypes for funcs exist (appease gcc)

(kre)

In the xxx_fini() routine we need to _un_set the hook!

(pgoyette)

Make the replacement/hook function return an int, to match the
expectations of the MP-safe HOOK mechanism.

(pgoyette)

And use proper macro names

(pgoyette)

If we're going to use the MP-safe HOOK macros, it helps to define them
by including the proper header!

(pgoyette)

Adjust declaration of netbsd32_sendsig_siginfo() too.

Remove the attempted initialization of the former function pointer; this
is now handled by the "default" argument in MODULE_HOOK_CALL() macro.

(pgoyette)

Make the netbsd32_sendsig_siginfo() routine return an int, to conform
to the MP-safe HOOKs mechanism.

(pgoyette)

Put the NTP issue in the "required before merge" section.

Note that the compat library is completely gone now, both .o and .a
variants.

(pgoyette)

Provide required args when calling the default (un-hooked) routine.

(pgoyette)

Update note on NTP

(pgoyette)

Revert previous.  We need a MP-safe way to determine _at_runtime_ if
the NTP option was included or not.

(pgoyette)

syntax: add missing )

(pgoyette)

Convert yet another function pointer to the MP-safe mechanism

(pgoyette)

Remove two unused bs handles

(jmcneill)

Resolvconf is (can be) used as part of network config during
system installation, so also must confine itself to POSIX specified
test ('[') usage.    Pointed out by roy@ - thanks.

(kre)

partial HEAD sync - ISDN device removal

(jdolecek)

fix DIOCLWEDGES to at least not fail with ENOTTY due to struct size difference,
but probably more needs to be done to work if any actual wedges are configured

(jdolecek)

add support for DIOCGSTRATEGY and DIOCSSTRATEGY

(jdolecek)

put back sparccap.c; too hard to iron out the bn_mont stuff.

(christos)

sparcv9cap is a c file.

(christos)

use sparcv9cap.S

(christos)

replace KASSERT() for READ LOG EXT returning without error indication
with just log to avoid triggering panic under QEMU - it doesn't implement
the command for AHCI emulation

(jdolecek)

Remove tests that don't exist anymore in OpenSSL-1.1.1 (We should add the
one ones it adds though :-)

(christos)

Note that we need to clean objdir for the OpenSSL update.

(martin)

Don't include the libc CRYPTO_memcmp file (it has a different prototype
now) and it is provided by openssl in C, or use the assembly versions.

(christos)

don't override the sha man page provided by libc.

(christos)

Work in progress

(pgoyette)

Call drm_fb_helper_prepare in tegra_drm_fb_init as required.

I can boot a kernel with this

jetsontk1# dmesg | grep fb
[  1.0000580] tegrafb0 at tegradrm0
[  1.0000580] tegrafb0: framebuffer at 0x9bc00000, size 1024x768, depth 32, stride 4096
[  1.0000580] wsdisplay0 at tegrafb0 kbdmux 1
jetsontk1#

(skrll)

Don't go beyond start().

(maxv)

sparc has memcmp and cpuid

(christos)

Add more casts to make clang happy.

(christos)

Make a note about clockctl vs NTP

(pgoyette)

For clockctl module purposes, assume that NTP is available.

XXX Will revisit this later

(pgoyette)

Don't compile NTP-related case if NTP is not available.

No HOOK required here since NTP is not modularized.

(pgoyette)

Convert uses of test (aka '[') to use only posix specified forms,
mostly just on general principle...  this resulted in one or two minor
code reformattings to keep 80 char limits - a few needless uses of
quotes ("no" ??) were also removed (sh is not C. strings are strings
without quotes around them...)

(kre)

Since we need an int paramater, and uio_resid is size_t cast it to int
to avoid warnings from the ever friendly compiler... (check that size if
in range was already made).

(kre)

Open code min() so we don't need to find its prototype ...

(kre)

Convert uses of test (ie: '[') to use only POSIX specified forms,
just on general principle (all but 3 of the changes are in code that
is commented out...)

(kre)

Convert to POSIX specified usage of test (even though it would probably
have never been an issue here, it is trivial..., and one of the two
uses that needed changing was (and still is) commented out anyway).

(kre)

rc.subr can be used in install images (from sysinst) so must use only
POSIX specified test uses (no -a or -o).  Also, use printf always,
rather than echo (replace echo as a function using echo with one which
uses printf).

