firefly-linux-kernel-4.4.55.git
10 years agospmi: Remove duplicate inclusion of module.h
Sachin Kamat [Mon, 23 Jun 2014 11:21:27 +0000 (16:51 +0530)]
spmi: Remove duplicate inclusion of module.h

module.h was included twice.

Signed-off-by: Sachin Kamat <sachin.kamat@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoGenWQE: Remove unnecessary include
Kleber Sacilotto de Souza [Thu, 10 Jul 2014 13:17:08 +0000 (10:17 -0300)]
GenWQE: Remove unnecessary include

The include for the UAPI header file from card_base.c can be removed
since it's already included on card_base.h.

Signed-off-by: Kleber Sacilotto de Souza <klebers@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoti-st: st-kim: Dont let probe fail when debugfs is disabled
Robin van der Gracht [Mon, 16 Jun 2014 14:38:56 +0000 (16:38 +0200)]
ti-st: st-kim: Dont let probe fail when debugfs is disabled

Signed-off-by: Robin van der Gracht <robin@protonic.nl>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agouio: uio_pruss: use struct device
Andre Heider [Sun, 29 Jun 2014 16:21:35 +0000 (18:21 +0200)]
uio: uio_pruss: use struct device

Get rid of the repeating &dev->dev constructs and prevent introducing
new ones.

Signed-off-by: Andre Heider <a.heider@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoipoctal: request_irq after configuration
Federico Vaga [Thu, 3 Jul 2014 08:53:58 +0000 (10:53 +0200)]
ipoctal: request_irq after configuration

The request for an IRQ handler must be done after whole configuration. This
was not the case for this driver which request the IRQ in the middle of
the configuration. Sometimes, it happens that something is not completely
configured, we recieve an interrupt thus we stumble into troubles in the
IRQ handler.

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
Acked-by: Samuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoipoctal: protect only the real critical section
Federico Vaga [Thu, 26 Jun 2014 07:46:24 +0000 (09:46 +0200)]
ipoctal: protect only the real critical section

In some conditions (echo or particular sequence of special
characters), on buffer push, the tty layer calls the write operation
while we are holding the spinlock. This means deadlock within the same
process on kernels version < 3.12. It seems not a problem on recent
kernel, but the patch still valid as locking optimization.

The protected variables by the spinlock are: xmit_buf, nb_bytes,
pointer_read and pointer_write. So, this patch reduces the locked area
in the IRQ handler only to these variables. Most of the code inside the
locked area in the IRQ handler is not protected elsewhere; it means
that it is not protected at all.

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
Acked-by: Samuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agomfd: vexpress: fix error handling vexpress_syscfg_regmap_init()
Dan Carpenter [Wed, 11 Jun 2014 10:17:41 +0000 (13:17 +0300)]
mfd: vexpress: fix error handling vexpress_syscfg_regmap_init()

This function should be returning an ERR_PTR() on failure instead of
NULL.  Also there is a use after free bug if regmap_init() fails because
we free "func" and then dereference doing the return.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Pawel Moll <pawel.moll@arm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agochar: xilinx_hwicap: missing error code if ioremap() fails
Dan Carpenter [Wed, 11 Jun 2014 06:04:33 +0000 (09:04 +0300)]
char: xilinx_hwicap: missing error code if ioremap() fails

Return -ENOMEM instead of success if ioremap() fails.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agobsr: avoid format string leaking into device name
Kees Cook [Tue, 10 Jun 2014 17:51:08 +0000 (10:51 -0700)]
bsr: avoid format string leaking into device name

This makes sure a format string cannot accidentally leak into a device name.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoi8k: Add support for Dell Precision 490 and Latitude D520
Guenter Roeck [Sat, 21 Jun 2014 15:08:12 +0000 (08:08 -0700)]
i8k: Add support for Dell Precision 490 and Latitude D520

Both systems need non-standard parameters for fan multiplier
and maximum fan speed.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Cc: Andreas Mohr <andi@lisas.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoi8k: uapi: Introduce define for new highest fan speed
Guenter Roeck [Sat, 21 Jun 2014 15:08:11 +0000 (08:08 -0700)]
i8k: uapi: Introduce define for new highest fan speed

Some Dell laptops support fan speeds of {0, 1, 2, 3} instead of {0, 1, 2}.
Add a define for it.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Cc: Andreas Mohr <andi@lisas.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoi8k: Add support for configurable maximum fan speed value
Guenter Roeck [Sat, 21 Jun 2014 15:08:10 +0000 (08:08 -0700)]
i8k: Add support for configurable maximum fan speed value

Newer Dell systems provide more granular fan speed selection.
Add support for it.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Cc: Andreas Mohr <andi@lisas.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoi8k: Drop all labels
Guenter Roeck [Sat, 21 Jun 2014 15:08:09 +0000 (08:08 -0700)]
i8k: Drop all labels

Labels are known to be wrong for several Dell laptops.
For example, a single fan may be shown as right fan when in reality
it sits on the left side of the chassis. Drop all labels to avoid
such inaccuracies. Users can select labels in the sensors configuration
file instead if desired.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Cc: Andreas Mohr <andi@lisas.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoVME: remove duplicate CA91CX42_DCTL_VDW_M define
Dan Carpenter [Mon, 9 Jun 2014 15:12:12 +0000 (18:12 +0300)]
VME: remove duplicate CA91CX42_DCTL_VDW_M define

The CA91CX42_DCTL_VDW_M define is cut and pasted twice so we can delete
the second instance.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agopcmcia: journada720: use sa1100 pin interfaces correctly
Arnd Bergmann [Thu, 5 Jun 2014 21:05:47 +0000 (23:05 +0200)]
pcmcia: journada720: use sa1100 pin interfaces correctly

commit dabd14684bc2 "PCMCIA: sa1111: remove duplicated initializers"
incorrectly moved some code into the pcmcia_jornada720_init, causing
a few build errors, and for unknown reasons, the driver lacks
an inclusion of <linux/io.h>, so we get the build errors, and more:

sa1111_jornada720.c: In function 'pcmcia_jornada720_init':
sa1111_jornada720.c:101:3: error: implicit declaration of function 'IOMEM' [-Werror=implicit-function-declaration]
   GRER |= 0x00000002;
   ^
sa1111_jornada720.c:104:3: warning: passing argument 1 of 'sa1111_set_io_dir' from incompatible pointer type [enabled by default]
   sa1111_set_io_dir(dev, pin, 0, 0);
   ^

This patch uses the SA1111_DEV() to convert the dev pointer to the
correct type before passing it and adds the missing include.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: Kristoffer Ericson <kristoffer.ericson@gmail.com>
Cc: linux-pcmcia@lists.infradead.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agopcmcia: pxa2xx: fix logic for lubbock
Arnd Bergmann [Thu, 5 Jun 2014 21:05:48 +0000 (23:05 +0200)]
pcmcia: pxa2xx: fix logic for lubbock

The lubbock platform uses the sa1111 companion chip with a pxa250
CPU, which means it requires both the PCMCIA_SA1111 and the
PCMCIA_PXA2XX code to be built into the kernel. Unfortunately,
the Makefile and Kconfig don't agree on how this is accomplished,
leading to a situation where you get this link error when building
a lubbock kernel with PCMCIA_SA1111 enabled but PCMCIA_PXA2XX
disabled:

ERROR: "pxa2xx_configure_sockets" [drivers/pcmcia/sa1111_cs.ko] undefined!
ERROR: "pxa2xx_drv_pcmcia_ops" [drivers/pcmcia/sa1111_cs.ko] undefined!
ERROR: "pxa2xx_drv_pcmcia_add_one" [drivers/pcmcia/sa1111_cs.ko] undefined!

This patch changes the Kconfig code to disallow that particular
configuration.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agopcmcia: sa1100: H3100 and H3600 share a driver
Arnd Bergmann [Thu, 5 Jun 2014 21:05:49 +0000 (23:05 +0200)]
pcmcia: sa1100: H3100 and H3600 share a driver

When building a iPAQ H3100-only kernel with PCMCIA enabled,
we get this build error:

ERROR: "pcmcia_h3600_init" [drivers/pcmcia/sa1100_cs.ko] undefined!

