openwrt/staging/blogic.git
12 years agoMerge branch 'lantiq' of git://dev.phrozen.org/mips-next into mips-for-linux-next
Ralf Baechle [Thu, 27 Sep 2012 16:07:25 +0000 (18:07 +0200)]
Merge branch 'lantiq' of git://dev.phrozen.org/mips-next into mips-for-linux-next

12 years agoMerge branch 'cn6xxx-mgmt' of git://git.linux-mips.org/pub/scm/daney/upstream-daney...
Ralf Baechle [Thu, 27 Sep 2012 16:00:34 +0000 (18:00 +0200)]
Merge branch 'cn6xxx-mgmt' of git://git.linux-mips.org/daney/upstream-daney into mips-for-linux-next

12 years agoMerge branch 'cn68xx-ciu2' of git://git.linux-mips.org/pub/scm/daney/upstream-daney...
Ralf Baechle [Thu, 27 Sep 2012 16:00:07 +0000 (18:00 +0200)]
Merge branch 'cn68xx-ciu2' of git://git.linux-mips.org/daney/upstream-daney into mips-for-linux-next

12 years agoMIPS: Replace `-' in defconfig filename wth `_' for consistency.
Ralf Baechle [Wed, 26 Sep 2012 20:00:05 +0000 (22:00 +0200)]
MIPS: Replace `-' in defconfig filename wth `_' for consistency.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMIPS: Wire kcmp syscall.
Ralf Baechle [Fri, 14 Sep 2012 20:47:15 +0000 (22:47 +0200)]
MIPS: Wire kcmp syscall.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMIPS: MIPSsim: Remove the MIPSsim platform.
Steven J. Hill [Mon, 24 Sep 2012 14:48:53 +0000 (09:48 -0500)]
MIPS: MIPSsim: Remove the MIPSsim platform.

The MIPSsim platform is no longer supported or used.

[ralf@linux-mips.org: Also remove mipssim from arch/mips/Kbuild.platforms
and delete arch/mips/include/asm/mach-mipssim/*.]

Signed-off-by: Steven J. Hill <sjhill@mips.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/4350/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMIPS: NOTIFY_RESUME is not needed in TIF masks
Al Viro [Sat, 5 May 2012 20:24:40 +0000 (16:24 -0400)]
MIPS: NOTIFY_RESUME is not needed in TIF masks

If it's set, SIGPENDING is also set.  And SIGPENDING is present in
the masks...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMIPS: Merge the identical "return from syscall" per-ABI code
Al Viro [Sat, 5 May 2012 20:11:35 +0000 (16:11 -0400)]
MIPS: Merge the identical "return from syscall" per-ABI code

No need to keep 4 copies of that stuff; merged and taken to
entry.S, unused public symbols there killed off.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMIPS: Unobfuscate _TIF..._MASK
Al Viro [Sat, 5 May 2012 19:57:00 +0000 (15:57 -0400)]
MIPS: Unobfuscate _TIF..._MASK

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMIPS: Prevent hitting do_notify_resume() with !user_mode(regs).
Al Viro [Thu, 3 May 2012 01:45:12 +0000 (21:45 -0400)]
MIPS: Prevent hitting do_notify_resume() with !user_mode(regs).

Too late to do anything there...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMIPS: Malta: Don't crash on spurious interrupt.
Ralf Baechle [Sun, 16 Sep 2012 23:23:21 +0000 (01:23 +0200)]
MIPS: Malta: Don't crash on spurious interrupt.

48d480b0bde794781fcae9501fb043c1bac0e523 [[MIPS] Malta: Fix off by one bug in interrupt
handler.] did not take in account that irq_ffs() will also return 0 if for some reason
the set of pending interrupts happens to be empty.

This is trivial to trigger with a RM5261 CPU module running a 64-bit kernel and results
in something like the following:

CPU 0 Unable to handle kernel paging request at virtual address 0000000000000000, epc == ffffffff801772d0, ra == ffffffff8017ad24
Oops[#1]:
Cpu 0
$ 0   : 0000000000000000 ffffffff9000a4e0 ffffffff9000a4e0 ffffffff9000a4e0
$ 4   : ffffffff80592be0 0000000000000000 00000000000000d6 ffffffff80322ed0
$ 8   : ffffffff805fe538 0000000000000000 ffffffff9000a4e0 ffffffff80590000
$12   : 00000000000000d6 0000000000000000 ffffffff80600000 ffffffff805fe538
$16   : 0000000000000000 0000000000000010 ffffffff80592be0 0000000000000010
$20   : 0000000000000000 0000000000500001 0000000000000000 ffffffff8051e078
$24   : 0000000000000028 ffffffff803226e8
$28   : 9800000003828000 980000000382b900 ffffffff8051e060 ffffffff8017ad24
Hi    : 0000000000000000
Lo    : 0000006388974000
epc   : ffffffff801772d0 handle_irq_event_percpu+0x70/0x2f0
    Not tainted
ra    : ffffffff8017ad24 handle_percpu_irq+0x54/0x88
Status: 9000a4e2    KX SX UX KERNEL EXL
Cause : 00808008
BadVA : 0000000000000000
PrId  : 000028a0 (Nevada)
Modules linked in:
Process init (pid: 1, threadinfo=9800000003828000, task=9800000003827968, tls=0000000077087490)
Stack : ffffffff80592be0 ffffffff8058d248 0000000000000040 0000000000000000
        ffffffff80613340 0000000000500001 ffffffff805a0000 0000000000000882
        9800000003b89000 ffffffff8017ad24 00000000000000d5 0000000000000010
        ffffffff9000a4e1 ffffffff801769f4 ffffffff9000a4e0 ffffffff801037f8
        0000000000000000 ffffffff80101c44 0000000000000000 ffffffff9000a4e0
        0000000000000000 9000000018000000 90000000180003f9 0000000000000001
        0000000000000000 00000000000000ff 0000000000000018 0000000000000001
        0000000000000001 00000000003fffff 0000000000000020 ffffffff802cf7ac
        ffffffff80208918 000000007fdadf08 ffffffff80612d88 ffffffff9000a4e1
        0000000000000040 0000000000000000 ffffffff80613340 0000000000500001
        ...
Call Trace:
[<ffffffff801772d0>] handle_irq_event_percpu+0x70/0x2f0
[<ffffffff8017ad24>] handle_percpu_irq+0x54/0x88
[<ffffffff801769f4>] generic_handle_irq+0x44/0x60
[<ffffffff801037f8>] do_IRQ+0x48/0x70
[<ffffffff80101c44>] ret_from_irq+0x0/0x4
[<ffffffff80326170>] serial8250_startup+0x310/0x870
[<ffffffff8032175c>] uart_startup.part.7+0x9c/0x330
[<ffffffff80321b4c>] uart_open+0x15c/0x1b0
[<ffffffff80302034>] tty_open+0x1fc/0x720
[<ffffffff801bffac>] chrdev_open+0x7c/0x180
[<ffffffff801b9ab8>] do_dentry_open.isra.14+0x288/0x390
[<ffffffff801bac5c>] nameidata_to_filp+0x5c/0xc0
[<ffffffff801ca700>] do_last.isra.33+0x330/0x8f0
[<ffffffff801caf3c>] path_openat+0xbc/0x440
[<ffffffff801cb3c8>] do_filp_open+0x38/0xa8
[<ffffffff801bade4>] do_sys_open+0x124/0x218
[<ffffffff80110538>] handle_sys+0x118/0x13c

Code: 02d5a825  12800012  02a0b02d <de820000de850008  0040f809  0220202d  0040a82d  40026000
---[ end trace 5d8e7b9a86badd2d ]---
Kernel panic - not syncing: Fatal exception in interrupt

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMIPS: Malta: Remove RTC Data Mode bootstrap breakage
Maciej W. Rozycki [Thu, 13 Sep 2012 20:23:25 +0000 (21:23 +0100)]
MIPS: Malta: Remove RTC Data Mode bootstrap breakage

 YAMON requires and enforces the RTC Data Mode (Register B, DM bit) to
binary, that is the bit is set every time the board goes through the
firmware bootstrap sequence.  Likewise its calendar manipulation commands
interpret or set the RTC registers unconditionally as binary, never
actually checking what the value of the DM bit is, under the (correct)
assumption that it has been previously set, to indicate the binary mode.

 A change to Linux a while ago however introduced a platform-specific
tweak that clears that bit and therefore forces the data mode to BCD.
This causes clock corruption and misinterpretation that has to be fixed up
by user-mode tools in system startup scripts as the initial clock is often
incorrect according to the BCD interpretation forced.

 This change removes the hack; a comment included refers to alarm code,
but even if it was broken at one point by requiring the BCD mode, it
should have been trivially corrected and even if not, given how rarely the
alarm feature is used, that was not really a reasonable justification to
break the system clock that is indeed used by virtually everything.  And
either way the alarm code has been since fixed anyway.

Signed-off-by: Maciej W. Rozycki <macro@codesourcery.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/4336/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoGPIO: MIPS: lantiq: fix overflow inside stp-xway driver
John Crispin [Wed, 11 Jul 2012 14:33:43 +0000 (16:33 +0200)]
GPIO: MIPS: lantiq: fix overflow inside stp-xway driver

The driver was using a 16 bit field for storing the shadow value of the shift
register cascade. This resulted in only the first 2 shift registeres receiving
the correct data. The third shift register would always receive 0x00.

Fix this by using a 32bit field for the shadow value.

