irq: refactor and clean up the free_irq() code flow

Impact: cleanup

- separate out the loop from the actual freeing logic, this wins us
  two indentation levels allowing a number of followup prettifications

- turn the WARN_ON() into a more informative WARN().

- clean up the comments and the code flow some more

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

irq: clean up manage.c

- make printk message git-greppable
- fix a few style details

Signed-off-by: Ingo Molnar <mingo@elte.hu>

irq: use GFP_KERNEL for action allocation in request_irq()

request_irq() calls into proc code via __setup_irq() which is not safe
in an atomic context, so request_irq() can itself use the more
reliable GFP_KERNEL allocation for the action descriptor.

Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

irq: provide debug_poll_all_shared_irqs() method under CONFIG_DEBUG_SHIRQ

Provide a shared interrupt debug facility under CONFIG_DEBUG_SHIRQ:
it uses the existing irqpoll facilities to iterate through all
registered interrupt handlers and call those which can handle shared
IRQ lines.

This can be handy for suspend/resume debugging: if we call this function
early during resume we can trigger crashes in those drivers which have
incorrect assumptions about when exactly their ISRs will be called
during suspend/resume.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

Merge branch 'linus' into irq/genirq

Merge branch 'sched-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  sched: sched_slice() fixlet
  sched: fix update_min_vruntime
  sched: SCHED_OTHER vs SCHED_IDLE isolation
  sched: SCHED_IDLE weight change
  sched: fix bandwidth validation for UID grouping
  Revert "sched: improve preempt debugging"

Merge git://git./linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  sparc64: Fix UP build failure.

Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (95 commits)
  b44: GFP_DMA skb should not escape from driver
  korina: do not use IRQF_SHARED with IRQF_DISABLED
  korina: do not stop queue here
  korina: fix handling tx_chain_tail
  korina: do tx at the right position
  korina: do schedule napi after testing for it
  korina: rework korina_rx() for use with napi
  korina: disable napi on close and restart
  korina: reset resource buffer size to 1536
  korina: fix usage of driver_data
  bnx2x: First slow path interrupt race
  bnx2x: MTU Filter
  bnx2x: Indirection table initialization index
  bnx2x: Missing brackets
  bnx2x: Fixing the doorbell size
  bnx2x: Endianness issues
  bnx2x: VLAN tagged packets without VLAN offload
  bnx2x: Protecting the link change indication
  bnx2x: Flow control updated before reporting the link
  bnx2x: Missing mask when calculating flow control
  ...

Merge branch 'hwmon-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6

* 'hwmon-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6:
  hwmon: (abituguru3) Fix CONFIG_DMI=n fallback to probe
  hwmon: (abituguru3) Enable DMI probing feature on IN9 32X MAX
  hwmon: (abituguru3) Match partial DMI board name strings
  hwmon: Add a driver for the ADT7475 hardware monitoring chip
  hwmon: (k8temp) Fix temperature reporting for (most) K8 RevG CPUs
  hwmon: (k8temp) Fix wrong sensor selection for AMD K8 RevF/RevG CPUs
  hwmon: (k8temp) Warn about fam F rev F errata

9p: disallow RDMA if RDMA CM isn't available

If INET=y and INFINIBAND=y, but IPV6=m then INFINIBAND_ADDR_TRANS is set
to n and the RDMA CM functions rdma_connect() et al are not built.
However, the current config dependencies allow NET_9P_RDMA to be selected
in this, which leads to a build failure.  Fix this by adding a dependency
on INFINIBAND_ADDR_TRANS to disallow NET_9P_RDMA in this case.

Reported-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Tested-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

ext2: also update the inode on disk when dir is IS_DIRSYNC

We used to just write changed page for IS_DIRSYNC inodes.  But we also
have to update the directory inode itself just for the case that we've
allocated a new block and changed i_size.

[akpm@linux-foundation.org: still sync the data page]
Signed-off-by: Jan Kara <jack@suse.cz>
Tested-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/usb/core/hub.c: fix CONFIG_USB_OTG=y build

Carry out the PM-routine interface change in the USB OTG pathway.  This
was omitted from the earlier interface-change patch by mistake.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Cc: Greg KH <greg@kroah.com>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Felipe Balbi <felipe.balbi@nokia.com>
Cc: Tony Lindgren <tony@atomide.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

include of <linux/types.h> is preferred over <asm/types.h>

Impact: fix 15 make headers_check warnings:

include of <linux/types.h> is preferred over <asm/types.h>

Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

memcg: fix a race when setting memory.swappiness

(suppose: memcg->use_hierarchy == 0 and memcg->swappiness == 60)

echo 10 > /memcg/0/swappiness   |
  mem_cgroup_swappiness_write() |
    ...                         | echo 1 > /memcg/0/use_hierarchy
                                | mkdir /mnt/0/1
                                |   sub_memcg->swappiness = 60;
    memcg->swappiness = 10;     |

In the above scenario, we end up having 2 different swappiness
values in a single hierarchy.

We should hold cgroup_lock() when cheking cgrp->children list.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Paul Menage <menage@google.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

memcg: fix section mismatch

At system boot when creating the top cgroup, mem_cgroup_create() calls
enable_swap_cgroup() which is marked as __init, so mark
mem_cgroup_create() as __ref to avoid false section mismatch warning.

Reported-by: Rakib Mullick <rakib.mullick@gmail.com>
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by; KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

hp_accel: do not call ACPI from invalid context

The LED on HP notebooks is connected through ACPI.  That unfortunately
means that it needs to be delayed by using schedule_work() to avoid
calling the ACPI interpreter from an invalid context.