The defconfig normally works fine because it enables both H3100
and H3600 support. This patch fixes the Makefile to build the
driver if at least one of the two machines are selected.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agohyperv: remove meaningless pr_err() in vmbus_recvpacket_raw()
Jason Wang [Mon, 30 Jun 2014 06:14:48 +0000 (14:14 +0800)]
hyperv: remove meaningless pr_err() in vmbus_recvpacket_raw()

All its callers depends on the return value of -ENOBUFS to reallocate a
bigger buffer and retry the receiving. So there's no need to call
pr_err() here since it was not a real issue, otherwise syslog will be
flooded by this false warning.

Cc: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoTools: hv: fix file overwriting of hv_fcopy_daemon
Yue Zhang [Sat, 28 Jun 2014 01:19:48 +0000 (18:19 -0700)]
Tools: hv: fix file overwriting of hv_fcopy_daemon

hv_fcopy_daemon fails to overwrite a file if the target file already
exits.

Add O_TRUNC flag on opening.

Signed-off-by: Yue Zhang <yuezha@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoAdd MODULE_DESCRIPTION to dummy-irq.c and lkdtm.c in drivers/misc
Terry Chia [Wed, 2 Jul 2014 13:02:25 +0000 (21:02 +0800)]
Add MODULE_DESCRIPTION to dummy-irq.c and lkdtm.c in drivers/misc

This starts to address
https://bugzilla.kernel.org/show_bug.cgi?id=10770

Signed-off-by: Terry Chia <terrycwk1994@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agodrivers/misc/carma/carma-fpga.c: use PTR_ERR_OR_ZERO
Fabian Frederick [Sun, 15 Jun 2014 06:22:42 +0000 (08:22 +0200)]
drivers/misc/carma/carma-fpga.c: use PTR_ERR_OR_ZERO

replace IS_ERR/PTR_ERR

Cc: Grant Likely <grant.likely@linaro.org>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agomei: add WPT second mei interface
Alexander Usyskin [Mon, 23 Jun 2014 12:10:37 +0000 (15:10 +0300)]
mei: add WPT second mei interface

Add WPT second mei interface.

Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agomei: sysfs: add Documentation mei class attributes
Tomas Winkler [Mon, 23 Jun 2014 12:10:36 +0000 (15:10 +0300)]
mei: sysfs: add Documentation mei class attributes

Add sysfs attributes Documentation entries
for /sys/class/mei

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agomei: move from misc to char device
Alexander Usyskin [Mon, 23 Jun 2014 12:10:35 +0000 (15:10 +0300)]
mei: move from misc to char device

We need to support more then one mei interface
hence the simple misc devices is not longer an option

In order not break the user space a device with pci function 0
need to be linked to /dev/mei

Signed-off-by: Alexander Usyskin <alexander.usyskin@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agomisc/GenWQE: fix pci_enable_msi usage
Sebastian Ott [Wed, 9 Jul 2014 10:46:39 +0000 (12:46 +0200)]
misc/GenWQE: fix pci_enable_msi usage

GenWQE used to call pci_enable_msi_block to allocate a desired number
of MSI's. If that was not possible pci_enable_msi_block returned with a
smaller number which might be possible to allocate. GenWQE then called
pci_enable_msi_block with that number.

Since commit a30d0108b
"GenWQE: Use pci_enable_msi_exact() instead of pci_enable_msi_block()"
pci_enable_msi_exact is used which fails if the desired number of MSI's
was not possible to allocate. Change GenWQE to use pci_enable_msi_range
to restore the old behavior.

Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Reviewed-by: Alexander Gordeev <agordeev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoGenWQE: Increase driver version number
Kleber Sacilotto de Souza [Wed, 4 Jun 2014 13:57:53 +0000 (10:57 -0300)]
GenWQE: Increase driver version number

Increase genwqe driver version number.

Signed-off-by: Kleber Sacilotto de Souza <klebers@linux.vnet.ibm.com>
Acked-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoGenWQE: Improve hardware error recovery
Kleber Sacilotto de Souza [Wed, 4 Jun 2014 13:57:52 +0000 (10:57 -0300)]
GenWQE: Improve hardware error recovery

Currently, in the event of a fatal hardware error, the driver tries a
recovery procedure that calls pci_reset_function() to reset the card.
This is not sufficient in some cases, needing a fundamental reset to
bring the card back.

This patch implements a call to the platform fundamental reset procedure
on the error recovery path if GENWQE_PLATFORM_ERROR_RECOVERY is enabled.
This is implemented by default only on PPC64, since this can cause
problems on other archs, e.g. zSeries, where the platform has its own
recovery procedures, leading to a potencial race conditition. For these
cases, the recovery is kept as it was before.

Signed-off-by: Kleber Sacilotto de Souza <klebers@linux.vnet.ibm.com>
Acked-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoGenWQE: Add support for EEH error recovery
Kleber Sacilotto de Souza [Wed, 4 Jun 2014 13:57:51 +0000 (10:57 -0300)]
GenWQE: Add support for EEH error recovery

This patch implements the callbacks and functions necessary to have EEH
recovery support.

It adds a config option to enable or disable explicit calls to trigger
platform specific mechanisms on error recovery paths. This option is
enabled by default only on PPC64 systems and can be overritten via
debugfs. If this option is enabled, on the error recovery path the
driver will call pci_channel_offline() to check for error condition and
issue non-raw MMIO reads to trigger early EEH detection in case of
hardware failures. This is necessary since the driver MMIO helper
funtions use raw accessors.

Signed-off-by: Kleber Sacilotto de Souza <klebers@linux.vnet.ibm.com>
Acked-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoGenWQE: Add sysfs interface for bitstream reload
Kleber Sacilotto de Souza [Wed, 4 Jun 2014 13:57:50 +0000 (10:57 -0300)]
GenWQE: Add sysfs interface for bitstream reload

This patch adds an interface on sysfs for userspace to request a card
bitstream reload. It sets the appropriate register and try to perform a
fundamental reset on the PCIe slot for the card to reload the bitstream
from the chosen partition.

Signed-off-by: Kleber Sacilotto de Souza <klebers@linux.vnet.ibm.com>
Acked-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Correct the size argument to devm_kzalloc
Himangi Saraogi [Sun, 6 Jul 2014 16:13:42 +0000 (21:43 +0530)]
thunderbolt: Correct the size argument to devm_kzalloc

nhi->rx_rings does not have type as struct tb_ring *, as it is a
double pointer so the elements of the array should have pointer type,
not structure type.

The Coccinelle semantic patch that makes this change is as follows:

// <smpl>
@disable sizeof_type_expr@
type T;
T **x;
@@

  x =
  <+...sizeof(
- T
+ *x
  )...+>
// </smpl>

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Cc: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agow1: select crc16 library for ds2406
Arnd Bergmann [Mon, 23 Jun 2014 14:31:47 +0000 (16:31 +0200)]
w1: select crc16 library for ds2406

commit 94859308a21b "w1: new w1_ds2406 driver" added a new driver
that uses the crc16 library, but didn't ensure that the core is
there. This adds the necessary Kconfig statements, just like we
have it for other w1 drivers.

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Scott Alfter <scott@alfter.us>
Cc: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: select CRC32 in Kconfig
Andreas Noever [Sat, 21 Jun 2014 10:15:44 +0000 (12:15 +0200)]
thunderbolt: select CRC32 in Kconfig

We use __crc32c_le in ctl.c. So make sure that the dependency is there.

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Make tb_eeprom_get_drom_offset static
Andreas Noever [Fri, 20 Jun 2014 19:42:25 +0000 (21:42 +0200)]
thunderbolt: Make tb_eeprom_get_drom_offset static

tb_eeprom_get_drom_offset is local to this file.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Make enum tb_drom_entry_type unsigned
Andreas Noever [Fri, 20 Jun 2014 19:42:24 +0000 (21:42 +0200)]
thunderbolt: Make enum tb_drom_entry_type unsigned

Force enum tb_drom_entry_type to unsigned to fix the following error:

drivers/thunderbolt/eeprom.c:202:39: error: dubious one-bit signed bitfield

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Fix header declaration of tb_find_cap
Andreas Noever [Fri, 20 Jun 2014 19:42:23 +0000 (21:42 +0200)]
thunderbolt: Fix header declaration of tb_find_cap

tb_find_cap in cap.c takes an enum tb_cap and not an u32. Fix the
declaration in tb.h.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Add casts to prevent endianness warnings
Andreas Noever [Fri, 20 Jun 2014 19:42:22 +0000 (21:42 +0200)]
thunderbolt: Add casts to prevent endianness warnings

Thunderbolt packets are big endian. Cast pkg->buffer to __be32* when
accessing the checksum.

