Prevent signed integer left shift UB in FD_SET(), FD_CLR(), FD_ISSET()

Set the type of shifted integer 1 to unsigned int.

Detected with MKSANITIZER/UBSan in sysinst(8).

(kamil)

provide pci_intr_establish_xname() on x86 independantly from MSI,
so it's available on XEN too; change also the stub to use weak
symbol instead #ifdef

(jdolecek)

Welcome to 8.0 RC2!

(martin)

Fix a typo

(leot)

Tickets #898 and #899

(martin)

Pull up following revision(s) (requested by kamil in ticket #899):

lib/libutil/passwd.c: revision 1.53

Prevent underflow buffer read in trim_whitespace() in libutil/passwd.c

If a string is empty or contains only white characters, the algorithm of
removal of white characters at the end of the passed string will read
buffer at index -1 and keep iterating backward.

Detected with MKSANITIZER/ASan when executing passwd(1).

(martin)

Pull up following revision(s) (requested by kamil in ticket #898):

usr.sbin/installboot/arch/i386.c: revision 1.41

Fix integer overflow in installboot(8)
Add a sanity check of the disk_buf first three bytes. The original code on
a disk with nul bytes was causing integer overflow and thus calling the
memcmp(3) functin in is_zero() with enormous length.
Verity that the 0th byte is JMP, 1th a signed byte >= 9 to prevent overflow
and 2th byte NOP.

Add a comment explaining the check.

Detected with MKSANITIZER and ASan.

(martin)

Fix stack use after scope in libutil/pty

The pt variable's elements are used after the end of the pt scope.
A move of pt to outer scope fixes this.

Detected with MKSANITIZER/ASan with tmux(1), a forkpty(3) user.

(kamil)

Install note changes

(martin)

Pull up the following changes with minor adjustement:

distrib/notes/acorn26/contents (same as acorn32 1.3)
distrib/notes/acorn32/contents 1.3
distrib/notes/alpha/contents 1.17
distrib/notes/amd64/contents 1.6
distrib/notes/amiga/contents 1.21
distrib/notes/arc/contents 1.3
distrib/notes/atari/contents 1.21
distrib/notes/bebox/contents 1.3
distrib/notes/cats/contents 1.3
distrib/notes/common/contents 1.173
distrib/notes/common/xfer 1.75
distrib/notes/emips/contents 1.3
distrib/notes/evbarm/contents 1.3
distrib/notes/evbppc/contents 1.4
distrib/notes/ews4800mips/contents 1.3
distrib/notes/hp300/contents 1.18
distrib/notes/hpcarm/contents 1.4
distrib/notes/hpcmips/contents 1.11
distrib/notes/hpcsh/contents 1.4
distrib/notes/hppa/contents 1.3
distrib/notes/i386/contents 1.30
distrib/notes/landisk/contents 1.4
distrib/notes/mac68k/contents 1.21
distrib/notes/macppc/contents 1.13
distrib/notes/mmeye/contents 1.4
distrib/notes/mvme68k/contents 1.14
distrib/notes/news68k/contents 1.7
distrib/notes/newsmips/contents 1.3
distrib/notes/next68k/contents 1.8
distrib/notes/ofppc/contents 1.4
distrib/notes/pmax/contents 1.18
distrib/notes/prep/contents 1.3
distrib/notes/rs6000/contents 1.3
distrib/notes/sandpoint/contents 1.3
distrib/notes/sgimips/contents 1.3
distrib/notes/shark/contents 1.3
distrib/notes/sparc/contents 1.22
distrib/notes/sparc64/contents 1.6
distrib/notes/sun2/contents 1.3
distrib/notes/sun3/contents 1.17
distrib/notes/vax/contents 1.12
distrib/notes/x68k/contents 1.18

Document debug sets and kernel module directory.

(martin)

Document debug sets and kernel module directory.

(martin)

Add temporary hint at manual installation instructions for UEFI setups.

(martin)

Prevent underflow buffer read in trim_whitespace() in libutil/passwd.c

If a string is empty or contains only white characters, the algorithm of
removal of white characters at the end of the passed string will read
buffer at index -1 and keep iterating backward.

