Skip to contentSkip to conversation

ARM: OMAP: clock data conversion to DT

PATCH Superseded Tero Kristo 5 participants 68 messages Sep 25, 2013 – Oct 14, 2013

Subsystems (8)

v7 of 8 Superseded Latest is v8 →

lore
Share & tools
lore
#0 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Hi all,

Version 7 contains following high level changes:
- Dropped support for basic bindings from Mike Turquette, instead using vendor specific bindings for all clocks
- Mux clock + divider clock vendor specific bindings get rid of use of the bit-masks, instead these are generated runtime based on parent clock / divider min/max info, see individual patches for details
- Added support for dummy_ck nodes, was missing from previous version
- Fixed support for omap3630
- Added support for new clock node type: ti,mux-gate-clock (using composite clock type)
- OMAP4 / OMAP5 only build fixes (compiler flag checks added to some files)
- most of the of_omap_* calls renamed to of_ti_*
- Rebased on top of v3.12-rc2

Some detailed comments about patch level changes (if no mention, no major changes):
- Patch #1:
* removed use of reg-names property, instead registers mapped using compatible string
* removed ti,modes property, instead uses DPLL mode flags now
* separated AM33XX DPLLs under their own compatible strings
- Patch #4 & #5: new patches
- Patch #6: merged basic gate support from Mike to this patch as vendor specific gate clock support
- Patch #9 & #10: new patches
- Patch #11: dummy clocks added, USB / ABE DPLL setup ordering changed
- Patch #14: dummy clocks added
- Patch #20: dropped usage of reg-names property
- Patch #21: dummy clocks added
- Patch #22 & #23: new patches
- Patch #30: AM35XX clock data added to this patch
- Patch #31:
* dummy clocks added
* added missing 3630 clocks

Test branch available here:
https://github.com/t-kristo/linux-pm.git
branch: mainline-3.12-rc2-ti-dt-clks-v7

Testing done:
- am335x-bone : boot only
- omap5-sevm : boot only
- dra7-evm : boot only
- omap3-beagle : boot + suspend/resume (ret + off)
- omap4-panda-es : boot + suspend/resume

Nishanth has also been doing some tests with omap3-beagle-xm with some WIP versions of this branch, but not with this latest one. Nishanth, maybe you can provide test results to this one also?

-Tero

#1 lore raw

PATCH v7 1/36 CLK: TI: Add DPLL clock support

The OMAP clock driver now supports DPLL clock type. This patch also adds support for DT DPLL nodes.

---
 .../devicetree/bindings/clock/ti/dpll.txt          |   80 +++
 arch/arm/mach-omap2/clock.h                        |  144 +----
 arch/arm/mach-omap2/clock3xxx.h                    |    2 -
 drivers/clk/Makefile                               |    1 +
 drivers/clk/ti/Makefile                            |    3 +
 drivers/clk/ti/dpll.c                              |  580 ++++++++++++++++++++
 include/linux/clk/ti.h                             |  164 ++++++
 7 files changed, 829 insertions(+), 145 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/clock/ti/dpll.txt
 create mode 100644 drivers/clk/ti/Makefile
 create mode 100644 drivers/clk/ti/dpll.c
 create mode 100644 include/linux/clk/ti.h
Documentation/devicetree/bindings/clock/ti/dpll.txt new file +80 -0
new file mode 100644
index 0000000..993879f
@@ -0,0 +1,80 @@
+Binding for Texas Instruments DPLL clock.
+
+This binding uses the common clock binding[1].  It assumes a
+register-mapped DPLL with usually two selectable input clocks
+(reference clock and bypass clock), with digital phase locked
+loop logic for multiplying the input clock to a desired output
+clock. This clock also typically supports different operation
+modes (locked, low power stop etc.) This binding has several
+sub-types, which effectively result in slightly different setup
+for the actual DPLL clock.
+
+[1] Documentation/devicetree/bindings/clock/clock-bindings.txt
+
+Required properties:
+- compatible : shall be one of:
+		"ti,omap3-dpll-clock",
+		"ti,omap3-dpll-core-clock",
+		"ti,omap3-dpll-per-clock",
+		"ti,omap3-dpll-per-j-type-clock",
+		"ti,omap4-dpll-clock",
+		"ti,omap4-dpll-x2-clock",
+		"ti,omap4-dpll-core-clock",
+		"ti,omap4-dpll-m4xen-clock",
+		"ti,omap4-dpll-j-type-clock",
+		"ti,am3-dpll-no-gate-clock",
+		"ti,am3-dpll-no-gate-j-type-clock",
+		"ti,am3-dpll-clock",
+		"ti,am3-dpll-core-clock",
+		"ti,am3-dpll-x2-clock",
+
+- #clock-cells : from common clock binding; shall be set to 0.
+- clocks : link phandles of parent clocks, first entry lists reference clock
+  and second entry bypass clock
+- reg : address and length of the register set for controlling the DPLL.
+  Registers are listed in following order:
+	"control" - contains the control register base address
+	"idlest" - contains the idle status register base address
+	"autoidle" - contains the autoidle register base address
+	"mult-div1" - contains the multiplier / divider register base address
+  ti,am3-* dpll types list the registers in the same order, except "autoidle"
+  register is left out as this hardware does not have it, e.g.:
+	reg = <0x4000 0x4>, <0x5000 0x4>, <0x6000 0x4>;
+  results in following register map:
+	0x4000 - control
+	0x5000 - idlest
+	0x6000 - mult-div1
+
+Optional properties:
+- DPLL mode setting - defining any one or more of the following overrides
+  default setting.
+	- ti,low-power-stop : DPLL supports low power stop mode, gating output
+	- ti,low-power-bypass : DPLL output matches rate of parent bypass clock
+	- ti,lock : DPLL locks in programmed rate
+
+Examples:
+	dpll_core_ck: dpll_core_ck at 44e00490 {
+		#clock-cells = <0>;
+		compatible = "ti,omap4-dpll-core-clock";
+		clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+		reg = <0x44e00490 0x4>, <0x44e0045c 0x4>, <0x0 0x4>,
+				<0x44e00468 0x4>;
+	};
+
+	dpll2_ck: dpll2_ck at 48004004 {
+		#clock-cells = <0>;
+		compatible = "ti,omap3-dpll-clock";
+		clocks = <&sys_ck>, <&dpll2_fck>;
+		ti,low-power-stop;
+		ti,low-power-bypass;
+		ti,lock;
+		reg = <0x48004004 0x4>, <0x48004024 0x4>, <0x48004034 0x4>,
+			<0x48004040 0x4>;
+	};
+
+	dpll_core_ck: dpll_core_ck at 44e00490 {
+		#clock-cells = <0>;
+		compatible = "ti,am3-dpll-core-clock";
+		clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+		reg = <0x44e00490 0x4>, <0x44e0045c 0x4>, <0x44e00468 0x4>;
+	};
arch/arm/mach-omap2/clock.h +1 -143

195 lines not on this page · raw message

arch/arm/mach-omap2/clock3xxx.h +0 -2
index 8cd4b0a..dab90e2 100644
@@ -9,8 +9,6 @@
 #define __ARCH_ARM_MACH_OMAP2_CLOCK3XXX_H
 
 int omap3xxx_clk_init(void);
-int omap3_dpll4_set_rate(struct clk_hw *clk, unsigned long rate,
-					unsigned long parent_rate);
 int omap3_core_dpll_m2_set_rate(struct clk_hw *clk, unsigned long rate,
 					unsigned long parent_rate);
 void omap3_clk_lock_dpll5(void);
drivers/clk/Makefile +1 -0
index 7b11106..981e14ac 100644
@@ -32,6 +32,7 @@ obj-$(CONFIG_ARCH_VT8500)	+= clk-vt8500.o
 obj-$(CONFIG_ARCH_ZYNQ)		+= zynq/
 obj-$(CONFIG_ARCH_TEGRA)	+= tegra/
 obj-$(CONFIG_PLAT_SAMSUNG)	+= samsung/
+obj-$(CONFIG_ARCH_OMAP)		+= ti/
 
 obj-$(CONFIG_X86)		+= x86/
 
drivers/clk/ti/Makefile new file +3 -0
new file mode 100644
index 0000000..93177987
@@ -0,0 +1,3 @@
+ifneq ($(CONFIG_OF),)
+obj-y					+= dpll.o
+endif
drivers/clk/ti/dpll.c new file +580 -0

580 lines not on this page · raw message

include/linux/clk/ti.h new file +164 -0

164 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#2 lore raw

PATCH v7 2/36 CLK: TI: add DT alias clock registration mechanism

In reply to Tero Kristo

Some devices require their clocks to be available with a specific dev-id con-id mapping. With DT, the clocks can be found by default only with their name, or alternatively through the device node of the consumer. With drivers, that don't support DT fully yet, add mechanism to register specific clock names.

---
 drivers/clk/ti/Makefile |    2 +-
 drivers/clk/ti/clk.c    |   52 +++++++++++++++++++++++++++++++++++++++++++++++
 include/linux/clk/ti.h  |   23 +++++++++++++++++++++
 3 files changed, 76 insertions(+), 1 deletion(-)
 create mode 100644 drivers/clk/ti/clk.c
drivers/clk/ti/Makefile +1 -1
index 93177987..05af5d8 100644
@@ -1,3 +1,3 @@
 ifneq ($(CONFIG_OF),)
-obj-y					+= dpll.o
+obj-y					+= clk.o dpll.o
 endif
drivers/clk/ti/clk.c new file +52 -0

52 lines not on this page · raw message

include/linux/clk/ti.h +23 -0
index 385384a..c89574e 100644
@@ -15,6 +15,8 @@
 #ifndef __LINUX_CLK_TI_H__
 #define __LINUX_CLK_TI_H__
 
+#include <linux/clkdev.h>
+
 /**
  * struct dpll_data - DPLL registers and integration data
  * @mult_div1_reg: register containing the DPLL M and N bitfields
@@ -138,6 +140,25 @@ struct clk_hw_omap {
 /* DPLL Type and DCO Selection Flags */
 #define DPLL_J_TYPE		0x1
 
+/**
+ * struct ti_dt_clk - OMAP DT clock alias declarations
+ * @lk: clock lookup definition
+ * @node_name: clock DT node to map to
+ */
+struct ti_dt_clk {
+	struct clk_lookup		lk;
+	char				*node_name;
+};
+
+#define DT_CLK(dev, con, name)		\
+	{				\
+		.lk = {			\
+			.dev_id = dev,	\
+			.con_id = con,	\
+		},			\
+		.node_name = name,	\
+	}
+
 void omap2_init_clk_hw_omap_clocks(struct clk *clk);
 int omap3_noncore_dpll_enable(struct clk_hw *hw);
 void omap3_noncore_dpll_disable(struct clk_hw *hw);
@@ -158,6 +179,8 @@ unsigned long omap3_clkoutx2_recalc(struct clk_hw *hw,
 int omap3_dpll4_set_rate(struct clk_hw *clk, unsigned long rate,
 			 unsigned long parent_rate);
 
+void ti_dt_clocks_register(struct ti_dt_clk *oclks);
+
 extern const struct clk_hw_omap_ops clkhwops_omap3_dpll;
 extern const struct clk_hw_omap_ops clkhwops_omap4_dpllmx;
 
Signature · 1 line
1.7.9.5
#3 lore raw

PATCH v7 3/36 CLK: TI: add autoidle support

In reply to Tero Kristo

TI clk driver now routes some of the basic clocks through own registration routine to allow autoidle support. This routine just checks a couple of device node properties and adds autoidle support if required, and just passes the registration forward to basic clocks.

---
 arch/arm/mach-omap2/clock.c |    6 +++
 drivers/clk/ti/Makefile     |    2 +-
 drivers/clk/ti/autoidle.c   |  107 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/clk/ti.h      |    9 ++++
 4 files changed, 123 insertions(+), 1 deletion(-)
 create mode 100644 drivers/clk/ti/autoidle.c
arch/arm/mach-omap2/clock.c +6 -0
index 0c38ca9..223f432b 100644
@@ -520,6 +520,9 @@ int omap2_clk_enable_autoidle_all(void)
 	list_for_each_entry(c, &clk_hw_omap_clocks, node)
 		if (c->ops && c->ops->allow_idle)
 			c->ops->allow_idle(c);
+
+	of_ti_clk_allow_autoidle_all();
+
 	return 0;
 }
 
@@ -539,6 +542,9 @@ int omap2_clk_disable_autoidle_all(void)
 	list_for_each_entry(c, &clk_hw_omap_clocks, node)
 		if (c->ops && c->ops->deny_idle)
 			c->ops->deny_idle(c);
+
+	of_ti_clk_deny_autoidle_all();
+
 	return 0;
 }
 
drivers/clk/ti/Makefile +1 -1
index 05af5d8..533efb4 100644
@@ -1,3 +1,3 @@
 ifneq ($(CONFIG_OF),)
-obj-y					+= clk.o dpll.o
+obj-y					+= clk.o dpll.o autoidle.o
 endif
drivers/clk/ti/autoidle.c new file +107 -0

107 lines not on this page · raw message

include/linux/clk/ti.h +9 -0
index c89574e..6d552c1 100644
@@ -180,6 +180,15 @@ int omap3_dpll4_set_rate(struct clk_hw *clk, unsigned long rate,
 			 unsigned long parent_rate);
 
 void ti_dt_clocks_register(struct ti_dt_clk *oclks);
+void of_ti_autoidle_setup(struct device_node *node);
+
+#ifdef CONFIG_OF
+void of_ti_clk_allow_autoidle_all(void);
+void of_ti_clk_deny_autoidle_all(void);
+#else
+static inline void of_ti_clk_allow_autoidle_all(void) { }
+static inline void of_ti_clk_deny_autoidle_all(void) { }
+#endif
 
 extern const struct clk_hw_omap_ops clkhwops_omap3_dpll;
 extern const struct clk_hw_omap_ops clkhwops_omap4_dpllmx;
Signature · 1 line
1.7.9.5
#4 lore raw

PATCH v7 4/36 CLK: ti: add support for ti divider-clock

In reply to Tero Kristo

This patch adds support for TI divider clock binding, which simply uses the basic clock divider to provide the features needed.

---
 .../devicetree/bindings/clock/ti/divider.txt       |   85 ++++++++++
 drivers/clk/ti/Makefile                            |    2 +-
 drivers/clk/ti/divider.c                           |  164 ++++++++++++++++++++
 include/linux/clk/ti.h                             |    2 +
 4 files changed, 252 insertions(+), 1 deletion(-)
 create mode 100644 Documentation/devicetree/bindings/clock/ti/divider.txt
 create mode 100644 drivers/clk/ti/divider.c
Documentation/devicetree/bindings/clock/ti/divider.txt new file +85 -0
new file mode 100644
index 0000000..c5eb1f3
@@ -0,0 +1,85 @@
+Binding for TI divider clock
+
+This binding uses the common clock binding[1].  It assumes a
+register-mapped adjustable clock rate divider that does not gate and has
+only one input clock or parent.  By default the value programmed into
+the register is one less than the actual divisor value.  E.g:
+
+register value		actual divisor value
+0			1
+1			2
+2			3
+
+This assumption may be modified by the following optional properties:
+
+ti,index-starts-at-one - valid divisor values start at 1, not the default
+of 0.  E.g:
+register value		actual divisor value
+1			1
+2			2
+3			3
+
+ti,index-power-of-two - valid divisor values are powers of two.  E.g:
+register value		actual divisor value
+0			1
+1			2
+2			4
+
+Additionally an array of valid dividers may be supplied like so:
+
+	dividers = <4>, <8>, <0>, <16>;
+
+Which will map the resulting values to a divisor table by their index:
+register value		actual divisor value
+0			4
+1			8
+2			<invalid divisor, skipped>
+3			16
+
+Any zero value in this array means the corresponding bit-value is invalid
+and must not be used.
+
+The binding must also provide the register to control the divider and
+unless the divider array is provided, min and max dividers. Optionally
+the number of bits to shift that mask, if necessary. If the shift value
+is missing it is the same as supplying a zero shift.
+
+Required properties:
+- compatible : shall be "divider-clock".
+- #clock-cells : from common clock binding; shall be set to 0.
+- clocks : link to phandle of parent clock
+- reg : base address for register controlling adjustable divider
+- ti,bit-mask : arbitrary bitmask for programming the adjustable divider
+
+Optional properties:
+- clock-output-names : from common clock binding.
+- ti,dividers : array of integers defining divisors
+- ti,bit-shift : number of bits to shift the divider value, defaults to 0
+- ti,min-div : min divisor for dividing the input clock rate, only
+  needed if the first divisor is offset from the default value (1)
+- ti,max-div : max divisor for dividing the input clock rate, only needed
+  if ti,dividers is not defined.
+- ti,index-starts-at-one : valid divisor programming starts at 1, not zero
+- ti,index-power-of-two : valid divisor programming must be a power of two
+- ti,autoidle-shift : bit shift of the autoidle enable bit for the clock
+- ti,autoidle-low : autoidle is enabled by setting the bit to 0
+- ti,set-rate-parent : clk_set_rate is propagated to parent
+
+Examples:
+dpll_usb_m2_ck: dpll_usb_m2_ck at 4a008190 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_usb_ck>;
+	ti,max-div = <127>;
+	reg = <0x4a008190 0x4>;
+	ti,index-starts-at-one;
+};
+
+aess_fclk: aess_fclk at 4a004528 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&abe_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a004528 0x4>;
+	ti,max-div = <2>;
+};
drivers/clk/ti/Makefile +1 -1
index 533efb4..72a410b 100644
@@ -1,3 +1,3 @@
 ifneq ($(CONFIG_OF),)
-obj-y					+= clk.o dpll.o autoidle.o
+obj-y					+= clk.o dpll.o autoidle.o divider.o
 endif
drivers/clk/ti/divider.c new file +164 -0

164 lines not on this page · raw message

include/linux/clk/ti.h +2 -0
index 6d552c1..7f58f11 100644
@@ -182,6 +182,8 @@ int omap3_dpll4_set_rate(struct clk_hw *clk, unsigned long rate,
 void ti_dt_clocks_register(struct ti_dt_clk *oclks);
 void of_ti_autoidle_setup(struct device_node *node);
 
+struct clk_div_table *ti_clk_get_div_table(struct device_node *node);
+
 #ifdef CONFIG_OF
 void of_ti_clk_allow_autoidle_all(void);
 void of_ti_clk_deny_autoidle_all(void);
Signature · 1 line
1.7.9.5
#5 lore raw

PATCH v7 5/36 clk: ti: add support for TI fixed factor clock

In reply to Tero Kristo

This behaves exactly in similar manner to basic fixed-factor-clock, but adds a few properties on top for handling clock hardware autoidling.

---
 .../bindings/clock/ti/fixed-factor-clock.txt       |   27 +++++++++
 drivers/clk/ti/Makefile                            |    3 +-
 drivers/clk/ti/fixed-factor.c                      |   62 ++++++++++++++++++++
 3 files changed, 91 insertions(+), 1 deletion(-)
 create mode 100644 Documentation/devicetree/bindings/clock/ti/fixed-factor-clock.txt
 create mode 100644 drivers/clk/ti/fixed-factor.c
Documentation/devicetree/bindings/clock/ti/fixed-factor-clock.txt new file +27 -0
new file mode 100644
index 0000000..e58762b
@@ -0,0 +1,27 @@
+Binding for TI fixed factor rate clock sources.
+
+This binding uses the common clock binding[1].
+
+[1] Documentation/devicetree/bindings/clock/clock-bindings.txt
+
+Required properties:
+- compatible : shall be "ti,fixed-factor-clock".
+- #clock-cells : from common clock binding; shall be set to 0.
+- ti,clock-div: fixed divider.
+- ti,clock-mult: fixed multiplier.
+- clocks: parent clock.
+
+Optional properties:
+- ti,autoidle-shift: bit shift of the autoidle enable bit for the clock
+- reg: base address for the autoidle register of this clock
+- ti,autoidle-low: autoidle is enabled by setting the bit to 0
+- ti,set-rate-parent: clk_set_rate is propagated to parent
+
+Example:
+	clock {
+		compatible = "ti,fixed-factor-clock";
+		clocks = <&parentclk>;
+		#clock-cells = <0>;
+		ti,clock-div = <2>;
+		ti,clock-mult = <1>;
+	};
drivers/clk/ti/Makefile +2 -1
index 72a410b..a9b986a 100644
@@ -1,3 +1,4 @@
 ifneq ($(CONFIG_OF),)
-obj-y					+= clk.o dpll.o autoidle.o divider.o
+obj-y					+= clk.o dpll.o autoidle.o divider.o \
+					   fixed-factor.o
 endif
drivers/clk/ti/fixed-factor.c new file +62 -0

62 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#6 lore raw

PATCH v7 6/36 CLK: TI: add support for OMAP gate clock

In reply to Tero Kristo

This patch adds support for TI specific gate clocks. These behave as basic gate-clock, but have different ops / hw-ops for controlling the actual gate, for example waiting until the clock is ready. Several sub-types are supported:
- ti,gate-clock: basic gate clock with default ops/hwops
- ti,clkdm-gate-clock: clockdomain level gate control
- ti,dss-gate-clock: gate clock with DSS specific hardware handling
- ti,am35xx-gate-clock: gate clock with AM35xx specific hardware handling
- ti,hsdiv-gate-clock: gate clock with OMAP36xx hardware errata handling