(kre)

Actually allocate/define the kern_proc_32 hook

(pgoyette)

Restore compat_netbsd32 for amd64, too.

(pgoyette)

we provide memcmp and rdtsc

(christos)

Use a .h file to define the xxx_{init,fini} routines, and include
the compat_stub.h header to get the definition of the kern_proc_32
hook.

(pgoyette)

Kill stack protector warnings.
XXX: need to understand why gcc complains; it is not obvious to me.

(christos)

Reenable the compat_netbsd32 module.  It is _NOT_ an all-version monolithic
module (as was compat_netbsd).

(pgoyette)

nvlist_copyin: fix the flag handling.

(rmind)

pad(4) mixer has only 1 channel, so return EINVAL in the case other than 1.

This fixes the following strange output of mixerctl(1):

outputs.master=255,0
inputs.dac=255,0

(nakayama)

Support rescan and child detachment

(jmcneill)

libnv: add a wrapper around free(9) since in FreeBSD it can take NULL.

(rmind)

Tickets #1020 - #1029, #1031 - #1035

(martin)

Pull up following revision(s) (requested by nakayama in ticket #1035):

sys/dev/pci/files.pci: revision 1.403

ixg and ixv depend on mii and mii_phy.

(martin)

Apply patch, requested by manu in ticket #1034:

Avoid sleeping with a spin lock held (solved differently on HEAD).

(martin)

Pull up following revision(s) (requested by tsutsui in ticket #1033):

sys/arch/pmax/pmax/dec_3min.c: revision 1.74

Fix hangup after framebuffers are attached on 3MIN.  PR port-pmax/53611

Ok'ed by mrg@.  Should be pulled up to netbsd-7 and netbsd-8.

(martin)

Pull up following revision(s) (requested by tsutsui in ticket #1032):

usr.sbin/sysinst/arch/sgimips/md.c: revision 1.5
usr.sbin/sysinst/arch/sgimips/md.c: revision 1.6

PR port-sgimips/53583: fetch kernel name always before using it.

Make the struct utsname local to avoid future similar issues.
Make sure to install a bootloader even on upgrade installation.

-

Fixes another part of port-sgimips/53583.
Should be pulled up (with rev 1.5) to netbsd-7 and netbsd-8.

(martin)

Pull up following revision(s) (requested by bouyer in ticket #1031):

sys/kern/kern_synch.c: revision 1.317

In mi_switch(), also call pserialize_switchpoint() if we're not switching
to another lwp, as proposed on
http://mail-index.netbsd.org/tech-kern/2018/07/20/msg023709.html

Without it, on a SMP machine with few processes running (e.g while
running sysinst), pserialize could hang for a long time until all
CPUs got a LWP to run (or, eventually, forever).

Tested on Xen domUs with 4 CPUs, and on a 64-threads AMD machine.

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #1029):

sys/dev/pci/ixgbe/ixgbe.c: revision 1.166

- Fix a bug that Denverton accessed wrong register in
  ixgbe_setup_low_power_mode(). The location of some registers in Denverton
  were changed. The GRC register is one of them, so use IXGBE_GRC_BY_MAC(hw)
  instead of IXGBE_GRC.

- ixgbe_check_wol_support() tests the capability of WOL, but the sysctl
  knobs for WOL were added only for IXGBE_DEV_ID_X550EM_X_10G_T. Add the
  knobs when adapter->wol_support is not 0.

XXX pullup-8

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #1028):

sys/dev/pci/pci_subr.c: revision 1.203
sys/dev/pci/pcireg.h: revision 1.140

  Add ATS Relaxed Ordering supported bit described in Address Translation
Relaxed Ordering ECN.

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #1027):
sys/dev/pci/if_wm.c: revision 1.586
sys/dev/pci/if_wm.c: revision 1.587
sys/dev/pci/if_wm.c: revision 1.588

- Split txdrop evcnt into pcqdrop and descdrop.
- Simplify evcnt name by removing duplicated "tx" or "rx".
- s/tu/underrun/

- Try m_defrag() to reduce the number of DMA segment if bus_dmamap_load_mbuf()
  returned EFBIG. When m_defrag() is called, txqNNdefrag event counter is
  incremented. If the 2nd try of bus_dmamap_load_mbuf() failed, txqNNtoomanyseg
  event counter is incremented.

