Make setsoftast take a cpu_info *

(matt)

remove various HAVE_GCC=45 fragments.

(mrg)

Add separate IPI routines for IPI_AST and IPI_KPREEMPT.

(matt)

define __HAVE_PREEMPTION

(matt)

Put the lock error info in the panicstr so ddb can get at it.

(riastradh)

Release the glock on VOP_GETPAGES failure.

Tripped over by nick@'s failing disk, missing unlock in error branch
discovered by jmcneill@.

(riastradh)

Remove FAST_FPE code

(matt)

KASSERT x then y, not x && y, to give more specific errors.

(riastradh)

add L2 cache write eviction buffer sync barrier

(hsuenaga)

clean up cpufuncs of CPU_PJ4B.

PJ4B is a ARMv7 compatible CPU, so most of cpufuncs are just redundant.
we need funcs for:
  - Marvell specific registers
  - workaround of errata
  - and Marvell specific L2 cache maintainance
if I/O coherency fabric is enabled(option AURORA_IO_CACHE_COHERENCY),
probaly we don't need to maintain L2 cache by software.

(hsuenaga)

implement L2 cache maintenance operations of ARMADA XP.
the L2 cahce maintenance operations are defined on SoC internal registers.

(hsuenaga)

lookup clock frequencies of ARMADA 370 correctly.

(hsuenaga)

add u-boot argument parser for Marvell ARMADA XP/370.
use 'ethaddr' and 'eth1addr' in u-boot argument to setup MAC address of mvgbe.

(hsuenaga)

Use LIST_FOREACH_SAFE

We have to use LIST_FOREACH_SAFE because LIST_REMOVE is used
inside the loop through encap_remove.

(ozaki-r)

On big endian machines needs to include sys/endian.h for le32dec

Mainly affects cross builds on big endian; tested on FreeBSD mips
and Linux ppc.

(justin)

Register iwm(4) with pmf(9).

Patch from Pierre Pronchery.

(nonaka)

ticket 832

(snj)

Pull up following revision(s) (requested by riastradh in ticket #832):
crypto/external/bsd/openssl/lib/libcrypto/evp.inc: revision 1.5
crypto/external/bsd/openssl/lib/libcrypto/modes.inc: revision 1.3
crypto/external/bsd/openssl/lib/libcrypto/arch/i386/aes.inc: revision 1.3
Fix build goo for OpenSSL AES-NI support.
OpenSSL now supports AES-NI in evp, not in an engine.  We can now get
rid of the no longer maintained aesni engine, which was broken last
summer.  Not only can OpenSSL now use AES-NI for everything it did
before we broke it last summer, but it can also use AES-NI for more
encryption modes than before, such as CTR.
Tested on amd64, both vanilla and in an i386 chroot.
ok christos

(snj)

Replace DIAGNOSTIC & panic with KASSERT/KASSERTMSG

(ozaki-r)

Add $NetBSD$ at the top of the file

(ozaki-r)

Coverity CID 1264915, Via FreeBSD (Xin Li)

When reading in the original file name from gzip header, we read
in PATH_MAX + 1 bytes from the file.  In r281500, strrchr() is
used to strip possible path portion of the file name to mitigate
a possible attack.  Unfortunately, strrchr() expects a buffer
that is NUL-terminated, and since we are processing potentially
untrusted data, we can not assert that be always true.

Solve this by reading in one less byte (now PATH_MAX) and
explicitly terminate the buffer after the read size with NUL.

(christos)

__HAVE_PREEEMPTION -> __HAVE_PREEMPTION

(jmcneill)

Forgot commit: Move another <sys/rndsource.h> to where it belongs.

(riastradh)

Include <sys/rndsource.h> where it is actually used.

I had removed <sys/rnd.h> from files that didn't mention anything of
the rnd(9) API.  But they included other files which assumed
<sys/rnd.h> had already been included.

