linux-drm-fsl-dcu.git
9 years agoDrivers: hv: vmbus: Properly protect calls to smp_processor_id()
K. Y. Srinivasan [Fri, 29 Aug 2014 01:29:53 +0000 (18:29 -0700)]
Drivers: hv: vmbus: Properly protect calls to smp_processor_id()

Disable preemption when sampling current processor ID when preemption
is otherwise possible.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Tested-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoDrivers: hv: vmbus: Cleanup hv_post_message()
K. Y. Srinivasan [Fri, 29 Aug 2014 01:29:52 +0000 (18:29 -0700)]
Drivers: hv: vmbus: Cleanup hv_post_message()

Minimize failures in this function by pre-allocating the buffer
for posting messages. The hypercall for posting the message can fail
for a number of reasons:

        1. Transient resource related issues
        2. Buffer alignment
        3. Buffer cannot span a page boundry

We address issues 2 and 3 by preallocating a per-cpu page for the buffer.
Transient resource related failures are handled by retrying by the callers
of this function.

This patch is based on the investigation
done by Dexuan Cui <decui@microsoft.com>.

I would like to thank Sitsofe Wheeler <sitsofe@yahoo.com>
for reporting the issue and helping in debuggging.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Reported-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Cc: <stable@vger.kernel.org>
Tested-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoDrivers: hv: vmbus: Cleanup vmbus_close_internal()
K. Y. Srinivasan [Wed, 27 Aug 2014 23:25:33 +0000 (16:25 -0700)]
Drivers: hv: vmbus: Cleanup vmbus_close_internal()

Eliminate calls to BUG_ON() in vmbus_close_internal().
We have chosen to potentially leak memory, than crash the guest
in case of failures.

In this version of the patch I have addressed comments from
Dan Carpenter (dan.carpenter@oracle.com).

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Cc: <stable@vger.kernel.org>
Tested-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoDrivers: hv: vmbus: Fix a bug in vmbus_open()
K. Y. Srinivasan [Wed, 27 Aug 2014 23:25:35 +0000 (16:25 -0700)]
Drivers: hv: vmbus: Fix a bug in vmbus_open()

Fix a bug in vmbus_open() and properly propagate the error. I would
like to thank Dexuan Cui <decui@microsoft.com> for identifying the
issue.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Cc: <stable@vger.kernel.org>
Tested-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoDrivers: hv: vmbus: Cleanup vmbus_establish_gpadl()
K. Y. Srinivasan [Wed, 27 Aug 2014 23:25:34 +0000 (16:25 -0700)]
Drivers: hv: vmbus: Cleanup vmbus_establish_gpadl()

Eliminate the call to BUG_ON() by waiting for the host to respond. We are
trying to reclaim the ownership of memory that was given to the host and so
we will have to wait until the host responds.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Cc: <stable@vger.kernel.org>
Tested-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoDrivers: hv: vmbus: Cleanup vmbus_teardown_gpadl()
K. Y. Srinivasan [Wed, 27 Aug 2014 23:25:32 +0000 (16:25 -0700)]
Drivers: hv: vmbus: Cleanup vmbus_teardown_gpadl()

Eliminate calls to BUG_ON() by properly handling errors. In cases where
rollback is possible, we will return the appropriate error to have the
calling code decide how to rollback state. In the case where we are
transferring ownership of the guest physical pages to the host,
we will wait for the host to respond.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Cc: <stable@vger.kernel.org>
Tested-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoDrivers: hv: vmbus: Cleanup vmbus_post_msg()
K. Y. Srinivasan [Wed, 27 Aug 2014 23:25:31 +0000 (16:25 -0700)]
Drivers: hv: vmbus: Cleanup vmbus_post_msg()

Posting messages to the host can fail because of transient resource
related failures. Correctly deal with these failures and increase the
number of attempts to post the message before giving up.

In this version of the patch, I have normalized the error code to
Linux error code.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Cc: <stable@vger.kernel.org>
Tested-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoi8k: Add support for Dell Latitude E6540
Steven Honeyman [Fri, 5 Sep 2014 17:03:42 +0000 (18:03 +0100)]
i8k: Add support for Dell Latitude E6540

Add support for the Dell Latitude E6540 which needs a different fan speed multiplier.

Signed-off-by: Steven Honeyman <stevenhoneyman@gmail.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoVMCI: integer overflow in vmci_datagram_dispatch()
Dan Carpenter [Fri, 29 Aug 2014 08:42:56 +0000 (11:42 +0300)]
VMCI: integer overflow in vmci_datagram_dispatch()

This is untrusted user data from vmci_host_do_send_datagram() so the
VMCI_DG_SIZE() macro can have an integer overflow.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoipoctal: clear break interrupt as soon as it occurs
Federico Vaga [Mon, 1 Sep 2014 11:49:56 +0000 (13:49 +0200)]
ipoctal: clear break interrupt as soon as it occurs

In some condition we receive the break interrupt but nothing is putted
in the Rx FIFO and the correspondend bit in the status register is not
set. Thus, no-one clear the interrupt and the handler will be called
forever.

This patch clear the break interrupt as soon as it occurs. Then, if the
break character '\0' is putted in the fifo we will manage it.

We can also unmask the Break interrupt but its bit in ISR is still set
on break. So I think is better to keep the registers clean.

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>
9 years agoGenWQE: Support blocking when DDCB queue is busy
Frank Haverkamp [Wed, 10 Sep 2014 14:37:53 +0000 (16:37 +0200)]
GenWQE: Support blocking when DDCB queue is busy

When the GenWQE hardware queue was busy, the driver returned simply
-EBUSY. This caused polling by applications which increased the load
on the already busy system. This change implements the possiblity to
sleep on a waitqueue instead when the DDCB queue is busy. The
requestor is woken up when there is free space on the queue again.
The old way to get -EBUSY is still available if the device is openend
with O_NONBLOCKING. The default is now blocking behavior.

Signed-off-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoGenWQE: Fix problem when reading HSI and Retc
Eberhard S. Amann [Wed, 10 Sep 2014 14:37:52 +0000 (16:37 +0200)]
GenWQE: Fix problem when reading HSI and Retc

This patch fixes a problem we found during debug on PPC64 when
reading HSI status and Retc.

Signed-off-by: Eberhard S. Amann <esa@linux.vnet.ibm.com>
Signed-off-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoGenWQE: Fix checkpatch complaints
Frank Haverkamp [Wed, 10 Sep 2014 14:37:51 +0000 (16:37 +0200)]
GenWQE: Fix checkpatch complaints

The checkpatch.pl script got improved. I ran it on the latest GenWQE
sources and fixed what it complained about.

Signed-off-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoGenWQE: Check return code of pci_sriov_enable
Frank Haverkamp [Wed, 10 Sep 2014 14:37:50 +0000 (16:37 +0200)]
GenWQE: Check return code of pci_sriov_enable

Forgetting to check this, can lead to problems on systems which
do not support SRIOV.

Signed-off-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoGenWQE: Do not modify return code of genwqe_set_interrupt_capability
Frank Haverkamp [Wed, 10 Sep 2014 14:37:49 +0000 (16:37 +0200)]
GenWQE: Do not modify return code of genwqe_set_interrupt_capability

Follow up patch to the one from Sebastian Ott. There is no need to
change the return code once it fails. And Sebastians version is tested
now and works nicely on our test-system.

Signed-off-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoGenWQE: Update author information
Frank Haverkamp [Wed, 10 Sep 2014 14:37:48 +0000 (16:37 +0200)]
GenWQE: Update author information

Updated email address of co-author.

Signed-off-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Michael Jung <mijung@gmx.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoGenWQE: Remove sysfs entry for driver version
Frank Haverkamp [Wed, 10 Sep 2014 14:37:47 +0000 (16:37 +0200)]
GenWQE: Remove sysfs entry for driver version

A special sysfs entry to display the driver version is not
needed. We left the driver version and adjusted it to the
naming a lot of other drivers use. The information can be
retrieved by using modinfo genwqe_card.

modinfo genwqe_card will provide the same information.

Signed-off-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoGenWQE: Check pci_get_totalvfs return code
Frank Haverkamp [Wed, 10 Sep 2014 14:37:46 +0000 (16:37 +0200)]
GenWQE: Check pci_get_totalvfs return code

Currently the driver is using the pci_get_totalvfs() return code
directly in a loop. To avoid problems with potentially negative
returns in case of errors, we are adding some more sanity checking
code.

