refactor the bitstring/mask operations to be behind an API. Make pending interrupt marking cpu aware.

(cherry)

KNF police (rmind@ :-)

(cherry)

Update the (shared) v_interlock if the upper node changes.

(hannken)

Sync usage with man page.

(wiz)

xen ipi infrastructure

(cherry)

newfs_v7fs(8)'s -v and -P options are obsolete. changed to newfs(8) compatible -V option.

(uch)

Disable bzip2 support in freetype on xorg tools build.
Fixes PR xsrc/45223 (xorg tools build fails on CentOS due to
lack of bz2 support).

(tsutsui)

Make it possible to disable #define FT_CONFIG_OPTION_USER_BZIP2
setting by -DFT_CONFIG_OPTION_DISABLE_BZIP2 for tools build
on systems which don't have native bzip2 support.
Ok'ed by mrg@ in PR xsrc/45223.

(tsutsui)

Introduce locking primitives for Xen pte operations, and xen helper calls for MP related MMU ops

(cherry)

fix root.cache

(christos)

disable some warnings to appease gcc-4.5

(plunky)

use offsetof() rather than rolling our own, to appease gcc 4.5

(plunky)

fix up some undefined behaviour, to appease gcc 4.5

(plunky)

Obsolete some include files.

(kiyohara)

Add Xen specific members to struct cpu_info, Add proper per-cpu curcpu() functionality

(cherry)

tweak the xen specific startup path to not use %fs before it is setup. Minor refactoring. Use Xen specific ipi calls.

(cherry)

Correct offset calculation for ci

(cherry)

Include Xen specific definitions.

(cherry)

Add Xen specific ipi bitmasks

(cherry)

For devices, sockets and fifos ignore setting the file size to zero to make
open(..., O_TRUNC) happy and allow them to write through the lower layer.

Fixes PR #43560 (writing to null device in unionfs fails)

(hannken)

Use LK_SHARED, it is sufficient for VOP_GETATTR() and VOP_READDIR().

(hannken)

Revert previous, it breaks nullfs. (And I guess there are no tests for
nullfs?)

(dholland)

fix build

(jmcneill)

Fail namei immediately if searchdir is unlinked / has been rmdir'd.
Do this by checking if v_size == 0. Should fix PR 44658 (and PR 32661).

(dholland)

Include missing part of previous commit to this file. (sigh)

(dholland)

Correct a comment on config_found_sm_loc().

(dyoung)

Add support for the Oxford Semiconductor OXPCIe952, a 2-port serial chip
with a PCI Express interface, in legacy mode.  Cursorily tested with a
StarTech PEX2S952LP.

See the manual for your adapter to see how to put it into legacy mode.
There is also a "native" mode that uses a different clock multiplier
and (I think) provides a richer feature set, but also requires a bit of
additional initialization (see FreeBSD's puc(4)).

The native-mode PCI functions use different PCI Device IDs than the
legacy-mode functions.

(dyoung)

__typeof__, not __typeof, to avoid confusing passersby.

(dholland)

Ticket 1653.

(riz)

Apply patches (requested by tron in ticket #1653):
external/ibm-public/postfix/dist/HISTORY patch
external/ibm-public/postfix/dist/makedefs patch
external/ibm-public/postfix/dist/postfix-install patch
external/ibm-public/postfix/dist/src/cleanup/cleanup_milter.c patch
external/ibm-public/postfix/dist/src/global/abounce.c patch
external/ibm-public/postfix/dist/src/global/abounce.h patch
external/ibm-public/postfix/dist/src/global/mail_version.h patch
external/ibm-public/postfix/dist/src/local/Makefile.in patch
external/ibm-public/postfix/dist/src/local/bounce_workaround.c patch
external/ibm-public/postfix/dist/src/local/mailbox.c patch
external/ibm-public/postfix/dist/src/local/unknown.c patch
external/ibm-public/postfix/dist/src/master/master_ent.c patch
external/ibm-public/postfix/dist/src/milter/milter8.c patch
external/ibm-public/postfix/dist/src/oqmgr/qmgr_active.c patch
external/ibm-public/postfix/dist/src/postdrop/postdrop.c patch
external/ibm-public/postfix/dist/src/qmgr/qmgr_active.c patch
external/ibm-public/postfix/dist/src/smtpd/smtpd_check.c patch
external/ibm-public/postfix/dist/src/util/sys_defs.h patch