(riastradh)

Note DMA support for sdhc on beaglebone

(bouyer)

Add the edma controller, and enable DMA for sdhc0 and sdhc1

(bouyer)

Add support for DMA transfers. From Jared D. McNeill, with final debug by me.
With this I can get nearly 20MB/s from my sdcard on the BB black at 1Ghz
(not bad for a 50Mhz 4-bits bus), and still 15MB/s on the BB white at
low speed (275Mhz).

(bouyer)

Ops, remove empty files that should not be there.

(bouyer)

Add a driver for the Enhanced Direct Memory Access controller found
in the AM335x SoC. Written by Jared D. McNeill, with some final debug
by me.
Supports only DMA (not QDMA) yet, and there's no support for DMA event
matrix yet (this means that only primary DMA events can be used)

(bouyer)

Add a new SDHC_FLAG, SDHC_FLAG_EXTDMA_DMAEN, which request that the
SDHC_DMA_ENABLE bit be set in the command, even if we're using an
external DMA engine. Needed by the upcoming DMA support for AM335x
(beaglebone).

(bouyer)

Ticket 1279.

(msaitoh)

Pull up following revision(s) (requested by snj in ticket #1279):
distrib/sets/makesrctars: revision 1.40
GPL code isn't just in src/gnu these days (i.e., the last 6 years).
move stuff from src/external/gpl[23] out of the src tarball and into
gnusrc.

(msaitoh)

Ticket 1279.

(msaitoh)

Pull up following revision(s) (requested by snj in ticket #1279):
distrib/sets/makesrctars: revision 1.40
GPL code isn't just in src/gnu these days (i.e., the last 6 years).
move stuff from src/external/gpl[23] out of the src tarball and into
gnusrc.

(msaitoh)

Summarize ticket 1279.

(msaitoh)

Ticket 1264, 1272-1726, 1728-1279.

(msaitoh)

Pull up following revision(s) (requested by snj in ticket #1279):
distrib/sets/makesrctars: revision 1.40
GPL code isn't just in src/gnu these days (i.e., the last 6 years).
move stuff from src/external/gpl[23] out of the src tarball and into
gnusrc.

(msaitoh)

Pull up following revision(s) (requested by khorben in ticket #1264):

include/unistd.h 1.144
sys/kern/init_sysent.c regen
sys/kern/syscalls.c regen
sys/kern/syscalls.master 1.272
sys/rump/include/rump/rump_syscalls.h regen
sys/rump/librump/rumpkern/rump_syscalls.c regen
sys/sys/syscall.h regen
sys/sys/syscallargs.h regen
sys/compat/netbds32/netbsd32_syscall.h regen
sys/compat/netbds32/netbsd32_syscallargs.h regen
sys/compat/netbds32/netbsd32_syscalls.c regen
sys/compat/netbds32/netbsd32_sysent.c regen
sys/compat/netbsd32/syscalls.master 1.102

Fix return type of the readlinkat(2) syscall. PR#49684.

(msaitoh)

Pull up following revision(s) (requested by bouyer in ticket #1278):
sys/arch/xen/include/evtchn.h: revision 1.23
sys/arch/xen/xen/evtchn.c: revision 1.71
sys/arch/xen/xen/pci_intr_machdep.c: revision 1.17
Properly implemement pci_intr_disestablish(9), so that interrupt
handlers stop being called when the device has been detached.
Should fix PR port-xen/47720 (which turns out to not be related to raidframe).
While there fix possible races in event_remove_handler() and pirq_establish().

(msaitoh)

mention yacc requirement for acpica

(christos)

Omit needless variable.

(riastradh)

Omit needless rnd_wakeup_readers.

(riastradh)

Centralize bookkeeping of rnd_initial_entropy.

It is not adjusted everywhere it should be, although the only places
it perhaps formally ought to be are places with questionable entropy
estimates in the first place.

(riastradh)

use the newer skeleton file.

(christos)

- remove malloc/realloc casts.
- use YYINT instead of short.

(christos)

Make the softint cookies static.

