firefly-linux-kernel-4.4.55.git
10 years agoclk: rockchip: fix rk3066 pll status register location
Heiko Stuebner [Mon, 1 Sep 2014 21:52:40 +0000 (23:52 +0200)]
clk: rockchip: fix rk3066 pll status register location

The register providing the pll lock status is at a different address on the
rk3066. The error became apparent while working on cpufreq support for
the rockchip SoCs.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
10 years agoclk: rockchip: change pll rate without a clk-notifier
Doug Anderson [Tue, 16 Sep 2014 04:07:57 +0000 (21:07 -0700)]
clk: rockchip: change pll rate without a clk-notifier

The Rockchip PLL code switches into slow mode (AKA bypass more AKA
24MHz mode) before actually changing the PLL.  This keeps anyone from
using the PLL while it's changing.  However, in all known Rockchip
SoCs nobody should ever see the 24MHz when changing the PLL supplying
the armclk because we should reparent children to an alternate
(faster than 24MHz) PLL.

One problem is that the code to switch to an alternate parent was
running in PRE_RATE_CHANGE.  ...and the code to switch to slow mode
was _also_ running in PRE_RATE_CHANGE.  That meant there was no real
guarantee that we would switch to an alternate parent before switching
to 24MHz mode.

Let's move the switch to "slow mode" straight into
rockchip_rk3066_pll_set_rate().  That means we're guaranteed that the
24MHz is really a last-resort.

Note that without this change on real systems we were the code to
switch to an alternate parent at 24MHz.  In some older versions of
that code we'd appy a (temporary) / 5 to the 24MHz causing us to run
at 4.8MHz.  That wasn't enough to service USB interrupts in some cases
and could lead to a system hang.

Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
10 years agoMerge branch 'v3.18-next/armclkid' into v3.18-next/cpuclk
Heiko Stuebner [Sat, 27 Sep 2014 15:56:55 +0000 (17:56 +0200)]
Merge branch 'v3.18-next/armclkid' into v3.18-next/cpuclk

10 years agoclk: rockchip: add binding id for ARMCLK
Heiko Stuebner [Fri, 5 Sep 2014 09:28:12 +0000 (11:28 +0200)]
clk: rockchip: add binding id for ARMCLK

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
10 years agoasm-generic: COMMON_CLK defines __clk_{get,put}
Mike Turquette [Tue, 9 Sep 2014 06:11:26 +0000 (23:11 -0700)]
asm-generic: COMMON_CLK defines __clk_{get,put}

If CONFIG_COMMON_CLK is selected then __clk_get and __clk_put are
defined in drivers/clk/clk.c and declared in include/linux/clkdev.h.

Sylwester's series[0] to properly support clk_{get,put} in the common
clock framework made changes to the asm-specific clkdev.h headers, but
not the asm-generic version. Tomeu's recent changes[1] to introduce a
provider/consumer split in the clock framework uncovered this problem,
causing the following build error on any architecture using the
asm-generic clkdev.h (e.g. x86 architecture and the ACPI LPSS driver):

In file included from drivers/acpi/acpi_lpss.c:15:0:
include/linux/clkdev.h:59:5: error: conflicting types for ‘__clk_get’
 int __clk_get(struct clk_core *clk);
     ^
In file included from arch/x86/include/generated/asm/clkdev.h:1:0,
                 from include/linux/clkdev.h:15,
                 from drivers/acpi/acpi_lpss.c:15:
include/asm-generic/clkdev.h:20:19: note: previous definition of ‘__clk_get’ was here
 static inline int __clk_get(struct clk *clk) { return 1; }
                   ^

Fixed by only declarating  __clk_get and __clk_put when
CONFIG_COMMON_CLK is set.

[0] http://lkml.kernel.org/r/<1386177127-2894-5-git-send-email-s.nawrocki@samsung.com>
[1] http://lkml.kernel.org/r/<1409758148-20104-1-git-send-email-tomeu.vizoso@collabora.com>

Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: Remove .owner field for driver
Kiran Padwal [Wed, 24 Sep 2014 09:45:29 +0000 (15:15 +0530)]
clk: 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: Kiran Padwal <kiran.padwal@smartplayin.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoMerge branch 'clk-next-rockchip' into clk-next
Mike Turquette [Thu, 25 Sep 2014 22:48:04 +0000 (15:48 -0700)]
Merge branch 'clk-next-rockchip' into clk-next

10 years agoclk: rockchip: add clock node in PD_VIDEO
Kever Yang [Thu, 25 Sep 2014 07:48:47 +0000 (15:48 +0800)]
clk: rockchip: add clock node in PD_VIDEO

This patch add the clock node in PD_VIDEO

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: rockchip: use the clock id for nodes init
Kever Yang [Thu, 25 Sep 2014 07:48:46 +0000 (15:48 +0800)]
clk: rockchip: use the clock id for nodes init

This patch use the new defined clock ID to initial the clock nodes.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: rockchip: add some needed clock binding id for rk3288
Kever Yang [Wed, 24 Sep 2014 13:36:34 +0000 (21:36 +0800)]
clk: rockchip: add some needed clock binding id for rk3288

This patch add some clock binding id for different modules
that under development and going to send upstream.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: rockchip: add missing rk3288 npll rate table
Heiko Stübner [Wed, 24 Sep 2014 21:41:54 +0000 (23:41 +0200)]
clk: rockchip: add missing rk3288 npll rate table

The npll on rk3288 is exactly the same pll type as the other 4. Yet it
was missing the link to the rate table, making rate changes impossible.
Change that by setting the table.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: rockchip: rk3288: fix softreset register count
Mark yao [Fri, 12 Sep 2014 09:24:46 +0000 (17:24 +0800)]
clk: rockchip: rk3288: fix softreset register count

The rk3288 actually has 12 softresets, so fix the register count.

Signed-off-by: Mark yao <mark.yao@rock-chips.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: rockchip: rk3288: add reset indices for SOFTRST9-11
Mark yao [Fri, 12 Sep 2014 11:45:27 +0000 (19:45 +0800)]
clk: rockchip: rk3288: add reset indices for SOFTRST9-11

The patch add the rest of the indices of the additional reset
registers from the updated TRM.

Signed-off-by: Mark yao <mark.yao@rock-chips.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoMerge branch 'clk-fixes' into clk-next
Mike Turquette [Wed, 17 Sep 2014 18:47:56 +0000 (11:47 -0700)]
Merge branch 'clk-fixes' into clk-next

10 years agoMerge branch 'clk-next-debugfs-lock' into clk-next
Mike Turquette [Wed, 10 Sep 2014 21:42:37 +0000 (14:42 -0700)]
Merge branch 'clk-next-debugfs-lock' into clk-next

10 years agoclk: Don't hold prepare_lock across debugfs creation
Stephen Boyd [Fri, 5 Sep 2014 06:37:49 +0000 (23:37 -0700)]
clk: Don't hold prepare_lock across debugfs creation

Rob Clark reports a lockdep splat that involves the prepare_lock
chained with the mmap semaphore.

======================================================
[ INFO: possible circular locking dependency detected ]
3.17.0-rc1-00050-g07a489b #802 Tainted: G        W
-------------------------------------------------------
Xorg.bin/5413 is trying to acquire lock:
 (prepare_lock){+.+.+.}, at: [<c0781280>] clk_prepare_lock+0x88/0xfc

but task is already holding lock:
 (qcom_iommu_lock){+.+...}, at: [<c079f664>] qcom_iommu_unmap+0x1c/0x1f0

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #4 (qcom_iommu_lock){+.+...}:
       [<c079f860>] qcom_iommu_map+0x28/0x450
       [<c079eb50>] iommu_map+0xc8/0x12c
       [<c056c1fc>] msm_iommu_map+0xb4/0x130
       [<c05697bc>] msm_gem_get_iova_locked+0x9c/0xe8
       [<c0569854>] msm_gem_get_iova+0x4c/0x64
       [<c0562208>] mdp4_kms_init+0x4c4/0x6c0
       [<c056881c>] msm_load+0x2ac/0x34c
       [<c0545724>] drm_dev_register+0xac/0x108
       [<c0547510>] drm_platform_init+0x50/0xf0
       [<c0578a60>] try_to_bring_up_master.part.3+0xc8/0x108
       [<c0578b48>] component_master_add_with_match+0xa8/0x104
       [<c0568294>] msm_pdev_probe+0x64/0x70
       [<c057e704>] platform_drv_probe+0x2c/0x60
       [<c057cff8>] driver_probe_device+0x108/0x234
       [<c057b65c>] bus_for_each_drv+0x64/0x98
       [<c057cec0>] device_attach+0x78/0x8c
       [<c057c590>] bus_probe_device+0x88/0xac
       [<c057c9b8>] deferred_probe_work_func+0x68/0x9c
       [<c0259db4>] process_one_work+0x1a0/0x40c
       [<c025a710>] worker_thread+0x44/0x4d8
       [<c025ec54>] kthread+0xd8/0xec
       [<c020e9a8>] ret_from_fork+0x14/0x2c

-> #3 (&dev->struct_mutex){+.+.+.}:
       [<c0541188>] drm_gem_mmap+0x38/0xd0
       [<c05695b8>] msm_gem_mmap+0xc/0x5c
       [<c02f0b6c>] mmap_region+0x35c/0x6c8
       [<c02f11ec>] do_mmap_pgoff+0x314/0x398
       [<c02de1e0>] vm_mmap_pgoff+0x84/0xb4
       [<c02ef83c>] SyS_mmap_pgoff+0x94/0xbc
       [<c020e8e0>] ret_fast_syscall+0x0/0x48

-> #2 (&mm->mmap_sem){++++++}:
       [<c0321138>] filldir64+0x68/0x180
       [<c0333fe0>] dcache_readdir+0x188/0x22c
       [<c0320ed0>] iterate_dir+0x9c/0x11c
       [<c03213b0>] SyS_getdents64+0x78/0xe8
       [<c020e8e0>] ret_fast_syscall+0x0/0x48

-> #1 (&sb->s_type->i_mutex_key#3){+.+.+.}:
       [<c03fc544>] __create_file+0x58/0x1dc
       [<c03fc70c>] debugfs_create_dir+0x1c/0x24
       [<c0781c7c>] clk_debug_create_subtree+0x20/0x170
       [<c0be2af8>] clk_debug_init+0xec/0x14c
       [<c0208c70>] do_one_initcall+0x8c/0x1c8
       [<c0b9cce4>] kernel_init_freeable+0x13c/0x1dc
       [<c0877bc4>] kernel_init+0x8/0xe8
       [<c020e9a8>] ret_from_fork+0x14/0x2c