Detected with MKSANITIZER/ASan when executing passwd(1).

(kamil)

Enlarge the set_status[] array by a single element

In the get_and_unpack_sets() function there is accessed the
set_status[SET_GROUP_END] element in the array. The array is allocated on
the stack with SET_GROUP_END elements. This means that it is 1 element too
short.

This has been reported with MKSANITIZER=yes with Address Sanitizer.

(kamil)

Specify -Wno-format-extra-args for Clang/LLVM in gpl2/gettext

This is needed with Clang v. 7svn (HEAD) 2018-06-23 snapshot.

(kamil)

Add MDC, MDI, MDI-X, MDIO

(sevan)

remove core list, portmaster list, releng list, and developer list.

...as proposed several times in several places back in 2009/2010.

this info can all be found on the website (where it's more accurate) and
has no place in a document describing the installation of netbsd.

(snj)

Disable all contemporary mode 1 quirks.

(jakllsch)

If mode 1 enable check fails, give mode 1 a second chance by trying to
use it to locate a PCI Host Bridge or device from vendor that produced
a chipset lacking a Host Bridge class device.

Should allow us to remove most all the mode 1 quirks added in the last
two decades.

(jakllsch)

make compile without DDB

PR port-xen/50282

(jdolecek)

The s in AST is system.
Add RC and common use of IPL, previous entry was from the world of s/360

(sevan)

add a kobj_error() to a recently added error case

(jakllsch)

Fix integer overflow in installboot(8)

Add a sanity check of the disk_buf first three bytes. The original code on
a disk with nul bytes was causing integer overflow and thus calling the
memcmp(3) functin in is_zero() with enormous length.

Verity that the 0th byte is JMP, 1th a signed byte >=9 to prevent overflow
and 2th byte NOP.

Add a comment explaining the check.

Detected with MKSANITIZER and ASan.

(kamil)

Add acpiecdt* at acpi?.

(jakllsch)

Fix current version of tzcode for ticket #809.

(martin)

remove the xen XSAVE entry, needs more work before it can be enabled

(jdolecek)

Tickets #891, #892, #893, #894, #895, #896 and #897.

(martin)

Pull up the following, via patch, requested by maxv in ticket #897:

sys/arch/amd64/amd64/locore.S          1.166 (patch)
sys/arch/i386/i386/locore.S            1.157 (patch)
sys/arch/x86/include/cpu.h              1.92 (patch)
sys/arch/x86/include/fpu.h              1.9 (patch)
sys/arch/x86/x86/fpu.c                  1.33-1.39 (patch)
sys/arch/x86/x86/identcpu.c            1.72 (patch)
sys/arch/x86/x86/vm_machdep.c          1.34 (patch)
sys/arch/x86/x86/x86_machdep.c          1.116,1.117 (patch)

Support eager fpu switch, to work around INTEL-SA-00145.
Provide a sysctl machdep.fpu_eager, which gets automatically
initialized to 1 on affected CPUs.

(martin)

Note import of dhcpcd-7.0.6.

(martin)

Pull up the following, requested by roy in ticket #896:

external/bsd/dhcpcd/dist/src/common.c          up to 1.1.1.5
external/bsd/dhcpcd/dist/src/defs.h            up to 1.1.1.14
external/bsd/dhcpcd/dist/src/dhcpcd.8.in        up to 1.1.1.10
external/bsd/dhcpcd/dist/src/dhcpcd.c          up to 1.13
external/bsd/dhcpcd/dist/src/dhcpcd.conf.5.in  up to 1.1.1.10
external/bsd/dhcpcd/dist/src/duid.c            up to 1.1.1.3

Import dhcpcd 7.0.6.

(martin)

fix intrctl_io_firstline() to properly return NULL if there are no records
to show

(jdolecek)

Pull up following revision(s) (requested by kamil in ticket #895):

usr.sbin/sysinst/util.c: revision 1.9

Fix invalid free(3) in sysinst(8)

The path variable is assigned with an allocation on the heap with
strdup(3). Later this pointer is changed with strsep(3) and this caused
invalid free(3).

Store the original pointer in a new helper variable opath and pass it to
free(3). With this change, the problem is going away.

