cascardo/linux.git
9 years agoMerge tag 'for-v3.19/omap-a' of git://git.kernel.org/pub/scm/linux/kernel/git/pjw...
Michael Turquette [Mon, 15 Dec 2014 21:30:05 +0000 (13:30 -0800)]
Merge tag 'for-v3.19/omap-a' of git://git./linux/kernel/git/pjw/omap-pending into tmp

Some OMAP clock/hwmod patches for v3.19.

Most of the patches are clock-related.  The DPLL implementation is
changed to better align to the common clock framework.
There is also a patch that removes a few lines from the hwmod code -
this patch should have no functional effect.

Basic build, boot, and PM test logs for these patches can be found here:

http://www.pwsan.com/omap/testlogs/omap-a-for-v3.19/20141113094101/

9 years agoclk: Really fix deadlock with mmap_sem
Stephen Boyd [Fri, 12 Dec 2014 23:04:16 +0000 (15:04 -0800)]
clk: Really fix deadlock with mmap_sem

Commit 6314b6796e3c (clk: Don't hold prepare_lock across debugfs
creation, 2014-09-04) forgot to update one place where we hold
the prepare_lock while creating debugfs directories. This means
we still have the chance of a deadlock that the commit was trying
to fix. Actually fix it by moving the debugfs creation outside
the prepare_lock.

Cc: <stable@vger.kernel.org> # 3.18
Reported-by: Russell King <rmk+kernel@arm.linux.org.uk>
Fixes: 6314b6796e3c "clk: Don't hold prepare_lock across debugfs creation"
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
[mturquette@linaro.org: removed lockdep_assert]

9 years agoMerge tag 'for-v3.19-exynos-clk-2' of git://linuxtv.org/snawrocki/samsung into clk...
Michael Turquette [Thu, 11 Dec 2014 20:17:15 +0000 (12:17 -0800)]
Merge tag 'for-v3.19-exynos-clk-2' of git://linuxtv.org/snawrocki/samsung into clk-next

- exynos4415 and exynos audio subsystem clk driver (build
  with PM_SLEEP disabled, resource release) fixes
- minor cleanups in drivers/clk/samsung/clk.c (spelling,
  includes)
- modification of the exynos4 HDMI PHY clock definition to
  model dependency of "sclk_hdmiphy" on the "hdmi" clock

9 years agoclk: mmp: fix sparse non static symbol warning
Wei Yongjun [Tue, 9 Dec 2014 01:07:30 +0000 (09:07 +0800)]
clk: mmp: fix sparse non static symbol warning

Fixes the following sparse warnings:

drivers/clk/mmp/clk-frac.c:113:6: warning:
 symbol 'clk_factor_init' was not declared. Should it be static?

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: Change clk_ops->determine_rate to return a clk_hw as the best parent
Tomeu Vizoso [Tue, 2 Dec 2014 07:54:22 +0000 (08:54 +0100)]
clk: Change clk_ops->determine_rate to return a clk_hw as the best parent

This is in preparation for clock providers to not have to deal with struct clk.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: change clk_debugfs_add_file to take a struct clk_hw
Tomeu Vizoso [Tue, 2 Dec 2014 07:54:21 +0000 (08:54 +0100)]
clk: change clk_debugfs_add_file to take a struct clk_hw

Instead of struct clk, as this should be only used by providers.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: Don't expose __clk_get_accuracy
Tomeu Vizoso [Tue, 2 Dec 2014 07:54:20 +0000 (08:54 +0100)]
clk: Don't expose __clk_get_accuracy

As it's only used internally, in drivers/clk/clk.c.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: Don't try to use a struct clk* after it could have been freed
Tomeu Vizoso [Tue, 2 Dec 2014 07:54:19 +0000 (08:54 +0100)]
clk: Don't try to use a struct clk* after it could have been freed

As __clk_release could call kfree on clk and then we wouldn't have a safe way
of getting the module that owns the clock.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Fixes: fcb0ee6a3d33 ("clk: Implement clk_unregister")
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: Remove unused function __clk_get_prepare_count
Tomeu Vizoso [Tue, 2 Dec 2014 07:54:18 +0000 (08:54 +0100)]
clk: Remove unused function __clk_get_prepare_count

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: samsung: Fix double add of syscore ops after driver rebind
Krzysztof Kozlowski [Wed, 26 Nov 2014 14:24:13 +0000 (15:24 +0100)]
clk: samsung: Fix double add of syscore ops after driver rebind

During driver unbind the syscore ops were not unregistered which lead to
double add on syscore list:

$ echo "3810000.audss-clock-controller" > /sys/bus/platform/drivers/exynos-audss-clk/unbind
$ echo "3810000.audss-clock-controller" > /sys/bus/platform/drivers/exynos-audss-clk/bind
[ 1463.044061] ------------[ cut here ]------------
[ 1463.047255] WARNING: CPU: 0 PID: 1 at lib/list_debug.c:36 __list_add+0x8c/0xc0()
[ 1463.054613] list_add double add: new=c06e52c0, prev=c06e52c0, next=c06d5f84.
[ 1463.061625] Modules linked in:
[ 1463.064623] CPU: 0 PID: 1 Comm: bash Tainted: G        W      3.18.0-rc5-next-20141121-00005-ga8fab06eab42-dirty #1022
[ 1463.075338] [<c0014e2c>] (unwind_backtrace) from [<c0011d80>] (show_stack+0x10/0x14)
[ 1463.083046] [<c0011d80>] (show_stack) from [<c048bb70>] (dump_stack+0x70/0xbc)
[ 1463.090236] [<c048bb70>] (dump_stack) from [<c00233d4>] (warn_slowpath_common+0x74/0xb0)
[ 1463.098295] [<c00233d4>] (warn_slowpath_common) from [<c00234a4>] (warn_slowpath_fmt+0x30/0x40)
[ 1463.106962] [<c00234a4>] (warn_slowpath_fmt) from [<c020fe80>] (__list_add+0x8c/0xc0)
[ 1463.114760] [<c020fe80>] (__list_add) from [<c0282094>] (register_syscore_ops+0x30/0x3c)
[ 1463.122819] [<c0282094>] (register_syscore_ops) from [<c0392f20>] (exynos_audss_clk_probe+0x36c/0x460)
[ 1463.132091] [<c0392f20>] (exynos_audss_clk_probe) from [<c0283084>] (platform_drv_probe+0x48/0xa4)
[ 1463.141013] [<c0283084>] (platform_drv_probe) from [<c0281a14>] (driver_probe_device+0x13c/0x37c)
[ 1463.149852] [<c0281a14>] (driver_probe_device) from [<c0280560>] (bind_store+0x90/0xe0)
[ 1463.157822] [<c0280560>] (bind_store) from [<c027fd10>] (drv_attr_store+0x20/0x2c)
[ 1463.165363] [<c027fd10>] (drv_attr_store) from [<c0143898>] (sysfs_kf_write+0x4c/0x50)
[ 1463.173252] [<c0143898>] (sysfs_kf_write) from [<c0142c80>] (kernfs_fop_write+0xbc/0x198)
[ 1463.181395] [<c0142c80>] (kernfs_fop_write) from [<c00e2be0>] (vfs_write+0xa0/0x1a8)
[ 1463.189104] [<c00e2be0>] (vfs_write) from [<c00e2f00>] (SyS_write+0x40/0x8c)
[ 1463.196122] [<c00e2f00>] (SyS_write) from [<c000f2a0>] (ret_fast_syscall+0x0/0x48)
[ 1463.203655] ---[ end trace 08f6710c9bc8d8f3 ]---
[ 1463.208244] exynos-audss-clk 3810000.audss-clock-controller: setup completed

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Fixes: 1241ef94ccc3 ("clk: samsung: register audio subsystem clocks using common clock framework")
Cc: <stable@vger.kernel.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
9 years agoclk: samsung: exynos4: set parent of sclk_hdmiphy to hdmi
Andrzej Hajda [Mon, 24 Nov 2014 07:30:50 +0000 (08:30 +0100)]
clk: samsung: exynos4: set parent of sclk_hdmiphy to hdmi

sclk_hdmiphy clock is generated by HDMI-PHY and depends on hdmi gate clock.
The patch models this dependency using parent/child hirerarchy.

