When finding a present and acceptable 2M/1G mapping, the number
of pages mapped this way shouldn't be incremented (as it was
already incremented when the earlier part of the mapping was
established). Instead, last_map_addr needs to be updated in this
case.

Further, address increments were wrong in one place each in both
phys_pmd_init() and phys_pud_init() (lacking the aligning down
to the respective page boundary).

As we're now doing the same calculation several times, fold it
into a single instance using a local variable (matching how
kernel_physical_mapping_init() itself does it at the PGD level).

Observed during code inspection, not because of an actual
problem.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/4FB3C27202000078000841A0@nat28.tlf.novell.comSigned-off-by: Ingo Molnar <mingo@kernel.org>

Since sizeof(long) is 4 in x86_32 mode, and it's 8 in x86_64
mode, sizeof(long long) is also 8 byte in x86_64 mode.
use long mode can fit TLB_FLUSH_ALL defination here both in 32
or 64 bits mode.
Signed-off-by: Alex Shi <alex.shi@intel.com>
Link: http://lkml.kernel.org/n/tip-evv5bekiipi2pmyzdsy8lkkw@git.kernel.orgSigned-off-by: Ingo Molnar <mingo@kernel.org>

x86 unconditionally uses NO_BOOTMEM so there is no use
of the HAVE_ARCH_BOOTMEM support as mm/bootmem.c is the
only file referencing this symbol.

bootmem_arch_preferred_node() is the function referred
in the mm/bootmem.c code and can thuis be dropped too.

x86 was the sole user of HAVE_ARCH_BOOTMEM - so there is
an opportunity to clean up a little in mm/bootmem.c too
if we do not expect other users to emerge.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Tejun Heo <tj@kernel.org>
Link: http://lkml.kernel.org/r/20120406124735.GA6920@merkur.ravnborg.orgSigned-off-by: Ingo Molnar <mingo@kernel.org>

Currently leave_mm() unconditionally switches the cr3 to swapper_pg_dir.
But there is no need to change the cr3, if we already left that mm.

intel_idle() for example calls leave_mm() on every deep c-state entry where
the CPU flushes the TLB for us. Similarly flush_tlb_all() was also calling
leave_mm() whenever the TLB is in LAZY state. Both these paths will be
improved with this change.
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1332460885.16101.147.camel@sbsiddha-desk.sc.intel.comSigned-off-by: H. Peter Anvin <hpa@zytor.com>

For machines that enable PSE, the first 2/4M memory region still uses
4K pages, so needs more PTEs in this case, but
find_early_table_space() doesn't count this.

This patch fixes it.

The bug was found via code review, no misbehavior of the kernel
was observed.
Signed-off-by: WANG Cong <xiyou.wangcong@gmail.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: <ianfang.cn@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Link: http://lkml.kernel.org/n/tip-kq6a00qe33h7c7ais2xsywnh@git.kernel.orgSigned-off-by: Ingo Molnar <mingo@elte.hu>

MMC fixes from Chris Ball for 3.3:
 - atmel-mci: oops fix against regression introduced in 3.2
 - core: power saving regression fix against 3.3-rc1
 - core: suspend/resume fix for UHS-I cards
 - esdhc-imx: MMC card regression fix against 3.0
 - mmci: oops fix for ARM systems with large (64k) pages
 - MAINTAINERS update for atmel-mci.

* tag 'mmc-fixes-for-3.3' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc:
  mmc: core: Fixup suspend/resume issues for UHS-I cards
  mmc: mmci: reduce max_blk_count to avoid overflowing max_req_size
  mmc: sdhci-esdhc-imx: fix for mmc cards on i.MX5
  mmc: core: fix regression: set default clock gating delay to 0
  MAINTAINERS: hand over atmel-mci (sd/mmc interface)
  mmc: atmel-mci: don't use dma features when using DMA with no chan available

Pull from Jiri Kosina:
 "Please pull to receive updates for HID layer.  Nikolai's patch is
  rather important and should still go in for 3.3, as it's a regression
  fix for commit b4b583d4."

* 'upstream-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  HID: hid-input: allow array fields out of range
  HID: usbhid: Add NOGET quirk for the AIREN Slim+ keyboard

Allow array field values out of range as per HID 1.11 specification,
section 6.2.25:

	Rather than returning a single bit for each button in the group, an
	array returns an index in each field that corresponds to the pressed
	button (like keyboard scan codes). An out-of range value in and array
	field is considered no controls asserted.

Apparently, "and" above is a typo and should be "an".

This fixes at least Waltop tablet pen clicks - otherwise BTN_TOUCH is never
released.