Detected with MKSANITIZER=yes with AddressSanitizer.

(martin)

Pull up following revision(s) (requested by pgoyette in ticket #894):

sys/dev/ata/ata_raid.c: revision 1.40

Call config_cfattach_attach() regardless of whether the module is
being built as part of a kernel or as a external module.

Addresses kern/53389 - thanks for the report!

XXX pullup-8

(martin)

Pull up following revision(s) (requested by maxv in ticket #893):

sys/netinet6/icmp6.c: revision 1.228,1.230

Remove the RH0 code from ICMPv6. RH0 is deprecated by RFC5095 (2007) for
security reasons. We already removed it in Route6.

In addition there was an mbuf bug here: calling IP6_EXTHDR_GET twice with
the same offset, but still using the pointer from the first call, which
could have been made invalid. By luck, m_pulldown leaves zero-sized mbufs
in place, instead of freeing them.

And in general, using a 'finaldst' pointer on the mbuf, and then modifying
that mbuf with IP6_EXTHDR_GET with a smaller offset, was really error-
prone.

Fix 'icmp6len', it shouldn't be ip6_plen, because we may not be at the
beginning of the packet (off+ip6_plen is beyond the end of the mbuf). By
luck, the IP6_EXTHDR_GET that follows will fail and prevent buffer
overflows in non-jumbogram packets.

For jumbograms we will probably be in trouble here; but it doesn't seem
possible to craft reliably a jumbogram for a non-jumbogram-enabled device.

So I don't think it's a huge problem.

(martin)

Pull up following revision(s) (requested by maya in ticket #892):

lib/libc/sys/sched.c: revision 1.5

PR/52826: Onno van der Linden: Return PRI_NONE for sched_get_priority_m{in,ax}
and SCHED_OTHER.

(martin)

Apply patch, requested by christos in ticket #891:

Provide future compatibility API when OPENSSL_API_COMPAT is
defined to >= 0x10100000L.

(martin)

Ticket #1617

(martin)

Pull up following revision(s) (requested by maya in ticket #1617):

external/gpl3/gcc/dist/libstdc++-v3/include/std/complex: revision 1.2

PR/50646: Use the c99 cabs math builtins directly, because cabs and cabsf
have broken ABI's can cabsl does not exist. The correct(?) probably fix is
to change the cabs builtins to point to the c99 variants directly...

XXX: pullup-7 and the same file from the gcc.old tree.

(martin)

re-do the XEN XSAVE support, this time to leave all probe code in
cpu_probe_fpu(), and have XEN cpu_init() just act

the xen probe is now guarded by XEN_USE_XSAVE option and XSAVE
support is thus still disabled by default (same as before), so it
wouldn't interfere with maxv's eager fpu rototil, while still
allowing testing for others

PR kern/50332

(jdolecek)

Add XXX in fpuinit_mxcsr_mask.

(maxv)

Reorder the code a little. On Xen, return earlier, we don't need to do
the XSAVE-related initialization if we don't support XSAVE.

(maxv)

Revert the rest of jdolecek's changes. This puts us back in a clean,
sensical state.

(maxv)

No semicolon after macro do ... while (0) wrapper.

(gson)

No semicolon after macro do ... while (0) wrapper.

(gson)

constify

(maxv)

constify

(maxv)

constify

(maxv)

Add some debugging in case someone else wants to debug gdb...

(christos)

Add all encoding supported by the mixer.

Addresses PR kern/52585.

(nat)

Fix thread debugging.

(christos)

locore.S is a MD_SFILES.

This keeps the dependency handling in the loop, so rebuilds after
changing options, say EARLYCONS, don't fail.

(jakllsch)

Remove commented out MATH_EMULATE from kernel configs that got it
by copying an older i386 config.

(maya)

Match generic arm,arm-v8 compatible string

(jmcneill)

Handle more Vortex CPU's from Andrius V.
While here refactor the code to make it smaller.

(christos)

fix free() on unitialized value for list -c when there are no
assigned interrupts

(jdolecek)

Explain the -Bstatic -Bdynamic dance.

(christos)

match our number of registers.

(christos)

use the right emulation

(christos)

Describe something for npf, npfd, blacklistd.

