Report AUXUART fifo size on rpi3 to 1 byte.  This prevents bluetooth
errors using the h5 protocol.

(nat)

Add support for btuart h5 3-wire protocol.

(nat)

Save and restore xcr0 when doing ACPI sleeps. Should fix PR/49174.

(maxv)

Don't include opt_compat_freebsd.h.

(maxv)

Don't include opt_compat_ibcs2.h. No idea what it was doing in amd64,
since it never got implemented there.

(maxv)

Remove the svr4/ibcs2 fpu flags.

(maxv)

Write known issues on scalability

(ozaki-r)

Update

ipsec(4), opencrypto(9) and vlan(4) are now MP-safe.

(ozaki-r)

prune redundant line

(snj)

1467

(snj)

Apply patch (requested by mrg in ticket #1467):
Update tzdata to 2017b.

(snj)

1467

(snj)

Apply patch (requested by mrg in ticket #1467):
Update tzdata to 2017b.

(snj)

Apply patch (requested by mrg in ticket #1467):
Update tzdata to 2017b.

(snj)

Use pool_cache(9) instead of pool(9) for tdb_crypto objects

The change improves network throughput especially on multi-core systems.

(ozaki-r)

Add per-CPU rtcache to ipsec_reinject_ipstack

It reduces route lookups and also reduces rtcache lock contentions
when NET_MPSAFE is enabled.

(ozaki-r)

Simplify ipsec_reinject_ipstack (NFC)

(ozaki-r)

1467

(snj)

Apply patch (requested by mrg in ticket #1467):
Update tzdata to 2017b.

(snj)

Add support IP_PKTINFO for sendmsg(2).

The source address or output interface can be specified by adding IP_PKTINFO
to the control part of the message on a SOCK_DGRAM or SOCK_RAW socket.

Reviewed by ozaki-r@ and christos@. thanks.

(ryo)

evbmips can "boot root FFSv2"

It typically uses a non-NetBSD bootloader (PMON, u-boot...) and those
don't differentiate from v1 from v2 - both are unsupported and it requires
a separate boot partition.

(maya)

Remove references to svr4 and ibcs2, they are not supported here.

(maxv)

Remove several dead entries from the x86 makefiles. Looks like people (me
included) regularly forget to take care of this.

(maxv)

Remove ibcs2_machdep.h on i386, and don't install it.

(maxv)

Remove __i386__.

(maxv)

Remove __i386__.

(maxv)

Remove references to compat_ibcs2.

(maxv)

Remove compat_ibcs2 from i386. After a discussion on port-vax, it turns
out that compat_ibcs2 does not implement the iBCS2 standard - which is
x86-specific - but rather SVR3. Our real iBCS2 implementation was a
mixture of compat_ibcs2 and compat_svr4, and was only partial. Keeping
support for this in i386 is totally irrelevant today. I also asked on
port-i386 but didn't wait long.

The main issue is that compat_ibcs2 should have been called compat_svr3.
But CVS does not support renaming files, and moving things around is both
painful and tiring, even more so when no one seems to be interested in
doing this work or in the feature at all. For now compat_ibcs2 is available
on Vax and will stay, until someone (not me) cleans it up.

(maxv)

If a DF_1_NODELETE DSO has been removed from the DAG list, it still
needs to be able to resolve relocations against itself. As such, search
the referencing object explicitly last, if it hasn't been searched
before.

(joerg)

PR/52472: Edgar Fuss: Document handling of scoped IPv6 addresses by embedding
ASCII art from:
IPv6 Core Protocols Implementation
By Qing Li, Tatuya Jinmei, Keiichi Shima
Page 56, Figure 2.12

(christos)

While ld(4) is MP safe, many backends are not.

Add a flag for backends that are MP safe. Take KERNEL_LOCK when calling
into a backend that doesn't have the flag set. Do the same for the
discard routine.

Fixes PR 52462.

(mlelstv)

Don't wait for allocation in polling mode.

(mlelstv)

Ammend ticket #189

(martin)

Additionally pull up r1.30 for ticket #189 to unbreak the build.