Signed-off-by: Frank Haverkamp <haver@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomisc: st_kim: Increase size of dev_name buffer to incorporate termination
Lee Jones [Thu, 28 Aug 2014 13:14:09 +0000 (14:14 +0100)]
misc: st_kim: Increase size of dev_name buffer to incorporate termination

Calling strncpy with a maximum size argument of 32 bytes on destination
array kim_gdata->dev_name of size 32 bytes might leave the destination
string unterminated.

Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomisc: st_core: Protect unsigned value against becoming negative
Lee Jones [Thu, 28 Aug 2014 13:14:08 +0000 (14:14 +0100)]
misc: st_core: Protect unsigned value against becoming negative

Coverity reported:
  This less-than-zero comparison of an unsigned value is never true.

In answer to that, we only ever decrement if protos_registered is
positive.  We can subsequently remove the paranoid checking during
unregister.

Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoparport: parport_pc: Implement architecture and device check to cut off false-positives
Matwey V. Kornilov [Wed, 27 Aug 2014 08:07:44 +0000 (12:07 +0400)]
parport: parport_pc: Implement architecture and device check to cut off false-positives

We definitely know that only x86 (32-bit) architecture is affected by the issue, so implement a stub instead of the actual check for other architectures.

We also know that motherboard LPT chipset is affected, so the port is either come from
  parport_pc_init (when `io' module param is used) or
  parport_pc_find_isa_ports (when default LPT ports are probbed: 0x378, 0x278, 0x3bc).
In both cases the port considered as 'legacy' and `dev' member of struct parport is NULL. See also comments for `struct parport' in parport.h

Signed-off-by: Matwey V. Kornilov <matwey@sai.msu.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoparport: parport_pc: Introduce intel_bug_present function.
Matwey V. Kornilov [Wed, 27 Aug 2014 08:07:43 +0000 (12:07 +0400)]
parport: parport_pc: Introduce intel_bug_present function.

Put the code to check present of the Intel bug from parport_EPP_supported
into new intel_bug_present function. The later also return ECR register
to the state it has before function call.

Signed-off-by: Matwey V. Kornilov <matwey@sai.msu.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoipoctal: get carrier driver to avoid rmmod
Federico Vaga [Tue, 2 Sep 2014 15:31:41 +0000 (17:31 +0200)]
ipoctal: get carrier driver to avoid rmmod

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>
9 years agoipack: save carrier owner to allow device to get it
Federico Vaga [Tue, 2 Sep 2014 15:31:40 +0000 (17:31 +0200)]
ipack: save carrier owner to allow device to get it

There was not any kind of protection against carrier driver removal.
In this way, device driver can 'get' the carrier driver when it is
using it.

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>
9 years agoipoctal: reset function istead of duplicate code
Federico Vaga [Tue, 2 Sep 2014 15:31:39 +0000 (17:31 +0200)]
ipoctal: reset function istead of duplicate code

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>
9 years agodrivers/ipack/devices/ipoctal.h: Fix typo in include guard
Rasmus Villemoes [Fri, 22 Aug 2014 12:09:23 +0000 (14:09 +0200)]
drivers/ipack/devices/ipoctal.h: Fix typo in include guard

Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Acked-by: Samuel Iglesias Gonsalvez <siglesias@igalia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agodocumentation: devicetree: Added xillybus to vendor-prefixes
Eli Billauer [Sun, 17 Aug 2014 15:01:03 +0000 (18:01 +0300)]
documentation: devicetree: Added xillybus to vendor-prefixes

Signed-off-by: Eli Billauer <eli.billauer@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agochar: xilinx_hwicap: Remove .owner field for driver
Michal Simek [Wed, 13 Aug 2014 12:02:05 +0000 (14:02 +0200)]
char: xilinx_hwicap: Remove .owner field for driver

There is no need to init .owner field.

Based on the patch from Peter Griffin <peter.griffin@linaro.org>
"mmc: remove .owner field for drivers using module_platform_driver"

This patch removes the superflous .owner field for drivers which
use the module_platform_driver API, as this is overriden in
platform_driver_register anyway."

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agopcie-gadget-spear: use devm_ functions
Himangi Saraogi [Thu, 7 Aug 2014 12:13:28 +0000 (17:43 +0530)]
pcie-gadget-spear: use devm_ functions

The various devm_ functions allocate memory that is released when a
driver detaches. This patch uses these functions for data that is
allocated in the probe function of a platform device and is only freed
in the remove function. Also, the unnecessary labels are removed and
linux/device.h is added to make sure the devm_*() routine declarations
are unambiguously available.

The initial call to platform_get_resource is moved down to the
introduced call to devm_ioremap_resource that uses its result.

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomisc: lattice-ecp3-config: Add missing MODULE_FIRMWARE()
Takashi Iwai [Wed, 6 Aug 2014 15:37:41 +0000 (17:37 +0200)]
misc: lattice-ecp3-config: Add missing MODULE_FIRMWARE()

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Tested-by: Jean-Michel Hautbois <jean-michel.hautbois@vodalys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agoEEPROM: Introduce the use of devm_kzalloc
Himangi Saraogi [Fri, 1 Aug 2014 19:04:23 +0000 (00:34 +0530)]
EEPROM: Introduce the use of devm_kzalloc

This patch introduces the use of devm_kzalloc and does away with the
kfrees in the probe and remove functions. Also, a label and the err
variable are removed. The header device.h is included to make the devm_
function explicitly available and slab.h is done away with as it is no
longer needed.

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Reviewed-by: Jean Delvare <jdelvare@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: bus: fix possible boundaries violation
Alexander Usyskin [Mon, 25 Aug 2014 13:46:53 +0000 (16:46 +0300)]
mei: bus: fix possible boundaries violation

function 'strncpy' will fill whole buffer 'id.name' of fixed size (32)
with string value and will not leave place for NULL-terminator.
Possible buffer boundaries violation in following string operations.
Replace strncpy with strlcpy.

Cc: stable@vger.kernel.org # 3.10+
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>
9 years agomei: use connect_data on the stack
Tomas Winkler [Thu, 21 Aug 2014 11:29:23 +0000 (14:29 +0300)]
mei: use connect_data on the stack

There is no need for dynamic allocation for connect_data.
We can use variable on the stack and make code less
error prone and simple

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: enable adding more IOCTL handlers
Tomas Winkler [Thu, 21 Aug 2014 11:29:22 +0000 (14:29 +0300)]
mei: enable adding more IOCTL handlers

Handle ioctls in a switch statement so we can
add more commands easily

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: extract supported features from the hbm version
Tomas Winkler [Thu, 21 Aug 2014 11:29:21 +0000 (14:29 +0300)]
mei: extract supported features from the hbm version

extract supported hbm features and commands from the hbm version

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: simplify handling of hbm client events
Tomas Winkler [Thu, 21 Aug 2014 11:29:20 +0000 (14:29 +0300)]
mei: simplify handling of hbm client events

Add mei_hbm_cl_find_by_cmd handler to retrieve
the destination client

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: wait for hbm start non-interruptible
Alexander Usyskin [Thu, 21 Aug 2014 11:29:19 +0000 (14:29 +0300)]
mei: wait for hbm start non-interruptible

We cannot handle user interrupt in context of hbm start
so we only wait for time out which is reasonably short.

1. Add kdoc
2. Rename state to better reflect its function
3. Simplify wait condition and rename
   wait_recvd_msg to wait_hbm_start

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>
9 years agomei: revamp connect and disconnect response handling
Tomas Winkler [Thu, 21 Aug 2014 11:29:18 +0000 (14:29 +0300)]
mei: revamp connect and disconnect response handling

Both responses have same flow only the client status
update is different. We introduce handler mei_hbm_cl_res()
that handles both responses
Also we use per client wait queue  (cl->wait) rather then
global dev->wait_recvd_msg

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: use disconnect name consistently
Tomas Winkler [Thu, 21 Aug 2014 11:29:17 +0000 (14:29 +0300)]
mei: use disconnect name consistently

Rename mei_cl_irq_close to mei_cl_irq_disconnect
and MEI_FOP_CLOSE to MEI_FOP_DISCONNECT
Remove unused MEI_FOP_OPEN

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: add hbm commands return status values
Tomas Winkler [Thu, 21 Aug 2014 11:29:16 +0000 (14:29 +0300)]
mei: add hbm commands return status values

HBM uses global list of status values
from which the values of particular commands
are derived

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: add mei_me_cl_by_uuid_id function
Tomas Winkler [Thu, 21 Aug 2014 11:29:15 +0000 (14:29 +0300)]
mei: add mei_me_cl_by_uuid_id function

When handling dynamic clients there might be a race
scenario in which two me clients with the same me
address would be linked in the me clients list,
therefore we need to search by both uuid and me address.