(maya)

Remove documentation references to rtsol.

(maya)

rtsol(8) and rtsold(8) have been long removed.

(roy)

When processing character classes ([:xxx:] inside []), treat a class name
that is longer than we can handle the same way we treat an unknown
class name (as a valid char class which contains nothing, so never
matches).  Previously a "too long" class name invalidated the
class, so [:very-long-name:] would match any of  '[' ':' 'v'  ...
(note: "very-long-name" is not long enough to trigger this, but you
get the idea!)

However, the name itself has a restricted syntax ([[:***:]] is not a
character class, it is a match for one of a '[' ':' or '*', followed by
a ']') which we did not implement - check the syntax of the name before
treating it as a character class (but we do add '_' to alphanumerics
as legal class name characters).

(kre)

Tickets #887, #888, #889

(martin)

Pull up following revision(s) (requested by maxv in ticket #889):

sys/netinet6/ip6_output.c: revision 1.205
sys/netipsec/xform_ah.c: revision 1.90,1.93,1.102,1.103

Simplify the IPv4 parser. Get the option length in 'optlen', and sanitize
it earlier. A new check is added (off + optlen > skip).

In the IPv6 parser we reuse 'optlen', and remove 'ad' as a result.

Remove the kernel RH0 code. RH0 is deprecated by RFC5095, for security
reasons. RH0 was already removed in the kernel's input path, but some
parts were still present in the output path: they are now removed.
Sent on tech-net@ a few days ago.

Fix non-INET6 builds

Strengthen and simplify, once more.

(martin)

- Fix groff reproducibility:
    - specify mdoc and font paths
    - clear encoding
    - specify pagesize on the command line
XXX: pullup-8

(christos)

Don't look in random (legacy) places for fonts. We want reproducible builds
to work.
XXX: pullup-8

(christos)

When matching a char class ([[:name:]]) in a pattern (for filename
expansion, case patterrns, etc) do not force '[' to be a member of
every class.

Before this fix, try:
case [ in [[:alpha:]]) echo Huh\?;; esac

XXX pullup-8    (Perhaps -7 as well, though that shell version has
much more relevant bugs than this one.)  This bug is not in -6 as
that has no charclass support.

(kre)

iic_use_direct_match(): when iic_compatible_match() does not find a
comptible entry, set match quality to 0. Otherwise callers might use
random stack garbage.

(martin)

Add a kernel panic when we matched but can not handle a string - better
die explicitly instead of crashing a few lines later with strange NULL
derefs.

(martin)

Add comment describing why we remove -Wsystem-headers, requested by mrg.

(maya)

Deal with ref after free found by ASAN when a function redefines
itself, or some other function which is still active.
This was a long known bug (fixed ages ago in the FreeBSD sh) which
hadn't been fixed as in practice, the situation that causes the
problem simply doesn't arise .. ASAN found it in the sh dotcmd
tests which do have this odd "feature" in the way they are written
(but where it never caused a problem, as the tests are so simple
that no mem is ever allocated between when the old version of the
function was deleted, and when it finished executing, so its code
all remained intact, despite having been freed.)

The fix is taken from the FreeBSD sh.

XXX -- pullup-8 (after a while to ensure no other problems arise).

(kre)

If max-frequency is not specified, use UINT_MAX instead of the rate programmed by the bootloader.

(jmcneill)

Pull up following revision(s) (requested by jakllsch in ticket #888):

sys/dev/pci/ld_virtio.c: revision 1.19-1.21,1.23

remove trailing whitespace

add feature/register definitions from virtio-v1.0-cs04

unload payload dma map upon command completion

Make ld_virtio aware of a possible device-side write cache.

The virtio block device capacity config item is expressed always in
request protocol sector units of 512 bytes.

Also, add and use a symbolic constant to refer to request protocol
sector units.

destroy 'sc_sync_wait' condvar and mutex upon detach

(martin)

Pull up following revision(s) (requested by christos in ticket #887):

bin/ed/main.c: revision 1.30

Prevent shell execution also in the 'r ! dobad' case, pointed out by
Martijn van Duren, thanks!

XXX: pullup-8

(martin)

Remove bpf_tap() in iexmit() because ie_start() does bpf_mtap().

