add instruction cache throttling SPR found on 750

(macallan)

Missing includes for some architectures

(justin)

Add uname to rump sources

(justin)

Move hw.machine and hw.machine_arch sysctls to base so rump can use them

This allows uname(3) and uname(1) to work on rump kernels.

(justin)

This enum is likely to be made unsigned by the compiler, so the assertion
will not work and clang objects with -Wtautological-constant-out-of-range-compare

(justin)

don't forget to NULL terminate the mode list when adding a single mode from
EDID. Now this works again on my Pismo.

(macallan)

CID 1309355: check error return from closefrom(3) where appropriate.

(christos)

zgrep(1): suppress the prefixing of filename on output when only
one file is specified to match the grep(1)'s output.

(nakayama)

Fix typo in comment.

(wiz)

Change tmpfs to vcache.
- Use tmpfs node address as key.
- Remove tn_vlock, field tn_vnode now protected by vcache.
- Add a hold count to tmpfs node to prevent nodes from disappearing
  while tmpfs_fhtovp() trys to vcache_get() them.  Last holder
  destroys reclaimed nodes.
- Remove the now unneeded parent unlock/lock for lookup of '..'.

(hannken)

Use VFS_PROTOS() for tmpfs.

(hannken)

Fix one last time.  mark libarm_g.a libpmc_g.a as compatfile, add
oabi/libc_vfp_g.a as compat.

(matt)

Add some compat oabi debug libs

(martin)

Backout previous, needs to be done differently

(martin)

Add some more debug/compat debug files.

(martin)

Don't clear PSL_FP/PSL_VEC

(matt)

Remove PSL_SPV from BOOKE PSL_USERMOD

(matt)

Back out last change.  Clearing PSL_FP/PSL_VEC is handled by child_return.

(matt)

Don't preserve PSL_FP/PSL_VEC in cpu_setmcontext.

(matt)

Clear PSL_FP/PSL_VEC so the lwp won't think it owns them.

(matt)

Don't reload if just re-enabling

(matt)

Add a check to make sure that if PSL_FP is set, we own the FPU.

(matt)

Don't reload the FPU register if this is just a re-enable.

(matt)

reapply some fixes to make multi-threaded ctfmerge work
that were lost in the overzealous resync with freebsd.
remove the hacks which were applied to work around those problems.

(chs)

Some test programs require these.

(matt)

848, 852, 854, 856-859, 862-864

(snj)

Pull up following revision(s) (requested by chs in ticket #854):
sys/external/bsd/drm2/dist/drm/i915/i915_gem.c: revision 1.31
fix Xorg coredumps that have started happening recently.
the problem is that we get a SIGALRM while we're sleeping during a page fault
on a mapping of a GEM object, and since we're sleeping interruptibly,
the GEM operation fails with EINTR.  this error is returned all the way back
through uvm_fault() to the trap handler, which responds to that error
by delivering a SIGSEGV.
fix this by doing like the linux version of the GEM fault handler and converting
EINTR into success, which results in delivering the original signal and
retrying the fault.

(snj)

Pull up following revision(s) (requested by chs in ticket #848):
sys/external/bsd/drm2/drm/drm_vma_manager.c: revision 1.3
plug memory leak in drm_vma_node_revoke().

(snj)

Pull up following revision(s) (requested by maxv in ticket #858):
sys/dev/ic/ac97.c: revision 1.96
fix memory leak, found by Brainy.
XXX: is there anything else that we need to cleanup at this point?

(snj)

Pull up following revision(s) (requested by maxv in ticket #857):
sys/arch/arm/xscale/ixp425_if_npe.c: revision 1.29
Initialize 'error'.
Can't test, but obvious enough apparently.
Found by Brainy.

(snj)

Pull up following revision(s) (requested by maxv in ticket #856):
sys/arch/pmax/tc/dt.c: revision 1.12
Use-after-free.
ok christos@
Found by The Brainy Code Scanner.

(snj)

Pull up following revision(s) (requested by dholland in ticket #859):
sys/sys/dirent.h: revision 1.29
fix inadequate parens in macro (my fault originally)

(snj)

Pull up following revision(s) (requested by martin in ticket #852):
sys/arch/vax/boot/boot/boot.c: revision 1.33
Close the file descriptor used for loading before jumping to the kernel.
Pointed out by Felix Deichmann on port-vax.