The patch fixes issue with system hangs during mixer device access, mixer uses
sclk_hdmiphy descendant clock.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
9 years agoclk: samsung: exynos4415: Fix build with PM_SLEEP disabled
Krzysztof Kozlowski [Mon, 1 Dec 2014 09:12:54 +0000 (10:12 +0100)]
clk: samsung: exynos4415: Fix build with PM_SLEEP disabled

Fix following build errors when PM_SLEEP is disabled (e.g. by disabling
SUSPEND and HIBERNATION):

drivers/clk/samsung/clk-exynos4415.c: In function ‘exynos4415_cmu_init’:
drivers/clk/samsung/clk-exynos4415.c:982:2: error: ‘exynos4415_ctx’ undeclared (first use in this function)
drivers/clk/samsung/clk-exynos4415.c:982:2: note: each undeclared identifier is reported only once for each function it appears in
drivers/clk/samsung/clk-exynos4415.c: In function ‘exynos4415_cmu_dmc_init’:
drivers/clk/samsung/clk-exynos4415.c:1123:2: error: ‘exynos4415_dmc_ctx’ undeclared (first use in this function)
make[3]: *** [drivers/clk/samsung/clk-exynos4415.o] Error 1

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
9 years agoclk: samsung: remove unnecessary inclusion of header files from clk.h
Pankaj Dubey [Mon, 29 Sep 2014 07:47:48 +0000 (13:17 +0530)]
clk: samsung: remove unnecessary inclusion of header files from clk.h

Let's remove unnecessary include of header files from clk.h and add
required one in clk.c

Signed-off-by: Pankaj Dubey <pankaj.dubey@samsung.com>
[s.nawrocki@samsung.com: dropped removal of '#include <linux/syscore_ops.h>']
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
9 years agoclk: samsung: remove unnecessary CONFIG_OF from clk.c
Pankaj Dubey [Mon, 29 Sep 2014 07:47:47 +0000 (13:17 +0530)]
clk: samsung: remove unnecessary CONFIG_OF from clk.c

Remove unnecessary CONFIG_OF from samsung/clk.c.

Signed-off-by: Pankaj Dubey <pankaj.dubey@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
9 years agoclk: samsung: Spelling s/bwtween/between/
Pankaj Dubey [Mon, 29 Sep 2014 07:47:46 +0000 (13:17 +0530)]
clk: samsung: Spelling s/bwtween/between/

Fix a typo in comment section of "struct samsung_clk_provider".

Signed-off-by: Pankaj Dubey <pankaj.dubey@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
9 years agoMerge tag 'ib-mfd-regulator-clk-v3.19' of git://git.kernel.org/pub/scm/linux/kernel...
Michael Turquette [Sat, 29 Nov 2014 05:11:10 +0000 (21:11 -0800)]
Merge tag 'ib-mfd-regulator-clk-v3.19' of git://git./linux/kernel/git/lee/mfd into clk-next

Immutable branch between MFD, Regulator and Clk, due for v3.19

9 years agoMerge tag 'v3.19-rockchip-clk2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Michael Turquette [Sat, 29 Nov 2014 05:00:16 +0000 (21:00 -0800)]
Merge tag 'v3.19-rockchip-clk2' of git://git./linux/kernel/git/mmind/linux-rockchip into clk-next

- clock phase setting capability for the rk3288 mmc clocks
- pll init to allow syncing to actual rate table values
- some more exported clocks
- fixes for some clocks (typos etc) all of them not yet used
  in actual drivers

9 years agoclk: rockchip: Add support for the mmc clock phases using the framework
Alexandru M Stan [Thu, 27 Nov 2014 01:30:27 +0000 (17:30 -0800)]
clk: rockchip: Add support for the mmc clock phases using the framework

This patch adds the 2 physical clocks for the mmc (drive and sample). They're
mostly there for the phase properties, but they also show the true clock
(by dividing by RK3288_MMC_CLKGEN_DIV).

The drive and sample phases are generated by dividing an upstream parent clock
by 2, this allows us to adjust the phase by 90 deg.

There's also an option to have up to 255 delay elements (40-80 picoseconds long).
This driver uses those elements (under the assumption that they're 60 ps long)
to generate approximate 22.5 degrees options. 67.5 (22.5*3) might be as high as
90 deg if the delay elements are as big as 80 ps, so a finer division (smaller
than 22.5) was not picked because the phase might not be monotonic anymore.

Suggested-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Alexandru M Stan <amstan@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoclk: rockchip: add bindings for the mmc clocks
Alexandru M Stan [Thu, 27 Nov 2014 01:30:26 +0000 (17:30 -0800)]
clk: rockchip: add bindings for the mmc clocks

These clocks represent the physical clocks (including phases) and they will
later be used for clock phase tuning.

Suggested-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Alexandru M Stan <amstan@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoclk: rockchip: rk3288 export i2s0_clkout for use in DT
Sonny Rao [Wed, 19 Nov 2014 07:15:19 +0000 (23:15 -0800)]
clk: rockchip: rk3288 export i2s0_clkout for use in DT

This exposes the clock that comes out of the i2s block which generally
goes to the audio codec.

Signed-off-by: Sonny Rao <sonnyrao@chromium.org>
[removed CLK_SET_RATE_PARENT from original patch]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoclk: rockchip: use clock ID for DMC (memory controller) on rk3288
Jeff Chen [Wed, 26 Nov 2014 00:13:03 +0000 (16:13 -0800)]
clk: rockchip: use clock ID for DMC (memory controller) on rk3288

The DMC clocks need to be turned off at runtime.  Use the newly
assigned clock IDs to export them.

Signed-off-by: Jeff Chen <cym@rock-chips.com>
[dianders: split into two patches; adjusted commit msg]
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoclk: rockchip: add binding ID for DMC (memory controller) clocks on rk3288
Jeff Chen [Wed, 26 Nov 2014 00:13:02 +0000 (16:13 -0800)]
clk: rockchip: add binding ID for DMC (memory controller) clocks on rk3288

The DMC clocks need to be turned off at runtime, so we should have IDs
so we can export them.