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: add me client remove functions
Tomas Winkler [Thu, 21 Aug 2014 11:29:14 +0000 (14:29 +0300)]
mei: add me client remove functions

To support dynamic addition/remove we add wrappers
for removal of me clients

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: use list for me clients book keeping
Tomas Winkler [Thu, 21 Aug 2014 11:29:13 +0000 (14:29 +0300)]
mei: use list for me clients book keeping

To support dynamic addition/remove of clients
it is more convenient to use list instead of
static array

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: me_client lookup function to return me_client object
Tomas Winkler [Sun, 24 Aug 2014 09:08:55 +0000 (12:08 +0300)]
mei: me_client lookup function to return me_client object

For support of dynamic addition and removal of me clients
it is more convenient to use a list instead of static array
as is use now.
As the first step of the transition to the new data structure
we change the lookup function so it returns me client address
instead of an index.

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: use wrapper for simple hbm client message
Tomas Winkler [Thu, 21 Aug 2014 11:29:11 +0000 (14:29 +0300)]
mei: use wrapper for simple hbm client message

Reduce few code lines by using wrappers for sending
simple hbm client messages

Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
9 years agomei: use consistently me_addr in the hbm structures
Tomas Winkler [Thu, 21 Aug 2014 11:29:10 +0000 (14:29 +0300)]
mei: use consistently me_addr in the hbm structures

Use consistently me_addr name in hbm protocol structures
to represent in firmware client address

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>
9 years agomei: amthif: don't check amthif client identity on amthif queues
Alexander Usyskin [Thu, 14 Aug 2014 14:22:21 +0000 (17:22 +0300)]
mei: amthif: don't check amthif client identity on amthif queues

Amthif has its own queues therefore it is redundant
to check the client type when processing the queues

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>
9 years agomei: amthif: use service function to flush amthif queue
Alexander Usyskin [Thu, 14 Aug 2014 14:22:20 +0000 (17:22 +0300)]
mei: amthif: use service function to flush amthif queue

Replace open coded loop with an existing service function:
mei_io_list_flush

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>
9 years agomei: me: wait for hw ready non-interruptible
Alexander Usyskin [Tue, 12 Aug 2014 17:16:03 +0000 (20:16 +0300)]
mei: me: wait for hw ready non-interruptible

We cannot handle user interrupt in context of hw initialization
so we only wait for time out which is reasonably short
Also we don't need to check error from wait, only flag value.

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>
9 years agoMerge tag 'extcon-next-for-3.18' of git://git.kernel.org/pub/scm/linux/kernel/git...
Greg Kroah-Hartman [Tue, 23 Sep 2014 00:42:40 +0000 (17:42 -0700)]
Merge tag 'extcon-next-for-3.18' of git://git./linux/kernel/git/chanwoo/extcon into char-misc-next

Chanwoo writes:

Update extcon for v3.18

This patchset add new extcon provider driver and fix minor issue of extcon driver.

Detailed description for patchset:
1. Add new Richtek RT8973A extcon driver
This driver support for Richtek RT8973A which is Micro USB Switch OVP and
i2c interface. The RT8973A is a USB port accessory detector and switch that is
optimized to protect low voltage system from abnormal high input voltage
(up to 28V) and supports high speed USB operation. Also, RT8973A support
'auto-configuration' mode. If auto-configuration mode is enabled, RT8973A
would control internal h/w patch for USB D-/D+ switching.

2. Fix code cleanup for other extcon driver
- extcon-sm5502 driver
 : Fix bug to check cable type and build break.
 : Move header file from include/linux/extcon to drivers/extcon because this
   header file is only user for extcon-sm5502.c.
 : Clean up codes by using checkpatch script
- extcon-max77693 driver
 : Use resource managed interrupt function
 : Fix bug to set ADC debounce time
- extcon-gpio driver
 : Fix minor code cleanup

9 years agoextcon: gpio: Fix code cleanup
George Cherian [Tue, 9 Sep 2014 04:14:34 +0000 (09:44 +0530)]
extcon: gpio: Fix code cleanup

This patch fixes following minor cleanup:
 - Order the include files in alphabetical order.
 - Fix description of state_off in extcon_gpio.h
 - Add a descrition for check_on_resume in extcon_gpio.h

Signed-off-by: George Cherian <george.cherian@ti.com>
[Modify the name/description of patch to keep standary codiyg style by Chanwoo Choi]
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
9 years agoextcon: max77693: Fix a bug occured at changing ADC debounce time.
Jonghwa Lee [Wed, 17 Sep 2014 03:58:43 +0000 (12:58 +0900)]
extcon: max77693: Fix a bug occured at changing ADC debounce time.

When it writes some value other than 0 to BTLDset and JIGset, muic device
will be reset automatically. And it happens during updating ADC debounce time,
because it shares same register. To update ADC debounce time without reset,
set value only to ADCDbset and 0 to BTLDset and JIGset.

Signed-off-by: Jonghwa Lee <jonghwa3.lee@samsung.com>
[Remove un-needed masking operation by Chanwoo Choi]
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
9 years agoextcon: sm5502: Drop useless include
Jean Delvare [Tue, 16 Sep 2014 21:16:15 +0000 (23:16 +0200)]
extcon: sm5502: Drop useless include

Don't include <linux/input.h> when the driver does not use anything
from this header file.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Acked-by: MyungJoo Ham <myungjoo.ham@samsung.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Cc: Chanwoo Choi <cw00.choi@samsung.com>
Cc: MyungJoo Ham <myungjoo.ham@samsung.com>
9 years agoextcon: max77693: Use resource managed interrupt line
Krzysztof Kozlowski [Fri, 12 Sep 2014 13:16:37 +0000 (15:16 +0200)]
extcon: max77693: Use resource managed interrupt line

Use resource managed interrupt line devm_request_threaded_irq() to
simplify a little cleanup paths:
 - no goto to cleanup label,
 - simpler remove function.

Overall the driver size is decreased by 11 line of code.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
9 years agodt-bindings: extcon: Add support for Richtek RT8973A MUIC device
Chanwoo Choi [Mon, 4 Aug 2014 06:54:14 +0000 (15:54 +0900)]
dt-bindings: extcon: Add support for Richtek RT8973A MUIC device

This patch add documentation for binding of Richtek RT8973A (Micro USB Switch)
device which is using EXTCON subsystem. The RT8973A device can detect various
external accessories when external accessories is attached or detached.

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
9 years agoextcon: rt8973a: Add Richtek RT8973A extcon driver
Chanwoo Choi [Mon, 18 Aug 2014 00:05:21 +0000 (09:05 +0900)]
extcon: rt8973a: Add Richtek RT8973A extcon driver

This patch add support for Richtek RT8973A which is Micro USB Switch OVP
and i2c interface. The RT8973A is a USB port accessory detector and switch
that is optimized to protect low voltage system from abnormal high input
voltage (up to 28V) and supports high speed USB operation. Also, RT8973A
support 'auto-configuration' mode. If auto-configuration mode is enabled,
RT8973A would control internal h/w patch for USB D-/D+ switching.

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Seung-Woo Kim <sw0312.kim@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
9 years agoextcon: sm5502: Clean up codes by using checkpatch script
Chanwoo Choi [Tue, 12 Aug 2014 01:15:39 +0000 (10:15 +0900)]
extcon: sm5502: Clean up codes by using checkpatch script

This patch just clean up codes by using checkpatch script and fix warning
message about if statement.