(msaitoh)

Remove bpf_tap() call in rx. It's done in if_percpuq_softint().

(msaitoh)

Call config_cfattach_attach() regardless of whether the module is
being built as part of a kernel or as a external module.

Addresses kern/53389 - thanks for the report!

XXX pullup-8

(pgoyette)

Revert jdolecek's changes related to FXSAVE. They just didn't make any
sense and were trying to hide a real bug, which is, that there is for some
reason a wrong stack alignment that causes FXSAVE to fault in
fpuinit_mxcsr_mask. As seen in current-users@ yesterday, rdi % 16 = 8. And
as seen several months ago, as well.

The rest of the changes in XSAVE are wrong too, but I'll let him fix these
ones.

(maxv)

Simplify like other drivers. NULL check of ifp->if_bpf is done in
bpf_mtap(), so it's not required to do it here.

(msaitoh)

It's not required to include net/bpfdesc.h. Remove it.

(msaitoh)

PR port-xen/50446
With approval from bouyer@ and acceptance, sometimes grudgingly, from
others I have removed the AGP lines from the XEN3_DOM0 config.  I have
left a note in the PR suggesting that if it ever gets a proper fix that
those lines can be considered for re-inclusion.

(darcy)

Amend MIPS, there are multiple stages in a pipeline.

(maya)

Fix invalid free(3) in sysinst(8)

The path variable is assigned with an allocation on the heap with
strdup(3). Later this pointer is changed with strsep(3) and this caused
invalid free(3).

Store the original pointer in a new helper variable opath and pass it to
free(3). With this change, the problem is going away.

Detected with MKSANITIZER=yes with AddressSanitizer.

(kamil)

Add AFP, AMD, ASA, BSDP, DEC, IPL, MDMX, MIPS, PIX, PRU (found on the BeagleBone
Black), XHR.

(sevan)

split the port status reporting to new function sata_interpret_det()
so it can be called separately from sata_reset_interface()

do not treat PHY offline as an error, it's pretty normal when there
is no device actually connected

debugging aid for PR kern/52372

(jdolecek)

remove unused arguments

(maxv)

Fix use-after-free, m_cat can free m.

(maxv)

destroy 'sc_sync_wait' condvar and mutex upon detach

(jakllsch)

Note import of dhcpcd-7.0.6

(roy)

Sync

(roy)

* make to work printf() and panic() even before consinit().
* tidy up output for VERBOSE_INIT_ARM.

(ryo)

Specify NOSANITIZER in distrib/amd64/ramdisks/common

This option cannot be set in distrib/common/Makefile.distrib as it's too
late, after including <bsd.own.mk> that will set MKSANITIZER.

This makes generation of ramdisk for amd64 with MKSANITIZER=yes successful.

(kamil)

Refactor MKSANITIZER flags in mk rules

Introduce a new variable SANITIZERFLAGS that contains all the sanitizer
specific flags. These flags are designed to be passed to CFLAGS, CXXFLAGS
and LDFLAGS.

Pass SANITIZERFLAGS in bsd.prog.mk rules to CFLAGS, CXXFLAGS and LDFLAGS.

Pass SANITIZERFLAGS in bsd.lib.mk rules to *_pic.a build flags. The *_pic.a
libraries are designed to construct .so libraries. All .so libraries with
few exceptions are expected to be prebuilt with a sanitizer. Another
purpuse of *_pic.a files can be constructing static PIE programs, however
they are not supported as of now and there is no sanitization of static
programs either.

Once MKSTATICPIE will be supported, it will be initially incompatible with
MKSANITIZER.

This change also makes it easier to build other targets like kernel and
kernel modules without userland sanitizers.

No functional change intended for MKSANITIZER=no users.

(kamil)

Make building of dhcp compatible with MKSANITIZER

Disable LD flags (-Wl,-Bstatic and -Wl,-Bdynamic) with enabled MKSANITIZER.
These options are incompatible with the current design of sanitizers,
because they cause duplication of symbols into programs and thus symbols
from the interceptors from sanitizers cannot be linked.

This change makes effectively mounting /usr required for dhcp programs like
dhclient(8).