Signed-off-by: Jeff Chen <cym@rock-chips.com>
[dianders: split into two patches; adjusted commit msg]
Signed-off-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoclk: clk-s2mps11: fix semicolon.cocci warnings
kbuild test robot [Tue, 25 Nov 2014 23:45:12 +0000 (07:45 +0800)]
clk: clk-s2mps11: fix semicolon.cocci warnings

drivers/clk/clk-s2mps11.c:181:2-3: Unneeded semicolon

 Removes unneeded semicolon.

Generated by: scripts/coccinelle/misc/semicolon.cocci

CC: Tushar Behera <tushar.behera@linaro.org>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agomfd: s2mps11: Add binding documentation for Samsung S2MPS13 PMIC
Chanwoo Choi [Tue, 18 Nov 2014 08:59:43 +0000 (17:59 +0900)]
mfd: s2mps11: Add binding documentation for Samsung S2MPS13 PMIC

This patch adds the binding documentation for Samsung S2MPS13 PMIC
which is similiar with existing S2MPS14 PMIC. S2MPS13 has the different number
of regulators from S2MPS14 and RTC/Clock is the same with the S2MPS14.

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Acked-by: Sangbeom Kim <sbkim73@samsung.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
9 years agoclk: s2mps11: Add the support for S2MPS13 PMIC clock
Chanwoo Choi [Tue, 18 Nov 2014 08:59:41 +0000 (17:59 +0900)]
clk: s2mps11: Add the support for S2MPS13 PMIC clock

This patch adds the support for S2MPS13 PMIC clock which is same with existing
S2MPS14 RTC IP. But, S2MPS13 uses all of clocks (32khz_{ap|bt|cp}).

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Acked-by: Michael Turquette <mturquette@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
9 years agoregulator: s2mps11: Add support S2MPS13 regulator device
Chanwoo Choi [Tue, 18 Nov 2014 08:59:40 +0000 (17:59 +0900)]
regulator: s2mps11: Add support S2MPS13 regulator device

This patch adds S2MPS13 regulator device to existing S2MPS11 device driver.
The S2MPS13 has just different number of regulators from S2MPS14.
The S2MPS13 regulator device includes LDO[1-40] and BUCK[1-10].

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Acked-by: Sangbeom Kim <sbkim73@samsung.com>
Acked-by: Mark Brown <broonie@kernel.org>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
9 years agomfd: sec-core: Add support for S2MPS13 device
Chanwoo Choi [Tue, 18 Nov 2014 08:59:39 +0000 (17:59 +0900)]
mfd: sec-core: Add support for S2MPS13 device

This patch adds the support for Samsung S2MPS13 PMIC device to the sec-core MFD
driver. The S2MPS13 is very similar with existing S2MPS14 and includes PMIC/
RTC/CLOCK devices.

Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Acked-by: Sangbeom Kim <sbkim73@samsung.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
9 years agoclk: rockchip: add ROCKCHIP_PLL_SYNC_RATE flag to some plls
Heiko Stuebner [Thu, 20 Nov 2014 19:38:53 +0000 (20:38 +0100)]
clk: rockchip: add ROCKCHIP_PLL_SYNC_RATE flag to some plls

Add the new flag to gpll and cpll on rk3188 and similar and to
gpll, cpll and npll on rk3288.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
9 years agoclk: rockchip: add optional sync to pll rate parameters
Heiko Stuebner [Thu, 20 Nov 2014 19:38:52 +0000 (20:38 +0100)]
clk: rockchip: add optional sync to pll rate parameters

In some cases firmware brings up plls with different parameters than the ones
noted in the rate table for the specific frequency. These firmware-selected
parameters are worse than the tested ones in the pll rate tables but cannot
be changed by a simple clk_set_rate call when the rate stays the same.

Therefore add a ROCKCHIP_PLL_SYNC_RATE flag and implement an init callback
that checks the runtime-parameters against the matching rate table entry
and adjusts them to the table-ones if necessary.

If no rate table is set or the current rate does not match any rate-table
entry no changes are made.

Being able to limit this adjustment to specific plls is necessary to not
touch the ones supplying core components like the apll and dpll supplying
the armcores and dram.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
9 years agoclk: rockchip: setup pll_mux data earlier
Heiko Stuebner [Thu, 20 Nov 2014 19:38:51 +0000 (20:38 +0100)]
clk: rockchip: setup pll_mux data earlier

In some cases we might need to access the data of the pll mux before the actual
mux gets registered - like in the following patch adding an init-callback.

Therefore populate pll_mux before registering the core pll-clock.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
9 years agoclk: rockchip: add ability to specify pll-specific flags
Heiko Stuebner [Thu, 20 Nov 2014 19:38:50 +0000 (20:38 +0100)]
clk: rockchip: add ability to specify pll-specific flags

This adds a flag parameter to plls that allows us to create
special flags to tweak the behaviour of the plls if necessary.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Tested-by: Kever Yang <kever.yang@rock-chips.com>
9 years agoMerge tag 'sunxi-clocks-for-3.19' of https://git.kernel.org/pub/scm/linux/kernel...
Michael Turquette [Tue, 25 Nov 2014 02:08:53 +0000 (18:08 -0800)]
Merge tag 'sunxi-clocks-for-3.19' of https://git./linux/kernel/git/mripard/linux into clk-next

Allwinner Clocks additions for 3.19

A few patches that should go through the clock tree, mostly fixes, cleanups,
and new clocks additions to start to support the A80.

9 years agoMerge branch 'clk-fixes' into clk-next
Michael Turquette [Tue, 25 Nov 2014 01:45:33 +0000 (17:45 -0800)]
Merge branch 'clk-fixes' into clk-next

9 years agoclk: sunxi: gmac-tx-clk mux is not a CLK_MUX_INDEX_BIT mux
Hans de Goede [Sun, 16 Nov 2014 12:56:57 +0000 (13:56 +0100)]
clk: sunxi: gmac-tx-clk mux is not a CLK_MUX_INDEX_BIT mux

A CLK_MUX_INDEX_BIT mux has one bit per parent, but the sun7i-a20-gmac-clk
has 2 bits selecting between 3 possible parents using values of 0, 1, 2,
which makes it a regular mux which should not have CLK_MUX_INDEX_BIT set in
its flag.

However we do not support parent 1 (an external clock), so use a table to
select parent 0 or 2, which are the 2 parents we support.

Note this has not been causing any issues sofar, because we start with a
parent setting of parent 0, and only ever re-parent to parent 2 (for which
we use an index of 1 as we skip parent 1) and with CLK_MUX_INDEX_BIT set
we write a value of 2 for index 1.

Tested on both a cubietruck (which uses rgmii mode) as well as a cs908
(an a31s board which uses mii mode).

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
9 years agoclk: sunxi: Implement A31 PLL6 as a divs clock for 2x output
Chen-Yu Tsai [Wed, 12 Nov 2014 18:08:31 +0000 (02:08 +0800)]
clk: sunxi: Implement A31 PLL6 as a divs clock for 2x output

Some clock modules on the A31 use PLL6x2 as one of their inputs.
This patch changes the PLL6 implementation for A31 to a divs clock,
i.e. clock with multiple outputs that have different dividers.
The first output will be the normal PLL6 output, and the second
will be PLL6x2.