---
 .../devicetree/bindings/clock/ti/gate.txt          |   75 ++++++++
 arch/arm/mach-omap2/clock.h                        |   28 ---
 drivers/clk/ti/Makefile                            |    2 +-
 drivers/clk/ti/gate.c                              |  184 ++++++++++++++++++++
 include/linux/clk/ti.h                             |   35 ++++
 5 files changed, 295 insertions(+), 29 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/clock/ti/gate.txt
 create mode 100644 drivers/clk/ti/gate.c
Documentation/devicetree/bindings/clock/ti/gate.txt new file +75 -0
new file mode 100644
index 0000000..cf17534
@@ -0,0 +1,75 @@
+Binding for Texas Instruments gate clock.
+
+This binding uses the common clock binding[1]. This clock is
+quite much similar to the basic gate-clock [2], however,
+it supports a number of additional features. If no register
+is provided for this clock, the code assumes that a clockdomain
+will be controlled instead and the corresponding hw-ops for
+that is used.
+
+[1] Documentation/devicetree/bindings/clock/clock-bindings.txt
+[2] Documentation/devicetree/bindings/clock/gate-clock.txt
+[3] Documentation/devicetree/bindings/clock/ti/clockdomain.txt
+
+Required properties:
+- compatible : shall be one of:
+  "ti,gate-clock" - basic gate clock
+  "ti,wait-gate-clock" - gate clock which waits until clock is active before
+			 returning from clk_enable()
+  "ti,dss-gate-clock" - gate clock with DSS specific hardware handling
+  "ti,am35xx-gate-clock" - gate clock with AM35xx specific hardware handling
+  "ti,clkdm-gate-clock" - clockdomain gate clock, which derives its functional
+			  clock directly from a clockdomain, see [3] how
+			  to map clockdomains properly
+  "ti,hsdiv-gate-clock" - gate clock with OMAP36xx specific hardware handling,
+			  required for a hardware errata
+- #clock-cells : from common clock binding; shall be set to 0
+- clocks : link to phandle of parent clock
+- reg : base address for register controlling adjustable gate, not needed for
+	ti,clkdm-gate-clock type
+- ti,bit-shift : bit shift for programming the clock gate, not needed for
+		 ti,clkdm-gate-clock type
+
+Optional properties:
+- ti,set-bit-to-disable : inverts default gate programming. Setting the bit
+  gates the clock and clearing the bit ungates the clock.
+
+Examples:
+	mmchs2_fck: mmchs2_fck at 48004a00 {
+		#clock-cells = <0>;
+		compatible = "ti,gate-clock";
+		clocks = <&core_96m_fck>;
+		reg = <0x48004a00 0x4>;
+		ti,bit-shift = <25>;
+	};
+
+	dss1_alwon_fck_3430es2: dss1_alwon_fck_3430es2 at 48004e00 {
+		#clock-cells = <0>;
+		compatible = "ti,dss-gate-clock";
+		clocks = <&dpll4_m4x2_ck>;
+		reg = <0x48004e00 0x4>;
+		ti,bit-shift = <0>;
+	};
+
+	emac_ick: emac_ick at 4800259c {
+		#clock-cells = <0>;
+		compatible = "ti,am35xx-gate-clock";
+		clocks = <&ipss_ick>;
+		reg = <0x4800259c 0x4>;
+		ti,bit-shift = <1>;
+	};
+
+	emu_src_ck: emu_src_ck {
+		#clock-cells = <0>;
+		compatible = "ti,clkdm-gate-clock";
+		clocks = <&emu_src_mux_ck>;
+	};
+
+	dpll4_m2x2_ck: dpll4_m2x2_ck at 48004d00 {
+		#clock-cells = <0>;
+		compatible = "ti,hsdiv-gate-clock";
+		clocks = <&dpll4_m2x2_mul_ck>;
+		ti,bit-shift = <0x1b>;
+		reg = <0x48004d00 0x4>;
+		ti,set-bit-to-disable;
+	};
arch/arm/mach-omap2/clock.h +0 -28
index 079536a..f0b7218 100644
@@ -179,25 +179,6 @@ struct clksel {
 	const struct clksel_rate *rates;
 };
 
-/*
- * struct clk.flags possibilities
- *
- * XXX document the rest of the clock flags here
- *
- * CLOCK_CLKOUTX2: (OMAP4 only) DPLL CLKOUT and CLKOUTX2 GATE_CTRL
- *     bits share the same register.  This flag allows the
- *     omap4_dpllmx*() code to determine which GATE_CTRL bit field
- *     should be used.  This is a temporary solution - a better approach
- *     would be to associate clock type-specific data with the clock,
- *     similar to the struct dpll_data approach.
- */
-#define ENABLE_REG_32BIT	(1 << 0)	/* Use 32-bit access */
-#define CLOCK_IDLE_CONTROL	(1 << 1)
-#define CLOCK_NO_IDLE_PARENT	(1 << 2)
-#define ENABLE_ON_INIT		(1 << 3)	/* Enable upon framework init */
-#define INVERT_ENABLE		(1 << 4)	/* 0 enables, 1 disables */
-#define CLOCK_CLKOUTX2		(1 << 5)
-
 struct clk_hw_omap_ops {
 	void			(*find_idlest)(struct clk_hw_omap *oclk,
 					void __iomem **idlest_reg,
@@ -260,9 +241,6 @@ extern void omap2_clkt_iclk_deny_idle(struct clk_hw_omap *clk);
 
 unsigned long omap2_get_dpll_rate(struct clk_hw_omap *clk);
 
-int omap2_dflt_clk_enable(struct clk_hw *hw);
-void omap2_dflt_clk_disable(struct clk_hw *hw);
-int omap2_dflt_clk_is_enabled(struct clk_hw *hw);
 void omap2_clk_dflt_find_companion(struct clk_hw_omap *clk,
 				   void __iomem **other_reg,
 				   u8 *other_bit);
@@ -292,15 +270,12 @@ extern const struct clksel_rate dsp_ick_rates[];
 extern struct clk dummy_ck;
 
 extern const struct clk_hw_omap_ops clkhwops_iclk_wait;
-extern const struct clk_hw_omap_ops clkhwops_wait;
 extern const struct clk_hw_omap_ops clkhwops_iclk;
 extern const struct clk_hw_omap_ops clkhwops_omap3430es2_ssi_wait;
 extern const struct clk_hw_omap_ops clkhwops_omap3430es2_iclk_ssi_wait;
-extern const struct clk_hw_omap_ops clkhwops_omap3430es2_dss_usbhost_wait;
 extern const struct clk_hw_omap_ops clkhwops_omap3430es2_iclk_dss_usbhost_wait;
 extern const struct clk_hw_omap_ops clkhwops_omap3430es2_iclk_hsotgusb_wait;
 extern const struct clk_hw_omap_ops clkhwops_omap3430es2_hsotgusb_wait;
-extern const struct clk_hw_omap_ops clkhwops_am35xx_ipss_module_wait;
 extern const struct clk_hw_omap_ops clkhwops_am35xx_ipss_wait;
 extern const struct clk_hw_omap_ops clkhwops_apll54;
 extern const struct clk_hw_omap_ops clkhwops_apll96;
@@ -318,8 +293,5 @@ extern const struct clksel_rate div31_1to31_rates[];
 
 extern int am33xx_clk_init(void);
 
-extern int omap2_clkops_enable_clkdm(struct clk_hw *hw);
-extern void omap2_clkops_disable_clkdm(struct clk_hw *hw);
-
 extern void omap_clocks_register(struct omap_clk *oclks, int cnt);
 #endif
drivers/clk/ti/Makefile +1 -1
index a9b986a..63b1c45 100644
@@ -1,4 +1,4 @@
 ifneq ($(CONFIG_OF),)
 obj-y					+= clk.o dpll.o autoidle.o divider.o \
-					   fixed-factor.o
+					   fixed-factor.o gate.o
 endif
drivers/clk/ti/gate.c new file +184 -0

184 lines not on this page · raw message

include/linux/clk/ti.h +35 -0
index 7f58f11..e0a2d32 100644
@@ -132,6 +132,33 @@ struct clk_hw_omap {
 	const struct clk_hw_omap_ops	*ops;
 };
 
+/*
+ * struct clk_hw_omap.flags possibilities
+ *
+ * XXX document the rest of the clock flags here
+ *
+ * ENABLE_REG_32BIT: (OMAP1 only) clock control register must be accessed
+ *     with 32bit ops, by default OMAP1 uses 16bit ops.
+ * CLOCK_IDLE_CONTROL: (OMAP1 only) clock has autoidle support.
+ * CLOCK_NO_IDLE_PARENT: (OMAP1 only) when clock is enabled, its parent
+ *     clock is put to no-idle mode.
+ * ENABLE_ON_INIT: Clock is enabled on init.
+ * INVERT_ENABLE: By default, clock enable bit behavior is '1' enable, '0'
+ *     disable. This inverts the behavior making '0' enable and '1' disable.
+ * CLOCK_CLKOUTX2: (OMAP4 only) DPLL CLKOUT and CLKOUTX2 GATE_CTRL
+ *     bits share the same register.  This flag allows the
+ *     omap4_dpllmx*() code to determine which GATE_CTRL bit field
+ *     should be used.  This is a temporary solution - a better approach
+ *     would be to associate clock type-specific data with the clock,
+ *     similar to the struct dpll_data approach.
+ */
+#define ENABLE_REG_32BIT        (1 << 0)        /* Use 32-bit access */
+#define CLOCK_IDLE_CONTROL      (1 << 1)
+#define CLOCK_NO_IDLE_PARENT    (1 << 2)
+#define ENABLE_ON_INIT          (1 << 3)        /* Enable upon framework init */
+#define INVERT_ENABLE           (1 << 4)        /* 0 enables, 1 disables */
+#define CLOCK_CLKOUTX2          (1 << 5)
+
 /* CM_CLKEN_PLL*.EN* bit values - not all are available for every DPLL */
 #define DPLL_LOW_POWER_STOP	0x1
 #define DPLL_LOW_POWER_BYPASS	0x5
@@ -176,8 +203,13 @@ long omap2_dpll_round_rate(struct clk_hw *hw, unsigned long target_rate,
 void omap2_init_clk_clkdm(struct clk_hw *clk);
 unsigned long omap3_clkoutx2_recalc(struct clk_hw *hw,
 				    unsigned long parent_rate);
+int omap2_clkops_enable_clkdm(struct clk_hw *hw);
+void omap2_clkops_disable_clkdm(struct clk_hw *hw);
 int omap3_dpll4_set_rate(struct clk_hw *clk, unsigned long rate,
 			 unsigned long parent_rate);
+int omap2_dflt_clk_enable(struct clk_hw *hw);
+void omap2_dflt_clk_disable(struct clk_hw *hw);
+int omap2_dflt_clk_is_enabled(struct clk_hw *hw);
 
 void ti_dt_clocks_register(struct ti_dt_clk *oclks);
 void of_ti_autoidle_setup(struct device_node *node);
@@ -194,5 +226,8 @@ static inline void of_ti_clk_deny_autoidle_all(void) { }
 
 extern const struct clk_hw_omap_ops clkhwops_omap3_dpll;
 extern const struct clk_hw_omap_ops clkhwops_omap4_dpllmx;
+extern const struct clk_hw_omap_ops clkhwops_wait;
+extern const struct clk_hw_omap_ops clkhwops_omap3430es2_dss_usbhost_wait;
+extern const struct clk_hw_omap_ops clkhwops_am35xx_ipss_module_wait;
 
 #endif
Signature · 1 line
1.7.9.5
#7 lore raw

PATCH v7 7/36 CLK: TI: add support for clockdomain binding

In reply to Tero Kristo

Some OMAP clocks require knowledge about their parent clockdomain for book keeping purposes. This patch creates a new DT binding for TI clockdomains, which act as a collection of device clocks.

---
 .../devicetree/bindings/clock/ti/clockdomain.txt   |   19 ++++++++
 arch/arm/mach-omap2/clock.h                        |    1 -
 drivers/clk/ti/Makefile                            |    2 +-
 drivers/clk/ti/clockdomain.c                       |   46 ++++++++++++++++++++
 include/linux/clk/ti.h                             |    2 +
 5 files changed, 68 insertions(+), 2 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/clock/ti/clockdomain.txt
 create mode 100644 drivers/clk/ti/clockdomain.c
Documentation/devicetree/bindings/clock/ti/clockdomain.txt new file +19 -0
new file mode 100644
index 0000000..0dc5cd7
@@ -0,0 +1,19 @@
+Binding for Texas Instruments clockdomain.
+
+This binding uses the common clock binding[1]. Every clock on
+TI SoC belongs to one clockdomain, but software only needs this
+information for specific clocks which require their parent
+clockdomain to be controlled when the clock is enabled/disabled.
+
+[1] Documentation/devicetree/bindings/clock/clock-bindings.txt
+
+Required properties:
+- compatible : shall be "ti,clockdomain"
+- #clock-cells : from common clock binding; shall be set to 0.
+- clocks : link phandles of clocks within this domain
+
+Examples:
+	dss_clkdm: dss_clkdm {
+		compatible = "ti,clockdomain";
+		clocks = <&dss1_alwon_fck_3430es2>, <&dss_ick_3430es2>;
+	};
arch/arm/mach-omap2/clock.h +0 -1
index f0b7218..56d4664 100644
@@ -38,7 +38,6 @@ struct omap_clk {
 	}
 
 struct clockdomain;
-#define to_clk_hw_omap(_hw) container_of(_hw, struct clk_hw_omap, hw)
 
 #define DEFINE_STRUCT_CLK(_name, _parent_array_name, _clkops_name)	\
 	static struct clk _name = {				\
drivers/clk/ti/Makefile +1 -1
index 63b1c45..ce8bd06 100644
@@ -1,4 +1,4 @@
 ifneq ($(CONFIG_OF),)
 obj-y					+= clk.o dpll.o autoidle.o divider.o \
-					   fixed-factor.o gate.o
+					   fixed-factor.o gate.o clockdomain.o
 endif
drivers/clk/ti/clockdomain.c new file +46 -0

46 lines not on this page · raw message

include/linux/clk/ti.h +2 -0
index e0a2d32..e70a2eb 100644
@@ -186,6 +186,8 @@ struct ti_dt_clk {
 		.node_name = name,	\
 	}
 
+#define to_clk_hw_omap(_hw) container_of(_hw, struct clk_hw_omap, hw)
+
 void omap2_init_clk_hw_omap_clocks(struct clk *clk);
 int omap3_noncore_dpll_enable(struct clk_hw *hw);
 void omap3_noncore_dpll_disable(struct clk_hw *hw);
Signature · 1 line
1.7.9.5
#8 lore raw

PATCH v7 8/36 ARM: dts: omap4 clock data

In reply to Tero Kristo

This patch creates a unique node for each clock in the OMAP4 power, reset and clock manager (PRCM). OMAP443x and OMAP446x have slightly different clock tree which is taken into account in the data.

---
 arch/arm/boot/dts/omap443x-clocks.dtsi |   17 +
 arch/arm/boot/dts/omap443x.dtsi        |   20 +
 arch/arm/boot/dts/omap4460.dtsi        |   20 +
 arch/arm/boot/dts/omap446x-clocks.dtsi |   26 +
 arch/arm/boot/dts/omap44xx-clocks.dtsi | 1534 ++++++++++++++++++++++++++++++++
 5 files changed, 1617 insertions(+)
 create mode 100644 arch/arm/boot/dts/omap443x-clocks.dtsi
 create mode 100644 arch/arm/boot/dts/omap446x-clocks.dtsi
 create mode 100644 arch/arm/boot/dts/omap44xx-clocks.dtsi
arch/arm/boot/dts/omap443x-clocks.dtsi new file +17 -0
new file mode 100644
index 0000000..a538a07
@@ -0,0 +1,17 @@
+/*
+ * Device Tree Source for OMAP443x clock data
+ *
+ * Copyright (C) 2013 Texas Instruments, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+bandgap_fclk: bandgap_fclk at 4a307888 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a307888 0x4>;
+};
arch/arm/boot/dts/omap443x.dtsi +20 -0
index bcf455e..68464c5 100644
@@ -30,4 +30,24 @@
 		       0x4a00232C 0x4>;
 		compatible = "ti,omap4430-bandgap";
 	};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "omap44xx-clocks.dtsi"
+		/include/ "omap443x-clocks.dtsi"
+	};
+
+	clockdomains {
+		l3_init_clkdm: l3_init_clkdm {
+			compatible = "ti,clockdomain";
+			clocks = <&usb_host_fs_fck>, <&dpll_usb_ck>;
+		};
+
+		emu_sys_clkdm: emu_sys_clkdm {
+			compatible = "ti,clockdomain";
+			clocks = <&trace_clk_div_ck>;
+		};
+	};
 };
arch/arm/boot/dts/omap4460.dtsi +20 -0
index c2f0f39..f3c48d0 100644
@@ -38,4 +38,24 @@
 		interrupts = <0 126 IRQ_TYPE_LEVEL_HIGH>; /* talert */
 		gpios = <&gpio3 22 0>; /* tshut */
 	};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "omap44xx-clocks.dtsi"
+		/include/ "omap446x-clocks.dtsi"
+	};
+
+	clockdomains {
+		l3_init_clkdm: l3_init_clkdm {
+			compatible = "ti,clockdomain";
+			clocks = <&usb_host_fs_fck>, <&dpll_usb_ck>;
+		};
+
+		emu_sys_clkdm: emu_sys_clkdm {
+			compatible = "ti,clockdomain";
+			clocks = <&trace_clk_div_ck>;
+		};
+	};
 };
arch/arm/boot/dts/omap446x-clocks.dtsi new file +26 -0
new file mode 100644
index 0000000..96d5ddb
@@ -0,0 +1,26 @@
+/*
+ * Device Tree Source for OMAP446x clock data
+ *
+ * Copyright (C) 2013 Texas Instruments, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+div_ts_ck: div_ts_ck at 4a307888 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&l4_wkup_clk_mux_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a307888 0x4>;
+	ti,dividers = <8>, <16>, <32>;
+};
+
+bandgap_ts_fclk: bandgap_ts_fclk at 4a307888 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&div_ts_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a307888 0x4>;
+};
arch/arm/boot/dts/omap44xx-clocks.dtsi new file +1534 -0

1,535 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#9 lore raw

PATCH v7 9/36 clk: ti: add mux-gate clock support

In reply to Tero Kristo

This is a multipurpose clock node, which contains support for multiple sub-clocks. Uses composite clock type to implement the actual functionality.

---
 .../devicetree/bindings/clock/ti/mux-gate.txt      |   98 +++++++++
 drivers/clk/ti/Makefile                            |    3 +-
 drivers/clk/ti/mux-gate.c                          |  214 ++++++++++++++++++++
 include/linux/clk/ti.h                             |    1 +
 4 files changed, 315 insertions(+), 1 deletion(-)
 create mode 100644 Documentation/devicetree/bindings/clock/ti/mux-gate.txt
 create mode 100644 drivers/clk/ti/mux-gate.c
Documentation/devicetree/bindings/clock/ti/mux-gate.txt new file +98 -0
new file mode 100644
index 0000000..bda2b4c
@@ -0,0 +1,98 @@
+Binding for TI mux-gate clock.
+
+This binding uses the common clock binding[1]. It assumes a
+register-mapped composite clock with multiple different sub-types;
+
+a multiplexer clock with multiple input clock signals or parents, one
+of which can be selected as output, this behaves exactly as [2]
+
+an adjustable clock rate divider, this behaves exactly as [3]
+
+a gating function which can be used to enable and disable the output
+clock, this behaves exactly as [4]
+
+The binding must provide the register mapping to control the mux-gate.
+"reg-names" property is used to specify which sub-function are used
+for this composite clock, and to specify the corresponding control
+register addresses. Optional parameters are specified according to
+the sub-types used.
+
+[1] Documentation/devicetree/bindings/clock/clock-bindings.txt
+[2] Documentation/devicetree/bindings/clock/ti/mux.txt
+[3] Documentation/devicetree/bindings/clock/ti/divider.txt
+[4] Documentation/devicetree/bindings/clock/ti/gate.txt
+
+Required properties:
+- compatible : shall be one of:
+  "ti,mux-gate-clock" : mux-gate clock which waits until clock is active
+			before returning from clk_enable()
+  "ti,no-wait-mux-gate-clock" : mux-gate clock which does not wait for
+				clock activity
+  "ti,interface-mux-gate-clock" : mux-gate interface clock which waits until
+				  clock is active before returning from
+				  clk_enable()
+- #clock-cells : from common clock binding; shall be set to 0.
+- clocks : link phandles of parent clocks
+- reg : base address array for registers controlling the mux-gate
+  sub-functions. Ordered according to "reg-names" property.
+- reg-names : register names property, can have following values:
+	"gate-reg" : base address for gate function control register
+	"mux-reg" : base address for mux function control register
+	"div-reg" : base address for divider function control register
+  Any of these values can be left out, and the corresponding function
+  will not be present in the composite clock either.
+
+Optional properties:
+- ti,gate-bit-shift : number of bits to shift the enable bit for gating
+  function, defaults to 0 if not present
+- ti,mux-bit-shift : number of bits to shift the bit-field for mux
+  function, defaults to 0 if not present
+- ti,div-index-starts-at-one : min-div is mapped to bit-value 1, default 0
+- ti,max-div : maximum divider value for the divider function
+- ti,min-div : minimum divider value for the divider function, defaults to
+  1 if not present
+- ti,dividers : array of valid divider values for the clock
+
+Examples:
+
+dpll_core_m3x2_ck: dpll_core_m3x2_ck at 4a004134 {
+	#clock-cells = <0>;
+	compatible = "ti,no-wait-mux-gate-clock";
+	clocks = <&dpll_core_x2_ck>;
+	ti,max-div = <31>;
+	reg-names = "gate-reg", "div-reg";
+	reg = <0x4a004134 0x4>, <0x4a004134 0x4>;
+	ti,gate-bit-shift = <8>;
+	ti,div-index-starts-at-one;
+};
+
+auxclk2_src_ck: auxclk2_src_ck at 4a30a318 {
+	#clock-cells = <0>;
+	compatible = "ti,no-wait-mux-gate-clock";
+	clocks = <&sys_clkin_ck>, <&dpll_core_m3x2_ck>, <&dpll_per_m3x2_ck>;
+	reg-names = "gate-reg", "mux-reg";
+	reg = <0x4a30a318 0x4>, <0x4a30a318 0x4>;
+	ti,gate-bit-shift = <8>;
+	ti,mux-bit-shift = <1>;
+};
+
+gpt10_fck: gpt10_fck at 48004a00 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-gate-clock";
+	clocks = <&omap_32k_fck>, <&sys_ck>;
+	reg-names = "gate-reg", "mux-reg";
+	reg = <0x48004a00 0x4>, <0x48004a40 0x4>;
+	ti,gate-bit-shift = <11>;
+	ti,mux-bit-shift = <6>;
+};
+
+ssi_ssr_fck_3430es1: ssi_ssr_fck_3430es1 at 48004a00 {
+	#clock-cells = <0>;
+	compatible = "ti,no-wait-mux-gate-clock";
+	clocks = <&corex2_fck>;
+	ti,div-bit-shift = <8>;
+	ti,dividers = <0>, <1>, <2>, <3>, <4>, <0>, <6>, <0>, <8>;
+	reg-names = "gate-reg", "div-reg";
+	reg = <0x48004a00 0x4>, <0x48004a40 0x4>;
+	ti,gate-bit-shift = <0>;
+};
drivers/clk/ti/Makefile +2 -1
index ce8bd06..68d55c9 100644
@@ -1,4 +1,5 @@
 ifneq ($(CONFIG_OF),)
 obj-y					+= clk.o dpll.o autoidle.o divider.o \
-					   fixed-factor.o gate.o clockdomain.o
+					   fixed-factor.o gate.o clockdomain.o \
+					   mux-gate.o
 endif
drivers/clk/ti/mux-gate.c new file +214 -0

214 lines not on this page · raw message

include/linux/clk/ti.h +1 -0
index e70a2eb..ed4b36e 100644
@@ -229,6 +229,7 @@ static inline void of_ti_clk_deny_autoidle_all(void) { }
 extern const struct clk_hw_omap_ops clkhwops_omap3_dpll;
 extern const struct clk_hw_omap_ops clkhwops_omap4_dpllmx;
 extern const struct clk_hw_omap_ops clkhwops_wait;
+extern const struct clk_hw_omap_ops clkhwops_iclk_wait;
 extern const struct clk_hw_omap_ops clkhwops_omap3430es2_dss_usbhost_wait;
 extern const struct clk_hw_omap_ops clkhwops_am35xx_ipss_module_wait;
 
Signature · 1 line
1.7.9.5
#10 lore raw

PATCH v7 10/36 clk: ti: add support for basic mux clock

In reply to Tero Kristo

ti,mux-clock provides now a binding for basic mux support. This is just using the basic clock type.