The relevant part of Waltop tablet report descriptors is this:

	0x09, 0x42,         /*          Usage (Tip Switch),         */
	0x09, 0x44,         /*          Usage (Barrel Switch),      */
	0x09, 0x46,         /*          Usage (Tablet Pick),        */
	0x15, 0x01,         /*          Logical Minimum (1),        */
	0x25, 0x03,         /*          Logical Maximum (3),        */
	0x75, 0x04,         /*          Report Size (4),            */
	0x95, 0x01,         /*          Report Count (1),           */
	0x80,               /*          Input,                      */

This is a regression fix for commit b4b583d4 ("HID: be more strict when
ignoring out-of-range fields").
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

MFD fixes from Samuel Ortiz:
 "This is the pull request for the MFD fixes for 3.3.  We have a few
  NULL pointer dereferences fixes, an ACPI conflict check fix, and a
  couple of wm8994 fixes."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sameo/mfd-2.6:
  mfd: Correct readability of WM8994 DC servo 4E register
  mfd: Initialize tps65912 irq platform data properly
  mfd: Fix ACPI conflict check
  mfd: Fix ab8500 error path bug
  mfd: Test for jack detection when deciding if wm8994 should suspend
  mfd: Initialize tps65910 irq platform data properly
  mfd: Fix possible s5m null pointer dereference
  mfd: wm8350 variable dereferenced before check

It's only used inside fs/dcache.c, and we're going to play games with it
for the word-at-a-time patches.  This time we really don't even want to
export it, because it really is an internal function to fs/dcache.c, and
has been since it was introduced.

Having it in that extremely hot header file (it's included in pretty
much everything, thanks to <linux/fs.h>) is a disaster for testing
different versions, and is utterly pointless.

We really should have some kind of header file diet thing, where we
figure out which parts of header files are really better off private and
only result in more expensive compiles.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Even if cards supports 1.8V I/O voltage those should anyway be
initialized at 3.3V I/O according to (e)MMC, SD and SDIO specs.
Some eMMC and embedded SDIO devices are able to be initialized
at 1.8V as well, but it is better to be safe.

Do note that initialization in this context means that the card
has been completely powered off, otherwise the card will remain
at the last I/O voltage level that were negotitiated.

Due to the above being taken care of the suspend/resume issues
for UHS-I SD-cards has been fixed.
Signed-off-by: Ulf Hansson <ulf.hansson@stericsson.com>
Acked-by: Philip Rakity <prakity@marvell.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Tested-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

On a system with large pages (64k in my case), the following BUG is
triggered in MMC core:

[    2.338023] BUG: failure at drivers/mmc/core/core.c:221/mmc_start_request()!
[    2.338102] Kernel panic - not syncing: BUG!
[    2.338155] Call trace:
[    2.338228] [<ffffffc00008635c>] dump_backtrace+0x0/0x120
[    2.338317] [<ffffffc0003365ec>] dump_stack+0x14/0x1c
[    2.338403] [<ffffffc000336990>] panic+0xbc/0x1f0
[    2.338498] [<ffffffc00027a494>] mmc_start_request+0x154/0x184
[    2.338600] [<ffffffc00027abdc>] mmc_start_req+0x110/0x140
[    2.338701] [<ffffffc00028604c>] mmc_blk_issue_rw_rq+0x7c/0x39c
[    2.338804] [<ffffffc00028652c>] mmc_blk_issue_rq+0x1c0/0x468
[    2.338905] [<ffffffc000287564>] mmc_queue_thread+0x68/0x118
[    2.338995] [<ffffffc0000bc308>] kthread+0x84/0x8c

This is because of a 64k request with a max_req_size of 64k-1 bytes.

The following patch fixes the problem by limiting the max_blk_count
such that max_blk_count * max_blk_size == max_req_size. I couldn't
pursuade the compiler to emit a shift instead of a div without encoding
the shift explicitly.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

On i.MX53 we have to write a special SDHCI_CMD_ABORTCMD to the
SDHCI_TRANSFER_MODE register during a MMC_STOP_TRANSMISSION
command. This works for SD cards. However, with MMC cards
the MMC_SET_BLOCK_COUNT command is used instead, but this
needs the same handling. Fix MMC cards by testing for the
MMC_SET_BLOCK_COUNT command aswell. Tested on a custom i.MX53
board with a Transcend MMC+ card and eMMC.

The kernel started used MMC_SET_BLOCK_COUNT in 3.0, so this
is a regression for these boards introduced in 3.0; it should
go to 3.0/3.1/3.2-stable.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

A recent commit "mmc: core: Use delayed work in clock gating framework"
(597dd9d7) introduced a default 200ms delay before clock gating
actually takes place.  This means that every time an MMC interface
becomes idle it first stays on for 200ms before gating its clock. This
leads to increased power consumption and is therefore a clear regression.
This patch restores the original behaviour by setting the default delay
to 0. Users prioritising throughput over power efficiency can still
modify the delay via sysfs.
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Chris Ball <cjb@laptop.org>

Modify MAINTAINERS entry for Atmel SD/MMC drivers.
I hand the atmel-mci and at91_mci drivers over to Ludovic.
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Signed-off-by: Chris Ball <cjb@laptop.org>

SCSI fixes from James Bottomley:
 "There's just a single fix in here: the osd max device number fix."