Signed-off-by: John Crispin <blogic@openwrt.org>
Cc: linux-kernel@vger.kernel.org
12 years agoMIPS: lantiq: make use of __gpio_to_irq
John Crispin [Wed, 15 Aug 2012 13:41:50 +0000 (15:41 +0200)]
MIPS: lantiq: make use of __gpio_to_irq

The gpio_chip struct allows us to set a .to_irq callback. Once this is set
we can rely on the generic __gpio_to_irq() function to map gpio->irq allowing
more than one gpio_chip to register an interrupt

Signed-off-by: John Crispin <blogic@openwrt.org>
12 years agoDocument: devicetree: add OF documents for lantiq falcon pinctrl
John Crispin [Sun, 22 Jul 2012 07:23:50 +0000 (09:23 +0200)]
Document: devicetree: add OF documents for lantiq falcon pinctrl

Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: Thomas Langer <thomas.langer@lantiq.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Cc: devicetree-discuss@lists.ozlabs.org
Cc: linux-kernel@vger.kernel.org
12 years agoDocument: devicetree: add OF documents for lantiq xway pinctrl
John Crispin [Fri, 20 Jul 2012 17:01:00 +0000 (19:01 +0200)]
Document: devicetree: add OF documents for lantiq xway pinctrl

Signed-off-by: John Crispin <blogic@openwrt.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Cc: devicetree-discuss@lists.ozlabs.org
Cc: linux-kernel@vger.kernel.org
12 years agoOF: pinctrl: MIPS: lantiq: adds support for FALCON SoC
John Crispin [Sat, 19 May 2012 22:33:56 +0000 (00:33 +0200)]
OF: pinctrl: MIPS: lantiq: adds support for FALCON SoC

Implement support for pinctrl on lantiq/falcon socs. The FALCON has 5 banks
of up to 32 pins.

Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: Thomas Langer <thomas.langer@lantiq.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Cc: devicetree-discuss@lists.ozlabs.org
Cc: linux-kernel@vger.kernel.org
12 years agoOF: pinctrl: MIPS: lantiq: implement lantiq/xway pinctrl support
John Crispin [Tue, 28 Aug 2012 10:44:59 +0000 (12:44 +0200)]
OF: pinctrl: MIPS: lantiq: implement lantiq/xway pinctrl support

Implement support for pinctrl on lantiq/xway socs. The IO core found on these
socs has the registers for pinctrl, pinconf and gpio mixed up in the same
register range. As the gpio_chip handling is only a few lines, the driver also
implements the gpio functionality. This obseletes the old gpio driver that was
located in the arch/ folder.

Signed-off-by: John Crispin <blogic@openwrt.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Cc: devicetree-discuss@lists.ozlabs.org
Cc: linux-kernel@vger.kernel.org
12 years agonetdev: octeon_mgmt: Make multi-line comment style consistent.
David Daney [Tue, 21 Aug 2012 18:45:12 +0000 (11:45 -0700)]
netdev: octeon_mgmt: Make multi-line comment style consistent.

No code changes.  Recent patches have used the netdev style multi-line
comment formatting, making the style inconsistent within octeon_mgmt.c

Update the remaining comment blocks to achieve style harmony.

Signed-off-by: David Daney <david.daney@cavium.com>
Acked-by: David S. Miller <davem@davemloft.net>
12 years agonetdev: octeon_mgmt: Remove some useless 'inline'
David Daney [Tue, 21 Aug 2012 18:45:11 +0000 (11:45 -0700)]
netdev: octeon_mgmt: Remove some useless 'inline'

Signed-off-by: David Daney <david.daney@cavium.com>
Acked-by: David S. Miller <davem@davemloft.net>
12 years agonetdev: octeon_mgmt: Cleanup and modernize MAC address handling.
David Daney [Tue, 21 Aug 2012 18:45:10 +0000 (11:45 -0700)]
netdev: octeon_mgmt: Cleanup and modernize MAC address handling.

Use eth_mac_addr(), and generate a random address if none is otherwise
assigned.

Signed-off-by: David Daney <david.daney@cavium.com>
Acked-by: David S. Miller <davem@davemloft.net>
12 years agonetdev: octeon_mgmt: Set the parent device.
David Daney [Tue, 21 Aug 2012 18:45:09 +0000 (11:45 -0700)]
netdev: octeon_mgmt: Set the parent device.

This establishes useful links in sysfs.

Signed-off-by: David Daney <david.daney@cavium.com>
Acked-by: David S. Miller <davem@davemloft.net>
12 years agonetdev: octeon_mgmt: Improve ethtool_ops.
David Daney [Tue, 21 Aug 2012 18:45:08 +0000 (11:45 -0700)]
netdev: octeon_mgmt: Improve ethtool_ops.

Correctly show no link when the interface is down, and return
-EOPNOTSUPP for things that don't work.  This quiets the ethtool
program when run on down interfaces.

Signed-off-by: David Daney <david.daney@cavium.com>
Acked-by: David S. Miller <davem@davemloft.net>
12 years agonetdev: octeon_mgmt: Add hardware timestamp support.
Chad Reese [Tue, 21 Aug 2012 18:45:07 +0000 (11:45 -0700)]
netdev: octeon_mgmt: Add hardware timestamp support.

Octeon cn6XXX models have timestamp support on the mgmt ports, so hook
it up.

Signed-off-by: Chad Reese <kreese@caviumnetworks.com>
Signed-off-by: David Daney <david.daney@cavium.com>
Acked-by: David S. Miller <davem@davemloft.net>
12 years agonetdev: octeon_mgmt: Add support for 1Gig ports.
David Daney [Tue, 21 Aug 2012 18:45:06 +0000 (11:45 -0700)]
netdev: octeon_mgmt: Add support for 1Gig ports.

The original hardware only supported 10M and 100M.  Later versions
added 1G support.  Here we update the driver to make use of this.

Also minor logic clean-ups for testing PHY registration error codes
and TX complete high water marks.

Signed-off-by: David Daney <david.daney@cavium.com>
Acked-by: David S. Miller <davem@davemloft.net>
12 years agoMIPS: Octeon: Add octeon_io_clk_delay() function.
David Daney [Tue, 21 Aug 2012 18:45:05 +0000 (11:45 -0700)]
MIPS: Octeon: Add octeon_io_clk_delay() function.

Also cleanup and fix octeon_init_cvmcount()

Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Acked-by: David S. Miller <davem@davemloft.net>
12 years agoMIPS: OCTEON: Register ciu/ciu2 as the default irq_domain.
David Daney [Wed, 29 Aug 2012 23:09:22 +0000 (16:09 -0700)]
MIPS: OCTEON: Register ciu/ciu2 as the default irq_domain.

This makes it possible to call irq_create_mapping(NULL, ??)

Signed-off-by: David Daney <david.daney@cavium.com>
12 years agoMIPS: Octeon: Make interrupt controller work with threaded handlers.
David Daney [Thu, 5 Apr 2012 17:24:25 +0000 (10:24 -0700)]
MIPS: Octeon: Make interrupt controller work with threaded handlers.

For CIUv1 controllers, we were relying on all calls to the irq_chip
functions to be done from the CPU that received the irq, and that they
would all be done from interrupt contest.  These assumptions do not
hold for threaded handlers.

We make all the masking actually mask the irq source, and use real
raw_spin_locks instead of manually twiddling the Status[IE] bit.

Signed-off-by: David Daney <david.daney@cavium.com>
12 years agoMIPS: OCTEON: Add support for cn68XX interrupt controller.
David Daney [Wed, 4 Apr 2012 22:34:41 +0000 (15:34 -0700)]
MIPS: OCTEON: Add support for cn68XX interrupt controller.

The cn68XX has a new interrupt controller named CIU2, add support for
this, and use it if cn68XX detected at runtime.

Signed-off-by: David Daney <david.daney@cavium.com>
12 years agoMIPS: OCTEON: Add OCTEON_IRQ_* definitions for cn68XX chips.
David Daney [Tue, 3 Apr 2012 22:22:05 +0000 (15:22 -0700)]
MIPS: OCTEON: Add OCTEON_IRQ_* definitions for cn68XX chips.

There are 64 workqueue, 32 watchdog, and 4 mbox.

Signed-off-by: David Daney <david.daney@cavium.com>
12 years agoMIPS: OCTEON: Update register definitions.
David Daney [Tue, 3 Apr 2012 20:44:18 +0000 (13:44 -0700)]
MIPS: OCTEON: Update register definitions.

Add support for cn68xx, cn61xx, cn63xx, cn66xx and cnf71XX.

Add little-endian register layouts.

Patch cvmx-interrupt-rsl.c for changed definition.

Signed-off-by: David Daney <david.daney@cavium.com>
12 years agoMIPS: OCTEON: Add detection of cnf71xx parts.
David Daney [Tue, 3 Apr 2012 19:08:24 +0000 (12:08 -0700)]
MIPS: OCTEON: Add detection of cnf71xx parts.

Also add cvmx_get_octeon_family().

Both of these are needed by the upcoming register definition refresh
patch.

Signed-off-by: David Daney <david.daney@cavium.com>
12 years agoMIPS: mm: Add compound tail page _mapcount when mapped
Jovi Zhang [Wed, 22 Aug 2012 02:34:08 +0000 (10:34 +0800)]
MIPS: mm: Add compound tail page _mapcount when mapped

See commit b6999b191 which did the same modification for x86's mm/gup,

Quote from commit b6999b191:
    "If compound pages are used and the page is a
    tail page, gup_huge_pmd() increases _mapcount to record tail page are
    mapped while gup_huge_pud does not do that."

[ralf@linux-mips.org: fixed rejects caused by the original patch getting
linewrapped.]