- the result of checkpatch script as following:
WARNING: void function return statements are not generally useful
+ return;
+}

WARNING: quoted string split across lines
+ dev_err(info->dev, "failed: irq request (IRQ: %d,"
+ " error :%d)\n", muic_irq->irq, ret);

- warning message about coding style.

drivers/extcon/extcon-sm5502.c:398 sm5502_muic_cable_handler()
warn: we tested 'attached' before and it was 'false'

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
9 years agoextcon: sm5502: Move sm5502.h header file to extcon directory
Chanwoo Choi [Thu, 31 Jul 2014 07:32:46 +0000 (16:32 +0900)]
extcon: sm5502: Move sm5502.h header file to extcon directory

This patch move sm5502.h header file from 'include/linux/extcon' to
'driver/extcon' because sm5502.h is used for driver/extcon/extcon-sm5502.c.
and remove duplicate license description.

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
9 years agoextcon: sm5502: Add I2C dependency to fix build break
Chanwoo Choi [Tue, 12 Aug 2014 04:28:11 +0000 (13:28 +0900)]
extcon: sm5502: Add I2C dependency to fix build break

This patch add I2C configuration dependency to fix following build break.
If specific kernel build I2C as module, extcon-sm5502 have to depend on
I2C configuration.

drivers/built-in.o: In function `regmap_smbus_byte_reg_read':
regmap-i2c.c:(.text+0x5030a): undefined reference to `i2c_smbus_read_byte_data'
drivers/built-in.o: In function `regmap_smbus_byte_reg_write':
regmap-i2c.c:(.text+0x50338): undefined reference to `i2c_smbus_write_byte_data'
drivers/built-in.o: In function `regmap_smbus_word_reg_read':
regmap-i2c.c:(.text+0x50356): undefined reference to `i2c_smbus_read_word_data'
drivers/built-in.o: In function `regmap_smbus_word_reg_write':
regmap-i2c.c:(.text+0x50384): undefined reference to `i2c_smbus_write_word_data'
drivers/built-in.o: In function `regmap_i2c_read':
regmap-i2c.c:(.text+0x503cf): undefined reference to `i2c_transfer'
drivers/built-in.o: In function `regmap_i2c_gather_write':
regmap-i2c.c:(.text+0x50442): undefined reference to `i2c_transfer'
drivers/built-in.o: In function `regmap_i2c_write':
regmap-i2c.c:(.text+0x50474): undefined reference to `i2c_master_send'
drivers/built-in.o: In function `sm5502_muic_i2c_init':
extcon-sm5502.c:(.init.text+0x6630): undefined reference to `i2c_register_driver'

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
9 years agoextcon: sm5502: Fix bug to check cable type
Chanwoo Choi [Wed, 30 Jul 2014 06:39:02 +0000 (15:39 +0900)]
extcon: sm5502: Fix bug to check cable type

This patch fix bug when checking cable type. SM5502 have to use ADC value
to get correct cable type.

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
9 years agoLinux 3.17-rc5
Linus Torvalds [Mon, 15 Sep 2014 00:50:12 +0000 (17:50 -0700)]
Linux 3.17-rc5

9 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Mon, 15 Sep 2014 00:37:36 +0000 (17:37 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs

Pull vfs fixes from Al Viro:
 "double iput() on failure exit in lustre, racy removal of spliced
  dentries from ->s_anon in __d_materialise_dentry() plus a bunch of
  assorted RCU pathwalk fixes"

The RCU pathwalk fixes end up fixing a couple of cases where we
incorrectly dropped out of RCU walking, due to incorrect initialization
and testing of the sequence locks in some corner cases.  Since dropping
out of RCU walk mode forces the slow locked accesses, those corner cases
slowed down quite dramatically.

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  be careful with nd->inode in path_init() and follow_dotdot_rcu()
  don't bugger nd->seq on set_root_rcu() from follow_dotdot_rcu()
  fix bogus read_seqretry() checks introduced in b37199e
  move the call of __d_drop(anon) into __d_materialise_unique(dentry, anon)
  [fix] lustre: d_make_root() does iput() on dentry allocation failure

9 years agovfs: avoid non-forwarding large load after small store in path lookup
Linus Torvalds [Mon, 15 Sep 2014 00:28:32 +0000 (17:28 -0700)]
vfs: avoid non-forwarding large load after small store in path lookup

The performance regression that Josef Bacik reported in the pathname
lookup (see commit 99d263d4c5b2 "vfs: fix bad hashing of dentries") made
me look at performance stability of the dcache code, just to verify that
the problem was actually fixed.  That turned up a few other problems in
this area.

There are a few cases where we exit RCU lookup mode and go to the slow
serializing case when we shouldn't, Al has fixed those and they'll come
in with the next VFS pull.

But my performance verification also shows that link_path_walk() turns
out to have a very unfortunate 32-bit store of the length and hash of
the name we look up, followed by a 64-bit read of the combined hash_len
field.  That screws up the processor store to load forwarding, causing
an unnecessary hickup in this critical routine.

It's caused by the ugly calling convention for the "hash_name()"
function, and easily fixed by just making hash_name() fill in the whole
'struct qstr' rather than passing it a pointer to just the hash value.

With that, the profile for this function looks much smoother.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoMerge branch 'parisc-3.17-1' of git://git.kernel.org/pub/scm/linux/kernel/git/deller...
Linus Torvalds [Sun, 14 Sep 2014 19:28:08 +0000 (12:28 -0700)]
Merge branch 'parisc-3.17-1' of git://git./linux/kernel/git/deller/parisc-linux

Pull parisc updates from Helge Deller:
 "The most important patch is a new Light Weigth Syscall (LWS) for 8,
  16, 32 and 64 bit atomic CAS operations which is required in order to
  be able to implement the atomic gcc builtins on our platform.

  Other than that, we wire up the seccomp, getrandom and memfd_create
  syscalls, fixes a minor off-by-one bug and a wrong printk string"

* 'parisc-3.17-1' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: Implement new LWS CAS supporting 64 bit operations.
  parisc: Wire up seccomp, getrandom and memfd_create syscalls
  parisc: dino: fix %d confusingly prefixed with 0x in format string
  parisc: sys_hpux: NUL terminator is one past the end

9 years agobe careful with nd->inode in path_init() and follow_dotdot_rcu()
Al Viro [Sun, 14 Sep 2014 01:59:43 +0000 (21:59 -0400)]
be careful with nd->inode in path_init() and follow_dotdot_rcu()

in the former we simply check if dentry is still valid after picking
its ->d_inode; in the latter we fetch ->d_inode in the same places
where we fetch dentry and its ->d_seq, under the same checks.

Cc: stable@vger.kernel.org # 2.6.38+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
9 years agodon't bugger nd->seq on set_root_rcu() from follow_dotdot_rcu()
Al Viro [Sun, 14 Sep 2014 01:55:46 +0000 (21:55 -0400)]
don't bugger nd->seq on set_root_rcu() from follow_dotdot_rcu()

return the value instead, and have path_init() do the assignment.  Broken by
"vfs: Fix absolute RCU path walk failures due to uninitialized seq number",
which was Cc-stable with 2.6.38+ as destination.  This one should go where
it went.

To avoid dummy value returned in case when root is already set (it would do
no harm, actually, since the only caller that doesn't ignore the return value
is guaranteed to have nd->root *not* set, but it's more obvious that way),
lift the check into callers.  And do the same to set_root(), to keep them
in sync.