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: fix format string for size_t
Arnd Bergmann [Fri, 20 Jun 2014 13:52:11 +0000 (15:52 +0200)]
thunderbolt: fix format string for size_t

The result of "sizeof(struct tb_drom_entry_port)" is a size_t, which
is not necessarily the same as 'long', so we should use the appropriate
%z format string instead of %l.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: add PCI dependency
Arnd Bergmann [Fri, 20 Jun 2014 13:52:09 +0000 (15:52 +0200)]
thunderbolt: add PCI dependency

The thunderbolt drivers cannot be built if CONFIG_PCI is disabled,
better add an explicit Kconfig dependency.
The "default no" line is redundant and can be removed at the same
time.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Staticize nhi_ids
Sachin Kamat [Fri, 20 Jun 2014 09:02:34 +0000 (14:32 +0530)]
thunderbolt: Staticize nhi_ids

'nhi_ids' is local to this file.

Signed-off-by: Sachin Kamat <sachin.kamat@samsung.com>
Acked-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Use NULL instead of 0 in nhi.c
Sachin Kamat [Fri, 20 Jun 2014 09:02:33 +0000 (14:32 +0530)]
thunderbolt: Use NULL instead of 0 in nhi.c

'descriptors' is a pointer. Use NULL isntead of 0.

Signed-off-by: Sachin Kamat <sachin.kamat@samsung.com>
Acked-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Use NULL instead of 0 in ctl.c
Sachin Kamat [Fri, 20 Jun 2014 09:02:32 +0000 (14:32 +0530)]
thunderbolt: Use NULL instead of 0 in ctl.c

The function returns a pointer. Hence return NULL instead of 0.

Signed-off-by: Sachin Kamat <sachin.kamat@samsung.com>
Acked-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Use NULL instead of 0 in switch.c
Sachin Kamat [Fri, 20 Jun 2014 09:02:31 +0000 (14:32 +0530)]
thunderbolt: Use NULL instead of 0 in switch.c

The function returns a pointer. Hence return NULL instead of 0.

Signed-off-by: Sachin Kamat <sachin.kamat@samsung.com>
Acked-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Fix build error in switch.c
Sachin Kamat [Fri, 20 Jun 2014 09:02:30 +0000 (14:32 +0530)]
thunderbolt: Fix build error in switch.c

Fixes the below error:
drivers/thunderbolt/switch.c:347:2: error: implicit declaration of function â€˜kzalloc’ [-Werror=implicit-function-declaration]
drivers/thunderbolt/switch.c:381:2: error: implicit declaration of function â€˜kcalloc’ [-Werror=implicit-function-declaration]

Signed-off-by: Sachin Kamat <sachin.kamat@samsung.com>
Acked-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Fix build error in eeprom.c
Sachin Kamat [Fri, 20 Jun 2014 09:02:29 +0000 (14:32 +0530)]
thunderbolt: Fix build error in eeprom.c

Fixes the below error:
drivers/thunderbolt/eeprom.c:407:2: error: implicit declaration of function â€˜kzalloc’ [-Werror=implicit-function-declaration]
drivers/thunderbolt/eeprom.c:444:2: error: implicit declaration of function â€˜kfree’ [-Werror=implicit-function-declaration]

Signed-off-by: Sachin Kamat <sachin.kamat@samsung.com>
Acked-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agow1/masters: use pr_* instead of printk
Fjodor Schelichow [Thu, 19 Jun 2014 00:52:01 +0000 (02:52 +0200)]
w1/masters: use pr_* instead of printk

This patch replaces all calls to the "printk" function within the "masters"
subdirectory by calls to the appropriate "pr_*" function thus addressing
the following warning generated by the checkpatch script:

WARNING: Prefer [subsystem eg: netdev]_err([subsystem]dev, ...
        then dev_err(dev, ... then pr_err(...  to printk(KERN_ERR ...

Signed-off-by: Fjodor Schelichow <fjodor.schelichow@hotmail.com>
Signed-off-by: Roman Sommer <romsom2@yahoo.de>
Acked-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agow1: use pr_* instead of printk
Fjodor Schelichow [Thu, 19 Jun 2014 00:52:00 +0000 (02:52 +0200)]
w1: use pr_* instead of printk

This patch replaces all calls to the "printk" function within the main "w1"
directory by calls to the appropriate "pr_*" function thus addressing
the following warning generated by the checkpatch script:

WARNING: Prefer [subsystem eg: netdev]_err([subsystem]dev, ...
        then dev_err(dev, ... then pr_err(...  to printk(KERN_ERR ...

Signed-off-by: Fjodor Schelichow <fjodor.schelichow@hotmail.com>
Signed-off-by: Roman Sommer <romsom2@yahoo.de>
Acked-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agow1/slaves: use pr_* instead of printk
Fjodor Schelichow [Thu, 19 Jun 2014 00:51:59 +0000 (02:51 +0200)]
w1/slaves: use pr_* instead of printk

This patch replaces all calls to the "printk" function within the "slaves"
subdirectory by calls to the appropriate "pr_*" function thus addressing
the following warning generated by the checkpatch script:

WARNING: Prefer [subsystem eg: netdev]_err([subsystem]dev, ...
        then dev_err(dev, ... then pr_err(...  to printk(KERN_ERR ...

Signed-off-by: Fjodor Schelichow <fjodor.schelichow@hotmail.com>
Signed-off-by: Roman Sommer <romsom2@yahoo.de>
Acked-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agow1: new w1_ds2406 driver
Scott Alfter [Tue, 17 Jun 2014 20:42:05 +0000 (20:42 +0000)]
w1: new w1_ds2406 driver

Some preliminary work at making use of this driver led me to implement
CRC-16 checks on read and write to deal with the occasional glitchiness of
the 1-Wire bus.  The revised driver (attached) returns an I/O error if the
CRC check fails.  When reading the chip's state, either you get a valid
indication or you get an I/O error.  When changing its state, either the
change is successful or an I/O error is returned.

Signed-off-by: Scott Alfter <scott@alfter.us>
Acked-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agodrivers/w1/w1_int.c: Fix style errors.
Thomas Wood [Fri, 30 May 2014 23:10:20 +0000 (16:10 -0700)]
drivers/w1/w1_int.c: Fix style errors.

Replace spaces at beginning of the string with tabs, and replace foo * bar with foo *bar in a pointer declaration.

Signed-off-by: Thomas Wood <tommyandrena@gmail.com>
Acked-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agow1: mxc_w1: Optimize mxc_w1_ds2_touch_bit()
Alexander Shiyan [Thu, 8 May 2014 07:56:40 +0000 (11:56 +0400)]
w1: mxc_w1: Optimize mxc_w1_ds2_touch_bit()

According to the i.MX reference manual, the read/write bit operations
takes from 60 us to 120 us.
This patch optimizes mxc_w1_ds2_touch_bit() function to use proper
value for such delay. Nevertheless, a small margin for the timeout has
been added for the case if clock frequency is inaccurate.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agow1: mxc_w1: Perform a software reset at startup
Alexander Shiyan [Thu, 8 May 2014 07:56:39 +0000 (11:56 +0400)]
w1: mxc_w1: Perform a software reset at startup

This patch adds a software reset for 1-Wire module at driver startup.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agow1: mxc_w1: Optimize mxc_w1_ds2_reset_bus()
Alexander Shiyan [Thu, 8 May 2014 07:56:38 +0000 (11:56 +0400)]
w1: mxc_w1: Optimize mxc_w1_ds2_reset_bus()

According to the i.MX reference manual, the reset procedure and
"presence" pulse takes 511 and 512 us, respectively. Measurement for
i.MX27 is about 1100 us. There is no need to wait Reset+Presence
more than this time.
This patch optimizes mxc_w1_ds2_reset_bus() function to use proper
value for delay after w1 bus reset. Nevertheless, a small margin for
the timeout has been added for the case if clock frequency is inaccurate.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agow1: mxc_w1: Fix incorrect "presence" status
Alexander Shiyan [Thu, 8 May 2014 07:56:37 +0000 (11:56 +0400)]
w1: mxc_w1: Fix incorrect "presence" status

W1 reset_bus() should return zero if slave device is present.
This patch fix this issue.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Fix nontrivial endpoint devices.
Andreas Noever [Thu, 12 Jun 2014 21:11:47 +0000 (23:11 +0200)]
thunderbolt: Fix nontrivial endpoint devices.