Signed-off-by: Jovi Zhang <boojovi@gmail.com>
Cc: Youquan Song <youquan.song@intel.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: <stable@vger.kernel.org>
Patchwork: https://patchwork.linux-mips.org/patch/4291/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMIPS: CMP/SMTC: Fix tc_id calculation
RongQing.Li [Mon, 2 Jul 2012 04:34:30 +0000 (12:34 +0800)]
MIPS: CMP/SMTC: Fix tc_id calculation

Currently the tc_id code is:

  (read_c0_tcbind() >> TCBIND_CURTC_SHIFT) & TCBIND_CURTC;

After processing this becomes:

  (read_c0_tcbind() >> 21) & ((0xff) << 21)

But it should be:

  (read_c0_tcbind() & ((0xff)<< 21)) >> 21

Signed-off-by: RongQing.Li <roy.qing.li@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/4077/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoMerge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus
Linus Torvalds [Sat, 25 Aug 2012 18:45:04 +0000 (11:45 -0700)]
Merge branch 'upstream' of git://git.linux-mips.org/ralf/upstream-linus

Pull MIPS fixes from Ralf Baechle:
 "Random fixes across the MIPS tree.  The two hotspots are several bugs
  in the module loader and the ath79 SOC support; also noteworthy is the
  restructuring of the code to synchronize CPU timers across CPUs on
  startup; the old code recently ceased to work due to unrelated
  changes.

  All except one of these patches have sat for a significant time in
  linux-next for testing."

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
  MIPS: pci-ar724x: avoid data bus error due to a missing PCIe module
  MIPS: Malta: Delete duplicate PCI fixup.
  MIPS: ath79: don't hardcode the unavailability of the DSP ASE
  MIPS: Synchronize MIPS count one CPU at a time
  MIPS: BCM63xx: Fix SPI message control register handling for BCM6338/6348.
  MIPS: Module: Deal with malformed HI16/LO16 relocation sequences.
  MIPS: Fix race condition in module relocation code.
  MIPS: Fix memory leak in error path of HI16/LO16 relocation handling.
  MIPS: MTX-1: Add udelay to mtx1_pci_idsel
  MIPS: ath79: select HAVE_CLK
  MIPS: ath79: Use correct IRQ number for the OHCI controller on AR7240
  MIPS: ath79: Fix number of GPIO lines for AR724[12]
  MIPS: Octeon: Fix broken interrupt controller code.

12 years agoMerge branch 'for-3.6' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Sat, 25 Aug 2012 18:43:41 +0000 (11:43 -0700)]
Merge branch 'for-3.6' of git://linux-nfs.org/~bfields/linux

Pull nfsd bugfixes from J. Bruce Fields:
 "Particular thanks to Michael Tokarev, Malahal Naineni, and Jamie
  Heilman for their testing and debugging help."

* 'for-3.6' of git://linux-nfs.org/~bfields/linux:
  svcrpc: fix svc_xprt_enqueue/svc_recv busy-looping
  svcrpc: sends on closed socket should stop immediately
  svcrpc: fix BUG() in svc_tcp_clear_pages
  nfsd4: fix security flavor of NFSv4.0 callback

12 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-block
Linus Torvalds [Sat, 25 Aug 2012 18:36:43 +0000 (11:36 -0700)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-block

Pull block-related fixes from Jens Axboe:

 - Improvements to the buffered and direct write IO plugging from
   Fengguang.

 - Abstract out the mapping of a bio in a request, and use that to
   provide a blk_bio_map_sg() helper.  Useful for mapping just a bio
   instead of a full request.

 - Regression fix from Hugh, fixing up a patch that went into the
   previous release cycle (and marked stable, too) attempting to prevent
   a loop in __getblk_slow().

 - Updates to discard requests, fixing up the sizing and how we align
   them.  Also a change to disallow merging of discard requests, since
   that doesn't really work properly yet.

 - A few drbd fixes.

 - Documentation updates.

* 'for-linus' of git://git.kernel.dk/linux-block:
  block: replace __getblk_slow misfix by grow_dev_page fix
  drbd: Write all pages of the bitmap after an online resize
  drbd: Finish requests that completed while IO was frozen
  drbd: fix drbd wire compatibility for empty flushes
  Documentation: update tunable options in block/cfq-iosched.txt
  Documentation: update tunable options in block/cfq-iosched.txt
  Documentation: update missing index files in block/00-INDEX
  block: move down direct IO plugging
  block: remove plugging at buffered write time
  block: disable discard request merge temporarily
  bio: Fix potential memory leak in bio_find_or_create_slab()
  block: Don't use static to define "void *p" in show_partition_start()
  block: Add blk_bio_map_sg() helper
  block: Introduce __blk_segment_map_sg() helper
  fs/block-dev.c:fix performance regression in O_DIRECT writes to md block devices
  block: split discard into aligned requests
  block: reorganize rounding of max_discard_sectors

12 years agoMerge tag 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Sat, 25 Aug 2012 17:28:19 +0000 (10:28 -0700)]
Merge tag 'upstream-linus' of git://git./linux/kernel/git/jgarzik/libata-dev

Pull libata fixes from Jeff Garzik:
 - libata-acpi regression fix
 - additional or corrected drive quirks for ata_blacklist
 - Kconfig text tweaking
 - new PCI IDs
 - pata_atiixp: quirk for MSI motherboard
 - export ahci_dev_classify for an ahci_platform driver

* tag 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  libata: Add a space to " 2GB ATA Flash Disk" DMA blacklist entry
  [libata] new quirk, lift bridge limits for Buffalo DriveStation Quattro
  [libata] Kconfig: Elaborate that SFF is meant for legacy and PATA stuff
  [libata] acpi: call ata_acpi_gtm during ata port init time
  ata_piix: Add Device IDs for Intel Lynx Point-LP PCH
  ahci: Add Device IDs for Intel Lynx Point-LP PCH
  pata_atiixp: override cable detection on MSI E350DM-E33
  ahci: un-staticize ahci_dev_classify

12 years agolibata: Add a space to " 2GB ATA Flash Disk" DMA blacklist entry
Prarit Bhargava [Thu, 23 Aug 2012 19:11:52 +0000 (15:11 -0400)]
libata: Add a space to " 2GB ATA Flash Disk" DMA blacklist entry

commit d70e551c8e1ecb6f20422f8db6bfe6a0049edcb8, Add " 2GB ATA Flash
Disk"/"ADMA428M" to DMA blacklist, should have added a space before 2GB.

Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Fri, 24 Aug 2012 04:58:04 +0000 (21:58 -0700)]
Merge git://git./linux/kernel/git/herbert/crypto-2.6

Pull crypto fixes from Herbert Xu:
 "This push fixes a build error on 32-bit archs in the hifn driver as
  well as a potential deadlock in the caam driver."

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: caam - fix possible deadlock condition
  crypto: hifn_795x - fix 64bit division and undefined __divdi3 on 32bit archs

12 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs
Linus Torvalds [Fri, 24 Aug 2012 04:56:22 +0000 (21:56 -0700)]
Merge branch 'for_linus' of git://git./linux/kernel/git/jack/linux-fs

Pull UDF, ext3 & reiserfs fixes from Jan Kara:
 "A couple of fixes (udf, reiserfs, ext3) that accumulated over my
  vacation."

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  udf: fix retun value on error path in udf_load_logicalvol
  jbd: don't write superblock when unmounting an ro filesystem
  reiserfs: fix deadlocks with quotas
  quota: Move down dqptr_sem read after initializing default warn[] type at __dquot_alloc_space().
  UDF: During mount free lvid_bh before rescanning with different blocksize
  udf: fix udf_setsize() for file data in ICB

12 years agoMerge tag 'upstream-3.6-rc3' of git://git.infradead.org/linux-ubifs
Linus Torvalds [Fri, 24 Aug 2012 04:50:40 +0000 (21:50 -0700)]
Merge tag 'upstream-3.6-rc3' of git://git.infradead.org/linux-ubifs

Pull UBIFS fixes from Artem Bityutskiy:
 - Fix crash on error which prevents emulated power-cut testing.
 - Fix log reply regression introduced in 3.6-rc1.
 - Fix UBIFS complaints about too small debug buffer size which.
 - Fix error message spelling, and remove incorrect commentary.

* tag 'upstream-3.6-rc3' of git://git.infradead.org/linux-ubifs:
  UBIFS: fix error messages spelling
  UBIFS: fix complaints about too small debug buffer size
  UBIFS: fix replay regression
  UBIFS: fix crash on error path
  UBIFS: remove stale commentary

12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/ide
Linus Torvalds [Fri, 24 Aug 2012 04:49:56 +0000 (21:49 -0700)]
Merge git://git./linux/kernel/git/davem/ide

Pull IDE power management bugfix from David S. Miller.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/ide:
  ide: fix generic_ide_suspend/resume Oops

12 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 24 Aug 2012 04:48:41 +0000 (21:48 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Ingo Molnar:
 "This tree contains misc fixlets: a perf script python binding fix, a
  uprobes fix and a syscall tracing fix."

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf tools: Add missing files to build the python binding
  uprobes: Fix mmap_region()'s mm->mm_rb corruption if uprobe_mmap() fails
  tracing/syscalls: Fix perf syscall tracing when syscall_nr == -1

12 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 24 Aug 2012 04:47:54 +0000 (21:47 -0700)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Ingo Molnar:
 "This tree contains assorted fixlets: an alternatives patching crash
  fix, an irq migration/hotplug interaction fix, a fix for large AMD
  microcode images and a comment fixlet."