[akpm@linux-foundation.org: use flush_work() rather than sort-of reimplementing it]
Signed-off-by: Pavel Machek <pavel@suse.cz>
Cc: Éric Piel <eric.piel@tremplin-utc.net>
Cc: Len Brown <lenb@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lis3: fix documentation to fit into 80 columns

Fix lis3 documentation to fit into 80 columns.

Signed-off-by: Pavel Machek <pavel@suse.cz>
Cc: Éric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lis3lv02d: merge with leds hp disk

Move the second part of the HP laptop disk protection functionality (a red
led) to the same driver.  From a purely Linux developer's point of view,
the led and the accelerometer have nothing related.  However, they
correspond to the same ACPI functionality, and so will always be used
together, moreover as they share the same ACPI PNP alias, there is no
other simple to allow to have same loaded at the same time if they are not
in the same module.  Also make it requires the led class to compile and
update the Kconfig text.

Signed-off-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Eric Piel <eric.piel@tremplin-utc.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/idr.c: use kmem_cache_zalloc() for the idr_layer cache

David points out that the idr_remove_all() function returns unused slabs
to the kmem cache, but needs to zero them first or else they will be
uninitialized upon next use.  This causes crashes which have been observed
in the firewire subsystem.

He fixed this by zeroing the object before freeing it in idr_remove_all().

But we agree that simply removing the constructor and zeroing the object
at allocation time is simpler than relying upon slab constructor machinery
and might even be faster.

This problem was introduced by "idr: make idr_remove rcu-safe" (commit
cf481c20c476ad2c0febdace9ce23f5a4db19582), which was first released in
2.6.27.

There are no known codesites which trigger this bug in 2.6.27 or 2.6.28.
The post-2.6.28 firewire changes are the only known triggerer.

There might of course be not-yet-discovered triggerers in 2.6.27 and
2.6.28, and there might be out-of-tree triggerers which are added to those
kernel versions.  I'll let the -stable guys decide whether they want to
backport this fix.

Reported-by: David Moore <dcm@acm.org>
Cc: Stefan Richter <stefanr@s5r6.in-berlin.de>
Cc: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Paul E. McKenney <paulmck@us.ibm.com>
Cc: Manfred Spraul <manfred@colorfullife.com>
Cc: Kristian Hgsberg <krh@redhat.com>
Acked-by: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

alpha: make pte_alloc_one_kernel() inline

As it's just a single call to __get_free_page().

Signed-off-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

alpha: .gitignore vmlinux.lds

Signed-off-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

alpha: fix RTC on marvel

Unlike other alphas, marvel doesn't have real PC-style CMOS clock hardware
- RTC accesses are emulated via PAL calls.  Unfortunately, for unknown
reason these calls work only on CPU #0.  So current implementation for
arbitrary CPU makes CMOS_READ/WRITE to be executed on CPU #0 via IPI.
However, for obvious reason this doesn't work with standard
get/set_rtc_time() functions, where a bunch of CMOS accesses is done with
disabled interrupts.

Solved by making the IPI calls for entire get/set_rtc_time() functions,
not for individual CMOS accesses.  Which is also a lot more effective
performance-wise.

The patch is largely based on the code from Jay Estabrook.
My changes:
- tweak asm-generic/rtc.h by adding a couple of #defines to
  avoid a massive code duplication in arch/alpha/include/asm/rtc.h;
- sys_marvel.c: fix get/set_rtc_time() return values (Jay's FIXMEs).

NOTE: this fixes *only* LIB_RTC drivers.  Legacy (CONFIG_RTC) driver
wont't work on marvel.  Actually I think that we should just disable
CONFIG_RTC on alpha (maybe in 2.6.30?), like most other arches - AFAIK,
all modern distributions use LIB_RTC anyway.

Signed-off-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

alpha: nautilus - fix hang on boot

Recently introduced generic pci_common_swizzle() relies on bus->self
being NULL for the root PCI bus. But on nautilus bus->self points to
the host bridge device, which is necessary as we do a root bus sizing
on this system. As a result, pci_common_swizzle() loops infinitely.
This worked until 2.6.29-rc1 because the alpha-specific swizzle routine
checked for bus->parent == NULL (instead of bus->self).

Fixed by clearing bus->self after bus sizing is done.

Signed-off-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

alpha: nautilus - fix compile failure with gcc-4.3

init_srm_irq() deals with irq's #16 and above, but size of irq_desc
array on nautilus and some other system types is 16. So gcc-4.3
complains that "array subscript is above array bounds", even though
this function is never called on those systems.

This adds a check for NR_IRQS <= 16, which effectively optimizes
init_srm_irq() code away on problematic platforms.

Thanks to Daniel Drake <dsd@gentoo.org> for detailed analysis
of the problem.