Cc: stable@vger.kernel.org # 2.6.38+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
9 years agoMerge tag 'ntb-3.17' of git://github.com/jonmason/ntb
Linus Torvalds [Sun, 14 Sep 2014 17:54:12 +0000 (10:54 -0700)]
Merge tag 'ntb-3.17' of git://github.com/jonmason/ntb

Pull ntb driver bugfixes from Jon Mason:
 "NTB driver fixes for queue spread and buffer alignment.  Also, update
  to MAINTAINERS to reflect new e-mail address"

* tag 'ntb-3.17' of git://github.com/jonmason/ntb:
  ntb: Add alignment check to meet hardware requirement
  MAINTAINERS: update NTB info
  NTB: correct the spread of queues over mw's

9 years agoMerge branch 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 14 Sep 2014 17:37:10 +0000 (10:37 -0700)]
Merge branch 'irq-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull ARM irq chip fixes from Thomas Gleixner:
 "Another pile of ARM specific irq chip fixlets:

   - off by one bugs in the crossbar driver
   - missing annotations
   - a bunch of "make it compile" updates

  I pulled the lot today from Jason, but it has been in -next for at
  least a week"

* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  irqchip: gic-v3: Declare rdist as __percpu pointer to __iomem pointer
  irqchip: gic: Make gic_default_routable_irq_domain_ops static
  irqchip: exynos-combiner: Fix compilation error on ARM64
  irqchip: crossbar: Off by one bugs in init
  irqchip: gic-v3: Tag all low level accessors __maybe_unused
  irqchip: gic-v3: Only define gic_peek_irq() when building SMP

9 years agoMerge tag 'irqchip-urgent-3.17' of git://git.infradead.org/users/jcooper/linux into...
Thomas Gleixner [Sun, 14 Sep 2014 13:19:19 +0000 (15:19 +0200)]
Merge tag 'irqchip-urgent-3.17' of git://git.infradead.org/users/jcooper/linux into irq/urgent

irqchip fixes for v3.17 from Jason Cooper

 - GIC/GICV3: Various fixlets
 - crossbar: Fix off-by-one bug
 - exynos-combiner: Fix arm64 build error

9 years agontb: Add alignment check to meet hardware requirement
Dave Jiang [Thu, 28 Aug 2014 20:53:02 +0000 (13:53 -0700)]
ntb: Add alignment check to meet hardware requirement

The NTB translate register must have the value to be BAR size aligned.
This alignment check make sure that the DMA memory allocated has the
proper alignment. Another requirement for NTB to function properly with
memory window BAR size greater or equal to 4M is to use the CMA feature
in 3.16 kernel with the appropriate CONFIG_CMA_ALIGNMENT and
CONFIG_CMA_SIZE_MBYTES set.

Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
9 years agoMAINTAINERS: update NTB info
Jon Mason [Thu, 19 Jun 2014 17:44:24 +0000 (10:44 -0700)]
MAINTAINERS: update NTB info

Update my contact info to my personal email address and add Dave Jiang.

Signed-off-by: Jon Mason <jon.mason@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
9 years agoNTB: correct the spread of queues over mw's
Jon Mason [Thu, 19 Jun 2014 17:11:13 +0000 (10:11 -0700)]
NTB: correct the spread of queues over mw's

The detection of an uneven number of queues on the given memory windows
was not correct.  The mw_num is zero based and the mod should be
division to spread them evenly over the mw's.

Signed-off-by: Jon Mason <jon.mason@intel.com>
9 years agofix bogus read_seqretry() checks introduced in b37199e
Al Viro [Sun, 14 Sep 2014 01:50:45 +0000 (21:50 -0400)]
fix bogus read_seqretry() checks introduced in b37199e

read_seqretry() returns true on mismatch, not on match...

Cc: stable@vger.kernel.org # 3.15+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
9 years agomove the call of __d_drop(anon) into __d_materialise_unique(dentry, anon)
Al Viro [Thu, 11 Sep 2014 22:55:50 +0000 (18:55 -0400)]
move the call of __d_drop(anon) into __d_materialise_unique(dentry, anon)

and lock the right list there

Cc: stable@vger.kernel.org
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
9 years ago[fix] lustre: d_make_root() does iput() on dentry allocation failure
Al Viro [Wed, 3 Sep 2014 17:11:09 +0000 (13:11 -0400)]
[fix] lustre: d_make_root() does iput() on dentry allocation failure

double-free is a bad thing

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
9 years agoMerge branches 'locking-urgent-for-linus' and 'timers-urgent-for-linus' of git:/...
Linus Torvalds [Sat, 13 Sep 2014 21:22:12 +0000 (14:22 -0700)]
Merge branches 'locking-urgent-for-linus' and 'timers-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull futex and timer fixes from Thomas Gleixner:
 "A oneliner bugfix for the jinxed futex code:

   - Drop hash bucket lock in the error exit path.  I really could slap
     myself for intruducing that bug while fixing all the other horror
     in that code three month ago ...

  and the timer department is not too proud about the following fixes:

   - Deal with a long standing rounding bug in the timeval to jiffies
     conversion.  It's a real issue and this fix fell through the cracks
     for quite some time.

   - Another round of alarmtimer fixes.  Finally this code gets used
     more widely and the subtle issues hidden for quite some time are
     noticed and fixed.  Nothing really exciting, just the itty bitty
     details which bite the serious users here and there"

* 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  futex: Unlock hb->lock in futex_wait_requeue_pi() error path

* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  alarmtimer: Lock k_itimer during timer callback
  alarmtimer: Do not signal SIGEV_NONE timers
  alarmtimer: Return relative times in timer_gettime
  jiffies: Fix timeval conversion to jiffies

9 years agoparisc: Implement new LWS CAS supporting 64 bit operations.
Guy Martin [Fri, 12 Sep 2014 16:02:34 +0000 (18:02 +0200)]
parisc: Implement new LWS CAS supporting 64 bit operations.

The current LWS cas only works correctly for 32bit. The new LWS allows
for CAS operations of variable size.

Signed-off-by: Guy Martin <gmsoft@tuxicoman.be>
Cc: <stable@vger.kernel.org> # 3.13+
Signed-off-by: Helge Deller <deller@gmx.de>
9 years agovfs: fix bad hashing of dentries
Linus Torvalds [Sat, 13 Sep 2014 18:30:10 +0000 (11:30 -0700)]
vfs: fix bad hashing of dentries

Josef Bacik found a performance regression between 3.2 and 3.10 and
narrowed it down to commit bfcfaa77bdf0 ("vfs: use 'unsigned long'
accesses for dcache name comparison and hashing"). He reports:

 "The test case is essentially

      for (i = 0; i < 1000000; i++)
              mkdir("a$i");

  On xfs on a fio card this goes at about 20k dir/sec with 3.2, and 12k
  dir/sec with 3.10.  This is because we spend waaaaay more time in
  __d_lookup on 3.10 than in 3.2.

  The new hashing function for strings is suboptimal for <
  sizeof(unsigned long) string names (and hell even > sizeof(unsigned
  long) string names that I've tested).  I broke out the old hashing
  function and the new one into a userspace helper to get real numbers
  and this is what I'm getting:

      Old hash table had 1000000 entries, 0 dupes, 0 max dupes
      New hash table had 12628 entries, 987372 dupes, 900 max dupes
      We had 11400 buckets with a p50 of 30 dupes, p90 of 240 dupes, p99 of 567 dupes for the new hash

  My test does the hash, and then does the d_hash into a integer pointer
  array the same size as the dentry hash table on my system, and then
  just increments the value at the address we got to see how many
  entries we overlap with.

  As you can see the old hash function ended up with all 1 million
  entries in their own bucket, whereas the new one they are only
  distributed among ~12.5k buckets, which is why we're using so much
  more CPU in __d_lookup".

The reason for this hash regression is two-fold:

 - On 64-bit architectures the down-mixing of the original 64-bit
   word-at-a-time hash into the final 32-bit hash value is very
   simplistic and suboptimal, and just adds the two 32-bit parts
   together.

   In particular, because there is no bit shuffling and the mixing
   boundary is also a byte boundary, similar character patterns in the
   low and high word easily end up just canceling each other out.

 - the old byte-at-a-time hash mixed each byte into the final hash as it
   hashed the path component name, resulting in the low bits of the hash
   generally being a good source of hash data.  That is not true for the
   word-at-a-time case, and the hash data is distributed among all the
   bits.

The fix is the same in both cases: do a better job of mixing the bits up
and using as much of the hash data as possible.  We already have the
"hash_32|64()" functions to do that.

Reported-by: Josef Bacik <jbacik@fb.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Chris Mason <clm@fb.com>
Cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoMake hash_64() use a 64-bit multiply when appropriate
Linus Torvalds [Sat, 13 Sep 2014 18:24:03 +0000 (11:24 -0700)]
Make hash_64() use a 64-bit multiply when appropriate

The hash_64() function historically does the multiply by the
GOLDEN_RATIO_PRIME_64 number with explicit shifts and adds, because
unlike the 32-bit case, gcc seems unable to turn the constant multiply
into the more appropriate shift and adds when required.

However, that means that we generate those shifts and adds even when the
architecture has a fast multiplier, and could just do it better in
hardware.