(martin)

MP-ify SAD (savlist)

localcount(9) is used to protect savlist of sah. The basic design is
similar to MP-ifications of SPD and SAD sahlist. Please read the
locking notes of SAD for more details.

(ozaki-r)

Limit the number of mbufs queued for deferred key_sendup_mbuf

It's easy to be queued hundreds of mbufs on the list under heavy
network load.

(ozaki-r)

1460-1462

(snj)

1460-1462

(snj)

1460-1462, 1464-1466

(snj)

Pull up following revision(s) (requested by uwe in ticket #1466):
sys/arch/hpcsh/dev/pfckbd.c: revision 1.30
Don't reschedule the callout in the scanning function itself, as that
conflicts with using the same function for polled console.
Makes early ddb work with wscons console again on my Jornada 690.

(snj)

Pull up following revision(s) (requested by uwe in ticket #1465):
sys/dev/hpc/hpckbd.c: revision 1.32
XXX: On hpcsh only - bring back old incorrect hpckbd_keymap_setup()
fixed in 1.31.  The old one with UNCONST does work on hpcsh b/c the
kernel is directly mapped.  The new one does not work on hpcsh b/c it
calls consinit() very early when malloc() is not yet available.
The real fix for this is to fix the constness of wscons keymap
structures that is self-contradictory.

(snj)

Pull up following revision(s) (requested by jmcneill in ticket #1464):
sys/arch/evbarm/rpi/rpi_machdep.c: 1.71 via patch
Disable BSC0 on Raspberry Pi 3 and Zero W boards.

(snj)

Pull up following revision(s) (requested by skrll in ticket #1462):
sys/dev/usb/if_ural.c: revision 1.52
Free the RX list if ural_alloc_rx_list fails part way through.
Reported by Ilja Van Sprundel.

(snj)

Pull up following revision(s) (requested by skrll in ticket #1462):
sys/dev/usb/if_ural.c: revision 1.52
Free the RX list if ural_alloc_rx_list fails part way through.
Reported by Ilja Van Sprundel.

(snj)

Pull up following revision(s) (requested by skrll in ticket #1462):
sys/dev/usb/if_ural.c: revision 1.52
Free the RX list if ural_alloc_rx_list fails part way through.
Reported by Ilja Van Sprundel.

(snj)

Pull up following revision(s) (requested by martin in ticket #1461):
sys/kern/kern_malloc.c: revision 1.146
Avoid integer overflow in kern_malloc(). Reported by Ilja Van Sprundel.

(snj)

Pull up following revision(s) (requested by martin in ticket #1461):
sys/kern/kern_malloc.c: revision 1.146
Avoid integer overflow in kern_malloc(). Reported by Ilja Van Sprundel.

(snj)

Pull up following revision(s) (requested by martin in ticket #1461):
sys/kern/kern_malloc.c: revision 1.146
Avoid integer overflow in kern_malloc(). Reported by Ilja Van Sprundel.

(snj)

Pull up following revision(s) (requested by nat in ticket #1460):
sys/dev/auconv.c: revision 1.30 via patch
sys/dev/ic/ac97.c: revision 1.97 via patch
sys/dev/pci/azalia_codec.c: revision 1.81 via patch
Mixer device bounds checking.
Analysis by Ilja van Sprundel.

(snj)

Pull up following revision(s) (requested by nat in ticket #1460):
sys/dev/auconv.c: revision 1.30 via patch
sys/dev/ic/ac97.c: revision 1.97 via patch
sys/dev/pci/azalia_codec.c: revision 1.81 via patch
Mixer device bounds checking.
Analysis by Ilja van Sprundel.

(snj)

Pull up following revision(s) (requested by nat in ticket #1460):
sys/dev/auconv.c: revision 1.30 via patch
sys/dev/ic/ac97.c: revision 1.97 via patch
sys/dev/pci/azalia_codec.c: revision 1.81 via patch
Mixer device bounds checking.
Analysis by Ilja van Sprundel.