-> #0 (prepare_lock){+.+.+.}:
       [<c087c408>] mutex_lock_nested+0x70/0x3e8
       [<c0781280>] clk_prepare_lock+0x88/0xfc
       [<c0782c50>] clk_prepare+0xc/0x24
       [<c079f474>] __enable_clocks.isra.4+0x18/0xa4
       [<c079f614>] __flush_iotlb_va+0xe0/0x114
       [<c079f6f4>] qcom_iommu_unmap+0xac/0x1f0
       [<c079ea3c>] iommu_unmap+0x9c/0xe8
       [<c056c2fc>] msm_iommu_unmap+0x64/0x84
       [<c0569da4>] msm_gem_free_object+0x11c/0x338
       [<c05413ec>] drm_gem_object_handle_unreference_unlocked+0xfc/0x130
       [<c0541604>] drm_gem_object_release_handle+0x50/0x68
       [<c0447a98>] idr_for_each+0xa8/0xdc
       [<c0541c10>] drm_gem_release+0x1c/0x28
       [<c0540b3c>] drm_release+0x370/0x428
       [<c031105c>] __fput+0x98/0x1e8
       [<c025d73c>] task_work_run+0xb0/0xfc
       [<c02477ec>] do_exit+0x2ec/0x948
       [<c0247ec0>] do_group_exit+0x4c/0xb8
       [<c025180c>] get_signal+0x28c/0x6ac
       [<c0211204>] do_signal+0xc4/0x3e4
       [<c02116cc>] do_work_pending+0xb4/0xc4
       [<c020e938>] work_pending+0xc/0x20

other info that might help us debug this:

Chain exists of:
  prepare_lock --> &dev->struct_mutex --> qcom_iommu_lock

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(qcom_iommu_lock);
                               lock(&dev->struct_mutex);
                               lock(qcom_iommu_lock);
  lock(prepare_lock);

 *** DEADLOCK ***

3 locks held by Xorg.bin/5413:
 #0:  (drm_global_mutex){+.+.+.}, at: [<c0540800>] drm_release+0x34/0x428
 #1:  (&dev->struct_mutex){+.+.+.}, at: [<c05413bc>] drm_gem_object_handle_unreference_unlocked+0xcc/0x130
 #2:  (qcom_iommu_lock){+.+...}, at: [<c079f664>] qcom_iommu_unmap+0x1c/0x1f0

stack backtrace:
CPU: 1 PID: 5413 Comm: Xorg.bin Tainted: G        W      3.17.0-rc1-00050-g07a489b #802
[<c0216290>] (unwind_backtrace) from [<c0211d8c>] (show_stack+0x10/0x14)
[<c0211d8c>] (show_stack) from [<c087a078>] (dump_stack+0x98/0xb8)
[<c087a078>] (dump_stack) from [<c027f024>] (print_circular_bug+0x218/0x340)
[<c027f024>] (print_circular_bug) from [<c0283e08>] (__lock_acquire+0x1d24/0x20b8)
[<c0283e08>] (__lock_acquire) from [<c0284774>] (lock_acquire+0x9c/0xbc)
[<c0284774>] (lock_acquire) from [<c087c408>] (mutex_lock_nested+0x70/0x3e8)
[<c087c408>] (mutex_lock_nested) from [<c0781280>] (clk_prepare_lock+0x88/0xfc)
[<c0781280>] (clk_prepare_lock) from [<c0782c50>] (clk_prepare+0xc/0x24)
[<c0782c50>] (clk_prepare) from [<c079f474>] (__enable_clocks.isra.4+0x18/0xa4)
[<c079f474>] (__enable_clocks.isra.4) from [<c079f614>] (__flush_iotlb_va+0xe0/0x114)
[<c079f614>] (__flush_iotlb_va) from [<c079f6f4>] (qcom_iommu_unmap+0xac/0x1f0)
[<c079f6f4>] (qcom_iommu_unmap) from [<c079ea3c>] (iommu_unmap+0x9c/0xe8)
[<c079ea3c>] (iommu_unmap) from [<c056c2fc>] (msm_iommu_unmap+0x64/0x84)
[<c056c2fc>] (msm_iommu_unmap) from [<c0569da4>] (msm_gem_free_object+0x11c/0x338)
[<c0569da4>] (msm_gem_free_object) from [<c05413ec>] (drm_gem_object_handle_unreference_unlocked+0xfc/0x130)
[<c05413ec>] (drm_gem_object_handle_unreference_unlocked) from [<c0541604>] (drm_gem_object_release_handle+0x50/0x68)
[<c0541604>] (drm_gem_object_release_handle) from [<c0447a98>] (idr_for_each+0xa8/0xdc)
[<c0447a98>] (idr_for_each) from [<c0541c10>] (drm_gem_release+0x1c/0x28)
[<c0541c10>] (drm_gem_release) from [<c0540b3c>] (drm_release+0x370/0x428)
[<c0540b3c>] (drm_release) from [<c031105c>] (__fput+0x98/0x1e8)
[<c031105c>] (__fput) from [<c025d73c>] (task_work_run+0xb0/0xfc)
[<c025d73c>] (task_work_run) from [<c02477ec>] (do_exit+0x2ec/0x948)
[<c02477ec>] (do_exit) from [<c0247ec0>] (do_group_exit+0x4c/0xb8)
[<c0247ec0>] (do_group_exit) from [<c025180c>] (get_signal+0x28c/0x6ac)
[<c025180c>] (get_signal) from [<c0211204>] (do_signal+0xc4/0x3e4)
[<c0211204>] (do_signal) from [<c02116cc>] (do_work_pending+0xb4/0xc4)
[<c02116cc>] (do_work_pending) from [<c020e938>] (work_pending+0xc/0x20)

We can break this chain if we don't hold the prepare_lock while
creating debugfs directories. We only hold the prepare_lock right
now because we're traversing the clock tree recursively and we
don't want the hierarchy to change during the traversal.
Replacing this traversal with a simple linked list walk allows us
to only grab a list lock instead of the prepare_lock, thus
breaking the lock chain.

Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: rockchip: also protect hclk_peri as critical
Heiko Stübner [Wed, 10 Sep 2014 15:52:02 +0000 (17:52 +0200)]
clk: rockchip: also protect hclk_peri as critical

The dwc2 usb controller also uses agressive clock gating, which in this
case leads to hclk_peri getting disabled and hanging the system.
Therefore move it to the critical clocks until we also control that
part of the system.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: fractional-divider: cast parent_rate to u64 before multiplying
Heiko Stübner [Thu, 28 Aug 2014 10:46:10 +0000 (12:46 +0200)]
clk: fractional-divider: cast parent_rate to u64 before multiplying

On 32bit architectures, like ARM calculating the fractional rate will
do the multiplication before converting the value to u64 when it gets
assigned to ret, which can produce overflows.

The error in question happened with a parent_rate of 386MHz, m = 3000,
n = 60000, which resulted in a wrong rate value of 15812Hz.

Therefore cast parent_rate to u64 to make sure the multiplication
happens in a 64bit space and produces the correct 192MHz in the example.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: max77802: Add DT binding documentation
Javier Martinez Canillas [Mon, 18 Aug 2014 08:33:04 +0000 (10:33 +0200)]
clk: max77802: Add DT binding documentation

Add Device Tree binding documentation for the clocks
outputs in the Maxim 77802 Power Management IC.

Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: Add driver for Maxim 77802 PMIC clocks
Javier Martinez Canillas [Mon, 18 Aug 2014 08:33:03 +0000 (10:33 +0200)]
clk: Add driver for Maxim 77802 PMIC clocks

The MAX77802 PMIC has two 32.768kHz Buffered Clock Outputs with
Low Jitter Mode. This patch adds support for these two clocks.

Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: max77686: Improve Maxim 77686 PMIC clocks binding
Javier Martinez Canillas [Mon, 18 Aug 2014 08:33:02 +0000 (10:33 +0200)]
clk: max77686: Improve Maxim 77686 PMIC clocks binding

Like most clock drivers, the Maxim 77686 PMIC clock binding
follows the convention that the "#clock-cells" property is
used to specify the number of cells in a clock provider.

But the binding document is not clear enough that it shall
be set to 1 since the PMIC support multiple clocks outputs.

Also, explain that the clocks identifiers are defined in a
header file that can be included by Device Tree source with
client nodes to avoid using magic numbers.

Finally, add "clock-output-names" as an optional property
since now is supported by the clock driver.

Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Reviewed-by: Mike Turquette <mturquette@linaro.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: max77686: Convert to the generic max clock driver
Javier Martinez Canillas [Mon, 18 Aug 2014 08:33:01 +0000 (10:33 +0200)]
clk: max77686: Convert to the generic max clock driver

Clocks drivers for Maxim PMIC are very similar so they can
be converted to use the generic Maxim clock driver.

Also, while being there use module_platform_driver() helper
macro to eliminate more boilerplate code.

Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: Add generic driver for Maxim PMIC clocks
Javier Martinez Canillas [Mon, 18 Aug 2014 08:33:00 +0000 (10:33 +0200)]
clk: Add generic driver for Maxim PMIC clocks

Maxim Integrated Power Management ICs are very similar with
regard to their clock outputs. Most of the clock drivers for
these chips are duplicating code and are simpler enough that
can be converted to use a generic driver to consolidate code
and avoid duplication.

Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: max77686: Add DT include for MAX77686 PMIC clock
Javier Martinez Canillas [Mon, 18 Aug 2014 08:32:59 +0000 (10:32 +0200)]
clk: max77686: Add DT include for MAX77686 PMIC clock

This patch adds a dt-binding include for Maxim 77686
PMIC clock IDs that can be used by both the max77686
clock driver and Device Tree source files.

Signed-off-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Mike Turquette <mturquette@linaro.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk/efm32gg: fix dt init prototype
Uwe Kleine-König [Mon, 28 Jul 2014 14:38:21 +0000 (16:38 +0200)]
clk/efm32gg: fix dt init prototype

Since commit 54196ccbe0ba (of: consolidate linker section OF match table
declarations) which went into 3.16-rc1 the following compiler warning is
generated:

In file included from drivers/clk/clk-efm32gg.c:12:0: include/linux/of.h:772:20:
warning: comparison of distinct pointer types lacks a cast [enabled by default]
.data = (fn == (fn_type)NULL) ? fn : fn  }
    ^
include/linux/of.h:785:3: note: in expansion of macro '_OF_DECLARE'
   _OF_DECLARE(table, name, compat, fn, of_init_fn_1)
   ^
include/linux/clk-provider.h:545:42: note: in expansion of macro 'OF_DECLARE_1'
 #define CLK_OF_DECLARE(name, compat, fn) OF_DECLARE_1(clk, name, compat, fn)
  ^
drivers/clk/clk-efm32gg.c:81:1: note: in expansion of macro 'CLK_OF_DECLARE'
 CLK_OF_DECLARE(efm32ggcmu, "efm32gg,cmu", efm32gg_cmu_init);
 ^

Fix it by making efm32gg_cmu_init return void.

Cc: Rob Herring <robh@kernel.org>
Reported-by: Bryan Hundven <bryanhundven@gmail.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: zynq: Move const initdata into correct code section
Soren Brinkmann [Tue, 2 Sep 2014 23:02:09 +0000 (16:02 -0700)]
clk: zynq: Move const initdata into correct code section

Use __initconst instead of __initdata for constant init data.

Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: zynq: Remove pointless return at end of void function
Soren Brinkmann [Tue, 2 Sep 2014 23:02:08 +0000 (16:02 -0700)]
clk: zynq: Remove pointless return at end of void function

Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: zynq: Remove unnecessary OOM message
Soren Brinkmann [Tue, 2 Sep 2014 23:02:07 +0000 (16:02 -0700)]
clk: zynq: Remove unnecessary OOM message

As checkpatch suggests:
  WARNING: Possible unnecessary 'out of memory' message,
remove an error message after failing kmalloc() from the PLL driver.

Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: prevent erronous parsing of children during rate change
Tero Kristo [Thu, 21 Aug 2014 13:47:45 +0000 (16:47 +0300)]
clk: prevent erronous parsing of children during rate change

In some cases, clocks can switch their parent with clk_set_rate, for
example clk_mux can do this in some cases. Current implementation of
clk_change_rate uses un-safe list iteration on the clock children, which
will cause wrong clocks to be parsed in case any of the clock children
change their parents during the change rate operation. Fixed by using
the safe list iterator instead.

The problem was detected due to some divide by zero errors generated
by clock init on dra7-evm board, see discussion under
http://article.gmane.org/gmane.linux.ports.arm.kernel/349180 for details.

Fixes: 71472c0c06cf ("clk: add support for clock reparent on set_rate")
Signed-off-by: Tero Kristo <t-kristo@ti.com>
Reported-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: rockchip: Fix the clocks for i2c1 and i2c2
Doug Anderson [Fri, 22 Aug 2014 16:49:29 +0000 (09:49 -0700)]
clk: rockchip: Fix the clocks for i2c1 and i2c2

The clocks for i2c1 and i2c2 are flipped.  The clock tree matched the
Technical Reference Manual (TRM) but the TRM was wrong.  Swap them in
the clock tree.  This was determined experimentally (by Addy) and
confirmed by the Rockchip IC team.

Signed-off-by: Doug Anderson <dianders@chromium.org>
Reported-by: Addy Ke <addy.ke@rock-chips.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoMerge branch 'for-v3.17-rc/ti-clk-driver' of github.com:t-kristo/linux-pm into clk...
Mike Turquette [Wed, 3 Sep 2014 17:07:14 +0000 (10:07 -0700)]
Merge branch 'for-v3.17-rc/ti-clk-driver' of github.com:t-kristo/linux-pm into clk-fixes

10 years agoclk: qcom: Fix sdc 144kHz frequency entry
Stephen Boyd [Fri, 29 Aug 2014 19:49:26 +0000 (12:49 -0700)]
clk: qcom: Fix sdc 144kHz frequency entry

The pre-divider for the sdc clocks only has 2 bits in it, so we
can't possibly divide by anything larger than 4 here.
Furthermore, we program the value of ~(n - m) and the n value is
larger than 8 bits (max of 256). Replace this entry with 200kHz
which is close enough to 144kHz to be usable.

Cc: Kumar Gala <galak@codeaurora.org>
Cc: Andy Gross <agross@codeaurora.org>
Fixes: 24d8fba44af3 "clk: qcom: Add support for IPQ8064's global clock controller (GCC)"
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: at91: fix div by zero in USB clock driver
Boris BREZILLON [Tue, 2 Sep 2014 07:50:18 +0000 (09:50 +0200)]
clk: at91: fix div by zero in USB clock driver

Test rate value before calculating the div value to avoid div by zero.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Reported-by: Gaël PORTAY <gael.portay@gmail.com>
Tested-by: Gaël PORTAY <gael.portay@gmail.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: at91: rework rm9200 USB clock to propagate set_rate to the parent clk
Boris BREZILLON [Tue, 2 Sep 2014 07:50:17 +0000 (09:50 +0200)]
clk: at91: rework rm9200 USB clock to propagate set_rate to the parent clk

The RM9200 USB clock is actually connected to a single parent (the PLLB)
on which we can apply a specific divider.
The USB clock divider does not allow for fine grained control on the USB
clock frequency, hence propagating the set_rate request to the parent is
the only choice we have to properly configure the USB clock rate.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Reported-by: Gaël PORTAY <gael.portay@gmail.com>
Tested-by: Gaël PORTAY <gael.portay@gmail.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: at91: fix recalc_rate implementation of PLL driver
Boris BREZILLON [Tue, 2 Sep 2014 07:50:16 +0000 (09:50 +0200)]
clk: at91: fix recalc_rate implementation of PLL driver

Use the cached values to calculate PLL rate instead of the register values.
This is required to prevent erroneous PLL rate return when the PLL rate
has been configured but the PLL is not prepared yet.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Reported-by: Gaël PORTAY <gael.portay@gmail.com>
Tested-by: Gaël PORTAY <gael.portay@gmail.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: at91: rework PLL rate calculation
Boris BREZILLON [Tue, 2 Sep 2014 07:50:15 +0000 (09:50 +0200)]
clk: at91: rework PLL rate calculation

The AT91 PLL rate configuration is done by configuring a multiplier/divider
pair.
The previous calculation was over-complicated (and apparently buggy).
Simplify the implementation and add some comments to explain what is done
here.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Reported-by: Gaël PORTAY <gael.portay@gmail.com>
Tested-by: Gaël PORTAY <gael.portay@gmail.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: at91: fix PLL_MAX_COUNT macro definition
Boris BREZILLON [Tue, 2 Sep 2014 07:50:14 +0000 (09:50 +0200)]
clk: at91: fix PLL_MAX_COUNT macro definition

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Reported-by: Gaël PORTAY <gael.portay@gmail.com>
Tested-by: Gaël PORTAY <gael.portay@gmail.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: rockchip: protect critical clocks from getting disabled
Heiko Stübner [Thu, 14 Aug 2014 21:00:26 +0000 (23:00 +0200)]
clk: rockchip: protect critical clocks from getting disabled

The clock-tree contains clocks that should never get disabled automatically.
One example are the base ACLKs, the base supplies for all peripherals.

Therefore add a structure similar to the sunxi clock-tree to protect these
special clocks from being disabled.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Doug Anderson <dianders@chromium.org>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: rockchip: make rockchip_clk_register_branch static
Heiko Stübner [Tue, 26 Aug 2014 22:54:56 +0000 (00:54 +0200)]
clk: rockchip: make rockchip_clk_register_branch static

It is only used locally in clk/rockchip/clk.c and thus can be static.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: rockchip: implement the fraction divider branch type
Heiko Stübner [Tue, 26 Aug 2014 22:54:21 +0000 (00:54 +0200)]
clk: rockchip: implement the fraction divider branch type

Rockchip SoCs may provide fraction dividers for some clocks, mostly for
i2s and uarts. In contrast to the other registers, these do not use
the hiword-mask paradigm, but instead split the register into the upper
16 bit for the nominator and the lower 16 bit for the denominator.

The common clock framework got a generic fractional divider clock type
recently that can accomodate this setting easily. All currently known
fraction dividers have a separate gate too, therefore implement the
divider as composite using the ops-struct from fractional_divider clock
and add the gate if necessary.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: shmobile: add missing 0x0100 for SDCKCR
Kuninori Morimoto [Wed, 6 Aug 2014 01:24:00 +0000 (10:24 +0900)]
clk: shmobile: add missing 0x0100 for SDCKCR

10cdfe9f327ab8d120cf6957e58c6203e3a53847
(clk: shmobile: Add R-Car Gen2 clocks support)
added R-Car Gen2 clock support via DT for v3.14.

But, cpg_sd01_div_table is missing
"0x0100: x 1/8" division ratio.
This patch fixes hidden bug.
It is based on R-Car H2 v0.7, R-Car M2 v0.9.

Acked-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Simon Horman <horms+renesas@verge.net.au>
Reported-by: Yusuke Goda <yusuke.goda.sx@renesas.com>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoARM: shmobile: r8a7794: document CPG clock support
Ulrich Hecht [Fri, 29 Aug 2014 18:15:10 +0000 (20:15 +0200)]
ARM: shmobile: r8a7794: document CPG clock support

Signed-off-by: Ulrich Hecht <ulrich.hecht+renesas@gmail.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoARM: shmobile: r8a7794: document MSTP clock support
Ulrich Hecht [Fri, 29 Aug 2014 18:15:08 +0000 (20:15 +0200)]
ARM: shmobile: r8a7794: document MSTP clock support

Signed-off-by: Ulrich Hecht <ulrich.hecht+renesas@gmail.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agocpufreq: kirkwood: use the powersave multiplexer
Mike Turquette [Mon, 18 Aug 2014 15:30:29 +0000 (17:30 +0200)]
cpufreq: kirkwood: use the powersave multiplexer

The powersave clock acts like a multiplexer for the cpu, selecting
either the clock signal derived from the cpu pll or from the ddr clock.
This patch changes powersave from a gate clock to a mux clock to better
reflect this behavior.

This is a cleaner approach whereby the frequency of the cpu always
matches the rate of powersave_clk. The cpufreq driver for the kirkwood
platform no longer must parse this behavior out of various calls to
clk_enable and clk_disable, but can instead simply select the parent cpu
it wants when changing rate. Likewise when requesting the cpu rate we
need only query powersave_clk's rate through the usual call to
clk_get_rate.

The new clock data and corresponding changes to the cpufreq driver are
combined into this single commit to avoid a git bisect issue where this
cpufreq driver fails to work properly between the commit that updates
the kirkwood clock driver and the commit that changes how the cpufreq
driver uses that clock.

Cc: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Cc: Rafael J. Wysocki <rjw@rjwysocki.net>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: at91: fix num_parents test in at91sam9260 slow clk implementation
Boris BREZILLON [Tue, 2 Sep 2014 15:27:51 +0000 (17:27 +0200)]
clk: at91: fix num_parents test in at91sam9260 slow clk implementation

The slow clk block provided by at91sam9260 and derived SoCs should always
have 2 parents.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: mvebu: powersave clock is a multiplexer
Mike Turquette [Thu, 28 Aug 2014 00:11:38 +0000 (17:11 -0700)]
clk: mvebu: powersave clock is a multiplexer

Kirkwood is unique among the mvebu SoCs for having a clock multiplexer
that feeds into the cpu. This multiplexer can select either the cpu pll
or the ddr clock as its input signal, allowing for a choice between
performance and power savings.

This patch introduces the code needed to register the clock multiplexer
on Kirkwood SoCs but does not include the clock data to actually
register the clock. That will be done in a follow-up patch which is
necessary to prevent breaking git bisect.

Cc: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoclk: mvebu: share locks between gate clocks
Mike Turquette [Wed, 27 Aug 2014 22:36:37 +0000 (15:36 -0700)]
clk: mvebu: share locks between gate clocks

Refactor mvebu_clk_gating_setup() to use a common spinlock instead of a
unique lock for every instance of a struct clk_gating_ctrl object. This
will be used later for a separate mux clock type that shares a register
with gate clock types and needs to use the same lock to protect access
to the register.