Signed-off-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

revert "mm: vmalloc use mutex for purge"

Revert commit e97a630eb0f5b8b380fd67504de6cedebb489003 ("mm: vmalloc use
mutex for purge")

Bryan Donlan reports:

: After testing 2.6.29-rc1 on xen-x86 with a btrfs root filesystem, I
: got the OOPS quoted below and a hard freeze shortly after boot.
: Boot messages and config are attached.
:
: ------------[ cut here ]------------
: Kernel BUG at c05ef80d [verbose debug info unavailable]
: invalid opcode: 0000 [#1] SMP
: last sysfs file: /sys/block/xvdc/size
: Modules linked in:
:
: Pid: 0, comm: swapper Not tainted (2.6.29-rc1 #6)
: EIP: 0061:[<c05ef80d>] EFLAGS: 00010087 CPU: 2
: EIP is at schedule+0x7cd/0x950
: EAX: d5aeca80 EBX: 00000002 ECX: 00000000 EDX: d4cb9a40
: ESI: c12f5600 EDI: d4cb9a40 EBP: d6033fa4 ESP: d6033ef4
:  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0069
: Process swapper (pid: 0, ti=d6032000 task=d6020b70 task.ti=d6032000)
: Stack:
:  000d85bc 00000000 000186a0 00000000 0dd11410 c0105417 c12efe00 0dc367c3
:  00000011 c0105d46 d5a5d310 deadbeef d4cb9a40 c07cc600 c05f1340 c12e0060
:  deadbeef d6020b70 d6020d08 00000002 c014377d 00000000 c12f5600 00002c22
: Call Trace:
:  [<c0105417>] xen_force_evtchn_callback+0x17/0x30
:  [<c0105d46>] check_events+0x8/0x12
:  [<c05f1340>] _spin_unlock_irqrestore+0x20/0x40
:  [<c014377d>] hrtimer_start_range_ns+0x12d/0x2e0
:  [<c014c4f6>] tick_nohz_restart_sched_tick+0x146/0x160
:  [<c0107485>] cpu_idle+0xa5/0xc0

and bisected it to this commit.

Let's remove it now while we have a think about the problem.

Reported-by: Bryan Donlan <bdonlan@gmail.com>
Tested-by: Christophe Saout <christophe@saout.de>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

memcg: make oom less frequently

In previous implementation, mem_cgroup_try_charge checked the return
value of mem_cgroup_try_to_free_pages, and just retried if some pages
had been reclaimed.
But now, try_charge(and mem_cgroup_hierarchical_reclaim called from it)
only checks whether the usage is less than the limit.

This patch tries to change the behavior as before to cause oom less
frequently.

Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Acked-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

memcg: fix hierarchical reclaim

If root_mem has no children, last_scaned_child is set to root_mem itself.
But after some children added to root_mem, mem_cgroup_get_next_node can
mem_cgroup_put the root_mem although root_mem has not been mem_cgroup_get.

This patch fixes this behavior by:

- Set last_scanned_child to NULL if root_mem has no children or DFS
  search has returned to root_mem itself(root_mem is not a "child" of
  root_mem).  Make mem_cgroup_get_first_node return root_mem in this case.
   There are no mem_cgroup_get/put for root_mem.

- Rename mem_cgroup_get_next_node to __mem_cgroup_get_next_node, and
  mem_cgroup_get_first_node to mem_cgroup_get_next_node.  Make
  mem_cgroup_hierarchical_reclaim call only new mem_cgroup_get_next_node.

Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

memcg: fix error path of mem_cgroup_move_parent

There is a bug in error path of mem_cgroup_move_parent.

Extra refcnt got from try_charge should be dropped, and usages incremented
by try_charge should be decremented in both error paths:

    A: failure at get_page_unless_zero
    B: failure at isolate_lru_page

This bug makes this parent directory unremovable.

In case of A, rmdir doesn't return, because res.usage doesn't go down to 0
at mem_cgroup_force_empty even after all the pc in lru are removed.

In case of B, rmdir fails and returns -EBUSY, because it has extra ref
counts even after res.usage goes down to 0.

Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

memcg: fix mem_cgroup_get_reclaim_stat_from_page

In case of swapin, a new page is added to lru before it is charged,
so page->pc->mem_cgroup points to NULL or last mem_cgroup the page
was charged before.

In the latter case, if the mem_cgroup has already freed by rmdir,
the area pointed to by page->pc->mem_cgroup may have invalid data.

Actually, I saw general protection fault.

