Some have noticed that the overlapping sched domains code doesn't quite work
as intended (it results in disjoint domains on some machines), and that a top
level, machine spanning domain is needed.  This patch from John Hawkes adds
it to the ia64 code.  This allows processes to run on all CPUs in large
systems, though balancing is limited.  It should go to Linus soon now
otherwise large systems will only have ~16p (depending on topology) usable by
the scheduler.  I sanity checked it on a small system after rediffing John's
original, and he's done some testing on very large systems.
 
 Nick, can you buy off on the sched.c change?  Alternatively, do you want to
 send that fix separately John? Nick did indeed ACK this change, but it isn't
 dependent on this ia64 specific part ... so it's going to be submitted
 separately.
Signed-off-by: John Hawkes <hawkes@sgi.com>
Signed-off-by: Jesse Barnes <jbarnes@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

into intel.com:/data/home/aegl/BK/linux-ia64-release-2.6.10

Patch supplied by Suresh Siddha

This is mainly needed for EM64T platforms and makes sense for ia64 too.
Need of this was broughtup sometime(long time?) back on lkml.
http://www.ussg.iu.edu/hypermail/linux/kernel/0406.3/0112.htmlSigned-off-by: Tony Luck <tony.luck@intel.com>

There is a rare deadlock condition during unwind script creation.  If
build_script() is interrupted in the middle of creating the script, it
holds the script write lock.  If the interrupt handler needs to call
unwind for some failure condition, unwind will try to read the
incomplete script and will deadlock on the script lock.

The fix is to disable interrupts while building the script, so
interrupt handlers never see partial scripts.

Promoting spin_lock_irqsave() from script_new() to find_save_locs()
changes the indentation, so the patch looks bigger than it really is.
Signed-off-by: Keith Owens <kaos@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

window is not zeroed, so the flags should be assigned, not modified.
This can lead to crashes at boot if the IO and Memory resources overlap.

Patch supplied by Matthew Wilcox
Signed-off-by: Tony Luck <tony.luck@intel.com>

Signed-off-by: Arun Sharma <arun.sharma@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

Patch submitted by H. J. Lu

  Gcc 3.4.2 fixed ia64 -mtune=merced regressions on Linux 2.6 kernel:

  http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16278

  I have been using this patch for several months now.
Signed-off-by: Tony Luck <tony.luck@intel.com>

into agluck-lia64.sc.intel.com:/data/home/aegl/BK/linux-ia64-release-2.6.10

Signed-off-by: Patrick Gefre <pfg@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

During peak utilization periods, the first interface on a node has
an inordinately large amount of contention.  This is due to all
cpus starting their scan for an interface at 0.  This patch distributes
that based upon the slice the requesting cpu is attached to.
 
Signed-off-by: Robin Holt
Signed-off-by: Tony Luck <tony.luck@intel.com>

The SN2 Block Transfer Engine occassionally fails to send a notification
that it has completed a transfer to the kernel.  This patch adds a
timeout mechanism which will detect the failure, reset the interface,
and then retry the transfer.
 
Signed-off-by: Robin Holt
Signed-off-by: Tony Luck <tony.luck@intel.com>

Signed-off-by: Patrick Gefre <pfg@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

If all bte interfaces are in use, the current code will leave the
bte pointer set when exiting from the loop trying to locate an
interface to use.  This results in two processes using the same
interface and both trying to free the same one.  With a
premptible kernel, this results in the preempt count getting
off.
Signed-off-by: Robin Holt <holt@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

This includes a fuller definition of the 8250 and compatible bits,
and adding notes where the definition varies between different
chips.

This also fixes the following build error:

drivers/serial/8250.c:185: error: `UART_FCR_R_TRIG_10'

Some serial drivers receive their serial port device information via
the device model.  This unfortunately means that the selected port
may not be available when the console subsystem initialises, so we
must keep trying to register the console after each port is added.

This allows 8250 PCI ports to register with their correct device
structures.

serial8250_register_port()/serial8250_unregister_port() has the
capability of registering ports with their struct device nodes,
which allows sysfs to indicate which tty devices belong to which
hardware devices.

We also add a serial8250 platform device driver in an initial
attempt at PM for ISA ports.  However, I'm leaving out the
platform device for the time being since adding that would cause
potential oops issues.

Convert port autoprobing to set up->capabilities as it discovers
various capabilities of the port.  Warn when the detected
capabilities do not match those in the uart_config table.

It's pointless accessing the LSR value via a pointer all the time -
it prevents the compiler optimising it.  Also, ensure that we
recognise a break sent during a kernel printk correctly.

Add a bit of explaination why we only turn off the transmitter if
we are called from uart_stop().

Add UART_CAP_AFE, and use this to enable TI16C750 flow control,
but only if we have 32 bytes or more of FIFO.