Cc: Andrew Lunn <andrew@lunn.ch>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Mike Turquette <mturquette@linaro.org>
10 years agoLinux 3.17-rc3
Linus Torvalds [Mon, 1 Sep 2014 01:23:04 +0000 (18:23 -0700)]
Linux 3.17-rc3

10 years agoMerge tag 'xtensa-20140830' of git://github.com/czankel/xtensa-linux
Linus Torvalds [Mon, 1 Sep 2014 00:08:42 +0000 (17:08 -0700)]
Merge tag 'xtensa-20140830' of git://github.com/czankel/xtensa-linux

Pull Xtensa updates from Chris Zankel:
 "Xtensa improvements for 3.17:
   - support highmem on cores with aliasing data cache.  Enable highmem
     on kc705 by default
   - simplify addition of new core variants (no need to modify Kconfig /
     Makefiles)
   - improve robustness of unaligned access handler and its interaction
     with window overflow/underflow exception handlers
   - deprecate atomic and spill registers syscalls
   - clean up Kconfig: remove orphan MATH_EMULATION, sort 'select'
     statements
   - wire up renameat2 syscall.

  Various fixes:
   - fix address checks in dma_{alloc,free}_coherent (runtime BUG)
   - fix access to THREAD_RA/THREAD_SP/THREAD_DS (debug build breakage)
   - fix TLBTEMP_BASE_2 region handling in fast_second_level_miss
     (runtime unrecoverable exception)
   - fix a6 and a7 handling in fast_syscall_xtensa (runtime userspace
     register clobbering)
   - fix kernel/user jump out of fast_unaligned (potential runtime
     unrecoverabl exception)
   - replace termios IOCTL code definitions with constants (userspace
     build breakage)"

* tag 'xtensa-20140830' of git://github.com/czankel/xtensa-linux: (25 commits)
  xtensa: deprecate fast_xtensa and fast_spill_registers syscalls
  xtensa: don't allow overflow/underflow on unaligned stack
  xtensa: fix a6 and a7 handling in fast_syscall_xtensa
  xtensa: allow single-stepping through unaligned load/store
  xtensa: move invalid unaligned instruction handler closer to its users
  xtensa: make fast_unaligned store restartable
  xtensa: add double exception fixup handler for fast_unaligned
  xtensa: fix kernel/user jump out of fast_unaligned
  xtensa: configure kc705 for highmem
  xtensa: support highmem in aliasing cache flushing code
  xtensa: support aliasing cache in kmap
  xtensa: support aliasing cache in k[un]map_atomic
  xtensa: implement clear_user_highpage and copy_user_highpage
  xtensa: fix TLBTEMP_BASE_2 region handling in fast_second_level_miss
  xtensa: allow fixmap and kmap span more than one page table
  xtensa: make fixmap region addressing grow with index
  xtensa: fix access to THREAD_RA/THREAD_SP/THREAD_DS
  xtensa: add renameat2 syscall
  xtensa: fix address checks in dma_{alloc,free}_coherent
  xtensa: replace IOCTL code definitions with constants
  ...

10 years agounicore32: Fix build error
Guenter Roeck [Sun, 31 Aug 2014 18:14:26 +0000 (11:14 -0700)]
unicore32: Fix build error

unicore32 builds fail with

  arch/unicore32/kernel/signal.c: In function ‘setup_frame’:
  arch/unicore32/kernel/signal.c:257: error: ‘usig’ undeclared (first use in this function)
  arch/unicore32/kernel/signal.c:279: error: ‘usig’ undeclared (first use in this function)
  arch/unicore32/kernel/signal.c: In function ‘handle_signal’:
  arch/unicore32/kernel/signal.c:306: warning: unused variable ‘tsk’
  arch/unicore32/kernel/signal.c: In function ‘do_signal’:
  arch/unicore32/kernel/signal.c:376: error: implicit declaration of function ‘get_signsl’
  make[1]: *** [arch/unicore32/kernel/signal.o] Error 1
  make: *** [arch/unicore32/kernel/signal.o] Error 2

Bisect points to commit 649671c90eaf ("unicore32: Use get_signal()
signal_setup_done()").

This code never even compiled.  Reverting the patch does not work, since
previously used functions no longer exist, so try to fix it up.  Compile
tested only.

Fixes: 649671c90eaf ("unicore32: Use get_signal() signal_setup_done()")
Cc: Richard Weinberger <richard@nod.at>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoMerge branch 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm
Linus Torvalds [Mon, 1 Sep 2014 00:02:57 +0000 (17:02 -0700)]
Merge branch 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm

Pull ARM fixes from Russell King:
 "Various assorted fixes:

   - a couple of patches from Mark Rutland to resolve an errata with
     Cortex-A15 CPUs.
   - fix cpuidle for the CPU part ID changes in the last merge window
   - add support for a relocation which ARM binutils is generating in
     some circumstances"

* 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
  ARM: 8130/1: cpuidle/cpuidle-big_little: fix reading cpu id part number
  ARM: 8129/1: errata: work around Cortex-A15 erratum 830321 using dummy strex
  ARM: 8128/1: abort: don't clear the exclusive monitors
  ARM: 8127/1: module: add support for R_ARM_TARGET1 relocations

10 years agoMerge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm...
Linus Torvalds [Mon, 1 Sep 2014 00:01:19 +0000 (17:01 -0700)]
Merge tag 'fixes-for-linus' of git://git./linux/kernel/git/arm/arm-soc

Pull ARM SoC fixes from Olof Johansson:
 "Here's the weekly batch of fixes from arm-soc.

  The delta is a largeish negative delta, due to revert of SMP support
  for Broadcom's STB SoC -- it was accidentally merged before some
  issues had been addressed, so they will make a new attempt for 3.18.
  I didn't see a need for a full revert of the whole platform due to
  this, we're keeping the rest enabled.

  The rest is mostly:

   - a handful of DT fixes for i.MX (Hummingboard/Cubox-i in particular)
   - some MTD/NAND fixes for OMAP
   - minor DT fixes for shmobile
   - warning fix for UP builds on vexpress/spc

  There's also a couple of patches that wires up hwmod on TI's DRA7 SoC
  so it can boot.  Drivers and the rest had landed for 3.17, and it's
  small and isolated so it made sense to pick up now even if it's not a
  bugfix"

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (23 commits)
  vexpress/spc: fix a build warning on array bounds
  ARM: DRA7: hwmod: Add dra74x and dra72x specific ocp interface lists
  ARM: DRA7: Add support for soc_is_dra74x() and soc_is_dra72x() variants
  MAINTAINERS: catch special Rockchip code locations
  ARM: dts: microsom-ar8035: MDIO pad must be set open drain
  ARM: dts: omap54xx-clocks: Fix the l3 and l4 clock rates
  ARM: brcmstb: revert SMP support
  ARM: OMAP2+: hwmod: Rearm wake-up interrupts for DT when MUSB is idled
  ARM: dts: Enable UART wake-up events for beagleboard
  ARM: dts: Remove twl6030 clk32g "regulator"
  ARM: OMAP2+: omap_device: remove warning that clk alias already exists
  ARM: OMAP: fix %d confusingly prefixed with 0x in format string
  ARM: dts: DRA7: fix interrupt-cells for GPIO
  mtd: nand: omap: Fix 1-bit Hamming code scheme, omap_calculate_ecc()
  ARM: dts: omap3430-sdp: Revert to using software ECC for NAND
  ARM: OMAP2+: GPMC: Support Software ECC scheme via DT
  mtd: nand: omap: Revert to using software ECC by default
  ARM: dts: hummingboard/cubox-i: change SPDIF output to be more descriptive
  ARM: dts: hummingboard/cubox-i: add USB OC pinctrl configuration
  ARM: shmobile: r8a7791: add missing 0x0100 for SDCKCR
  ...

10 years agovexpress/spc: fix a build warning on array bounds
Alex Shi [Wed, 16 Jul 2014 11:21:56 +0000 (19:21 +0800)]
vexpress/spc: fix a build warning on array bounds

With ARCH_VEXPRESS_SPC option, kernel build has the following
warning:

arch/arm/mach-vexpress/spc.c: In function ‘ve_spc_clk_init’:
arch/arm/mach-vexpress/spc.c:431:38: warning: array subscript is below array bounds [-Warray-bounds]
  struct ve_spc_opp *opps = info->opps[cluster];
                                      ^
since 'cluster' maybe '-1' in UP system. This patch does a active
checking to fix this issue.

Signed-off-by: Alex Shi <alex.shi@linaro.org>
Acked-by: Pawel Moll <pawel.moll@arm.com>
Acked-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Olof Johansson <olof@lixom.net>
10 years agoMerge tag 'for-v3.17-rc/omap-dra72x-d74x-support-a' of git://git.kernel.org/pub/scm...
Olof Johansson [Sun, 31 Aug 2014 17:19:09 +0000 (10:19 -0700)]
Merge tag 'for-v3.17-rc/omap-dra72x-d74x-support-a' of git://git./linux/kernel/git/pjw/omap-pending into fixes

Pull "ARM: OMAP2+: DRA72x/DRA74x basic support" from Tony Lindgren:

Add basic subarchitecture support for the DRA72x and DRA74x.  These
are OMAP2+ derivative SoCs.  This should be low-risk to existing OMAP
platforms.

Basic build, boot, and PM test logs are available here:

http://www.pwsan.com/omap/testlogs/hwmod-a-early-v3.17-rc/20140827194314/

* tag 'for-v3.17-rc/omap-dra72x-d74x-support-a' of git://git.kernel.org/pub/scm/linux/kernel/git/pjw/omap-pending:
  ARM: DRA7: hwmod: Add dra74x and dra72x specific ocp interface lists
  ARM: DRA7: Add support for soc_is_dra74x() and soc_is_dra72x() variants

Signed-off-by: Olof Johansson <olof@lixom.net>
10 years agoMerge tag 'spi-v3.17-rc3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Linus Torvalds [Sun, 31 Aug 2014 17:09:07 +0000 (10:09 -0700)]
Merge tag 'spi-v3.17-rc3-2' of git://git./linux/kernel/git/broonie/spi

Pull spi bugfixes from Mark Brown:
 "A smattering of bug fixes for the SPI subsystem, all in driver code
  which has seen active work recently and none of them with any great
  global impact.