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86, microcode, AMD: Fix broken ucode patch size check
  x86/alternatives: Fix p6 nops on non-modular kernels
  x86/fixup_irq: Use cpu_online_mask instead of cpu_all_mask
  x86/spinlocks: Fix comment in spinlock.h

12 years agoMerge branch 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 24 Aug 2012 04:46:57 +0000 (21:46 -0700)]
Merge branch 'timers-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull timer fixes from Thomas Gleixner:
 "Mostly small fixes for the fallout of the timekeeping overhaul in 3.6
  along with stable fixes to address an accumulation problem and missing
  sanity checks for RTC readouts and user space provided values."

* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  time: Avoid making adjustments if we haven't accumulated anything
  time: Avoid potential shift overflow with large shift values
  time: Fix casting issue in timekeeping_forward_now
  time: Ensure we normalize the timekeeper in tk_xtime_add
  time: Improve sanity checking of timekeeping inputs

12 years agoMerge branch 'upstream-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
Linus Torvalds [Fri, 24 Aug 2012 04:45:54 +0000 (21:45 -0700)]
Merge branch 'upstream-fixes' of git://git./linux/kernel/git/jikos/hid

Pull HID fix from Jiri Kosina:
 "Fix for one particular device not being properly claimed by
  hid-multitouch driver"

* 'upstream-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  HID: Remove QUANTA from special drivers list

12 years agoMIPS: pci-ar724x: avoid data bus error due to a missing PCIe module
Gabor Juhos [Thu, 23 Aug 2012 13:35:26 +0000 (15:35 +0200)]
MIPS: pci-ar724x: avoid data bus error due to a missing PCIe module

If the controller has no PCIe module attached, accessing of the device
configuration space causes a data bus error. Avoid this by checking the
status of the PCIe link in advance, and indicate an error if the link
is down.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
Cc: stable@vger.kernel.org
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/4293/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
12 years agoblock: replace __getblk_slow misfix by grow_dev_page fix
Hugh Dickins [Thu, 23 Aug 2012 10:17:36 +0000 (12:17 +0200)]
block: replace __getblk_slow misfix by grow_dev_page fix

Commit 91f68c89d8f3 ("block: fix infinite loop in __getblk_slow")
is not good: a successful call to grow_buffers() cannot guarantee
that the page won't be reclaimed before the immediate next call to
__find_get_block(), which is why there was always a loop there.

Yesterday I got "EXT4-fs error (device loop0): __ext4_get_inode_loc:3595:
inode #19278: block 664: comm cc1: unable to read itable block" on console,
which pointed to this commit.

I've been trying to bisect for weeks, why kbuild-on-ext4-on-loop-on-tmpfs
sometimes fails from a missing header file, under memory pressure on
ppc G5.  I've never seen this on x86, and I've never seen it on 3.5-rc7
itself, despite that commit being in there: bisection pointed to an
irrelevant pinctrl merge, but hard to tell when failure takes between
18 minutes and 38 hours (but so far it's happened quicker on 3.6-rc2).

(I've since found such __ext4_get_inode_loc errors in /var/log/messages
from previous weeks: why the message never appeared on console until
yesterday morning is a mystery for another day.)

Revert 91f68c89d8f3, restoring __getblk_slow() to how it was (plus
a checkpatch nitfix).  Simplify the interface between grow_buffers()
and grow_dev_page(), and avoid the infinite loop beyond end of device
by instead checking init_page_buffers()'s end_block there (I presume
that's more efficient than a repeated call to blkdev_max_block()),
returning -ENXIO to __getblk_slow() in that case.

And remove akpm's ten-year-old "__getblk() cannot fail ... weird"
comment, but that is worrying: are all users of __getblk() really
now prepared for a NULL bh beyond end of device, or will some oops??

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: stable@vger.kernel.org # 3.0 3.2 3.4 3.5
Signed-off-by: Jens Axboe <axboe@kernel.dk>
12 years agox86, microcode, AMD: Fix broken ucode patch size check
Andreas Herrmann [Tue, 31 Jul 2012 13:41:45 +0000 (15:41 +0200)]
x86, microcode, AMD: Fix broken ucode patch size check

This issue was recently observed on an AMD C-50 CPU where a patch of
maximum size was applied.

Commit be62adb49294 ("x86, microcode, AMD: Simplify ucode verification")
added current_size in get_matching_microcode(). This is calculated as
size of the ucode patch + 8 (ie. size of the header). Later this is
compared against the maximum possible ucode patch size for a CPU family.
And of course this fails if the patch has already maximum size.

Cc: <stable@vger.kernel.org> [3.3+]
Signed-off-by: Andreas Herrmann <andreas.herrmann3@amd.com>
Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
Link: http://lkml.kernel.org/r/1344361461-10076-1-git-send-email-bp@amd64.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
12 years agoMIPS: lantiq: adds support for gptu timers
John Crispin [Thu, 16 Aug 2012 11:31:02 +0000 (11:31 +0000)]
MIPS: lantiq: adds support for gptu timers

Lantiq socs have a General Purpose Timer Unit (GPTU). This driver allows us to
initialize the timers. The voice firmware needs these timers as a reference.

Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/4236/

12 years agoMIPS: lantiq: enable pci clk conditional for xrx200 SoC
John Crispin [Thu, 16 Aug 2012 08:25:42 +0000 (08:25 +0000)]
MIPS: lantiq: enable pci clk conditional for xrx200 SoC

The xrx200 SoC family has the same PCI clock register layout as the AR9.
Enable the same quirk as for AR9

Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/4235/

12 years agoMIPS: lantiq: falcon clocks were not enabled properly
John Crispin [Thu, 16 Aug 2012 08:25:41 +0000 (08:25 +0000)]
MIPS: lantiq: falcon clocks were not enabled properly

As a result of a non populated ->bits field inside the clock struct, the clock
domains were never powered on the Falcon. Until now we only used domains that
were also used and powered by the bootloader.

Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/4234/

12 years agoMIPS: lantiq: adds support for nmi and ejtag bootrom vectors
John Crispin [Thu, 16 Aug 2012 11:39:56 +0000 (11:39 +0000)]
MIPS: lantiq: adds support for nmi and ejtag bootrom vectors

Register nmi and ejtag bootrom vectors for FALC-ON SoC.

Signed-off-by: Thomas Langer <thomas.langer@lantiq.com>
Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/4238/

12 years agoMIPS: lantiq: external irq sources are not loaded properly
John Crispin [Thu, 16 Aug 2012 08:09:22 +0000 (08:09 +0000)]
MIPS: lantiq: external irq sources are not loaded properly

Support for the external interrupt unit was broken when the code was converted
to devicetree support.

Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/4231/

12 years agoMIPS: lantiq: dont register irq_chip for the irq cascade
John Crispin [Thu, 16 Aug 2012 08:09:21 +0000 (08:09 +0000)]
MIPS: lantiq: dont register irq_chip for the irq cascade

We dont want to register the irq_chip for the MIPS IRQ cascade.

Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/4230/

12 years agoMIPS: lantiq: timer irq can be different to 7
John Crispin [Thu, 16 Aug 2012 08:09:20 +0000 (08:09 +0000)]
MIPS: lantiq: timer irq can be different to 7

The SVIP SoC has its timer IRQ on a different IRQ than 7. Fix up the irq
code to be able to handle this.

Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/4229/

12 years agoMIPS: lantiq: split up IRQ IM ranges
John Crispin [Thu, 16 Aug 2012 11:39:57 +0000 (11:39 +0000)]
MIPS: lantiq: split up IRQ IM ranges

Up to now all our SoCs had the 5 IM ranges in a consecutive order. To accomodate
the SVIP we need to support IM ranges that are scattered inside the register range.

Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/4237/

12 years agoLinux 3.6-rc3
Linus Torvalds [Wed, 22 Aug 2012 20:29:06 +0000 (13:29 -0700)]
Linux 3.6-rc3

12 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Wed, 22 Aug 2012 17:45:13 +0000 (10:45 -0700)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "Intel: edid fixes, power consumption fix, s/r fix, haswell fix

  Radeon: BIOS loading fixes for UEFI and Thunderbolt machines, better
  MSAA validation, lockup timeout fixes, modesetting fixes

  One udl dpms fix, one vmwgfx fix, a couple of trivial core changes.

  There is an export added to ACPI as part of the radeon bios fixes.

  I've also included the fbcon flashing cursor vs deinit race fix, that
  seems the simplest place to start"

Trivial conflict in drivers/video/console/fbcon.c due to me having
already applied the fbcon flashing cursor vs deinit race fix, and Dave
had added a comment in there too.

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: (22 commits)
  fbcon: fix race condition between console lock and cursor timer (v1.1)
  drm: Add missing static storage class specifiers in drm_proc.c file
  drm/udl: dpms off the crtc when disabled.
  drm: Remove two unused fields from struct drm_display_mode
  drm: stop vmgfx driver explosion
  drm/radeon/ss: use num_crtc rather than hardcoded 6
  Revert "drm/radeon: fix bo creation retry path"
  drm/i915: use hsw rps tuning values everywhere on gen6+
  drm/radeon: split ATRM support out from the ATPX handler (v3)
  drm/radeon: convert radeon vfct code to use acpi_get_table_with_size
  ACPI: export symbol acpi_get_table_with_size
  drm/radeon: implement ACPI VFCT vbios fetch (v3)
  drm/radeon/kms: extend the Fujitsu D3003-S2 board connector quirk to cover later silicon stepping
  drm/radeon: fix checking of MSAA renderbuffers on r600-r700
  drm/radeon: allow CMASK and FMASK in the CS checker on r600-r700
  drm/radeon: init lockup timeout on ring init
  drm/radeon: avoid turning off spread spectrum for used pll
  drm/i915: fall back to bit-banging if GMBUS fails in CRT EDID reads
  drm/i915: extract connector update from intel_ddc_get_modes() for reuse
  drm/i915: fix hsw uncached pte
  ...