(snj)

Add counter check to vlan(4) ATF. Implemented by s-yamaguchi@IIJ, thanks.

(knakahara)

Fix vlan(4) obytes counter. Implemented by s-yamaguchi@IIJ, thanks.

(knakahara)

tickets 192-204

(snj)

Fix that prev isn't cleared on retry

(ozaki-r)

Apply patch (requested by mrg in ticket #204):
Don't define _DIAGNOSTIC in libc.  When this was introduced in revision
1.9, it was intended to be turned off for release branches.

(snj)

Pull up following revision(s) (requested by uwe in ticket #203):
sys/arch/hpcsh/dev/pfckbd.c: revision 1.30
Don't reschedule the callout in the scanning function itself, as that
conflicts with using the same function for polled console.
Makes early ddb work with wscons console again on my Jornada 690.

(snj)

Pull up following revision(s) (requested by uwe in ticket #202):
sys/dev/hpc/hpckbd.c: revision 1.32
XXX: On hpcsh only - bring back old incorrect hpckbd_keymap_setup()
fixed in 1.31.  The old one with UNCONST does work on hpcsh b/c the
kernel is directly mapped.  The new one does not work on hpcsh b/c it
calls consinit() very early when malloc() is not yet available.
The real fix for this is to fix the constness of wscons keymap
structures that is self-contradictory.

(snj)

Pull up following revision(s) (requested by knakahara in ticket #201):
sys/net/if_gif.c: revision 1.128
fix leak when encap_attach() fails twice.

(snj)

Pull up following revision(s) (requested by jmcneill in ticket #200):
sys/arch/arm/sunxi/files.sunxi: 1.15-1.16
sys/arch/arm/sunxi/sun6i_dma.c: 1.1-1.2
sys/arch/arm/sunxi/sun8i_h3_ccu.c: 1.9-1.10
sys/arch/arm/sunxi/sun8i_h3_codec.c: 1.1-1.2
sys/arch/arm/sunxi/sunxi_ccu.h: 1.8
sys/arch/arm/sunxi/sunxi_ccu_nkmp.c: 1.5
sys/arch/arm/sunxi/sunxi_codec.c: 1.1
sys/arch/arm/sunxi/sunxi_codec.h: 1.1
sys/arch/evbarm/conf/SUNXI: 1.20-1.21
Add DMA controller driver for sun6i and later family SoCs.
--
Enable sun6idma
--
add DMA gate
--
Fix burst field encoding, and add a helper function to dump registers
from ddb.
--
Add support for H3 audio PLL and digital audio part.
--
Add support for Allwinner H3 audio codec.
--
Enable H3 audio support
--
h3_codec_pr_write: clear write mode bit after setting it; fixes an issue with output being mutex when skipping tracks in mpg123

(snj)

Pull up following revision(s) (requested by kre in ticket #199):
bin/sh/input.c: revision 1.61
bin/sh/parser.c: revision 1.143
PR bin/52458
Avoid mangling history when editing is enabled, and the prompt contains a \n
Also, allow empty input lines into history when they are being appended to
a previous (partial) command (but not when they would just make an empty entry)
.
For all the gory details, see the PR.
Note nothing here actually makes prompts containing \n work correctly
when editing is enabled, that's a libedit issue, which will be addressed
some other time.

(snj)

Pull up following revision(s) (requested by spz in ticket #198):
sys/altq/altq_cbq.c: revision 1.31
sys/altq/altq_hfsc.c: revision 1.27
sys/altq/altq_jobs.c: revision 1.11
sys/altq/altq_priq.c: revision 1.24
sys/altq/altq_wfq.c: revision 1.22
Zero buffers copied to userland to avoid stack disclosure.
From Ilja Van Sprundel.
--
Reject negative indices.
(Would be nice to change the types too, and it's *probably* safe to
replace int by u_int, but I'm reluctant to touch the ioctl
definitions without at least a modicum more thought.  Also one of
them is a u_long, because why not?)
From Ilja Van Sprundel.