---
 Documentation/devicetree/bindings/clock/ti/mux.txt |   65 +++++++++++++++
 drivers/clk/ti/Makefile                            |    2 +-
 drivers/clk/ti/mux.c                               |   83 ++++++++++++++++++++
 3 files changed, 149 insertions(+), 1 deletion(-)
 create mode 100644 Documentation/devicetree/bindings/clock/ti/mux.txt
 create mode 100644 drivers/clk/ti/mux.c
Documentation/devicetree/bindings/clock/ti/mux.txt new file +65 -0
new file mode 100644
index 0000000..5bb5fae
@@ -0,0 +1,65 @@
+Binding for TI mux clock.
+
+This binding uses the common clock binding[1].  It assumes a
+register-mapped multiplexer with multiple input clock signals or
+parents, one of which can be selected as output.  This clock does not
+gate or adjust the parent rate via a divider or multiplier.
+
+By default the "clocks" property lists the parents in the same order
+as they are programmed into the regster.  E.g:
+
+	clocks = <&foo_clock>, <&bar_clock>, <&baz_clock>;
+
+results in programming the register as follows:
+
+register value		selected parent clock
+0			foo_clock
+1			bar_clock
+2			baz_clock
+
+Some clock controller IPs do not allow a value of zero to be programmed
+into the register, instead indexing begins at 1.  The optional property
+"index-starts-at-one" modified the scheme as follows:
+
+register value		selected clock parent
+1			foo_clock
+2			bar_clock
+3			baz_clock
+
+The binding must provide the register to control the mux. Optionally
+the number of bits to shift the control field in the register can be
+supplied. If the shift value is missing it is the same as supplying
+a zero shift.
+
+[1] Documentation/devicetree/bindings/clock/clock-bindings.txt
+
+Required properties:
+- compatible : shall be "mux-clock".
+- #clock-cells : from common clock binding; shall be set to 0.
+- clocks : link phandles of parent clocks
+- reg : base address for register controlling adjustable mux
+
+Optional properties:
+- ti,bit-shift : number of bits to shift the bit-mask, defaults to
+  0 if not present
+- ti,index-starts-at-one : valid input select programming starts at 1, not
+  zero
+- ti,set-rate-parent : clk_set_rate is propagated to parent clock
+
+Examples:
+
+sys_clkin_ck: sys_clkin_ck at 4a306110 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&virt_12000000_ck>, <&virt_13000000_ck>, <&virt_16800000_ck>, <&virt_19200000_ck>, <&virt_26000000_ck>, <&virt_27000000_ck>, <&virt_38400000_ck>;
+	reg = <0x4a306110 0x4>;
+	ti,index-starts-at-one;
+};
+
+abe_dpll_bypass_clk_mux_ck: abe_dpll_bypass_clk_mux_ck at 4a306108 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clkin_ck>, <&sys_32k_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a306108 0x4>;
+};
drivers/clk/ti/Makefile +1 -1
index 68d55c9..f0986a5 100644
@@ -1,5 +1,5 @@
 ifneq ($(CONFIG_OF),)
 obj-y					+= clk.o dpll.o autoidle.o divider.o \
 					   fixed-factor.o gate.o clockdomain.o \
-					   mux-gate.o
+					   mux-gate.o mux.o
 endif
drivers/clk/ti/mux.c new file +83 -0

83 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#11 lore raw

PATCH v7 11/36 CLK: TI: add omap4 clock init file

In reply to Tero Kristo

clk-44xx.c now contains the clock init functionality for omap4, including DT clock registration and adding of static clkdev entries.

---
 arch/arm/mach-omap2/clock.h |    1 -
 drivers/clk/ti/clk-44xx.c   |  330 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/clk/ti.h      |    3 +
 3 files changed, 333 insertions(+), 1 deletion(-)
 create mode 100644 drivers/clk/ti/clk-44xx.c
arch/arm/mach-omap2/clock.h +0 -1
index 56d4664..106c5c2 100644
@@ -247,7 +247,6 @@ void omap2_clk_dflt_find_idlest(struct clk_hw_omap *clk,
 				void __iomem **idlest_reg,
 				u8 *idlest_bit, u8 *idlest_val);
 int omap2_clk_enable_autoidle_all(void);
-int omap2_clk_disable_autoidle_all(void);
 void omap2_clk_enable_init_clocks(const char **clk_names, u8 num_clocks);
 int omap2_clk_switch_mpurate_at_boot(const char *mpurate_ck_name);
 void omap2_clk_print_new_rates(const char *hfclkin_ck_name,
drivers/clk/ti/clk-44xx.c new file +330 -0

330 lines not on this page · raw message

include/linux/clk/ti.h +3 -0
index ed4b36e..c9a18b9 100644
@@ -207,6 +207,7 @@ unsigned long omap3_clkoutx2_recalc(struct clk_hw *hw,
 				    unsigned long parent_rate);
 int omap2_clkops_enable_clkdm(struct clk_hw *hw);
 void omap2_clkops_disable_clkdm(struct clk_hw *hw);
+int omap2_clk_disable_autoidle_all(void);
 int omap3_dpll4_set_rate(struct clk_hw *clk, unsigned long rate,
 			 unsigned long parent_rate);
 int omap2_dflt_clk_enable(struct clk_hw *hw);
@@ -218,6 +219,8 @@ void of_ti_autoidle_setup(struct device_node *node);
 
 struct clk_div_table *ti_clk_get_div_table(struct device_node *node);
 
+int omap4xxx_clk_init(void);
+
 #ifdef CONFIG_OF
 void of_ti_clk_allow_autoidle_all(void);
 void of_ti_clk_deny_autoidle_all(void);
Signature · 1 line
1.7.9.5
#12 lore raw

PATCH v7 12/36 ARM: OMAP4: remove old clock data and link in new clock init code

In reply to Tero Kristo

OMAP4 clocks have now been moved to DT, thus remove the old data file and use the new init code under drivers/clk/omap/clk-44xx.c.

---
 arch/arm/mach-omap2/Makefile          |    2 +-
 arch/arm/mach-omap2/cclock44xx_data.c | 1734 ---------------------------------
 drivers/clk/ti/Makefile               |    1 +
 3 files changed, 2 insertions(+), 1735 deletions(-)
 delete mode 100644 arch/arm/mach-omap2/cclock44xx_data.c
arch/arm/mach-omap2/Makefile +1 -1
index afb457c..f518360 100644
@@ -184,7 +184,7 @@ obj-$(CONFIG_ARCH_OMAP3)		+= clock34xx.o clkt34xx_dpll3m2.o
 obj-$(CONFIG_ARCH_OMAP3)		+= clock3517.o clock36xx.o
 obj-$(CONFIG_ARCH_OMAP3)		+= dpll3xxx.o cclock3xxx_data.o
 obj-$(CONFIG_ARCH_OMAP3)		+= clkt_iclk.o
-obj-$(CONFIG_ARCH_OMAP4)		+= $(clock-common) cclock44xx_data.o
+obj-$(CONFIG_ARCH_OMAP4)		+= $(clock-common)
 obj-$(CONFIG_ARCH_OMAP4)		+= dpll3xxx.o dpll44xx.o
 obj-$(CONFIG_SOC_AM33XX)		+= $(clock-common) dpll3xxx.o
 obj-$(CONFIG_SOC_AM33XX)		+= cclock33xx_data.o
arch/arm/mach-omap2/cclock44xx_data.c deleted +0 -1734

1,734 lines not on this page · raw message

drivers/clk/ti/Makefile +1 -0
index f0986a5..1ed6dd7 100644
@@ -2,4 +2,5 @@ ifneq ($(CONFIG_OF),)
 obj-y					+= clk.o dpll.o autoidle.o divider.o \
 					   fixed-factor.o gate.o clockdomain.o \
 					   mux-gate.o mux.o
+obj-$(CONFIG_ARCH_OMAP4)		+= clk-44xx.o
 endif
Signature · 1 line
1.7.9.5
#13 lore raw

PATCH v7 13/36 ARM: dts: omap5 clock data

In reply to Tero Kristo

This patch creates a unique node for each clock in the OMAP5 power, reset and clock manager (PRCM).

---
 arch/arm/boot/dts/omap5.dtsi           |   14 +
 arch/arm/boot/dts/omap54xx-clocks.dtsi | 1327 ++++++++++++++++++++++++++++++++
 2 files changed, 1341 insertions(+)
 create mode 100644 arch/arm/boot/dts/omap54xx-clocks.dtsi
arch/arm/boot/dts/omap5.dtsi +14 -0
index 156b587..f25adc5 100644
@@ -724,4 +724,18 @@
 			compatible = "ti,omap5430-bandgap";
 		};
 	};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "omap54xx-clocks.dtsi"
+	};
+
+	clockdomains {
+		l3init_clkdm: l3init_clkdm {
+			compatible = "ti,clockdomain";
+			clocks = <&dpll_usb_ck>;
+		};
+	};
 };
arch/arm/boot/dts/omap54xx-clocks.dtsi new file +1327 -0

1,328 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#14 lore raw

PATCH v7 14/36 CLK: TI: add omap5 clock init file

In reply to Tero Kristo

clk-54xx.c now contains the clock init functionality for omap5, including DT clock registration and adding of static clkdev entries.

---
 arch/arm/mach-omap2/io.c  |    1 +
 drivers/clk/ti/Makefile   |    1 +
 drivers/clk/ti/clk-54xx.c |  241 +++++++++++++++++++++++++++++++++++++++++++++
 include/linux/clk/ti.h    |    1 +
 4 files changed, 244 insertions(+)
 create mode 100644 drivers/clk/ti/clk-54xx.c
arch/arm/mach-omap2/io.c +1 -0
index ff2113c..205bb1d 100644
@@ -650,6 +650,7 @@ void __init omap5_init_early(void)
 	omap54xx_clockdomains_init();
 	omap54xx_hwmod_init();
 	omap_hwmod_init_postsetup();
+	omap_clk_init = omap5xxx_clk_init;
 }
 #endif
 
drivers/clk/ti/Makefile +1 -0
index 1ed6dd7..9d785e5 100644
@@ -3,4 +3,5 @@ obj-y					+= clk.o dpll.o autoidle.o divider.o \
 					   fixed-factor.o gate.o clockdomain.o \
 					   mux-gate.o mux.o
 obj-$(CONFIG_ARCH_OMAP4)		+= clk-44xx.o
+obj-$(CONFIG_SOC_OMAP5)			+= clk-54xx.o
 endif
drivers/clk/ti/clk-54xx.c new file +241 -0

241 lines not on this page · raw message

include/linux/clk/ti.h +1 -0
index c9a18b9..9a29e17 100644
@@ -220,6 +220,7 @@ void of_ti_autoidle_setup(struct device_node *node);
 struct clk_div_table *ti_clk_get_div_table(struct device_node *node);
 
 int omap4xxx_clk_init(void);
+int omap5xxx_clk_init(void);
 
 #ifdef CONFIG_OF
 void of_ti_clk_allow_autoidle_all(void);
Signature · 1 line
1.7.9.5
#15 lore raw

PATCH v7 15/36 CLK: TI: omap5: Initialize USB_DPLL at boot

In reply to Tero Kristo

From: Roger Quadros <[email protected]>

USB_DPLL must be initialized and locked at boot so that USB modules can work.

---
 drivers/clk/ti/clk-54xx.c |   18 +++++++++++++++++-
 1 file changed, 17 insertions(+), 1 deletion(-)
drivers/clk/ti/clk-54xx.c +17 -1
index efa81ab..e39e03d 100644
@@ -19,6 +19,12 @@
 
 #define OMAP5_DPLL_ABE_DEFFREQ				98304000
 
+/*
+ * OMAP543x TRM, section "3.6.3.9.5 DPLL_USB Preferred Settings"
+ * states it must be at 960MHz
+ */
+#define OMAP5_DPLL_USB_DEFFREQ				960000000
+
 static struct ti_dt_clk omap54xx_clks[] = {
 	DT_CLK(NULL, "pad_clks_src_ck", "pad_clks_src_ck"),
 	DT_CLK(NULL, "pad_clks_ck", "pad_clks_ck"),
@@ -220,7 +226,7 @@ static struct ti_dt_clk omap54xx_clks[] = {
 int __init omap5xxx_clk_init(void)
 {
 	int rc;
-	struct clk *abe_dpll_ref, *abe_dpll, *sys_32k_ck;
+	struct clk *abe_dpll_ref, *abe_dpll, *sys_32k_ck, *usb_dpll;
 
 	of_clk_init(NULL);
 
@@ -237,5 +243,15 @@ int __init omap5xxx_clk_init(void)
 	if (rc)
 		pr_err("%s: failed to configure ABE DPLL!\n", __func__);
 
+	usb_dpll = clk_get_sys(NULL, "dpll_usb_ck");
+	rc = clk_set_rate(usb_dpll, OMAP5_DPLL_USB_DEFFREQ);
+	if (rc)
+		pr_err("%s: failed to configure USB DPLL!\n", __func__);
+
+	usb_dpll = clk_get_sys(NULL, "dpll_usb_m2_ck");
+	rc = clk_set_rate(usb_dpll, OMAP5_DPLL_USB_DEFFREQ/2);
+	if (rc)
+		pr_err("%s: failed to set USB_DPLL M2 OUT\n", __func__);
+
 	return 0;
 }
Signature · 1 line
1.7.9.5
#16 lore raw

PATCH v7 16/36 ARM: dts: dra7 clock data

In reply to Tero Kristo

This patch creates a unique node for each clock in the DRA7 power, reset and clock manager (PRCM).

TODO: apll_pcie clock node is still a dummy in this version, and proper support for the APLL should be added.