Use the now-cleaned-up CONFIG_ARCH_HAS_FAST_MULTIPLIER (together with
"is it a 64-bit architecture") to decide whether to use an integer
multiply or the explicit sequence of shift/add instructions.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoMake ARCH_HAS_FAST_MULTIPLIER a real config variable
Linus Torvalds [Sat, 13 Sep 2014 18:14:53 +0000 (11:14 -0700)]
Make ARCH_HAS_FAST_MULTIPLIER a real config variable

It used to be an ad-hoc hack defined by the x86 version of
<asm/bitops.h> that enabled a couple of library routines to know whether
an integer multiply is faster than repeated shifts and additions.

This just makes it use the real Kconfig system instead, and makes x86
(which was the only architecture that did this) select the option.

NOTE! Even for x86, this really is kind of wrong.  If we cared, we would
probably not enable this for builds optimized for netburst (P4), where
shifts-and-adds are generally faster than multiplies.  This patch does
*not* change that kind of logic, though, it is purely a syntactic change
with no code changes.

This was triggered by the fact that we have other places that really
want to know "do I want to expand multiples by constants by hand or
not", particularly the hash generation code.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
9 years agoMerge tag 'dm-3.17-fix2' of git://git.kernel.org/pub/scm/linux/kernel/git/device...
Linus Torvalds [Sat, 13 Sep 2014 17:04:10 +0000 (10:04 -0700)]
Merge tag 'dm-3.17-fix2' of git://git./linux/kernel/git/device-mapper/linux-dm

Pull device mapper fix from Mike Snitzer:
 "Fix a race in the DM cache target that caused dirty blocks to be
  marked as clean.  This could cause no writeback to occur or spurious
  dirty block counts"

* tag 'dm-3.17-fix2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm cache: fix race causing dirty blocks to be marked as clean

9 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-block
Linus Torvalds [Sat, 13 Sep 2014 16:39:55 +0000 (09:39 -0700)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:
 "A small collection of fixes for the current rc series.  This contains:

   - Two small blk-mq patches from Rob Elliott, cleaning up error case
     at init time.

   - A fix from Ming Lei, fixing SG merging for blk-mq where
     QUEUE_FLAG_SG_NO_MERGE is the default.

   - A dev_t minor lifetime fix from Keith, fixing an issue where a
     minor might be reused before all references to it were gone.

   - Fix from Alan Stern where an unbalanced queue bypass caused SCSI
     some headaches when it does a series of add/del on devices without
     fully registrering the queue.

   - A fix from me for improving the scaling of tag depth in blk-mq if
     we are short on memory"

* 'for-linus' of git://git.kernel.dk/linux-block:
  blk-mq: scale depth and rq map appropriate if low on memory
  Block: fix unbalanced bypass-disable in blk_register_queue
  block: Fix dev_t minor allocation lifetime
  blk-mq: cleanup after blk_mq_init_rq_map failures
  blk-mq: pass along blk_mq_alloc_tag_set return values
  blk-merge: fix blk_recount_segments

9 years agoMerge tag 'stable/for-linus-3.17-b-rc4-arm-tag' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Sat, 13 Sep 2014 00:45:27 +0000 (17:45 -0700)]
Merge tag 'stable/for-linus-3.17-b-rc4-arm-tag' of git://git./linux/kernel/git/xen/tip

Pull Xen ARM bugfix from Stefano Stabellini:
 "The patches fix the "xen_add_mach_to_phys_entry: cannot add" bug that
  has been affecting xen on arm and arm64 guests since 3.16.  They
  require a few hypervisor side changes that just went in xen-unstable.

  A couple of days ago David sent out a pull request with a few other
  Xen fixes (it is already in master).  Sorry we didn't synchronized
  better among us"

* tag 'stable/for-linus-3.17-b-rc4-arm-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen/arm: remove mach_to_phys rbtree
  xen/arm: reimplement xen_dma_unmap_page & friends
  xen/arm: introduce XENFEAT_grant_map_identity

9 years agoalarmtimer: Lock k_itimer during timer callback
Richard Larocque [Wed, 10 Sep 2014 01:31:05 +0000 (18:31 -0700)]
alarmtimer: Lock k_itimer during timer callback

Locks the k_itimer's it_lock member when handling the alarm timer's
expiry callback.

The regular posix timers defined in posix-timers.c have this lock held
during timout processing because their callbacks are routed through
posix_timer_fn().  The alarm timers follow a different path, so they
ought to grab the lock somewhere else.

Cc: stable@vger.kernel.org
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Sharvil Nanavati <sharvil@google.com>
Signed-off-by: Richard Larocque <rlarocque@google.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
9 years agoalarmtimer: Do not signal SIGEV_NONE timers
Richard Larocque [Wed, 10 Sep 2014 01:31:04 +0000 (18:31 -0700)]
alarmtimer: Do not signal SIGEV_NONE timers

Avoids sending a signal to alarm timers created with sigev_notify set to
SIGEV_NONE by checking for that special case in the timeout callback.

The regular posix timers avoid sending signals to SIGEV_NONE timers by
not scheduling any callbacks for them in the first place.  Although it
would be possible to do something similar for alarm timers, it's simpler
to handle this as a special case in the timeout.

Prior to this patch, the alarm timer would ignore the sigev_notify value
and try to deliver signals to the process anyway.  Even worse, the
sanity check for the value of sigev_signo is skipped when SIGEV_NONE was
specified, so the signal number could be bogus.  If sigev_signo was an
unitialized value (as it often would be if SIGEV_NONE is used), then
it's hard to predict which signal will be sent.

Cc: stable@vger.kernel.org
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Sharvil Nanavati <sharvil@google.com>
Signed-off-by: Richard Larocque <rlarocque@google.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
9 years agoalarmtimer: Return relative times in timer_gettime
Richard Larocque [Wed, 10 Sep 2014 01:31:03 +0000 (18:31 -0700)]
alarmtimer: Return relative times in timer_gettime

Returns the time remaining for an alarm timer, rather than the time at
which it is scheduled to expire.  If the timer has already expired or it
is not currently scheduled, the it_value's members are set to zero.

This new behavior matches that of the other posix-timers and the POSIX
specifications.

This is a change in user-visible behavior, and may break existing
applications.  Hopefully, few users rely on the old incorrect behavior.

Cc: stable@vger.kernel.org
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Sharvil Nanavati <sharvil@google.com>
Signed-off-by: Richard Larocque <rlarocque@google.com>
[jstultz: minor style tweak]
Signed-off-by: John Stultz <john.stultz@linaro.org>
9 years agojiffies: Fix timeval conversion to jiffies
Andrew Hunter [Thu, 4 Sep 2014 21:17:16 +0000 (14:17 -0700)]
jiffies: Fix timeval conversion to jiffies

timeval_to_jiffies tried to round a timeval up to an integral number
of jiffies, but the logic for doing so was incorrect: intervals
corresponding to exactly N jiffies would become N+1. This manifested
itself particularly repeatedly stopping/starting an itimer:

setitimer(ITIMER_PROF, &val, NULL);
setitimer(ITIMER_PROF, NULL, &val);

would add a full tick to val, _even if it was exactly representable in
terms of jiffies_ (say, the result of a previous rounding.)  Doing
this repeatedly would cause unbounded growth in val.  So fix the math.

Here's what was wrong with the conversion: we essentially computed
(eliding seconds)

jiffies = usec  * (NSEC_PER_USEC/TICK_NSEC)

by using scaling arithmetic, which took the best approximation of
NSEC_PER_USEC/TICK_NSEC with denominator of 2^USEC_JIFFIE_SC =
x/(2^USEC_JIFFIE_SC), and computed:

jiffies = (usec * x) >> USEC_JIFFIE_SC

and rounded this calculation up in the intermediate form (since we
can't necessarily exactly represent TICK_NSEC in usec.) But the
scaling arithmetic is a (very slight) *over*approximation of the true
value; that is, instead of dividing by (1 usec/ 1 jiffie), we
effectively divided by (1 usec/1 jiffie)-epsilon (rounding
down). This would normally be fine, but we want to round timeouts up,
and we did so by adding 2^USEC_JIFFIE_SC - 1 before the shift; this
would be fine if our division was exact, but dividing this by the
slightly smaller factor was equivalent to adding just _over_ 1 to the
final result (instead of just _under_ 1, as desired.)