- Reduce the max number of DMA segments from 256 to 64 (it's the same value
  as other BSD's (EM_MAX_SCATTER) and more than before if_wm.c rev. 1.75's
  value (40)) because we do m_defrag() now.

  Fix comment.

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #1026):

sys/arch/x86/x86/procfs_machdep.c: revision 1.24
sys/arch/x86/include/specialreg.h: revision 1.130

OK'd by maxv:
- Add cpuid 7 edx L1D_FLUSH bit.
- Add IA32_ARCH_SKIP_L1DFL_VMENTRY bit.
- Add IA32_FLUSH_CMD MSR.

(martin)

Pull up following revision(s) (requested by mrg in ticket #1025):

sys/kern/subr_autoconf.c: revision 1.263
sys/kern/kern_drvctl.c: revision 1.44
sys/sys/device.h: revision 1.156
sys/sys/systm.h: revision 1.278

- move export for devmon_insert_vec into sys/device.h.
- export root_is_mounted for future USB RB_ASKNAME hack.
- make some things in subr_autoconf.c static
- move device_printf() prototype out from the middle of two sets of
  aprint_*() prototypes.

(martin)

Pull up following revision(s) (requested by mrg in ticket #1024):

sys/ddb/db_command.c: revision 1.157
sys/ddb/db_command.c: revision 1.158
sys/kern/subr_lockdebug.c: revision 1.67

always call lockdebug_dismiss() from DDB -- there are always some
minimal lockdebug checks in place, even without LOCKDEBUG.

adjust lockdebug_abort() to ignore problems after ld_panic is set

so that there's a chance of this working.
this fixes ddb 'reboot' on softiron od1000.

call spl0() before cpu_reboot(), so that there's a chance that:
- interrupts can work afterwards
- this also means if IO stalls, serial break might work again.

this mimics how reboot(2) ends up calling cpu_reboot().

(martin)

Pull up following revision(s) (requested by mrg in ticket #1023):

sys/dev/scsipi/scsipi_verbose.c: revision 1.34

don't print a blank line that ends up in logs and console output.

(martin)

Pull up following revision(s) (requested by mrg in ticket #1022):

sys/dev/pci/hdaudio_pci.c: revision 1.9
sys/dev/pci/hdaudio_pci.c: revision 1.10

use pci_mapreg_type() instead of assuming PCI_MAPREG_TYPE_MEM.
suggested by riastradh@.

fixes hdaudio on thinkpad P51 where the registers are mapped
over 4GB range.  probably fixes intel X10-based systems that
map hdaudio high as well, but i was able to work around that
in the BIOS.  not so for the P51.

XXX: pullup-7, pullup-8.

use pa_dmat64 if pci_dma64_available().

(martin)

Pull up following revision(s) (requested by rin in ticket #1021):

sys/dev/hpc/hpckbd.c: revision 1.33
sys/dev/hpc/hpckbd.c: revision 1.34
sys/dev/hpc/hpckbd.c: revision 1.35

Apply workaround introduced in r1.31 to hpcmips.

Fix kernel crash when console is attached. Now, my MC/R550 boots multiuser.
Thanks Masahiko Ito for encouraging me to do bisection to find this out.
XXX pullup-8, pullup-7

Fix comment formatting in previous.

Clarify the comment in the "unconst" version of hpckbd_keymap_setup.
Both sh3 (P1) and mips (kseg0) have the kernel directly mapped.

(martin)

Pull up following revision(s) (requested by kre in ticket #1020):
usr.bin/printf/printf.c: revision 1.46

A truly ancient bug found by Edgar Fuss

When printf is running builtin in a sh, global vars aren't reset to
0 between invocations.  This affects "rval" which remembers state
from a previous %b \c and thereafter always exits after the first
format conversion, until we get a conversion that generates an
error (which resets the flag almost by accident)

printf %b abc\\c
abc (no \n)
printf %s%s hello world
hello (no \n, of course, no world ...)
printf %s%s hello world
hello
printf %s%s hello world
hello
printf %d hello
printf: hello: expected numeric value
0 (no \n)
printf %s%s hello world
helloworld (no \n, and we are back!)

This affects both /bin/sh and /bin/csh (and has for a very long time).