---
 arch/arm/boot/dts/dra7xx-clocks.dtsi | 1975 ++++++++++++++++++++++++++++++++++
 1 file changed, 1975 insertions(+)
 create mode 100644 arch/arm/boot/dts/dra7xx-clocks.dtsi
arch/arm/boot/dts/dra7xx-clocks.dtsi new file +1975 -0
new file mode 100644
index 0000000..0d83f95
@@ -0,0 +1,1975 @@
+/*
+ * Device Tree Source for DRA7xx clock data
+ *
+ * Copyright (C) 2013 Texas Instruments, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+atl_clkin0_ck: atl_clkin0_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+atl_clkin1_ck: atl_clkin1_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+atl_clkin2_ck: atl_clkin2_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+atlclkin3_ck: atlclkin3_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+hdmi_clkin_ck: hdmi_clkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+mlb_clkin_ck: mlb_clkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+mlbp_clkin_ck: mlbp_clkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+pciesref_acs_clk_ck: pciesref_acs_clk_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <100000000>;
+};
+
+ref_clkin0_ck: ref_clkin0_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+ref_clkin1_ck: ref_clkin1_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+ref_clkin2_ck: ref_clkin2_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+ref_clkin3_ck: ref_clkin3_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+rmii_clk_ck: rmii_clk_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+sdvenc_clkin_ck: sdvenc_clkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+secure_32k_clk_src_ck: secure_32k_clk_src_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <32768>;
+};
+
+sys_32k_ck: sys_32k_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <32768>;
+};
+
+virt_12000000_ck: virt_12000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <12000000>;
+};
+
+virt_13000000_ck: virt_13000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <13000000>;
+};
+
+virt_16800000_ck: virt_16800000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <16800000>;
+};
+
+virt_19200000_ck: virt_19200000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <19200000>;
+};
+
+virt_20000000_ck: virt_20000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <20000000>;
+};
+
+virt_26000000_ck: virt_26000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <26000000>;
+};
+
+virt_27000000_ck: virt_27000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <27000000>;
+};
+
+virt_38400000_ck: virt_38400000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <38400000>;
+};
+
+sys_clkin1: sys_clkin1 at 4ae06110 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&virt_12000000_ck>, <&virt_20000000_ck>, <&virt_16800000_ck>, <&virt_19200000_ck>, <&virt_26000000_ck>, <&virt_27000000_ck>, <&virt_38400000_ck>;
+	reg = <0x4ae06110 0x4>;
+	ti,index-starts-at-one;
+};
+
+sys_clkin2: sys_clkin2 {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <22579200>;
+};
+
+usb_otg_clkin_ck: usb_otg_clkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+video1_clkin_ck: video1_clkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+video1_m2_clkin_ck: video1_m2_clkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+video2_clkin_ck: video2_clkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+video2_m2_clkin_ck: video2_m2_clkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
+
+abe_dpll_sys_clk_mux: abe_dpll_sys_clk_mux at 4ae06118 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clkin1>, <&sys_clkin2>;
+	reg = <0x4ae06118 0x4>;
+};
+
+abe_dpll_bypass_clk_mux: abe_dpll_bypass_clk_mux at 4ae06114 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_dpll_sys_clk_mux>, <&sys_32k_ck>;
+	reg = <0x4ae06114 0x4>;
+};
+
+abe_dpll_clk_mux: abe_dpll_clk_mux at 4ae0610c {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_dpll_sys_clk_mux>, <&sys_32k_ck>;
+	reg = <0x4ae0610c 0x4>;
+};
+
+dpll_abe_ck: dpll_abe_ck at 4a0051e0 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-m4xen-clock";
+	clocks = <&abe_dpll_clk_mux>, <&abe_dpll_bypass_clk_mux>;
+	reg = <0x4a0051e0 0x4>, <0x4a0051e4 0x4>, <0x4a0051e8 0x4>, <0x4a0051ec 0x4>;
+};
+
+dpll_abe_x2_ck: dpll_abe_x2_ck {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-x2-clock";
+	clocks = <&dpll_abe_ck>;
+};
+
+dpll_abe_m2x2_ck: dpll_abe_m2x2_ck at 4a0051f0 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_abe_x2_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a0051f0 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+abe_24m_fclk: abe_24m_fclk at 4ae0611c {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_abe_m2x2_ck>;
+	reg = <0x4ae0611c 0x4>;
+	ti,dividers = <8>, <16>;
+};
+
+abe_clk: abe_clk at 4a005108 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_abe_m2x2_ck>;
+	reg = <0x4a005108 0x4>;
+	ti,max-div = <4>;
+	ti,index-power-of-two;
+};
+
+aess_fclk: aess_fclk at 4ae06178 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&abe_clk>;
+	reg = <0x4ae06178 0x4>;
+	ti,max-div = <2>;
+};
+
+abe_giclk_div: abe_giclk_div at 4ae06174 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&aess_fclk>;
+	reg = <0x4ae06174 0x4>;
+	ti,max-div = <2>;
+};
+
+abe_lp_clk_div: abe_lp_clk_div at 4ae061d8 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_abe_m2x2_ck>;
+	reg = <0x4ae061d8 0x4>;
+	ti,dividers = <16>, <32>;
+};
+
+abe_sys_clk_div: abe_sys_clk_div at 4ae06120 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&sys_clkin1>;
+	reg = <0x4ae06120 0x4>;
+	ti,max-div = <2>;
+};
+
+adc_gfclk_mux: adc_gfclk_mux at 4ae061dc {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clkin1>, <&sys_clkin2>, <&sys_32k_ck>;
+	reg = <0x4ae061dc 0x4>;
+};
+
+dpll_pcie_ref_ck: dpll_pcie_ref_ck at 4a008200 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-clock";
+	clocks = <&sys_clkin1>, <&sys_clkin1>;
+	reg = <0x4a008200 0x4>, <0x4a008204 0x4>, <0x4a008208 0x4>, <0x4a00820c 0x4>;
+};
+
+dpll_pcie_ref_m2ldo_ck: dpll_pcie_ref_m2ldo_ck at 4a008210 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_pcie_ref_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a008210 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+apll_pcie_ck: apll_pcie_ck at 4a008200 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-clock";
+	clocks = <&dpll_pcie_ref_ck>, <&dpll_pcie_ref_ck>;
+	reg = <0x4a008200 0x4>, <0x4a008204 0x4>, <0x4a008208 0x4>, <0x4a00820c 0x4>;
+};
+
+apll_pcie_clkvcoldo: apll_pcie_clkvcoldo {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&apll_pcie_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+apll_pcie_clkvcoldo_div: apll_pcie_clkvcoldo_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&apll_pcie_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+apll_pcie_m2_ck: apll_pcie_m2_ck at 4a008224 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&apll_pcie_ck>;
+	ti,max-div = <127>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a008224 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+sys_clk1_dclk_div: sys_clk1_dclk_div at 4ae061c8 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&sys_clkin1>;
+	reg = <0x4ae061c8 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+sys_clk2_dclk_div: sys_clk2_dclk_div at 4ae061cc {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&sys_clkin2>;
+	reg = <0x4ae061cc 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+dpll_abe_m2_ck: dpll_abe_m2_ck at 4a0051f0 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_abe_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a0051f0 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+per_abe_x1_dclk_div: per_abe_x1_dclk_div at 4ae061bc {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_abe_m2_ck>;
+	reg = <0x4ae061bc 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+dpll_abe_m3x2_ck: dpll_abe_m3x2_ck at 4a0051f4 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_abe_x2_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a0051f4 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_core_ck: dpll_core_ck at 4a005120 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-core-clock";
+	clocks = <&sys_clkin1>, <&dpll_abe_m3x2_ck>;
+	reg = <0x4a005120 0x4>, <0x4a005124 0x4>, <0x4a005128 0x4>, <0x4a00512c 0x4>;
+};
+
+dpll_core_x2_ck: dpll_core_x2_ck {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-x2-clock";
+	clocks = <&dpll_core_ck>;
+};
+
+dpll_core_h12x2_ck: dpll_core_h12x2_ck at 4a00513c {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a00513c 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+mpu_dpll_hs_clk_div: mpu_dpll_hs_clk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_h12x2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dpll_mpu_ck: dpll_mpu_ck at 4a005160 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-clock";
+	clocks = <&sys_clkin1>, <&mpu_dpll_hs_clk_div>;
+	reg = <0x4a005160 0x4>, <0x4a005164 0x4>, <0x4a005168 0x4>, <0x4a00516c 0x4>;
+};
+
+dpll_mpu_m2_ck: dpll_mpu_m2_ck at 4a005170 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_mpu_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005170 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+mpu_dclk_div: mpu_dclk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_mpu_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dsp_dpll_hs_clk_div: dsp_dpll_hs_clk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_h12x2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dpll_dsp_ck: dpll_dsp_ck at 4a005234 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-clock";
+	clocks = <&sys_clkin1>, <&dsp_dpll_hs_clk_div>;
+	reg = <0x4a005234 0x4>, <0x4a005238 0x4>, <0x4a00523c 0x4>, <0x4a005240 0x4>;
+};
+
+dpll_dsp_m2_ck: dpll_dsp_m2_ck at 4a005244 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_dsp_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005244 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dsp_gclk_div: dsp_gclk_div at 4ae0618c {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_dsp_m2_ck>;
+	reg = <0x4ae0618c 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+iva_dpll_hs_clk_div: iva_dpll_hs_clk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_h12x2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dpll_iva_ck: dpll_iva_ck at 4a0051a0 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-clock";
+	clocks = <&sys_clkin1>, <&iva_dpll_hs_clk_div>;
+	reg = <0x4a0051a0 0x4>, <0x4a0051a4 0x4>, <0x4a0051a8 0x4>, <0x4a0051ac 0x4>;
+};
+
+dpll_iva_m2_ck: dpll_iva_m2_ck at 4a0051b0 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_iva_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a0051b0 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+iva_dclk: iva_dclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_iva_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dpll_gpu_ck: dpll_gpu_ck at 4a0052d8 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-clock";
+	clocks = <&sys_clkin1>, <&dpll_abe_m3x2_ck>;
+	reg = <0x4a0052d8 0x4>, <0x4a0052dc 0x4>, <0x4a0052e0 0x4>, <0x4a0052e4 0x4>;
+};
+
+dpll_gpu_m2_ck: dpll_gpu_m2_ck at 4a0052e8 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_gpu_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a0052e8 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+gpu_dclk: gpu_dclk at 4ae061a0 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_gpu_m2_ck>;
+	reg = <0x4ae061a0 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+dpll_core_m2_ck: dpll_core_m2_ck at 4a005130 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005130 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+core_dpll_out_dclk_div: core_dpll_out_dclk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dpll_ddr_ck: dpll_ddr_ck at 4a005210 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-clock";
+	clocks = <&sys_clkin1>, <&dpll_abe_m3x2_ck>;
+	reg = <0x4a005210 0x4>, <0x4a005214 0x4>, <0x4a005218 0x4>, <0x4a00521c 0x4>;
+};
+
+dpll_ddr_m2_ck: dpll_ddr_m2_ck at 4a005220 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_ddr_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005220 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+emif_phy_dclk_div: emif_phy_dclk_div at 4ae06190 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_ddr_m2_ck>;
+	reg = <0x4ae06190 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+dpll_gmac_ck: dpll_gmac_ck at 4a0052a8 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-clock";
+	clocks = <&sys_clkin1>, <&dpll_abe_m3x2_ck>;
+	reg = <0x4a0052a8 0x4>, <0x4a0052ac 0x4>, <0x4a0052b0 0x4>, <0x4a0052b4 0x4>;
+};
+
+dpll_gmac_m2_ck: dpll_gmac_m2_ck at 4a0052b8 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_gmac_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a0052b8 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+gmac_250m_dclk_div: gmac_250m_dclk_div at 4ae0619c {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_gmac_m2_ck>;
+	reg = <0x4ae0619c 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+video2_dclk_div: video2_dclk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&video2_m2_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+video1_dclk_div: video1_dclk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&video1_m2_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+hdmi_dclk_div: hdmi_dclk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&hdmi_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+per_dpll_hs_clk_div: per_dpll_hs_clk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_abe_m3x2_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+dpll_per_ck: dpll_per_ck at 4a008140 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-clock";
+	clocks = <&sys_clkin1>, <&per_dpll_hs_clk_div>;
+	reg = <0x4a008140 0x4>, <0x4a008144 0x4>, <0x4a008148 0x4>, <0x4a00814c 0x4>;
+};
+
+dpll_per_m2_ck: dpll_per_m2_ck at 4a008150 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_per_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a008150 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+func_96m_aon_dclk_div: func_96m_aon_dclk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+usb_dpll_hs_clk_div: usb_dpll_hs_clk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_abe_m3x2_ck>;
+	clock-mult = <1>;
+	clock-div = <3>;
+};
+
+dpll_usb_ck: dpll_usb_ck at 4a008180 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-j-type-clock";
+	clocks = <&sys_clkin1>, <&usb_dpll_hs_clk_div>;
+	reg = <0x4a008180 0x4>, <0x4a008184 0x4>, <0x4a008188 0x4>, <0x4a00818c 0x4>;
+};
+
+dpll_usb_m2_ck: dpll_usb_m2_ck at 4a008190 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_usb_ck>;
+	ti,max-div = <127>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a008190 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+l3init_480m_dclk_div: l3init_480m_dclk_div at 4ae061ac {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_usb_m2_ck>;
+	reg = <0x4ae061ac 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+usb_otg_dclk_div: usb_otg_dclk_div at 4ae06184 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&usb_otg_clkin_ck>;
+	reg = <0x4ae06184 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+sata_dclk_div: sata_dclk_div at 4ae061c0 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&sys_clkin1>;
+	reg = <0x4ae061c0 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+dpll_pcie_ref_m2_ck: dpll_pcie_ref_m2_ck at 4a008210 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_pcie_ref_ck>;
+	ti,max-div = <127>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a008210 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+pcie2_dclk_div: pcie2_dclk_div at 4ae061b8 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_pcie_ref_m2_ck>;
+	reg = <0x4ae061b8 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+pcie_dclk_div: pcie_dclk_div at 4ae061b4 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&apll_pcie_m2_ck>;
+	reg = <0x4ae061b4 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+emu_dclk_div: emu_dclk_div at 4ae06194 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&sys_clkin1>;
+	reg = <0x4ae06194 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+secure_32k_dclk_div: secure_32k_dclk_div at 4ae061c4 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&secure_32k_clk_src_ck>;
+	reg = <0x4ae061c4 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+eve_dpll_hs_clk_div: eve_dpll_hs_clk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_h12x2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dpll_eve_ck: dpll_eve_ck at 4a005284 {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-clock";
+	clocks = <&sys_clkin1>, <&eve_dpll_hs_clk_div>;
+	reg = <0x4a005284 0x4>, <0x4a005288 0x4>, <0x4a00528c 0x4>, <0x4a005290 0x4>;
+};
+
+dpll_eve_m2_ck: dpll_eve_m2_ck at 4a005294 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_eve_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005294 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+eve_dclk_div: eve_dclk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_eve_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+clkoutmux0_clk_mux: clkoutmux0_clk_mux at 4ae06158 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clk1_dclk_div>, <&sys_clk2_dclk_div>, <&per_abe_x1_dclk_div>, <&mpu_dclk_div>, <&dsp_gclk_div>, <&iva_dclk>, <&gpu_dclk>, <&core_dpll_out_dclk_div>, <&emif_phy_dclk_div>, <&gmac_250m_dclk_div>, <&video2_dclk_div>, <&video1_dclk_div>, <&hdmi_dclk_div>, <&func_96m_aon_dclk_div>, <&l3init_480m_dclk_div>, <&usb_otg_dclk_div>, <&sata_dclk_div>, <&pcie2_dclk_div>, <&pcie_dclk_div>, <&emu_dclk_div>, <&secure_32k_dclk_div>, <&eve_dclk_div>;
+	reg = <0x4ae06158 0x4>;
+};
+
+clkoutmux1_clk_mux: clkoutmux1_clk_mux at 4ae0615c {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clk1_dclk_div>, <&sys_clk2_dclk_div>, <&per_abe_x1_dclk_div>, <&mpu_dclk_div>, <&dsp_gclk_div>, <&iva_dclk>, <&gpu_dclk>, <&core_dpll_out_dclk_div>, <&emif_phy_dclk_div>, <&gmac_250m_dclk_div>, <&video2_dclk_div>, <&video1_dclk_div>, <&hdmi_dclk_div>, <&func_96m_aon_dclk_div>, <&l3init_480m_dclk_div>, <&usb_otg_dclk_div>, <&sata_dclk_div>, <&pcie2_dclk_div>, <&pcie_dclk_div>, <&emu_dclk_div>, <&secure_32k_dclk_div>, <&eve_dclk_div>;
+	reg = <0x4ae0615c 0x4>;
+};
+
+clkoutmux2_clk_mux: clkoutmux2_clk_mux at 4ae06160 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clk1_dclk_div>, <&sys_clk2_dclk_div>, <&per_abe_x1_dclk_div>, <&mpu_dclk_div>, <&dsp_gclk_div>, <&iva_dclk>, <&gpu_dclk>, <&core_dpll_out_dclk_div>, <&emif_phy_dclk_div>, <&gmac_250m_dclk_div>, <&video2_dclk_div>, <&video1_dclk_div>, <&hdmi_dclk_div>, <&func_96m_aon_dclk_div>, <&l3init_480m_dclk_div>, <&usb_otg_dclk_div>, <&sata_dclk_div>, <&pcie2_dclk_div>, <&pcie_dclk_div>, <&emu_dclk_div>, <&secure_32k_dclk_div>, <&eve_dclk_div>;
+	reg = <0x4ae06160 0x4>;
+};
+
+custefuse_sys_gfclk_div: custefuse_sys_gfclk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin1>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+dpll_core_h13x2_ck: dpll_core_h13x2_ck at 4a005140 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005140 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_core_h14x2_ck: dpll_core_h14x2_ck at 4a005144 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005144 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_core_h22x2_ck: dpll_core_h22x2_ck at 4a005154 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005154 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_core_h23x2_ck: dpll_core_h23x2_ck at 4a005158 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005158 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_core_h24x2_ck: dpll_core_h24x2_ck at 4a00515c {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a00515c 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_ddr_x2_ck: dpll_ddr_x2_ck {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-x2-clock";
+	clocks = <&dpll_ddr_ck>;
+};
+
+dpll_ddr_h11x2_ck: dpll_ddr_h11x2_ck at 4a005228 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_ddr_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005228 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_dsp_x2_ck: dpll_dsp_x2_ck {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-x2-clock";
+	clocks = <&dpll_dsp_ck>;
+};
+
+dpll_dsp_m3x2_ck: dpll_dsp_m3x2_ck at 4a005248 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_dsp_x2_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a005248 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_gmac_x2_ck: dpll_gmac_x2_ck {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-x2-clock";
+	clocks = <&dpll_gmac_ck>;
+};
+
+dpll_gmac_h11x2_ck: dpll_gmac_h11x2_ck at 4a0052c0 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_gmac_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a0052c0 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_gmac_h12x2_ck: dpll_gmac_h12x2_ck at 4a0052c4 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_gmac_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a0052c4 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_gmac_h13x2_ck: dpll_gmac_h13x2_ck at 4a0052c8 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_gmac_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a0052c8 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_gmac_m3x2_ck: dpll_gmac_m3x2_ck at 4a0052bc {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_gmac_x2_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a0052bc 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_per_x2_ck: dpll_per_x2_ck {
+	#clock-cells = <0>;
+	compatible = "ti,omap4-dpll-x2-clock";
+	clocks = <&dpll_per_ck>;
+};
+
+dpll_per_h11x2_ck: dpll_per_h11x2_ck at 4a008158 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_per_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a008158 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_per_h12x2_ck: dpll_per_h12x2_ck at 4a00815c {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_per_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a00815c 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_per_h13x2_ck: dpll_per_h13x2_ck at 4a008160 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_per_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a008160 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_per_h14x2_ck: dpll_per_h14x2_ck at 4a008164 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_per_x2_ck>;
+	ti,max-div = <63>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a008164 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_per_m2x2_ck: dpll_per_m2x2_ck at 4a008150 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_per_x2_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x4a008150 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_usb_clkdcoldo: dpll_usb_clkdcoldo {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_usb_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+eve_clk: eve_clk at 4ae06180 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&dpll_eve_m2_ck>, <&dpll_dsp_m3x2_ck>;
+	reg = <0x4ae06180 0x4>;
+};
+
+func_128m_clk: func_128m_clk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_h11x2_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+func_12m_fclk: func_12m_fclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2x2_ck>;
+	clock-mult = <1>;
+	clock-div = <16>;
+};
+
+func_24m_clk: func_24m_clk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <4>;
+};
+
+func_48m_fclk: func_48m_fclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2x2_ck>;
+	clock-mult = <1>;
+	clock-div = <4>;
+};
+
+func_96m_fclk: func_96m_fclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2x2_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+gmii_m_clk_div: gmii_m_clk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_gmac_h11x2_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+hdmi_clk2_div: hdmi_clk2_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&hdmi_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+hdmi_div_clk: hdmi_div_clk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&hdmi_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+hdmi_dpll_clk_mux: hdmi_dpll_clk_mux at 4ae061a4 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clkin1>, <&sys_clkin2>;
+	reg = <0x4ae061a4 0x4>;
+};
+
+l3_iclk_div: l3_iclk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_h12x2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+l3init_60m_fclk: l3init_60m_fclk at 4a008104 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_usb_m2_ck>;
+	reg = <0x4a008104 0x4>;
+	ti,dividers = <1>, <8>;
+};
+
+l4_root_clk_div: l4_root_clk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&l3_iclk_div>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+mlb_clk: mlb_clk at 4ae06134 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&mlb_clkin_ck>;
+	reg = <0x4ae06134 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+mlbp_clk: mlbp_clk at 4ae06130 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&mlbp_clkin_ck>;
+	reg = <0x4ae06130 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+per_abe_x1_gfclk2_div: per_abe_x1_gfclk2_div at 4ae06138 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_abe_m2_ck>;
+	reg = <0x4ae06138 0x4>;
+	ti,max-div = <64>;
+	ti,index-power-of-two;
+};
+
+timer_sys_clk_div: timer_sys_clk_div at 4ae06144 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&sys_clkin1>;
+	reg = <0x4ae06144 0x4>;
+	ti,max-div = <2>;
+};
+
+video1_clk2_div: video1_clk2_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&video1_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+video1_div_clk: video1_div_clk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&video1_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+video1_dpll_clk_mux: video1_dpll_clk_mux at 4ae061d0 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clkin1>, <&sys_clkin2>;
+	reg = <0x4ae061d0 0x4>;
+};
+
+video2_clk2_div: video2_clk2_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&video2_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+video2_div_clk: video2_div_clk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&video2_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+video2_dpll_clk_mux: video2_dpll_clk_mux at 4ae061d4 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clkin1>, <&sys_clkin2>;
+	reg = <0x4ae061d4 0x4>;
+};
+
+wkupaon_iclk_mux: wkupaon_iclk_mux at 4ae06108 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clkin1>, <&abe_lp_clk_div>;
+	reg = <0x4ae06108 0x4>;
+};
+
+dss_32khz_clk: dss_32khz_clk at 4a009120 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <11>;
+	reg = <0x4a009120 0x4>;
+};
+
+dss_48mhz_clk: dss_48mhz_clk at 4a009120 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&func_48m_fclk>;
+	ti,bit-shift = <9>;
+	reg = <0x4a009120 0x4>;
+};
+
+dss_dss_clk: dss_dss_clk at 4a009120 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&dpll_per_h12x2_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009120 0x4>;
+};
+
+dss_hdmi_clk: dss_hdmi_clk at 4a009120 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&hdmi_dpll_clk_mux>;
+	ti,bit-shift = <10>;
+	reg = <0x4a009120 0x4>;
+};
+
+dss_video1_clk: dss_video1_clk at 4a009120 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&video1_dpll_clk_mux>;
+	ti,bit-shift = <12>;
+	reg = <0x4a009120 0x4>;
+};
+
+dss_video2_clk: dss_video2_clk at 4a009120 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&video2_dpll_clk_mux>;
+	ti,bit-shift = <13>;
+	reg = <0x4a009120 0x4>;
+};
+
+gpio1_dbclk: gpio1_dbclk at 4ae07838 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4ae07838 0x4>;
+};
+
+gpio2_dbclk: gpio2_dbclk at 4a009760 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009760 0x4>;
+};
+
+gpio3_dbclk: gpio3_dbclk at 4a009768 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009768 0x4>;
+};
+
+gpio4_dbclk: gpio4_dbclk at 4a009770 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009770 0x4>;
+};
+
+gpio5_dbclk: gpio5_dbclk at 4a009778 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009778 0x4>;
+};
+
+gpio6_dbclk: gpio6_dbclk at 4a009780 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009780 0x4>;
+};
+
+gpio7_dbclk: gpio7_dbclk at 4a009810 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009810 0x4>;
+};
+
+gpio8_dbclk: gpio8_dbclk at 4a009818 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009818 0x4>;
+};
+
+mmc1_clk32k: mmc1_clk32k at 4a009328 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009328 0x4>;
+};
+
+mmc2_clk32k: mmc2_clk32k at 4a009330 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009330 0x4>;
+};
+
+mmc3_clk32k: mmc3_clk32k at 4a009820 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009820 0x4>;
+};
+
+mmc4_clk32k: mmc4_clk32k at 4a009828 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009828 0x4>;
+};
+
+sata_ref_clk: sata_ref_clk at 4a009388 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_clkin1>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009388 0x4>;
+};
+
+usb_otg_ss1_refclk960m: usb_otg_ss1_refclk960m at 4a0093f0 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&dpll_usb_clkdcoldo>;
+	ti,bit-shift = <8>;
+	reg = <0x4a0093f0 0x4>;
+};
+
+usb_otg_ss2_refclk960m: usb_otg_ss2_refclk960m at 4a009340 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&dpll_usb_clkdcoldo>;
+	ti,bit-shift = <8>;
+	reg = <0x4a009340 0x4>;
+};
+
+usb_phy1_always_on_clk32k: usb_phy1_always_on_clk32k at 4a008640 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a008640 0x4>;
+};
+
+usb_phy2_always_on_clk32k: usb_phy2_always_on_clk32k at 4a008688 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a008688 0x4>;
+};
+
+usb_phy3_always_on_clk32k: usb_phy3_always_on_clk32k at 4a008698 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x4a008698 0x4>;
+};
+
+atl_dpll_clk_mux: atl_dpll_clk_mux at 4a008c00 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_32k_ck>, <&video1_clkin_ck>, <&video2_clkin_ck>, <&hdmi_clkin_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a008c00 0x4>;
+};
+
+atl_gfclk_mux: atl_gfclk_mux at 4a008c00 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&l3_iclk_div>, <&dpll_abe_m2_ck>, <&atl_dpll_clk_mux>;
+	ti,bit-shift = <26>;
+	reg = <0x4a008c00 0x4>;
+};
+
+dcan1_sys_clk_mux: dcan1_sys_clk_mux at 4ae07888 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clkin1>, <&sys_clkin2>;
+	ti,bit-shift = <24>;
+	reg = <0x4ae07888 0x4>;
+};
+
+gmac_gmii_ref_clk_div: gmac_gmii_ref_clk_div at 4a0093d0 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_gmac_m2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a0093d0 0x4>;
+	ti,dividers = <2>;
+};
+
+gmac_rft_clk_mux: gmac_rft_clk_mux at 4a0093d0 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&video1_clkin_ck>, <&video2_clkin_ck>, <&dpll_abe_m2_ck>, <&hdmi_clkin_ck>, <&l3_iclk_div>;
+	ti,bit-shift = <25>;
+	reg = <0x4a0093d0 0x4>;
+};
+
+gpu_core_gclk_mux: gpu_core_gclk_mux at 4a009220 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&dpll_core_h14x2_ck>, <&dpll_per_h14x2_ck>, <&dpll_gpu_m2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009220 0x4>;
+};
+
+gpu_hyd_gclk_mux: gpu_hyd_gclk_mux at 4a009220 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&dpll_core_h14x2_ck>, <&dpll_per_h14x2_ck>, <&dpll_gpu_m2_ck>;
+	ti,bit-shift = <26>;
+	reg = <0x4a009220 0x4>;
+};
+
+ipu1_gfclk_mux: ipu1_gfclk_mux at 4a005520 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&dpll_abe_m2x2_ck>, <&dpll_core_h22x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a005520 0x4>;
+};
+
+l3instr_ts_gclk_div: l3instr_ts_gclk_div at 4a008e50 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&wkupaon_iclk_mux>;
+	ti,bit-shift = <24>;
+	reg = <0x4a008e50 0x4>;
+	ti,dividers = <8>, <16>, <32>;
+};
+
+mcasp1_ahclkr_mux: mcasp1_ahclkr_mux at 4a005550 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_24m_fclk>, <&abe_sys_clk_div>, <&func_24m_clk>, <&atlclkin3_ck>, <&atl_clkin2_ck>, <&atl_clkin1_ck>, <&atl_clkin0_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&mlb_clk>, <&mlbp_clk>;
+	ti,bit-shift = <28>;
+	reg = <0x4a005550 0x4>;
+};
+
+mcasp1_ahclkx_mux: mcasp1_ahclkx_mux at 4a005550 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_24m_fclk>, <&abe_sys_clk_div>, <&func_24m_clk>, <&atlclkin3_ck>, <&atl_clkin2_ck>, <&atl_clkin1_ck>, <&atl_clkin0_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&mlb_clk>, <&mlbp_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a005550 0x4>;
+};
+
+mcasp1_aux_gfclk_mux: mcasp1_aux_gfclk_mux at 4a005550 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&per_abe_x1_gfclk2_div>, <&video1_clk2_div>, <&video2_clk2_div>, <&hdmi_clk2_div>;
+	ti,bit-shift = <22>;
+	reg = <0x4a005550 0x4>;
+};
+
+mcasp2_ahclkr_mux: mcasp2_ahclkr_mux at 4a009860 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_24m_fclk>, <&abe_sys_clk_div>, <&func_24m_clk>, <&atlclkin3_ck>, <&atl_clkin2_ck>, <&atl_clkin1_ck>, <&atl_clkin0_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&mlb_clk>, <&mlbp_clk>;
+	ti,bit-shift = <28>;
+	reg = <0x4a009860 0x4>;
+};
+
+mcasp2_ahclkx_mux: mcasp2_ahclkx_mux at 4a009860 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_24m_fclk>, <&abe_sys_clk_div>, <&func_24m_clk>, <&atlclkin3_ck>, <&atl_clkin2_ck>, <&atl_clkin1_ck>, <&atl_clkin0_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&mlb_clk>, <&mlbp_clk>;
+	ti,bit-shift = <28>;
+	reg = <0x4a009860 0x4>;
+};
+
+mcasp2_aux_gfclk_mux: mcasp2_aux_gfclk_mux at 4a009860 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&per_abe_x1_gfclk2_div>, <&video1_clk2_div>, <&video2_clk2_div>, <&hdmi_clk2_div>;
+	ti,bit-shift = <22>;
+	reg = <0x4a009860 0x4>;
+};
+
+mcasp3_ahclkx_mux: mcasp3_ahclkx_mux at 4a009868 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_24m_fclk>, <&abe_sys_clk_div>, <&func_24m_clk>, <&atlclkin3_ck>, <&atl_clkin2_ck>, <&atl_clkin1_ck>, <&atl_clkin0_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&mlb_clk>, <&mlbp_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009868 0x4>;
+};
+
+mcasp3_aux_gfclk_mux: mcasp3_aux_gfclk_mux at 4a009868 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&per_abe_x1_gfclk2_div>, <&video1_clk2_div>, <&video2_clk2_div>, <&hdmi_clk2_div>;
+	ti,bit-shift = <22>;
+	reg = <0x4a009868 0x4>;
+};
+
+mcasp4_ahclkx_mux: mcasp4_ahclkx_mux at 4a009898 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_24m_fclk>, <&abe_sys_clk_div>, <&func_24m_clk>, <&atlclkin3_ck>, <&atl_clkin2_ck>, <&atl_clkin1_ck>, <&atl_clkin0_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&mlb_clk>, <&mlbp_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009898 0x4>;
+};
+
+mcasp4_aux_gfclk_mux: mcasp4_aux_gfclk_mux at 4a009898 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&per_abe_x1_gfclk2_div>, <&video1_clk2_div>, <&video2_clk2_div>, <&hdmi_clk2_div>;
+	ti,bit-shift = <22>;
+	reg = <0x4a009898 0x4>;
+};
+
+mcasp5_ahclkx_mux: mcasp5_ahclkx_mux at 4a009878 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_24m_fclk>, <&abe_sys_clk_div>, <&func_24m_clk>, <&atlclkin3_ck>, <&atl_clkin2_ck>, <&atl_clkin1_ck>, <&atl_clkin0_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&mlb_clk>, <&mlbp_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009878 0x4>;
+};
+
+mcasp5_aux_gfclk_mux: mcasp5_aux_gfclk_mux at 4a009878 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&per_abe_x1_gfclk2_div>, <&video1_clk2_div>, <&video2_clk2_div>, <&hdmi_clk2_div>;
+	ti,bit-shift = <22>;
+	reg = <0x4a009878 0x4>;
+};
+
+mcasp6_ahclkx_mux: mcasp6_ahclkx_mux at 4a009904 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_24m_fclk>, <&abe_sys_clk_div>, <&func_24m_clk>, <&atlclkin3_ck>, <&atl_clkin2_ck>, <&atl_clkin1_ck>, <&atl_clkin0_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&mlb_clk>, <&mlbp_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009904 0x4>;
+};
+
+mcasp6_aux_gfclk_mux: mcasp6_aux_gfclk_mux at 4a009904 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&per_abe_x1_gfclk2_div>, <&video1_clk2_div>, <&video2_clk2_div>, <&hdmi_clk2_div>;
+	ti,bit-shift = <22>;
+	reg = <0x4a009904 0x4>;
+};
+
+mcasp7_ahclkx_mux: mcasp7_ahclkx_mux at 4a009908 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_24m_fclk>, <&abe_sys_clk_div>, <&func_24m_clk>, <&atlclkin3_ck>, <&atl_clkin2_ck>, <&atl_clkin1_ck>, <&atl_clkin0_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&mlb_clk>, <&mlbp_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009908 0x4>;
+};
+
+mcasp7_aux_gfclk_mux: mcasp7_aux_gfclk_mux at 4a009908 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&per_abe_x1_gfclk2_div>, <&video1_clk2_div>, <&video2_clk2_div>, <&hdmi_clk2_div>;
+	ti,bit-shift = <22>;
+	reg = <0x4a009908 0x4>;
+};
+
+mcasp8_ahclk_mux: mcasp8_ahclk_mux at 4a009890 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&abe_24m_fclk>, <&abe_sys_clk_div>, <&func_24m_clk>, <&atlclkin3_ck>, <&atl_clkin2_ck>, <&atl_clkin1_ck>, <&atl_clkin0_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&mlb_clk>, <&mlbp_clk>;
+	ti,bit-shift = <22>;
+	reg = <0x4a009890 0x4>;
+};
+
+mcasp8_aux_gfclk_mux: mcasp8_aux_gfclk_mux at 4a009890 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&per_abe_x1_gfclk2_div>, <&video1_clk2_div>, <&video2_clk2_div>, <&hdmi_clk2_div>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009890 0x4>;
+};
+
+mmc1_fclk_mux: mmc1_fclk_mux at 4a009328 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_128m_clk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009328 0x4>;
+};
+
+mmc1_fclk_div: mmc1_fclk_div at 4a009328 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&mmc1_fclk_mux>;
+	ti,bit-shift = <25>;
+	ti,max-div = <4>;
+	reg = <0x4a009328 0x4>;
+	ti,index-power-of-two;
+};
+
+mmc2_fclk_mux: mmc2_fclk_mux at 4a009330 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_128m_clk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009330 0x4>;
+};
+
+mmc2_fclk_div: mmc2_fclk_div at 4a009330 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&mmc2_fclk_mux>;
+	ti,bit-shift = <25>;
+	ti,max-div = <4>;
+	reg = <0x4a009330 0x4>;
+	ti,index-power-of-two;
+};
+
+mmc3_gfclk_mux: mmc3_gfclk_mux at 4a009820 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009820 0x4>;
+};
+
+mmc3_gfclk_div: mmc3_gfclk_div at 4a009820 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&mmc3_gfclk_mux>;
+	ti,bit-shift = <25>;
+	ti,max-div = <4>;
+	reg = <0x4a009820 0x4>;
+	ti,index-power-of-two;
+};
+
+mmc4_gfclk_mux: mmc4_gfclk_mux at 4a009828 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009828 0x4>;
+};
+
+mmc4_gfclk_div: mmc4_gfclk_div at 4a009828 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&mmc4_gfclk_mux>;
+	ti,bit-shift = <25>;
+	ti,max-div = <4>;
+	reg = <0x4a009828 0x4>;
+	ti,index-power-of-two;
+};
+
+qspi_gfclk_mux: qspi_gfclk_mux at 4a009838 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_128m_clk>, <&dpll_per_h13x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009838 0x4>;
+};
+
+qspi_gfclk_div: qspi_gfclk_div at 4a009838 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&qspi_gfclk_mux>;
+	ti,bit-shift = <25>;
+	ti,max-div = <4>;
+	reg = <0x4a009838 0x4>;
+	ti,index-power-of-two;
+};
+
+timer10_gfclk_mux: timer10_gfclk_mux at 4a009728 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009728 0x4>;
+};
+
+timer11_gfclk_mux: timer11_gfclk_mux at 4a009730 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009730 0x4>;
+};
+
+timer13_gfclk_mux: timer13_gfclk_mux at 4a0097c8 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a0097c8 0x4>;
+};
+
+timer14_gfclk_mux: timer14_gfclk_mux at 4a0097d0 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a0097d0 0x4>;
+};
+
+timer15_gfclk_mux: timer15_gfclk_mux at 4a0097d8 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a0097d8 0x4>;
+};
+
+timer16_gfclk_mux: timer16_gfclk_mux at 4a009830 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009830 0x4>;
+};
+
+timer1_gfclk_mux: timer1_gfclk_mux at 4ae07840 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4ae07840 0x4>;
+};
+
+timer2_gfclk_mux: timer2_gfclk_mux at 4a009738 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009738 0x4>;
+};
+
+timer3_gfclk_mux: timer3_gfclk_mux at 4a009740 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009740 0x4>;
+};
+
+timer4_gfclk_mux: timer4_gfclk_mux at 4a009748 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009748 0x4>;
+};
+
+timer5_gfclk_mux: timer5_gfclk_mux at 4a005558 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>, <&clkoutmux0_clk_mux>;
+	ti,bit-shift = <24>;
+	reg = <0x4a005558 0x4>;
+};
+
+timer6_gfclk_mux: timer6_gfclk_mux at 4a005560 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>, <&clkoutmux0_clk_mux>;
+	ti,bit-shift = <24>;
+	reg = <0x4a005560 0x4>;
+};
+
+timer7_gfclk_mux: timer7_gfclk_mux at 4a005568 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>, <&clkoutmux0_clk_mux>;
+	ti,bit-shift = <24>;
+	reg = <0x4a005568 0x4>;
+};
+
+timer8_gfclk_mux: timer8_gfclk_mux at 4a005570 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>, <&clkoutmux0_clk_mux>;
+	ti,bit-shift = <24>;
+	reg = <0x4a005570 0x4>;
+};
+
+timer9_gfclk_mux: timer9_gfclk_mux at 4a009750 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&timer_sys_clk_div>, <&sys_32k_ck>, <&sys_clkin2>, <&ref_clkin0_ck>, <&ref_clkin1_ck>, <&ref_clkin2_ck>, <&ref_clkin3_ck>, <&abe_giclk_div>, <&video1_div_clk>, <&video2_div_clk>, <&hdmi_div_clk>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009750 0x4>;
+};
+
+uart10_gfclk_mux: uart10_gfclk_mux at 4ae07880 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4ae07880 0x4>;
+};
+
+uart1_gfclk_mux: uart1_gfclk_mux at 4a009840 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009840 0x4>;
+};
+
+uart2_gfclk_mux: uart2_gfclk_mux at 4a009848 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009848 0x4>;
+};
+
+uart3_gfclk_mux: uart3_gfclk_mux at 4a009850 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009850 0x4>;
+};
+
+uart4_gfclk_mux: uart4_gfclk_mux at 4a009858 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009858 0x4>;
+};
+
+uart5_gfclk_mux: uart5_gfclk_mux at 4a009870 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009870 0x4>;
+};
+
+uart6_gfclk_mux: uart6_gfclk_mux at 4a005580 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a005580 0x4>;
+};
+
+uart7_gfclk_mux: uart7_gfclk_mux at 4a0098d0 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a0098d0 0x4>;
+};
+
+uart8_gfclk_mux: uart8_gfclk_mux at 4a0098e0 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a0098e0 0x4>;
+};
+
+uart9_gfclk_mux: uart9_gfclk_mux at 4a0098e8 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&func_48m_fclk>, <&dpll_per_m2x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a0098e8 0x4>;
+};
+
+vip1_gclk_mux: vip1_gclk_mux at 4a009020 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&l3_iclk_div>, <&dpll_core_h23x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009020 0x4>;
+};
+
+vip2_gclk_mux: vip2_gclk_mux at 4a009028 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&l3_iclk_div>, <&dpll_core_h23x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009028 0x4>;
+};
+
+vip3_gclk_mux: vip3_gclk_mux at 4a009030 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&l3_iclk_div>, <&dpll_core_h23x2_ck>;
+	ti,bit-shift = <24>;
+	reg = <0x4a009030 0x4>;
+};
+
+dummy_ck: dummy_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <0>;
+};
\ No newline at end of file
Signature · 1 line
1.7.9.5
#17 lore raw