    general protection fault: 0000 [#1] SMP
    last sysfs file: /sys/devices/system/cpu/cpu15/cache/index1/shared_cpu_map
    CPU 4
    Modules linked in: ipt_REJECT xt_tcpudp iptable_filter ip_tables x_tables bridge stp ipv6 autofs4 hidp rfcomm l2cap bluetooth sunrpc dm_mirror dm_region_hash dm_log dm_multipath dm_mod rfkill input_polldev sbs sbshc battery ac lp sg ide_cd_mod cdrom button serio_raw acpi_memhotplug parport_pc e1000 rtc_cmos parport rtc_core rtc_lib i2c_i801 i2c_core shpchp pcspkr ata_piix libata megaraid_mbox megaraid_mm sd_mod scsi_mod ext3 jbd ehci_hcd ohci_hcd uhci_hcd [last unloaded: microcode]
    Pid: 26038, comm: page01 Tainted: G        W  2.6.28-rc9-mm1-mmotm-2008-12-22-16-14-f2ab3dea #1
    RIP: 0010:[<ffffffff8028e710>]  [<ffffffff8028e710>] update_page_reclaim_stat+0x2f/0x42
    RSP: 0000:ffff8801ee457da8  EFLAGS: 00010002
    RAX: 32353438312021c8 RBX: 0000000000000000 RCX: 32353438312021c8
    RDX: 0000000000000000 RSI: ffff8800cb0b1000 RDI: ffff8801164d1d28
    RBP: ffff880110002cb8 R08: ffff88010f2eae23 R09: 0000000000000001
    R10: ffff8800bc514b00 R11: ffff880110002c00 R12: 0000000000000000
    R13: ffff88000f484100 R14: 0000000000000003 R15: 00000000001200d2
    FS:  00007f8a261726f0(0000) GS:ffff88010f2eaa80(0000) knlGS:0000000000000000
    CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
    CR2: 00007f8a25d22000 CR3: 00000001ef18c000 CR4: 00000000000006e0
    DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
    DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
    Process page01 (pid: 26038, threadinfo ffff8801ee456000, task ffff8800b585b960)
    Stack:
     ffffe200071ee568 ffff880110001f00 0000000000000000 ffffffff8028ea17
     ffff88000f484100 0000000000000000 0000000000000020 00007f8a25d22000
     ffff8800bc514b00 ffffffff8028ec34 0000000000000000 0000000000016fd8
    Call Trace:
     [<ffffffff8028ea17>] ? ____pagevec_lru_add+0xc1/0x13c
     [<ffffffff8028ec34>] ? drain_cpu_pagevecs+0x36/0x89
     [<ffffffff802a4f8c>] ? swapin_readahead+0x78/0x98
     [<ffffffff8029a37a>] ? handle_mm_fault+0x3d9/0x741
     [<ffffffff804da654>] ? do_page_fault+0x3ce/0x78c
     [<ffffffff804d7a42>] ? trace_hardirqs_off_thunk+0x3a/0x3c
     [<ffffffff804d860f>] ? page_fault+0x1f/0x30
    Code: cc 55 48 8d af b8 0d 00 00 48 89 f7 53 89 d3 e8 39 85 02 00 48 63 d3 48 ff 44 d5 10 45 85 e4 74 05 48 ff 44 d5 00 48 85 c0 74 0e <48> ff 44 d0 10 45 85 e4 74 04 48 ff 04 d0 5b 5d 41 5c c3 41 54
    RIP  [<ffffffff8028e710>] update_page_reclaim_stat+0x2f/0x42
     RSP <ffff8801ee457da8>

Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

hwmon: applesmc: fix light sensor readings on newer MacBooks

The light sensors ALV0 and ALV1 on newer MacBooks (early 2008 and later)
changed to report 10 bytes instead the earlier 6, and the sensor encoding
subsequently changed.  As a result, the reported light sensors readings
are much too low.

Via experiments leading up to this patch, it seems only the ALV0 is
reporting data, and the most useful value therein is a 10-bit big-endian
value at offset 6.  This suggests that a new protocol was added as a
backward-compatible replacement on top of the old one.

This patch makes applesmc report the improved light sensor reading for the
new machines, on a scale in conformance with earlier ones.

Signed-off-by: Alex Murray <murray.alex@gmail.com>
Signed-off-by: Henrik Rydberg <rydberg@euromail.se>
Cc: Nicolas Boichat <nicolas@boichat.ch>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: version: 0.27

Signed-off-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: struct seq_operations should normally be const

In the general use case struct seq_operations should be a const object.
Check for and warn where it is not.

Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: if should not continue a preceeding brace

We should not be continuing a braced section with an if, for example:

if (...) {
} if (...) {
}

Detect this and suggest adding a newline.

Signed-off-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: allow parentheses on return handle array values

When we allow return to have surrounding parentheses when containing
comparison operators we are not correctly handling the case where the
values contain array sufffixes.  Squash them.

Signed-off-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: type/cast spacing should not check prefix spacing

We should not be complaining about the prefix spacing for types and casts.
 We are triggering here because the check for spacing between '*'s is
overly loose.  Tighten this up.

Signed-off-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

checkpatch: handle missing #if open in context

If the #if opening statement is not in the context then the context stack
can be empty.  Handle this by ensuring there is always a blank entry in
the stack.

Signed-off-by: Andy Whitcroft <apw@canonical.com>
Tested-by: Dhaval Giani <dhaval@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

nbd: do not allow two clients at the same time

Two nbd-clients at same time are bad idea, and cause WARN_ON from nbd in
2.6.28-rc7 from sysfs_add_one.  This simply prevents that from happening.

To reproduce:

 cat /dev/zero | head -c 10000000 > /tmp/delme.fstest.fs
 nbd-server 9100 -l /anyone.can.connect > /tmp/delme.fstest.fs &
 sleep 1
 nbd-client localhost 9100 /dev/nd0 &
 nbd-client localhost 9100 /dev/nd0 &

Signed-off-by: Pavel Machek <pavel@suse.cz>
Acked-by: Paul Clements <paul.clements@steeleye.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

btrfs & squashfs: Move btrfs and squashfsto's magic number to <linux/magic.h>

Use the standard magic.h for btrfs and squashfs.

Signed-off-by: Qinghuang Feng <qhfeng.kernel@gmail.com>
Cc: Phillip Lougher <phillip@lougher.demon.co.uk>
Cc: Chris Mason <chris.mason@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

MAINTAINERS: add entry for freezer

Now that people are using freezer for non-suspend/hibernation stuff, it
should have separate maintainers entry so that it is easier to find.

Signed-off-by: Pavel Machek <pavel@suse.cz>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

resources: fix parameter name and kernel-doc

Fix __request_region() parameter kernel-doc notation and parameter name:

Warning(linux-2.6.28-git10//kernel/resource.c:627): No description found for parameter 'flags'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

jbd: fix missing kernel-doc

Fix jbd header file kernel-doc notation:

Warning(linux-2.6.28-git13//include/linux/jbd.h:823): No description found for parameter 'j_average_commit_time'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Documentation/accounting/getdelays.c: fix endless loop

When no option is passed to getdelays it just hangs, waiting
for a reply which will never come.