Update Postfix to version 2.7.5. Changes since version 2.7.4:
- Performance: a high load of DSN success notification requests could
  slow down the queue manager. Solution: make the trace client
  asynchronous, just like the bounce and defer clients.
- The local(8) delivery agent ignored table lookup errors in
  mailbox_command_maps, mailbox_transport_maps,
  fallback_transport_maps and (while bouncing mail to alias) alias
  owner lookup.
- Workaround: dbl.spamhaus.org rejects lookups with "No IP queries"
  even if the name has an alphanumerical prefix. We play safe, and
  skip both RHSBL and RHSWL queries for names ending in a numerical
  suffix.
- The Postfix Milter client reported a temporary error instead of
  "file too large" in three cases.
- The "sendmail -t" command reported "protocol error" instead of
  "file too large", "no space left on device" etc.

(riz)

Simplify handling of slashes. Provides a proper fix for PR 44961.

(dholland)

Check (and fix) that unprivileged-user has been changed from _atf to _tests
in /etc/atf/common.conf.  Requested by martin@.

(jmmv)

Xref limits(3).

(jruoho)

Note the over fifty new manual pages.

(jruoho)

Add a summary page for <limits.h>.

(jruoho)

Tickets 1651, 1652.

(riz)

Pull up following revision(s) (requested by matt in ticket #1652):
sys/sys/kthread.h: revision 1.7.10.1.8.1
sys/kern/kern_kthread.c: revision 1.24.10.2.8.1
Pullup <a  rel="nofollow" href="http://mail-index.netbsd.org/source-changes/2010/05/11/msg009810.html">http://mail-index.netbsd.org/source-changes/2010/05/11/msg009810.html</a>
to this branch because the new flash framework will need it.

(riz)

Also: netpgp(1).

(jruoho)

Remove few entries that clearly have been done (aio(3), ssp(3), tests(7)).

(jruoho)

Pull up following revision(s) (requested by manu in ticket #1651):
usr.bin/extattr/getextattr.1: revision 1.4
usr.bin/extattr/getextattr.c: revision 1.8
Improve extended attribute commands:
- allow namespace to be omitted when using a namespace-prefixed attribute
  name, a la Linux.
- Improve hexdump for getextattr -x
- Add more display option for binary attributes, using vis(3)
- Add a -i option to setextattr to get the attribute value from file, so
  that we can set binary values

(riz)

when sorting the metalog for unpriv builds, first run it through sort so
mtree doesn't fail with:
  nbmtree: .: missing directory in specification
  nbmtree: failed at line 1 of the specification

(jmcneill)

While the GENERIC/MONOLITHIC/MODULAR issue remains to be sorted out,
MONOLITHIC is still a part of the release build and needs to remain
config(1)urable.

(jakllsch)

Add missing RUMP_SERVER export, so the "normal" test does not fail with
a stupid error.

(martin)

No longer needs -Wno-array-bounds with clang.

(joerg)

Requires -Wno-array-bounds with clang now.

(joerg)

Passes with the relaxed array boundary check in clang

(joerg)

libc passes the relaxed array boundary check in clang now.

(joerg)

Restrict -Wno-array-bounds to tulip.c with gcc.

(joerg)

Passes relaxed array boundary checks with clang now.

(joerg)

Requires -Wno-unused-value -Wno-tautological-compare when building with
clang.