(riastradh)

Pull up following revision(s) (requested by riastradh in ticket #1276):
usr.bin/ktrace/ktrace.1: revision 1.43
Omit no longer existing l, m from default trace points.

(msaitoh)

Pull up following revision(s) (requested by riastradh in ticket #1275):
share/man/man9/uvm_km.9: revision 1.4
Clarify uvm_km_alloc is contiguous and zero result means fail.

(msaitoh)

Pull up following revision(s) (requested by riastradh in ticket #1274):
doc/BUILDING.mdoc: revision 1.111
Use `It Sy', not just `It', for live-image, to match other targets.
No change to plain-text BUILDING file.

(msaitoh)

Fix rump build: rndpseudo_50.c now needed by kernel, not rnd device.

(riastradh)

Pull up following revision(s) (requested by riastradh in ticket #1273):
lib/libc/stdio/fopen.3: revision 1.30
Use Pq to avoid space before O_EXCL.

(msaitoh)

Pull up following revision(s) (requested by riastradh in ticket #1272):
share/man/man9/pci_intr.9: revision 1.18 via patch
Fix type of pc argument in pci_intr(9) man page.

(msaitoh)

Group initialization of rnd_samples and rnd_global.

(riastradh)

Gather global pool and sources state into a cacheline-aligned struct.

(riastradh)

Note about PR#49864.

(msaitoh)

Omit unused rndpool_cv.

(riastradh)

Move RND_EXTRACT_{ANY,GOOD} to rndpool.h whose API they are part of.

(riastradh)

Make various things now private to kern_rndq.c static.

(riastradh)

Make rnd_getmore acquire rndpool_mtx itself.

(riastradh)

#ifdef RND_VERBOSE printf -> rnd_printf_verbose

(riastradh)

Sort includes.

(riastradh)

Move substantive part of rnd_ioctl to kern_rndq.c.

(riastradh)

Omit unused rndsource `anonymous'.

(riastradh)

Remove null pointer dereference in the code that never worked before...

(riastradh)

Eliminate remaining cases of u_int*_t in kern_rndq.c.

(riastradh)

Eliminate last two cases of u_int*_t in rndpseudo.c.

(riastradh)

Use binuptime, not microtime/nanotime as substitute cycle counter.

(riastradh)

Include rndpool.h, rndsource.h here because we use them.

(riastradh)

No need for <sys/rnd.h> here either.  Missed this one yesterday too.

(riastradh)

moar because moar

(rodent)

Need only <sys/rndsource.h>, not <sys/rnd.h>.

Made this change yesterday, failed to commit it -- sorry!

(riastradh)

Use rnd_add_data, not rndpool_mtx and rndpool_add_data.

(riastradh)

Fix comment.

(riastradh)

Minor grammatical update.  (I wonder if wizd(8) thinks I need to update
the revision date?)

(pgoyette)

Fix a bug that ifconfig -z (SOICZIFDATA) doesn't work.

(msaitoh)

Sync ixg(4) up to FreeBSD r243716:
- A lot of bugfixes. Some of them are realted to multi queue and those
  have not affected in NetBSD because we have not used it yet.
- Show 1000Base-SX correctly.
- Fix if_baudrate from 1G to 10G.
- Improve performance.

(msaitoh)

Include rndio.h for missing defines.

(nat)

Include sys/rndsource.h for rnd_add_data.

(nat)

Fix typo expliict -> explicit.

(nat)

tickets 675-681, 683-692

(snj)

Pull up following revision(s) (requested by christos in ticket #679):
lib/libedit/chartype.c: revisions 1.11, 1.12
lib/libedit/chartype.h: revisions 1.12, 1.13
PR/49683: Amir Plivatsky: Off-by-one comparison in ct_decode_string() leading
to out of bounds referrence.
--
split the allocation functions, their mixed usage was too confusing.