  There's also a new ACPI ID for the pxa2xx driver which required no
  code changes and the addition of kerneldoc for some structure fields
  that were missing it and generating warnings during documentation
  builds as a result"

* tag 'spi-v3.17-rc3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
  spi: sh-msiof: Fix transmit-only DMA transfers
  spi/rockchip: Avoid accidentally turning off the clock
  spi: dw: fix kernel crash due to NULL pointer dereference
  spi: dw-pci: fix bug when regs left uninitialized
  spi: davinci: fix SPI_NO_CS functionality
  spi/rockchip: fixup incorrect dma direction setting
  spi/pxa2xx: Add ACPI ID for Intel Braswell
  spi: spi-au1550: fix build failure
  spi: rspi: Fix leaking of unused DMA descriptors
  spi: sh-msiof: Fix leaking of unused DMA descriptors
  spi: Add missing kerneldoc bits
  spi/omap-mcspi: Fix the spi task hangs waiting dma_rx

10 years agoMerge remote-tracking branch 'spi/fix/sh-msiof' into spi-linus
Mark Brown [Sun, 31 Aug 2014 12:46:19 +0000 (13:46 +0100)]
Merge remote-tracking branch 'spi/fix/sh-msiof' into spi-linus

10 years agoMerge tag 'spi-v3.17-rc3' into spi-linus
Mark Brown [Sun, 31 Aug 2014 12:46:19 +0000 (13:46 +0100)]
Merge tag 'spi-v3.17-rc3' into spi-linus

spi: Bug fixes for v3.17

A smattering of bug fixes for the SPI subsystem, all in driver code
which has seen active work recently and none of them with any great
global impact.

There's also a new ACPI ID for the pxa2xx driver which required no code
changes and the addition of kerneldoc for some structure fields that
were missing it and generating warnings during documentation builds as a
result.

# gpg: Signature made Sun 31 Aug 2014 13:19:12 BST using RSA key ID 7EA229BD
# gpg: Good signature from "Mark Brown <broonie@sirena.org.uk>"
# gpg:                 aka "Mark Brown <broonie@debian.org>"
# gpg:                 aka "Mark Brown <broonie@kernel.org>"
# gpg:                 aka "Mark Brown <broonie@tardis.ed.ac.uk>"
# gpg:                 aka "Mark Brown <broonie@linaro.org>"
# gpg:                 aka "Mark Brown <Mark.Brown@linaro.org>"

10 years agospi: sh-msiof: Fix transmit-only DMA transfers
Geert Uytterhoeven [Thu, 7 Aug 2014 12:07:43 +0000 (14:07 +0200)]
spi: sh-msiof: Fix transmit-only DMA transfers

Fix tx/rx mixup, which broke transmit-only transfers.

Introduced by commit 4240305f7cbdc7782aa8bc40cc702775d9ac0839
("spi: sh-msiof: Fix leaking of unused DMA descriptors").

Reported-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
10 years agoMerge tag 'locks-v3.17-3' of git://git.samba.org/jlayton/linux
Linus Torvalds [Sun, 31 Aug 2014 04:04:37 +0000 (21:04 -0700)]
Merge tag 'locks-v3.17-3' of git://git.samba.org/jlayton/linux

Pull file locking bugfx from Jeff Layton:
 "Just a bugfix for a bug that crept in to v3.15.  It's in a rather rare
  error path, and I'm not aware of anyone having hit it, but it's worth
  fixing for v3.17"

* tag 'locks-v3.17-3' of git://git.samba.org/jlayton/linux:
  locks: pass correct "before" pointer to locks_unlink_lock in generic_add_lease

10 years agoMerge remote-tracking branches 'spi/fix/au1550', 'spi/fix/davinci', 'spi/fix/doc...
Mark Brown [Sat, 30 Aug 2014 18:28:04 +0000 (19:28 +0100)]
Merge remote-tracking branches 'spi/fix/au1550', 'spi/fix/davinci', 'spi/fix/doc', 'spi/fix/dw', 'spi/fix/omap-mcspi', 'spi/fix/pxa2xx', 'spi/fix/rockchip' and 'spi/fix/rspi' into spi-linus

10 years agoMerge remote-tracking branch 'spi/fix/sh-msiof' into spi-linus
Mark Brown [Sat, 30 Aug 2014 18:28:03 +0000 (19:28 +0100)]
Merge remote-tracking branch 'spi/fix/sh-msiof' into spi-linus

10 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 30 Aug 2014 00:22:27 +0000 (17:22 -0700)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Peter Anvin:
 "One patch to avoid assigning interrupts we don't actually have on
  non-PC platforms, and two patches that addresses bugs in the new
  IOAPIC assignment code"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86, irq, PCI: Keep IRQ assignment for runtime power management
  x86: irq: Fix bug in setting IOAPIC pin attributes
  x86: Fix non-PC platform kernel crash on boot due to NULL dereference

10 years agoMerge tag 'pm+acpi-3.17-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Sat, 30 Aug 2014 00:18:48 +0000 (17:18 -0700)]
Merge tag 'pm+acpi-3.17-rc3' of git://git./linux/kernel/git/rafael/linux-pm

Pull ACPI and power management fixes from Rafael Wysocki:

 - Fix for an ACPI regression related to the handling of fixed events
   that caused netlink routines to be (incorrectly) run in interrupt
   context from Lan Tianyu

 - Fix for an ACPI EC driver regression on Acer Aspire V5-573G that
   caused AC/battery plug/unplug and video brightness change
   notifications to be delayed on that machine from Lv Zheng

 - Fix for an ACPI device enumeration regression that caused ACPI driver
   probe to fail for some devices where it succeeded before (Rafael J
   Wysocki)

 - intel_pstate driver fix to prevent it from printing an information
   message for every CPU in the system on every boot from Andi Kleen

 - s5pv210 cpufreq driver fix to remove an __init annotation from a
   routine that in fact can be called at any time after init too from
   Mark Brown

 - New Intel Braswell device ID for the ACPI LPSS (Low-Power Subsystem)
   driver from Alan Cox

 - New Intel Braswell CPU ID for intel_pstate from Mika Westerberg

* tag 'pm+acpi-3.17-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  cpufreq: s5pv210: Remove spurious __init annotation
  cpufreq: intel_pstate: Add CPU ID for Braswell processor
  intel_pstate: Turn per cpu printk into pr_debug
  ACPI / LPSS: Add ACPI IDs for Intel Braswell
  ACPI / EC: Add support to disallow QR_EC to be issued before completing previous QR_EC
  ACPI / EC: Add support to disallow QR_EC to be issued when SCI_EVT isn't set
  ACPI: Run fixed event device notifications in process context
  ACPI / scan: Allow ACPI drivers to bind to PNP device objects

10 years agoMerge branch 'akpm' (fixes from Andrew Morton)
Linus Torvalds [Fri, 29 Aug 2014 23:28:29 +0000 (16:28 -0700)]
Merge branch 'akpm' (fixes from Andrew Morton)

Merge patches from Andrew Morton:
 "22 fixes"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (22 commits)
  kexec: purgatory: add clean-up for purgatory directory
  Documentation/kdump/kdump.txt: add ARM description
  flush_icache_range: export symbol to fix build errors
  tools: selftests: fix build issue with make kselftests target
  ocfs2: quorum: add a log for node not fenced
  ocfs2: o2net: set tcp user timeout to max value
  ocfs2: o2net: don't shutdown connection when idle timeout
  ocfs2: do not write error flag to user structure we cannot copy from/to
  x86/purgatory: use approprate -m64/-32 build flag for arch/x86/purgatory
  drivers/rtc/rtc-s5m.c: re-add support for devices without irq specified
  xattr: fix check for simultaneous glibc header inclusion
  kexec: remove CONFIG_KEXEC dependency on crypto
  kexec: create a new config option CONFIG_KEXEC_FILE for new syscall
  x86,mm: fix pte_special versus pte_numa
  hugetlb_cgroup: use lockdep_assert_held rather than spin_is_locked
  mm/zpool: use prefixed module loading
  zram: fix incorrect stat with failed_reads
  lib: turn CONFIG_STACKTRACE into an actual option.
  mm: actually clear pmd_numa before invalidating
  memblock, memhotplug: fix wrong type in memblock_find_in_range_node().
  ...

10 years agokexec: purgatory: add clean-up for purgatory directory
Michael Welling [Fri, 29 Aug 2014 22:19:13 +0000 (15:19 -0700)]
kexec: purgatory: add clean-up for purgatory directory

Without this patch the kexec-purgatory.c and purgatory.ro files are not
removed after make mrproper.

Signed-off-by: Michael Welling <mwelling@ieee.org>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoDocumentation/kdump/kdump.txt: add ARM description
HuKeping [Fri, 29 Aug 2014 22:19:11 +0000 (15:19 -0700)]
Documentation/kdump/kdump.txt: add ARM description

Add arm specific parts to kdump kernel documentation.

Signed-off-by: Hu Keping <hukeping@huawei.com>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Haren Myneni <hbabu@us.ibm.com>
Cc: Rob Landley <rob@landley.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoflush_icache_range: export symbol to fix build errors
Pranith Kumar [Fri, 29 Aug 2014 22:19:09 +0000 (15:19 -0700)]
flush_icache_range: export symbol to fix build errors

Fix building errors occuring due to a missing export of
flush_icache_range() in

kisskb.ellerman.id.au/kisskb/buildresult/11677809/

ERROR: "flush_icache_range" [drivers/misc/lkdtm.ko] undefined!

Signed-off-by: Pranith Kumar <bobby.prani@gmail.com>
Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Vineet Gupta <vgupta@synopsys.com> [arc]
Acked-by: Richard Kuo <rkuo@codeaurora.org> [hexagon]
Cc: Chris Metcalf <cmetcalf@tilera.com>
Cc: Chris Zankel <chris@zankel.net>
Acked-by: Max Filippov <jcmvbkbc@gmail.com> [xtensa]
Cc: Noam Camus <noamc@ezchip.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Acked-by: Zhigang Lu <zlu@tilera.com> [tile]
Cc: Kirill Tkhai <tkhai@yandex.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agotools: selftests: fix build issue with make kselftests target
Phong Tran [Fri, 29 Aug 2014 22:19:06 +0000 (15:19 -0700)]
tools: selftests: fix build issue with make kselftests target

Fix the typo of ARCH when running 'make kselftests'.  Change the 'X86'
to 'x86'.  Test by compilation.

Signed-off-by: Phong Tran <tranmanphong@gmail.com>
Cc: David Herrmann <dh.herrmann@gmail.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Shuah Khan <shuah.kh@samsung.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Michal Marek <mmarek@suse.cz>
Cc: Shuah Khan <shuahkh@osg.samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: quorum: add a log for node not fenced
Junxiao Bi [Fri, 29 Aug 2014 22:19:04 +0000 (15:19 -0700)]
ocfs2: quorum: add a log for node not fenced

For debug use, we can see from the log whether the fence decision is
made and why it is not fenced.

Signed-off-by: Junxiao Bi <junxiao.bi@oracle.com>
Reviewed-by: Srinivas Eeda <srinivas.eeda@oracle.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: o2net: set tcp user timeout to max value
Junxiao Bi [Fri, 29 Aug 2014 22:19:02 +0000 (15:19 -0700)]
ocfs2: o2net: set tcp user timeout to max value

When tcp retransmit timeout(15mins), the connection will be closed.
Pending messages may be lost during this time.  So we set tcp user
timeout to override the retransmit timeout to the max value.  This is OK
for ocfs2 since we have disk heartbeat, if peer crash, the disk
heartbeat will timeout and it will be evicted, if disk heartbeat not
timeout and connection idle for a long time, then this means the cluster
enters split-brain state, since fence can't happen, we'd better keep the
connection and wait network recover.