In particular, with HZ=1000, we consistently computed that 10000 usec
was 11 jiffies; the same was true for any exact multiple of
TICK_NSEC.

We could possibly still round in the intermediate form, adding
something less than 2^USEC_JIFFIE_SC - 1, but easier still is to
convert usec->nsec, round in nanoseconds, and then convert using
time*spec*_to_jiffies.  This adds one constant multiplication, and is
not observably slower in microbenchmarks on recent x86 hardware.

Tested: the following program:

int main() {
  struct itimerval zero = {{0, 0}, {0, 0}};
  /* Initially set to 10 ms. */
  struct itimerval initial = zero;
  initial.it_interval.tv_usec = 10000;
  setitimer(ITIMER_PROF, &initial, NULL);
  /* Save and restore several times. */
  for (size_t i = 0; i < 10; ++i) {
    struct itimerval prev;
    setitimer(ITIMER_PROF, &zero, &prev);
    /* on old kernels, this goes up by TICK_USEC every iteration */
    printf("previous value: %ld %ld %ld %ld\n",
           prev.it_interval.tv_sec, prev.it_interval.tv_usec,
           prev.it_value.tv_sec, prev.it_value.tv_usec);
    setitimer(ITIMER_PROF, &prev, NULL);
  }
    return 0;
}

Cc: stable@vger.kernel.org
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Paul Turner <pjt@google.com>
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Reviewed-by: Paul Turner <pjt@google.com>
Reported-by: Aaron Jacobs <jacobsa@google.com>
Signed-off-by: Andrew Hunter <ahh@google.com>
[jstultz: Tweaked to apply to 3.17-rc]
Signed-off-by: John Stultz <john.stultz@linaro.org>
9 years agofutex: Unlock hb->lock in futex_wait_requeue_pi() error path
Thomas Gleixner [Thu, 11 Sep 2014 21:44:35 +0000 (23:44 +0200)]
futex: Unlock hb->lock in futex_wait_requeue_pi() error path

futex_wait_requeue_pi() calls futex_wait_setup(). If
futex_wait_setup() succeeds it returns with hb->lock held and
preemption disabled. Now the sanity check after this does:

        if (match_futex(&q.key, &key2)) {
    ret = -EINVAL;
goto out_put_keys;
}

which releases the keys but does not release hb->lock.

So we happily return to user space with hb->lock held and therefor
preemption disabled.

Unlock hb->lock before taking the exit route.

Reported-by: Dave "Trinity" Jones <davej@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Darren Hart <dvhart@linux.intel.com>
Reviewed-by: Davidlohr Bueso <dave@stgolabs.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/alpine.DEB.2.10.1409112318500.4178@nanos
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
9 years agoMerge tag 'char-misc-3.17-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregk...
Linus Torvalds [Fri, 12 Sep 2014 19:00:49 +0000 (12:00 -0700)]
Merge tag 'char-misc-3.17-rc5' of git://git./linux/kernel/git/gregkh/char-misc

Pull char/misc driver fix from Greg KH:
 "Here is one misc driver fix for 3.17-rc5.  It resolves a kernel oops
  that can happen in the lattice FPGA driver if the firmware isn't
  present on the system.

  It's been in the linux-next tree for a while now"

* tag 'char-misc-3.17-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  Lattice ECP3 FPGA: Check firmware pointer

9 years agoMerge tag 'staging-3.17-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Fri, 12 Sep 2014 19:00:18 +0000 (12:00 -0700)]
Merge tag 'staging-3.17-rc5' of git://git./linux/kernel/git/gregkh/staging

Pull staging driver fixes from Greg KH:
 "Here are 3 tiny staging driver fixes for 3.17-rc5.

  Two are fixes for the imx-drm driver, resolving issues that have been
  reported.  The other is a memory leak fix for the Android sync driver,
  due to changes that went into 3.17-rc1.

  All have been in linux-next for a while"

* tag 'staging-3.17-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  android: fix reference leak in sync_fence_create
  imx-drm: imx-ldb: fix NULL pointer in imx_ldb_unbind()
  imx-drm: ipuv3-plane: fix ipu_plane_dpms()

9 years agoMerge tag 'tty-3.17-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Fri, 12 Sep 2014 18:59:48 +0000 (11:59 -0700)]
Merge tag 'tty-3.17-rc5' of git://git./linux/kernel/git/gregkh/tty

Pull tty/serial fixes from Greg KH:
 "Here are 3 patches for 3.17-rc5.  Two serial driver fixes that resolve
  some reported issues, and one new device id.

  All have been in linux-next just fine"

* tag 'tty-3.17-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  tty: xuartps: Fix tx_emtpy() callback
  tty/serial: at91: BUG: disable interrupts when !UART_ENABLE_MS()
  serial: 8250_dw: Add ACPI ID for Intel Braswell

9 years agoMerge tag 'usb-3.17-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Fri, 12 Sep 2014 18:59:10 +0000 (11:59 -0700)]
Merge tag 'usb-3.17-rc5' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are some USB and PHY fixes for 3.17-rc5.

  Nothing major here, just a number of tiny fixes for reported issues,
  and some new device ids as well.

  All have been tested in linux-next"

* tag 'usb-3.17-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (46 commits)
  xhci: fix oops when xhci resumes from hibernate with hw lpm capable devices
  usb: xhci: Fix OOPS in xhci error handling code
  xhci: Fix null pointer dereference if xhci initialization fails
  storage: Add single-LUN quirk for Jaz USB Adapter
  uas: Add missing le16_to_cpu calls to asm1051 / asm1053 usb-id check
  usb: chipidea: msm: Initialize PHY on reset event
  usb: chipidea: msm: Use USB PHY API to control PHY state
  usb: hub: take hub->hdev reference when processing from eventlist
  uas: Disable uas on ASM1051 devices
  usb: dwc2/gadget: avoid disabling ep0
  usb: dwc2/gadget: delay enabling irq once hardware is configured properly
  usb: dwc2/gadget: do not call disconnect method in pullup
  usb: dwc2/gadget: break infinite loop in endpoint disable code
  usb: dwc2/gadget: fix phy initialization sequence
  usb: dwc2/gadget: fix phy disable sequence
  uwb: init beacon cache entry before registering uwb device
  USB: ftdi_sio: Add support for GE Healthcare Nemo Tracker device
  USB: document the 'u' flag for usb-storage quirks parameter
  usb: host: xhci: fix compliance mode workaround
  usb: dwc3: fix TRB completion when multiple TRBs are started
  ...

9 years agoMerge tag 'nfs-for-3.17-4' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Fri, 12 Sep 2014 18:54:54 +0000 (11:54 -0700)]
Merge tag 'nfs-for-3.17-4' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client fixes from Trond Myklebust:
 "Highlights:
   - fix a kernel warning when removing /proc/net/nfsfs
   - revert commit 49a4bda22e18 due to Oopses
   - fix a typo in the pNFS file layout commit code"

* tag 'nfs-for-3.17-4' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  pnfs: fix filelayout_retry_commit when idx > 0
  nfs: revert "nfs4: queue free_lock_state job submission to nfsiod"
  nfs: fix kernel warning when removing proc entry

9 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux...
Linus Torvalds [Fri, 12 Sep 2014 18:53:30 +0000 (11:53 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mason/linux-btrfs

Pull btrfs fixes from Chris Mason:
 "Filipe is doing a careful pass through fsync problems, and these are
  the fixes so far.  I'll have one more for rc6 that we're still
  testing.