Fix issues observed with the Startech docking station:

Fix the type of the route parameter in tb_ctl_rx. It should be u64 and not
u8 (which only worked for short routes).

A thunderbolt cable contains two lanes. If both endpoints support it a
connection will be established on both lanes. Previously we tried to
scan below both "dual link ports". Use the information extracted from
the drom to only scan behind ports with lane_nr == 0.

Endpoints with more complex thunderbolt controllers have some of their
ports disabled (for example the NHI port or one of the HDMI/DP ports).
Accessing them results in an error so we now ignore ports which are
marked as disabled in the drom.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Read port configuration from eeprom.
Andreas Noever [Thu, 12 Jun 2014 21:11:46 +0000 (23:11 +0200)]
thunderbolt: Read port configuration from eeprom.

All Thunderbolt switches (except the root switch) contain a drom which
contains information about the device. Right now we only read the UID.

Add code to read and parse this drom. For now we are only interested in
which ports are disabled and which ports are "dual link ports" (a
physical thunderbolt port/socket contains two such ports).

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Add suspend/hibernate support
Andreas Noever [Tue, 3 Jun 2014 20:04:12 +0000 (22:04 +0200)]
thunderbolt: Add suspend/hibernate support

We use _noirq since we have to restore the pci tunnels before the pci
core wakes the tunneled devices.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Read switch uid from EEPROM
Andreas Noever [Tue, 3 Jun 2014 20:04:11 +0000 (22:04 +0200)]
thunderbolt: Read switch uid from EEPROM

Add eeprom access code and read the uid during switch initialization.
The UID will be used to check device identity after suspend.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoPCI: Suspend/resume quirks for Apple thunderbolt
Andreas Noever [Tue, 3 Jun 2014 20:04:10 +0000 (22:04 +0200)]
PCI: Suspend/resume quirks for Apple thunderbolt

Add two quirks to support thunderbolt suspend/resume on Apple systems.
We need to perform two different actions during suspend and resume:

The whole controller has to be powered down before suspend. If this is
not done then the native host interface device will be gone after resume
if a thunderbolt device was plugged in before suspending. The controller
represents itself as multiple PCI devices/bridges. To power it down we
hook into the upstream bridge of the controller and call the magic ACPI
methods.  Power will be restored automatically during resume (by the
firmware presumably).

During resume we have to wait for the native host interface to
reestablish all pci tunnels. Since there is no parent-child relationship
between the NHI and the bridges we have to explicitly wait for them
using device_pm_wait_for_dev. We do this in the resume_noirq phase of
the downstream bridges of the controller (which lead into the
thunderbolt tunnels).

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoPCI: Add pci_fixup_suspend_late quirk pass
Andreas Noever [Tue, 3 Jun 2014 20:04:09 +0000 (22:04 +0200)]
PCI: Add pci_fixup_suspend_late quirk pass

Add pci_fixup_suspend_late as a new pci_fixup_pass. The pass is called
from suspend_noirq and poweroff_noirq. Using the same pass for suspend
and hibernate is consistent with resume_early which is called by
resume_noirq and restore_noirq.

The new quirk pass is required for Thunderbolt support on Apple
hardware.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Add support for simple pci tunnels
Andreas Noever [Tue, 3 Jun 2014 20:04:08 +0000 (22:04 +0200)]
thunderbolt: Add support for simple pci tunnels

A pci downstream and pci upstream port can be connected through a
tunnel. To establish the tunnel we have to setup two unidirectional
paths between the two ports.

Right now we only support paths with two hops (i.e. no chaining) and at
most one pci device per thunderbolt device.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Add path setup code.
Andreas Noever [Tue, 3 Jun 2014 20:04:07 +0000 (22:04 +0200)]
thunderbolt: Add path setup code.

A thunderbolt path is a unidirectional channel between two thunderbolt
ports. Two such paths are needed to establish a pci tunnel.

This patch introduces struct tb_path as well as a set of tb_path_*
methods which are used to activate & deactivate paths.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Handle hotplug events
Andreas Noever [Tue, 3 Jun 2014 20:04:06 +0000 (22:04 +0200)]
thunderbolt: Handle hotplug events

We receive a plug event callback whenever a thunderbolt device is added
or removed. This patch fills in the tb_handle_hotplug method and starts
reacting to these events by adding/removing switches from the hierarchy.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Scan for downstream switches
Andreas Noever [Tue, 3 Jun 2014 20:04:05 +0000 (22:04 +0200)]
thunderbolt: Scan for downstream switches

Add utility methods tb_port_state and tb_wait_for_port. Add
tb_scan_switch which recursively checks for downstream switches.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Enable plug events
Andreas Noever [Tue, 3 Jun 2014 20:04:04 +0000 (22:04 +0200)]
thunderbolt: Enable plug events

Thunderbolt switches have a plug events capability. This patch adds the
tb_plug_events_active method and uses it to activate plug events during
switch allocation.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Add thunderbolt capability handling
Andreas Noever [Tue, 3 Jun 2014 20:04:03 +0000 (22:04 +0200)]
thunderbolt: Add thunderbolt capability handling

Thunderbolt config areas contain capability lists similar to those found
on pci devices. This patch introduces a tb_find_cap utility method to
search for capabilities.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Initialize root switch and ports
Andreas Noever [Tue, 3 Jun 2014 20:04:02 +0000 (22:04 +0200)]
thunderbolt: Initialize root switch and ports

This patch adds the structures tb_switch and tb_port as well as code to
initialize the root switch.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Add tb_regs.h
Andreas Noever [Tue, 3 Jun 2014 20:04:01 +0000 (22:04 +0200)]
thunderbolt: Add tb_regs.h

Every thunderbolt device consists (logically) of a switch with multiple
ports. Every port contains up to four config regions (HOPS, PORT,
SWITCH, COUNTERS) which are used to configure the device.

The tb_regs.h file contains all known registers and capabilities from
these config regions.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Setup control channel
Andreas Noever [Tue, 3 Jun 2014 20:04:00 +0000 (22:04 +0200)]
thunderbolt: Setup control channel

Add struct tb which will contain our view of the thunderbolt bus.  For
now it just contains a pointer to the control channel and a workqueue
for hotplug events.

Add thunderbolt_alloc_and_start() and thunderbolt_shutdown_and_free()
which are responsible for setup and teardown of struct tb.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Add control channel interface
Andreas Noever [Tue, 3 Jun 2014 20:03:59 +0000 (22:03 +0200)]
thunderbolt: Add control channel interface

Thunderbolt devices are configured by reading/writing into their
configuration space (similar to pci). This is done by sending packets
through the NHI (native host interface) onto the control channel.

This patch handles the low level packet based protocol and exposes
higher level operations like tb_cfg_read/tb_cfg_write.

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agothunderbolt: Add initial cactus ridge NHI support
Andreas Noever [Tue, 3 Jun 2014 20:03:58 +0000 (22:03 +0200)]
thunderbolt: Add initial cactus ridge NHI support

Thunderbolt hotplug is supposed to be handled by the firmware. But Apple
decided to implement thunderbolt at the operating system level. The
firmare only initializes thunderbolt devices that are present at boot
time. This driver enables hotplug of thunderbolt of non-chained
thunderbolt devices on Apple systems with a cactus ridge controller.

This first patch adds the Kconfig file as well the parts of the driver
which talk directly to the hardware (that is pci device setup, interrupt
handling and RX/TX ring management).

Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
10 years agoLinux 3.16-rc1
Linus Torvalds [Mon, 16 Jun 2014 03:45:28 +0000 (17:45 -1000)]
Linux 3.16-rc1

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Mon, 16 Jun 2014 02:37:03 +0000 (16:37 -1000)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Fix checksumming regressions, from Tom Herbert.

 2) Undo unintentional permissions changes for SCTP rto_alpha and
    rto_beta sysfs knobs, from Denial Borkmann.

 3) VXLAN, like other IP tunnels, should advertize it's encapsulation
    size using dev->needed_headroom instead of dev->hard_header_len.
    From Cong Wang.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
  net: sctp: fix permissions for rto_alpha and rto_beta knobs
  vxlan: Checksum fixes
  net: add skb_pop_rcv_encapsulation
  udp: call __skb_checksum_complete when doing full checksum
  net: Fix save software checksum complete
  net: Fix GSO constants to match NETIF flags
  udp: ipv4: do not waste time in __udp4_lib_mcast_demux_lookup
  vxlan: use dev->needed_headroom instead of dev->hard_header_len
  MAINTAINERS: update cxgb4 maintainer

10 years agoMerge tag 'clk-for-linus-3.16-part2' of git://git.linaro.org/people/mike.turquette...
Linus Torvalds [Mon, 16 Jun 2014 02:02:20 +0000 (16:02 -1000)]
Merge tag 'clk-for-linus-3.16-part2' of git://git.linaro.org/people/mike.turquette/linux

Pull more clock framework updates from Mike Turquette:
 "This contains the second half the of the clk changes for 3.16.

  They are simply fixes and code refactoring for the OMAP clock drivers.
  The sunxi clock driver changes include splitting out the one
  mega-driver into several smaller pieces and adding support for the A31
  SoC clocks"

* tag 'clk-for-linus-3.16-part2' of git://git.linaro.org/people/mike.turquette/linux: (25 commits)
  clk: sunxi: document PRCM clock compatible strings
  clk: sunxi: add PRCM (Power/Reset/Clock Management) clks support
  clk: sun6i: Protect SDRAM gating bit
  clk: sun6i: Protect CPU clock
  clk: sunxi: Rework clock protection code
  clk: sunxi: Move the GMAC clock to a file of its own
  clk: sunxi: Move the 24M oscillator to a file of its own
  clk: sunxi: Remove calls to clk_put
  clk: sunxi: document new A31 USB clock compatible
  clk: sunxi: Implement A31 USB clock
  ARM: dts: OMAP5/DRA7: use omap5-mpu-dpll-clock capable of dealing with higher frequencies
  CLK: TI: dpll: support OMAP5 MPU DPLL that need special handling for higher frequencies
  ARM: OMAP5+: dpll: support Duty Cycle Correction(DCC)
  CLK: TI: clk-54xx: Set the rate for dpll_abe_m2x2_ck
  CLK: TI: Driver for DRA7 ATL (Audio Tracking Logic)
  dt:/bindings: DRA7 ATL (Audio Tracking Logic) clock bindings
  ARM: dts: dra7xx-clocks: Correct name for atl clkin3 clock
  CLK: TI: gate: add composite interface clock to OMAP2 only build
  ARM: OMAP2: clock: add DT boot support for cpufreq_ck
  CLK: TI: OMAP2: add clock init support
  ...

10 years agoMerge git://git.infradead.org/users/willy/linux-nvme
Linus Torvalds [Mon, 16 Jun 2014 01:58:03 +0000 (15:58 -1000)]
Merge git://git.infradead.org/users/willy/linux-nvme

Pull NVMe update from Matthew Wilcox:
 "Mostly bugfixes again for the NVMe driver.  I'd like to call out the
  exported tracepoint in the block layer; I believe Keith has cleared
  this with Jens.

  We've had a few reports from people who're really pounding on NVMe
  devices at scale, hence the timeout changes (and new module
  parameters), hotplug cpu deadlock, tracepoints, and minor performance
  tweaks"

[ Jens hadn't seen that tracepoint thing, but is ok with it - it will
  end up going away when mq conversion happens ]

* git://git.infradead.org/users/willy/linux-nvme: (22 commits)
  NVMe: Fix START_STOP_UNIT Scsi->NVMe translation.
  NVMe: Use Log Page constants in SCSI emulation
  NVMe: Define Log Page constants
  NVMe: Fix hot cpu notification dead lock
  NVMe: Rename io_timeout to nvme_io_timeout
  NVMe: Use last bytes of f/w rev SCSI Inquiry
  NVMe: Adhere to request queue block accounting enable/disable
  NVMe: Fix nvme get/put queue semantics
  NVMe: Delete NVME_GET_FEAT_TEMP_THRESH
  NVMe: Make admin timeout a module parameter
  NVMe: Make iod bio timeout a parameter
  NVMe: Prevent possible NULL pointer dereference
  NVMe: Fix the buffer size passed in GetLogPage(CDW10.NUMD)
  NVMe: Update data structures for NVMe 1.2
  NVMe: Enable BUILD_BUG_ON checks
  NVMe: Update namespace and controller identify structures to the 1.1a spec
  NVMe: Flush with data support
  NVMe: Configure support for block flush
  NVMe: Add tracepoints
  NVMe: Protect against badly formatted CQEs
  ...

10 years agonet: sctp: fix permissions for rto_alpha and rto_beta knobs
Daniel Borkmann [Sat, 14 Jun 2014 22:59:14 +0000 (00:59 +0200)]
net: sctp: fix permissions for rto_alpha and rto_beta knobs

Commit 3fd091e73b81 ("[SCTP]: Remove multiple levels of msecs
to jiffies conversions.") has silently changed permissions for
rto_alpha and rto_beta knobs from 0644 to 0444. The purpose of
this was to discourage users from tweaking rto_alpha and
rto_beta knobs in production environments since they are key
to correctly compute rtt/srtt.

RFC4960 under section 6.3.1. RTO Calculation says regarding
rto_alpha and rto_beta under rule C3 and C4:

  [...]
  C3)  When a new RTT measurement R' is made, set

       RTTVAR <- (1 - RTO.Beta) * RTTVAR + RTO.Beta * |SRTT - R'|

       and

       SRTT <- (1 - RTO.Alpha) * SRTT + RTO.Alpha * R'

       Note: The value of SRTT used in the update to RTTVAR
       is its value before updating SRTT itself using the
       second assignment. After the computation, update
       RTO <- SRTT + 4 * RTTVAR.

  C4)  When data is in flight and when allowed by rule C5
       below, a new RTT measurement MUST be made each round
       trip. Furthermore, new RTT measurements SHOULD be
       made no more than once per round trip for a given
       destination transport address. There are two reasons
       for this recommendation: First, it appears that
       measuring more frequently often does not in practice
       yield any significant benefit [ALLMAN99]; second,
       if measurements are made more often, then the values
       of RTO.Alpha and RTO.Beta in rule C3 above should be
       adjusted so that SRTT and RTTVAR still adjust to
       changes at roughly the same rate (in terms of how many
       round trips it takes them to reflect new values) as
       they would if making only one measurement per
       round-trip and using RTO.Alpha and RTO.Beta as given
       in rule C3. However, the exact nature of these
       adjustments remains a research issue.
  [...]

While it is discouraged to adjust rto_alpha and rto_beta
and not further specified how to adjust them, the RFC also
doesn't explicitly forbid it, but rather gives a RECOMMENDED
default value (rto_alpha=3, rto_beta=2). We have a couple
of users relying on the old permissions before they got
changed. That said, if someone really has the urge to adjust
them, we could allow it with a warning in the log.

Fixes: 3fd091e73b81 ("[SCTP]: Remove multiple levels of msecs to jiffies conversions.")
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'csum_fixes'
David S. Miller [Sun, 15 Jun 2014 08:00:56 +0000 (01:00 -0700)]
Merge branch 'csum_fixes'

Tom Herbert says:

====================
Fixes related to some recent checksum modifications.