This patch fixes the PLL6 N factor in the clock driver, and removes
any /2 dividers in the PLL6 factors clock part. The N factor counts
from 1 to 32, mapping to values 0 to 31, as shown in the A31 manual.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
9 years agoclk: sunxi: Specify number of child clocks for divs clocks
Chen-Yu Tsai [Wed, 12 Nov 2014 18:08:30 +0000 (02:08 +0800)]
clk: sunxi: Specify number of child clocks for divs clocks

Currently sunxi_divs_clk_setup assumes the number of child clocks
to be the same as the number of clock-output-names, and a maximum
of SUNXI_DIVS_MAX_QTY child clocks.

On sun6i, PLL6 only has 1 child clock, but the parent would be used
as well, thereby also having it's own clock-output-names entry. This
results in an extra bogus clock being registered.

This patch adds an entry for the number of child clocks alongside
the data structures for them.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
9 years agoclk: sunxi: Removed unused/incorrect sun6i-a31-apb2-clk driver
Chen-Yu Tsai [Thu, 6 Nov 2014 03:40:32 +0000 (11:40 +0800)]
clk: sunxi: Removed unused/incorrect sun6i-a31-apb2-clk driver

This driver does not match the hardware, which is actually compatible
to sun4i-a10-apb1-clk. Since we've switch to the correct one, drop
this driver.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
9 years agoARM: dts: sunxi: Use sun4i-a10-apb1-clk for sun6i/sun8i apb2 clocks.
Chen-Yu Tsai [Thu, 6 Nov 2014 03:40:31 +0000 (11:40 +0800)]
ARM: dts: sunxi: Use sun4i-a10-apb1-clk for sun6i/sun8i apb2 clocks.

The apb2 clocks are actually the same as apb1 clocks on the other sunxi
platforms, hence compatible with "allwinner,sun4i-a10-apb1-clk".

Update the dtsi to use the new unified apb1 clk.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
9 years agoARM: dts: sunxi: unify APB1 clock
Emilio López [Thu, 6 Nov 2014 03:40:30 +0000 (11:40 +0800)]
ARM: dts: sunxi: unify APB1 clock

With the new factors infrastructure in place, we can unify apb1 and
apb1_mux as a single clock now.

Signed-off-by: Emilio López <emilio@elopez.com.ar>
[wens@csie.org: Change apb1 node label to "apb1"; reword commit title]
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
9 years agoclk: rockchip: fix rk3188 USB HSIC PHY clock divider
Julien CHAUVEAU [Fri, 21 Nov 2014 10:08:47 +0000 (11:08 +0100)]
clk: rockchip: fix rk3188 USB HSIC PHY clock divider

The USB HSIC PHY clock divider is set in the register RK2928_CLKSEL_CON(11).

Signed-off-by: Julien CHAUVEAU <julien.chauveau@neo-technologies.fr>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoclk: rockchip: fix clock gate for rk3188 spdif_pre
Julien CHAUVEAU [Fri, 21 Nov 2014 09:27:41 +0000 (10:27 +0100)]
clk: rockchip: fix clock gate for rk3188 spdif_pre

In rk3188 clock branches, spdif_pre gate was set to RK2928_CLKGATE_CON(13) bit 13.
This appears to be a copy-paste error because such a register does not exist.
We correct it to RK2928_CLKGATE_CON(0) and find out that the rk3188 spdif clock
is the same as the rk3066 spdif clock, so we move it to the common clock branches.

Signed-off-by: Julien CHAUVEAU <julien.chauveau@neo-technologies.fr>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoclk: hi3620: Move const initdata into correct code section
Bintian Wang [Mon, 10 Nov 2014 12:58:31 +0000 (20:58 +0800)]
clk: hi3620: Move const initdata into correct code section

Use __initconst instead of __initdata for constant init data.

Signed-off-by: Bintian Wang <bintian.wang@huawei.com>
Acked-by: Haojian Zhuang <haojian.zhuang@linaro.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk_mux: Fix set_parent doing the wrong thing when INDEX_BIT && index >= 3
Hans de Goede [Wed, 19 Nov 2014 13:48:59 +0000 (14:48 +0100)]
clk_mux: Fix set_parent doing the wrong thing when INDEX_BIT && index >= 3

If CLK_MUX_INDEX_BIT is set, then each bit turns on / off a single parent,
so theoretically multiple parents could be enabled at the same time, but in
practice only one bit should ever be 1. So to select parent 0, set
the register (*) to 0x01, to select parent 1 set it 0x02, parent 2, 0x04,
parent 3, 0x08, etc.

But the current code does:

                if (mux->flags & CLK_MUX_INDEX_BIT)
                        index = (1 << ffs(index));

Which means that:

For an input index of 0, ffs returns 0, so we set the register
to 0x01, ok.

For an input index of 1, ffs returns 1, so we set the register
to 0x02, ok.

For an input index of 2, ffs returns 2, so we set the register
to 0x04, ok.

For an input index of 3, ffs returns 1, so we set the register
to 0x02, not good!

The code should simply be:

                if (mux->flags & CLK_MUX_INDEX_BIT)
                        index = 1 << index;

Which always does the right thing, this commit fixes this.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoMerge branch 'for-v3.19/exynos-clk' of git://linuxtv.org/snawrocki/samsung into clk...
Michael Turquette [Wed, 19 Nov 2014 19:25:37 +0000 (11:25 -0800)]
Merge branch 'for-v3.19/exynos-clk' of git://linuxtv.org/snawrocki/samsung into clk-next-exynos

9 years agoclk: shmobile: Deprecate renesas,clock-indices
Geert Uytterhoeven [Mon, 10 Nov 2014 18:49:34 +0000 (19:49 +0100)]
clk: shmobile: Deprecate renesas,clock-indices

Commit 8e33f91a0b84ae19 ("clk: shmobile: clk-mstp: change to using
clock-indices") forgot to replace all occurrences of
"renesas,clock-indices" in the driver-specific binding documentation.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Simon Horman <horms+renesas@verge.net.au>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoMerge branch 'clk-next-shmobile' into clk-next
Michael Turquette [Tue, 18 Nov 2014 23:11:52 +0000 (15:11 -0800)]
Merge branch 'clk-next-shmobile' into clk-next

9 years agoclk: delete a local variable's repeated assignment
Zhen Lei [Fri, 14 Nov 2014 02:10:40 +0000 (10:10 +0800)]
clk: delete a local variable's repeated assignment

It's the same to the next statement, "ret = clk->parent". I think compiler will
optimize it, it's just not looking well.

Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoMerge tag 'v3.19-rockchip-clk1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Michael Turquette [Tue, 18 Nov 2014 22:00:44 +0000 (14:00 -0800)]
Merge tag 'v3.19-rockchip-clk1' of git://git./linux/kernel/git/mmind/linux-rockchip into clk-next-rockchip

- fixes for clock ordering/rate issues
- do not keep all clocks enabled anymore
- allow special pll rates for special cases

9 years agoclk: rockchip: fix parent clock for rk3188 hclk_lcdc1
Julien CHAUVEAU [Tue, 18 Nov 2014 11:10:43 +0000 (12:10 +0100)]
clk: rockchip: fix parent clock for rk3188 hclk_lcdc1

The parent clock for hclk_lcdc1 was set to aclk_cpu instead of hclk_cpu.