  My big commit is fixing up some inode hash races that Al Viro found
  (thanks Al)"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  Btrfs: use insert_inode_locked4 for inode creation
  Btrfs: fix fsync data loss after a ranged fsync
  Btrfs: kfree()ing ERR_PTRs
  Btrfs: fix crash while doing a ranged fsync
  Btrfs: fix corruption after write/fsync failure + fsync + log recovery
  Btrfs: fix autodefrag with compression

9 years agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Fri, 12 Sep 2014 16:53:47 +0000 (09:53 -0700)]
Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fixes from Will Deacon:
 "Just a couple of stragglers here:

   - fix an issue migrating interrupts on CPU hotplug
   - fix a potential information leak of TLS registers across an exec
     (Nathan has sent a corresponding patch for arch/arm/ to rmk)"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: flush TLS registers during exec
  arm64: use irq_set_affinity with force=false when migrating irqs

9 years agoMerge tag 'iommu-fixes-v3.17-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 12 Sep 2014 16:26:49 +0000 (09:26 -0700)]
Merge tag 'iommu-fixes-v3.17-rc4' of git://git./linux/kernel/git/joro/iommu

Pull iommu fixes from Joerg Roedel:
 - two fixes for issues found by Coverity
 - various fixes for the ARM SMMU driver
 - a warning fix for the FSL PAMU driver

* tag 'iommu-fixes-v3.17-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
  iommu/fsl: Fix warning resulting from adding PCI device twice
  iommu/arm-smmu: fix corner cases in address size calculations
  iommu/arm-smmu: fix decimal printf format specifiers prefixed with 0x
  iommu/arm-smmu: Do not access non-existing S2CR registers
  iommu/arm-smmu: fix s2cr and smr teardown on device detach from domain
  iommu/arm-smmu: remove pgtable_page_{c,d}tor()
  iommu/arm-smmu: fix programming of SMMU_CBn_TCR for stage 1
  iommu/arm-smmu: avoid calling request_irq in atomic context
  iommu/vt-d: Check return value of acpi_bus_get_device()
  iommu/core: Make iommu_group_get_for_dev() more robust

9 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris...
Linus Torvalds [Fri, 12 Sep 2014 16:24:46 +0000 (09:24 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jmorris/linux-security

Pull assoc array garbage collection fix from James Morris.

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security:
  KEYS: Fix termination condition in assoc array garbage collection

9 years agoMerge tag 'fbdev-fixes-3.17' of git://git.kernel.org/pub/scm/linux/kernel/git/tomba...
Linus Torvalds [Fri, 12 Sep 2014 16:11:37 +0000 (09:11 -0700)]
Merge tag 'fbdev-fixes-3.17' of git://git./linux/kernel/git/tomba/linux

Pull fbdev fixes from Tomi Valkeinen:
 "Minor fixes for amba-clcd and video DT bindings"

* tag 'fbdev-fixes-3.17' of git://git.kernel.org/pub/scm/linux/kernel/git/tomba/linux:
  video: ARM CLCD: Fix color model capabilities for DT platforms
  video: fix composite video connector compatible string

9 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Fri, 12 Sep 2014 15:27:40 +0000 (08:27 -0700)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "AST, i915, radeon and msm fixes, all over the place.

  All fixing build issues, regressions, oopses or failure to detect
  cards"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/ast: AST2000 cannot be detected correctly
  drm/ast: open key before detect chips
  drm/msm: don't crash if no msm.vram param
  drm/msm/hdmi: fix build break on non-CCF platforms
  drm/msm: Change nested function to static function
  drm/radeon/dpm: set the thermal type properly for special configs
  drm/radeon: reduce memory footprint for debugging
  drm/radeon: add connector quirk for fujitsu board
  drm/radeon: fix semaphore value init
  drm/radeon: only use me/pfp sync on evergreen+
  drm/i915: Wait for vblank before enabling the TV encoder
  drm/i915: Evict CS TLBs between batches
  drm/i915: Fix irq enable tracking in driver load
  drm/i915: Fix EIO/wedged handling in gem fault handler
  drm/i915: Prevent recursive deadlock on releasing a busy userptr

9 years agoKEYS: Fix termination condition in assoc array garbage collection
David Howells [Wed, 10 Sep 2014 21:22:00 +0000 (22:22 +0100)]
KEYS: Fix termination condition in assoc array garbage collection

This fixes CVE-2014-3631.

It is possible for an associative array to end up with a shortcut node at the
root of the tree if there are more than fan-out leaves in the tree, but they
all crowd into the same slot in the lowest level (ie. they all have the same
first nibble of their index keys).

When assoc_array_gc() returns back up the tree after scanning some leaves, it
can fall off of the root and crash because it assumes that the back pointer
from a shortcut (after label ascend_old_tree) must point to a normal node -
which isn't true of a shortcut node at the root.

Should we find we're ascending rootwards over a shortcut, we should check to
see if the backpointer is zero - and if it is, we have completed the scan.

This particular bug cannot occur if the root node is not a shortcut - ie. if
you have fewer than 17 keys in a keyring or if you have at least two keys that
sit into separate slots (eg. a keyring and a non keyring).

This can be reproduced by:

ring=`keyctl newring bar @s`
for ((i=1; i<=18; i++)); do last_key=`keyctl newring foo$i $ring`; done
keyctl timeout $last_key 2

Doing this:

echo 3 >/proc/sys/kernel/keys/gc_delay

first will speed things up.

If we do fall off of the top of the tree, we get the following oops:

BUG: unable to handle kernel NULL pointer dereference at 0000000000000018
IP: [<ffffffff8136cea7>] assoc_array_gc+0x2f7/0x540
PGD dae15067 PUD cfc24067 PMD 0
Oops: 0000 [#1] SMP
Modules linked in: xt_nat xt_mark nf_conntrack_netbios_ns nf_conntrack_broadcast ip6t_rpfilter ip6t_REJECT xt_conntrack ebtable_nat ebtable_broute bridge stp llc ebtable_filter ebtables ip6table_ni
CPU: 0 PID: 26011 Comm: kworker/0:1 Not tainted 3.14.9-200.fc20.x86_64 #1
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
Workqueue: events key_garbage_collector
task: ffff8800918bd580 ti: ffff8800aac14000 task.ti: ffff8800aac14000
RIP: 0010:[<ffffffff8136cea7>] [<ffffffff8136cea7>] assoc_array_gc+0x2f7/0x540
RSP: 0018:ffff8800aac15d40  EFLAGS: 00010206
RAX: 0000000000000000 RBX: 0000000000000000 RCX: ffff8800aaecacc0
RDX: ffff8800daecf440 RSI: 0000000000000001 RDI: ffff8800aadc2bc0
RBP: ffff8800aac15da8 R08: 0000000000000001 R09: 0000000000000003
R10: ffffffff8136ccc7 R11: 0000000000000000 R12: 0000000000000000
R13: 0000000000000000 R14: 0000000000000070 R15: 0000000000000001
FS:  0000000000000000(0000) GS:ffff88011fc00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000018 CR3: 00000000db10d000 CR4: 00000000000006f0
Stack:
 ffff8800aac15d50 0000000000000011 ffff8800aac15db8 ffffffff812e2a70
 ffff880091a00600 0000000000000000 ffff8800aadc2bc3 00000000cd42c987
 ffff88003702df20 ffff88003702dfa0 0000000053b65c09 ffff8800aac15fd8
Call Trace:
 [<ffffffff812e2a70>] ? keyring_detect_cycle_iterator+0x30/0x30
 [<ffffffff812e3e75>] keyring_gc+0x75/0x80
 [<ffffffff812e1424>] key_garbage_collector+0x154/0x3c0
 [<ffffffff810a67b6>] process_one_work+0x176/0x430
 [<ffffffff810a744b>] worker_thread+0x11b/0x3a0
 [<ffffffff810a7330>] ? rescuer_thread+0x3b0/0x3b0
 [<ffffffff810ae1a8>] kthread+0xd8/0xf0
 [<ffffffff810ae0d0>] ? insert_kthread_work+0x40/0x40
 [<ffffffff816ffb7c>] ret_from_fork+0x7c/0xb0
 [<ffffffff810ae0d0>] ? insert_kthread_work+0x40/0x40
Code: 08 4c 8b 22 0f 84 bf 00 00 00 41 83 c7 01 49 83 e4 fc 41 83 ff 0f 4c 89 65 c0 0f 8f 5a fe ff ff 48 8b 45 c0 4d 63 cf 49 83 c1 02 <4e> 8b 34 c8 4d 85 f6 0f 84 be 00 00 00 41 f6 c6 01 0f 84 92
RIP  [<ffffffff8136cea7>] assoc_array_gc+0x2f7/0x540
 RSP <ffff8800aac15d40>
CR2: 0000000000000018
---[ end trace 1129028a088c0cbd ]---

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: James Morris <james.l.morris@oracle.com>