While sending interrupts to a cpu to repeatedly wake a thread, on occasion
that thread will take a full timer tick cycle (4002 usec in my case)
to wakeup.

The problem concerns a race condition in the code around the safe_halt()
call in the default_idle() routine.  Setting 'nohalt' on the kernel
command line causes the long wakeups to disappear.

void
default_idle (void)
{
        local_irq_enable();
        while (!need_resched()) {
-->             if (can_do_pal_halt)
-->                     safe_halt();
                else

A timer tick could arrive between the check for !need_resched and the
actual call to safe_halt() (which does a pal call to PAL_HALT_LIGHT).
By the time the timer tick completes, a thread that might now need to run
could get held up for as long as a timer tick waiting for the halted cpu.

I'm proposing that we disable irq's and check need_resched again before
calling safe_halt().  Does anyone see any problem with this approach?
Signed-off-by: Dimitri Sivanich <sivanich@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

* git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched:
  sched: run_rebalance_domains: s/SCHED_IDLE/CPU_IDLE/
  sched: fix sleeper bonus
  sched: make global code static

Level type interrupts do not need to be resent.  It was also found that
some chipsets get confused in case of the resend.

Mark the ioapic level type interrupts as such to avoid the resend
functionality in the generic irq code.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Level type interrupts are resent by the interrupt hardware when they are
still active at irq_enable().

Suppress the resend mechanism for interrupts marked as level.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Commit 5a43a066: "genirq: Allow fasteoi
handler to retrigger disabled interrupts" was erroneously applied to
handle_level_irq().  This added the irq retrigger / resend functionality
to the level irq handler.

Revert the offending bits.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

rebalance_domains(SCHED_IDLE) looks strange (typo), change it to CPU_IDLE.

the effect of this bug was slightly more agressive idle-balancing on
SMP than intended.
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Peter Ziljstra noticed that the sleeper bonus deduction code
was not properly rate-limited: a task that scheduled more
frequently would get a disproportionately large deduction.
So limit the deduction to delta_exec.
Signed-off-by: Ingo Molnar <mingo@elte.hu>

This patch makes the following needlessly global code static:

- arch_reinit_sched_domains()
- struct attr_sched_mc_power_savings
- struct attr_sched_smt_power_savings
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6: (28 commits)
  ACPI: thermal: add DMI hooks to handle AOpen's broken Award BIOS
  ACPI: thermal: create "thermal.act=" to disable or override active trip point
  ACPI: thermal: create "thermal.nocrt" to disable critical actions
  ACPI: thermal: create "thermal.psv=" to override passive trip points
  ACPI: thermal: expose "thermal.tzp=" to set global polling frequency
  ACPI: thermal: create "thermal.off=1" to disable ACPI thermal support
  ACPI: thinkpad-acpi: fix sysfs paths in documentation
  ACPI: static
  ACPI EC: remove potential deadlock from EC
  ACPI: dock: Send key=value pair instead of plain value
  ACPI: bay: send envp with uevent - fix
  acpi-cpufreq: Fix some x86/x86-64 acpi-cpufreq driver issues
  ACPI: fix "Time Problems with 2.6.23-rc1-gf695baf2"
  ACPI: thinkpad-acpi: change thinkpad-acpi input default and kconfig help
  ACPI: EC: fix run-together printk lines
  ACPI: sbs: remove dead code
  ACPI: EC: acpi_ec_remove(): fix use-after-free
  ACPI: EC: Switch from boot_ec as soon as we find its desc in DSDT.
  ACPI: EC: fix build warning
  ACPI: EC: If ECDT is not found, look up EC in DSDT.
  ...

Commit 3320ad99 broke mmio config space
accesses totally on i386 - it dropped the "reg" offset to the address.

Cc: dean gaudet <dean@arctic.org>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

apply_alternatives uses memcpy() to apply alternatives.  Which has the
unfortunate effect that while applying memcpy alternative to memcpy
itself it tries to overwrite itself with nops - which causes #UD fault
as it overwrites half of an instruction in copy loop, and from this
point on only possible outcome is triplefault and reboot.

So let's overwrite only first two instructions of memcpy - as long as
the main memcpy loop is not in first two bytes it will work fine.
Signed-off-by: Petr Vandrovec <petr@vandrovec.name>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Use DMI to:
1. enable polling (BIOS thermal events are broken)
2. disable active trip points (BIOS fan control is broken)
3. disable passive trip point (BIOS hard-codes it too low)

The actual temperature reading does work,
and with the aid of polling, the critical
trip point should work too.

http://bugzilla.kernel.org/show_bug.cgi?id=8842Signed-off-by: Len Brown <len.brown@intel.com>

thermal.act=-1 disables all active trip points
in all ACPI thermal zones.

thermal.act=C, where C > 0, overrides all lowest temperature
active trip points in all thermal zones to C degrees Celsius.
Raising this trip-point may allow you to keep your system silent
up to a higher temperature.  However, it will not allow you to
raise the lowest temperature trip point above the next higher
trip point (if there is one).  Lowering this trip point may
kick in the fan sooner.