(snj)

Pull up following revision(s) (requested by spz in ticket #197):
sys/netsmb/smb_dev.c: revision 1.50
sys/netsmb/smb_subr.c: revision 1.38
sys/netsmb/smb_subr.h: revision 1.22
sys/netsmb/smb_usr.c: revision 1.17-1.19
Reject allocations for too-small buffers from userland.
From Ilja Van Sprundel.
--
Plug another overflow: refuse bogus sa_len from user.
--
Reject negative ioc_setupcnt.
--
Reject negative offset/count for smb read/write.
Not clear that this is actually a problem for the kernel -- might
overwrite user's buffers or return garbage to user, but that's their
own damn fault.  But it's hard to imagine that negative offset/count
ever makes sense, and I haven't ruled out a problem for the kernel.

(snj)

Pull up following revision(s) (requested by spz in ticket #196):
sys/dev/ic/ciss.c: revision 1.37
Reject negative indices from userland.

(snj)

Pull up following revision(s) (requested by spz in ticket #195):
sys/dev/ic/isp_netbsd.c: revision 1.89
Reject out-of-bounds channel index.
From Ilja Van Sprundel.

(snj)

Pull up following revision(s) (requested by spz in ticket #194):
sys/kern/kern_ktrace.c: revision 1.171
Clamp the length we use, not the length we don't.
Avoids uninitialized memory disclosure to userland.
From Ilja Van Sprundel.

(snj)

Pull up following revision(s) (requested by spz in ticket #193):
sys/compat/common/vfs_syscalls_12.c: revision 1.34
sys/rump/kern/lib/libsys_sunos/rump_sunos_compat.c: revision 1.2
sys/compat/svr4_32/svr4_32_misc.c: revision 1.78
sys/compat/sunos32/sunos32_misc.c: revision 1.78
sys/compat/linux/common/linux_misc.c: revision 1.239
sys/compat/osf1/osf1_file.c: revision 1.44
sys/compat/common/vfs_syscalls_43.c: revision 1.60
sys/compat/svr4/svr4_misc.c: revision 1.158
sys/compat/ibcs2/ibcs2_misc.c: revision 1.114
sys/compat/linux/common/linux_file64.c: revision 1.59
sys/compat/linux32/common/linux32_dirent.c: revision 1.18
sys/compat/sunos/sunos_misc.c: revision 1.171
Fail, don't panic, on bad dirents from file system.
Controllable via puffs from userland.
From Ilja Van Sprundel.

(snj)

Pull up following revision(s) (requested by spz in ticket #192):
sys/kern/vfs_getcwd.c: revision 1.52
Don't walk off the end of the dirent buffer.
From Ilja Van Sprundel.

(snj)

- Sprinkle static and const. No functional change.
- Print chip ID in hexadecimal instead of octal in def_match().

(msaitoh)

Fix deadlock between key_sendup_mbuf called from key_acquire and localcount_drain

If we call key_sendup_mbuf from key_acquire that is called on packet
processing, a deadlock can happen like this:
- At key_acquire, a reference to an SP (and an SA) is held
- key_sendup_mbuf will try to take key_so_mtx
- Some other thread may try to localcount_drain to the SP with
  holding key_so_mtx in say key_api_spdflush
- In this case localcount_drain never return because key_sendup_mbuf
  that has stuck on key_so_mtx never release a reference to the SP

Fix the deadlock by deferring key_sendup_mbuf to the timer
(key_timehandler).

(ozaki-r)

Fix locking notes of SAD

(ozaki-r)

Make use of the defined YP errors when ypbind encounters one rather than
just returning an empty (successful) RPC result.

Don't exit when allocating memory for a new bound domain received via a RPC.

(ginsbach)

If the harware ring is starved of data...Insert silence into the mix ring
so mix_write can function properly.

(nat)

Mmh, don't overwrite tf_err and tf_trapno. Looks like it can be used to
exploit the intel sysret vulnerability once again.

(maxv)

Remove dumb debug code and outdated comment.