(joerg)

-Wno-array-bounds is no longer needed for clang here

(joerg)

Can't build libstdc++ and libsupc++ with clang since it depends on using
volatiles as initializers of non-volatile fields, void * arithmetic and
some new attributes for the unwinding.

(joerg)

Disable -Wunused-value for clang, it triggers on the bcopy replacement.

(joerg)

Since cproj and friends are used, link explicitly against libm.

(joerg)

Don't use array access to compute addresses that are definitely beyond
the static array boundaries.

(joerg)

Update LLVM/Clang snapshot to r137038. This brings in the extended array
boundary checker and the white listing of pre-c99 flexible array member
idiom foo[1] as last element of a struct.

(joerg)

Build acpiwdrt(4) as a module on x86.

(jruoho)

Two placeholder comments.

(jruoho)

add CLK_DELAY

(jmcneill)

Fix warning message.

(uch)

include cx24227, lg3303, mt2131, nxt2k, tvpll, xc3028, coram, cxdtv, and emdtv
modules in x86 build

(jmcneill)

lg3303 should depend on dtv_math, not xc3028

(jmcneill)

add dtv_math.h

(jmcneill)

Check whether superblock is writable sector.

(uch)

Fix uninitiaized variable usage (never though lint would miss that when
used by return statement) that caused unprivilegied user to fail on
unlink(2) and rename(2) operations.

(manu)

As wmieeepc(4) was added to amd64 GENERIC, build it also as a amd64 module.

(jruoho)

re-gen after PWDOG1 addition.

(mbalmer)

Add the QUANCOM PWDOG1.

(mbalmer)

Print 'Keyboard encoding -> sg' instead of just 'encoding -> sg' on the
console.

(mbalmer)

Add the USENIX paper of Watanabe.

(jruoho)

Fix a typo.

(jruoho)

Finally, add missing %D.

(jruoho)

One more.

(jruoho)

Add few papers to SEE ALSO. Particularly the first one demonstrates
the diminishing value of P-states compared to C-states on new CPUs.

(jruoho)

Fix warning

(manu)

Remove PUFFS_KFLAG_WTCACHE, which caused data corruption and slowdown

(manu)

No need to lock the selcluster in selscan() if either
NO_DIRECT_SELECT is defined or all polls return an event.

(hannken)

improve comment about AT_* defines: they are not only used by linkat(2)

(manu)

remove commented out entries for emdtv, auvitek, and dtv

(jmcneill)

add dtv_math module

(jmcneill)

modify the dtv device api so hardware drivers can be loaded independently
of the dtv module

(jmcneill)

Fix build failed.  mouse.h and pccons.h already removed.

(kiyohara)

COMPAT_16 code only exists on amd64 in the presence of COMPAT_NETBSD32.

(jakllsch)

Fix typo that prevented installmodules from working with -U

(jmcneill)

add an installmodules=<dir> command to build.sh

(jmcneill)

Tickets 1648-1650.

(riz)

Pull up following revision(s) (requested by tron in ticket #1650):
lib/librefuse/refuse.c: revision 1.94
Call the FUSE init operation before we try to access the file-system.
This prevents a crash in "fuse_ext2" which I previously worked around
with a patch.

(riz)

Pull up following revision(s) (requested by tron in ticket #1649):
lib/librefuse/fuse_opt.h: revision 1.5
PR/41250: David H. Gutteridge: librefuse fuse_opt.h header doesn't accommod=
ate
C++ compilation

(riz)

Try to fix a sentence. New sentence, new line.
Add a comment where the example is unclear. Use more markup.

(wiz)

Pull up following revision(s) (requested by tron in ticket #1648):
lib/librefuse/refuse.c: revision 1.93
Don't ignore "userdata" argument in "fuse_main_real". This can crash
a FUSE file-system that passes a non-NULL argument here.

(riz)

Tickets 1643, 1644, 1646.