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6:
  [SCSI] osd_uld: Bump MAX_OSD_DEVICES from 64 to 1,048,576

PARISC fixes from James Bottomley:
 "This is a set of build fixes to get the cross compiled architecture
  testbeds building again"

* tag 'parisc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/parisc-2.6:
  [PARISC] don't unconditionally override CROSS_COMPILE for 64 bit.
  [PARISC] include <linux/prefetch.h> in drivers/parisc/iommu-helpers.h
  [PARISC] fix compile break caused by iomap: make IOPORT/PCI mapping functions conditional

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/kvm: Fix Host-Only/Guest-Only counting with SVM disabled

Pull from Herbert Xu:
  "This push fixes a bug in mv_cesa that causes all hash operations
   that supply data on a final operation to fail."

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: mv_cesa - fix final callback not ignoring input data

Commit 5707c87f "vfs: uninline full_name_hash()" broke the modular
build, because it needs exporting now that it isn't inlined any more.
Reported-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

hhwmon fixes for 3.3-rc6 from Guenter Roeck:

These patches are necessary for correct operation and management of
F75387.

* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (f75375s) Catch some attempts to write to r/o registers
  hwmon: (f75375s) Properly map the F75387 automatic modes to pwm_enable
  hwmon: (f75375s) Make pwm*_mode writable for the F75387
  hwmon: (f75375s) Fix writes to the pwm* attribute for the F75387

fbdev fixes for 3.3 from Florian Tobias Schandinat

It includes:
 - two fixes for OMAP HDMI
 - one fix to make new OMAP functions behave as they are supposed to
 - one Kconfig dependency fix
 - two fixes for viafb for modesetting on VX900 hardware

* tag 'fbdev-fixes-for-3.3-2' of git://github.com/schandinat/linux-2.6:
  OMAPDSS: APPLY: make ovl_enable/disable synchronous
  OMAPDSS: panel-dvi: Add Kconfig dependency on I2C
  viafb: fix IGA1 modesetting on VX900
  viafb: select HW scaling on VX900 for IGA2
  OMAPDSS: HDMI: hot plug detect fix
  OMAPDSS: HACK: Ensure DSS clock domain gets out of idle when HDMI is enabled

sound fixes for 3.3-rc6 from Takashi Iwai

This contains again regression fixes for various HD-audio and ASoC
regarding SSI and dapm shutdown path.  In addition, a minor azt3328
fix and the correction of the new jack-notification strings in HD-audio.

* tag 'sound-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda - Kill hyphenated names
  ALSA: hda - Add a fake mute feature
  ALSA: hda - Always set HP pin in unsol handler for STAC/IDT codecs
  ALSA: azt3328 - Fix NULL ptr dereference on cards without OPL3
  ALSA: hda/realtek - Fix resume of multiple input sources
  ASoC: i.MX SSI: Fix DSP_A format.
  ASoC: dapm: Check for bias level when powering down

The code in link_path_walk() that finds out the length and the hash of
the next path component is some of the hottest code in the kernel.  And
I have a version of it that does things at the full width of the CPU
wordsize at a time, but that means that we *really* want to split it up
into a separate helper function.

So this re-organizes the code a bit and splits the hashing part into a
helper function called "hash_name()".  It returns the length of the
pathname component, while at the same time computing and writing the
hash to the appropriate location.

The code generation is slightly changed by this patch, but generally for
the better - and the added abstraction actually makes the code easier to
read too.  And the new interface is well suited for replacing just the
"hash_name()" function with alternative implementations.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

It did some odd things for unclear reasons.  As this is one of the
functions that gets changed when doing word-at-a-time compares, this is
yet another of the "don't change any semantics, but clean things up so
that subsequent patches don't get obscured by the cleanups".
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

.. and also use it in lookup_one_len() rather than open-coding it.

There aren't any performance-critical users, so inlining it is silly.
But it wouldn't matter if it wasn't for the fact that the word-at-a-time
dentry name patches want to conditionally replace the function, and
uninlining it sets the stage for that.

So again, this is a preparatory patch that doesn't change any semantics,
and only prepares for a much cleaner and testable word-at-a-time dentry
name accessor patch.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

These don't change any semantics, but they clean up the code a bit and
mark some arguments appropriately 'const'.