(maxv)

Remove compat_svr4, compat_svr4_32 and compat_ibcs2 from the list of
autoloaded modules. These options are disabled everywhere (except ibcs2
on Vax, but Vax does not support kernel modules, so doesn't matter),
therefore there is no issue in removing them from the list. Interested
users will now have to do a 'modload' first, or uncomment the entries in
GENERIC.

(maxv)

Tickets #188 - #191

(martin)

Pull up following revision(s) (requested by spz in ticket #191):
sys/compat/ibcs2/ibcs2_ioctl.c: revision 1.46
sys/compat/ibcs2/ibcs2_stat.c: revision 1.50
sys/compat/ibcs2/ibcs2_exec_coff.c: revision 1.27
sys/compat/ibcs2/ibcs2_exec_coff.c: revision 1.28
sys/compat/ibcs2/ibcs2_exec_coff.c: revision 1.29
sys/compat/ibcs2/ibcs2_stat.c: revision 1.49
Check for NUL termination within the buffer we have.
From Ilja Van Sprundel.
Make sure we have enough space in the buffer before reading it.
From Ilja Van Sprundel.
Make sure we move forward over the buffer.
From Ilja Van Sprundel.
Zero buffers in ibcs2 ioctl to avoid disclosing stack to userland.
From Ilja Van Sprundel.
Don't drop vnode ref until we're done with mount in ibcs2_stat(v)fs.
Nothing else guarantees the mount will stick around.
From Ilja Van Sprundel.
Little happy on the commit trigger.  Actually use the out label.

(martin)

Pull up following revision(s) (requested by spz in ticket #190):
sys/dev/vnd.c: revision 1.260
sys/dev/vnd.c: revision 1.262
Put in a litany of judicious bounds checks around vnd headers.
Thought I was done with this crap after I rewrote vndcompress(1)!
From Ilja Van Sprundel.
Appease toxic bullshit warning from gcc.
If you have a better way to write a useful bounds check that happens
to always pass on LP64 but doesn't always on LP32, without making it
fail to compile on LP64 or making it an #ifdef conditional on LP32,
please put it in here instead.

(martin)

Pull up following revision(s) (requested by spz in ticket #189):
sys/compat/svr4/svr4_stream.c: revision 1.89
sys/compat/svr4/svr4_signal.c: revision 1.67
sys/compat/svr4/svr4_stream.c: revision 1.90
sys/compat/svr4/svr4_stream.c: revision 1.91
sys/compat/svr4_32/svr4_32_signal.c: revision 1.29
sys/compat/svr4/svr4_lwp.c: revision 1.20
Fix some of the multitudinous holes in svr4 streams.
We should never have enabled this by default; it is a minefield.
From Ilja Van Sprundel.
Zero stack data before copyout.
From Ilja Van Sprundel.
Fix indexing of svr4 signals.
From Ilja Van Sprundel.
Feebly attempt to get this reference counting less bad.
This svr4 streams code is bad and it should feel bad.
From Ilja Van Sprundel.
Check bounds in svr4_sys_putmsg.  Check more svr4_strmcmd bounds.
svr4 streams code is still a disaster.
From Ilja Van Sprundel.

(martin)

Pull up following revision(s) (requested by spz in ticket #188):
sys/compat/linux/common/linux_time.c: revision 1.38
sys/compat/linux/common/linux_time.c: revision 1.39
Only let the superuser set the compat_linux timezone.
Not really keen to invent a new kauth cookie for this useless purpose.
From Ilja Van Sprundel.
Put suser check in the right function: settimeofday, not gettimeofday.
While here, remove wrong comment.
Noted by kre@.