This patch prints usage() when no output marker is specified.

Signed-off-by: Marcus Meissner <meissner@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

idr: fix wrong kernel-doc

idr_get_new_above() and ida_get_new_above() return an id in the range of
@staring_id ... 0x7fffffff, not 0 ... 0x7fffffff.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

cgroups: consolidate cgroup documents

Move Documentation/cpusets.txt and Documentation/controllers/* to
Documentation/cgroups/

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Acked-by: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

cgroups: clean up Kconfig

- move CONFIG_PROC_PID_CPUSET into cgroup menu
- move MM_OWNER to the bottom for better menu indent
- fix typos
- use tabs not spaces

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: Paul Menage <menage@google.com>
Acked-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

synclink_gt: enable RI interrupt

- Enable ring indicator interrupt.

- Remove vendor specific CVS version tags.

Signed-off-by: Paul Fulghum <paulkf@microgate.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sgi-xp: eliminate false detection of no heartbeat

After XPC has been up and running on multiple partitions for any length of
time, if XPC on one of the partitions is stopped and restarted (either by
a rmmod/insmod or a system restart), it is possible for the XPCs running
on the other partitions to falsely detect a lack of heartbeat from the XPC
that was just restarted.  This false detection will occur if the restarted
XPC comes up within the five-seconds preceding one of the other XPC's
heartbeat check (which occurs once every twenty seconds).

The detection of no heartbeat results in the detecting XPC deactivating
from the just restarted XPC.  The only remedy is to restart one of the
XPCs and hope that one doesn't hit this five-second window on any of the
other partitions.

Signed-off-by: Dean Nelson <dcn@sgi.com>
Signed-off-by: Robin Holt <holt@sgi.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

rtc: rtc-twl4030 don't mask alarm interrupts on shutdown

A triggering RTC alarm should be able to power on a device that has been
powered off. This patch enables that on twl4030 by not masking the alarm
interrupt at shutdown.

Signed-off-by: Matti Halme <matti.halme@nokia.com>
Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Acked-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Tony Lindgren <tony@atomide.com>
Cc: Samuel Ortiz <sameo@openedhand.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

rtc-pxa: fix build failure

Fix these build errors:

  CC      drivers/rtc/rtc-pxa.o
drivers/rtc/rtc-pxa.c: In function `pxa_rtc_init':
drivers/rtc/rtc-pxa.c:472: error: implicit declaration of function `cpu_is_pxa27x'
drivers/rtc/rtc-pxa.c:472: error: implicit declaration of function `cpu_is_pxa3xx'

Signed-off-by: Antonio Ospite <ao2@openezx.org>
Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

rtc: tw4030 add alarm/update interfaces

- implement alarm_irq_enable
- return correct error code when registering fails

[dbrownell@users.sourceforge.net: build fixes, force 1/sec irqs]
Signed-off-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Tony Lindgren <tony@atomide.com>
Cc: Samuel Ortiz <sameo@openedhand.com>
Cc: rtc-linux@googlegroups.com
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sysrq: add commentary on why we use the console loglevel over using KERN_EMERG

Add an explanitory comment as to why we modify the kernel console loglevel
rather than simply moving sysrq messages to KERN_EMERG level.

Signed-off-by: Andy Whitcroft <apw@canonical.com>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sysrq documentation: document why the command header only is shown

Document the interactions between loglevel and the sysrq output.  Also
document how to work round it should output be required on the console.

Signed-off-by: Andy Whitcroft <apw@canonical.com>
Cc: Martin Mares <mj@ucw.cz>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sysrq documentation: remove the redundant updated date

git is maintaining the last update time much more accuratly than the
internal update time.  Remove it.

Signed-off-by: Andy Whitcroft <apw@canonical.com>
Cc: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

edac: add MAINTAINERS entry for i5400 EDAC driver

i5400 EDAC driver were added upstream by those changesets:

 - 920c8df6ac678fdb8c49a6ce2e47a98e62757d77 "edac: driver for i5400 MCH (Seaburg)"
 - 8375d4909aee4c18798f373ecf24a79f040f75fc "edac: driver for i5400 MCH (update)"

Update MAINTAINERS entry for this file to correspond to the driver
maintainer.

Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
Cc: Ben Woodard <woodard@redhat.com>
Cc: Doug Thompson <norsk5@yahoo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

alpha: fix vmalloc breakage

On alpha, we have to map some stuff in the VMALLOC space very early in the
boot process (to make SRM console callbacks work and so on, see
arch/alpha/mm/init.c).  For old VM allocator, we just manually placed a
vm_struct onto the global vmlist and this worked for ages.

Unfortunately, the new allocator isn't aware of this, so it constantly
tries to allocate the VM space which is already in use, making vmalloc on
alpha defunct.

This patch forces KVA to import vmlist entries on init.

[akpm@linux-foundation.org: remove unneeded check (per Johannes)]
Signed-off-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Nick Piggin <npiggin@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

video/framebuffer: fix bug: jpegview cannot work on framebuffer device other than 16BPP

Force fb_var_screeninfo color format on all Blackfin Framebuffer Drivers.

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
Cc: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

gpio: fix probe() error return in gpio driver probes

A number of drivers in drivers/gpio return -ENODEV when confronted with
missing setup parameters such as the platform data.  However, returning
-ENODEV causes the driver layer to silently ignore the driver as it
assumes the probe did not find anything and was only speculative.

To make life easier to discern why a driver is not being attached, change
to returning -EINVAL, which is a better description of the fact that the
driver data was not valid.