Signed-off-by: Julien CHAUVEAU <julien.chauveau@neo-technologies.fr>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoclk: pxa: keep clocks initialization separated per variant
Robert Jarzmik [Mon, 6 Oct 2014 23:07:59 +0000 (01:07 +0200)]
clk: pxa: keep clocks initialization separated per variant

Have each pxa variant (pxa25x, pxa27x, pxa3xx) have its own device-tree
clock initializing function, to be able to register its own specific
core clocks.

Apply that change specifically to pxa27x.

Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: pxa: declare init function and data __init
Robert Jarzmik [Mon, 6 Oct 2014 23:07:58 +0000 (01:07 +0200)]
clk: pxa: declare init function and data __init

As the clock descriptions are constant and only usefull at init time,
mark them as such by :
 - spliting clock description (desc) and clock private data (dynamic)
 - mark __initdata clock descriptions

This makes all the register and descriptions of the clocks to go after
kernel init phase.

Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: pxa: fix pxa27x CCCR bit usage
Robert Jarzmik [Mon, 6 Oct 2014 23:07:57 +0000 (01:07 +0200)]
clk: pxa: fix pxa27x CCCR bit usage

Trivial fix to check the A bit of CCCR for memory frequency
calculations, where the shift of the bit index was missing, triggering a
wrong calculation of memory frequency.

Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: add pxa25x clock drivers
Robert Jarzmik [Tue, 11 Nov 2014 14:53:25 +0000 (15:53 +0100)]
clk: add pxa25x clock drivers

Move pxa25x clock drivers from arch/arm/mach-pxa to driver/clk.
In the move :
 - convert to new clock framework legacy clocks
 - provide clocks as before for platform data based boards
 - provide clocks through devicetree with clk-pxa-dt

This is the preliminary step in the conversion. The remaining steps are
:
 - pxa3xx
 - once PXA is fully converted to device tree, if that happens,
   clk-pxa2* and clk-pxa3* should only hold the core clocks which cannot
   be described in devicetree.

Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Tested-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk-divider: Fix READ_ONLY when divider > 1
James Hogan [Fri, 14 Nov 2014 15:32:09 +0000 (15:32 +0000)]
clk-divider: Fix READ_ONLY when divider > 1

Commit 79c6ab509558 (clk: divider: add CLK_DIVIDER_READ_ONLY flag) in
v3.16 introduced the CLK_DIVIDER_READ_ONLY flag which caused the
recalc_rate() and round_rate() clock callbacks to be omitted.

However using this flag has the unfortunate side effect of causing the
clock recalculation code when a clock rate change is attempted to always
treat it as a pass-through clock, i.e. with a fixed divide of 1, which
may not be the case. Child clock rates are then recalculated using the
wrong parent rate.

Therefore instead of dropping the recalc_rate() and round_rate()
callbacks, alter clk_divider_bestdiv() to always report the current
divider as the best divider so that it is never altered.

For me the read only clock was the system clock, which divided the PLL
rate by 2, from which both the UART and the SPI clocks were divided.
Initial setting of the UART rate set it correctly, but when the SPI
clock was set, the other child clocks were miscalculated. The UART clock
was recalculated using the PLL rate as the parent rate, resulting in a
UART new_rate of double what it should be, and a UART which spewed forth
garbage when the rate changes were propagated.

Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Thomas Abraham <thomas.ab@samsung.com>
Cc: Tomasz Figa <t.figa@samsung.com>
Cc: Max Schwarz <max.schwarz@online.de>
Cc: <stable@vger.kernel.org> # v3.16+
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: qcom: Fix duplicate rbcpr clock name
Georgi Djakov [Fri, 10 Oct 2014 13:57:24 +0000 (16:57 +0300)]
clk: qcom: Fix duplicate rbcpr clock name

There is a duplication in a clock name for apq8084 platform that causes
the following warning: "RBCPR_CLK_SRC" redefined

Resolve this by adding a MMSS_ prefix to this clock and making its name
coherent with msm8974 platform.

Fixes: 2b46cd23a5a2 ("clk: qcom: Add APQ8084 Multimedia Clock Controller (MMCC) support")
Signed-off-by: Georgi Djakov <gdjakov@mm-sol.com>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: at91: usb: fix at91sam9x5 recalc, round and set rate
Boris Brezillon [Mon, 17 Nov 2014 13:16:56 +0000 (14:16 +0100)]
clk: at91: usb: fix at91sam9x5 recalc, round and set rate

First check for rate == 0 in set_rate and round_rate to avoid div by zero.
Then, in order to get the closest rate, round all divisions to the closest
result instead of rounding them down.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: at91: usb: fix at91rm9200 round and set rate
Boris Brezillon [Fri, 14 Nov 2014 18:54:49 +0000 (19:54 +0100)]
clk: at91: usb: fix at91rm9200 round and set rate

at91rm9200_clk_usb_set_rate might fail depending on the requested rate,
because the parent_rate / rate remainder is not necessarily zero.
Moreover, when rounding down the calculated rate we might alter the
divisor calculation and end up with an invalid divisor.

To solve those problems, accept a non zero remainder, and always round
division to the closest result.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Reported-by: Andreas Henriksson <andreas.henriksson@endian.se>
Tested-by: Andreas Henriksson <andreas.henriksson@endian.se>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: rockchip: fix clock select order for rk3288 usbphy480m_src
Kever Yang [Thu, 13 Nov 2014 08:11:49 +0000 (16:11 +0800)]
clk: rockchip: fix clock select order for rk3288 usbphy480m_src

According to rk3288 trm, the mux selector locate at bit[12:11]
of CRU_CLKSEL13_CON shows:
2'b00: select HOST0 USB pll clock (clk_otgphy1)
2'b01: select HOST1 USB pll clock (clk_otgphy2)
2'b10: select OTG USB pll clock   (clk_otgphy0)

The clock map is in Fig. 3-4 CRU Clock Architecture Diagram 3
- clk_otgphy0 -> USB PHY OTG
- clk_otgphy1 -> USB PHY host0
- clk_otgphy2 -> USB PHY host1

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoclk: rockchip: fix rk3288 clk_usbphy480m_gate bit location in register
Kever Yang [Thu, 13 Nov 2014 07:19:21 +0000 (15:19 +0800)]
clk: rockchip: fix rk3288 clk_usbphy480m_gate bit location in register

According to rk3288 trm, the clk_usbphy480m_gate is located at
bit 14 of CRU_CLKGATE5_CON register.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoMerge branch 'clk-next-mmp' into clk-next
Michael Turquette [Fri, 14 Nov 2014 01:20:39 +0000 (17:20 -0800)]
Merge branch 'clk-next-mmp' into clk-next

9 years agoMAINTAINERS: clk framework git tree moved to kernel.org
Michael Turquette [Fri, 14 Nov 2014 01:18:20 +0000 (17:18 -0800)]
MAINTAINERS: clk framework git tree moved to 

Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoARM: OMAP2+: hwmod: drop unnecessary list initialization
Felipe Balbi [Thu, 13 Nov 2014 16:36:55 +0000 (09:36 -0700)]
ARM: OMAP2+: hwmod: drop unnecessary list initialization

ml->node and sl->node are currently initialized
by means of INIT_LIST_HEAD(). That initialiation
is followed by a list_add() call.