12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending
Linus Torvalds [Wed, 22 Aug 2012 17:42:30 +0000 (10:42 -0700)]
Merge git://git./linux/kernel/git/nab/target-pending

Pull SCSI target fixes from Nicholas Bellinger:
 "The executive summary includes:

   - Post-merge review comments for tcm_vhost (MST + nab)
   - Avoid debugging overhead when not debugging for tcm-fc(FCoE) (MDR)
   - Fix NULL pointer dereference bug on alloc_page failulre (Yi Zou)
   - Fix REPORT_LUNs regression bug with pSCSI export (AlexE + nab)
   - Fix regression bug with handling of zero-length data CDBs (nab)
   - Fix vhost_scsi_target structure alignment (MST)

  Thanks again to everyone who contributed a bugfix patch, gave review
  feedback on tcm_vhost code, and/or reported a bug during their own
  testing over the last weeks.

  There is one other outstanding bug reported by Roland recently related
  to SCSI transfer length overflow handling, for which the current
  proposed bugfix has been left in queue pending further testing with
  other non iscsi-target based fabric drivers.

  As the patch is verified with loopback (local SGL memory from SCSI
  LLD) + tcm_qla2xxx (TCM allocated SGL memory mapped to PCI HW) fabric
  ports, it will be included into the next 3.6-rc-fixes PULL request."

* git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending:
  target: Remove unused se_cmd.cmd_spdtl
  tcm_fc: rcu_deref outside rcu lock/unlock section
  tcm_vhost: Fix vhost_scsi_target structure alignment
  target: Fix regression bug with handling of zero-length data CDBs
  target/pscsi: Fix bug with REPORT_LUNs handling for SCSI passthrough
  tcm_vhost: Change vhost_scsi_target->vhost_wwpn to char *
  target: fix NULL pointer dereference bug alloc_page() fails to get memory
  tcm_fc: Avoid debug overhead when not debugging
  tcm_vhost: Post-merge review changes requested by MST
  tcm_vhost: Fix incorrect IS_ERR() usage in vhost_scsi_map_iov_to_sgl

12 years agoMerge branch 'i2c-embedded/for-current' of git://git.pengutronix.de/git/wsa/linux
Linus Torvalds [Wed, 22 Aug 2012 17:41:36 +0000 (10:41 -0700)]
Merge branch 'i2c-embedded/for-current' of git://git.pengutronix.de/git/wsa/linux

Pull i2c-embedded fixes from Wolfram Sang:
 "Some bugfixes for the "embedded" part of the I2C subsystem.  The fixes
  affect mostly drivers which have been largely reworked lately and
  where regressions appeared."

* 'i2c-embedded/for-current' of git://git.pengutronix.de/git/wsa/linux:
  i2c: tegra: protect suspend/resume callbacks with CONFIG_PM_SLEEP
  i2c: diolan-u2c: Fix master_xfer return code
  I2C: OMAP: xfer: fix runtime PM get/put balance on error
  i2c: nomadik: Add default configuration into the Nomadik I2C driver

12 years agoMerge tag 'for-3.6-rc3' of git://gitorious.org/linux-pwm/linux-pwm
Linus Torvalds [Wed, 22 Aug 2012 17:27:12 +0000 (10:27 -0700)]
Merge tag 'for-3.6-rc3' of git://gitorious.org/linux-pwm/linux-pwm

Pull pwm fixes from Thierry Reding:
 "These patches fix the Samsung PWM driver and perform some minor
  cleanups like fixing checkpatch and sparse warnings.

  Two redundant error messages are removed and the Kconfig help text for
  the PWM subsystem is made more descriptive."

* tag 'for-3.6-rc3' of git://gitorious.org/linux-pwm/linux-pwm:
  pwm: Improve Kconfig help text
  pwm: core: Fix coding style issues
  pwm: vt8500: Fix coding style issue
  pwm: Remove a redundant error message when devm_request_and_ioremap fails
  pwm: samsung: add missing device pointer to struct pwm_chip
  pwm: Add missing static storage class specifiers in core.c file

12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph...
Linus Torvalds [Wed, 22 Aug 2012 16:58:05 +0000 (09:58 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/sage/ceph-client

Pull ceph fixes from Sage Weil:
 "Jim's fix closes a narrow race introduced with the msgr changes.  One
  fix resolves problems with debugfs initialization that Yan found when
  multiple client instances are created (e.g., two clusters mounted, or
  rbd + cephfs), another one fixes problems with mounting a nonexistent
  server subdirectory, and the last one fixes a divide by zero error
  from unsanitized ioctl input that Dan Carpenter found."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
  ceph: avoid divide by zero in __validate_layout()
  libceph: avoid truncation due to racing banners
  ceph: tolerate (and warn on) extraneous dentry from mds
  libceph: delay debugfs initialization until we learn global_id

12 years agoMerge tag 'nfs-for-3.6-3' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Wed, 22 Aug 2012 16:57:25 +0000 (09:57 -0700)]
Merge tag 'nfs-for-3.6-3' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client bugfixes from Trond Myklebust:
 - NFSv3 mounts need to fail if the FSINFO rpc call fails
 - Ensure that the NFS commit cache gets torn down when we unload the
   NFS module.
 - Fix memory scribble issues when interrupting a LAYOUTGET rpc call
 - Fix NFSv4 legacy idmapper regressions
 - Fix issues with the NFSv4 getacl command
 - Fix a regression when using the legacy "mount -t nfs4"

* tag 'nfs-for-3.6-3' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  NFSv3: Ensure that do_proc_get_root() reports errors correctly
  NFSv4: Ensure that nfs4_alloc_client cleans up on error.
  NFS: return -ENOKEY when the upcall fails to map the name
  NFS: Clear key construction data if the idmap upcall fails
  NFSv4: Don't use private xdr_stream fields in decode_getacl
  NFSv4: Fix the acl cache size calculation
  NFSv4: Fix pointer arithmetic in decode_getacl
  NFS: Alias the nfs module to nfs4
  NFS: Fix a regression when loading the NFS v4 module
  NFSv4.1: Remove a bogus BUG_ON() in nfs4_layoutreturn_done
  pnfs-obj: Better IO pattern in case of unaligned offset
  NFS41: add pg_layout_private to nfs_pageio_descriptor
  pnfs: nfs4_proc_layoutget returns void
  pnfs: defer release of pages in layoutget
  nfs: tear down caches in nfs_init_writepagecache when allocation fails

12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Wed, 22 Aug 2012 16:56:06 +0000 (09:56 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs

Pull assorted fixes - mostly vfs - from Al Viro:
 "Assorted fixes, with an unexpected detour into vfio refcounting logics
  (fell out when digging in an analog of eventpoll race in there)."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  task_work: add a scheduling point in task_work_run()
  fs: fix fs/namei.c kernel-doc warnings
  eventpoll: use-after-possible-free in epoll_create1()
  vfio: grab vfio_device reference *before* exposing the sucker via fd_install()
  vfio: get rid of vfio_device_put()/vfio_group_get_device* races
  vfio: get rid of open-coding kref_put_mutex
  introduce kref_put_mutex()
  vfio: don't dereference after kfree...
  mqueue: lift mnt_want_write() outside ->i_mutex, clean up a bit

12 years agoHID: Remove QUANTA from special drivers list
Simon Farnsworth [Wed, 22 Aug 2012 10:17:17 +0000 (11:17 +0100)]
HID: Remove QUANTA from special drivers list

This QUANTA device is driven by the generic hid-multitouch.ko driver, and
therefore shouldn't be in the special drivers list.

This has been an oversight in 4fa3a58 ("HID: hid-multitouch: Switch to
device groups").

Signed-off-by: Simon Farnsworth <simon.farnsworth@onelan.co.uk>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
12 years agoUBIFS: fix error messages spelling
Artem Bityutskiy [Wed, 22 Aug 2012 13:47:28 +0000 (16:47 +0300)]
UBIFS: fix error messages spelling

Corruptio -> corruption.

Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
12 years agotask_work: add a scheduling point in task_work_run()
Eric Dumazet [Tue, 21 Aug 2012 13:05:14 +0000 (15:05 +0200)]
task_work: add a scheduling point in task_work_run()

It seems commit 4a9d4b02 (switch fput to task_work_add) reintroduced
the problem addressed in commit 944be0b2 (close_files(): add scheduling
point)

If a server process with a lot of files (say 2 million tcp sockets)
is killed, we can spend a lot of time in task_work_run() and trigger
a soft lockup.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agofs: fix fs/namei.c kernel-doc warnings
Randy Dunlap [Sun, 19 Aug 2012 00:39:25 +0000 (17:39 -0700)]
fs: fix fs/namei.c kernel-doc warnings

Fix kernel-doc warnings in fs/namei.c:

Warning(fs/namei.c:360): No description found for parameter 'inode'
Warning(fs/namei.c:672): No description found for parameter 'nd'

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agoeventpoll: use-after-possible-free in epoll_create1()
Al Viro [Sat, 18 Aug 2012 02:42:36 +0000 (22:42 -0400)]
eventpoll: use-after-possible-free in epoll_create1()

As soon as we'd installed the file into descriptor table, it can
get closed by another thread.  Freeing ep in process...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agovfio: grab vfio_device reference *before* exposing the sucker via fd_install()
Al Viro [Sat, 18 Aug 2012 01:32:56 +0000 (21:32 -0400)]
vfio: grab vfio_device reference *before* exposing the sucker via fd_install()

It's not critical (anymore) since another thread closing the file will block
on ->device_lock before it gets to dropping the final reference, but it's
definitely cleaner that way...