(snj)

Pull up following revision(s) (requested by sborrill in ticket #863):
sbin/raidctl/raidctl.8: revisions 1.68, 1.69
Clarify that what was previously known as -A root is now -A forceroot, not
-A softroot and that -A root can still be used for historical reasons.
--
Bump date for previous. Add whitespace.
Fix some mandoc warnings.

(snj)

Pull up following revision(s) (requested by sborrill in ticket #862):
sbin/raidctl/raidctl.c: revision 1.61
Compare correct length string for force option to -A

(snj)

Make clang happy by adding -mfpu=none to -mfloat-abi=soft

(matt)

Pull up following revision(s) (requested by martin in ticket #864):
usr.sbin/postinstall/postinstall: revision 1.196 via patch
PR install/50020: postinstall does not check for mandoc in /etc/man.conf.
Add a check, but ask the user to fix manually (the file could have local
modifications).

(snj)

Revert patch from PR/48715 as it is incorrect.

(skrll)

Deregister pmf hook when attach failed.

(nonaka)

Typo fxi.

(matt)

Revamped QVSS (VCB01) support.  (from Charles Dickman)

(matt)

Configury support for QVSS (from Charles Dickman)

(matt)

Add code to probe the Q-bus/Unibus for memory.  (From Charles Dickman)

(matt)

Reorganize a bit.  Add QVSS (from Charles Dickman).

(matt)

Hooks needed for QVSS support

(matt)

fix infinite loop

(christos)

fix memory leaks on error; found by brainy.

(christos)

If we fail 5 times in a row, fail correctly (Steffen Nurpmeso).

(christos)

strptime now sets tm_yday

(christos)

Improve the code in __sync_bool_compare_and_swap_{1,2,4}.
- bccs is smaller and faster than bcc(.w) in this case.
- it can be used movql in this case (assembler optimise it though).

(isaki)

atomic_cas_{8,16}:
- Correct the wrong offset in stack.
__sync_bool_compare_and_swap_{1,2,4}:
- Correct the wrong offset in stack.
- D3 must be preserved in subroutines.
PR/49995.

(isaki)

Remove a dead continue.

Harmless, found by Brainy

(maxv)

Remove a dead continue.

Harmless, found by Brainy

(maxv)

New sentence, new line.

(wiz)

/usr/sbin/chown -> /sbin/chown
(not sure why these need explicit paths at all, but whatevs)

(dholland)

Remove entries made redundnat by compatdir/compatfile

(matt)

Remove entries made redundant by compatdir/compatfile.  (Leaves only
libexec/ld.elf_so-rv32 as the only compat files).

(matt)

Remove entries made redundant by compatdir/compatfile

(matt)

Add basic support for tap detection, emulating a left-click.

(bouyer)

Regen

(skrll)

Lumix Camera DMC-FS45 from kern/49809

(skrll)

PR/50009: David CARLIER: Enhance strptime to fill out more fields when it
can. From FreeBSD.

(christos)

Add an explicit note about the search path used by dlopen(3).

(joerg)

PR/50021: Rin Okuyama: Fix linux affinity syscalls
XXX: pullup-7

(christos)

new libssh

(christos)

merge conflicts

(christos)

td_close() needs to be able to write to the process image, so call sooner,
before we detach from it.

(christos)

PR install/50020: postinstall does not check for mandoc in /etc/man.conf.
Add a check, but ask the user to fix manually (the file could have local
modifications).

(martin)

Simplify previous, pass awk as TOOL_AWK (that is what the makefile uses),
create an obj dir for make and force it to use that.
This version now works with read-only source again.

(martin)

Somehow the ; disappeared from the previouis commit

(skrll)

From FreeBSD rev1.30 via OpenBSD:
- always initialise the pipe multiplier to a valid value
- do not unchain the the interrupt QHs immediately after initialisation

Prompted by msaitoh

(skrll)

Use armreg_*_{read,write} instead of handcrafted asm

No functional change.

(skrll)

Fix bus clock for Airmont from the latest Intel SDM.

(msaitoh)

document PIOD_READ_AUXV.

(christos)

Support PIOD_READ_AUXV so that gdb can handle PIE binaries. From OpenBSD.

(christos)

fix lint sets.

(christos)

Add Xeon E5-4600 v3,  Xeon E3-1200 v4 etc. from the latest Intel SDM.

(msaitoh)

remove lint hack.