Note that overriding this trip-point will disable any BIOS attempts
to implement hysteresis around the lowest temperature trip point.
This may result in the fan starting and stopping frequently
if temperature frequently crosses C.

WARNING: raising trip points above the manufacturer's defaults
may cause the system to run at higher temperature and shorten
its life.
Signed-off-by: Len Brown <len.brown@intel.com>

thermal.nocrt=1 disables actions on _CRT and _HOT
ACPI thermal zone trip-points.  They will be marked
as <disabled> in /proc/acpi/thermal_zone/*/trip_points.

There are two cases where this option is used:

1. Debugging a hot system crossing valid trip point.

   If your system fan is spinning at full speed,
   be sure that the vent is not clogged with dust.
   Many laptops have very fine thermal fins that are easily blocked.

   Check that the processor fan-sink is properly seated,
   has the proper thermal grease, and is really spinning.

   Check for fan related options in BIOS SETUP.
   Sometimes there is a performance vs quiet option.
   Defaults are generally the most conservative.

   If your fan is not spinning, yet /proc/acpi/fan/
   has files in it, please file a Linux/ACPI bug.

   WARNING: you risk shortening the lifetime of your
   hardware if you use this parameter on a hot system.
   Note that this refers to all system components,
   including the disk drive.

2. Working around a cool system crossing critical
   trip point due to erroneous temperature reading.

   Try again with CONFIG_HWMON=n
   There is known potential for conflict between the
   the hwmon sub-system and the ACPI BIOS.
   If this fixes it, notify lm-sensors@lm-sensors.org
   and linux-acpi@vger.kernel.org

   Otherwise, file a Linux/ACPI bug, or notify
   just linux-acpi@vger.kernel.org.
Signed-off-by: Len Brown <len.brown@intel.com>

"thermal.psv=-1" disables passive trip points
for all ACPI thermal zones.

"thermal.psv=C", where 'C' is degrees Celsius,
overrides all existing passive trip points
for all ACPI thermal zones.

thermal.psv is checked at module load time,
and in response to trip-point change events.

Note that if the system does not deliver thermal zone
temperature change events near the new trip-point,
then it will not be noticed.  To force your custom
trip point to be noticed, you may need to enable polling:
eg. thermal.tzp=3000 invokes polling every 5 minutes.

Note that once passive thermal throttling is invoked,
it has its own internal Thermal Sampling Period (_TSP),
that is unrelated to _TZP.

WARNING: disabling or raising a thermal trip point
may result in increased running temperature and
shorter hardware lifetime on some systems.
Signed-off-by: Len Brown <len.brown@intel.com>

Thermal Zone Polling frequency (_TZP) is an optional ACPI object
recommending the rate that the OS should poll the associated thermal zone.

If _TZP is 0, no polling should be used.
If _TZP is non-zero, then the platform recommends that
the OS poll the thermal zone at the specified rate.
The minimum period is 30 seconds.
The maximum period is 5 minutes.

(note _TZP and thermal.tzp units are in deci-seconds,
 so _TZP = 300 corresponds to 30 seconds)

If _TZP is not present, ACPI 3.0b recommends that the
thermal zone be polled at an "OS provided default frequency".

However, common industry practice is:
1. The BIOS never specifies any _TZP
2. High volume OS's from this century never poll any thermal zones

Ie. The OS depends on the platform's ability to
provoke thermal events when necessary, and
the "OS provided default frequency" is "never":-)

There is a proposal that ACPI 4.0 be updated to reflect
common industry practice -- ie. no _TZP, no polling.

The Linux kernel already follows this practice --
thermal zones are not polled unless _TZP is present and non-zero.

But thermal zone polling is useful as a workaround for systems
which have ACPI thermal control, but have an issue preventing
thermal events.  Indeed, some Linux distributions still
set a non-zero thermal polling frequency for this reason.

But rather than ask the user to write a polling frequency
into all the /proc/acpi/thermal_zone/*/polling_frequency
files, here we simply document and expose the already
existing module parameter to do the same at system level,
to simplify debugging those broken platforms.

Note that thermal.tzp is a module-load time parameter only.
Signed-off-by: Len Brown <len.brown@intel.com>

"thermal.off=1" disables all ACPI thermal support at boot time.

CONFIG_ACPI_THERMAL=n can do this at build time.
"# rmmod thermal" can do this at run time,
as long as thermal is built as a module.