(martin)

Build the loongson kernels with all the binutils loongson2f errata workarounds

From Zhang Fuxin via https://sourceware.org/ml/binutils/2009-11/msg00387.html

  - The NOP issue

"The nature of the erratum is deeply related to the microarchitecture of
Loongson-2. It uses roughly a 4-way superscalar dynamically scheduled core,
instructions are excuted as much as possible in parallel with technics like
branch prediction etc. We use a 8-entry internal branch prediction queue to
keep track of each predicted branches, if some branches are proved to be
wrongly predicted, all the instructions following it will be cancelled,together
with the resources used by them, including the registers used for renaming, and
the queue entry will be freeed. There is a bug that might cause a hang when the
queue is full(some resources might been leaked due to conflict branch entries),
the workaround is to reduce the possiblity of branch queue full by using
renaming registers(they are also limited, can prevent too many simutaneos
branches). In theory this is still not enough to fully eliminate possible
hangs, but the possiblity is extremely low now and hard to be hit in real
code."

  - The JUMP instructions issue

"The Loongson-2 series processors have quite complex micro-architecture, it will
try to execute instructions from the predicated branch of coming instruction
stream before they are confirmed to be run, if the predication of branch
direction is proved wrong later, the instructions will be cancelled, but if the
instructions is a read from memory, the read action might not be cancelled(but
the changes to register will) to enable some prefetch. This will lead to some
problems when compining with some chipsets. E.g. the AMD CS5536 used in
Yeeloong/Fuloong will hang if it gets an address in the physical address range
of 0x100000-0x200000(might be more other ranges). Speculative reads can perform
read at any address in theory(due to wrong prediction of branch directions and
the use of branch target buffer), thus in very few occasions they might cause a
hard lock of the machine.

To prevent this, we need to prevent some addresses from entering branch
target buffers. A way to do this is that to modify all jump targets, e.g.,
    calulations of t9
    ...
    jalr t9  =>
    calculations of t9
    or t9, t9, 0x80000000;  // to make sure t9 is in kseg0
    jalr t9
Of course, we have to consider 64/32bit, and modules addresses etc.

This only need to be performed on kernel code, because only there we can have
accesses not translated/limited by TLB. For user code, it is impossible to
generate accesses to unwanted physical address. So it is safe.

Also, to prevent addresses generated by user mode code to be used by the
kernel, we add a few empty jumps to flush the BTB upon entrance to kernel."

(maya)

Ticket #1464

(martin)

Pull up following revision(s) (requested by maxv in ticket #1464):

sys/arch/i386/i386/trap.c:    revision 1.288 (patch)
sys/arch/i386/i386/machdep.c:  revision 1.783 (patch)
sys/arch/i386/i386/locore.S:  revision 1.146 (patch)
sys/arch/amd64/amd64/locore.S: revision 1.122,1.124 (patch)
sys/arch/amd64/amd64/machdep.c revision 1.254 (patch)
sys/arch/amd64/amd64/trap.c:  revision 1.95-1.96 (patch)

Remove the osyscall call gate and emulate it. There is a
one-instruction race in it that could panic the kernel.

Restore the ability to run netbsd 1.0 32-bit executables by checking
for the relevant lcall instruction in the trap handler and treating it
as a syscall.

(martin)

Ticket #1464

(martin)

Pull up following revision(s) (requested by maxv in ticket #1464):

sys/arch/i386/i386/trap.c:    revision 1.288 (patch)
sys/arch/i386/i386/machdep.c:  revision 1.783 (patch)
sys/arch/i386/i386/locore.S:  revision 1.146 (patch)
sys/arch/amd64/amd64/locore.S: revision 1.122,1.124 (patch)
sys/arch/amd64/amd64/machdep.c revision 1.254 (patch)
sys/arch/amd64/amd64/trap.c:  revision 1.95-1.96 (patch)

Remove the osyscall call gate and emulate it. There is a
one-instruction race in it that could panic the kernel.

Restore the ability to run netbsd 1.0 32-bit executables by checking
for the relevant lcall instruction in the trap handler and treating it
as a syscall.

(martin)

Ticket #1464

(martin)

Pull up following revision(s) (requested by maxv in ticket #1464):

sys/arch/i386/i386/trap.c:    revision 1.288 (patch)
sys/arch/i386/i386/machdep.c:  revision 1.783 (patch)
sys/arch/i386/i386/locore.S:  revision 1.146 (patch)
sys/arch/amd64/amd64/locore.S: revision 1.122,1.124 (patch)
sys/arch/amd64/amd64/machdep.c revision 1.254 (patch)
sys/arch/amd64/amd64/trap.c:  revision 1.95-1.96 (patch)

Remove the osyscall call gate and emulate it. There is a
one-instruction race in it that could panic the kernel.