Also add a set of dev_dbg() statements to the error paths to provide an
better explanation of the error as there may be more that one point in the
driver.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

atmel_spi: allow transfer when max_speed_hz = 0

For some reason I have to slowdown clock to touchscreen device.

In atmel_spi_setup() there is comment that max_speed_hz == 0 means as slow
as possible and divider is set to maximum value.  But in
atmel_spi_transfer() function is check against not zero max_speed_hz with
EINVAL returned.

Probably driver should setup divider for each transfer based on
transfer->speed_hz value, but I think that would be not necessary overhead
as all used devices have constant clock.

Below patch works fine for me.

Signed-off-by: Stanislaw Gruszka <stf_xl@wp.pl>
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Cc: David Brownell <david-b@pacbell.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

atmel_serial: fix flow control bug

Fix the following problem, related to hardware flow control (CTS/RTS):
Transmitting while CTS line is asserted in DMA mode, due to not checking
for tx-stopped condition.

We found these problems while testing the UARTs with hardware
flow-control.

Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Cc: "Andrew Victor" <avictor.za@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Update of Documentation: vm.txt and proc.txt

Update Documentation/sysctl/vm.txt and Documentation/filesystems/proc.txt.
 More specifically, the section on /proc/sys/vm in
Documentation/filesystems/proc.txt was removed and a link to
Documentation/sysctl/vm.txt added.

Most of the verbiage from proc.txt was simply moved in vm.txt, with new
addtional text for "swappiness" and "stat_interval".

Signed-off-by: Peter W Morreale <pmorreale@novell.com>
Acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Revert "x86 PAT: remove CPA WARN_ON for zero pte"

This reverts commit 58dab916dfb57328d50deb0aa9b3fc92efa248ff, which
makes my Nehalem come to a nasty crawling almost-halt.  It looks like it
turns off caching of regular kernel RAM, with the understandable
slowdown of a few orders of magnitude as a result.

Acked-by: Ingo Molnar <mingo@elte.hu>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Peter Anvin <hpa@zytor.com>
Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

b44: GFP_DMA skb should not escape from driver

b44 chip has some hardware limitations, that need GFP_DMA bounce
buffers in some situations.

In order to not deplete DMA zone, we should keep allocated GFP_DMA skb
only for driver use. At rx time, we copy such skb to newly allocated
skb, reusing existing copybreak infrastructure.

On machines with low amount of memory, all skb meet the hardware limitation,
so no copy is needed. We detect this situation using a new device flag, set
to one if one GFP_DMA skb was ever allocated by b44_alloc_rx_skb().

Previously allocated skb, even outside from DMA zone will then be recycled,
to have minimal impact on DMA zone use.

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Tested-by: Ionut Leonte <ionut.leonte@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

hwmon: (abituguru3) Fix CONFIG_DMI=n fallback to probe

When CONFIG_DMI is not enabled, dmi detection should flag that no board
could be detected (err=1) rather than another error condition (err<0).

This fixes the fallback to manual probing for all motherboards, even
those without DMI strings, when CONFIG_DMI=n.

Signed-off-by: Alistair John Strachan <alistair@devzero.co.uk>
Cc: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

hwmon: (abituguru3) Enable DMI probing feature on IN9 32X MAX

Switch the IN9 32X MAX over from port probing to the preferred DMI
probe method.

Signed-off-by: Alistair John Strachan <alistair@devzero.co.uk>
Tested-by: Paul Hartman <paul.hartman+gentoo@gmail.com>
Cc: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

hwmon: (abituguru3) Match partial DMI board name strings

The switch-over to using DMI board strings to identify abituguru3 compatible
mainboards works most of the time, but sometimes the vendor has substantially
modified the board string between BIOS revisions.

We have found that the vendor chipset identification string (provided in
brackets) changes frequently and is of no use to us. The rest of the board
string sometimes changes in subtle ways, e.g. whitespace or variations in
capitalization.

The new comparison code checks only a part of the supplied DMI board string,
trimming the bracketed content, whitespace, and ignoring case as necessary.

This fixes a bug where an IP35 Pro running an early BIOS would not be
detected without the force=1 module parameter, and also speculatively
fixes other similiar issues.

Signed-off-by: Alistair John Strachan <alistair@devzero.co.uk>
Reported-by: Nick Pasich <NewsLetters@nickandbarb.net>
Cc: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

hwmon: Add a driver for the ADT7475 hardware monitoring chip

Hwmon driver for the ADT7475 chip.

Signed-off-by: Jordan Crouse <jordan.crouse@amd.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

hwmon: (k8temp) Fix temperature reporting for (most) K8 RevG CPUs

Current Temperature for K8 RevG desktop CPUs is a "normalized value"
which can be below ambient temperature.

As a consequence lots of RevG systems report temperatures like:

$ sensors
k8temp-pci-00c3
Adapter: PCI adapter
Core0 Temp:    +17 C
Core0 Temp:     +3 C
Core1 Temp:    +21 C
Core1 Temp:     +5 C

being quite below ambient temperature.
There are even reports of negative temperature values.

This patch corrects the temperature reporting of k8temp for
RevG desktop CPUs.

Cc: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: Andreas Herrmann <andreas.herrmann3@amd.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

hwmon: (k8temp) Fix wrong sensor selection for AMD K8 RevF/RevG CPUs

Meaning of ThermSenseCoreSel bit was inverted beginning with K8 RevF.
That means with current driver temp1/temp2 belong to core 1 and
temp3/temp4 belong to core 0 on a K8 RevF/RevG CPU.

