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>

add extern for pci_root_ops
delete our version of the pci_root_ops code
make pci_root_ops non static

Ack-by: Matthew Wilcox
Ack-by: Jesse Barnes
Ack-by: Grant Grundler
Signed-off-by: Patrick Gefre <pfg@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

Fix a possible memory leak.  Change the name of the 'new_sn_irq_info'
variable to 'tmp_sn_irq_info' since we're not really allocating a new
kernel sn_irq_info, we just getting new target information from the
PROM and replacing it in the kernel's sn_irq_info for the given irq
Signed-off-by: Patrick Gefre <pfg@sgi.com>
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>

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

Changes inside CONFIG_IA64_GENERIC CONFIG_IA64_SGI_SN2
Need to include sn/io.h for generic builds (sn_+pci_set_vchan def)
Remove rrb alloc - now done in PROM
Remove endian setting - now done in PROM
Change arg 2 for sn_pci_set_vchan()

Acked-by: Jes Sorenson
Signed-off-by: Patrick Gefre <pfg@sgi.com>
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>

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

Fixes a data corruption issue. Because of a typo in the driver, IO packets
were wrongly shared by the IOCTL path. This caused the whole IO command
to be replaced by an incoming IOCTL command.

ACPI still explodes on my old PII and stops it booting.

Anyway, it is oopsing in drivers/acpi/scan.c line 207 where element
(which is NULL) gets dereferenced.

The ACPI bios on this thing has always seemed to be pretty broken, but
this at least allows the 'power' button to continue to work (the only
reason why I want ACPI).

As spotted by one of our Fedora users, we sometimes
oops during shutdown (http://www.roberthancock.com/kerneloops.png)
because disable_IO_APIC() wants to call find_isa_irq_pin(),
which we threw away during init.
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

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

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

"lapic" is available to force enabling the LAPIC
in the event you know more than your BIOS vendor.
http://bugzilla.kernel.org/show_bug.cgi?id=3238

This is made possible by the previous resource allocation ordering
patch.

Now that we run the pre-existing PCI resource discovery _before_
ACPI resources have been added, we should not "insert" the resources
into any existing tree. If they clash with existing resources, we
should re-allocate them (later, when we have the full resource map).

This re-orders the PCI and ACPI IO resource assignment as suggested by
Linus.

With this patch, now the sequence of reserving resources is:
 1. PCI claim BAR
 2. ACPI reserve motherboard resources
 3. PNP reserve motherboard resources
 4. PCI allocate resources for uninitialized PCI devices

This way the kernel allocates new PCI resources after it has full
knowledge of the resource state, and at the same time allows ACPI and
PnP to be run _after_ we've filled in our knowledge about pre-allocated
resources. 

The way it is done is to make the last phase of the original PCI assign
resources code to be an 'fs_initcall', along with the ACPI and PnP
initializations.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

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

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

From: Jens Axboe <axboe@suse.de>

The blacklist stuff is broken. When set_using_dma() calls into
->ide_dma_check(), it returns __ide_dma_off() for a blacklisted drive.
This of course succeeds, returning success to the caller of ->ide_dma_check().
Not so good... It then uncondtionally calls ->ide_dma_on(), which turns on
DMA for the drive.

This moves the check to __ide_dma_on() so we also catch the buggy
->ide_dma_check() defined by various chipset drivers.
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>

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

Doing so apparently breaks every Dell on Earth.
http://bugzilla.kernel.org/show_bug.cgi?id=3238

FB_NEOMAGIC needs FB_MODE_HELPER (for vesa_modes).

Reported by Michel Angelo da Silva Pereira.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

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

Patch from Sascha Hauer

This patch supports the Hynix developer boards h7201 and h7202 for the
HMS30C7201 and HMS30C7202 CPUs.

Changes to previous patch:
- reorganized file structure. We now have common.c which shares stuff
  common to all h720x processors and cpu-h720[12].c files for the cpu
  specific stuff
- fixed timer handling for timers 1/2
- cleaned up h720x register definitions and splitted in files specific
  to each cpu and one generic part

Signed-off-by: Thomas Gleixner
Signed-off-by: Robert Schwebel
Signed-off-by: Sascha Hauer

This fixes the "Ignoring changed section attributes" warning by commenting
out the attributes appended by gcc.

The previous 'fix' created a section '.text,#alloc' (literally), which I
guess was not intended.
Signed-off-by: Clemens Buchacher <drizzd@aon.at>
Signed-off-by: William Irwin <wli@holomorphy.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Since get_cpu_ptr() went away, we probably shouldn't tease developers by
telling them to use it in the comments.
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

The cpu mask fix-ups in 2.6.8 broke SMP kernels booting on a DP264.
Instead of not setting the DIM for cpus that did not exit, the patch
inadvertently doesn't set the DIM for CPUs that do exist.  Thus no device
interrupts get to the cpu.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

qeth network driver changes:
 - Unlock queue in qeth_do_sent_packet if there is no empty buffer
   in packing state.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

DCSS block device driver changes:
 - Add module/kernel parameter for loading segments.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

z/VM monitor stream changes:
 - Reduce stack usage of appldata_get_mem_data.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

dasd driver changes:
 - Free memory returned by read_conf_data.
 - Add warning message if use_diag is specified with a non-CMS
   formatted device.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Common i/o layer changes:
 - Fix error handling in io_subchannel_register.
 - Fix __MAX_SUBCHANNELS limit checking.
 - Clear slow_subchannel structure after kmalloc.
 - Update ssd_info if a different device appears at an already known
   subchannel to get the correct set of chpids.
 - Avoid struct initializers to reduce stack usage of ccwgroup_create,
   readall_cmb and io_subchannel_recog,
 - Setup fields in pmcw in each retry because msch might fail and a stsch
   overwrites them.
 - Prevent irq_exit() in cio_tpi from calling do_softirq by adding a
   local_bh_disable/__local_bh_enable pair.
 - Retry sense id after receiving an unsolicited interrupt.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>