- Fix GSO constants to match NETIF flags
- Fix logic in saving checksum complete in __skb_checksum_complete
- Call __skb_checksum_complete from UDP if we are checksumming over
  whole packet in order to save checksum.
- Fixes to VXLAN to work correctly with checksum complete
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agovxlan: Checksum fixes
Tom Herbert [Sun, 15 Jun 2014 06:24:36 +0000 (23:24 -0700)]
vxlan: Checksum fixes

Call skb_pop_rcv_encapsulation and postpull_rcsum for the Ethernet
header to work properly with checksum complete.

Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: add skb_pop_rcv_encapsulation
Tom Herbert [Sun, 15 Jun 2014 06:24:28 +0000 (23:24 -0700)]
net: add skb_pop_rcv_encapsulation

This function is used by UDP encapsulation protocols in RX when
crossing encapsulation boundary. If ip_summed is set to
CHECKSUM_UNNECESSARY and encapsulation is not set, change to
CHECKSUM_NONE since the checksum has not been validated within the
encapsulation. Clears csum_valid by the same rationale.

Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoudp: call __skb_checksum_complete when doing full checksum
Tom Herbert [Sun, 15 Jun 2014 06:24:20 +0000 (23:24 -0700)]
udp: call __skb_checksum_complete when doing full checksum

In __udp_lib_checksum_complete check if checksum is being done over all
the data (len is equal to skb->len) and if it is call
__skb_checksum_complete instead of __skb_checksum_complete_head. This
allows checksum to be saved in checksum complete.

Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: Fix save software checksum complete
Tom Herbert [Sun, 15 Jun 2014 06:24:03 +0000 (23:24 -0700)]
net: Fix save software checksum complete

Geert reported issues regarding checksum complete and UDP.
The logic introduced in commit 7e3cead5172927732f51fde
("net: Save software checksum complete") is not correct.

This patch:
1) Restores code in __skb_checksum_complete_header except for setting
   CHECKSUM_UNNECESSARY. This function may be calculating checksum on
   something less than skb->len.
2) Adds saving checksum to __skb_checksum_complete. The full packet
   checksum 0..skb->len is calculated without adding in pseudo header.
   This value is saved in skb->csum and then the pseudo header is added
   to that to derive the checksum for validation.
3) In both __skb_checksum_complete_header and __skb_checksum_complete,
   set skb->csum_valid to whether checksum of zero was computed. This
   allows skb_csum_unnecessary to return true without changing to
   CHECKSUM_UNNECESSARY which was done previously.
4) Copy new csum related bits in __copy_skb_header.

Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: Fix GSO constants to match NETIF flags
Tom Herbert [Sun, 15 Jun 2014 06:23:52 +0000 (23:23 -0700)]
net: Fix GSO constants to match NETIF flags

Joseph Gasparakis reported that VXLAN GSO offload stopped working with
i40e device after recent UDP changes. The problem is that the
SKB_GSO_* bits are out of sync with the corresponding NETIF flags. This
patch fixes that. Also, we add BUILD_BUG_ONs in net_gso_ok for several
GSO constants that were missing to avoid the problem in the future.

Reported-by: Joseph Gasparakis <joseph.gasparakis@intel.com>
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge tag 'scsi-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb...
Linus Torvalds [Sun, 15 Jun 2014 00:49:48 +0000 (19:49 -0500)]
Merge tag 'scsi-for-linus' of git://git./linux/kernel/git/jejb/scsi

Pull more SCSI updates from James Bottomley:
 "This is just a couple of drivers (hpsa and lpfc) that got left out for
  further testing in linux-next.  We also have one fix to a prior
  submission (qla2xxx sparse)"

* tag 'scsi-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (36 commits)
  qla2xxx: fix sparse warnings introduced by previous target mode t10-dif patch
  lpfc: Update lpfc version to driver version 10.2.8001.0
  lpfc: Fix ExpressLane priority setup
  lpfc: mark old devices as obsolete
  lpfc: Fix for initializing RRQ bitmap
  lpfc: Fix for cleaning up stale ring flag and sp_queue_event entries
  lpfc: Update lpfc version to driver version 10.2.8000.0
  lpfc: Update Copyright on changed files from 8.3.45 patches
  lpfc: Update Copyright on changed files
  lpfc: Fixed locking for scsi task management commands
  lpfc: Convert runtime references to old xlane cfg param to fof cfg param
  lpfc: Fix FW dump using sysfs
  lpfc: Fix SLI4 s abort loop to process all FCP rings and under ring_lock
  lpfc: Fixed kernel panic in lpfc_abort_handler
  lpfc: Fix locking for postbufq when freeing
  lpfc: Fix locking for lpfc_hba_down_post
  lpfc: Fix dynamic transitions of FirstBurst from on to off
  hpsa: fix handling of hpsa_volume_offline return value
  hpsa: return -ENOMEM not -1 on kzalloc failure in hpsa_get_device_id
  hpsa: remove messages about volume status VPD inquiry page not supported
  ...

10 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux...
Linus Torvalds [Sun, 15 Jun 2014 00:48:43 +0000 (19:48 -0500)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mason/linux-btrfs

Pull more btrfs updates from Chris Mason:
 "This has a few fixes since our last pull and a new ioctl for doing
  btree searches from userland.  It's very similar to the existing
  ioctl, but lets us return larger items back down to the app"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  btrfs: fix error handling in create_pending_snapshot
  btrfs: fix use of uninit "ret" in end_extent_writepage()
  btrfs: free ulist in qgroup_shared_accounting() error path
  Btrfs: fix qgroups sanity test crash or hang
  btrfs: prevent RCU warning when dereferencing radix tree slot
  Btrfs: fix unfinished readahead thread for raid5/6 degraded mounting
  btrfs: new ioctl TREE_SEARCH_V2
  btrfs: tree_search, search_ioctl: direct copy to userspace
  btrfs: new function read_extent_buffer_to_user
  btrfs: tree_search, copy_to_sk: return needed size on EOVERFLOW
  btrfs: tree_search, copy_to_sk: return EOVERFLOW for too small buffer
  btrfs: tree_search, search_ioctl: accept varying buffer
  btrfs: tree_search: eliminate redundant nr_items check

10 years agoMerge git://git.kvack.org/~bcrl/aio-next
Linus Torvalds [Sun, 15 Jun 2014 00:43:27 +0000 (19:43 -0500)]
Merge git://git.kvack.org/~bcrl/aio-next

Pull aio fix and cleanups from Ben LaHaise:
 "This consists of a couple of code cleanups plus a minor bug fix"

* git://git.kvack.org/~bcrl/aio-next:
  aio: cleanup: flatten kill_ioctx()
  aio: report error from io_destroy() when threads race in io_destroy()
  fs/aio.c: Remove ctx parameter in kiocb_cancel

10 years agofix __swap_writepage() compile failure on old gcc versions
Al Viro [Sat, 14 Jun 2014 06:12:41 +0000 (07:12 +0100)]
fix __swap_writepage() compile failure on old gcc versions

Tetsuo Handa wrote:
 "Commit 62a8067a7f35 ("bio_vec-backed iov_iter") introduced an unnamed
  union inside a struct which gcc-4.4.7 cannot handle.  Name the unnamed
   union as u in order to fix build failure"

Let's do this instead: there is only one place in the entire tree that
steps into this breakage.  Anon structs and unions work in older gcc
versions; as the matter of fact, we have those in the tree - see e.g.
struct ieee80211_tx_info in include/net/mac80211.h

What doesn't work is handling their initializers:

struct {
int a;
union {
int b;
char c;
};
} x[2] = {{.a = 1, .c = 'a'}, {.a = 0, .b = 1}};

is the obvious syntax for initializer, perfectly fine for C11 and
handled correctly by gcc-4.7 or later.

Earlier versions, though, break on it - declaration is fine and so's
access to fields (i.e.  x[0].c = 'a'; would produce the right code), but
members of the anon structs and unions are not inserted into the right
namespace.  Tellingly, those older versions will not barf on struct {int
a; struct {int a;};}; - looks like they just have it hacked up somewhere
around the handling of .  and -> instead of doing the right thing.

The easiest way to deal with that crap is to turn initialization of
those fields (in the only place where we have such initializer of
iov_iter) into plain assignment.

