Remove end of line spaces - one (two in one line) added during recent merge,
one older.

(kre)

Fix merge botch

(kre)

Fix merge botch

(kre)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

Fix merge botch

(kre)

Restore comment accidentally lost in merge.

(dholland)

Fix merge botch

(kre)

Fix merge botch.

(kre)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

Fix merge botch

(kre)

fix duplicated chunk from merge

(dholland)

Fix merge botch

(kre)

Restore accidentally-dropped opt_octeon.h.
(or at least it really looks like it was accidental)

(dholland)

Fix merge botch

(kre)

Fix merge botch.

(kre)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

restore accidentally-removed rcsid

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

fix duplicated chunk from merge

(dholland)

Fix merge issue. (leftover line)

(oster)

Remove extra #ifdef and duplicated set of #ifdef's.

(oster)

Remove extra #endif.

(oster)

Remove extra #endif.

(alnsn)

make this compile again (enable #if 0 code)

(christos)

remove extra ifdef.

(christos)

remove extra #endif

(christos)

remove dup

(christos)

lint understands attributes.

(christos)

fix another mergeo

(dholland)

fix mergeo

(dholland)

Update version on the branch, too

(pgoyette)

Belated bump of kernel version, for merge of [pgoyette-compat]

(pgoyette)

Merge the [pgoyette-compat] branch

(pgoyette)

cast to intmax_t instead of long, since time_t is "long long"

(christos)

apparently the fdwatch crap is needed, so undo the ifdefs.

(christos)

graft back the fdwatch crap that dhcpd needs.

(christos)

Sync with HEAD

(pgoyette)

Regen.

(msaitoh)

Add some RADEON devices. Mainly taken from OpenBSD.

(msaitoh)

Explicitly cast time_t to match format string - should fix the build on
some 32bit architectures.

(martin)

Some arm CPUs do not implement traps on floating point exceptions.

(martin)

aarch64 does not trap on unaligned acces

(martin)

Optimize: keep a per-VCPU buffer for the state, and copy in and out
directly on it. The VCPUs are protected by mutexes, so nothing to worry
about.

This saves two kmem_allocs in {get,set}state.

(maxv)

Always convert input numbers (from the command line) in the C
locale, not as set in the environment.  Conforms with POSIX spec.

(kre)

While cute, the previous version is not really safe.
After all, a system might want to sleep for several
thousand years on a spaceship headed to a distant
solar system...

So, remove the pause() code, deal with limits on the
range (it is just an int) that can be passed to sleep()
by looping, and do a much better job of checking for
out of range input values.

With this change sleep(1) should work for durations
up to something more than 250 billion years.  It
fails (at startup, with an error) if the requested
duration is beyond what can be handled.

Here no changes at all related to locales and arg
parsing.    Still for another day.

(kre)

Adjust the way the arg string is parsed in the "not entirely
integer" case, so we avoid adjusting the locale of sleep,
and generally be more reliable and simpler.

In addition, deal with weirdness in nanosleep() when the
interval gets long, by avoiding using it.  In this version
when the sleep interval < 10000 seconds, we use nanosleep()
as before, for delays longer than that we use sleep() instead,
and ignore any fractional seconds.

We avoid overflow problems here by not bothering to sleep
at all for delays longer than 135 years (approx) and simply
pause() instead.  That sleep never terminates in such a
case is unlikely to ever be observed.

This commit makes no decision on the question of whether
the arg should be interpreted in the locale of the user,
or always in the C locale.  That is for another day.

(kre)

Remove nvmm_exit_memory.npc, useless.

(maxv)

Ah, fix bug: when the opcode has an immediate, we fill the src with a
register storage, but then we overwrite it without zeroing out the highest
bits of the resulting immediate (which may contain garbage from the union).

(maxv)

Define constants for representing the standard interrupt types
({pos,neg,double}-edge, {high,low}-level) from the FDT "interrupts"
bindings.  Use these defined constants rather than magic numbers.

(thorpej)

Define constants for representing the standard interrupt types
({pos,neg,double}-edge, {high,low}-level) from the FDT "interrupts"
bindings.  Use these defined constants rather than magic numbers.