This patch ensures that temp1/temp2 always belong to core 0 and
temp3/temp4 to core 1 for all K8 revisions.

Cc: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: Andreas Herrmann <andreas.herrmann3@amd.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

hwmon: (k8temp) Warn about fam F rev F errata

Add warning about wrong CPU temperature readouts on all fam F rev F.

The allowed combinations of processors ensure that all processors
in a multisocket system have similar characteristics, e.g.

(1) provide temperature sensor interface (>=RevC && <RevF)
(2) are affected by erratum #141 (>=RevF)

Thus it is sufficient to check the revision of the boot CPU.

For "mixed silicon support" refer to
"Revision Guide for AMD Athlon 64 and AMD Opteron Processors" (RevA-E) and
"Revision Guide for AMD NPT Family 0Fh Processors" (RefF-G).

Cc: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: Andreas Herrmann <andreas.herrmann3@amd.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] SN specific version of dma_get_required_mask()
  [IA64] generic_defconfig: Enable SATA_VITESSE
  [IA64] dump stack on kernel unaligned warnings
  [IA64] Turn on CONFIG_HAVE_UNSTABLE_CLOCK
  [IA64] Update to use account_{steal,idle}_ticks

CREDITS address update for dwmw2.

Update employer's care-of address in CREDITS file, and remove references
to some _very_ old stuff I'd forgotten I'd ever done.

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Merge branch 'for-linus' of git://git./linux/kernel/git/tiwai/sound-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ALSA: hda - Fix invalid amp value for STAC925x
  ASoC: Fix the power update function for snd_soc_dapm_value_mux
  sound: virtuoso: do not overwrite EEPROM on Xonar D2/D2X
  ALSA: hda - Fix HP dv5 mic input
  ALSA: hda - Fix missing initialization of NID 0x0e for STAC925x
  ALSA: USB quirk for Logitech Quickcam Pro 9000 name
  ALSA: hda - Fix stac92hd83xxx_amp_nids[]
  ALSA: hda - Add automatic model setting for Samsung Q45
  ALSA: hda - Don't reset HP pinctl in patch_sigmatel.c
  ALSA: hda: stac92hd8xxx amp mixers
  ALSA: hda - Fix silent headphone output on Panasonic CF-74
  ALSA: hda - Update model descriptions in patch_sigmatel.c
  ALSA: hda - Use queue_delayed_work()
  ALSA: hda - Add quirk for another HP dv5
  ALSA: hda - Add support of NVidia MCP78 HDMI
  ALSA: hda - Fix a typo
  ALSA: hda - More fixes on Gateway entries
  ALSA: patch_sigmatel: Add missing Gateway entries and autodetection
  ALSA: hda - Add a new function to seek for a codec ID

hso serial throttled tty kref fix.

This patch is for Alan Cox as it related to the tty layer.
Hopefully the hso driver is again relatively stable with this fix.

Signed-off-by: Denis Joseph Barrow <D.Barow@option.com>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

tty: Fix double grabbing of a spinlock

The HSO changes for kref introduced a recursive spinlock take. All
functions which call put_rxbuf_data already have serial->serial_lock
grabbed.

[Comment to code added-AC]

Signed-off-by: Denis Joseph Barrow <D.Barrow@option.com>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

tty: Fix a kref leak in the HSO driver on re-open

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

ftdi_sio: fix kref leak

Commit 4a90f09b20f4622dcbff1f0e1e6bae1704f8ad8c added kref stuff to
ftdi_sio, but missed tty_kref_put at one exit point in
ftdi_process_read.

Signed-off-by: Jim Paris <jim@jtan.com>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

When a break signal is detected, the next character should be ignored.

This was not implemented correctly for the pnx8xxx_uart driver.

[From further discussion:
Correct, you can look to it as two separate bugs:
a) the next character is not ignored while it should;
b) the status bits 31-8 are copied to the 'ch' variable while they shouldn't.

Both bugs prevent correct break signal handling (and therefore correct
behaviour of the magic SysRq key). Bug b didn't cause too much trouble
earlier because in most situations the status bits are all zero; for
this case they unfortunately aren't.
]

Signed-off-by: Mischa Jonker <mischa.jonker@nxp.com>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

serial: Add SupraExpress 336i PnP Voice Modem

Add SupraExpress 336i PnP Voice Modem

Tested and working with the following device: (output from lspnp -v)
01:01.00 SUP1381 (unknown)
    state = active
io 0x2f8-0x2ff
irq 3

Signed-off-by: Daniel Gagnon <daniel.gagnon@yahoo.com>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8250_pci: add support for netmos 9835 IBM devices

Most of netmos 9835 hardware is handled by parport-serial.  IBM introduces
a device which doesn't have any parallel ports and have screwed subdevice
PCI id (not corresponding to port numbers).

Handle this device (9710:9835 1014:0299) properly.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

tty: Fix race in the flush for some ldiscs

If you issue an ioctl to flush a tty as the line discipline is changing or
otherwise unplugged you can get a crash. The bug is very old but the rest
of the BKL lock dropping and some very "good" luck on Ingo's part caught
an example.

Use the correct ldisc_ref form so that we wait for the ldisc change to
complete and then flush

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

sched: sched_slice() fixlet

Mike's change: 0a582440f "sched: fix sched_slice())" broke group
scheduling by forgetting to reload cfs_rq on each loop.

This patch fixes aim7 regression and specjbb2005 regression becomes
less than 1.5% on 8-core stokley.