Reported-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Reported-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoMerge tag 'hsi-for-3.16-fixes1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 14 Jun 2014 21:51:25 +0000 (14:51 -0700)]
Merge tag 'hsi-for-3.16-fixes1' of git://git./linux/kernel/git/sre/linux-hsi

Pull HSI build fixes from Sebastian Reichel:
 - tighten dependency between ssi-protocol and omap-ssi to fix build
   failures with randconfig.
 - use normal module refcounting in omap driver to fix build with
   disabled module support

* tag 'hsi-for-3.16-fixes1' of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-hsi:
  hsi: omap_ssi_port: use normal module refcounting
  HSI: fix omap ssi driver dependency

10 years agoMerge tag 'gpio-v3.16-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw...
Linus Torvalds [Sat, 14 Jun 2014 21:49:51 +0000 (14:49 -0700)]
Merge tag 'gpio-v3.16-2' of git://git./linux/kernel/git/linusw/linux-gpio

Pull GPIO fix from Linus Walleij:
 "A first GPIO fix for the v3.16 series, this was serious since it
  blocks the OMAP boot.

  Sending you this vital fix before leaving for a short vacation so it
  does not sit collecting dust in my tree for no good reason.

  Apart from this, our v3.16 cycle looks like a good start"

* tag 'gpio-v3.16-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:
  gpio: of: Fix handling for deferred probe for -gpio suffix

10 years agoMerge branch 'x86-vdso-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 14 Jun 2014 21:46:29 +0000 (14:46 -0700)]
Merge branch 'x86-vdso-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 vdso fixes from Peter Anvin:
 "Fixes for x86/vdso.

  One is a simple build fix for bigendian hosts, one is to make "make
  vdso_install" work again, and the rest is about working around a bug
  in Google's Go language -- two are documentation patches that improves
  the sample code that the Go coders took, modified, and broke; the
  other two implements a workaround that keeps existing Go binaries from
  segfaulting at least"

* 'x86-vdso-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/vdso: Fix vdso_install
  x86/vdso: Hack to keep 64-bit Go programs working
  x86/vdso: Add PUT_LE to store little-endian values
  x86/vdso/doc: Make vDSO examples more portable
  x86/vdso/doc: Rename vdso_test.c to vdso_standalone_test_x86.c
  x86, vdso: Remove one final use of htole16()

10 years agoMerge tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck...
Linus Torvalds [Sat, 14 Jun 2014 21:43:23 +0000 (14:43 -0700)]
Merge tag 'hwmon-for-linus' of git://git./linux/kernel/git/groeck/linux-staging

Pull hwmon updates from Guenter Roeck:
 - new driver for Sensirion SHTC1 humidity / temperature sensor
 - convert ltc4151 and vexpress drivers to use devm functions
 - drop generic chip detection from lm85 driver
 - avoid forward declarations in atxp1 driver
 - fix sign extensions in ina2xx driver

* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: vexpress: Use devm helper for hwmon device registration
  hwmon: (atxp1) Avoid forward declaration
  hwmon: add support for Sensirion SHTC1 sensor
  hwmon: (ltc4151) Convert to devm_hwmon_device_register_with_groups
  hwmon: (lm85) Drop generic detection
  hwmon: (ina2xx) Cast to s16 on shunt and current regs

10 years agoudp: ipv4: do not waste time in __udp4_lib_mcast_demux_lookup
Eric Dumazet [Thu, 12 Jun 2014 23:13:06 +0000 (16:13 -0700)]
udp: ipv4: do not waste time in __udp4_lib_mcast_demux_lookup

Its too easy to add thousand of UDP sockets on a particular bucket,
and slow down an innocent multicast receiver.

Early demux is supposed to be an optimization, we should avoid spending
too much time in it.

It is interesting to note __udp4_lib_demux_lookup() only tries to
match first socket in the chain.

10 is the threshold we already have in __udp4_lib_lookup() to switch
to secondary hash.

Fixes: 421b3885bf6d5 ("udp: ipv4: Add udp early demux")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: David Held <drheld@google.com>
Cc: Shawn Bohrer <sbohrer@rgmadvisors.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agovxlan: use dev->needed_headroom instead of dev->hard_header_len
Cong Wang [Thu, 12 Jun 2014 18:53:10 +0000 (11:53 -0700)]
vxlan: use dev->needed_headroom instead of dev->hard_header_len

When we mirror packets from a vxlan tunnel to other device,
the mirror device should see the same packets (that is, without
outer header). Because vxlan tunnel sets dev->hard_header_len,
tcf_mirred() resets mac header back to outer mac, the mirror device
actually sees packets with outer headers

Vxlan tunnel should set dev->needed_headroom instead of
dev->hard_header_len, like what other ip tunnels do. This fixes
the above problem.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: stephen hemminger <stephen@networkplumber.org>
Cc: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: Cong Wang <cwang@twopensource.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMAINTAINERS: update cxgb4 maintainer
Dimitris Michailidis [Fri, 13 Jun 2014 21:11:14 +0000 (14:11 -0700)]
MAINTAINERS: update cxgb4 maintainer

Hari's been doing the patch submissions for a while now and he'll be
taking over as maintainer.

Signed-off-by: Dimitris Michailidis <dm@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agox86/vdso: Fix vdso_install
Andy Lutomirski [Thu, 12 Jun 2014 15:28:10 +0000 (08:28 -0700)]
x86/vdso: Fix vdso_install

"make vdso_install" installs unstripped versions of the vdso objects
for the benefit of the debugger.  This was broken by checkin:

6f121e548f83 x86, vdso: Reimplement vdso.so preparation in build-time C

The filenames are different now, so update the Makefile to cope.

This still installs the 64-bit vdso as vdso64.so.  We believe this
will be okay, as the only known user is a patched gdb which is known
to use build-ids, but if it turns out to be a problem we may have to
add a link.

Inspired by a patch from Sam Ravnborg.

Acked-by: Sam Ravnborg <sam@ravnborg.org>
Reported-by: Josh Boyer <jwboyer@fedoraproject.org>
Tested-by: Josh Boyer <jwboyer@fedoraproject.org>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Link: http://lkml.kernel.org/r/b10299edd8ba98d17e07dafcd895b8ecf4d99eff.1402586707.git.luto@amacapital.net
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
10 years agoNVMe: Fix START_STOP_UNIT Scsi->NVMe translation.
Dan McLeran [Fri, 6 Jun 2014 14:27:27 +0000 (08:27 -0600)]
NVMe: Fix START_STOP_UNIT Scsi->NVMe translation.

This patch contains several fixes for Scsi START_STOP_UNIT. The previous
code did not account for signed vs. unsigned arithmetic which resulted
in an invalid lowest power state caculation when the device only supports
1 power state.

The code for Power Condition == 2 (Idle) was not following the spec. The
spec calls for setting the device to specific power states, depending
upon Power Condition Modifier, without accounting for the number of
power states supported by the device.

The code for Power Condition == 3 (Standby) was using a hard-coded '0'
which is replaced with the macro POWER_STATE_0.

Signed-off-by: Dan McLeran <daniel.mcleran@intel.com>
Reviewed-by: Vishal Verma <vishal.l.verma@linux.intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
10 years agobtrfs: fix error handling in create_pending_snapshot
Eric Sandeen [Thu, 12 Jun 2014 05:53:44 +0000 (00:53 -0500)]
btrfs: fix error handling in create_pending_snapshot

fcebe456 cut and pasted some code to a later point
in create_pending_snapshot(), but didn't switch
to the appropriate error handling for this stage
of the function.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Chris Mason <clm@fb.com>
10 years agobtrfs: fix use of uninit "ret" in end_extent_writepage()
Eric Sandeen [Thu, 12 Jun 2014 05:39:58 +0000 (00:39 -0500)]
btrfs: fix use of uninit "ret" in end_extent_writepage()

If this condition in end_extent_writepage() is false:

if (tree->ops && tree->ops->writepage_end_io_hook)

we will then test an uninitialized "ret" at:

ret = ret < 0 ? ret : -EIO;

The test for ret is for the case where ->writepage_end_io_hook
failed, and we'd choose that ret as the error; but if
there is no ->writepage_end_io_hook, nothing sets ret.