(thorpej)

Add a note concerning MPSAFE

(pgoyette)

Note the future need to sort out placement of the netbsd32_machine32_hook

(pgoyette)

PR/53908: Alex Raschi: One more file needs to move.

(christos)

PR/53908: Alex Raschi: Test that require modules belong in modules; move the
threadpool test from kernel to modules.

(christos)

Make sure ci_{smt,core,package}_id are all filled in appropriately.

Previously the BP was missed and __SHIFTOUT wasn't used.

(skrll)

Explicitly include the headers in the right sequence to make the magic happen,
explaining why.

(christos)

Use _XTYPEDEF_POINTER directly to determine if the typedef is available.

(christos)

no need for curses.h

(christos)

PR/53906: Paul Ripke: systat(1) vmstat & syscall get
"alternate system clock has died" on terminal resize (SIGWINCH)

(christos)

PR/53908: Alex Raschi: Include <sys/socket.h> for socketpair(2)

(christos)

Revert previous change(s).  There's no good reason to use the hook when
the required value is available locally (in the current module).

Thanks mrg@ for pointing this out.

(pgoyette)

Move the netbsd32_machine32_hook into the main kernel with most of
the other hooks.

Although this hook might better belong in compat/netbsd32/ code,
there are some machines without a netbsd32 module (for example, i386
and sgimips) which still have consumers/users of this hook.  :(

(pgoyette)

Declare the variable which we're using as the return value of the
hook function.

(pgoyette)

Add __cacheline_aligned to ppp softc list and its mutex just in case.

(knakahara)

80003's SERDES is not the same as 82575's but the same as legacy devices.
Use the old methods on 80003.

XXX The reason why this bug existed is that our order of WM_T_* was little
different from FreeBSD's enum e1000_mac_type. From 80003 to PCH_CNP and from
82575 to I211 are swapped. I carefully read through if_wm.c and knew only
serdes related test were wrong. When we change WM_T_* to match FreeBSD's,
additional changes also will be required.

(msaitoh)

Use __nothing (== ((void)0)) instead of "do {} while (/*CONSTCOND*/false)".

(msaitoh)

remove the new check for tlen >= PTRDIFF_MAX.  the prior
checks assure the condition is met and GCC7 is happy.

suggested by kre.

(mrg)

set md_unwind_header=aarch64/aarch64-unwind.h for netbsd arm64.

(mrg)

adjust libstdc++ headers for gcc7.

(mrg)

PR/53904: Jintao Zhu: Use a mutex instead of an rwlock to assure thread safety

(christos)

Ticket #1168

(martin)

Pull up following revision(s) (requested by maxv in ticket #1168):

sys/kern/kern_ntptime.c: revision 1.60

Zero out the ntptimeval structure to prevent a 4 byte kernel stack disclosure.

Reported by Thomas Barabosch.

(martin)

Tickets #1674 and #1675

(martin)

Pull up following revision(s) (requested by maxv in ticket #1675):

sys/kern/kern_ntptime.c: revision 1.60

Zero out the ntptimeval structure to prevent a 4 byte kernel stack disclosure.
Reported by Thomas Barabosch.

(martin)

Tickets #1674 and #1675

(martin)

Pull up following revision(s) (requested by maxv in ticket #1675):

sys/kern/kern_ntptime.c: revision 1.60

Zero out the ntptimeval structure to prevent a 4 byte kernel stack disclosure.
Reported by Thomas Barabosch.

(martin)

Tickets #1674 and #1675

(martin)

Pull up following revision(s) (requested by maxv in ticket #1675):

sys/kern/kern_ntptime.c: revision 1.60

Zero out the ntptimeval structure to prevent a 4 byte kernel stack disclosure.
Reported by Thomas Barabosch.

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #1674):

regress/sys/arch/m68k/060sp/Makefile: revision 1.15

Replace manual realdepend dependency with DPSRCS.

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #1674):

regress/sys/arch/m68k/060sp/Makefile: revision 1.15

Replace manual realdepend dependency with DPSRCS.