(snj)

Pull up following revision(s) (requested by christos in ticket #692):
sbin/ping/ping.c: revisions 1.108, 1.109
PR/49423: Martin Husemann: ping for small packets does not work in -7 or
-current
--
Adjust default packet size to 56 data bytes (64 total).
Make error messages consistent.

(snj)

Pull up following revision(s) (requested by christos in ticket #691):
lib/libc/gen/opendir.c: revision 1.39
- Use O_DIRECTORY to open the file, so that we don't need to stat() after
  that.
- Move the stat() call to fdopendir() and change it's error handling so that
  it does not hide errors.
- According to POSIX, fdopendir() transfers ownership of the fd only on
  success, so don't close it on failure. XXX: We still make it non-blocking
  on failure, but that's nitpicking.

(snj)

Pull up following revision(s) (requested by christos in ticket #690):
external/bsd/nvi/dist/ex/ex_mkexrc.c: revision 1.4
exrc expects 1 argument.

(snj)

Pull up following revision(s) (requested by christos in ticket #689):
sys/netinet/in.c: revision 1.149
Don't pass junk in sin_family and sin_len for SIOCGIFNETMASK, and explain why.
XXX: pullup 7?

(snj)

Pull up following revision(s) (requested by christos in ticket #688):
sbin/fsck_ext2fs/setup.c: revision 1.32
Instead of zerodivide, give a useful error message.

(snj)

Pull up following revision(s) (requested by christos in ticket #687):
usr.sbin/nfsd/nfsd.c: revision 1.63
Memset the whole cfg array, not piecemeal because otherwise you'll end
up with random data depending on the options set.

(snj)

Pull up following revision(s) (requested by christos in ticket #686):
sys/kern/kern_exec.c: revision 1.411
sys/sys/exec.h: revision 1.147
PR/49287: Masao Uebayashi: Handle exec_script argument vector from the 32 ->
64 bit case. When execing a 64 bit shell from a 32 bit binary the argument
vector was still incorrect.

(snj)

Pull up following revision(s) (requested by christos in ticket #685):
usr.bin/ftp/ftp_var.h: revision 1.83
Increase the buffer limit; otherwise files in:
    http://www.taxdetective.ca/Samples/sampledatafiles.html
fail.

(snj)

Pull up following revision(s) (requested by christos in ticket #684):
usr.sbin/makefs/ffs.c: revision 1.64
PR/49559: Christian Brueffer: Fix typo maxbpf instead of maxbpg.

(snj)

Pull up following revision(s) (requested by christos in ticket #683):
sys/dev/dkwedge/dkwedge_apple.c: revision 1.2
Look at the bzb flags to mark partition as swap. Requested by John D. Baker.
XXX: Pullup-7

(snj)

Pull up following revision(s) (requested by christos in ticket #681):
external/mit/xorg-server/dist/include/regionstr.h: revisions 1.3, 1.4
fix signed/unsigned comparison.
--
fix previous.

(snj)

Pull up following revision(s) (requested by christos in ticket #680):
external/bsd/openldap/dist/servers/slapd/filter.c: revision 1.2
external/bsd/openldap/dist/servers/slapd/overlays/deref.c: revision 1.2
Apply: http://www.openldap.org/devel/gitweb.cgi?p=openldap.git;a=commit;h=c32e74763f77675b9e144126e375977ed6dc562c
The deref overlay in slapd 2.4.13 through 2.4.40 dereferences a NULL
pointer when a search request includes the Deref control with an empty
list of attributes to return (missing input validation). [CVE-2015-1545]
--
Apply: http://www.openldap.org/devel/gitweb.cgi?p=openldap.git;a=commit;h=2f1a2dd329b91afe561cd06b872d09630d4edb6a
Certain search queries including the Matched Values control can trigger
a double free in slapd 2.4.40 when freeing operation controls. This is a
regression in 2.4.40, no earlier releases are affected. [CVE-1546]

(snj)

Pull up following revision(s) (requested by christos in ticket #678):
sys/kern/uipc_usrreq.c: revision 1.175
PR/39918: ITOH Yasufumi: Replace KASSERT with continue, since the file
descriptor can be closed since closef() does not pay attention to FDEFER.
XXX: Pullup-7

(snj)

Pull up following revision(s) (requested by christos in ticket #677):
lib/libc/sys/kqueue.2: revision 1.34
sys/kern/kern_event.c: revision 1.83
put the exit code of the process in data, like FreeBSD does.
--
say that we put the exit code in data.