Acked-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agovfio: get rid of vfio_device_put()/vfio_group_get_device* races
Al Viro [Sat, 18 Aug 2012 01:29:06 +0000 (21:29 -0400)]
vfio: get rid of vfio_device_put()/vfio_group_get_device* races

we really need to make sure that dropping the last reference happens
under the group->device_lock; otherwise a loop (under device_lock)
might find vfio_device instance that is being freed right now, has
already dropped the last reference and waits on device_lock to exclude
the sucker from the list.

Acked-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agovfio: get rid of open-coding kref_put_mutex
Al Viro [Sat, 18 Aug 2012 01:27:32 +0000 (21:27 -0400)]
vfio: get rid of open-coding kref_put_mutex

Acked-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agointroduce kref_put_mutex()
Al Viro [Sat, 18 Aug 2012 00:10:46 +0000 (20:10 -0400)]
introduce kref_put_mutex()

equivalent of
mutex_lock(mutex);
if (!kref_put(kref, release))
mutex_unlock(mutex);

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agovfio: don't dereference after kfree...
Al Viro [Fri, 17 Aug 2012 23:49:09 +0000 (19:49 -0400)]
vfio: don't dereference after kfree...

Acked-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agox86/alternatives: Fix p6 nops on non-modular kernels
Avi Kivity [Wed, 22 Aug 2012 10:03:48 +0000 (13:03 +0300)]
x86/alternatives: Fix p6 nops on non-modular kernels

Probably a leftover from the early days of self-patching, p6nops
are marked __initconst_or_module, which causes them to be
discarded in a non-modular kernel.  If something later triggers
patching, it will overwrite kernel code with garbage.

Reported-by: Tomas Racek <tracek@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Cc: Michael Tokarev <mjt@tls.msk.ru>
Cc: Borislav Petkov <borislav.petkov@amd.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: qemu-devel@nongnu.org
Cc: Anthony Liguori <anthony@codemonkey.ws>
Cc: H. Peter Anvin <hpa@linux.intel.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Alan Cox <alan@linux.intel.com>
Link: http://lkml.kernel.org/r/5034AE84.90708@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
12 years agoUBIFS: fix complaints about too small debug buffer size
Artem Bityutskiy [Tue, 21 Aug 2012 18:50:58 +0000 (21:50 +0300)]
UBIFS: fix complaints about too small debug buffer size

When debugging is enabled, we use a temporary on-stack buffer for formatting
the key strings like "(11368871, direntry, 0xcd0750)". The buffer size is
32 bytes and sometimes it is not enough to fit the key string - e.g., when
inode numbers are high. This is not fatal, but the key strings are incomplete
and UBIFS complains like this:

UBIFS assert failed in dbg_snprintf_key at 137 (pid 1)

This is a regression caused by "515315a UBIFS: fix key printing".

Fix the issue by increasing the buffer to 48 bytes.

Reported-by: Michael Hench <michaelhench@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Tested-by: Michael Hench <michaelhench@gmail.com>
Cc: stable@vger.kernel.org [v3.3+]
12 years agotime: Avoid making adjustments if we haven't accumulated anything
John Stultz [Wed, 22 Aug 2012 00:30:49 +0000 (20:30 -0400)]
time: Avoid making adjustments if we haven't accumulated anything

If update_wall_time() is called and the current offset isn't large
enough to accumulate, avoid re-calling timekeeping_adjust which may
change the clock freq and can cause 1ns inconsistencies with
CLOCK_REALTIME_COARSE/CLOCK_MONOTONIC_COARSE.

Signed-off-by: John Stultz <john.stultz@linaro.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/1345595449-34965-5-git-send-email-john.stultz@linaro.org
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
12 years agotime: Avoid potential shift overflow with large shift values
John Stultz [Wed, 22 Aug 2012 00:30:48 +0000 (20:30 -0400)]
time: Avoid potential shift overflow with large shift values

Andreas Schwab noticed that the 1 << tk->shift could overflow if the
shift value was greater than 30, since 1 would be a 32bit long on
32bit architectures. This issue was introduced by 1e75fa8be (time:
Condense timekeeper.xtime into xtime_sec)

Use 1ULL instead to ensure we don't overflow on the shift.

Reported-by: Andreas Schwab <schwab@linux-m68k.org>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Link: http://lkml.kernel.org/r/1345595449-34965-4-git-send-email-john.stultz@linaro.org
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
12 years agotime: Fix casting issue in timekeeping_forward_now
Andreas Schwab [Wed, 22 Aug 2012 00:30:47 +0000 (20:30 -0400)]
time: Fix casting issue in timekeeping_forward_now

arch_gettimeoffset returns a u32 value which when shifted by tk->shift
can overflow. This issue was introduced with 1e75fa8be (time: Condense
timekeeper.xtime into xtime_sec)

Cast it to u64 first.

Signed-off-by: Andreas Schwab <schwab@linux-m68k.org>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Link: http://lkml.kernel.org/r/1345595449-34965-3-git-send-email-john.stultz@linaro.org
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
12 years agotime: Ensure we normalize the timekeeper in tk_xtime_add
John Stultz [Wed, 22 Aug 2012 00:30:46 +0000 (20:30 -0400)]
time: Ensure we normalize the timekeeper in tk_xtime_add

Andreas noticed problems with resume on specific hardware after commit
1e75fa8b (time: Condense timekeeper.xtime into xtime_sec) combined
with commit b44d50dca (time: Fix casting issue in tk_set_xtime and
tk_xtime_add)

After some digging I realized we aren't normalizing the timekeeper
after the add. Add the missing normalize call.

Reported-by: Andreas Schwab <schwab@linux-m68k.org>
Tested-by: Andreas Schwab <schwab@linux-m68k.org>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Link: http://lkml.kernel.org/r/1345595449-34965-2-git-send-email-john.stultz@linaro.org
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
12 years agox86/fixup_irq: Use cpu_online_mask instead of cpu_all_mask
Liu, Chuansheng [Tue, 14 Aug 2012 06:55:01 +0000 (06:55 +0000)]
x86/fixup_irq: Use cpu_online_mask instead of cpu_all_mask

When one CPU is going down and this CPU is the last one in irq
affinity, current code is setting cpu_all_mask as the new
affinity for that irq.

But for some systems (such as in Medfield Android mobile) the
firmware sends the interrupt to each CPU in the irq affinity
mask, averaged, and cpu_all_mask includes all potential CPUs,
i.e. offline ones as well.

So replace cpu_all_mask with cpu_online_mask.

Signed-off-by: liu chuansheng <chuansheng.liu@intel.com>
Acked-by: Yanmin Zhang <yanmin_zhang@linux.intel.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/27240C0AC20F114CBF8149A2696CBE4A137286@SHSMSX101.ccr.corp.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
12 years agox86/spinlocks: Fix comment in spinlock.h
Richard Weinberger [Tue, 14 Aug 2012 21:47:37 +0000 (14:47 -0700)]
x86/spinlocks: Fix comment in spinlock.h

This comment is no longer true.  We support up to 2^16 CPUs
because __ticket_t is an u16 if NR_CPUS is larger than 256.

Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
12 years agofbcon: fix race condition between console lock and cursor timer (v1.1)
Dave Airlie [Tue, 21 Aug 2012 06:29:47 +0000 (16:29 +1000)]
fbcon: fix race condition between console lock and cursor timer (v1.1)

So we've had a fair few reports of fbcon handover breakage between
efi/vesafb and i915 surface recently, so I dedicated a couple of
days to finding the problem.

Essentially the last thing we saw was the conflicting framebuffer
message and that was all.

So after much tracing with direct netconsole writes (printks
under console_lock not so useful), I think I found the race.

Thread A (driver load)    Thread B (timer thread)
  unbind_con_driver ->              |
  bind_con_driver ->                |
  vc->vc_sw->con_deinit ->          |
  fbcon_deinit ->                   |
  console_lock()                    |
      |                             |
      |                       fbcon_flashcursor timer fires
      |                       console_lock() <- blocked for A
      |
      |
fbcon_del_cursor_timer ->
  del_timer_sync
  (BOOM)

Of course because all of this is under the console lock,
we never see anything, also since we also just unbound the active
console guess what we never see anything.

Hopefully this fixes the problem for anyone seeing vesafb->kms
driver handoff.

v1.1: add comment suggestion from Alan.

Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agoMerge branch 'akpm' (Andrew's patch-bomb)
Linus Torvalds [Wed, 22 Aug 2012 00:22:22 +0000 (17:22 -0700)]
Merge branch 'akpm' (Andrew's patch-bomb)

Merge fixes from Andrew Morton.

Random drivers and some VM fixes.