Signed-off-by: Junxiao Bi <junxiao.bi@oracle.com>
Reviewed-by: Srinivas Eeda <srinivas.eeda@oracle.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: o2net: don't shutdown connection when idle timeout
Junxiao Bi [Fri, 29 Aug 2014 22:19:00 +0000 (15:19 -0700)]
ocfs2: o2net: don't shutdown connection when idle timeout

This patch series is to fix a possible message lost bug in ocfs2 when
network go bad.  This bug will cause ocfs2 hung forever even network
become good again.

The messages may lost in this case.  After the tcp connection is
established between two nodes, an idle timer will be set to check its
state periodically, if no messages are received during this time, idle
timer will timeout, it will shutdown the connection and try to
reconnect, so pending messages in tcp queues will be lost.  This
messages may be from dlm.  Dlm may get hung in this case.  This may
cause the whole ocfs2 cluster hung.

This is very possible to happen when network state goes bad.  Do the
reconnect is useless, it will fail if network state is still bad.  Just
waiting there for network recovering may be a good idea, it will not
lost messages and some node will be fenced until cluster goes into
split-brain state, for this case, Tcp user timeout is used to override
the tcp retransmit timeout.  It will timeout after 25 days, user should
have notice this through the provided log and fix the network, if they
don't, ocfs2 will fall back to original reconnect way.

This patch (of 3):

Some messages in the tcp queue maybe lost if we shutdown the connection
and reconnect when idle timeout.  If packets lost and reconnect success,
then the ocfs2 cluster maybe hung.

To fix this, we can leave the connection there and do the fence decision
when idle timeout, if network recover before fence dicision is made, the
connection survive without lost any messages.

This bug can be saw when network state go bad.  It may cause ocfs2 hung
forever if some packets lost.  With this fix, ocfs2 will recover from
hung if network becomes good again.

Signed-off-by: Junxiao Bi <junxiao.bi@oracle.com>
Reviewed-by: Srinivas Eeda <srinivas.eeda@oracle.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: do not write error flag to user structure we cannot copy from/to
Ben Hutchings [Fri, 29 Aug 2014 22:18:58 +0000 (15:18 -0700)]
ocfs2: do not write error flag to user structure we cannot copy from/to

If we failed to copy from the structure, writing back the flags leaks 31
bits of kernel memory (the rest of the ir_flags field).

In any case, if we cannot copy from/to the structure, why should we
expect putting just the flags to work?

Also make sure ocfs2_info_handle_freeinode() returns the right error
code if the copy_to_user() fails.

Fixes: ddee5cdb70e6 ('Ocfs2: Add new OCFS2_IOC_INFO ioctl for ocfs2 v8.')
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Cc: Joel Becker <jlbec@evilplan.org>
Acked-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agox86/purgatory: use approprate -m64/-32 build flag for arch/x86/purgatory
Vivek Goyal [Fri, 29 Aug 2014 22:18:55 +0000 (15:18 -0700)]
x86/purgatory: use approprate -m64/-32 build flag for arch/x86/purgatory

Thomas reported that build of x86_64 kernel was failing for him.  He is
using 32bit tool chain.

Problem is that while compiling purgatory, I have not specified -m64
flag.  And 32bit tool chain must be assuming -m32 by default.

Following is error message.

(mini) [~/work/linux-2.6] make
scripts/kconfig/conf --silentoldconfig Kconfig
  CHK     include/config/kernel.release
  UPD     include/config/kernel.release
  CHK     include/generated/uapi/linux/version.h
  CHK     include/generated/utsrelease.h
  UPD     include/generated/utsrelease.h
  CC      arch/x86/purgatory/purgatory.o
arch/x86/purgatory/purgatory.c:1:0: error: code model 'large' not supported in
the 32 bit mode

Fix it by explicitly passing appropriate -m64/-m32 build flag for
purgatory.

Reported-by: Thomas Glanzmann <thomas@glanzmann.de>
Tested-by: Thomas Glanzmann <thomas@glanzmann.de>
Suggested-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agodrivers/rtc/rtc-s5m.c: re-add support for devices without irq specified
Bartlomiej Zolnierkiewicz [Fri, 29 Aug 2014 22:18:53 +0000 (15:18 -0700)]
drivers/rtc/rtc-s5m.c: re-add support for devices without irq specified

The rtc-s5m driver used to support devices without irq specified in the
past.  Re-add this support.

The patch fixes boot for Insignal's Exynos4412 based Origen board.

Error messages before the patch:

  ...
  Unable to handle kernel NULL pointer dereference at virtual address 00000094
  pgd = c0004000
  [00000094] *pgd=00000000
  Internal error: Oops: 5 [#1] PREEMPT SMP ARM
  Modules linked in:
  CPU: 1 PID: 1 Comm: swapper/0 Not tainted 3.16.0-next-20140804-00008-ga59480f-dirty #701
  task: ea80f000 ti: ea882000 task.ti: ea882000
  PC is at regmap_irq_get_virq+0x0/0x28
  LR is at s5m_rtc_probe+0xdc/0x310
  pc : [<c02574fc>]    lr : [<c02de75c>]    psr: 80000153
  sp : ea883e48  ip : 00000000  fp : 00000000
  r10: 0000000c  r9 : c05de7ac  r8 : eaabc600
  r7 : eaa6b4d0  r6 : c0439e8c  r5 : eaabc610  r4 : eab30e50
  r3 : 00000000  r2 : 00000000  r1 : 0000000c  r0 : 00000000
  Flags: Nzcv  IRQs on  FIQs off  Mode SVC_32  ISA ARM  Segment kernel
  Control: 10c5387d  Table: 4000404a  DAC: 00000015
  Process swapper/0 (pid: 1, stack limit = 0xea882240)
  Backtrace:
    regmap_irq_get_virq
    s5m_rtc_probe
    platform_drv_probe
    driver_probe_device
    __driver_attach
    bus_for_each_dev
    bus_add_driver
    driver_register
    do_one_initcall
    kernel_init_freeable
    kernel_init
  ---[ end trace a954d7f019122700 ]---
  Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
  ...

Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Tested-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoxattr: fix check for simultaneous glibc header inclusion
Filipe Brandenburger [Fri, 29 Aug 2014 22:18:51 +0000 (15:18 -0700)]
xattr: fix check for simultaneous glibc header inclusion

The guard was introduced in commit ea1a8217b06b ("xattr: guard against
simultaneous glibc header inclusion") but it is using #ifdef to check
for a define that is either set to 1 or 0.  Fix it to use #if instead.

* Without this patch:

  $ { echo "#include <sys/xattr.h>"; echo "#include <linux/xattr.h>"; } | gcc -E -Iinclude/uapi - >/dev/null
  include/uapi/linux/xattr.h:19:0: warning: "XATTR_CREATE" redefined [enabled by default]
   #define XATTR_CREATE 0x1 /* set value, fail if attr already exists */
   ^
  /usr/include/x86_64-linux-gnu/sys/xattr.h:32:0: note: this is the location of the previous definition
   #define XATTR_CREATE XATTR_CREATE
   ^

* With this patch:

  $ { echo "#include <sys/xattr.h>"; echo "#include <linux/xattr.h>"; } | gcc -E -Iinclude/uapi - >/dev/null
  (no warnings)

Signed-off-by: Filipe Brandenburger <filbranden@google.com>
Acked-by: Serge E. Hallyn <serge.hallyn@ubuntu.com>
Cc: Allan McRae <allan@archlinux.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agokexec: remove CONFIG_KEXEC dependency on crypto
Vivek Goyal [Fri, 29 Aug 2014 22:18:49 +0000 (15:18 -0700)]
kexec: remove CONFIG_KEXEC dependency on crypto

New system call depends on crypto.  As it did not have a separate config
option, CONFIG_KEXEC was modified to select CRYPTO and CRYPTO_SHA256.

But now previous patch introduced a new config option for new syscall.
So CONFIG_KEXEC does not require crypto.  Remove that dependency.

Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Shaun Ruffell <sruffell@digium.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agokexec: create a new config option CONFIG_KEXEC_FILE for new syscall
Vivek Goyal [Fri, 29 Aug 2014 22:18:46 +0000 (15:18 -0700)]
kexec: create a new config option CONFIG_KEXEC_FILE for new syscall

Currently new system call kexec_file_load() and all the associated code
compiles if CONFIG_KEXEC=y.  But new syscall also compiles purgatory
code which currently uses gcc option -mcmodel=large.  This option seems
to be available only gcc 4.4 onwards.

Hiding new functionality behind a new config option will not break
existing users of old gcc.  Those who wish to enable new functionality
will require new gcc.  Having said that, I am trying to figure out how
can I move away from using -mcmodel=large but that can take a while.

I think there are other advantages of introducing this new config
option.  As this option will be enabled only on x86_64, other arches
don't have to compile generic kexec code which will never be used.  This
new code selects CRYPTO=y and CRYPTO_SHA256=y.  And all other arches had
to do this for CONFIG_KEXEC.  Now with introduction of new config
option, we can remove crypto dependency from other arches.

Now CONFIG_KEXEC_FILE is available only on x86_64.  So whereever I had
CONFIG_X86_64 defined, I got rid of that.

For CONFIG_KEXEC_FILE, instead of doing select CRYPTO=y, I changed it to
"depends on CRYPTO=y".  This should be safer as "select" is not
recursive.

Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Tested-by: Shaun Ruffell <sruffell@digium.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agox86,mm: fix pte_special versus pte_numa
Hugh Dickins [Fri, 29 Aug 2014 22:18:44 +0000 (15:18 -0700)]
x86,mm: fix pte_special versus pte_numa

Sasha Levin has shown oopses on ffffea0003480048 and ffffea0003480008 at
mm/memory.c:1132, running Trinity on different 3.16-rc-next kernels:
where zap_pte_range() checks page->mapping to see if PageAnon(page).

Those addresses fit struct pages for pfns d2001 and d2000, and in each
dump a register or a stack slot showed d2001730 or d2000730: pte flags
0x730 are PCD ACCESSED PROTNONE SPECIAL IOMAP; and Sasha's e820 map has
a hole between cfffffff and 100000000, which would need special access.

Commit c46a7c817e66 ("x86: define _PAGE_NUMA by reusing software bits on
the PMD and PTE levels") has broken vm_normal_page(): a PROTNONE SPECIAL
pte no longer passes the pte_special() test, so zap_pte_range() goes on
to try to access a non-existent struct page.

Fix this by refining pte_special() (SPECIAL with PRESENT or PROTNONE) to
complement pte_numa() (SPECIAL with neither PRESENT nor PROTNONE).  A
hint that this was a problem was that c46a7c817e66 added pte_numa() test
to vm_normal_page(), and moved its is_zero_pfn() test from slow to fast
path: This was papering over a pte_special() snag when the zero page was
encountered during zap.  This patch reverts vm_normal_page() to how it
was before, relying on pte_special().