Restore the ability to run netbsd 1.0 32-bit executables by checking
for the relevant lcall instruction in the trap handler and treating it
as a syscall.

(martin)

Fix KASSERT(solocked(sb->sb_so)) failure in sbappendaddr that is called eventually from key_sendup_mbuf

If key_sendup_mbuf isn't passed a socket, the assertion fails.
Originally in this case sb->sb_so was softnet_lock and callers
held softnet_lock so the assertion was magically satisfied.
Now sb->sb_so is key_so_mtx and also softnet_lock isn't always
held by callers so the assertion can fail.

Fix it by holding key_so_mtx if key_sendup_mbuf isn't passed a socket.

Reported by knakahara@
Tested by knakahara@ and ozaki-r@

(ozaki-r)

Remove whitespace I just introduced

(maya)

In working around loongson errata clear BTB and RAS, same as
other operating systems.

15 Errata: Issue of Out-of-order in loongson (translated)

In loongson 2F, because of the branch prediction, sometimes the CPU
may fetch the instructions from some unexpected area (for example I/O
space). It is an invalid operation. There are two ways for the CPU to
choose the branch target. The first one is predicting the branch
target according to the branch target history. The second one is
calculating the branch target by the ALU. There are most 8
instructions in the instruction window at the same time in loongson2f
(Remember the loongson 2f is superscalar, right?). Hence, the
branch target of an indirect branch(such as jr) could be got(may be
predicted by the branch target history) earlier and the instrctions of
the branch target could be prefetched even if there are branch
instructions before it. As a result, it is possible to fetch the
instructions from I/O region( say out-of the physical address range of
[0- 0x100000]) in kernel model because of the instruction prefetch of
the branch target.

There are some suggestions to prevent prefetching instructions from
the I/O region in kernel mode.

(1) When switching from user model to kernel model, you should flush
the branch target history such as BTB and RAS.
(2) Doing some tricks to the indirect branch target to make sure that
the indirect branch target can not be in the I/O region.

(maya)

Destroy sav only in the loop for DEAD sav

(ozaki-r)

Introduce key_sa_refcnt and replace sav->refcnt with it (NFC)

(ozaki-r)

Remove compat_freebsd from the list of autoloaded modules. Interested users
will now have to type 'modload' to use it, or uncomment the entry in
GENERIC. I should have removed it when I disabled COMPAT_FREEBSD by
default, sorry about that.

(maxv)

Move freebsd_machdep.h into sys/compat/freebsd, and don't install it. Now,
the compat_freebsd files are all contained in sys/compat/freebsd.

(maxv)

Remove dead codes. chan will never be NULL in SIMPLEQ_FOREACH.

(isaki)

Introduce audio_destroy_pfilters()/audio_destroy_rfilters()
and use it.

(isaki)

Add missing initialization of sc_rfilters in audioattach().

(isaki)

Remove a duplicated line.

(isaki)

MP-ify SAD (key_sad.sahlist and sah entries)

localcount(9) is used to protect key_sad.sahlist and sah entries
as well as SPD (and will be used for SAD sav).

Please read the locking notes of SAD for more details.

(ozaki-r)

s/tail queue/simple queue/ in SIMPLE QUEUES section.

(isaki)

fix leak when encap_attach() fails twice.

XXX need pullup to -8 branch

(knakahara)

Fix setkey -D -P outputs

The outputs were tweaked (by me), but I forgot updating libipsec
in my local ATF environment...