They came up as I was doing the word-at-a-time dcache name accessor
code, and cleaning this up now allows me to send out a smaller relevant
interesting patch for the experimental stuff.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Some callbacks are set too early -- i.e. we can have dma capabilities but
we can't get a dma channel. So wait to get the dma channel before setting
callbacks and change logs consequently.
Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
[Should be applied to 3.2-stable.]
Cc: <stable@vger.kernel.org>
Signed-off-by: Chris Ball <cjb@laptop.org>

It makes no sense to attempt to manually configure the fan in auto mode,
or set the duty cycle directly in closed loop mode.  The corresponding
registers are then read-only.  If the user tries it nonetheless, error out
with EINVAL instead of silently doing nothing.
Signed-off-by: Nikolaus Schulz <mail@microschulz.de>
[guenter.roeck@ericsson.com: Minor formatting cleanup]
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

The F75387 supports automatic fan control using either PWM duty cycle or
RPM speed values.  Make the driver detect the latter mode, and expose the
different modes in sysfs as per pwm_enable, so that the user can switch
between them.

The interpretation of the pwm_enable attribute for the F75387 is adjusted
to be a superset of those values used for similar Fintek chips which do
not support automatic duty mode, with 2 mapping to automatic speed mode,
and moving automatic duty mode to the new value 4.

Toggling the duty mode via pwm_enable is currently denied for the F75387,
as the chip then simply reinterprets the fan configuration register values
according to the new mode, switching between RPM and PWM units, which
makes this a dangerous operation.

This patch introduces a new pwm mode into the driver. This is necessary
because the new mode (automatic pwm mode, 4) may already be enabled by the
BIOS, and the driver should not break existing functionality. This was seen
on at least one board.
Signed-off-by: Nikolaus Schulz <mail@microschulz.de>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

Merge branches 'core-urgent-for-linus', 'perf-urgent-for-linus' and 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pulling latest branches from Ingo:

* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  memblock: Fix size aligning of memblock_alloc_base_nid()

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf probe: Ensure offset provided is not greater than function length without DWARF info too
  perf tools: Ensure comm string is properly terminated
  perf probe: Ensure offset provided is not greater than function length
  perf evlist: Return first evsel for non-sample event on old kernel
  perf/hwbp: Fix a possible memory leak

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  CPU hotplug, cpusets, suspend: Don't touch cpusets during suspend/resume

There is only one error code to return for a bad user-space buffer
pointer passed to a system call in the same address space as the
system call is executed, and that is EFAULT.  Furthermore, the
low-level access routines, which catch most of the faults, return
EFAULT already.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Roland McGrath <roland@hack.frob.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

The regset common infrastructure assumed that regsets would always
have .get and .set methods, but not necessarily .active methods.
Unfortunately people have since written regsets without .set methods.

Rather than putting in stub functions everywhere, handle regsets with
null .get or .set methods explicitly.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Roland McGrath <roland@hack.frob.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

It turned out that a performance counter on AMD does not
count at all when the GO or HO bit is set in the control
register and SVM is disabled in EFER.

This patch works around this issue by masking out the HO bit
in the performance counter control register when SVM is not
enabled.

The GO bit is not touched because it is only set when the
user wants to count in guest-mode only. So when SVM is
disabled the counter should not run at all and the
not-counting is the intended behaviour.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Avi Kivity <avi@redhat.com>
Cc: Stephane Eranian <eranian@google.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Gleb Natapov <gleb@redhat.com>
Cc: Robert Richter <robert.richter@amd.com>
Cc: stable@vger.kernel.org # v3.2
Link: http://lkml.kernel.org/r/1330523852-19566-1-git-send-email-joerg.roedel@amd.comSigned-off-by: Ingo Molnar <mingo@elte.hu>

Kill hyphens from "Line-Out" name strings, as suggested by Mark Brown.
Signed-off-by: Takashi Iwai <tiwai@suse.de>

Merge tag 'perf-urgent-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/urgent

Various smaller perf/urgent fixes.
Signed-off-by: Ingo Molnar <mingo@elte.hu>

Watchdog updates from Wim Van Sebroeck:

* git://www.linux-watchdog.org/linux-watchdog:
  watchdog: fix GETTIMEOUT ioctl in booke_wdt
  watchdog: update maintainers git entry
  watchdog: Fix typo in pnx4008_wdt.c
  watchdog: Fix typo in Kconfig
  watchdog: fix error in probe() of s3c2410_wdt (reset at booting)
  watchdog: hpwdt: clean up set_memory_x call for 32 bit

Pull from Mark Brown:
 "A simple, driver specific fix.  This device isn't widely used outside
  of Marvell reference boards most of which are probably used with their
  BSPs rather than with mainline so low risk."

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: fix the ldo configure according to 88pm860x spec

i2c bugfix from Wolfram Sang:
  "This patch fixes a wrong assumption in the mxs-i2c-driver about a
   command queue being done.  Without it, we have seen races when the
   bus was under load."

* 'i2c-embedded/for-3.3' of git://git.pengutronix.de/git/wsa/linux-2.6:
  i2c: mxs: only flag completion when queue is completely done