It still appears that this patch may be incomplete: aren't there other
places which need to be handling PROTNONE along with PRESENT?  For
example, pte_mknuma() clears _PAGE_PRESENT and sets _PAGE_NUMA, but on a
PROT_NONE area, that would make it pte_special().  This is side-stepped
by the fact that NUMA hinting faults skipped PROT_NONE VMAs and there
are no grounds where a NUMA hinting fault on a PROT_NONE VMA would be
interesting.

Fixes: c46a7c817e66 ("x86: define _PAGE_NUMA by reusing software bits on the PMD and PTE levels")
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Tested-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Matthew Wilcox <matthew.r.wilcox@intel.com>
Cc: <stable@vger.kernel.org> [3.16]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agohugetlb_cgroup: use lockdep_assert_held rather than spin_is_locked
Michal Hocko [Fri, 29 Aug 2014 22:18:42 +0000 (15:18 -0700)]
hugetlb_cgroup: use lockdep_assert_held rather than spin_is_locked

spin_lock may be an empty struct for !SMP configurations and so
arch_spin_is_locked may return unconditional 0 and trigger the VM_BUG_ON
even when the lock is held.

Replace spin_is_locked by lockdep_assert_held.  We will not BUG anymore
but it is questionable whether crashing makes a lot of sense in the
uncharge path.  Uncharge happens after the last page reference was
released so nobody should touch the page and the function doesn't update
any shared state except for res counter which uses synchronization of
its own.

Signed-off-by: Michal Hocko <mhocko@suse.cz>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm/zpool: use prefixed module loading
Kees Cook [Fri, 29 Aug 2014 22:18:40 +0000 (15:18 -0700)]
mm/zpool: use prefixed module loading

To avoid potential format string expansion via module parameters, do not
use the zpool type directly in request_module() without a format string.
Additionally, to avoid arbitrary modules being loaded via zpool API
(e.g.  via the zswap_zpool_type module parameter) add a "zpool-" prefix
to the requested module, as well as module aliases for the existing
zpool types (zbud and zsmalloc).

Signed-off-by: Kees Cook <keescook@chromium.org>
Cc: Seth Jennings <sjennings@variantweb.net>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Nitin Gupta <ngupta@vflare.org>
Acked-by: Dan Streetman <ddstreet@ieee.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agozram: fix incorrect stat with failed_reads
Chao Yu [Fri, 29 Aug 2014 22:18:37 +0000 (15:18 -0700)]
zram: fix incorrect stat with failed_reads

Since we allocate a temporary buffer in zram_bvec_read to handle partial
page operations in commit 924bd88d703e ("Staging: zram: allow partial
page operations"), our ->failed_reads value may be incorrect as we do
not increase its value when failing to allocate the temporary buffer.

Let's fix this issue and correct the annotation of failed_reads.

Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Cc: Nitin Gupta <ngupta@vflare.org>
Acked-by: Jerome Marchand <jmarchan@redhat.com>
Acked-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agolib: turn CONFIG_STACKTRACE into an actual option.
Dave Jones [Fri, 29 Aug 2014 22:18:35 +0000 (15:18 -0700)]
lib: turn CONFIG_STACKTRACE into an actual option.

I was puzzled why /proc/$$/stack had disappeared, until I figured out I
had disabled the last debug option that did a 'select STACKTRACE'.  This
patch makes the option show up at config time, so it can be enabled
without enabling any of the more heavyweight debug options.

Signed-off-by: Dave Jones <davej@redhat.com>
Acked-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: actually clear pmd_numa before invalidating
Matthew Wilcox [Fri, 29 Aug 2014 22:18:33 +0000 (15:18 -0700)]
mm: actually clear pmd_numa before invalidating

Commit 67f87463d3a3 ("mm: clear pmd_numa before invalidating") cleared
the NUMA bit in a copy of the PMD entry, but then wrote back the
original

Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomemblock, memhotplug: fix wrong type in memblock_find_in_range_node().
Tang Chen [Fri, 29 Aug 2014 22:18:31 +0000 (15:18 -0700)]
memblock, memhotplug: fix wrong type in memblock_find_in_range_node().

In memblock_find_in_range_node(), we defined ret as int.  But it should
be phys_addr_t because it is used to store the return value from
__memblock_find_range_bottom_up().

The bug has not been triggered because when allocating low memory near
the kernel end, the "int ret" won't turn out to be negative.  When we
started to allocate memory on other nodes, and the "int ret" could be
minus.  Then the kernel will panic.

A simple way to reproduce this: comment out the following code in
numa_init(),

        memblock_set_bottom_up(false);

and the kernel won't boot.

Reported-by: Xishi Qiu <qiuxishi@huawei.com>
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Tested-by: Xishi Qiu <qiuxishi@huawei.com>
Cc: <stable@vger.kernel.org> [3.13+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoresource: fix the case of null pointer access
Vivek Goyal [Fri, 29 Aug 2014 22:18:29 +0000 (15:18 -0700)]
resource: fix the case of null pointer access

Richard and Daniel reported that UML is broken due to changes to
resource traversal functions.  Problem is that iomem_resource.child can
be null and new code does not consider that possibility.  Old code used
a for loop and that loop will not even execute if p was null.

Revert back to for() loop logic and bail out if p is null.

I also moved sibling_only check out of resource_lock. There is no
reason to keep it inside the lock.

Following is backtrace of the UML crash.

RIP: 0033:[<0000000060039b9f>]
RSP: 0000000081459da0  EFLAGS: 00010202
RAX: 0000000000000000 RBX: 00000000219b3fff RCX: 000000006010d1d9
RDX: 0000000000000001 RSI: 00000000602dfb94 RDI: 0000000081459df8
RBP: 0000000081459de0 R08: 00000000601b59f4 R09: ffffffff0000ff00
R10: ffffffff0000ff00 R11: 0000000081459e88 R12: 0000000081459df8
R13: 00000000219b3fff R14: 00000000602dfb94 R15: 0000000000000000
Kernel panic - not syncing: Segfault with no mm
CPU: 0 PID: 1 Comm: swapper Not tainted 3.16.0-10454-g58d08e3 #13
Stack:
 00000000 000080d0 81459df0 219b3fff
 81459e70 6010d1d9 ffffffff 6033e010
 81459e50 6003a269 81459e30 00000000
Call Trace:
 [<6010d1d9>] ? kclist_add_private+0x0/0xe7
 [<6003a269>] walk_system_ram_range+0x61/0xb7
 [<6000e859>] ? proc_kcore_init+0x0/0xf1
 [<6010d574>] kcore_update_ram+0x4c/0x168
 [<6010d72e>] ? kclist_add+0x0/0x2e
 [<6000e943>] proc_kcore_init+0xea/0xf1
 [<6000e859>] ? proc_kcore_init+0x0/0xf1
 [<6000e859>] ? proc_kcore_init+0x0/0xf1
 [<600189f0>] do_one_initcall+0x13c/0x204
 [<6004ca46>] ? parse_args+0x1df/0x2e0
 [<6004c82d>] ? parameq+0x0/0x3a
 [<601b5990>] ? strcpy+0x0/0x18
 [<60001e1a>] kernel_init_freeable+0x240/0x31e
 [<6026f1c0>] kernel_init+0x12/0x148
 [<60019fad>] new_thread_handler+0x81/0xa3

Fixes 8c86e70acead629aacb4a ("resource: provide new functions to walk
through resources").

Reported-by: Daniel Walter <sahne@0x90.at>
Tested-by: Richard Weinberger <richard@nod.at>
Tested-by: Toralf Förster <toralf.foerster@gmx.de>
Tested-by: Daniel Walter <sahne@0x90.at>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agocheckpatch: relax check for length of git commit IDs
Joe Perches [Fri, 29 Aug 2014 22:18:26 +0000 (15:18 -0700)]
checkpatch: relax check for length of git commit IDs

Checkpatch currently warns if a git commit ID (in the changelog,
usually) is less than 12 characters or more than 16.  The "more than 16"
is excessive.  Change the check so we accept IDs from 12 to 40 chars in
length.

Cc: Geert Uytterhoeven <geert@linux-m68k.org
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoMerge tag 'nfs-for-3.17-3' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Fri, 29 Aug 2014 20:04:13 +0000 (13:04 -0700)]
Merge tag 'nfs-for-3.17-3' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client fixes from Trond Myklebust:
 "Highlights:
   - NFSv3 stable fix for another POSIX ACL regression
   - NFSv4 stable fix for a regression with OPEN_DOWNGRADE
   - NFSv4 stable fix for bad close() behaviour when holding a delegation"

* tag 'nfs-for-3.17-3' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  NFSv3: Fix another acl regression
  NFSv4: Don't clear the open state when we just did an OPEN_DOWNGRADE
  NFSv4: Fix problems with close in the presence of a delegation

10 years agoMerge tag 'usb-3.17-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Fri, 29 Aug 2014 19:10:03 +0000 (12:10 -0700)]
Merge tag 'usb-3.17-rc3' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are a bunch of fixes for the USB drivers for 3.17-rc3.

  Also in here is the movement of the usbip driver out of staging, into
  the "real" part of the kernel, it had to wait until after -rc1 to
  handle the merge issues involved between the USB and staging trees.
  The code is identical, just file movements there.

  The USB fixes are all over the place, new device ids, xhci fixes for
  reported issues and the usual gadget driver fixes as well.  All have
  been in linux-next for a while now"

* tag 'usb-3.17-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (46 commits)
  USB: fix build error with CONFIG_PM_RUNTIME disabled
  Revert "usb: ehci/ohci-exynos: Fix PHY getting sequence"
  xhci: Disable streams on Via XHCI with device-id 0x3432
  USB: serial: fix potential heap buffer overflow
  USB: serial: fix potential stack buffer overflow
  usb: ehci/ohci-exynos: Fix PHY getting sequence
  usb: hub: Prevent hub autosuspend if usbcore.autosuspend is -1
  USB: sisusb: add device id for Magic Control USB video
  usb: dwc2: gadget: Set the default EP max packet value as 8 bytes
  usb: ehci: using wIndex + 1 for hub port
  USB: storage: add quirk for Newer Technology uSCSI SCSI-USB converter
  MAINTAINERS: Add an entry for USB/IP driver
  usbip: remove struct usb_device_id table
  usbip: move usbip kernel code out of staging
  usbip: move usbip userspace code out of staging
  USB: whiteheat: Added bounds checking for bulk command response
  usb: gadget: remove $(PWD) in ccflags-y
  usb: pch_udc: usb gadget device support for Intel Quark X1000
  usb: gadget: uvc: fix possible lockup in uvc gadget
  usb: wusbcore: fix below build warning
  ...

10 years agoMerge tag 'staging-3.17-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Fri, 29 Aug 2014 19:04:10 +0000 (12:04 -0700)]
Merge tag 'staging-3.17-rc3' of git://git./linux/kernel/git/gregkh/staging

Pull staging driver fixes from Greg KH:
 "Here are some staging driver fixes for your tree.  Nothing huge, just
  some fixes for issues that have been reported and a few new device ids
  added.