(christos)

new test

(christos)

new test.

(christos)

Fix segmentation fault caused by freeing prematurely function parameter types
inside cast expressions.

(christos)

Clean up the src/etc/mtree directory after using it to generate the mtree
spec file.
XXX this is fishy, we should find a better way.

(martin)

Add tests of interface creation/destruction

(ozaki-r)

Unfortunately MFS uses v_data of its anonymous device vnode so
it cannot be used as vcache key.  Use v_interlock as key ...

(hannken)

tickets 860 and 861

(bouyer)

Pull up following revision(s) (requested by dholland in ticket #861):
usr.bin/calendar/calendar.c: revision 1.52
Repair accidental regression in -r1.49: for -a mode, don't allow
calendars to be other than regular files unless the -x option is in
effect.
(If not in -a mode, we're running purely as the user whose calendar it
is and if they want to DoS themselves with named pipes it's their own
lookout.)

(bouyer)

Pull up following revision(s) (requested by dholland in ticket #860):
usr.bin/calendar/calendar.c: revision 1.51
Correct privilege handling problems in calendar -a (which runs as root
from /etc/daily); do not exec other programs while the real uid is
still 0.
Also, clear the supplementary groups list up front and call initgroups
when becoming another user, to avoid leaking any extra group
privileges that we might have.
And finally, don't silently ignore errors changing uid and gid; those
are serious if they happen.

(bouyer)

ticket 1309

(bouyer)

ticket 1309

(bouyer)

ticket 1309

(bouyer)

Pull up following revision(s) (requested by dholland in ticket #1309):
usr.bin/calendar/calendar.c: revision 1.51
Correct privilege handling problems in calendar -a (which runs as root
from /etc/daily); do not exec other programs while the real uid is
still 0.
Also, clear the supplementary groups list up front and call initgroups
when becoming another user, to avoid leaking any extra group
privileges that we might have.
And finally, don't silently ignore errors changing uid and gid; those
are serious if they happen.

(bouyer)

Pull up following revision(s) (requested by dholland in ticket #1309):
usr.bin/calendar/calendar.c: revision 1.51
Correct privilege handling problems in calendar -a (which runs as root
from /etc/daily); do not exec other programs while the real uid is
still 0.
Also, clear the supplementary groups list up front and call initgroups
when becoming another user, to avoid leaking any extra group
privileges that we might have.
And finally, don't silently ignore errors changing uid and gid; those
are serious if they happen.

(bouyer)

Pull up following revision(s) (requested by dholland in ticket #1309):
usr.bin/calendar/calendar.c: revision 1.51
Correct privilege handling problems in calendar -a (which runs as root
from /etc/daily); do not exec other programs while the real uid is
still 0.
Also, clear the supplementary groups list up front and call initgroups
when becoming another user, to avoid leaking any extra group
privileges that we might have.
And finally, don't silently ignore errors changing uid and gid; those
are serious if they happen.

(bouyer)

ticket 1971

(bouyer)

ticket 1971

(bouyer)

ticket 1971

(bouyer)

Pull up following revision(s) (requested by dholland in ticket #1971):
usr.bin/calendar/calendar.c: revision 1.51
Correct privilege handling problems in calendar -a (which runs as root
from /etc/daily); do not exec other programs while the real uid is
still 0.
Also, clear the supplementary groups list up front and call initgroups
when becoming another user, to avoid leaking any extra group
privileges that we might have.
And finally, don't silently ignore errors changing uid and gid; those
are serious if they happen.

(bouyer)

Pull up following revision(s) (requested by dholland in ticket #1971):
usr.bin/calendar/calendar.c: revision 1.51
Correct privilege handling problems in calendar -a (which runs as root
from /etc/daily); do not exec other programs while the real uid is
still 0.
Also, clear the supplementary groups list up front and call initgroups
when becoming another user, to avoid leaking any extra group
privileges that we might have.
And finally, don't silently ignore errors changing uid and gid; those
are serious if they happen.

(bouyer)

Pull up following revision(s) (requested by dholland in ticket #1971):
usr.bin/calendar/calendar.c: revision 1.51
Correct privilege handling problems in calendar -a (which runs as root
from /etc/daily); do not exec other programs while the real uid is
still 0.
Also, clear the supplementary groups list up front and call initgroups
when becoming another user, to avoid leaking any extra group
privileges that we might have.
And finally, don't silently ignore errors changing uid and gid; those
are serious if they happen.