(martin)

Pull up following revision(s) (requested by msaitoh in ticket #1674):

regress/sys/arch/m68k/060sp/Makefile: revision 1.15

Replace manual realdepend dependency with DPSRCS.

(martin)

Ticket #1167

(martin)

Pull up following revision(s) (requested by christos in ticket #1167):

usr.sbin/rpcbind/util.c: revision 1.23

- fix wrong size allocation that triggers buffer overflow
- remove unneeded casts and assertions

XXX: pullup-8

(martin)

Optimize: change the behavior of the HLT vmexit, make it a "change in vcpu
state" which occurs after the instruction executed, rather than an
instruction intercept which occurs before. Disable the shadow and the intr
window in kernel mode, and advance the RIP, so that the virtualizer doesn't
have to do it itself. This saves two syscalls and one VMCB cache flush.

Provide npc for other instruction intercepts, in case someone is
interested.

(maxv)

Add comments about D_MPSAFE to functions called as struct linesw.l_ioctl.

(knakahara)

Add KERNEL_LOCK in ppptioctl() to protect struct ppp_softc members.

struct linesw.i_ioctl can be called without any preservation when the caller's
struct cdevsw is set D_MPSAFE such as ucom(4).

(knakahara)

mknative gcc7 for ppc and arm64.

(mrg)

Restore the MD-vs-REQ requirements structure, just in case

(pgoyette)

Clean up and reorganize the requirements for compat_linux32 module

(pgoyette)

Move the definition around to avoid strange compiler error (something
about trying to set a section attribute on a local variable).

(pgoyette)

revert previous, just define the hook here

(pgoyette)

No functional change intended:
- Use "do {} while (/*CONSTCOND*/false)" for null DPRINTF().
- Reduce indent level of wm_linkintr_gmii().

(msaitoh)

And include the definition

(pgoyette)

Make sure we declare the hook variable

(pgoyette)

I must be getting tired - use the correct name of the hook

(pgoyette)

I wrote this stuff, when will I ever learn to use MODULE_CALL_HOOK
instead of MODULE_HOOK_CALL ?  :)

(pgoyette)

remove obsolete rcsid

(lukem)

sync to NetBSD glob.c 1.38

Update from NetBSD src/lib/libc/gen/glob.c 1.34 to 1.38:
- Minimize changes from upstream NetBSD source.
- 1.35: Use unsigned foo not u_FOO. [Already done]
- 1.36: Bump the glob limits to 512KB for total string size and 64K path
  entries. The old limits were too small for some important FTP use cases
  like a pkgsrc repository.
- 1.37: Switch from a recursive pattern matching algorithm to handle '*'
  to a backtracking one. Avoids DoS attacks with patterns "a*a*a*a*a*...b"
  matching against "aaaaaaaaaaaa..." https://research.swtch.com/glob
- 1.38: Use the symbolic M_ALL and trim with M_MASK.

(lukem)

Use the hook to get the value of machine32

(pgoyette)

The event_channel_op hypercall uses a newer API since
__XEN_INTERFACE_VERSION__  0x00030202

Since hvm_op only supports event_channel_op via the newer API, we
can't get away with our current event_channel_op_compat shim.

We thus introduce the new API to our internal hypercall C API
interface.

This change should have no effect on the PV kernels, since they will
continue to use the pre 0x00030202 API.

(cherry)

Use correct macro name

(pgoyette)

hvm_op returns a signed value.

The pattern is that a hypercall which returns a value < 0 may imply an
error.

(cherry)

Rework placement of the new netbsd32_machine32_hook

(pgoyette)

Found a couple other consumers of machine32 - adapt them to use the hook

(pgoyette)

Replace weak symbol definition for machine32 with a hook.  Our in-kernel
linker doesn't handle weak symbols, so this prevented us from loading the
compat_linux32 module.

XXX There don't seem to be any other consumers of machine32 (nor does
XXX there seem to be any consumers of machine_arch32), even though it
XXX is defined for aarch64, arm32, and riscv).