This allows us to adapt the FCR for each port type in a much more
flexible way, and allows us to set the transmit trigger levels.

Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

The patch below adds power management support to the i2c bus.
It adds just two small functions which call down to the devices
power management functions if they are present, so the i2c device
drivers will receive the suspend and resume events.

From: Gerd Knorr <kraxel@bytesex.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

This is my port of the lm87 driver to Linux 2.6. It is based on the
preliminary work of Jeff Oliver. I then significantly improved the code,
added functionality, tested the whole thing on a real motherboard, fixed
a couple remaining bugs, and here we are.

I'll port a number of improvements and fixes back to the 2.4 version of
the driver after lm_sensors 2.8.8 is released (i.e. soon).
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Bus driver for the Samsung S3C2410 SoC onboard I2C controller
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

A number of messages in the i2c-amd756 and i2c-prosavage drivers have a
leading ": " (especially the former). This is a legacy from lm_sensors'
printks of the 2.4 times. This patch cleans them up. While I was there,
I dropped a couple useless white spaces and dots as well.
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

This sets the proper name for busses supported by the i2c-amd756 driver.
So far, all busses were named AMD756 regardless of the real hardware.
Setting the real name is certainly less confusing for the user, and the
sensors-detect script expects this too.
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

This updates the AMD entries i2c/busses/Kconfig in two ways:
* Add missing PCI dependancy.
* Reword the help so that users know exactly what is supported by each
  driver.
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Use msleep_interruptible() instead of schedule_timeout() to
guarantee the task delays as expected. Remove the unnecessary
set_current_state() following the if, as schedule_timeout() [and thus,
mlseep_interruptible()] is guaranteed to return in TASK_RUNNING.
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Following patch fixes the bug introduced by me in VID VRM patch.
Spotted (and later reviewed) by Jean Delvare. This bug is non-fatal,
it8712 will be just treated as it was before my VID VRM patch.

Tested on it8705 and it8712  hardware.
Signed-off-by: Rudolf Marek <r.marek@sh.cvut.cz>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Here comes my fourth (and hopefully last) sysfs interface proposal for
implementing auto-fan control in 2.6. Previous proposals have been
discussed here:

[1] http://archives.andrew.net.au/lm-sensors/msg07517.html
[2] http://archives.andrew.net.au/lm-sensors/msg08049.html
[3] http://archives.andrew.net.au/lm-sensors/msg18008.html

The interface is still made up of two parts: per fan temp channels
selection, and trip points selection. Changes from the third proposal:

pwm[1-*]_enable value 2 is now used to explicitely state the auto pwm
mode. This was proposed by Mark D. Studebaker [4].

[4] http://archives.andrew.net.au/lm-sensors/msg18011.html

Temp channels selection
=======================

Renamed files from fan[1-*]_auto_channels to
pwm[1-*]_auto_channels_temp. The change from fan tp pwm is to match the
recent renaming suggested by Mark M. Hoffman [5]. The "_temp" suffix is
to leave some room for a "_fan" suffix at a later time if new chips
drive auto pwm according to fan speeds instead of temperature.

[5] http://archives.andrew.net.au/lm-sensors/msg18797.html

Trip points
===========

Trip points are now numbered (point1, point2, etc...) instead of named
(_off, _min, _max, _full...). This solves the problem of various chips
having a different number of trip points. The interface is still chip
independent in that it doesn't require chip-specific knowledge to be
used by user-space apps.

The reason for this change is that newer chips tend to have more trip
points. the LM63 has 8, the LM93 has no less than 12. Also, I read in
the LM63 datasheet that ideal pwm vs temperature curve were parabolic in
shape. Seems hard to achieve this if we arbitrarily lock the number of
trip points to 3 ;)

I also introduced an optional hysteresis temperature for trip points.
The LM63 has this. Since it makes full sense I'd expect other chips to
propose this as well.

As before, there are two sets of files, each chip driver picks the one
matching its internal model: trip points are either temperature
channel-dependent (ADM1031...) or pwm channel-dependent (IT87xx...). If
we ever come accross fan speed-driven pwm outputs where trip points are
fan channel-dependent we may have to offer a third set of files. We'll
see when/if this happens.

I hope I have taken everyone's comments and advice into account and we
can make this interface proposal part of the sysfs interface standard
now. I'm sorry it took so long. Comments welcome.
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

I just noticed the other day that the lm90 driver uses an u16 to store
the value of the 8-bit alarms register. This is most probably due to the
fact that I originally copied the lm90 driver from the lm83 driver,
which actually has two 8-bit registers for alarms, and obviously forgot
to change the variable type.
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Make ->ide_dma_begin() functions void and rename them to ->dma_start().
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>