PATCH v7 17/36 ARM: dts: clk: Add apll related clocks

In reply to Tero Kristo

From: Keerthy <[email protected]>

The patch adds a mux node to choose the parent of apll_pcie_ck node.

---
 arch/arm/boot/dts/dra7xx-clocks.dtsi |   14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)
arch/arm/boot/dts/dra7xx-clocks.dtsi +11 -3
index 0d83f95..c830e15 100644
@@ -315,11 +315,19 @@ dpll_pcie_ref_m2ldo_ck: dpll_pcie_ref_m2ldo_ck at 4a008210 {
 	ti,autoidle-low;
 };
 
+apll_pcie_in_clk_mux: apll_pcie_in_clk_mux at 4ae06118 {
+	compatible = "ti,mux-clock";
+	clocks = <&dpll_pcie_ref_ck>, <&pciesref_acs_clk_ck>;
+	#clock-cells = <0>;
+	reg = <0x4a00821c 0x4>;
+	ti,bit-shift = <7>;
+};
+
 apll_pcie_ck: apll_pcie_ck at 4a008200 {
 	#clock-cells = <0>;
-	compatible = "ti,omap4-dpll-clock";
-	clocks = <&dpll_pcie_ref_ck>, <&dpll_pcie_ref_ck>;
-	reg = <0x4a008200 0x4>, <0x4a008204 0x4>, <0x4a008208 0x4>, <0x4a00820c 0x4>;
+	compatible = "ti,dra7-apll-clock";
+	clocks = <&apll_pcie_in_clk_mux>, <&dpll_pcie_ref_ck>;
+	reg = <0x4a00821c 0x4>, <0x4a008220 0x4>;
 };
 
 apll_pcie_clkvcoldo: apll_pcie_clkvcoldo {
Signature · 1 line
1.7.9.5
#18 lore raw

PATCH v7 18/36 ARM: dts: DRA7: Change apll_pcie_m2_ck to fixed factor clock

In reply to Tero Kristo

From: Keerthy <[email protected]>

This patch changes apll_pcie_m2_ck to fixed factor clock as there are no configurable divider associated to m2.

---
 arch/arm/boot/dts/dra7xx-clocks.dtsi |    9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)
arch/arm/boot/dts/dra7xx-clocks.dtsi +3 -6
index c830e15..75d5e1b 100644
@@ -348,13 +348,10 @@ apll_pcie_clkvcoldo_div: apll_pcie_clkvcoldo_div {
 
 apll_pcie_m2_ck: apll_pcie_m2_ck at 4a008224 {
 	#clock-cells = <0>;
-	compatible = "ti,divider-clock";
+	compatible = "fixed-factor-clock";
 	clocks = <&apll_pcie_ck>;
-	ti,max-div = <127>;
-	ti,autoidle-shift = <8>;
-	reg = <0x4a008224 0x4>;
-	ti,index-starts-at-one;
-	ti,autoidle-low;
+	clock-mult = <1>;
+	clock-div = <1>;
 };
 
 sys_clk1_dclk_div: sys_clk1_dclk_div at 4ae061c8 {
Signature · 1 line
1.7.9.5
#19 lore raw

PATCH v7 19/36 ARM: dts: DRA7: Add PCIe related clock nodes

In reply to Tero Kristo

From: Keerthy <[email protected]>

This patch adds optfclk_pciephy_clk and optfclk_pciephy_div_clk which are used by PCIe phy. It also adds a mux clock to choose the source of optfclk_pciephy_div_clk clock.

---
 arch/arm/boot/dts/dra7xx-clocks.dtsi |   27 ++++++++++++++++++++++++++-
 1 file changed, 26 insertions(+), 1 deletion(-)
arch/arm/boot/dts/dra7xx-clocks.dtsi +26 -1
index 75d5e1b..d83afd3 100644
@@ -1973,8 +1973,33 @@ vip3_gclk_mux: vip3_gclk_mux at 4a009030 {
 	reg = <0x4a009030 0x4>;
 };
 
+optfclk_pciephy_div: optfclk_pciephy_div at 4a00821c {
+	compatible = "ti,divider-clock";
+	clocks = <&apll_pcie_ck>;
+	#clock-cells = <0>;
+	reg = <0x4a00821c 0x4>;
+	ti,bit-shift = <8>;
+	ti,max-div = <2>;
+};
+
+optfclk_pciephy_clk: optfclk_pciephy_clk at 4a0093b0 {
+	compatible = "ti,gate-clock";
+	clocks = <&apll_pcie_ck>;
+	#clock-cells = <0>;
+	reg = <0x4a0093b0 0x4>;
+	ti,bit-shift = <9>;
+};
+
+optfclk_pciephy_div_clk: optfclk_pciephy_div_clk at 4a0093b0 {
+	compatible = "ti,gate-clock";
+	clocks = <&optfclk_pciephy_div>;
+	#clock-cells = <0>;
+	reg = <0x4a0093b0 0x4>;
+	ti,bit-shift = <10>;
+};
+
 dummy_ck: dummy_ck {
 	#clock-cells = <0>;
 	compatible = "fixed-clock";
 	clock-frequency = <0>;
-};
\ No newline at end of file
+};
Signature · 1 line
1.7.9.5
#20 lore raw

PATCH v7 20/36 CLK: TI: DRA7: Add APLL support

In reply to Tero Kristo

From: Keerthy <[email protected]>

The patch adds support for DRA7 PCIe APLL. The APLL sources the optional functional clocks for PCIe module.

APLL stands for Analog PLL. This is different when comapred with DPLL meaning Digital PLL, the phase detection is done using an analog circuit.

---
 .../devicetree/bindings/clock/ti/apll.txt          |   29 +++
 drivers/clk/ti/Makefile                            |    2 +-
 drivers/clk/ti/apll.c                              |  207 ++++++++++++++++++++
 3 files changed, 237 insertions(+), 1 deletion(-)
 create mode 100644 Documentation/devicetree/bindings/clock/ti/apll.txt
 create mode 100644 drivers/clk/ti/apll.c
Documentation/devicetree/bindings/clock/ti/apll.txt new file +29 -0
new file mode 100644
index 0000000..02649cd
@@ -0,0 +1,29 @@
+Binding for Texas Instruments APLL clock.
+
+This binding uses the common clock binding[1].  It assumes a
+register-mapped APLL with usually two selectable input clocks
+(reference clock and bypass clock), with analog phase locked
+loop logic for multiplying the input clock to a desired output
+clock. This clock also typically supports different operation
+modes (locked, low power stop etc.) APLL mostly behaves like
+a subtype of a DPLL [2], although a simplified one at that.
+
+[1] Documentation/devicetree/bindings/clock/clock-bindings.txt
+[2] Documentation/devicetree/bindings/clock/ti/dpll.txt
+
+Required properties:
+- compatible : shall be "ti,dra7-apll-clock"
+- #clock-cells : from common clock binding; shall be set to 0.
+- clocks : link phandles of parent clocks (clk-ref and clk-bypass)
+- reg : address and length of the register set for controlling the APLL.
+  It contains the information of registers in the following order:
+	"control" - contains the control register base address
+	"idlest" - contains the idlest register base address
+
+Examples:
+	apll_pcie_ck: apll_pcie_ck at 4a008200 {
+		#clock-cells = <0>;
+		clocks = <&apll_pcie_in_clk_mux>, <&dpll_pcie_ref_ck>;
+		reg = <0x4a00821c 0x4>, <0x4a008220 0x4>;
+		compatible = "ti,dra7-apll-clock";
+	};
drivers/clk/ti/Makefile +1 -1
index 9d785e5..dba8b3e 100644
@@ -1,7 +1,7 @@
 ifneq ($(CONFIG_OF),)
 obj-y					+= clk.o dpll.o autoidle.o divider.o \
 					   fixed-factor.o gate.o clockdomain.o \
-					   mux-gate.o mux.o
+					   mux-gate.o mux.o apll.o
 obj-$(CONFIG_ARCH_OMAP4)		+= clk-44xx.o
 obj-$(CONFIG_SOC_OMAP5)			+= clk-54xx.o
 endif
drivers/clk/ti/apll.c new file +207 -0

207 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#21 lore raw

PATCH v7 20/36 CLK: TI: DRA7: Add APLL support

Quoting Tero Kristo (2013-09-25 01:48:26)

drivers/clk/ti/apll.c line 105

Quoted 4 lines
drivers/clk/ti/apll.c new file +4 -0
+
+static const struct clk_ops apll_ck_ops = {
+       .enable         = &dra7_apll_enable,
+       .disable        = &dra7_apll_disable,

Looks like .is_enabled is missing?

Also have you thought about using .prepare or .unprepare for these PLLs which might take some time to lock? The code there doesn't sleep or schedule, but it does poll for some time while under a spinlock. Something to think about for a future patch.

Regards,
Mike

#22 lore raw

PATCH v7 21/36 CLK: TI: add dra7 clock init file

In reply to Tero Kristo

clk-7xx.c now contains the clock init functionality for dra7, including DT clock registration and adding of static clkdev entries.

---
 drivers/clk/ti/Makefile  |    1 +
 drivers/clk/ti/clk-7xx.c |  334 ++++++++++++++++++++++++++++++++++++++++++++++
 include/linux/clk/ti.h   |    1 +
 3 files changed, 336 insertions(+)
 create mode 100644 drivers/clk/ti/clk-7xx.c
drivers/clk/ti/Makefile +1 -0
index dba8b3e..1a83c3c 100644
@@ -4,4 +4,5 @@ obj-y					+= clk.o dpll.o autoidle.o divider.o \
 					   mux-gate.o mux.o apll.o
 obj-$(CONFIG_ARCH_OMAP4)		+= clk-44xx.o
 obj-$(CONFIG_SOC_OMAP5)			+= clk-54xx.o
+obj-$(CONFIG_SOC_DRA7XX)		+= clk-7xx.o
 endif
drivers/clk/ti/clk-7xx.c new file +334 -0

334 lines not on this page · raw message

include/linux/clk/ti.h +1 -0
index 9a29e17..1b2dc51 100644
@@ -221,6 +221,7 @@ struct clk_div_table *ti_clk_get_div_table(struct device_node *node);
 
 int omap4xxx_clk_init(void);
 int omap5xxx_clk_init(void);
+int dra7xx_clk_init(void);
 
 #ifdef CONFIG_OF
 void of_ti_clk_allow_autoidle_all(void);
Signature · 1 line
1.7.9.5
#23 lore raw

PATCH v7 22/36 ARM: OMAP: DRA7: Enable clock init

In reply to Tero Kristo

Initializes clock data from device tree.