(bouyer)

Repair accidental regression in -r1.49: for -a mode, don't allow
calendars to be other than regular files unless the -x option is in
effect.

(If not in -a mode, we're running purely as the user whose calendar it
is and if they want to DoS themselves with named pipes it's their own
lookout.)

(dholland)

Correct privilege handling problems in calendar -a (which runs as root
from /etc/daily); do not exec other programs while the real uid is
still 0.

Also, clear the supplementary groups list up front and call initgroups
when becoming another user, to avoid leaking any extra group
privileges that we might have.

And finally, don't silently ignore errors changing uid and gid; those
are serious if they happen.

(dholland)

Use ip_hresolv_output for if_token as well

I thought we cannot apply ip_hresolv_output to if_token because
rt0 looked being needed by arpresolve in token_output. However,
rt0 is actually not used by arpresolve in NetBSD (see obsolete
ARPRESOLVE macro).

(ozaki-r)

Bump date for previous. Add whitespace.
Fix some mandoc warnings.

(wiz)

handle PIC compilation (if we are building a PIE system; this is used by tests)

(christos)

Clarify that what was previously known as -A root is now -A forceroot, not
-A softroot and that -A root can still be used for historical reasons.

(sborrill)

Compare correct length string for force option to -A

(sborrill)

Fix logic inversion in 1.107

(skrll)

whitespace police.

(christos)

Allow /var/shm to be a symlink to a properly set up directory.

(martin)

Use KASSERT for argument NULL checks

(ozaki-r)

Fix nd6_numroutes counting

nd6_numroutes is intended to be incremented when a route is added via RA
and decremented when a RA route is deleted. However, a decrement of a RA
route was skipped when there remained references to the RA route.

(ozaki-r)

Redo previous again, v_specnode is invariant but not unique.

Set "vp->v_data = vp" and use v_data as key.

(hannken)

Fix reversed arguments in copyin(9), c.f. kern/sys_mqueue.c rev 1.39.

(martin)

Make vmparam.h change work with RUMP

(matt)

don't mess with the cycle counter event counter, out timer interrupt comes
from elsewhere and is counted there

(macallan)

shorten the interrupt name so the level field is actually readable in systat

(macallan)

don't map the framebuffer as cacheable

(macallan)

We support multiple page sizes so let modules know it.

(matt)

Use cpu_startup_common()

(matt)

use cpu_startup_common

(matt)

841, 847, 849, 853, 855

(snj)

Pull up following revision(s) (requested by jnemeth in ticket #847):
sbin/gpt/recover.c: revision 1.6
Instruct user to use resizedisk if media size has changed, as
resizedisk will adjust the media size in the headers, whereas
recover simply copies the existing header over the missing one.
XXX recover and resizedisk should probably be merged (even if just
partially, so that recover can properly handle media size changes).
Also, reading in the GPT should probably be centralised so that
error handling can be centralised, and users aren't given misleading
messages (i.e. they aren't told to run recover when they should be
running resizedisk).  However, something that can be quickly pulled
up to netbsd-7 was needed, and there isn't time for a major overhaul
or rewrite.

(snj)

Pull up following revision(s) (requested by phx in ticket #841):
sys/arch/amiga/include/vmparam.h: revision 1.43
Remove unused KUSER_AREA, SYSPTSIZE, USRPTSIZE.
Bump MAXTSIZ and MAXDSIZ to the same values atari is using.
This makes gcc 4.8 (/usr/libexec/cc1) load and execute.

(snj)

Pull up following revision(s) (requested by roy in ticket #853):
usr.bin/ldd/Makefile.inc: revision 1.6
Do not set MLIBDIR if MKCOMPAT=no

(snj)

Pull up following revision(s) (requested by chs in ticket #855):
external/cddl/osnet/dist/tools/ctf/cvt/dwarf.c: revision 1.11
don't fail for anonymous unions.

(snj)

Pull up following revision(s) (requested by mrg in ticket #849):
sys/conf/copyright: revision 1.13
Welcome to 2015!

(snj)

Use the address of vp->v_specnode as vcache key.  It is invariant
over the lifetime of the vnode.

The previous worked by luck, it took the first sizeof(void *) bytes
of struct vnode as key.