  All have been in linux-next for a while"

* tag 'staging-3.17-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging: r8188eu: Add new USB ID
  staging/rtl8188eu: add 0df6:0076 Sitecom Europe B.V.
  staging: android: fix a possible memory leak
  staging: lustre: lustre: libcfs: workitem.c: Cleaning up missing null-terminate after strncpy call
  staging: et131x: Fix errors caused by phydev->addr accesses before initialisation
  staging: lustre: Remove circular dependency on header

10 years agoMerge tag 'char-misc-3.17-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregk...
Linus Torvalds [Fri, 29 Aug 2014 19:01:22 +0000 (12:01 -0700)]
Merge tag 'char-misc-3.17-rc3' of git://git./linux/kernel/git/gregkh/char-misc

Pull char/misc driver fixes from Greg KH:
 "Here are 3 fixes for the mei and thunderbolt drivers that resolve some
  reported issues.

  All have been in linux-next for a while"

* tag 'char-misc-3.17-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  thunderbolt: Clear hops before overwriting
  mei: nfc: fix memory leak in error path
  mei: reset client state on queued connect request

10 years agoMerge tag 'fbdev-fixes-3.17' of git://git.kernel.org/pub/scm/linux/kernel/git/tomba...
Linus Torvalds [Fri, 29 Aug 2014 18:59:46 +0000 (11:59 -0700)]
Merge tag 'fbdev-fixes-3.17' of git://git./linux/kernel/git/tomba/linux

Pull fbdev fixes from Tomi Valkeinen:
 "Minor fbdev fixes for da8xx-fb, atmel_lcdfb, arm clcd and chipsfb"

* tag 'fbdev-fixes-3.17' of git://git.kernel.org/pub/scm/linux/kernel/git/tomba/linux:
  video: da8xx-fb: preserve display width when changing HSYNC
  video: of: display_timing: double free on error
  drivers: video: fbdev: atmel_lcdfb.c: fix error return code
  video: ARM CLCD: Fix calculation of bits-per-pixel
  fbdev: Remove __init from chips_hw_init() to fix build failure

10 years agoMerge tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 29 Aug 2014 18:52:46 +0000 (11:52 -0700)]
Merge tag 'ext4_for_linus_stable' of git://git./linux/kernel/git/tytso/ext4

Pull ext4 bugfixes from Ted Ts'o:
 "Ext4 bug fixes for 3.17, to provide better handling of memory
  allocation failures, and to fix some journaling bugs involving
  journal checksums and FALLOC_FL_ZERO_RANGE"

* tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: fix same-dir rename when inline data directory overflows
  jbd2: fix descriptor block size handling errors with journal_csum
  jbd2: fix infinite loop when recovering corrupt journal blocks
  ext4: update i_disksize coherently with block allocation on error path
  ext4: fix transaction issues for ext4_fallocate and ext_zero_range
  ext4: fix incorect journal credits reservation in ext4_zero_range
  ext4: move i_size,i_disksize update routines to helper function
  ext4: fix BUG_ON in mb_free_blocks()
  ext4: propagate errors up to ext4_find_entry()'s callers

10 years agoMerge tag 'dm-3.17-fix' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mappe...
Linus Torvalds [Fri, 29 Aug 2014 18:49:10 +0000 (11:49 -0700)]
Merge tag 'dm-3.17-fix' of git://git./linux/kernel/git/device-mapper/linux-dm

Pull device mapper fix from Mike Snitzer:
 "Fix a 3.17-rc1 regression introduced by switching the DM crypt target
  to using per-bio data"

* tag 'dm-3.17-fix' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm crypt: fix access beyond the end of allocated space

10 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-block
Linus Torvalds [Fri, 29 Aug 2014 18:21:48 +0000 (11:21 -0700)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-block

Pull block layer fixes from Jens Axboe:
 "A smaller collection of fixes that have come up since the initial
  merge window pull request.  This contains:

   - error handling cleanup and support for larger than 16 byte cdbs in
     sg_io() from Christoph.  The latter just matches what bsg and
     friends support, sg_io() got left out in the merge.

   - an option for brd to expose partitions in /proc/partitions.  They
     are hidden by default for compat reasons.  From Dmitry Monakhov.

   - a few blk-mq fixes from me - killing a dead/unused flag, fix for
     merging happening even if turned off, and correction of a few
     comments.

   - removal of unnecessary ->owner setting in systemace.  From Michal
     Simek.

   - two related fixes for a problem with nesting freezing of queues in
     blk-mq.  One from Ming Lei removing an unecessary freeze operation,
     and another from Tejun fixing the nesting regression introduced in
     the merge window.

   - fix for a BUG_ON() at bio_endio time when protection info is
     attached and the IO has an error.  From Sagi Grimberg.

   - two scsi_ioctl bug fixes for regressions with scsi-mq from Tony
     Battersby.

   - a cfq weight update fix and subsequent comment update from Toshiaki
     Makita"

* 'for-linus' of git://git.kernel.dk/linux-block:
  cfq-iosched: Add comments on update timing of weight
  cfq-iosched: Fix wrong children_weight calculation
  block: fix error handling in sg_io
  fix regression in SCSI_IOCTL_SEND_COMMAND
  scsi-mq: fix requests that use a separate CDB buffer
  block: support > 16 byte CDBs for SG_IO
  block: cleanup error handling in sg_io
  brd: add ram disk visibility option
  block: systemace: Remove .owner field for driver
  blk-mq: blk_mq_freeze_queue() should allow nesting
  blk-mq: correct a few wrong/bad comments
  block: Fix BUG_ON when pi errors occur
  blk-mq: don't allow merges if turned off for the queue
  blk-mq: get rid of unused BLK_MQ_F_SHOULD_SORT flag
  blk-mq: fix WARNING "percpu_ref_kill() called more than once!"

10 years agoalpha: io: implement relaxed accessor macros for writes
Will Deacon [Fri, 25 Jul 2014 00:53:54 +0000 (17:53 -0700)]
alpha: io: implement relaxed accessor macros for writes

write{b,w,l,q}_relaxed are implemented by some architectures in order to
permit memory-mapped I/O writes with weaker barrier semantics than the
non-relaxed variants.

This patch implements these write macros for Alpha, in the same vein as
the relaxed read macros, which are already implemented.

Acked-by: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoalpha: Wire up sched_setattr, sched_getattr, and renameat2 syscalls.
Michael Cree [Fri, 25 Jul 2014 00:53:53 +0000 (17:53 -0700)]
alpha: Wire up sched_setattr, sched_getattr, and renameat2 syscalls.

Signed-off-by: Michael Cree <mcree@orcon.net.nz>
Signed-off-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agox86, irq, PCI: Keep IRQ assignment for runtime power management
Jiang Liu [Fri, 29 Aug 2014 09:26:23 +0000 (17:26 +0800)]
x86, irq, PCI: Keep IRQ assignment for runtime power management

Now IOAPIC driver dynamically allocates IRQ numbers for IOAPIC pins.
We need to keep IRQ assignment for PCI devices during runtime power
management, otherwise it may cause failure of device wakeups.

Commit 3eec595235c17a7 "x86, irq, PCI: Keep IRQ assignment for PCI
devices during suspend/hibernation" has fixed the issue for suspend/
hibernation, we also need the same fix for runtime device sleep too.

Fix: https://bugzilla.kernel.org/show_bug.cgi?id=83271
Reported-and-Tested-by: EmanueL Czirai <amanual@openmailbox.org>
Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: EmanueL Czirai <amanual@openmailbox.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Rafael J. Wysocki <rjw@rjwysocki.net>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Grant Likely <grant.likely@linaro.org>
Link: http://lkml.kernel.org/r/1409304383-18806-1-git-send-email-jiang.liu@linux.intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
10 years agospi/rockchip: Avoid accidentally turning off the clock
Doug Anderson [Thu, 28 Aug 2014 23:43:48 +0000 (16:43 -0700)]
spi/rockchip: Avoid accidentally turning off the clock

If our client is requesting a clock that is above the maximum clock
then the following division will result in 0:
  rs->max_freq / rs->speed

We'll then program 0 into the SPI_BAUDR register.  The Rockchip TRM
says: "If the value is 0, the serial output clock (sclk_out) is
disabled."

It's much better to end up with the fastest possible clock rather than
a clock that is off, so enforce a minimum value.

Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
10 years agoext4: fix same-dir rename when inline data directory overflows
Darrick J. Wong [Wed, 27 Aug 2014 22:40:09 +0000 (18:40 -0400)]
ext4: fix same-dir rename when inline data directory overflows

When performing a same-directory rename, it's possible that adding or
setting the new directory entry will cause the directory to overflow
the inline data area, which causes the directory to be converted to an
extent-based directory.  Under this circumstance it is necessary to
re-read the directory when deleting the old dirent because the "old
directory" context still points to i_block in the inode table, which
is now an extent tree root!  The delete fails with an FS error, and
the subsequent fsck complains about incorrect link counts and
hardlinked directories.

Test case (originally found with flat_dir_test in the metadata_csum
test program):

# mkfs.ext4 -O inline_data /dev/sda
# mount /dev/sda /mnt
# mkdir /mnt/x
# touch /mnt/x/changelog.gz /mnt/x/copyright /mnt/x/README.Debian
# sync
# for i in /mnt/x/*; do mv $i $i.longer; done
# ls -la /mnt/x/
total 0
-rw-r--r-- 1 root root 0 Aug 25 12:03 changelog.gz.longer
-rw-r--r-- 1 root root 0 Aug 25 12:03 copyright
-rw-r--r-- 1 root root 0 Aug 25 12:03 copyright.longer
-rw-r--r-- 1 root root 0 Aug 25 12:03 README.Debian.longer

(Hey!  Why are there four files now??)

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org
10 years agojbd2: fix descriptor block size handling errors with journal_csum
Darrick J. Wong [Wed, 27 Aug 2014 22:40:07 +0000 (18:40 -0400)]
jbd2: fix descriptor block size handling errors with journal_csum

It turns out that there are some serious problems with the on-disk
format of journal checksum v2.  The foremost is that the function to
calculate descriptor tag size returns sizes that are too big.  This
causes alignment issues on some architectures and is compounded by the
fact that some parts of jbd2 use the structure size (incorrectly) to
determine the presence of a 64bit journal instead of checking the
feature flags.

Therefore, introduce journal checksum v3, which enlarges the
descriptor block tag format to allow for full 32-bit checksums of
journal blocks, fix the journal tag function to return the correct
sizes, and fix the jbd2 recovery code to use feature flags to
determine 64bitness.

Add a few function helpers so we don't have to open-code quite so
many pieces.

Switching to a 16-byte block size was found to increase journal size
overhead by a maximum of 0.1%, to convert a 32-bit journal with no
checksumming to a 32-bit journal with checksum v3 enabled.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reported-by: TR Reardon <thomas_reardon@hotmail.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org