Looking at what both these functions do we will have:

ml->node.next = &ml->node;
ml->node.prev = &ml->node;
oi->master->master_ports.next.prev = &ml->node;
ml->node.next = &oi->master->master_ports.next;
ml->node.prev = &oi->master->master_ports;
oi->master->master_ports.next = &ml->node;

from this, it's clear that both INIT_LIST_HEAD() calls
are unnecessary and can be safely removed.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
9 years agoARM: OMAP3+: DPLL: use determine_rate() and set_rate_and_parent()
Tero Kristo [Fri, 3 Oct 2014 13:57:14 +0000 (16:57 +0300)]
ARM: OMAP3+: DPLL: use determine_rate() and set_rate_and_parent()

Currently, DPLLs are hiding the gory details of switching parent
within set_rate, which confuses the common clock code and is wrong.
Fixed by applying the new determine_rate() and set_rate_and_parent()
functionality to any clock-ops previously using the broken approach.
This patch also removes the broken legacy code.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
9 years agoARM: OMAP3: clock: add support for dpll4_set_rate_and_parent
Tero Kristo [Fri, 3 Oct 2014 13:57:13 +0000 (16:57 +0300)]
ARM: OMAP3: clock: add support for dpll4_set_rate_and_parent

Expand the support of omap4 per-dpll to provide set_rate_and_parent.
This is required for proper behavior of clk_change_rate with
determine_rate support.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
9 years agoARM: OMAP4: clock: add support for determine_rate for omap4 regm4xen DPLL
Tero Kristo [Fri, 3 Oct 2014 13:57:12 +0000 (16:57 +0300)]
ARM: OMAP4: clock: add support for determine_rate for omap4 regm4xen DPLL

Similarly to OMAP3 noncore DPLL, the implementation of this DPLL clock
type is wrong. This patch adds basic functionality for determine_rate
for this clock type which will be taken into use in the patches following
later.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
9 years agoARM: OMAP3: clock: add new rate changing logic support for noncore DPLLs
Tero Kristo [Fri, 3 Oct 2014 13:57:11 +0000 (16:57 +0300)]
ARM: OMAP3: clock: add new rate changing logic support for noncore DPLLs

Currently, DPLL code hides the re-parenting within its internals, which
is wrong. This needs to be exposed to the common clock code via
determine_rate and set_rate_and_parent APIs. This patch adds support
for these, which will be taken into use in the following patches.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
9 years agoARM: OMAP3: clock: use clk_features flags for omap3 DPLL4 checks
Tero Kristo [Fri, 3 Oct 2014 13:57:10 +0000 (16:57 +0300)]
ARM: OMAP3: clock: use clk_features flags for omap3 DPLL4 checks

DPLL4 can't be reprogrammed on OMAP3430 ES1.0 due to hardware limitation.
Currently, the code does runtime omap_rev() check to see the chip it is
being executed on, instead, change this to use clk_features flags.
This avoids need for runtime omap_rev() checks.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Signed-off-by: Paul Walmsley <paul@pwsan.com>
9 years agoclk: rockchip: ensure HCLK_VIO2_H2P and PCLK_VIO2_H2P stay enabled
Dmitry Torokhov [Wed, 12 Nov 2014 21:38:45 +0000 (13:38 -0800)]
clk: rockchip: ensure HCLK_VIO2_H2P and PCLK_VIO2_H2P stay enabled

Currently there is no driver owning these clocks and they have to stay
up for the system to function properly, so let's mark them as
CLK_IGNORE_UNUSED.

Without this patch we have trouble with suspend/resume and we have
trouble turning the eDP back on if it ever idles off.

Signed-off-by: Dmitry Torokhov <dtor@chromium.org>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Reviewed-by: Kever Yang <kever.yang@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoMAINTAINERS: add Stephen Boyd as clk co-maintainer
Michael Turquette [Thu, 13 Nov 2014 00:43:47 +0000 (16:43 -0800)]
MAINTAINERS: add Stephen Boyd as clk co-maintainer

Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoarm: mmp: Make use of the DT supported clock
Chao Xie [Fri, 31 Oct 2014 02:13:53 +0000 (10:13 +0800)]
arm: mmp: Make use of the DT supported clock

Change the dtsi and dts file, soc initialization code to make
use of DT support clock.
So now in the code we do only need call of_clk_init to initialize
the clocks.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoarm: mmp: Make all the dts file to be compiled by Makefile
Chao Xie [Fri, 31 Oct 2014 02:13:52 +0000 (10:13 +0800)]
arm: mmp: Make all the dts file to be compiled by Makefile

Add items in arch/arm/boot/dt/Makefile to compile the dtb
for mach-mmp.
Change the dts and dtsi file to use #include instead of \include\

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: add mmp2 DT support for clock driver
Chao Xie [Fri, 31 Oct 2014 02:13:51 +0000 (10:13 +0800)]
clk: mmp: add mmp2 DT support for clock driver

It adds the DT support for mmp2 clock subsystem.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: add pxa910 DT support for clock driver
Chao Xie [Fri, 31 Oct 2014 02:13:50 +0000 (10:13 +0800)]
clk: mmp: add pxa910 DT support for clock driver

It adds the DT support for pxa910 clock subsystem.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: add pxa168 DT support for clock driver
Chao Xie [Fri, 31 Oct 2014 02:13:49 +0000 (10:13 +0800)]
clk: mmp: add pxa168 DT support for clock driver

It adds the DT support for pxa168 clock subsystem.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: add reset support
Chao Xie [Fri, 31 Oct 2014 02:13:48 +0000 (10:13 +0800)]
clk: mmp: add reset support

Some clock control regsiter has bit to reset the cotroller.
So before enable the clock, we need deassert the reset pin.
Make use of reset controller framework to export reset interface
for device drivers, then device driver can control the reset action.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: add basic support functions for DT support
Chao Xie [Fri, 31 Oct 2014 02:13:47 +0000 (10:13 +0800)]
clk: mmp: add basic support functions for DT support

In order to support DT for mmp SOC clocks, it defines
some basic APIs which are shared by all mmp SOC clock
units.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: add mmp private gate clock
Chao Xie [Fri, 31 Oct 2014 02:13:46 +0000 (10:13 +0800)]
clk: mmp: add mmp private gate clock

Some SOCes have this kind of the gate clock
1. There are some bits to control the gate not only one bit.
2. It is not always that "1" is to enable while "0" is to disable
   when write register.

So we have to define the "mask", "enable_val", "disable_val" for
this kind of gate clock.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: add clock type mix
Chao Xie [Fri, 31 Oct 2014 02:13:45 +0000 (10:13 +0800)]
clk: mmp: add clock type mix

The clock type mix is a kind of clock combines "div" and "mux".
This kind of clock can not allow to change div first then
mux or change mux first or div.
The reason is
1. Some clock has frequency change bit. Each time want to change
   the frequency, there are some operations based on this bit, and
   these operations are time-cost.
   Seperating div and mux change will make the process longer, and
   waste more time.
2. Seperting the div and mux may generate middle clock that the
   peripharals do not support. It may make the peripharals hang.

There are three kinds of this type of clock in all SOCes.
1. The clock has bit to trigger the frequency change.
2. Same as #1, but the operations for the bit is different
3. Do not have frequency change bit.