XXX pullup -8

(martin)

Fix for i386, functionality intended in:
http://mail-index.netbsd.org/source-changes/2018/09/23/msg099357.html

This should fix the build for both GENERIC and XEN3PAE_DOM0

This has not been boot tested on native or xen3pae

Notes: pmap_changeprot_local() seems to be x86_64 only.
I was a bit surprised by this initially, but I suspect that the table
protections are enforced via ring0/ring1 fencing rather than page protections

the gdt registration code in i386 is still messy. I will leave it as is
for now - to avoid a rabbit hole.

(cherry)

oops forgot to commit (add libcryptotest)

(christos)

remove references to isdn

(maxv)

note removal of isdn

(maxv)

merge openssl-1.1.1

(christos)

Merge openssl-1.1.1

(christos)

merge conflicts

(christos)

The kern_proc_32 hook is defined in non-compat code.  Don't redefine.

(pgoyette)

Use the proper test: _RUMPKERNEL vs _RUMP_KERNEL

(pgoyette)

This hook doesn't exist for _RUMPSERVER

(pgoyette)

The kern_proc_32 hook is in the main kernel, so don't make another
one here.

(pgoyette)

More work on kern_proc_32

(pgoyette)

Use restricted namespace identifiers for variables within the code
generated by macro expansion.

(pgoyette)

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.

(maxv)

White space for readability

(pgoyette)

Be selective about including the NTP syscall ntp_gettime

(pgoyette)

Remove the isdn sets from syspkg, now that they have no user left.

(maxv)

Revert previous.  Instead, don't put an entry for ntp_gettime in the
syscall package if we don't have NTP.

(pgoyette)

Provide a dummy version of netbsd32_ntp_gettime for systems where NTP
isn't defined.

(pgoyette)

Summary: Remove older and now conflicting type declaration for global
variable 'idt'.

This should fix the i386 build failure now.

(cherry)

Don't bother to include COMPAT_50 related headers, since all the COMPAT_50
code has been relocated.

(pgoyette)

Don't try to include usb_verbose.h - it might not exist on platforms
that don't have USB!

It turns out that nothing uses it any.

(pgoyette)

Revert previous - not appropriate

(pgoyette)

Put the USBVERBOSE option into a header file so it can be found by
modules.

(pgoyette)

Remove the userland part of ISDN. The kernel part is untouched for now.
ipppctl was actually an exact copy of pppoectl; there is no functional
change in pppoectl in this commit.

(maxv)

Use a space to separate fields, not _

(pgoyette)

Include header to get prototypes for the empty xxx_init() and xxx_fini()

(pgoyette)

Remove ISDN.

(rin)

No need to declare the crypto_50 init/fini routines.  The compat_50
module no longer needs them, and they're declared static in  the
ocryptodev.c code

(pgoyette)

Now that cryptodev_50 is a separate module we can't initialize it
from the generic compat_50 module.

(pgoyette)

Use the correct module name

(pgoyette)

Mark the correct module obsolete!

(pgoyette)

Fix some thinkos/pastos/typos

(pgoyette)

Revert:
http://mail-index.netbsd.org/source-changes/2018/09/10/msg098995.html

It is incorrect to infer semantics from usage.

the problem for which the original commit was intended should be fixed
within the callee intr_establish_xname() and not the caller:
pci_intr_find_intx_irq()

This was accomplished via:
http://mail-index.netbsd.org/source-changes/2018/09/20/msg099286.html

(cherry)

Encapsulate pre-processing of registered (pirq,evtchn) pair in
preparation for API reorg.

(cherry)

Switch from getopt to getopts - and as part of that, update the usage,
and the man page, to reflect modern usage rather than some ancient style
that getopt[s] does not support (options, and their args when they exist)
always come before non-option agrs.

While here, update uses of test(1) to avoid obsolescent forms (and
even when still defined, prefer sh structures over test when possible).

(kre)

Split the compat_crypto_50 from the rest of the crypto module

Cleanup some stuff left over from similar changes to raid modules.

(pgoyette)

Make XEN use the same api as native, for idt vector allocation
and registration.

lidt() placed in xenfunc() on maxv@ suggestion.

There should be no functional change due to this commit.

Tested on amd64 native and XEN.

(cherry)

Add dummy file since the compat_16 code assumes that there might be
some MD code here.

(pgoyette)

Split the raidframe compatability code out into separate modules, since
the driver is already a module.  This avoids having to pass a bunch of
function pointers as arguments to raidframe's compat_ioctl().

(pgoyette)