(pgoyette)

fnmatch.c: minimise changes from NetBSD src

(lukem)

sync to fts.c 1.49, fts.h 1.19

Update from NetBSD src/lib/libc/gen/fts.c 1.38 to 1.49.
Update from NetBSD src/include/fts.h 1.18 to 1.19.

Notable changes:
- Minimize changes from upstream NetBSD methods.
- Bump fts_level from short to int to avoid a potential DoS.
- Make sure fts_statp is NULL when FTS_NOSTAT is set.
- Use [non-standard] O_CLOEXEC instead of fcntl(FD_CLOEXEC).
- Avoid potential integer overflow.
- Fix double free in fts_read()/fts_close().
- Use calloc(1, ...) instead of malloc(...) & memset.

(lukem)

Need sys/module.h here

(pgoyette)

The raidframe compat code lives in sys/dev/raidframe/ so move all the
module stuff there.  Update the Makefiles appropriately.

(pgoyette)

Reduce the list of required modules to what is really needed

(pgoyette)

Disable more COMPAT_xx stuff for the NOCOMPAT config
cvs: ----------------------------------------------------------------------

(pgoyette)

get rid of homegrown pci_conf_hook(), use the mechanism from arm32_pci_chipset
instead
less difference to evbarm

(macallan)

- fix wrong size allocation that triggers buffer overflow
- remove unneeded casts and assertions
XXX: pullup-8

(christos)

Another missing semicolon.  grrrrr.

(pgoyette)

It's section 3 page, not section 2.  While here, xref puffs(3) in the
SEE ALSO section too.

(uwe)

skip clflush on aarch64, too

(jmcneill)

There are aarch64 Cortex cpus that do not trap on some floating point
exceptions - so apply the arm handling for aarch64 as well.

(martin)

Avoid dup inconsistent definitions.

(christos)

Some acronyms related to SIM cards; PR 53901 from Nikolai Lifanov.

(dholland)

Don't call OS_ReadSysInfo Platform class call on RISC OS 3

(sborrill)

Add xosmodule_enumeratewithversion() to allow operating systems modules
to be enumerated along with their version numbers.

(sborrill)

Change the place of initproc initialization

The initproc variable cannot be initialized in start_init as there
is a race between vfs_mountroot and start_init.

PR kern/53817 by Andreas Gustafsson

(kamil)

Fix no options DEBUG nor DIAGNOSTIC build

(ryoon)

- Move the location of calling setup_eee() in ixgbe_attach(). The previous
  location was too early. The new location is before calling
  ixgbe_set_phy_power().
- Restore call of setup_eee() in ixgbe_init_locked(). It was removed in
  rev. 1.99 (FreeBSD r320688).
- When calling setup_eee(), pass enable/disable correctly.

(msaitoh)

Ooops, missed a semicolon - add it back

(pgoyette)

Call ixgbe_init_swfw_semaphore(hw) before calling ixgbe_check_wol_support()
because ixgbe_check_wol_support()->ixgbe_check_wol_support()->
ixgbe_get_device_caps() accesses semaphore.

(msaitoh)

Fix mvfr0.fptrap = 0 description

(skrll)

Fix build break (return type of mii_writereg).

(msaitoh)

Welcome to 8.99.31 (MII {read,write}reg API change).

(msaitoh)

provide compat _DIAGASSERT()

define no-op _DIAGASSERT() in tnftpd.h, so that I can perform
less changes from "upstream" NetBSD when I merge code.

(lukem)

Nowadays some UEFI BIOSes don't enable some PCI devices' address decoding.
To resolve this problem, pci_map.c rev. 1.34-1.36 changed the
pci_mapreg_(sub)map()'s to set the decode bit if it's not set. It's good for
almost all drivers, but some other drivers don't use pci_mapreg_map().
In drivers which don't use pci_mapreg_map(), some of them expilicitly enable
decoding but others don't. Add code to enable decoding to them.

See also the following discussion:
http://mail-index.netbsd.org/tech-kern/2017/03/22/msg021678.html

(msaitoh)

KNF. No functional change.