(riz)

Pull up following revision(s) (requested by dholland in ticket #1646):
share/mk/bsd.x11.mk: revision 1.98
Fix the value of "appdefaultdir" that ends up in native X .pc files.
(It should be "app-defaults" but was "app-default", probably because
of a typo; it's been wrong ever since this logic was put in two and a
half years ago, and as app-defaults files are a legacy mechanism it
took this long for anyone to notice.) Should be pulled up to -5.

(riz)

Sort sections. Remove trailing whitespace. Sort errors.
New sentence, new line.

(wiz)

Ticket 1644.

(riz)

Pull up following revision(s) (requested by bouyer in ticket #1644):
sys/sys/socketvar.h: revision 1.126
sys/kern/init_main.c: revision 1.433
sys/kern/uipc_socket.c: revision 1.205
Fix kern/45093 as discussed on tech-kern@:
http://mail-index.netbsd.org/tech-kern/2011/06/17/msg010734.html
The cause of the problem is that the so_pendfree is processed with
the softnet_lock held at one point, and processing the list
calls sodoloanfree() which may kpause(). As the thread sleeps with
softnet_lock held, it ultimately cause a deadlock (see the PR or tech-kern
thread for details).
Although it should be possible to call sodopendfree() after releasing
the socket lock, it's not so easy to know where he socket lock is held and
where it's not, so we may hit the issue again later.
Add a kernel thread to handle the so_pendfree list, and wake up this
thread when adding mbufs to this list. Get rid of the various sodopendfree()
calls, hopefully fixing definitively the problem.

(riz)

Pull up following revision(s) (requested by bouyer in ticket #1644):
sys/sys/socketvar.h: revision 1.126
sys/kern/init_main.c: revision 1.433
sys/kern/uipc_socket.c: revision 1.205
Fix kern/45093 as discussed on tech-kern@:
http://mail-index.netbsd.org/tech-kern/2011/06/17/msg010734.html
The cause of the problem is that the so_pendfree is processed with
the softnet_lock held at one point, and processing the list
calls sodoloanfree() which may kpause(). As the thread sleeps with
softnet_lock held, it ultimately cause a deadlock (see the PR or tech-kern
thread for details).
Although it should be possible to call sodopendfree() after releasing
the socket lock, it's not so easy to know where he socket lock is held and
where it's not, so we may hit the issue again later.
Add a kernel thread to handle the so_pendfree list, and wake up this
thread when adding mbufs to this list. Get rid of the various sodopendfree()
calls, hopefully fixing definitively the problem.

(riz)

Ticket 1643.

(riz)

Pull up following revision(s) (requested by sborrill in ticket #1643):
sys/net/if.c: revision 1.243
Prevent if_detach() from crashing while it walks the routing table
to find and unlink routes that reference the detached ifnet: make
if_rt_walktree() return ERESTART whenever it has deleted a route.
Whenever rt_walktree() returns ERESTART, if_detach() restarts it.
I believe that this fix resembles one by Jonathan Kollasch or by someone
else, which has languished in a PR for too long.  Sorry!
Tested by me and by Jeff Rizzo.
XXX It's supposed to be safe for rn_walktree() to apply to the routing
XXX table a routine that may delete routes.  Why isn't it safe in
XXX practice?

(riz)

Ticket 1643

(riz)

Pull up following revision(s) (requested by sborrill in ticket #1643):
sys/net/if.c: revision 1.243
Prevent if_detach() from crashing while it walks the routing table
to find and unlink routes that reference the detached ifnet: make
if_rt_walktree() return ERESTART whenever it has deleted a route.
Whenever rt_walktree() returns ERESTART, if_detach() restarts it.
I believe that this fix resembles one by Jonathan Kollasch or by someone
else, which has languished in a PR for too long.  Sorry!
Tested by me and by Jeff Rizzo.
XXX It's supposed to be safe for rn_walktree() to apply to the routing
XXX table a routine that may delete routes.  Why isn't it safe in
XXX practice?