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (17 commits)
  mm: compaction: Abort async compaction if locks are contended or taking too long
  mm: have order > 0 compaction start near a pageblock with free pages
  rapidio/tsi721: fix unused variable compiler warning
  rapidio/tsi721: fix inbound doorbell interrupt handling
  drivers/rtc/rtc-rs5c348.c: fix hour decoding in 12-hour mode
  mm: correct page->pfmemalloc to fix deactivate_slab regression
  drivers/rtc/rtc-pcf2123.c: initialize dynamic sysfs attributes
  mm/compaction.c: fix deferring compaction mistake
  drivers/misc/sgi-xp/xpc_uv.c: SGI XPC fails to load when cpu 0 is out of IRQ resources
  string: do not export memweight() to userspace
  hugetlb: update hugetlbpage.txt
  checkpatch: add control statement test to SINGLE_STATEMENT_DO_WHILE_MACRO
  mm: hugetlbfs: correctly populate shared pmd
  cciss: fix incorrect scsi status reporting
  Documentation: update mount option in filesystem/vfat.txt
  mm: change nr_ptes BUG_ON to WARN_ON
  cs5535-clockevt: typo, it's MFGPT, not MFPGT

12 years agoMerge branch 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Tue, 21 Aug 2012 23:54:38 +0000 (16:54 -0700)]
Merge branch 'v4l_for_linus' of git://git./linux/kernel/git/mchehab/linux-media

Pull media fixes from Mauro Carvalho Chehab:
 "For bug fixes, at soc_camera, si470x, uvcvideo, iguanaworks IR driver,
  radio_shark Kbuild fixes, and at the V4L2 core (radio fixes)."

* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
  [media] media: soc_camera: don't clear pix->sizeimage in JPEG mode
  [media] media: mx2_camera: Fix clock handling for i.MX27
  [media] video: mx2_camera: Use clk_prepare_enable/clk_disable_unprepare
  [media] video: mx1_camera: Use clk_prepare_enable/clk_disable_unprepare
  [media] media: mx3_camera: buf_init() add buffer state check
  [media] radio-shark2: Only compile led support when CONFIG_LED_CLASS is set
  [media] radio-shark: Only compile led support when CONFIG_LED_CLASS is set
  [media] radio-shark*: Call cancel_work_sync from disconnect rather then release
  [media] radio-shark*: Remove work-around for dangling pointer in usb intfdata
  [media] Add USB dependency for IguanaWorks USB IR Transceiver
  [media] Add missing logging for rangelow/high of hwseek
  [media] VIDIOC_ENUM_FREQ_BANDS fix
  [media] mem2mem_testdev: fix querycap regression
  [media] si470x: v4l2-compliance fixes
  [media] DocBook: Remove a spurious character
  [media] uvcvideo: Reset the bytesused field when recycling an erroneous buffer

12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Tue, 21 Aug 2012 23:46:08 +0000 (16:46 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking update from David Miller:
 "A couple weeks of bug fixing in there.  The largest chunk is all the
  broken crap Amerigo Wang found in the netpoll layer."

 1) netpoll and it's users has several serious bugs:
    a) uses GFP_KERNEL with locks held
    b) interfaces requiring interrupts disabled are called with them
       enabled
    c) and vice versa
    d) VLAN tag demuxing, as per all other RX packet input paths, is not
       applied

    All from Amerigo Wang.

 2) Hopefully cure the ipv4 mapped ipv6 address TCP early demux bugs for
    good, from Neal Cardwell.

 3) Unlike AF_UNIX, AF_PACKET sockets don't set a default credentials
    when the user doesn't specify one explicitly during sendmsg().
    Instead we attach an empty (zero) SCM credential block which is
    definitely not what we want.  Fix from Eric Dumazet.

 4) IPv6 illegally invokes netdevice notifiers with RCU lock held, fix
    from Ben Hutchings.

 5) inet_csk_route_child_sock() checks wrong inet options pointer, fix
    from Christoph Paasch.

 6) When AF_PACKET is used for transmit, packet loopback doesn't behave
    properly when a socket fanout is enabled, from Eric Leblond.

 7) On bluetooth l2cap channel create failure, we leak the socket, from
    Jaganath Kanakkassery.

 8) Fix all the netprio file handling bugs found by Al Viro, from John
    Fastabend.

 9) Several error return and NULL deref bug fixes in networking drivers
    from Julia Lawall.

10) A large smattering of struct padding et al.  kernel memory leaks to
    userspace found of Mathias Krause.

11) Conntrack expections in netfilter can access an uninitialized timer,
    fix from Pablo Neira Ayuso.

12) Several netfilter SIP tracker bug fixes from Patrick McHardy.

13) IPSEC ipv6 routes are not initialized correctly all the time,
    resulting in an OOPS in inet_putpeer().  Also from Patrick McHardy.

14) Bridging does rcu_dereference() outside of RCU protected area, from
    Stephen Hemminger.

15) Fix routing cache removal performance regression when looking up
    output routes that have a local destination.  From Zheng Yan.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (87 commits)
  af_netlink: force credentials passing [CVE-2012-3520]
  ipv4: fix ip header ident selection in __ip_make_skb()
  ipv4: Use newinet->inet_opt in inet_csk_route_child_sock()
  tcp: fix possible socket refcount problem
  net: tcp: move sk_rx_dst_set call after tcp_create_openreq_child()
  net/core/dev.c: fix kernel-doc warning
  netconsole: remove a redundant netconsole_target_put()
  net: ipv6: fix oops in inet_putpeer()
  net/stmmac: fix issue of clk_get for Loongson1B.
  caif: Do not dereference NULL in chnl_recv_cb()
  af_packet: don't emit packet on orig fanout group
  drivers/net/irda: fix error return code
  drivers/net/wan/dscc4.c: fix error return code
  drivers/net/wimax/i2400m/fw.c: fix error return code
  smsc75xx: add missing entry to MAINTAINERS
  net: qmi_wwan: new devices: UML290 and K5006-Z
  net: sh_eth: Add eth support for R8A7779 device
  netdev/phy: skip disabled mdio-mux nodes
  dt: introduce for_each_available_child_of_node, of_get_next_available_child
  net: netprio: fix cgrp create and write priomap race
  ...

12 years agomm: compaction: Abort async compaction if locks are contended or taking too long
Mel Gorman [Tue, 21 Aug 2012 23:16:17 +0000 (16:16 -0700)]
mm: compaction: Abort async compaction if locks are contended or taking too long

Jim Schutt reported a problem that pointed at compaction contending
heavily on locks.  The workload is straight-forward and in his own words;

The systems in question have 24 SAS drives spread across 3 HBAs,
running 24 Ceph OSD instances, one per drive.  FWIW these servers
are dual-socket Intel 5675 Xeons w/48 GB memory.  I've got ~160
Ceph Linux clients doing dd simultaneously to a Ceph file system
backed by 12 of these servers.

Early in the test everything looks fine

  procs -------------------memory------------------ ---swap-- -----io---- --system-- -----cpu-------
   r  b       swpd       free       buff      cache   si   so    bi    bo   in   cs  us sy  id wa st
  31 15          0     287216        576   38606628    0    0     2  1158    2   14   1  3  95  0  0
  27 15          0     225288        576   38583384    0    0    18 2222016 203357 134876  11 56  17 15  0
  28 17          0     219256        576   38544736    0    0    11 2305932 203141 146296  11 49  23 17  0
   6 18          0     215596        576   38552872    0    0     7 2363207 215264 166502  12 45  22 20  0
  22 18          0     226984        576   38596404    0    0     3 2445741 223114 179527  12 43  23 22  0

and then it goes to pot

  procs -------------------memory------------------ ---swap-- -----io---- --system-- -----cpu-------
   r  b       swpd       free       buff      cache   si   so    bi    bo   in   cs  us sy  id wa st
  163  8          0     464308        576   36791368    0    0    11 22210  866  536   3 13  79  4  0
  207 14          0     917752        576   36181928    0    0   712 1345376 134598 47367   7 90   1  2  0
  123 12          0     685516        576   36296148    0    0   429 1386615 158494 60077   8 84   5  3  0
  123 12          0     598572        576   36333728    0    0  1107 1233281 147542 62351   7 84   5  4  0
  622  7          0     660768        576   36118264    0    0   557 1345548 151394 59353   7 85   4  3  0
  223 11          0     283960        576   36463868    0    0    46 1107160 121846 33006   6 93   1  1  0

Note that system CPU usage is very high blocks being written out has
dropped by 42%. He analysed this with perf and found

  perf record -g -a sleep 10
  perf report --sort symbol --call-graph fractal,5
    34.63%  [k] _raw_spin_lock_irqsave
            |
            |--97.30%-- isolate_freepages
            |          compaction_alloc
            |          unmap_and_move
            |          migrate_pages
            |          compact_zone
            |          compact_zone_order
            |          try_to_compact_pages
            |          __alloc_pages_direct_compact
            |          __alloc_pages_slowpath
            |          __alloc_pages_nodemask
            |          alloc_pages_vma
            |          do_huge_pmd_anonymous_page
            |          handle_mm_fault
            |          do_page_fault
            |          page_fault
            |          |
            |          |--87.39%-- skb_copy_datagram_iovec
            |          |          tcp_recvmsg
            |          |          inet_recvmsg
            |          |          sock_recvmsg
            |          |          sys_recvfrom
            |          |          system_call
            |          |          __recv
            |          |          |
            |          |           --100.00%-- (nil)
            |          |
            |           --12.61%-- memcpy
             --2.70%-- [...]

There was other data but primarily it is all showing that compaction is
contended heavily on the zone->lock and zone->lru_lock.

commit [b2eef8c0: mm: compaction: minimise the time IRQs are disabled
while isolating pages for migration] noted that it was possible for
migration to hold the lru_lock for an excessive amount of time. Very
broadly speaking this patch expands the concept.

This patch introduces compact_checklock_irqsave() to check if a lock
is contended or the process needs to be scheduled. If either condition
is true then async compaction is aborted and the caller is informed.
The page allocator will fail a THP allocation if compaction failed due
to contention. This patch also introduces compact_trylock_irqsave()
which will acquire the lock only if it is not contended and the process
does not need to schedule.