---
 arch/arm/mach-omap2/io.c |    1 +
 1 file changed, 1 insertion(+)
arch/arm/mach-omap2/io.c +1 -0
index 205bb1d..1f23182 100644
@@ -671,6 +671,7 @@ void __init dra7xx_init_early(void)
 	dra7xx_clockdomains_init();
 	dra7xx_hwmod_init();
 	omap_hwmod_init_postsetup();
+	omap_clk_init = dra7xx_clk_init;
 }
 #endif
 
Signature · 1 line
1.7.9.5
#24 lore raw

PATCH v7 23/36 ARM: dts: DRA7: link in clock DT data

In reply to Tero Kristo

Clock tree DT data is now included from base dra7.dtsi file.

---
 arch/arm/boot/dts/dra7.dtsi |   14 ++++++++++++++
 1 file changed, 14 insertions(+)
arch/arm/boot/dts/dra7.dtsi +14 -0
index c01ef76..c98997b 100644
@@ -572,4 +572,18 @@
 			status = "disabled";
 		};
 	};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "dra7xx-clocks.dtsi"
+	};
+
+	clockdomains {
+		coreaon_clkdm: coreaon_clkdm {
+			compatible = "ti,clockdomain";
+			clocks = <&dpll_usb_ck>;
+		};
+	};
 };
Signature · 1 line
1.7.9.5
#25 lore raw

PATCH v7 24/36 ARM: dts: am33xx clock data

In reply to Tero Kristo

This patch creates a unique node for each clock in the AM33xx power, reset and clock manager (PRCM).

---
 arch/arm/boot/dts/am33xx-clocks.dtsi |  640 ++++++++++++++++++++++++++++++++++
 arch/arm/boot/dts/am33xx.dtsi        |   16 +-
 2 files changed, 655 insertions(+), 1 deletion(-)
 create mode 100644 arch/arm/boot/dts/am33xx-clocks.dtsi
arch/arm/boot/dts/am33xx-clocks.dtsi new file +640 -0
new file mode 100644
index 0000000..9a79e49
@@ -0,0 +1,640 @@
+/*
+ * Device Tree Source for AM33xx clock data
+ *
+ * Copyright (C) 2013 Texas Instruments, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+clk_32768_ck: clk_32768_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <32768>;
+};
+
+clk_rc32k_ck: clk_rc32k_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <32000>;
+};
+
+virt_19200000_ck: virt_19200000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <19200000>;
+};
+
+virt_24000000_ck: virt_24000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <24000000>;
+};
+
+virt_25000000_ck: virt_25000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <25000000>;
+};
+
+virt_26000000_ck: virt_26000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <26000000>;
+};
+
+sys_clkin_ck: sys_clkin_ck at 44e10040 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&virt_19200000_ck>, <&virt_24000000_ck>, <&virt_25000000_ck>, <&virt_26000000_ck>;
+	ti,bit-shift = <22>;
+	reg = <0x44e10040 0x4>;
+};
+
+tclkin_ck: tclkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <12000000>;
+};
+
+dpll_core_ck: dpll_core_ck at 44e00490 {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-core-clock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44e00490 0x4>, <0x44e0045c 0x4>, <0x44e00468 0x4>;
+};
+
+dpll_core_x2_ck: dpll_core_x2_ck {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-x2-clock";
+	clocks = <&dpll_core_ck>;
+};
+
+dpll_core_m4_ck: dpll_core_m4_ck at 44e00480 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	reg = <0x44e00480 0x4>;
+	ti,max-div = <31>;
+	ti,index-starts-at-one;
+};
+
+dpll_core_m5_ck: dpll_core_m5_ck at 44e00484 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	reg = <0x44e00484 0x4>;
+	ti,max-div = <31>;
+	ti,index-starts-at-one;
+};
+
+dpll_core_m6_ck: dpll_core_m6_ck at 44e004d8 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	reg = <0x44e004d8 0x4>;
+	ti,max-div = <31>;
+	ti,index-starts-at-one;
+};
+
+dpll_mpu_ck: dpll_mpu_ck at 44e00488 {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-clock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44e00488 0x4>, <0x44e00420 0x4>, <0x44e0042c 0x4>;
+};
+
+dpll_mpu_m2_ck: dpll_mpu_m2_ck at 44e004a8 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_mpu_ck>;
+	reg = <0x44e004a8 0x4>;
+	ti,max-div = <31>;
+	ti,index-starts-at-one;
+};
+
+dpll_ddr_ck: dpll_ddr_ck at 44e00494 {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-no-gate-clock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44e00494 0x4>, <0x44e00434 0x4>, <0x44e00440 0x4>;
+};
+
+dpll_ddr_m2_ck: dpll_ddr_m2_ck at 44e004a0 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_ddr_ck>;
+	reg = <0x44e004a0 0x4>;
+	ti,max-div = <31>;
+	ti,index-starts-at-one;
+};
+
+dpll_ddr_m2_div2_ck: dpll_ddr_m2_div2_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_ddr_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+dpll_disp_ck: dpll_disp_ck at 44e00498 {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-no-gate-clock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44e00498 0x4>, <0x44e00448 0x4>, <0x44e00454 0x4>;
+};
+
+dpll_disp_m2_ck: dpll_disp_m2_ck at 44e004a4 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_disp_ck>;
+	ti,max-div = <31>;
+	reg = <0x44e004a4 0x4>;
+	ti,index-starts-at-one;
+	ti,set-rate-parent;
+};
+
+dpll_per_ck: dpll_per_ck at 44e0048c {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-no-gate-j-type-clock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44e0048c 0x4>, <0x44e00470 0x4>, <0x44e0049c 0x4>;
+};
+
+dpll_per_m2_ck: dpll_per_m2_ck at 44e004ac {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_per_ck>;
+	reg = <0x44e004ac 0x4>;
+	ti,max-div = <31>;
+	ti,index-starts-at-one;
+};
+
+dpll_per_m2_div4_wkupdm_ck: dpll_per_m2_div4_wkupdm_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <4>;
+};
+
+dpll_per_m2_div4_ck: dpll_per_m2_div4_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <4>;
+};
+
+adc_tsc_fck: adc_tsc_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+cefuse_fck: cefuse_fck at 44e00a20 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_clkin_ck>;
+	ti,bit-shift = <1>;
+	reg = <0x44e00a20 0x4>;
+};
+
+clk_24mhz: clk_24mhz {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <8>;
+};
+
+clkdiv32k_ck: clkdiv32k_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&clk_24mhz>;
+	clock-mult = <1>;
+	clock-div = <732>;
+};
+
+clkdiv32k_ick: clkdiv32k_ick at 44e0014c {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkdiv32k_ck>;
+	ti,bit-shift = <1>;
+	reg = <0x44e0014c 0x4>;
+};
+
+dcan0_fck: dcan0_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dcan1_fck: dcan1_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+l3_gclk: l3_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+pruss_ocp_gclk: pruss_ocp_gclk at 44e00530 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&l3_gclk>, <&dpll_disp_m2_ck>;
+	reg = <0x44e00530 0x4>;
+};
+
+mcasp0_fck: mcasp0_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+mcasp1_fck: mcasp1_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+mmu_fck: mmu_fck at 44e00914 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&dpll_core_m4_ck>;
+	ti,bit-shift = <1>;
+	reg = <0x44e00914 0x4>;
+};
+
+smartreflex0_fck: smartreflex0_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+smartreflex1_fck: smartreflex1_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+sha0_fck: sha0_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+aes0_fck: aes0_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+rng_fck: rng_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+timer1_fck: timer1_fck at 44e00528 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clkin_ck>, <&clkdiv32k_ick>, <&tclkin_ck>, <&clk_rc32k_ck>, <&clk_32768_ck>;
+	reg = <0x44e00528 0x4>;
+};
+
+timer2_fck: timer2_fck at 44e00508 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44e00508 0x4>;
+};
+
+timer3_fck: timer3_fck at 44e0050c {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44e0050c 0x4>;
+};
+
+timer4_fck: timer4_fck at 44e00510 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44e00510 0x4>;
+};
+
+timer5_fck: timer5_fck at 44e00518 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44e00518 0x4>;
+};
+
+timer6_fck: timer6_fck at 44e0051c {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44e0051c 0x4>;
+};
+
+timer7_fck: timer7_fck at 44e00504 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44e00504 0x4>;
+};
+
+usbotg_fck: usbotg_fck at 44e0047c {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&dpll_per_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x44e0047c 0x4>;
+};
+
+dpll_core_m4_div2_ck: dpll_core_m4_div2_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+ieee5000_fck: ieee5000_fck at 44e000e4 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&dpll_core_m4_div2_ck>;
+	ti,bit-shift = <1>;
+	reg = <0x44e000e4 0x4>;
+};
+
+wdt1_fck: wdt1_fck at 44e00538 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&clk_rc32k_ck>, <&clkdiv32k_ick>;
+	reg = <0x44e00538 0x4>;
+};
+
+l4_rtc_gclk: l4_rtc_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+l4hs_gclk: l4hs_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+l3s_gclk: l3s_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_div2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+l4fw_gclk: l4fw_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_div2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+l4ls_gclk: l4ls_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_div2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+sysclk_div_ck: sysclk_div_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+cpsw_125mhz_gclk: cpsw_125mhz_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m5_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+cpsw_cpts_rft_clk: cpsw_cpts_rft_clk at 44e00520 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&dpll_core_m5_ck>, <&dpll_core_m4_ck>;
+	reg = <0x44e00520 0x4>;
+};
+
+gpio0_dbclk_mux_ck: gpio0_dbclk_mux_ck at 44e0053c {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&clk_rc32k_ck>, <&clk_32768_ck>, <&clkdiv32k_ick>;
+	reg = <0x44e0053c 0x4>;
+};
+
+gpio0_dbclk: gpio0_dbclk at 44e00408 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&gpio0_dbclk_mux_ck>;
+	ti,bit-shift = <18>;
+	reg = <0x44e00408 0x4>;
+};
+
+gpio1_dbclk: gpio1_dbclk at 44e000ac {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkdiv32k_ick>;
+	ti,bit-shift = <18>;
+	reg = <0x44e000ac 0x4>;
+};
+
+gpio2_dbclk: gpio2_dbclk at 44e000b0 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkdiv32k_ick>;
+	ti,bit-shift = <18>;
+	reg = <0x44e000b0 0x4>;
+};
+
+gpio3_dbclk: gpio3_dbclk at 44e000b4 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkdiv32k_ick>;
+	ti,bit-shift = <18>;
+	reg = <0x44e000b4 0x4>;
+};
+
+lcd_gclk: lcd_gclk at 44e00534 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&dpll_disp_m2_ck>, <&dpll_core_m5_ck>, <&dpll_per_m2_ck>;
+	reg = <0x44e00534 0x4>;
+	ti,set-rate-parent;
+};
+
+mmc_clk: mmc_clk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+gfx_fclk_clksel_ck: gfx_fclk_clksel_ck at 44e0052c {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&dpll_core_m4_ck>, <&dpll_per_m2_ck>;
+	ti,bit-shift = <1>;
+	reg = <0x44e0052c 0x4>;
+};
+
+gfx_fck_div_ck: gfx_fck_div_ck at 44e0052c {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&gfx_fclk_clksel_ck>;
+	reg = <0x44e0052c 0x4>;
+	ti,dividers = <1>, <2>;
+};
+
+sysclkout_pre_ck: sysclkout_pre_ck at 44e00700 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&clk_32768_ck>, <&l3_gclk>, <&dpll_ddr_m2_ck>, <&dpll_per_m2_ck>, <&lcd_gclk>;
+	reg = <0x44e00700 0x4>;
+};
+
+clkout2_div_ck: clkout2_div_ck at 44e00700 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&sysclkout_pre_ck>;
+	ti,bit-shift = <3>;
+	reg = <0x44e00700 0x4>;
+	ti,dividers = <1>, <2>, <3>, <4>, <5>, <6>, <7>, <8>;
+};
+
+dbg_sysclk_ck: dbg_sysclk_ck at 44e00414 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_clkin_ck>;
+	ti,bit-shift = <19>;
+	reg = <0x44e00414 0x4>;
+};
+
+dbg_clka_ck: dbg_clka_ck at 44e00414 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&dpll_core_m4_ck>;
+	ti,bit-shift = <30>;
+	reg = <0x44e00414 0x4>;
+};
+
+stm_pmd_clock_mux_ck: stm_pmd_clock_mux_ck at 44e00414 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&dbg_sysclk_ck>, <&dbg_clka_ck>;
+	ti,bit-shift = <22>;
+	reg = <0x44e00414 0x4>;
+};
+
+trace_pmd_clk_mux_ck: trace_pmd_clk_mux_ck at 44e00414 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&dbg_sysclk_ck>, <&dbg_clka_ck>;
+	ti,bit-shift = <20>;
+	reg = <0x44e00414 0x4>;
+};
+
+stm_clk_div_ck: stm_clk_div_ck at 44e00414 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&stm_pmd_clock_mux_ck>;
+	ti,bit-shift = <27>;
+	ti,max-div = <64>;
+	reg = <0x44e00414 0x4>;
+	ti,index-power-of-two;
+};
+
+trace_clk_div_ck: trace_clk_div_ck at 44e00414 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&trace_pmd_clk_mux_ck>;
+	ti,bit-shift = <24>;
+	ti,max-div = <64>;
+	reg = <0x44e00414 0x4>;
+	ti,index-power-of-two;
+};
+
+clkout2_ck: clkout2_ck at 44e00700 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkout2_div_ck>;
+	ti,bit-shift = <7>;
+	reg = <0x44e00700 0x4>;
+};
+
+ehrpwm0_tbclk: ehrpwm0_tbclk at 44e10664 {
+	#clock-cells = <0>;
+	compatible = "ti,no-wait-mux-gate-clock";
+	clocks = <&dpll_per_m2_ck>;
+	ti,gate-bit-shift = <0>;
+	reg = <0x44e10664 0x4>;
+	reg-names = "gate-reg";
+};
+
+ehrpwm1_tbclk: ehrpwm1_tbclk at 44e10664 {
+	#clock-cells = <0>;
+	compatible = "ti,no-wait-mux-gate-clock";
+	clocks = <&dpll_per_m2_ck>;
+	ti,gate-bit-shift = <1>;
+	reg = <0x44e10664 0x4>;
+	reg-names = "gate-reg";
+};
+
+ehrpwm2_tbclk: ehrpwm2_tbclk at 44e10664 {
+	#clock-cells = <0>;
+	compatible = "ti,no-wait-mux-gate-clock";
+	clocks = <&dpll_per_m2_ck>;
+	ti,gate-bit-shift = <2>;
+	reg = <0x44e10664 0x4>;
+	reg-names = "gate-reg";
+};
arch/arm/boot/dts/am33xx.dtsi +15 -1
index 553adc6..4fb2bf3 100644
@@ -711,4 +711,18 @@
 			status = "disabled";
 		};
 	};
-};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "am33xx-clocks.dtsi"
+	};
+
+	clockdomains {
+		clk_24mhz_clkdm: clk_24mhz_clkdm {
+			compatible = "ti,clockdomain";
+			clocks = <&clkdiv32k_ick>;
+		};
+	};
+};
\ No newline at end of file
Signature · 1 line
1.7.9.5
#26 lore raw

PATCH v7 25/36 CLK: TI: add am33xx clock init file

In reply to Tero Kristo

clk-33xx.c now contains the clock init functionality for am33xx, including DT clock registration and adding of static clkdev entries.

This patch also moves the omap2_clk_enable_init_clocks declaration to the driver include, as this is needed by the am33xx clock init code.

---
 arch/arm/mach-omap2/clock.h |    1 -
 drivers/clk/ti/clk-33xx.c   |  163 +++++++++++++++++++++++++++++++++++++++++++
 include/linux/clk/ti.h      |    2 +
 3 files changed, 165 insertions(+), 1 deletion(-)
 create mode 100644 drivers/clk/ti/clk-33xx.c
arch/arm/mach-omap2/clock.h +0 -1
index 106c5c2..2e6cbeb 100644
@@ -247,7 +247,6 @@ void omap2_clk_dflt_find_idlest(struct clk_hw_omap *clk,
 				void __iomem **idlest_reg,
 				u8 *idlest_bit, u8 *idlest_val);
 int omap2_clk_enable_autoidle_all(void);
-void omap2_clk_enable_init_clocks(const char **clk_names, u8 num_clocks);
 int omap2_clk_switch_mpurate_at_boot(const char *mpurate_ck_name);
 void omap2_clk_print_new_rates(const char *hfclkin_ck_name,
 			       const char *core_ck_name,
drivers/clk/ti/clk-33xx.c new file +163 -0

163 lines not on this page · raw message

include/linux/clk/ti.h +2 -0
index 1b2dc51..0748581 100644
@@ -208,6 +208,7 @@ unsigned long omap3_clkoutx2_recalc(struct clk_hw *hw,
 int omap2_clkops_enable_clkdm(struct clk_hw *hw);
 void omap2_clkops_disable_clkdm(struct clk_hw *hw);
 int omap2_clk_disable_autoidle_all(void);
+void omap2_clk_enable_init_clocks(const char **clk_names, u8 num_clocks);
 int omap3_dpll4_set_rate(struct clk_hw *clk, unsigned long rate,
 			 unsigned long parent_rate);
 int omap2_dflt_clk_enable(struct clk_hw *hw);
@@ -222,6 +223,7 @@ struct clk_div_table *ti_clk_get_div_table(struct device_node *node);
 int omap4xxx_clk_init(void);
 int omap5xxx_clk_init(void);
 int dra7xx_clk_init(void);
+int am33xx_clk_init(void);
 
 #ifdef CONFIG_OF
 void of_ti_clk_allow_autoidle_all(void);
Signature · 1 line
1.7.9.5
#27 lore raw

PATCH v7 26/36 ARM: AM33xx: remove old clock data and link in new clock init code

In reply to Tero Kristo

AM33xx clocks have now been moved to DT, thus remove the old data file and use the new init code under OMAP clock driver.

---
 arch/arm/mach-omap2/Makefile          |    1 -
 arch/arm/mach-omap2/cclock33xx_data.c | 1064 ---------------------------------
 drivers/clk/ti/Makefile               |    1 +
 3 files changed, 1 insertion(+), 1065 deletions(-)
 delete mode 100644 arch/arm/mach-omap2/cclock33xx_data.c
arch/arm/mach-omap2/Makefile +0 -1
index f518360..020079e 100644
@@ -187,7 +187,6 @@ obj-$(CONFIG_ARCH_OMAP3)		+= clkt_iclk.o
 obj-$(CONFIG_ARCH_OMAP4)		+= $(clock-common)
 obj-$(CONFIG_ARCH_OMAP4)		+= dpll3xxx.o dpll44xx.o
 obj-$(CONFIG_SOC_AM33XX)		+= $(clock-common) dpll3xxx.o
-obj-$(CONFIG_SOC_AM33XX)		+= cclock33xx_data.o
 obj-$(CONFIG_SOC_OMAP5)			+= $(clock-common)
 obj-$(CONFIG_SOC_OMAP5)			+= dpll3xxx.o dpll44xx.o
 
arch/arm/mach-omap2/cclock33xx_data.c deleted +0 -1064

1,064 lines not on this page · raw message

drivers/clk/ti/Makefile +1 -0
index 1a83c3c..63b77f0 100644
@@ -2,6 +2,7 @@ ifneq ($(CONFIG_OF),)
 obj-y					+= clk.o dpll.o autoidle.o divider.o \
 					   fixed-factor.o gate.o clockdomain.o \
 					   mux-gate.o mux.o apll.o
+obj-$(CONFIG_SOC_AM33XX)		+= clk-33xx.o
 obj-$(CONFIG_ARCH_OMAP4)		+= clk-44xx.o
 obj-$(CONFIG_SOC_OMAP5)			+= clk-54xx.o
 obj-$(CONFIG_SOC_DRA7XX)		+= clk-7xx.o
Signature · 1 line
1.7.9.5
#28 lore raw

PATCH v7 27/36 CLK: TI: add interface clock support for OMAP3

In reply to Tero Kristo

OMAP3 has interface clocks in addition to functional clocks, which require special handling for the autoidle and idle status register offsets mainly.