(riz)

Pull up following revision(s) (requested by sborrill in ticket #1643):
sys/net/if.c: revision 1.243
Prevent if_detach() from crashing while it walks the routing table
to find and unlink routes that reference the detached ifnet: make
if_rt_walktree() return ERESTART whenever it has deleted a route.
Whenever rt_walktree() returns ERESTART, if_detach() restarts it.
I believe that this fix resembles one by Jonathan Kollasch or by someone
else, which has languished in a PR for too long.  Sorry!
Tested by me and by Jeff Rizzo.
XXX It's supposed to be safe for rn_walktree() to apply to the routing
XXX table a routine that may delete routes.  Why isn't it safe in
XXX practice?

(riz)

Miscellaneous mbuf changes:

1 Add some protection against double-freeing mbufs in DIAGNOSTIC kernels.

2 Add a m_defrag() that's derived from
  sys/dev/pci/if_vge.c:vge_m_defrag().  This one copies the packet
  header.

3 Constify m_tag_find().

(dyoung)

remove dtv (available as a module)

(jmcneill)

remove dtv (available as a module)

(jmcneill)

#include <sys/sched.h> since it's a dependency

(matt)

Global inlines are not allowed.

(matt)

We have PUFFS, we need putter.

(jakllsch)

revert previous

(jmcneill)

Add P1003_1B_SEMAPHORE to match i386.

(jakllsch)

Finish reverting modularization of i386 GENERIC.

(jakllsch)

Fix bigendian build of iscsictl.

(riz)

Finish reverting premature modularization of amd64 kernels.

(jakllsch)

nfs_namei() should not return a non-null path buffer except on success,
even though the callers are apparently prepared to cope.

Fixes last tidyup part of PR 44625.

(dholland)

inline -> static inline
For successful gcc 4.5 builds.

(jakllsch)

regen

(manu)

First stage of support for Extended API set 2. Most of the think is
unimplemented, except enough of linkat(2) to hardlink to a symlink.

Everything new in headers is guarded #ifdef _INCOMPLETE_XOPEN_C063 since
some software (e.g.: xcvs in our own tree) will assume they can use openat(2)
when AT_FDCWD is defined. _INCOMPLETE_XOPEN_C063 will go away once support
will be completed.

(manu)

Remove suspicious warning message.

(uch)

The absence of USE_PLATFORM_CLOCK does not imply that the PM Timer is
unreliable; in fact, the PM Timer has to be reliable by definition! Remove
the incorrect warning message that was added in r1.19

(jmcneill)

If the USE_PLATFORM_CLOCK flag is set in the FADT, it indicates that OSPM
should use a platform provided timer (either HPET or the PM timer). A
platform may set this flag if internal processor clock(s) cannot provide
consistent monotonically non-decreasing counters. Set TSC quality to -100
if this flag is set.

(jmcneill)

Minor edits (I don't know enough about how thi works to help that much)

(dholland)

i've created the initial draft of the man page - take this off the list of
things to do.

(agc)

Also make the man page for iscsictl(8)

(agc)

Add a first pass at a manual page, iscsictl(8).

(agc)

Bounds process's stack size with max_stack_size so that 32bit
binary works regardless of stack size limit setting.

(enami)

Make this compile WARNS=4

(agc)

Make this compile WARNS=4

(agc)

- Add an argument to kcpuset_create() for zeroing.
- Add kcpuset_atomic_set(), kcpuset_atomic_clear() and kcpuset_merge().

(rmind)

Remove LW_AFFINITY flag and fix some bugs affinity mask handling.

(rmind)

New linux syscalls: fadvise64() and fadvise64_64().

(alnsn)

Another instance of REG_OFFSET that needs to be consistent with the rest.

(jakllsch)

Appease _LP64 build with GCC 4.5.

(jakllsch)

Fix XEN3PAE_DOMx build

(cherry)

Appease _LP64 build with GCC 4.5.

(jakllsch)

Make uhci_find_prev_qh() always 'static inline', just 'inline" leads to trouble.