So this type of clock has implemented the callbacks
->determine_rate
->set_rate_and_parent
These callbacks can help to change the div and mux together.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: move definiton of mmp_clk_frac to clk.h
Chao Xie [Fri, 31 Oct 2014 02:13:44 +0000 (10:13 +0800)]
clk: mmp: move definiton of mmp_clk_frac to clk.h

Move the definition of structure of mmp_clk_frac to
clk.h.
So device tree support can use this structure.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: add init callback for clk-frac
Chao Xie [Fri, 31 Oct 2014 02:13:43 +0000 (10:13 +0800)]
clk: mmp: add init callback for clk-frac

For the clk-frac, we need to make sure that the initial
clock rate is one item of the table.
If it is not, we use the first item in the table by default.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: add spin lock for clk-frac
Chao Xie [Fri, 31 Oct 2014 02:13:42 +0000 (10:13 +0800)]
clk: mmp: add spin lock for clk-frac

The register used by clk-frac may be shared with
other clocks.
So it needs to use spin lock to protect the register
access.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: mmp: add prefix "mmp" for structures defined for clk-frac
Chao Xie [Fri, 31 Oct 2014 02:13:41 +0000 (10:13 +0800)]
clk: mmp: add prefix "mmp" for structures defined for clk-frac

The structures defined for clk-frac will be used out side
of clk-frac.c.
To avoid conflicts, add prefix "mmp" for these structures'
name.

Signed-off-by: Chao Xie <chao.xie@marvell.com>
Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Signed-off-by: Michael Turquette <mturquette@linaro.org>
9 years agoclk: shmobile: document DIV6 clock parent bindings
Ulrich Hecht [Fri, 7 Nov 2014 15:51:08 +0000 (16:51 +0100)]
clk: shmobile: document DIV6 clock parent bindings

Describes how to specify the parents for clocks with EXSRC bits.

Signed-off-by: Ulrich Hecht <ulrich.hecht+renesas@gmail.com>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
9 years agoclk: shmobile: div6: support selectable-input clocks
Ulrich Hecht [Fri, 7 Nov 2014 15:51:07 +0000 (16:51 +0100)]
clk: shmobile: div6: support selectable-input clocks

Support for setting the parent at initialization time based on the current
hardware configuration in DIV6 clocks with selectable parents as found in
the r8a73a4, r8a7740, sh73a0, and other SoCs.

Signed-off-by: Ulrich Hecht <ulrich.hecht+renesas@gmail.com>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
9 years agoclk: sunxi: unify APB1 clock
Emilio López [Thu, 6 Nov 2014 03:40:29 +0000 (11:40 +0800)]
clk: sunxi: unify APB1 clock

This commit unifies the APB1 mux with the APB1 clock, using the new
factors infrastructure.

Signed-off-by: Emilio López <emilio@elopez.com.ar>
[wens@csie.org: Add mux mask bits]
Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
9 years agoclk: rockchip: rk3288: add suspend and resume
Chris Zhong [Fri, 7 Nov 2014 13:49:33 +0000 (21:49 +0800)]
clk: rockchip: rk3288: add suspend and resume

save and restore some clks, which might be changed in suspend.

Signed-off-by: Tony Xie <xxx@rock-chips.com>
Signed-off-by: Chris Zhong <zyw@rock-chips.com>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Tested-by: Doug Anderson <dianders@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
9 years agoLinux 3.18-rc4 v3.18-rc4
Linus Torvalds [Sun, 9 Nov 2014 22:55:29 +0000 (14:55 -0800)]
Linux 3.18-rc4

9 years agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Sun, 9 Nov 2014 22:49:56 +0000 (14:49 -0800)]
Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fixes from Catalin Marinas:
 - enable bpf syscall for compat
 - cpu_suspend fix when checking the idle state type
 - defconfig update

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: defconfig: update defconfig for 3.18
  arm64: compat: Enable bpf syscall
  arm64: psci: fix cpu_suspend to check idle state type for index

9 years agoMerge tag 'armsoc-for-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Linus Torvalds [Sun, 9 Nov 2014 22:46:36 +0000 (14:46 -0800)]
Merge tag 'armsoc-for-rc4' of git://git./linux/kernel/git/arm/arm-soc

Pull ARM SoC fixes from Olof Johansson:
 "Another quiet week:

   - a fix to silence edma probe error on non-supported platforms from
     Arnd
   - a fix to enable the PL clock for Parallella, to make mainline
     usable with the SDK.
   - a somewhat verbose fix for the PLL clock tree on VF610
   - enabling of SD/MMC on one of the VF610-based boards (for testing)
   - a fix for i.MX where CONFIG_SPI used to be implicitly enabled and
     now needs to be added to the defconfig instead
   - another maintainer added for bcm2835: Lee Jones"

* tag 'armsoc-for-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ARM: dts: zynq: Enable PL clocks for Parallella
  dma: edma: move device registration to platform code
  ARM: dts: vf610: add SD node to cosmic dts
  MAINTAINERS: update bcm2835 entry
  ARM: imx: Fix the removal of CONFIG_SPI option
  ARM: imx: clk-vf610: define PLL's clock tree

9 years agoMerge branch 'devicetree/merge' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 9 Nov 2014 22:33:49 +0000 (14:33 -0800)]
Merge branch 'devicetree/merge' of git://git./linux/kernel/git/glikely/linux

Pull devicetree bugfix from Grant Likely:
 "One buffer overflow bug that shouldn't be left around"

* 'devicetree/merge' of git://git.kernel.org/pub/scm/linux/kernel/git/glikely/linux:
  of: Fix overflow bug in string property parsing functions

9 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux...
Linus Torvalds [Sun, 9 Nov 2014 22:30:24 +0000 (14:30 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mason/linux-btrfs

Pull btrfs fix from Chris Mason:
 "It's a one liner for an error cleanup path that leads to crashes"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  Btrfs: fix kfree on list_head in btrfs_lookup_csums_range error cleanup

9 years agoMerge tag 'driver-core-3.18-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 9 Nov 2014 22:11:58 +0000 (14:11 -0800)]
Merge tag 'driver-core-3.18-rc4' of git://git./linux/kernel/git/gregkh/driver-core

Pull driver core fixes from Greg KH:
 "Here are 3 tiny fixes for 3.18-rc4.

  One fixes up a long-stading race condition in the driver core for
  removing directories in /sys/devices/virtual/ and the other 2 fix up
  the wording of a new Kconfig option that was added in 3.18-rc1"

* tag 'driver-core-3.18-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  tiny: rename ENABLE_DEV_COREDUMP to ALLOW_DEV_COREDUMP
  tiny: reverse logic for DISABLE_DEV_COREDUMP
  sysfs: driver core: Fix glue dir race condition by gdp_mutex

9 years agoMerge tag 'staging-3.18-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sun, 9 Nov 2014 22:11:07 +0000 (14:11 -0800)]
Merge tag 'staging-3.18-rc4' of git://git./linux/kernel/git/gregkh/staging

Pull staging driver fixes from Greg KH:
 "Here are some staging/iio fixes for 3.18-rc4.