(kamil)

Add new option -s to crunchgen(1) -- enable sanitization

As of today typical sanitizers require dynamic executables, while
crunchgen(1) programs are produced with static properties.

Lack of specified -s will:
- generate a Makefile file with NOSANITIZER=
- build programs that are dependencies with NOSANITIZER=

In future there is an option to handle sanitization in statically linked
programs.

An idea with -s LGTM by <christos>

(kamil)

Enable DWC3 USB DRD (host mode) on Rock64.

(jmcneill)

Add support for RK3328

(jmcneill)

Set NOSANITIZER in rescue/Makefile

This is required to stop sanitizing rcmd.o and liboverride.o. These object
files will be linked into statically linked rescue.

(kamil)

Disable SANITIZER for ldd(1)

These utilities (elf32, elf32_compat, elf64, liblldb) share code with the
ELF dynamic loader that is not being sanitized and its symbols are
installed into sanitized programs (in particular __tls_get_addr()).

Additionally libldd is used in rescue that is not expected to be sanitized
as of today.

(kamil)

sbappendaddr() is required any lock. Currently, softnet_lock is appropriate.

When rip_input() is called as inetsw[].pr_input, rip_iput() is always called
with holding softnet_lock, that is, in case of !defined(NET_MPSAFE) it is
acquired in ipintr(), otherwise(defined(NET_MPSAFE)) it is acquire in
PR_WRAP_INPUT macro.
However, some function calls rip_input() directly without holding softnet_lock.
That causes assertion failure in sbappendaddr().
rip6_input() and icmp6_rip6_input() are also required softnet_lock for the same
reason.

(knakahara)

Simplify & white space fix. No functional change.

(msaitoh)

An acronym entry needs to be all uppercase.

Thanks to Thorsten Glaser

(sevan)

as a stop-gap, make fpuinit_mxcsr_mask() for native independant of
XSAVE as it should be, only xen case checks the flag now; need to
investigate further why exactly the fault happens for the xen
no-xsave case

pointed out by maxv

(jdolecek)

Add missing return in error path

(jmcneill)

Revert previous.

(thorpej)

Note release note changes

(martin)

Update for NetBSD-8

(martin)

Use PMAP_DIRECT_UNMAP.

(maxv)

Add and use bootspace.smodule. Initialize it in locore/prekern to better
hide the specifics from the "upper" layers. This allows for greater
flexibility.

(maxv)

Put these arrays in .rodata, they aren't supposed to be executable.

(maxv)

Regen.

(msaitoh)

Add nSYS Technologies.

(msaitoh)

Add l2cc support.

(hkenken)

Move intr_establish() before imxuart_attach_subr().

(hkenken)

Modified CCM register defines.

(hkenken)

Rename a9tmr to arma9tmr.

(hkenken)

In fdtbus_print(), aprint_normal the path to the device (rather than
aprint_debug).  This info is every bit as useful as, say, PCI device
locations.

(thorpej)

Fix typo.

(hkenken)

Support Cortex-A9 (addr_d > addr_c).

(hkenken)

Style fix. No functional change.

(msaitoh)

Use mpcaa_off1 parameter for mapping subregion.

(hkenken)

Whitespace fix. No functional change.

(msaitoh)

Add test case for PR lib/50646.

Make sure that cabsl (aka __c99_cabsl) is usable from C++.

(maya)

Strip -Wsystem-headers from CXXFLAGS.

GCC's C++ headers are not clean (yet). They are trying, but haven't got there
yet.

Necessary for including <atf-c++.hpp> which uses <sstream>.

(maya)

Disable threshhold mode for TX/RX DMA and ignore tx_delay/rx_delay props
for now. Fixes GMAC on RK3328.

(jmcneill)

Make the error message from pms_alps_probe_init() an error again now
that we don't misidentify almost anything as an ALPS device.  Reset if
E6 check failed as well.

(uwe)

Don't pass response buffer to the reset command since we are not
interested in it (pckbport_poll_cmd() is smart enough).

(uwe)

If the first byte of E7 signature is not 0x73, it's not an ALPS
device, so don't complain we failed to initialize it.  Still need a
reset, b/c the E7 check will be interpreted as normal commands by a
normal device.