(jakllsch)

Guard against spurious xbdback_backend_changed() calls which would result
in the block device being opened twice. Fixes port-xen/45158,
although the underlying cause (multiple open of the same device not
properly handled any more) is not fixed.

(bouyer)

Add a comment explaing why a flush workqueue is handled differently from
read/write workqueue requests.

(bouyer)

Several fixes to the continuation engine:
- make sure to enter the continuation loop at splbio(), and add some
  KASSERT() for this.
- When a flush operation is enqueued to the workqueue, make sure the
  continuation loop can't be restarted by a previous workqueue
  completion or an event. We can't restart it at this point because
  the flush even is still recorded as the current I/O.
  For this add a xbdback_co_cache_doflush_wait() which acts as a noop;
  the workqueue callback will restart the loop once the flush is complete.
Should fix "kernel diagnostic assertion xbd_io->xio_mapped == 0" panics
reported by Jeff Rizzo on port-xen@.

(bouyer)

Cleanup spin loop for 2nd cpu.

(kiyohara)

Try to attach 2nd CPU.  But not support options MULTIPROCESSOR yet.

(kiyohara)

Cleanup headers and variables.

(kiyohara)

Not allow interrupt hear.

(kiyohara)

Use BEBOX_REG for mapped to BAT instead of bebox_mb_reg.

(kiyohara)

Use BEBOX_REG for mapped to BAT instead of bebox_mb_reg.

(kiyohara)

Set BeBox Mainboard Registers space to BAT register.

(kiyohara)

Try to ensure no SSE instructions are generated for kernel code,
as is already done on amd64.

(jakllsch)

Remove obsoleted headers.

(kiyohara)

Rename slightly misleading KTHREAD_JOINABLE to KTHREAD_MUSTJOIN.

(rmind)

Trailing whitespace, what trailing whitespace?

(jakllsch)

Convert ppp_list_lock to mutex(9).

(rmind)

PR# kern/30398: panic in ohci_softintr

If the host controller has a TD that the driver doesn't know about, ignore it
instead of calling panic. Patch from Karl Janmar.

(jmcneill)

Replace some wakeup_one(9) uses with mutex(9) or plain wakeup(9).

(rmind)

Add kcpuset(9) - a reworked dynamic CPU set implementation for kernel.
Suitable for use during the early boot.  MD and other implementations
should be replaced with this interface.

Discussed on: tech-kern@

(rmind)

add the -t option that modifies -l behaviour to recursively scan for
child devices and print them in tree format:

$ drvctl -lt usb0
uhub0
  uhub6
    uhidev0
      ukbd0
        wskbd1
    uhidev1
      ums0
        wsmouse1
      uhid0
    ubt0

(jmcneill)

Fix .Xr to membar_ops(3), not membar(9).  Spotted by wiz@.

(rmind)

re-run mknative for hppa.

(mrg)

add an optional argument to the -p flag that lets you extract specific
property values from the command-line:

  $ drvctl -p wd0 disk-info/geometry/cylinders-per-unit
  620181
  $ drvctl -p wd0 device-driver device-unit
  wd
  0
  $ drvctl -p wd0 nonexistent || echo "not found"
  not found

(jmcneill)

after all the recent fixes, let's call this "NetBSD nb2 20110806"

(mrg)

Add serial comma, remove trailing whitespace, bump date for previous.

(wiz)

prop_dictionary_get_dict: last argument is prop_dictionary_t *, not bool *

(jmcneill)

Merge the attributes into the returned characters.

(blymn)

Document the return values for innstr and friends.

(blymn)

Fix a bug that prevented instr and friends returning OK.

(blymn)

flush the output after writing the meta sequence.

(blymn)

Add commented out entries for V7FS and V7FS Endian Independent support.

(tron)

note switch to GCC 4.5 on x86

(mrg)

switch x86 to GCC 4.5.

(mrg)