  Nothing major, just a few bugfixes of things that have been reported"

* tag 'staging-3.18-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging:iio:ade7758: Remove "raw" from channel name
  staging:iio:ade7758: Fix check if channels are enabled in prenable
  staging:iio:ade7758: Fix NULL pointer deref when enabling buffer
  iio: as3935: allocate correct iio_device size
  io: accel: kxcjk-1013: Fix iio_event_spec direction
  iio: tsl4531: Fix compiler error when CONFIG_PM_OPS is not defined
  iio: adc: mxs-lradc: Disable the clock on probe failure
  iio: st_sensors: Fix buffer copy
  staging:iio:ad5933: Drop "raw" from channel names
  staging:iio:ad5933: Fix NULL pointer deref when enabling buffer

9 years agoMerge tag 'tty-3.18-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Sun, 9 Nov 2014 22:07:30 +0000 (14:07 -0800)]
Merge tag 'tty-3.18-rc4' of git://git./linux/kernel/git/gregkh/tty

Pull tty/serial fixes from Greg KH:
 "Here are some tiny serial/tty fixes for 3.18-rc4 that resolve some
  reported issues"

* tag 'tty-3.18-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  tty: Fix pty master poll() after slave closes v2
  serial: of-serial: fix uninitialized kmalloc variable
  tty/vt: don't set font mappings on vc not supporting this
  tty: serial: 8250_mtk: Fix quot calculation
  tty: Prevent "read/write wait queue active!" log flooding
  tty: Fix high cpu load if tty is unreleaseable
  serial: Fix divide-by-zero fault in uart_get_divisor()

9 years agoMerge tag 'usb-3.18-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Sun, 9 Nov 2014 22:05:53 +0000 (14:05 -0800)]
Merge tag 'usb-3.18-rc4' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are some USB fixes for 3.18-rc4.

  Just a bunch of little fixes resolving reported issues and new device
  ids for existing drivers.  Full details are in the shortlog"

* tag 'usb-3.18-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (26 commits)
  USB: Update default usb-storage delay_use value in kernel-parameters.txt
  USB: cdc-acm: add quirk for control-line state requests
  phy: omap-usb2: Enable runtime PM of omap-usb2 phy properly
  USB: storage: Fix timeout in usb_stor_euscsi_init() and usb_stor_huawei_e220_init()
  USB: cdc-acm: only raise DTR on transitions from B0
  Revert "storage: Replace magic number with define in usb_stor_euscsi_init()"
  usb: core: notify disconnection when core detects disconnect
  usb: core: need to call usb_phy_notify_connect after device setup
  uas: Add US_FL_NO_ATA_1X quirk for 2 more Seagate models
  xhci: no switching back on non-ULT Haswell
  USB: quirks: enable device-qualifier quirk for yet another Elan touchscreen
  USB: quirks: enable device-qualifier quirk for another Elan touchscreen
  MAINTAINERS: Remove duplicate entry for usbip driver
  usb: storage: fix build warnings !CONFIG_PM
  usb: Remove references to non-existent PLAT_S5P symbol
  uas: Add NO_ATA_1X for VIA VL711 devices
  xhci: Disable streams on Asmedia 1042 xhci controllers
  USB: HWA: fix a warning message
  uas: Add US_FL_NO_ATA_1X quirk for 1 more Seagate model
  usb-storage: handle a skipped data phase
  ...

9 years agoARM: dts: zynq: Enable PL clocks for Parallella
Andreas Färber [Thu, 6 Nov 2014 17:22:10 +0000 (18:22 +0100)]
ARM: dts: zynq: Enable PL clocks for Parallella

The Parallella board comes with a U-Boot bootloader that loads one of
two predefined FPGA bitstreams before booting the kernel. Both define an
AXI interface to the on-board Epiphany processor.

Enable clocks FCLK0..FCLK3 for the Programmable Logic by default.

Otherwise accessing, e.g., the ESYSRESET register freezes the board,
as seen with the Epiphany SDK tools e-reset and e-hw-rev, using /dev/mem.

Cc: <stable@vger.kernel.org> # 3.17.x
Signed-off-by: Andreas Färber <afaerber@suse.de>
Acked-by: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Olof Johansson <olof@lixom.net>
9 years agoMerge branch 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Sat, 8 Nov 2014 17:32:29 +0000 (09:32 -0800)]
Merge branch 'i2c/for-current' of git://git./linux/kernel/git/wsa/linux

Pull i2c bugfixes from Wolfram Sang:
 "One bigger cleanup (FSF address removal) and two bugfixes for I2C"

* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: core: Dispose OF IRQ mapping at client removal time
  i2c: at91: don't account as iowait
  i2c: remove FSF address

9 years agoMerge branch 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 8 Nov 2014 16:47:16 +0000 (08:47 -0800)]
Merge branch 'irq-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull irq fixes from Thomas Gleixner:
 "Two fixlets for the armada SoC interrupt controller"

* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  irqchip: armada-370-xp: Fix MPIC interrupt handling
  irqchip: armada-370-xp: Fix MSI interrupt handling

9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media
Linus Torvalds [Sat, 8 Nov 2014 16:45:20 +0000 (08:45 -0800)]
Merge git://git./linux/kernel/git/mchehab/linux-media

Pull media fixes from Mauro Carvalho Chehab:
 "For:
   - some regression fixes at the Remote Controller core and imon driver
   - a build fix for certain randconfigs with ir-hix5hd2
   - don't feed power to satellite system at ds3000 driver init

  It also contains some fixes for drivers added for Kernel 3.18:
   - some fixes at the new ISDB-S driver, and the corresponding bits to
     fix some descriptors for this Japanese TV standard at the DVB core
   - two warning cleanups for sp2 driver if PM is disabled
   - change the default mode for the new vivid driver"

* git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
  [media] sp2: sp2_init() can be static
  [media] dvb:tc90522: fix always-false expression
  [media] dvb-core: set default properties of ISDB-S
  [media] dvb:tc90522: fix stats report
  [media] vivid: default to single planar device instances
  [media] imon: fix other RC type protocol support
  [media] ir-hix5hd2 fix build warning
  [media] ds3000: fix LNB supply voltage on Tevii S480 on initialization
  [media] rc5-decoder: BZ#85721: Fix RC5-SZ decoding
  [media] rc-core: fix protocol_change regression in ir_raw_event_register

9 years agoMerge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus
Linus Torvalds [Sat, 8 Nov 2014 02:08:02 +0000 (18:08 -0800)]
Merge branch 'upstream' of git://git.linux-mips.org/ralf/upstream-linus

Pull MIPS updates from Ralf Baechle:
 "This weeks' round of MIPS bug fixes for 3.18:

   - wire up the bpf syscall
   - fix TLB dump output for R3000 class TLBs
   - fix strnlen_user return value if no NUL character was found.
   - fix build with binutils 2.24.51+.  While there is no binutils 2.25
     release yet, toolchains derived from binutils 2.24.51+ are already
     in common use.
   - the Octeon GPIO code forgot to offline GPIO IRQs.
   - fix build error for XLP.
   - fix possible BUG assertion with EVA for CMA"

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
  MIPS: Fix build with binutils 2.24.51+
  MIPS: R3000: Fix debug output for Virtual page number
  MIPS: Fix strnlen_user() return value in case of overlong strings.
  MIPS: CMA: Do not reserve memory if not required
  MIPS: Wire up bpf syscall.
  MIPS/Xlp: Remove the dead function destroy_irq() to fix build error
  MIPS: Octeon: Make Octeon GPIO IRQ chip CPU hotplug-aware