(ozaki-r)

Add missing mutex_exit

(ozaki-r)

Tweak wording (pserialize critical section => pserialize read section)

Suggested by riastradh@

(ozaki-r)

Add __read_mostly to key_psz

Suggested by riastradh@

(ozaki-r)

XXX: On hpcsh only - bring back old incorrect hpckbd_keymap_setup()
fixed in 1.31.  The old one with UNCONST does work on hpcsh b/c the
kernel is directly mapped.  The new one does not work on hpcsh b/c it
calls consinit() very early when malloc() is not yet available.

The real fix for this is to fix the constness of wscons keymap
structures that is self-contradictory.

(uwe)

Don't reschedule the callout in the scanning function itself, as that
conflicts with using the same function for polled console.

Makes early ddb work with wscons console again on my Jornada 690.

(uwe)

Use NULL instead of 0 for null pointer.

(uwe)

h3_codec_pr_write: clear write mode bit after setting it; fixes an issue with output being mutex when skipping tracks in mpg123

(jmcneill)

Fix GCC warning on NET4501, PR/52451.

(maxv)

Remove incorrect KASSERT, only the allocation is protected by cpu_lock.

(maxv)

Make it compile again when __STDC_ISO_10646__ is not defined.

(perseant)

Fix typo in comment.

(isaki)

Add support Microchip SST25VF016B.
- Fixed imxspi send and receive bugs.

(hkenken)

Clean up IOPORT settings.

(hkenken)

Hold key_sad.lock on SAVLIST_WRITER_INSERT_TAIL

(ozaki-r)

Tidy up ufs_readdir. First step only; there's plenty more that could be
done to improve this code.

(dholland)

don't build apm, ark or glint on cats for xorg-server 1.18.

(mrg)

Describe constraints of key_sp_ref and key_sp_unref

Requested by riastradh@

(ozaki-r)

Update the locking notes

- Add locking order
- Add locking notes for misc lists such as reglist
- Mention pserialize, key_sp_ref and key_sp_unref on SP operations

Requested by riastradh@

(ozaki-r)

Move locking notes

(ozaki-r)

Assemble global lists and related locks into cache lines (NFCI)

Also rename variable names from *tree to *list because they are
just lists, not trees.

Suggested by riastradh@

(ozaki-r)

Use KDASSERT instead of KASSERT for mutex_ownable

Because mutex_ownable is too heavy to run in a fast path
even for DIAGNOSTIC + LOCKDEBUG.

Suggested by riastradh@

(ozaki-r)

Remove out-of-date log output

Pointed out by riastradh@

(ozaki-r)

Enable H3 audio support

(jmcneill)

Add support for Allwinner H3 audio codec.

(jmcneill)

Add support for H3 audio PLL and digital audio part.

(jmcneill)

Fix burst field encoding, and add a helper function to dump registers
from ddb.

(jmcneill)

Turn on DIAGNOSTIC and DEBUG.

(christos)

Ensure to call mix write if inserting silence into the hw ring.

Addresses PR kern/52459.

Patch provided and tested by isaki@.

(nat)

use the same string for the log and uprintf.

(christos)

Mention high mem.

(maxv)

Mention PMCs.

(maxv)

Deprecate.

(maxv)

Regen from previous changes of usbdevs to fix kernel build

(ryoon)

Fix compile warnings.

(ginsbach)

Write the "warm start" file to /var/run rather than /tmp.

(ginsbach)

Sort; simplify; new sentence, new line; fix typo.

(wiz)

Fix punctuation markup.

(wiz)

Fix punctuation markup, remove superfluous Pp.

(wiz)

Remove superfluous Pp.

(wiz)

Use Fx/Nx.

(wiz)

fixes for xorg-server 1.18.

(mrg)

Enable sun6idma

(jmcneill)

Add DMA controller driver for sun6i and later family SoCs.

(jmcneill)

add DMA gate

(jmcneill)

Reset mixer ports on opening /dev/sound.

This fixes an issue reported on port-arm@ by Brian Buhrow.

(nat)

kill dead code

(jdolecek)

Automatically eject RALINK RT73 devices

This lets the MW-P54SS USB Wireless Broadband Router from Synet attach
as rum(4).

(khorben)

Sort the matching vendor list alphabetically

NFC.

(khorben)