Resolves CID 1308957: wrong sizeof()

(hannken)

Remove a dead branch. Could look like a memory leak, but ih cannot be
NULL.

Found by Brainy.

(maxv)

Revert previous, and explain why.

(christos)

Typos

(skrll)

CID 1308958: Fix reversed arguments in copyin(9)

(christos)

CID 1308957: Fix wrong sizeof()

(christos)

Use-after-free.

ok christos@

Found by Brainy.

(maxv)

Remove ifnet_addrs

We can assume that ifnet_addrs[ifp->if_index] is always the same as
ifp->if_dl, so we can replace ifnet_addrs[ifp->if_index] with ifp->if_dl
and remove ifnet_addrs entirely.

ok martin@

(ozaki-r)

add brconfig to userland utils

(pooka)

Need debug sets if ${MKDEBUGLIB} != "no"

(matt)

Print both user and kernel segtabs

(matt)

Add serial comma, use upper case for abbreviation, and remove apostrophe for plural.

(wiz)

fix ifstats command by relaxing invalid sanity check - ntp bug #2860

(kardel)

Initialize 'error'.

Can't test, but obvious enough apparently.

Found by Brainy.

(maxv)

Move makphy to after mv{gb,xp}e.
And indents.

(kiyohara)

Remove mvsdio.  OpenBlocks AX3 not have it.

(kiyohara)

Small fixes.

ok hannken@

(maxv)

Pass AWK to make when invoking emit_dist_file.
Part of fixing PR toolchain/50004.

(martin)

Guard a few $MK... != "no" tests by an additional defined(MK...) clause
and make the emit_dist_file target depend on EXTRA_DIST_FILES.
Part of fixing PR toolchain/50004.

(martin)

Use-after-free.

ok christos@

Found by The Brainy Code Scanner.

(maxv)

Remove unused struct pt_entry_t union.

(matt)

Need to test for MKCOMPATX11

(matt)

liblisp, libre, libmp are now private lirbaries and no longer installed.

(matt)

X11 libs liblisp, libmp, libre are no longer installed and are now obsolete.

(matt)

Use LIBISPRIVATE, not MKPRIVATELIB

(matt)

Add MKCOMPATX11 knob for making X11 compat libraries.

(matt)

p++ symlink is only create if MKCXX != no

(matt)

p++ is a symlink so it won't have a .debug

(matt)

Fix various missing or incorrect keywords.

(matt)

add pigzgzip to comment

(matt)

Terminate the obsole directory walking loop when we hit "/" or "."
(both relative to DESTDIR), I have seen it run into an endless loop with
_path=. when invoked from the top level make with some (valid) args.

(martin)

Remove useless and harmful sync(2) call in umount(8)

Remove sync(2) call before unmount(2) in umount(8). This sync(2) is useless
since unmount(2) will perform a VFS_SYNC anyway.

But moreover, this sync(2) may be harmful, as there are some situation where
it cannot return (unreachable NFS server, for instance), causing umount -f
to be uneffective.

(manu)

add one

(dholland)

note GCC 4.8.5 has been imported, and while i'm here, untag thorpej
and tag christos and skrll as being responsible :-)

(mrg)

Whitespace.

(skrll)

Make evbarm64 (little endian) the default for aarch64.

(matt)

add whitespace

(matt)

Cleanup includes.

(matt)

Decode jr.hb and ssnop and few more spec3 instructions
u_int32_t -> uint32_t

(matt)

More instructions

(matt)

Turn KX (no need for UX) when !O32 ABI and MULTIPROCESSOR

(matt)

ifdef out bitfield struct for pte (not used).

(matt)

Cleanup includes
Use cpu_startup_common()

(matt)

Cleanup includes

(matt)

Cleanup includes

(matt)

#include <mips/locore.h>

(matt)

#include <mips/locore.h> and other include cleanups.

(matt)

#include <mips/locore.h>

(matt)

Cleanup includes

(matt)

Cleanup includes

(matt)

Remove pmapdebug stuff

(matt)

Use cpu_startup_common()

(matt)

Just cpu_startup_common()

(matt)

Move <mips/locore.h> lower

(matt)

Add #include <mips/locore.h>

(matt)

if NEED_MDSETIMAGE != no and the ramdisk.fs exists
add ramdisk.fs to SYSTEM_DEP
if DEBUG is defined, mdsetimage the ramdisk image into the .gdb copy

(matt)