---
 .../devicetree/bindings/clock/ti/interface.txt     |   52 ++++++++
 arch/arm/mach-omap2/clock.h                        |    6 -
 drivers/clk/ti/Makefile                            |    1 +
 drivers/clk/ti/interface.c                         |  124 ++++++++++++++++++++
 include/linux/clk/ti.h                             |    5 +
 5 files changed, 182 insertions(+), 6 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/clock/ti/interface.txt
 create mode 100644 drivers/clk/ti/interface.c
Documentation/devicetree/bindings/clock/ti/interface.txt new file +52 -0
new file mode 100644
index 0000000..f3e43a9
@@ -0,0 +1,52 @@
+Binding for Texas Instruments interface clock.
+
+This binding uses the common clock binding[1]. This clock is
+quite much similar to the basic gate-clock [2], however,
+it supports a number of additional features, including
+companion clock finding (match corresponding functional gate
+clock) and hardware autoidle enable / disable.
+
+[1] Documentation/devicetree/bindings/clock/clock-bindings.txt
+[2] Documentation/devicetree/bindings/clock/gate-clock.txt
+
+Required properties:
+- compatible : shall be one of:
+  "ti,omap3-interface-clock" - basic OMAP3 interface clock
+  "ti,omap3-no-wait-interface-clock" - interface clock which has no hardware
+				       capability for waiting clock to be ready
+  "ti,omap3-hsotgusb-interface-clock" - interface clock with USB specific HW
+					handling
+  "ti,omap3-dss-interface-clock" - interface clock with DSS specific HW handling
+  "ti,omap3-ssi-interface-clock" - interface clock with SSI specific HW handling
+  "ti,am35xx-interface-clock" - interface clock with AM35xx specific HW handling
+- #clock-cells : from common clock binding; shall be set to 0
+- clocks : link to phandle of parent clock
+- reg : base address for the control register
+
+Optional properties:
+- ti,bit-shift : bit shift for the bit enabling/disabling the clock (default 0)
+
+Examples:
+	aes1_ick: aes1_ick at 48004a14 {
+		#clock-cells = <0>;
+		compatible = "ti,omap3-interface-clock";
+		clocks = <&security_l4_ick2>;
+		reg = <0x48004a14 0x4>;
+		ti,bit-shift = <3>;
+	};
+
+	cam_ick: cam_ick at 48004f10 {
+		#clock-cells = <0>;
+		compatible = "ti,omap3-no-wait-interface-clock";
+		clocks = <&l4_ick>;
+		reg = <0x48004f10 0x4>;
+		ti,bit-shift = <0>;
+	};
+
+	ssi_ick_3430es2: ssi_ick_3430es2 at 48004a10 {
+		#clock-cells = <0>;
+		compatible = "ti,omap3-ssi-interface-clock";
+		clocks = <&ssi_l4_ick>;
+		reg = <0x48004a10 0x4>;
+		ti,bit-shift = <0>;
+	};
arch/arm/mach-omap2/clock.h +0 -6
index 2e6cbeb..d1b227e 100644
@@ -266,14 +266,8 @@ extern const struct clksel_rate gfx_l3_rates[];
 extern const struct clksel_rate dsp_ick_rates[];
 extern struct clk dummy_ck;
 
-extern const struct clk_hw_omap_ops clkhwops_iclk_wait;
-extern const struct clk_hw_omap_ops clkhwops_iclk;
 extern const struct clk_hw_omap_ops clkhwops_omap3430es2_ssi_wait;
-extern const struct clk_hw_omap_ops clkhwops_omap3430es2_iclk_ssi_wait;
-extern const struct clk_hw_omap_ops clkhwops_omap3430es2_iclk_dss_usbhost_wait;
-extern const struct clk_hw_omap_ops clkhwops_omap3430es2_iclk_hsotgusb_wait;
 extern const struct clk_hw_omap_ops clkhwops_omap3430es2_hsotgusb_wait;
-extern const struct clk_hw_omap_ops clkhwops_am35xx_ipss_wait;
 extern const struct clk_hw_omap_ops clkhwops_apll54;
 extern const struct clk_hw_omap_ops clkhwops_apll96;
 extern const struct clk_hw_omap_ops clkhwops_omap2xxx_dpll;
drivers/clk/ti/Makefile +1 -0
index 63b77f0..1ecb1a4 100644
@@ -3,6 +3,7 @@ obj-y					+= clk.o dpll.o autoidle.o divider.o \
 					   fixed-factor.o gate.o clockdomain.o \
 					   mux-gate.o mux.o apll.o
 obj-$(CONFIG_SOC_AM33XX)		+= clk-33xx.o
+obj-$(CONFIG_ARCH_OMAP3)		+= interface.o
 obj-$(CONFIG_ARCH_OMAP4)		+= clk-44xx.o
 obj-$(CONFIG_SOC_OMAP5)			+= clk-54xx.o
 obj-$(CONFIG_SOC_DRA7XX)		+= clk-7xx.o
drivers/clk/ti/interface.c new file +124 -0

124 lines not on this page · raw message

include/linux/clk/ti.h +5 -0
index 0748581..88d7497 100644
@@ -239,5 +239,10 @@ extern const struct clk_hw_omap_ops clkhwops_wait;
 extern const struct clk_hw_omap_ops clkhwops_iclk_wait;
 extern const struct clk_hw_omap_ops clkhwops_omap3430es2_dss_usbhost_wait;
 extern const struct clk_hw_omap_ops clkhwops_am35xx_ipss_module_wait;
+extern const struct clk_hw_omap_ops clkhwops_am35xx_ipss_wait;
+extern const struct clk_hw_omap_ops clkhwops_iclk;
+extern const struct clk_hw_omap_ops clkhwops_omap3430es2_iclk_ssi_wait;
+extern const struct clk_hw_omap_ops clkhwops_omap3430es2_iclk_dss_usbhost_wait;
+extern const struct clk_hw_omap_ops clkhwops_omap3430es2_iclk_hsotgusb_wait;
 
 #endif
Signature · 1 line
1.7.9.5
#29 lore raw

PATCH v7 28/36 ARM: OMAP: hwmod: fix an incorrect clk type cast with _get_clkdm

In reply to Tero Kristo

If the main clock for a hwmod is of basic clock type, it is illegal to type cast this to clk_hw_omap and will result in bogus data. Fixed by checking the clock flags before attempting the type cast.

---
 arch/arm/mach-omap2/omap_hwmod.c |    2 ++
 1 file changed, 2 insertions(+)
arch/arm/mach-omap2/omap_hwmod.c +2 -0
index d35145d..5f8f587 100644
@@ -656,6 +656,8 @@ static struct clockdomain *_get_clkdm(struct omap_hwmod *oh)
 	if (oh->clkdm) {
 		return oh->clkdm;
 	} else if (oh->_clk) {
+		if (__clk_get_flags(oh->_clk) & CLK_IS_BASIC)
+			return NULL;
 		clk = to_clk_hw_omap(__clk_get_hw(oh->_clk));
 		return  clk->clkdm;
 	}
Signature · 1 line
1.7.9.5
#30 lore raw

PATCH v7 29/36 ARM: OMAP3: hwmod: initialize clkdm from clkdm_name

In reply to Tero Kristo

DT clocks are mostly missing clkdm info now, and this causes an issue with counter32k which makes its slave idlemode wrong and prevents core idle.

Fixed by initializing the hwmod clkdm pointers for omap3 also which makes sure the clkdm flag matching logic works properly.

This patch also changes the return value for _init_clkdm to 0 for incorrect clkdm_name, as this a warning, not a fatal error.

---
 arch/arm/mach-omap2/omap_hwmod.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/arm/mach-omap2/omap_hwmod.c +2 -1
index 5f8f587..ae33fae 100644
@@ -1546,7 +1546,7 @@ static int _init_clkdm(struct omap_hwmod *oh)
 	if (!oh->clkdm) {
 		pr_warning("omap_hwmod: %s: could not associate to clkdm %s\n",
 			oh->name, oh->clkdm_name);
-		return -EINVAL;
+		return 0;
 	}
 
 	pr_debug("omap_hwmod: %s: associated to clkdm %s\n",
@@ -4117,6 +4117,7 @@ void __init omap_hwmod_init(void)
 		soc_ops.assert_hardreset = _omap2_assert_hardreset;
 		soc_ops.deassert_hardreset = _omap2_deassert_hardreset;
 		soc_ops.is_hardreset_asserted = _omap2_is_hardreset_asserted;
+		soc_ops.init_clkdm = _init_clkdm;
 	} else if (cpu_is_omap44xx() || soc_is_omap54xx() || soc_is_dra7xx()) {
 		soc_ops.enable_module = _omap4_enable_module;
 		soc_ops.disable_module = _omap4_disable_module;
Signature · 1 line
1.7.9.5
#31 lore raw

PATCH v7 30/36 ARM: dts: omap3 clock data

In reply to Tero Kristo

This patch creates a unique node for each clock in the OMAP3 power, reset and clock manager (PRCM).

TODO: add still missing am35xx only clock nodes

---
 arch/arm/boot/dts/am35xx-clocks.dtsi               |  101 ++
 arch/arm/boot/dts/omap3.dtsi                       |    7 +
 arch/arm/boot/dts/omap3430es1-clocks.dtsi          |  145 ++
 arch/arm/boot/dts/omap34xx-omap36xx-clocks.dtsi    |  223 ++++
 arch/arm/boot/dts/omap34xx.dtsi                    |  121 +-
 .../omap36xx-am35xx-omap3430es2plus-clocks.dtsi    |  186 +++
 arch/arm/boot/dts/omap36xx-clocks.dtsi             |   78 ++
 .../boot/dts/omap36xx-omap3430es2plus-clocks.dtsi  |  158 +++
 arch/arm/boot/dts/omap36xx.dtsi                    |  122 +-
 arch/arm/boot/dts/omap3xxx-clocks.dtsi             | 1392 ++++++++++++++++++++
 10 files changed, 2531 insertions(+), 2 deletions(-)
 create mode 100644 arch/arm/boot/dts/am35xx-clocks.dtsi
 create mode 100644 arch/arm/boot/dts/omap3430es1-clocks.dtsi
 create mode 100644 arch/arm/boot/dts/omap34xx-omap36xx-clocks.dtsi
 create mode 100644 arch/arm/boot/dts/omap36xx-am35xx-omap3430es2plus-clocks.dtsi
 create mode 100644 arch/arm/boot/dts/omap36xx-clocks.dtsi
 create mode 100644 arch/arm/boot/dts/omap36xx-omap3430es2plus-clocks.dtsi
 create mode 100644 arch/arm/boot/dts/omap3xxx-clocks.dtsi
arch/arm/boot/dts/am35xx-clocks.dtsi new file +101 -0
new file mode 100644
index 0000000..7f3af1a
@@ -0,0 +1,101 @@
+/*
+ * Device Tree Source for AM35xx clock data
+ *
+ * Copyright (C) 2013 Texas Instruments, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+ipss_ick: ipss_ick at 48004a10 {
+	#clock-cells = <0>;
+	compatible = "ti,am35xx-interface-clock";
+	clocks = <&core_l3_ick>;
+	reg = <0x48004a10 0x4>;
+	ti,bit-shift = <4>;
+};
+
+rmii_ck: rmii_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <50000000>;
+};
+
+pclk_ck: pclk_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <27000000>;
+};
+
+emac_ick: emac_ick at 4800259c {
+	#clock-cells = <0>;
+	compatible = "ti,am35xx-gate-clock";
+	clocks = <&ipss_ick>;
+	reg = <0x4800259c 0x4>;
+	ti,bit-shift = <1>;
+};
+
+emac_fck: emac_fck at 4800259c {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&rmii_ck>;
+	reg = <0x4800259c 0x4>;
+	ti,bit-shift = <9>;
+};
+
+vpfe_ick: vpfe_ick at 4800259c {
+	#clock-cells = <0>;
+	compatible = "ti,am35xx-gate-clock";
+	clocks = <&ipss_ick>;
+	reg = <0x4800259c 0x4>;
+	ti,bit-shift = <2>;
+};
+
+vpfe_fck: vpfe_fck at 4800259c {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&pclk_ck>;
+	reg = <0x4800259c 0x4>;
+	ti,bit-shift = <10>;
+};
+
+hsotgusb_ick_am35xx: hsotgusb_ick_am35xx at 4800259c {
+	#clock-cells = <0>;
+	compatible = "ti,am35xx-gate-clock";
+	clocks = <&ipss_ick>;
+	reg = <0x4800259c 0x4>;
+	ti,bit-shift = <0>;
+};
+
+hsotgusb_fck_am35xx: hsotgusb_fck_am35xx at 4800259c {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&sys_ck>;
+	reg = <0x4800259c 0x4>;
+	ti,bit-shift = <8>;
+};
+
+hecc_ck: hecc_ck at 4800259c {
+	#clock-cells = <0>;
+	compatible = "ti,am35xx-gate-clock";
+	clocks = <&sys_ck>;
+	reg = <0x4800259c 0x4>;
+	ti,bit-shift = <3>;
+};
+
+uart4_ick_am35xx: uart4_ick_am35xx at 48004a10 {
+	#clock-cells = <0>;
+	compatible = "ti,omap3-interface-clock";
+	clocks = <&core_l4_ick>;
+	reg = <0x48004a10 0x4>;
+	ti,bit-shift = <23>;
+};
+
+uart4_fck_am35xx: uart4_fck_am35xx at 48004a00 {
+	#clock-cells = <0>;
+	compatible = "ti,wait-gate-clock";
+	clocks = <&core_48m_fck>;
+	reg = <0x48004a00 0x4>;
+	ti,bit-shift = <23>;
+};
arch/arm/boot/dts/omap3.dtsi +7 -0
index 16420ae..bc11b83 100644
@@ -533,4 +533,11 @@
 			ram-bits = <12>;
 		};
 	};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "omap3xxx-clocks.dtsi"
+	};
 };
arch/arm/boot/dts/omap3430es1-clocks.dtsi new file +145 -0

145 lines not on this page · raw message

arch/arm/boot/dts/omap34xx-omap36xx-clocks.dtsi new file +223 -0

223 lines not on this page · raw message

arch/arm/boot/dts/omap34xx.dtsi +120 -1

125 lines not on this page · raw message

arch/arm/boot/dts/omap36xx-am35xx-omap3430es2plus-clocks.dtsi new file +186 -0

186 lines not on this page · raw message

arch/arm/boot/dts/omap36xx-clocks.dtsi new file +78 -0

78 lines not on this page · raw message

arch/arm/boot/dts/omap36xx-omap3430es2plus-clocks.dtsi new file +158 -0

158 lines not on this page · raw message

arch/arm/boot/dts/omap36xx.dtsi +121 -1

126 lines not on this page · raw message

arch/arm/boot/dts/omap3xxx-clocks.dtsi new file +1392 -0

1,392 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#32 lore raw

PATCH v7 30/36 ARM: dts: omap3 clock data

Quoted 6 lines · Tero Kristo
arch/arm/boot/dts/omap3.dtsi +0 -0
index 16420ae..bc11b83 100644
@@ -533,4 +533,11 @@
 			ram-bits = <12>;

arch/arm/boot/dts/omap3.dtsi line 543

Quoted 10 lines
arch/arm/boot/dts/omap3.dtsi +7 -0
 		};
 	};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "omap3xxx-clocks.dtsi"
+	};
 };

Clocks are introduced towards the tail of the dts -> this has a problem associated with it -> device nodes should be able to reference phandle like:

devicex {
	clocks = <&sys_ck>;
}

Since all the devices on ocp and cpu0 node appears above the definition they fail to catch the phandle. instead, moving the clocks node as high in the tree as possible resolves this: something like:

arch/arm/boot/dts/omap3.dtsi +7 -6
index bc11b83..0b2161d 100644
@@ -24,6 +24,13 @@
 		serial2 = &uart3;
 	};
 
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "omap3xxx-clocks.dtsi"
+	};
+
 	cpus {
 		#address-cells = <1>;
 		#size-cells = <0>;
@@ -534,10 +541,4 @@
 		};
 	};
 
-	clocks {
-		#address-cells = <1>;
-		#size-cells = <1>;
-		ranges;
-		/include/ "omap3xxx-clocks.dtsi"
-	};
 };

What do you think of the change? applies for all clock nodes in other dtsi as well - I will mark them as I find them.

Further,

Quoted 5 lines
arch/arm/boot/dts/omap34xx.dtsi +0 -0
index 5355d61..2ed7c69 100644

arch/arm/boot/dts/omap34xx.dtsi line 32

Quoted 9 lines
arch/arm/boot/dts/omap34xx.dtsi +5 -1
@@ -25,4 +25,123 @@
 			clock-latency = <300000>; /* From legacy driver */
 		};
 	};
-};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;

dont need to state the above two -> already done in omap3.dtsi.

Quoted 5 lines
arch/arm/boot/dts/omap34xx.dtsi +5 -0
+		/include/ "omap34xx-omap36xx-clocks.dtsi"
+		/include/ "omap36xx-omap3430es2plus-clocks.dtsi"
+		/include/ "omap36xx-am35xx-omap3430es2plus-clocks.dtsi"
+	};
+};

[...]

Quoted 1 line

\ No newline at end of file

^^
[...]

Quoted 5 lines
arch/arm/boot/dts/omap36xx.dtsi +0 -0
index f8b3765..71fb6fb 100644

arch/arm/boot/dts/omap36xx.dtsi line 42

Quoted 9 lines
arch/arm/boot/dts/omap36xx.dtsi +5 -1
@@ -35,4 +35,124 @@
 			clock-frequency = <48000000>;
 		};
 	};
-};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;

^^ here as well.

Quoted 5 lines
arch/arm/boot/dts/omap36xx.dtsi +5 -0
+		/include/ "omap36xx-clocks.dtsi"
+		/include/ "omap34xx-omap36xx-clocks.dtsi"
+		/include/ "omap36xx-omap3430es2plus-clocks.dtsi"
+		/include/ "omap36xx-am35xx-omap3430es2plus-clocks.dtsi"
+	};

[...]

Quoted 2 lines

+};
\ No newline at end of file

^^ this need fix as well.

Signature · 2 lines
Regards,
Nishanth Menon
#33 lore raw

PATCH v7 30/36 ARM: dts: omap3 clock data

Quoted 58 lines · Nishanth Menon
Quoted 6 lines · Tero Kristo
arch/arm/boot/dts/omap3.dtsi +0 -0
index 16420ae..bc11b83 100644
@@ -533,4 +533,11 @@
 			ram-bits = <12>;

arch/arm/boot/dts/omap3.dtsi line 543

Quoted 10 lines
arch/arm/boot/dts/omap3.dtsi +7 -0
 		};
 	};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "omap3xxx-clocks.dtsi"
+	};
 };

Clocks are introduced towards the tail of the dts -> this has a problem associated with it -> device nodes should be able to reference phandle like:

devicex {
	clocks = <&sys_ck>;
}

Since all the devices on ocp and cpu0 node appears above the definition they fail to catch the phandle. instead, moving the clocks node as high in the tree as possible resolves this: something like:

arch/arm/boot/dts/omap3.dtsi +7 -6
index bc11b83..0b2161d 100644
@@ -24,6 +24,13 @@
 		serial2 = &uart3;
 	};
 
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "omap3xxx-clocks.dtsi"
+	};
+
 	cpus {
 		#address-cells = <1>;
 		#size-cells = <0>;
@@ -534,10 +541,4 @@
 		};
 	};
 
-	clocks {
-		#address-cells = <1>;
-		#size-cells = <1>;
-		ranges;
-		/include/ "omap3xxx-clocks.dtsi"
-	};
 };
Quoted 2 lines

What do you think of the change? applies for all clock nodes in other dtsi as well - I will mark them as I find them.

Digging further, the above is not needed. Apologies on the noise. below changes will be nice though..

Quoted 51 lines

Further,

Quoted 5 lines
arch/arm/boot/dts/omap34xx.dtsi +0 -0
index 5355d61..2ed7c69 100644

arch/arm/boot/dts/omap34xx.dtsi line 32

Quoted 9 lines
arch/arm/boot/dts/omap34xx.dtsi +5 -1
@@ -25,4 +25,123 @@
 			clock-latency = <300000>; /* From legacy driver */
 		};
 	};
-};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;

dont need to state the above two -> already done in omap3.dtsi.

Quoted 5 lines
arch/arm/boot/dts/omap34xx.dtsi +5 -0
+		/include/ "omap34xx-omap36xx-clocks.dtsi"
+		/include/ "omap36xx-omap3430es2plus-clocks.dtsi"
+		/include/ "omap36xx-am35xx-omap3430es2plus-clocks.dtsi"
+	};
+};

[...]

Quoted 1 line

\ No newline at end of file

^^
[...]

Quoted 5 lines
arch/arm/boot/dts/omap36xx.dtsi +0 -0
index f8b3765..71fb6fb 100644

arch/arm/boot/dts/omap36xx.dtsi line 42

Quoted 9 lines
arch/arm/boot/dts/omap36xx.dtsi +5 -1
@@ -35,4 +35,124 @@
 			clock-frequency = <48000000>;
 		};
 	};
-};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;

^^ here as well.

Quoted 5 lines
arch/arm/boot/dts/omap36xx.dtsi +5 -0
+		/include/ "omap36xx-clocks.dtsi"
+		/include/ "omap34xx-omap36xx-clocks.dtsi"
+		/include/ "omap36xx-omap3430es2plus-clocks.dtsi"
+		/include/ "omap36xx-am35xx-omap3430es2plus-clocks.dtsi"
+	};

[...]

Quoted 2 lines

+};
\ No newline at end of file

^^ this need fix as well.

Signature · 2 lines
Regards,
Nishanth Menon
#34 lore raw

PATCH v7 31/36 CLK: TI: add omap3 clock init file

In reply to Tero Kristo

clk-3xxx.c now contains the clock init functionality for omap3, including DT clock registration and adding of static clkdev entries.

---
 arch/arm/mach-omap2/clock3xxx.h |    1 -
 drivers/clk/ti/Makefile         |    2 +-
 drivers/clk/ti/clk-3xxx.c       |  402 +++++++++++++++++++++++++++++++++++++++
 include/linux/clk/ti.h          |    5 +
 4 files changed, 408 insertions(+), 2 deletions(-)
 create mode 100644 drivers/clk/ti/clk-3xxx.c
arch/arm/mach-omap2/clock3xxx.h +0 -1
index dab90e2..78d9f56 100644
@@ -11,7 +11,6 @@
 int omap3xxx_clk_init(void);
 int omap3_core_dpll_m2_set_rate(struct clk_hw *clk, unsigned long rate,
 					unsigned long parent_rate);
-void omap3_clk_lock_dpll5(void);
 
 extern struct clk *sdrc_ick_p;
 extern struct clk *arm_fck_p;
drivers/clk/ti/Makefile +1 -1
index 1ecb1a4..df226bc 100644
@@ -3,7 +3,7 @@ obj-y					+= clk.o dpll.o autoidle.o divider.o \
 					   fixed-factor.o gate.o clockdomain.o \
 					   mux-gate.o mux.o apll.o
 obj-$(CONFIG_SOC_AM33XX)		+= clk-33xx.o
-obj-$(CONFIG_ARCH_OMAP3)		+= interface.o
+obj-$(CONFIG_ARCH_OMAP3)		+= interface.o clk-3xxx.o
 obj-$(CONFIG_ARCH_OMAP4)		+= clk-44xx.o
 obj-$(CONFIG_SOC_OMAP5)			+= clk-54xx.o
 obj-$(CONFIG_SOC_DRA7XX)		+= clk-7xx.o
drivers/clk/ti/clk-3xxx.c new file +402 -0

402 lines not on this page · raw message

include/linux/clk/ti.h +5 -0
index 88d7497..bf8beff 100644
@@ -214,12 +214,17 @@ int omap3_dpll4_set_rate(struct clk_hw *clk, unsigned long rate,
 int omap2_dflt_clk_enable(struct clk_hw *hw);
 void omap2_dflt_clk_disable(struct clk_hw *hw);
 int omap2_dflt_clk_is_enabled(struct clk_hw *hw);
+void omap3_clk_lock_dpll5(void);
 
 void ti_dt_clocks_register(struct ti_dt_clk *oclks);
 void of_ti_autoidle_setup(struct device_node *node);
 
 struct clk_div_table *ti_clk_get_div_table(struct device_node *node);
 
+int omap3430_clk_init(void);
+int omap3630_clk_init(void);
+int am35xx_clk_init(void);
+int ti81xx_clk_init(void);
 int omap4xxx_clk_init(void);
 int omap5xxx_clk_init(void);
 int dra7xx_clk_init(void);
Signature · 1 line
1.7.9.5
#35 lore raw

PATCH v7 32/36 ARM: dts: AM35xx: use DT clock data

In reply to Tero Kristo

AM35xx now uses the clock data from device tree. Most of the data is shared with OMAP3xxx, but as there is some delta, a new base .dtsi file is also created for the SoC.