Initializing ret to 0 should be sufficient; if
writepage_end_io_hook wasn't set, (!uptodate) means
non-zero err was passed in, so we choose -EIO in that case.

Signed-of-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Chris Mason <clm@fb.com>
10 years agobtrfs: free ulist in qgroup_shared_accounting() error path
Eric Sandeen [Thu, 12 Jun 2014 05:14:59 +0000 (00:14 -0500)]
btrfs: free ulist in qgroup_shared_accounting() error path

If tmp = ulist_alloc(GFP_NOFS) fails, we return without
freeing the previously allocated qgroups = ulist_alloc(GFP_NOFS)
and cause a memory leak.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: Chris Mason <clm@fb.com>
10 years agoBtrfs: fix qgroups sanity test crash or hang
Filipe Manana [Thu, 12 Jun 2014 01:47:37 +0000 (02:47 +0100)]
Btrfs: fix qgroups sanity test crash or hang

Often when running the qgroups sanity test, a crash or a hang happened.
This is because the extent buffer the test uses for the root node doesn't
have an header level explicitly set, making it have a random level value.
This is a problem when it's not zero for the btrfs_search_slot() calls
the test ends up doing, resulting in crashes or hangs such as the following:

[ 6454.127192] Btrfs loaded, debug=on, assert=on, integrity-checker=on
(...)
[ 6454.127760] BTRFS: selftest: Running qgroup tests
[ 6454.127964] BTRFS: selftest: Running test_test_no_shared_qgroup
[ 6454.127966] BTRFS: selftest: Qgroup basic add
[ 6480.152005] BUG: soft lockup - CPU#0 stuck for 23s! [modprobe:5383]
[ 6480.152005] Modules linked in: btrfs(+) xor raid6_pq binfmt_misc nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc i2c_piix4 i2c_core pcspkr evbug psmouse serio_raw e1000 [last unloaded: btrfs]
[ 6480.152005] irq event stamp: 188448
[ 6480.152005] hardirqs last  enabled at (188447): [<ffffffff8168ef5c>] restore_args+0x0/0x30
[ 6480.152005] hardirqs last disabled at (188448): [<ffffffff81698e6a>] apic_timer_interrupt+0x6a/0x80
[ 6480.152005] softirqs last  enabled at (188446): [<ffffffff810516cf>] __do_softirq+0x1cf/0x450
[ 6480.152005] softirqs last disabled at (188441): [<ffffffff81051c25>] irq_exit+0xb5/0xc0
[ 6480.152005] CPU: 0 PID: 5383 Comm: modprobe Not tainted 3.15.0-rc8-fdm-btrfs-next-33+ #4
[ 6480.152005] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
[ 6480.152005] task: ffff8802146125a0 ti: ffff8800d0d00000 task.ti: ffff8800d0d00000
[ 6480.152005] RIP: 0010:[<ffffffff81349a63>]  [<ffffffff81349a63>] __write_lock_failed+0x13/0x20
[ 6480.152005] RSP: 0018:ffff8800d0d038e8  EFLAGS: 00000287
[ 6480.152005] RAX: 0000000000000000 RBX: ffffffff8168ef5c RCX: 000005deb8525852
[ 6480.152005] RDX: 0000000000000000 RSI: 0000000000001d45 RDI: ffff8802105000b8
[ 6480.152005] RBP: ffff8800d0d038e8 R08: fffffe12710f63db R09: ffffffffa03196fb
[ 6480.152005] R10: ffff8802146125a0 R11: ffff880214612e28 R12: ffff8800d0d03858
[ 6480.152005] R13: 0000000000000000 R14: ffff8800d0d00000 R15: ffff8802146125a0
[ 6480.152005] FS:  00007f14ff804700(0000) GS:ffff880215e00000(0000) knlGS:0000000000000000
[ 6480.152005] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[ 6480.152005] CR2: 00007fff4df0dac8 CR3: 00000000d1796000 CR4: 00000000000006f0
[ 6480.152005] Stack:
[ 6480.152005]  ffff8800d0d03908 ffffffff810ae967 0000000000000001 ffff8802105000b8
[ 6480.152005]  ffff8800d0d03938 ffffffff8168e57e ffffffffa0319c16 0000000000000007
[ 6480.152005]  ffff880210500000 ffff880210500100 ffff8800d0d039b8 ffffffffa0319c16
[ 6480.152005] Call Trace:
[ 6480.152005]  [<ffffffff810ae967>] do_raw_write_lock+0x47/0xa0
[ 6480.152005]  [<ffffffff8168e57e>] _raw_write_lock+0x5e/0x80
[ 6480.152005]  [<ffffffffa0319c16>] ? btrfs_tree_lock+0x116/0x270 [btrfs]
[ 6480.152005]  [<ffffffffa0319c16>] btrfs_tree_lock+0x116/0x270 [btrfs]
[ 6480.152005]  [<ffffffffa02b2acb>] btrfs_lock_root_node+0x3b/0x50 [btrfs]
[ 6480.152005]  [<ffffffffa02b81a6>] btrfs_search_slot+0x916/0xa20 [btrfs]
[ 6480.152005]  [<ffffffff811a727f>] ? create_object+0x23f/0x300
[ 6480.152005]  [<ffffffffa02b9958>] btrfs_insert_empty_items+0x78/0xd0 [btrfs]
[ 6480.152005]  [<ffffffffa036041a>] insert_normal_tree_ref.constprop.4+0xa2/0x19a [btrfs]
[ 6480.152005]  [<ffffffffa03605c3>] test_no_shared_qgroup+0xb1/0x1ca [btrfs]
[ 6480.152005]  [<ffffffff8108cad6>] ? local_clock+0x16/0x30
[ 6480.152005]  [<ffffffffa035ef8e>] btrfs_test_qgroups+0x1ae/0x1d7 [btrfs]
[ 6480.152005]  [<ffffffffa03a69d2>] ? ftrace_define_fields_btrfs_space_reservation+0xfd/0xfd [btrfs]
[ 6480.152005]  [<ffffffffa03a6a86>] init_btrfs_fs+0xb4/0x153 [btrfs]
[ 6480.152005]  [<ffffffff81000352>] do_one_initcall+0x102/0x150
[ 6480.152005]  [<ffffffff8103d223>] ? set_memory_nx+0x43/0x50
[ 6480.152005]  [<ffffffff81682668>] ? set_section_ro_nx+0x6d/0x74
[ 6480.152005]  [<ffffffff810d91cc>] load_module+0x1cdc/0x2630
(...)

Therefore initialize the extent buffer as an empty leaf (level 0).

Issue easy to reproduce when btrfs is built as a module via:

    $ for ((i = 1; i <= 1000000; i++)); do rmmod btrfs; modprobe btrfs; done

Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com>
Signed-off-by: Chris Mason <clm@fb.com>
10 years agobtrfs: prevent RCU warning when dereferencing radix tree slot
Sasha Levin [Wed, 11 Jun 2014 16:00:25 +0000 (12:00 -0400)]
btrfs: prevent RCU warning when dereferencing radix tree slot

Mark the dereference as protected by lock. Not doing so triggers
an RCU warning since the radix tree assumed that RCU is in use.

Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Chris Mason <clm@fb.com>
10 years agoBtrfs: fix unfinished readahead thread for raid5/6 degraded mounting
Wang Shilong [Wed, 11 Jun 2014 02:55:22 +0000 (10:55 +0800)]
Btrfs: fix unfinished readahead thread for raid5/6 degraded mounting

Steps to reproduce:

 # mkfs.btrfs -f /dev/sd[b-f] -m raid5 -d raid5
 # mkfs.ext4 /dev/sdc --->corrupt one of btrfs device
 # mount /dev/sdb /mnt -o degraded
 # btrfs scrub start -BRd /mnt

This is because readahead would skip missing device, this is not true
for RAID5/6, because REQ_GET_READ_MIRRORS return 1 for RAID5/6 block
mapping. If expected data locates in missing device, readahead thread
would not call __readahead_hook() which makes event @rc->elems=0
wait forever.

Fix this problem by checking return value of btrfs_map_block(),we
can only skip missing device safely if there are several mirrors.

Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com>
Signed-off-by: Chris Mason <clm@fb.com>