(uwe)

dwcmmc: set ciu clock rate rather than assuming fixed input rate

(jmcneill)

When E6 signature is checked, ignore pressed buttons.

(uwe)

Use PMS_SEND_DEV_STATUS for E9 instead of PMS_GET_SCALE (also defined
as E9).  This is more readable and is congruent with other drivers.
Same object code is generated.

(uwe)

actually Xen XSAVE support itself is bigger deal than the AVX alone, so mention
that more prominently

(jdolecek)

note the Xen AVX fix

(jdolecek)

fix FPU initialization on Xen to allow e.g. AVX when supported by hardware;
only use XSAVE when the the CPUID OSXSAVE bit is set, as this seems to be
reliable indication

tested with Xen 4.2.6 DOM0/DOMU on Intel CPU, without and with no-xsave flag,
so should work also on those AMD CPUs, which have XSAVE disabled by default;
also tested with Xen DOM0 4.8.3

fixes PR kern/50332 by Torbjorn Granlund; sorry it took three years to address

XXX pullup netbsd-8

(jdolecek)

Install RK3328 .dtb files to the correct location

(jmcneill)

Increase reserved space at start of image from 4MB to 16MB to make room
for Rockchip bootloaders.

(jmcneill)

Use symbolic names for chip revision. NFC.

(jmcneill)

When using EagerFPU, create the fpu state in execve at IPL_HIGH.

A preemption could occur in the middle, and we don't want that to happen,
because the context switch would use the partially-constructed fpu state.

The procedure becomes:

splhigh
unbusy the current cpu's fpu
create a new fpu state in memory
install the state on the current cpu's fpu
splx

Disabling preemption also ensures that x86_fpu_eager doesn't change in
the middle.

In LazyFPU mode we drop IPL_HIGH right away.

Add more KASSERTs.

(maxv)

No semicolon after macro do ... while (0) wrapper.

(gson)

Explicitly clear l2's pcb_fpcpu when forking.

A context switch (preemption) could occur between

fpusave_lwp(l1, true);
and
memcpy(pcb2, pcb1, sizeof(struct pcb));

In this case, l1's FPU state is re-installed on the current CPU, and
pcb1->pcb_fpcpu becomes non NULL. While it's fine to have l1's state
installed, we don't want to indicate l2's state is installed too.

With lazy fpu this was not a problem, because the context-switch would
not re-install the state, so pcb1->pcb_fpcpu was NULL.

Should fix PR/53383.

(maxv)

Regen.

(msaitoh)

All of 88E151[0248]'s model number is 0x001d.

(msaitoh)

Sigh, fix another stupid mistake in previous that squeaked by because,
again, I booted the wrong test kernel.

(thorpej)

Use the device_compatible_entry mechanism rather than of_compat_data;
all of the OF / FDT data we need is already in the i2c_attach_args.

(thorpej)

rk_cru_composite_set_rate: allow selection of parent clocks in different
domains

(jmcneill)

Write MAC address high register before low register. Apparently the
hardware updates the filter when the low register is written.

(jmcneill)

The whole point of my recent addition of the specificdata(9) man page
was so it could be cross-referenced here.  So, add the xref.

(pgoyette)

Clear IFF_RUNNING | IFF_OACTIVE when stopping interface.

(jmcneill)

Add more KASSERTs, see if they help PR/53383.

(maxv)

Keep things portable (requested by joerg) by not depending on reallocarr
and instead doing the overflow check ourselves.

(christos)

- Rename iic_compat_match() to iic_compatible_match() and change it
  to use the new device_compatible_match() routine.  A pointer to
  the matching device_compatible_entry is returned if a match is
  found.
- Adjust iic_use_direct_match() accordingly.
- i2c drivers now provide device_compatible_entry tables when performing
  direct-config matching.
- In the dsrtc driver, take advantage of this new capability to greatly
  simplify model selection.

(I'm coming for you next, of_compat_data...)

(thorpej)

Fix a silly mistake in device_compatible_entry_matches() that I made
while re-factoring this from a prior version.

(I booted the wrong kernel when testing, oops.)

(thorpej)

Tickets #882, #883, #884, #885 and #886

(martin)