Signed-off-by: Lin Ming <ming.m.lin@intel.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Tested-by: Jayson King <dev@jaysonking.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

[IA64] SN specific version of dma_get_required_mask()

Create a platform specific version of dma_get_required_mask()
for ia64 SN Altix. All SN Altix platforms support 64 bit DMA
addressing regardless of the size of system memory.
Create an ia64 machvec for dma_get_required_mask, with the
SN version unconditionally returning DMA_64BIT_MASK.

Signed-off-by: John Keller <jpk@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

[IA64] generic_defconfig: Enable SATA_VITESSE

CONFIG_SATA_VITESSE=y was not added to generic_defconfig when
sn2_defconfig was removed.  SGI Altix systems that use an IO10
base IO card to drive the root device are unable to boot without
the Vitesse controller.

Signed-off-by: Brent Casavant <bcasavan@sgi.com>
Signed-off-by: Robin Holt <holt@sgi.com>
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

[IA64] dump stack on kernel unaligned warnings

Often the cause of kernel unaligned access warnings is not
obvious from just the ip displayed in the warning.  This adds
the option via proc to dump the stack in addition to the warning.
The default is off (just display the 1 line warning).  To enable
the stack to be shown: echo 1 > /proc/sys/kernel/unaligned-dump-stack

Signed-off-by: Doug Chapman <doug.chapman@hp.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

[IA64] Turn on CONFIG_HAVE_UNSTABLE_CLOCK

sched_clock() on ia64 is based on ar.itc, so is never
completely synchronized between cpus. On some platforms
(e.g. certain models of SGI Altix) it may be running at
radically different frequencies.

Based on a patch from Dimitri Sivanich which set this
just for SN2 && GENERIC kernels ... it is needed for
all ia64 machines.

Signed-off-by: Tony Luck <tony.luck@intel.com>

[IA64] Update to use account_{steal,idle}_ticks

This patch fixes the following errors caused by
79741dd35713ff4f6fd0eafd59fa94e8a4ba922d which changed
the prototypes of account_steal_time() and account_idle_time().

>   CC      arch/ia64/xen/time.o
> arch/ia64/xen/time.c: In function 'consider_steal_time':
> arch/ia64/xen/time.c:132: warning: passing argument 1 of 'account_steal_time' makes integer from pointer without a cast
> arch/ia64/xen/time.c:132: error: too many arguments to function 'account_steal_time'
> arch/ia64/xen/time.c:133: warning: passing argument 1 of 'account_steal_time' makes integer from pointer without a cast
> arch/ia64/xen/time.c:133: error: too many arguments to function 'account_steal_time'

Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
Signed-off-by: Tony Luck <tony.luck@intel.com>

korina: do not use IRQF_SHARED with IRQF_DISABLED

As the kernel warning states: "IRQF_DISABLED is not guaranteed on shared
IRQs". Since these IRQs' values are hardcoded and my test system doesn't
show any shared use of IRQs at all, rather make them non-shared than
non-disabled.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
Acked-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

korina: do not stop queue here

Apparently this doesn't make sense. Otherwise the queue gets disabled as
soon as it's getting empty and can only be resurrected by a driver
restart.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
Acked-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

korina: fix handling tx_chain_tail

Originally this must have been a rewrite error when introducing
'chain_index'. But the original driver did not use the previous chain
item everywhere: when altering the address tx_chain_tail points to, it
should move forward, not backwards.
Also this is not an "index" but rather the penultimate element in the
chain, so rename it accordingly.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
Acked-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

korina: do tx at the right position

Triggering TX before the write to the DMA status mask register leads to
transferring packets with maximum payload no matter what the actual
packet size is.
While here, also trigger RX scheduling after writing the DMA status mask
register, like it was in the original driver before it was sent
upstream.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
Acked-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

korina: do schedule napi after testing for it

The called netif_rx_schedule() does all the work for us:
- it checks the return value of netif_rx_schedule_prep() and
- if everything is ok calls __netif_rx_schedule().
Before this change, the driver received absolutely nothing.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
Acked-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

korina: rework korina_rx() for use with napi

This function needs an early exit condition to function properly, or
else caller assumes napi workload wasn't enough to handle all received
packets and korina_rx is called again (and again and again and ...).

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
Acked-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

korina: disable napi on close and restart

Without this the driver will crash when the NIC is being restarted.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
Acked-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

korina: reset resource buffer size to 1536

The new value is the one used in the external patch before and allows at
least a standard MTU of 1500 to be handled correctly. Impact of this
change gets visible when bigger packets are to be received, issuing:
| ping -s 492 <IP>
and bigger payload sized led to 100% packet loss.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
Acked-by: Florian Fainelli <florian@openwrt.org
Signed-off-by: David S. Miller <davem@davemloft.net>

korina: fix usage of driver_data

Using platform_set_drvdata() here makes no sense, since the driver_data
field has already been filled with valuable data (i.e. the MAC address).
Also having driver_data point to the net_device is rather pointless
since struct korina_device contains an apropriate field for it.

Signed-off-by: Phil Sutter <n0-1@freewrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: First slow path interrupt race

The "read for interrupts" flag must be set before enabling slow-path
interrupts as well (and not just before fast-path interrupts)

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: MTU Filter

Too big packets could pass due to wrong filter size

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Indirection table initialization index

Wrong initialization of the multi-queue indirection table - it should
be using the function and not the port index

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Missing brackets

Calculation bug due to missing brackets

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Fixing the doorbell size

The size of the doorbell is 4KB, this bug become visible when using
more than 8 queues

Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>