WARNING: On some systems, disabling ACPI thermal support
will cause the system to run hotter and reduce the
lifetime of the hardware.
Signed-off-by: Len Brown <len.brown@intel.com>

The documentation used "thinkpad-acpi" to refer to the directories in
sysfs, while it should have been using "thinkpad_acpi".  Thanks to Hugh
Dickins for the error report.

I wish I could just call the module and everything else by the proper
name with the "-", instead of using these ugly translations to "_".
Signed-off-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Len Brown <len.brown@intel.com>

Make the needlessly global "acpi_event_seqnum" static.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Len Brown <len.brown@intel.com>

Signed-off-by: Alexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Len Brown <len.brown@intel.com>

Send key=value pair along with the uevent instead of a plain value so that
userspace (udev) can handle it like common environment variables.
Signed-off-by: Holger Macht <hmacht@suse.de>
Acked-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Cc: Stephan Berberig <s.berberig@arcor.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Len Brown <len.brown@intel.com>

There must not be a new-line character in the uevent.  Otherwise, udev gets
confused.  Thanks to Kay Sievers for pointing it out.
Signed-off-by: Stephan Berberig <s.berberig@arcor.de>
Cc: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Len Brown <len.brown@intel.com>

* 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6:
  [S390] monwriter: Serialization bug for multithreaded applications.
  [S390] vmur: diag14 only works with buffers below 2GB
  [S390] vmur: add "top of queue" sanity check for reader open
  [S390] vmur: reject open on z/VM reader files with status HOLD
  [S390] vmur: use DECLARE_COMPLETION_ONSTACK to keep lockdep happy
  [S390] vmur: allocate single record buffers instead of one big data buffer
  [S390] remove DEFAULT_MIGRATION_COST
  [S390] qdio: make sure data structures are correctly aligned.
  [S390] hypfs: implement show_options
  [S390] cio: avoid memory leak on error in css_alloc_subchannel().

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
  [POWERPC] Fix size check for hugetlbfs
  [POWERPC] Fix initialization and usage of dma_mask
  [POWERPC] Fix more section mismatches in head_64.S
  [POWERPC] Revert "[POWERPC] Add 'mdio' to bus scan id list for platforms with QE UEC"
  [POWERPC] PS3: Update ps3_defconfig
  [POWERPC] PS3: Remove text saying PS3 support is incomplete
  [POWERPC] PS3: Fix storage probe logic
  [POWERPC] cell: Move SPU affinity init to spu_management_of_ops
  [POWERPC] Fix potential duplicate entry in SLB shadow buffer

* 'async-tx-fixes-for-linus' of git://lost.foo-projects.org/~dwillia2/git/iop:
  async_tx: update MAINTAINERS for async_tx and iop-adma

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh/ocfs2:
  ocfs2: set non-default s_time_gran during mount
  ocfs2: Retry sendpage() if it returns EAGAIN
  ocfs2: Fix rename/extend race
  [2.6 patch] ocfs2_insert_extent(): remove dead code
  ocfs2: Fix max offset calculations
  ocfs2: check ia_size limits in setattr
  ocfs2: Fix some casting errors related to file writes
  ocfs2: use s_maxbytes directly in ocfs2_change_file_space()
  ocfs2: Restrict inode changes in ocfs2_update_inode_atime()

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  BLOCK: Hide the contents of linux/bio.h if CONFIG_BLOCK=n
  sysace: HDIO_GETGEO has it's own method for ages
  drivers/block/cpqarray.c: better error handling and kmalloc + memset conversion to k[cz]alloc
  drivers/block/cciss.c: kmalloc + memset conversion to kzalloc
  Clean up duplicate includes in drivers/block/
  Fix remap handling by blktrace
  [PATCH] remove mm/filemap.c:file_send_actor()

* git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched:
  sched debug: dont print kernel address in /proc/sched_debug
  sched: fix typo in the FAIR_GROUP_SCHED branch
  sched: improve rq-clock overflow logic

The new percpu code has apparently broken the doublefault handler
when CONFIG_DEBUG_SPINLOCK is set. Doublefault is handled by
a hardware task, making the check

        SPIN_BUG_ON(lock->owner == current, lock, "recursion");

fault because it uses the FS register to access the percpu data
for current, and that register is zero in the new TSS. (The trace
I saw was on 2.6.20 where it was GS, but it looks like this will
still happen with FS on 2.6.22.)

Initializing FS in the doublefault_tss should fix it.

AK: Also fix broken ptr_ok() and turn printks into KERN_EMERG
AK: And add a PANIC prefix to make clear the system will hang
AK: (e.g. x86-64 will recover)
Signed-off-by: Chuck Ebbert <cebbert@redhat.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Fix

WARNING: vmlinux.o(.text+0x183): Section mismatch: reference to .init.text:start_kernel (between 'is386' and 'check_x87')
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>