(snj)

Pull up following revision(s) (requested by christos in ticket #676):
usr.bin/man/man.conf.5: revision 1.25
usr.sbin/makemandb/makemandb.8: revision 1.8
mention that this replaces makewhatis(8)
--
there is no more makewhatis; it has been replaced by makemandb

(snj)

Pull up following revision(s) (requested by christos in ticket #675):
usr.sbin/makemandb/makemandb.c: revision 1.27
- handle section numbers that are not single digits
- don't allocate and free needlessly

(snj)

Pull up following revision(s) (requested by nonaka in ticket #672):
sys/dev/i2c/ds1307.c: revisions 1.19, 1.20
sys/dev/i2c/ds1307reg.h: revision 1.5
- support DS3231 ( more or less a DS3232 without NVRAM it seems )
- support the DS3231's temperature sensor
--
Initialize a variable that gcc thinks might be used uninitialized.

(snj)

Use explicit_memset.

(riastradh)

Omit needless declaration.

(riastradh)

Cull another unused rndpool routine.

(riastradh)

Cull some unused rndpool routines.

(riastradh)

Remove remaining includes from <sys/rnd.h>.

(riastradh)

Include <sys/rndsource.h> for rnd_add_data.

(riastradh)

More rnd.h user cleanup.

(riastradh)

Add required dependency on i2cexec

(pgoyette)

Use <sys/rndio.h> for rnd ioctls.

(riastradh)

Update module dependencies:

drmkms does not depend on iic
drmkms_linux depends on i2cexec

(pgoyette)

Update header file references in rnd man pages.

(riastradh)

Use <sys/rndio.h> instead of <sys/rnd.h>.

(riastradh)

Limit <sys/rndsource.h> include to kernel.

(riastradh)

Oops -- install <sys/rndio.h> now that ioctls live there.

(riastradh)

Fix location of comment for MIN/MAX.

(riastradh)

Missed a spot: rnd.h -> rndsource.h.

(riastradh)

MD rnd.h cleanups.  Please let me know if I broke anything!

(riastradh)

sync with new acpica struct shuffling.

(christos)

CID 1293655: Forward NULL deref

(christos)

CID 1293652: Forward NULL deref

(christos)

CID 1293644: Check returns

(christos)

CID 1293642: Check returns

(christos)

CID 1269537: Unreachable code

(christos)

CID 1199067: Unchecked return

(christos)

CID 1018734: Fix uninit

(christos)

CID 996110: Fix memory leak

(christos)

new acpica

(christos)

update for newest acpi
XXX: needs recompiled yacc

(christos)

update for new acpica

(christos)

resolve conflicts.

(christos)

bump table to the max value allowed

(christos)

Omit needless <sys/rnd.h>.

(riastradh)

Convert remaining MI <sys/rnd.h> stragglers.  Many MD ones left.

(riastradh)

<sys/rnd.h> -> <sys/rndsource.h> in kern_todr.c.

(riastradh)

<sys/rnd.h> not needed for kern_pax.c.

(riastradh)

<sys/rnd.h> not needed for pf_norm.c.

(riastradh)

Missed a spot: <sys/rndsource.h> for if_cdcereg.h.

(riastradh)

Convert sys/dev to use <sys/rndsource.h>.

(riastradh)

Need <sys/rndsource.h>, not <sys/cprng.h>, for rnd source stuff.

(riastradh)

Add pmap locking to pmap_kenter_pa/kremove

(matt)

Convert arch/x86 to use <sys/rnd*.h>.  Omit needless includes.