Reported-by: Jim Schutt <jaschut@sandia.gov>
Tested-by: Jim Schutt <jaschut@sandia.gov>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomm: have order > 0 compaction start near a pageblock with free pages
Mel Gorman [Tue, 21 Aug 2012 23:16:15 +0000 (16:16 -0700)]
mm: have order > 0 compaction start near a pageblock with free pages

Commit 7db8889ab05b ("mm: have order > 0 compaction start off where it
left") introduced a caching mechanism to reduce the amount work the free
page scanner does in compaction.  However, it has a problem.  Consider
two process simultaneously scanning free pages

     C
Process A M     S      F
|---------------------------------------|
Process B M  FS

C is zone->compact_cached_free_pfn
S is cc->start_pfree_pfn
M is cc->migrate_pfn
F is cc->free_pfn

In this diagram, Process A has just reached its migrate scanner, wrapped
around and updated compact_cached_free_pfn accordingly.

Simultaneously, Process B finishes isolating in a block and updates
compact_cached_free_pfn again to the location of its free scanner.

Process A moves to "end_of_zone - one_pageblock" and runs this check

                if (cc->order > 0 && (!cc->wrapped ||
                                      zone->compact_cached_free_pfn >
                                      cc->start_free_pfn))
                        pfn = min(pfn, zone->compact_cached_free_pfn);

compact_cached_free_pfn is above where it started so the free scanner
skips almost the entire space it should have scanned.  When there are
multiple processes compacting it can end in a situation where the entire
zone is not being scanned at all.  Further, it is possible for two
processes to ping-pong update to compact_cached_free_pfn which is just
random.

Overall, the end result wrecks allocation success rates.

There is not an obvious way around this problem without introducing new
locking and state so this patch takes a different approach.

First, it gets rid of the skip logic because it's not clear that it
matters if two free scanners happen to be in the same block but with
racing updates it's too easy for it to skip over blocks it should not.

Second, it updates compact_cached_free_pfn in a more limited set of
circumstances.

If a scanner has wrapped, it updates compact_cached_free_pfn to the end
of the zone. When a wrapped scanner isolates a page, it updates
compact_cached_free_pfn to point to the highest pageblock it
can isolate pages from.

If a scanner has not wrapped when it has finished isolated pages it
checks if compact_cached_free_pfn is pointing to the end of the
zone. If so, the value is updated to point to the highest
pageblock that pages were isolated from. This value will not
be updated again until a free page scanner wraps and resets
compact_cached_free_pfn.

This is not optimal and it can still race but the compact_cached_free_pfn
will be pointing to or very near a pageblock with free pages.

Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agorapidio/tsi721: fix unused variable compiler warning
Alexandre Bounine [Tue, 21 Aug 2012 23:16:12 +0000 (16:16 -0700)]
rapidio/tsi721: fix unused variable compiler warning

Fix unused variable compiler warning when built with CONFIG_RAPIDIO_DEBUG
option off.

This patch is applicable to kernel versions starting from v3.2

Signed-off-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agorapidio/tsi721: fix inbound doorbell interrupt handling
Alexandre Bounine [Tue, 21 Aug 2012 23:16:11 +0000 (16:16 -0700)]
rapidio/tsi721: fix inbound doorbell interrupt handling

Make sure that there is no doorbell messages left behind due to disabled
interrupts during inbound doorbell processing.

The most common case for this bug is loss of rionet JOIN messages in
systems with three or more rionet participants and MSI or MSI-X enabled.
As result, requests for packet transfers may finish with "destination
unreachable" error message.

This patch is applicable to kernel versions starting from v3.2.

Signed-off-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrivers/rtc/rtc-rs5c348.c: fix hour decoding in 12-hour mode
Atsushi Nemoto [Tue, 21 Aug 2012 23:16:10 +0000 (16:16 -0700)]
drivers/rtc/rtc-rs5c348.c: fix hour decoding in 12-hour mode

Correct the offset by subtracting 20 from tm_hour before taking the
modulo 12.

[ "Why 20?" I hear you ask. Or at least I did.

  Here's the reason why: RS5C348_BIT_PM is 32, and is - stupidly -
  included in the RS5C348_HOURS_MASK define.  So it's really subtracting
  out that bit to get "hour+12".  But then because it does things modulo
  12, it needs to add the 12 in again afterwards anyway.

  This code is confused.  It would be much clearer if RS5C348_HOURS_MASK
  just didn't include the RS5C348_BIT_PM bit at all, then it wouldn't
  need to do the silly subtract either.

  Whatever. It's all just math, the end result is the same.   - Linus ]

Reported-by: James Nute <newten82@gmail.com>
Tested-by: James Nute <newten82@gmail.com>
Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomm: correct page->pfmemalloc to fix deactivate_slab regression
Alex Shi [Tue, 21 Aug 2012 23:16:08 +0000 (16:16 -0700)]
mm: correct page->pfmemalloc to fix deactivate_slab regression

Commit cfd19c5a9ecf ("mm: only set page->pfmemalloc when
ALLOC_NO_WATERMARKS was used") tried to narrow down page->pfmemalloc
setting, but it missed some places the pfmemalloc should be set.

So, in __slab_alloc, the unalignment pfmemalloc and ALLOC_NO_WATERMARKS
cause incorrect deactivate_slab() on our core2 server:

    64.73%           fio  [kernel.kallsyms]     [k] _raw_spin_lock
                     |
                     --- _raw_spin_lock
                        |
                        |---0.34%-- deactivate_slab
                        |          __slab_alloc
                        |          kmem_cache_alloc
                        |          |

That causes our fio sync write performance to have a 40% regression.

Move the checking in get_page_from_freelist() which resolves this issue.

Signed-off-by: Alex Shi <alex.shi@intel.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Cc: David Miller <davem@davemloft.net
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Tested-by: Eric Dumazet <eric.dumazet@gmail.com>
Tested-by: Sage Weil <sage@inktank.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrivers/rtc/rtc-pcf2123.c: initialize dynamic sysfs attributes
Ilya Shchepetkov [Tue, 21 Aug 2012 23:16:06 +0000 (16:16 -0700)]
drivers/rtc/rtc-pcf2123.c: initialize dynamic sysfs attributes

Dynamically allocated sysfs attributes must be initialized using
sysfs_attr_init(), otherwise lockdep complains: BUG: key <address> not in
.data!

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Ilya Shchepetkov <shchepetkov@ispras.ru>
Cc: Chris Verges <chrisv@cyberswitching.com>
Cc: Christian Pellegrin <chripell@fsfe.org>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomm/compaction.c: fix deferring compaction mistake
Minchan Kim [Tue, 21 Aug 2012 23:16:03 +0000 (16:16 -0700)]
mm/compaction.c: fix deferring compaction mistake

Commit aff622495c9a ("vmscan: only defer compaction for failed order and
higher") fixed bad deferring policy but made mistake about checking
compact_order_failed in __compact_pgdat().  So it can't update
compact_order_failed with the new order.  This ends up preventing
correct operation of policy deferral.  This patch fixes it.

Signed-off-by: Minchan Kim <minchan@kernel.org>
Reviewed-by: Rik van Riel <riel@redhat.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrivers/misc/sgi-xp/xpc_uv.c: SGI XPC fails to load when cpu 0 is out of IRQ resources
Robin Holt [Tue, 21 Aug 2012 23:16:02 +0000 (16:16 -0700)]
drivers/misc/sgi-xp/xpc_uv.c: SGI XPC fails to load when cpu 0 is out of IRQ resources

On many of our larger systems, CPU 0 has had all of its IRQ resources
consumed before XPC loads.  Worst cases on machines with multiple 10
GigE cards and multiple IB cards have depleted the entire first socket
of IRQs.

This patch makes selecting the node upon which IRQs are allocated (as
well as all the other GRU Message Queue structures) specifiable as a
module load param and has a default behavior of searching all nodes/cpus
for an available resources.

[akpm@linux-foundation.org: fix build: include cpu.h and module.h]
Signed-off-by: Robin Holt <holt@sgi.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agostring: do not export memweight() to userspace
WANG Cong [Tue, 21 Aug 2012 23:16:00 +0000 (16:16 -0700)]
string: do not export memweight() to userspace

Fix the following warning:

  usr/include/linux/string.h:8: userspace cannot reference function or variable defined in the kernel

Signed-off-by: WANG Cong <xiyou.wangcong@gmail.com>
Acked-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agohugetlb: update hugetlbpage.txt
Zhouping Liu [Tue, 21 Aug 2012 23:15:57 +0000 (16:15 -0700)]
hugetlb: update hugetlbpage.txt

Commit f0f57b2b1488 ("mm: move hugepage test examples to
tools/testing/selftests/vm") moved map_hugetlb.c, hugepage-shm.c and
hugepage-mmap.c tests into tools/testing/selftests/vm/ directory, but it
didn't update hugetlbpage.txt

Signed-off-by: Zhouping Liu <sanweidaying@gmail.com>
Acked-by: Dave Young <dyoung@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agocheckpatch: add control statement test to SINGLE_STATEMENT_DO_WHILE_MACRO
Joe Perches [Tue, 21 Aug 2012 23:15:53 +0000 (16:15 -0700)]
checkpatch: add control statement test to SINGLE_STATEMENT_DO_WHILE_MACRO

Commit b13edf7ff2dd ("checkpatch: add checks for do {} while (0) macro
misuses") added a test that is overly simplistic for single statement
macros.

Macros that start with control tests should be enclosed in a do {} while
(0) loop.

Add the necessary control tests to the check.

Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Andy Whitcroft <apw@canonical.com>
Tested-by: Franz Schrober <franzschrober@yahoo.de>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>