(msaitoh)

sync to fparseln.c 1.10

Update from NetBSD src/lib/libc/stdio/fparseln.c 1.5 to 1.10:
- Don't report spurious empty lines eg after 2 comment lines, or on
  EOF after a single comment line.
- No escape character means no escaped characters.
- Remove 3rd and 4th clauses in christos' license.  OK christos.

(lukem)

Remove the 3rd clause from my BSD-derived license.

(lukem)

sync to fnmatch.c 1.26, fnmatch.h

Update from NetBSD src/lib/libc/gen/fnmatch.c 1.18 to 1.26.
Update from NetBSD src/include/fnmatch.h 1.8 to 1.13.

Notable changes:
- Implement FNM_LEADING_DIR.
- Limit recursions to avoid DoS attacks. From Maksymilian Arciemowicz.
- Treat a backslash followed by NUL as NUL.

(lukem)

Implement subroutines for parsing out some of the generic properties
specified in the pinctrl bindings, and adapt Meson, Rockchip, and
Allwinner pinctrl back-ends to use them.

Ok jmcneill@

(thorpej)

Implement subroutines for parsing out some of the generic properties
specified in the pinctrl bindings, and adapt Meson, Rockchip, and
Allwinner pinctrl back-ends to use them.

Ok jmcneill@

(thorpej)

Improve a comment

(pgoyette)

Make evbmips/conf/WGT624V3 compilable:
- Modify for new read/write API
- Do "#if 0" for unused structures.

(msaitoh)

Convert the macros for setting and unsetting a hook to generate
in-line code rather than using an intermediary hook##set routine.
Hooks are set and unset only in one place, so the intermediary
routine provides no benefit.  IMHO using the macro at the point-
of-call is more readable than using it elsewhere in the code and
then calling the generated intermediary routine (for which you
won't even find its declaration or definition unless you remember
to search for the HOOK_SET macro instead).

NFC intended, will verify with a bulk build and an atf test run.

(pgoyette)

Add compat_raid_80 to the required list

(pgoyette)

And as long as we're attempting to achieve perfection in code
that is never going to be executed, let's also check for possible
overflow in a sum that will never be computed...

(kre)

whitespace fixup

consistent tab indenting, not a mix of spaces and tabs

(lukem)

sync to daemon.c 1.10

Update to NetBSD src/lib/libc/gen/daemon.c 1.10 from 1.9:
- Update old-style definitions to ANSI

(lukem)

sync to arpa/ftp.h 1.7

Update to NetBSD src/include/arpa/ftp.h revision 1.7 from revision 1.5:
- 1.6: Move UCB-licensed code from 4-clause to 3-clause licence. [already done]
- 1.7: Welcome to the constant age!

(lukem)

There's no point setting errno, it is just overwritten by err
in the exit path ... this function never fails, it simply sometimes
doesn't actually expand the %m and just leaves the format string
intact.

And declare variables at the head of their scope, not at some random
place in the middle of the code, whatever C allows, that is just ugly.

(kre)

Revert previous.  Both opencrypto/ocryptodev and raidframe/rf_compat_xx
already exist in their own modules and do not need to be part of the
compat_{50,80} modules.

(pgoyette)

Ensure that all the compat code gets into the appropriate modules, even
compat code that doesn't live in sys/compat

(pgoyette)

limit allocation to PTRDIFF_T to appease gcc-7, from mrg@

(christos)

memset to zero the entire array, not just 8 bytes of it.

picked up by gcc7's checker that the array size has not
been multipled by the member size.

here, we had 8 but should have had 8 * sizeof(int).

XXX: real bug - pullup-7, pullup-8

(mrg)

remove diag assert - the compiler demands fp is non NULL now.

(mrg)

- fix Makefile.gthr CONFIG_LINKS handling.  demand only that
  gthr-default.h source be found.  fix installing the symlink.
- -Wno-error=incompatible-pointer-types needed for libgomp/lock.c
- build new libsupc++ files with -std=gnu++1z
- install new libstdc++ bits files. (XXX use bits_sup_headers
  from libstdc++/include/Makefile.am with mknative-gccm)
- start to get libstdc++ to build: various flag updates.

(mrg)