(riastradh)

<sys/rndsource.h> needs <sys/rndio.h> for RND_TYPE_*, RND_FLAG_*.

(riastradh)

Omit needless include of <sys/fcntl.h> from <sys/cprng.h>.

(riastradh)

cprng_strong(kern_cprng, ...) never blocks, pass 0 for flags.

FASYNC was wrong anyway!  It's FNONBLOCK.

(riastradh)

Remove bogus #include <sys/rnd.h> from <sys/cprng.h>.

(riastradh)

Add header guards and necessary includes.

(riastradh)

Split <sys/rnd.h> into several header files.

It is silly that changing it causes ~the whole kernel to rebuild.
New header files:

- <sys/rndio.h> (user/kernel) for the ioctl interface.
- <sys/rndsource.h> (kernel-only) for the entropy source API.
- <sys/rndpool.h> (kernel-only) for the entropy pool algorithm.

(We already had <sys/rndsink.h> for the asynchronous reseed API.)

For now, <sys/rnd.h> includes all of these, until all users are
converted to include what they actually need.

(riastradh)

Little rnd.h cleanup.

- Make rnd_printing static in kern_rndq.c.
- Merge some #ifdef _KERNEL blocks.
- Fix comments about RND_DEV_(U)RANDOM.
- Move rnd_ready to rnd_private.h.

(riastradh)

Gather rnd-private declarations into <dev/rnd_private.h>.

Let's try to avoid putting externs in .c files where the compiler
can't check them.

(riastradh)

If we're going to use the queue macros, use them as intended.

(riastradh)

Move rndpool_(maybe_)extract to rndq, rename to rnd_(try)extract.

Make rnd_extract_data static.

(riastradh)

Omit needless define.

(riastradh)

Last ones... for some time to come.

(rodent)

By popular request.

(rodent)

Welcome to 2015, acronyms.

(rodent)

Use a bit per IPI type in local mailbox 0 registers. Ok matt@

(jmcneill)

Strip IO_JOURNALLOCKED, PGO_JOURNALLOCKED out of ulfs_readwrite.c.

These are vestigial from ufs_readwrite.c with wapbl -- lfs does not
have a journal but only the explicit wapbl calls, not these flags,
got ripped out in the transition to ulfs_readwrite.c.

(riastradh)

Same putpages->kassert in ulfs_readwrite.c

(riastradh)

Omit now-unused variable.  rump build didn't catch this...

(riastradh)

Don't putpages in ufs buffercached writes: kassert there are none.

(riastradh)

Don't break when hardclock_ticks wraps around.

Since we now only count time spent in wait, rather than determining
the end time and checking whether we've passed it, timeouts might be
marginally longer in effect.  Unlikely to be an issue.

(riastradh)

flushing the framebuffer from interrupt has several problems:
- As pointed out by Jared McNeill, this doesn't work for cold
  or ddb. Some specific handing for these cases would be needed,
  but there's no easy way to know if we're in ddb or not.
- periodic bus_dmamap_sync() of the whole framebuffer isn't cheap (about
  3% CPU constant on the beaglebone back).

instead map the framebuffer BUS_DMA_PREFETCHABLE (undocumented flag, but
is MI and properly handled by arm). This allows write-combining,
which, in addition with a shadow framebuffer for rasops(9) gives
a good performance boost too (difference with the cached+sync from intr
isn't noticeable on a boot sequence).
For X11, the framebuffer was already mapped BUS_DMA_PREFETCHABLE.
A few tests shows that mapping it fully cached doesn't make a noticeable
difference either (with or without xorg's shadow framebuffer).

(bouyer)

RPI2 MP support.

Thanks to Matt Thomas for making this possible with his changes to pic.c

(skrll)

Don't adjust interval if we are too far into the next tick.

(matt)

document ulimit -r

(jmcneill)

Don't send IPIs to ourselves if sending to everyone.

(matt)

replace NFS_BOOT_BOOTP with NFS_BOOT_DHCP

(jmcneill)