---
 arch/arm/boot/dts/am3517-evm.dts        |    2 +-
 arch/arm/boot/dts/am3517.dtsi           |  126 +++++++++++++++++++++++++++++++
 arch/arm/boot/dts/am3517_mt_ventoux.dts |    2 +-
 3 files changed, 128 insertions(+), 2 deletions(-)
 create mode 100644 arch/arm/boot/dts/am3517.dtsi
arch/arm/boot/dts/am3517-evm.dts +1 -1
index e99dfaf..9ff51d7 100644
@@ -7,7 +7,7 @@
  */
 /dts-v1/;
 
-#include "omap34xx.dtsi"
+#include "am3517.dtsi"
 
 / {
 	model = "TI AM3517 EVM (AM3517/05)";
arch/arm/boot/dts/am3517.dtsi new file +126 -0

126 lines not on this page · raw message

arch/arm/boot/dts/am3517_mt_ventoux.dts +1 -1
index fdf5ce6..d00e934 100644
@@ -7,7 +7,7 @@
  */
 /dts-v1/;
 
-#include "omap34xx.dtsi"
+#include "am3517.dtsi"
 
 / {
 	model = "TeeJet Mt.Ventoux";
Signature · 1 line
1.7.9.5
#36 lore raw

PATCH v7 33/36 ARM: OMAP3: use DT clock init if DT data is available

In reply to Tero Kristo

OMAP3 platforms support both DT and non-DT boot at the moment, make the clock init work according to the used setup.

---
 arch/arm/mach-omap2/clock3xxx.h |    4 ++++
 arch/arm/mach-omap2/io.c        |   13 ++++++++++++-
 2 files changed, 16 insertions(+), 1 deletion(-)
arch/arm/mach-omap2/clock3xxx.h +4 -0
index 78d9f56..a5e2308 100644
@@ -9,6 +9,10 @@
 #define __ARCH_ARM_MACH_OMAP2_CLOCK3XXX_H
 
 int omap3xxx_clk_init(void);
+int omap3430_clk_init(void);
+int omap3630_clk_init(void);
+int ti81xx_clk_init(void);
+int am35xx_clk_init(void);
 int omap3_core_dpll_m2_set_rate(struct clk_hw *clk, unsigned long rate,
 					unsigned long parent_rate);
 
arch/arm/mach-omap2/io.c +12 -1
index 1f23182..315a148 100644
@@ -488,21 +488,29 @@ void __init omap3_init_early(void)
 void __init omap3430_init_early(void)
 {
 	omap3_init_early();
+	if (of_have_populated_dt())
+		omap_clk_init = omap3430_clk_init;
 }
 
 void __init omap35xx_init_early(void)
 {
 	omap3_init_early();
+	if (of_have_populated_dt())
+		omap_clk_init = omap3430_clk_init;
 }
 
 void __init omap3630_init_early(void)
 {
 	omap3_init_early();
+	if (of_have_populated_dt())
+		omap_clk_init = omap3630_clk_init;
 }
 
 void __init am35xx_init_early(void)
 {
 	omap3_init_early();
+	if (of_have_populated_dt())
+		omap_clk_init = am35xx_clk_init;
 }
 
 void __init ti81xx_init_early(void)
@@ -520,7 +528,10 @@ void __init ti81xx_init_early(void)
 	omap3xxx_clockdomains_init();
 	omap3xxx_hwmod_init();
 	omap_hwmod_init_postsetup();
-	omap_clk_init = omap3xxx_clk_init;
+	if (of_have_populated_dt())
+		omap_clk_init = ti81xx_clk_init;
+	else
+		omap_clk_init = omap3xxx_clk_init;
 }
 
 void __init omap3_init_late(void)
Signature · 1 line
1.7.9.5
#37 lore raw

PATCH v7 34/36 ARM: dts: am43xx clock data

In reply to Tero Kristo

This patch creates a unique node for each clock in the AM43xx power, reset and clock manager (PRCM).

---
 arch/arm/boot/dts/am43xx-clocks.dtsi |  656 ++++++++++++++++++++++++++++++++++
 1 file changed, 656 insertions(+)
 create mode 100644 arch/arm/boot/dts/am43xx-clocks.dtsi
arch/arm/boot/dts/am43xx-clocks.dtsi new file +656 -0
new file mode 100644
index 0000000..09eaa2a
@@ -0,0 +1,656 @@
+/*
+ * Device Tree Source for AM43xx clock data
+ *
+ * Copyright (C) 2013 Texas Instruments, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+clk_32768_ck: clk_32768_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <32768>;
+};
+
+clk_rc32k_ck: clk_rc32k_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <32768>;
+};
+
+virt_19200000_ck: virt_19200000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <19200000>;
+};
+
+virt_24000000_ck: virt_24000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <24000000>;
+};
+
+virt_25000000_ck: virt_25000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <25000000>;
+};
+
+virt_26000000_ck: virt_26000000_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <26000000>;
+};
+
+sys_clkin_ck: sys_clkin_ck at 44e10040 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&virt_19200000_ck>, <&virt_24000000_ck>, <&virt_25000000_ck>, <&virt_26000000_ck>;
+	ti,bit-shift = <22>;
+	reg = <0x44e10040 0x4>;
+};
+
+tclkin_ck: tclkin_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <26000000>;
+};
+
+dpll_core_ck: dpll_core_ck at 44df2d20 {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-core-lock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44df2d20 0x4>, <0x44df2d24 0x4>, <0x44df2d2c 0x4>;
+	ti,low-power-bypass;
+	ti,lock;
+};
+
+dpll_core_x2_ck: dpll_core_x2_ck {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-x2-clock";
+	clocks = <&dpll_core_ck>;
+};
+
+dpll_core_m4_ck: dpll_core_m4_ck at 44df2d38 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x44df2d38 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_core_m5_ck: dpll_core_m5_ck at 44df2d3c {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x44df2d3c 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_core_m6_ck: dpll_core_m6_ck at 44df2d40 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_core_x2_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x44df2d40 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_mpu_ck: dpll_mpu_ck at 44df2d60 {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-clock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44df2d60 0x4>, <0x44df2d64 0x4>, <0x44df2d6c 0x4>;
+};
+
+dpll_mpu_m2_ck: dpll_mpu_m2_ck at 44df2d70 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_mpu_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x44df2d70 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_ddr_ck: dpll_ddr_ck at 44df2da0 {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-clock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44df2da0 0x4>, <0x44df2da4 0x4>, <0x44df2dac 0x4>;
+};
+
+dpll_ddr_m2_ck: dpll_ddr_m2_ck at 44df2db0 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_ddr_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x44df2db0 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_disp_ck: dpll_disp_ck at 44df2e20 {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-clock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44df2e20 0x4>, <0x44df2e24 0x4>, <0x44df2e2c 0x4>;
+};
+
+dpll_disp_m2_ck: dpll_disp_m2_ck at 44df2e30 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_disp_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x44df2e30 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_per_ck: dpll_per_ck at 44df2de0 {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-j-type-clock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44df2de0 0x4>, <0x44df2de4 0x4>, <0x44df2dec 0x4>;
+};
+
+dpll_per_m2_ck: dpll_per_m2_ck at 44df2df0 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_per_ck>;
+	ti,max-div = <127>;
+	ti,autoidle-shift = <8>;
+	reg = <0x44df2df0 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_per_m2_div4_wkupdm_ck: dpll_per_m2_div4_wkupdm_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <4>;
+};
+
+dpll_per_m2_div4_ck: dpll_per_m2_div4_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <4>;
+};
+
+adc_tsc_fck: adc_tsc_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+clk_24mhz: clk_24mhz {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <8>;
+};
+
+clkdiv32k_ck: clkdiv32k_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&clk_24mhz>;
+	clock-mult = <1>;
+	clock-div = <732>;
+};
+
+clkdiv32k_ick: clkdiv32k_ick at 44df2a38 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkdiv32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x44df2a38 0x4>;
+};
+
+dcan0_fck: dcan0_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dcan1_fck: dcan1_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+sysclk_div: sysclk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+pruss_ocp_gclk: pruss_ocp_gclk at 44df4248 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sysclk_div>, <&dpll_disp_m2_ck>;
+	reg = <0x44df4248 0x4>;
+};
+
+mcasp0_fck: mcasp0_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+mcasp1_fck: mcasp1_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+smartreflex0_fck: smartreflex0_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+smartreflex1_fck: smartreflex1_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+sha0_fck: sha0_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+aes0_fck: aes0_fck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+clk_32k_tpm_ck: clk_32k_tpm_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <32768>;
+};
+
+timer1_fck: timer1_fck at 44df4200 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sys_clkin_ck>, <&clkdiv32k_ick>, <&tclkin_ck>, <&clk_rc32k_ck>, <&clk_32768_ck>, <&clk_32k_tpm_ck>;
+	reg = <0x44df4200 0x4>;
+};
+
+timer2_fck: timer2_fck at 44df4204 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44df4204 0x4>;
+};
+
+timer3_fck: timer3_fck at 44df4208 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44df4208 0x4>;
+};
+
+timer4_fck: timer4_fck at 44df420c {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44df420c 0x4>;
+};
+
+timer5_fck: timer5_fck at 44df4210 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44df4210 0x4>;
+};
+
+timer6_fck: timer6_fck at 44df4214 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44df4214 0x4>;
+};
+
+timer7_fck: timer7_fck at 44df4218 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>;
+	reg = <0x44df4218 0x4>;
+};
+
+wdt1_fck: wdt1_fck at 44df422c {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&clk_rc32k_ck>, <&clkdiv32k_ick>;
+	reg = <0x44df422c 0x4>;
+};
+
+l3_gclk: l3_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dpll_core_m4_div2_ck: dpll_core_m4_div2_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sysclk_div>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+l4hs_gclk: l4hs_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+l3s_gclk: l3s_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_div2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+l4ls_gclk: l4ls_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m4_div2_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+cpsw_125mhz_gclk: cpsw_125mhz_gclk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m5_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+cpsw_cpts_rft_clk: cpsw_cpts_rft_clk at 44df4238 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sysclk_div>, <&dpll_core_m5_ck>, <&dpll_disp_m2_ck>;
+	reg = <0x44df4238 0x4>;
+};
+
+clk_32k_mosc_ck: clk_32k_mosc_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-clock";
+	clock-frequency = <32768>;
+};
+
+gpio0_dbclk_mux_ck: gpio0_dbclk_mux_ck at 44df4240 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&clk_rc32k_ck>, <&clk_32768_ck>, <&clkdiv32k_ick>, <&clk_32k_mosc_ck>, <&clk_32k_tpm_ck>;
+	reg = <0x44df4240 0x4>;
+};
+
+gpio0_dbclk: gpio0_dbclk at 44df2b68 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&gpio0_dbclk_mux_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x44df2b68 0x4>;
+};
+
+gpio1_dbclk: gpio1_dbclk at 44df8c78 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkdiv32k_ick>;
+	ti,bit-shift = <8>;
+	reg = <0x44df8c78 0x4>;
+};
+
+gpio2_dbclk: gpio2_dbclk at 44df8c80 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkdiv32k_ick>;
+	ti,bit-shift = <8>;
+	reg = <0x44df8c80 0x4>;
+};
+
+gpio3_dbclk: gpio3_dbclk at 44df8c88 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkdiv32k_ick>;
+	ti,bit-shift = <8>;
+	reg = <0x44df8c88 0x4>;
+};
+
+gpio4_dbclk: gpio4_dbclk at 44df8c90 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkdiv32k_ick>;
+	ti,bit-shift = <8>;
+	reg = <0x44df8c90 0x4>;
+};
+
+gpio5_dbclk: gpio5_dbclk at 44df8c98 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&clkdiv32k_ick>;
+	ti,bit-shift = <8>;
+	reg = <0x44df8c98 0x4>;
+};
+
+mmc_clk: mmc_clk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+gfx_fclk_clksel_ck: gfx_fclk_clksel_ck at 44df423c {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&sysclk_div>, <&dpll_per_m2_ck>;
+	ti,bit-shift = <1>;
+	reg = <0x44df423c 0x4>;
+};
+
+gfx_fck_div_ck: gfx_fck_div_ck at 44df423c {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&gfx_fclk_clksel_ck>;
+	reg = <0x44df423c 0x4>;
+	ti,max-div = <2>;
+};
+
+disp_clk: disp_clk at 44df4244 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&dpll_disp_m2_ck>, <&dpll_core_m5_ck>, <&dpll_per_m2_ck>;
+	reg = <0x44df4244 0x4>;
+};
+
+dpll_extdev_ck: dpll_extdev_ck at 44df2e60 {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-clock";
+	clocks = <&sys_clkin_ck>, <&sys_clkin_ck>;
+	reg = <0x44df2e60 0x4>, <0x44df2e64 0x4>, <0x44df2e6c 0x4>;
+};
+
+dpll_extdev_m2_ck: dpll_extdev_m2_ck at 44df2e70 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_extdev_ck>;
+	ti,max-div = <127>;
+	ti,autoidle-shift = <8>;
+	reg = <0x44df2e70 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+mux_synctimer32k_ck: mux_synctimer32k_ck at 44df4230 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&clk_32768_ck>, <&clk_32k_tpm_ck>, <&clkdiv32k_ick>;
+	reg = <0x44df4230 0x4>;
+};
+
+synctimer_32kclk: synctimer_32kclk at 44df2a30 {
+	#clock-cells = <0>;
+	compatible = "ti,gate-clock";
+	clocks = <&mux_synctimer32k_ck>;
+	ti,bit-shift = <8>;
+	reg = <0x44df2a30 0x4>;
+};
+
+timer8_fck: timer8_fck at 44df421c {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>, <&clk_32k_tpm_ck>;
+	reg = <0x44df421c 0x4>;
+};
+
+timer9_fck: timer9_fck at 44df4220 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>, <&clk_32k_tpm_ck>;
+	reg = <0x44df4220 0x4>;
+};
+
+timer10_fck: timer10_fck at 44df4224 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>, <&clk_32k_tpm_ck>;
+	reg = <0x44df4224 0x4>;
+};
+
+timer11_fck: timer11_fck at 44df4228 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&tclkin_ck>, <&sys_clkin_ck>, <&clkdiv32k_ick>, <&clk_32k_tpm_ck>;
+	reg = <0x44df4228 0x4>;
+};
+
+cpsw_50m_clkdiv: cpsw_50m_clkdiv {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_core_m5_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+cpsw_5m_clkdiv: cpsw_5m_clkdiv {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&cpsw_50m_clkdiv>;
+	clock-mult = <1>;
+	clock-div = <10>;
+};
+
+dpll_ddr_x2_ck: dpll_ddr_x2_ck {
+	#clock-cells = <0>;
+	compatible = "ti,am3-dpll-x2-clock";
+	clocks = <&dpll_ddr_ck>;
+};
+
+dpll_ddr_m4_ck: dpll_ddr_m4_ck at 44df2db8 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&dpll_ddr_x2_ck>;
+	ti,max-div = <31>;
+	ti,autoidle-shift = <8>;
+	reg = <0x44df2db8 0x4>;
+	ti,index-starts-at-one;
+	ti,autoidle-low;
+};
+
+dpll_per_clkdcoldo: dpll_per_clkdcoldo {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_ck>;
+	clock-mult = <1>;
+	clock-div = <1>;
+};
+
+dll_aging_clk_div: dll_aging_clk_div at 44df4250 {
+	#clock-cells = <0>;
+	compatible = "ti,divider-clock";
+	clocks = <&sys_clkin_ck>;
+	reg = <0x44df4250 0x4>;
+	ti,dividers = <8>, <16>, <32>;
+};
+
+div_core_25m_ck: div_core_25m_ck {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sysclk_div>;
+	clock-mult = <1>;
+	clock-div = <8>;
+};
+
+func_12m_clk: func_12m_clk {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&dpll_per_m2_ck>;
+	clock-mult = <1>;
+	clock-div = <16>;
+};
+
+vtp_clk_div: vtp_clk_div {
+	#clock-cells = <0>;
+	compatible = "fixed-factor-clock";
+	clocks = <&sys_clkin_ck>;
+	clock-mult = <1>;
+	clock-div = <2>;
+};
+
+usbphy_32khz_clkmux: usbphy_32khz_clkmux at 44df4260 {
+	#clock-cells = <0>;
+	compatible = "ti,mux-clock";
+	clocks = <&clk_32768_ck>, <&clk_32k_tpm_ck>;
+	reg = <0x44df4260 0x4>;
+};
Signature · 1 line
1.7.9.5
#38 lore raw

PATCH v7 35/36 ARM: dts: AM43xx: link in clock DT data

In reply to Tero Kristo

Clock tree data is now included from base am4372.dtsi file.

---
 arch/arm/boot/dts/am4372.dtsi |    7 +++++++
 1 file changed, 7 insertions(+)
arch/arm/boot/dts/am4372.dtsi +7 -0
index 0fe393a..218c1bd 100644
@@ -412,4 +412,11 @@
 			status = "disabled";
 		};
 	};
+
+	clocks {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		ranges;
+		/include/ "am43xx-clocks.dtsi"
+	};
 };
Signature · 1 line
1.7.9.5
#39 lore raw

PATCH v7 36/36 CLK: TI: add am43xx clock init file

In reply to Tero Kristo

clk-43xx.c now contains the clock init functionality for am43xx, including DT clock registration and adding of static clkdev entries.

---
 drivers/clk/ti/Makefile   |    2 +-
 drivers/clk/ti/clk-43xx.c |  120 +++++++++++++++++++++++++++++++++++++++++++++
 include/linux/clk/ti.h    |    1 +
 3 files changed, 122 insertions(+), 1 deletion(-)
 create mode 100644 drivers/clk/ti/clk-43xx.c
drivers/clk/ti/Makefile +1 -1
index df226bc..1137c1d 100644
@@ -1,7 +1,7 @@
 ifneq ($(CONFIG_OF),)
 obj-y					+= clk.o dpll.o autoidle.o divider.o \
 					   fixed-factor.o gate.o clockdomain.o \
-					   mux-gate.o mux.o apll.o
+					   mux-gate.o mux.o apll.o clk-43xx.o
 obj-$(CONFIG_SOC_AM33XX)		+= clk-33xx.o
 obj-$(CONFIG_ARCH_OMAP3)		+= interface.o clk-3xxx.o
 obj-$(CONFIG_ARCH_OMAP4)		+= clk-44xx.o
drivers/clk/ti/clk-43xx.c new file +120 -0

120 lines not on this page · raw message

include/linux/clk/ti.h +1 -0
index bf8beff..21ceed6 100644
@@ -229,6 +229,7 @@ int omap4xxx_clk_init(void);
 int omap5xxx_clk_init(void);
 int dra7xx_clk_init(void);
 int am33xx_clk_init(void);
+int am43xx_clk_init(void);
 
 #ifdef CONFIG_OF
 void of_ti_clk_allow_autoidle_all(void);
Signature · 1 line
1.7.9.5
#40 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

In reply to Tero Kristo

On 09/25/2013 03:48 AM, Tero Kristo wrote:
[...]

Quoted 11 lines

Test branch available here:
https://github.com/t-kristo/linux-pm.git
branch: mainline-3.12-rc2-ti-dt-clks-v7

Testing done:
- am335x-bone : boot only
- omap5-sevm : boot only
- dra7-evm : boot only
- omap3-beagle : boot + suspend/resume (ret + off)
- omap4-panda-es : boot + suspend/resume

Quoted 3 lines

Nishanth has also been doing some tests with omap3-beagle-xm with some WIP versions of this branch, but not with this latest one. Nishanth, maybe you can provide test results to this one also?

I rebased the branch on top of benoit's for_3.13/dts merged with v3.12-rc2 tag
45646cd ARM: dts: AM33XX: don't redefine OCP bus and device nodes

All platforms were to MMC filesystem boot (SD card). u-boot used v2013.10-rc3 on all platforms (mainline u-boot).

OMAP3630: Beagle-XM:
http://pastebin.pandaboard.org/index.php/view/96577127
(needs additional patch
-> https://patchwork.kernel.org/patch/2919661/ )

OMAP4460: Panda-ES: http://pastebin.pandaboard.org/index.php/view/85143661

OMAP5432: OMAP5-UEVM:
http://pastebin.pandaboard.org/index.php/view/91361843
(nice to have patch
-> https://patchwork.kernel.org/patch/2907761/ )

AM335x: BeagleBone Black:
http://pastebin.pandaboard.org/index.php/view/42000597
(needs additional patches:
https://patchwork.kernel.org/patch/2902711/
https://patchwork.kernel.org/patch/2902711/
)

DRA7: DRA7-EVM: http://pastebin.pandaboard.org/index.php/view/38362805
(needs additional patches:
-> https://patchwork.kernel.org/patch/2849391/
-> https://patchwork.kernel.org/patch/2849601/
-> https://patchwork.kernel.org/patch/2849602/
-> https://patchwork.kernel.org/patch/2907761/
)

So, for the entire series:

Signature · 2 lines
Regards,
Nishanth Menon
#41 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 20 lines · Nishanth Menon

On 09/25/2013 03:48 AM, Tero Kristo wrote:
[...]

Quoted 11 lines

Test branch available here:
https://github.com/t-kristo/linux-pm.git
branch: mainline-3.12-rc2-ti-dt-clks-v7

Testing done:
- am335x-bone : boot only
- omap5-sevm : boot only
- dra7-evm : boot only
- omap3-beagle : boot + suspend/resume (ret + off)
- omap4-panda-es : boot + suspend/resume

Quoted 3 lines

Nishanth has also been doing some tests with omap3-beagle-xm with some WIP versions of this branch, but not with this latest one. Nishanth, maybe you can provide test results to this one also?

I rebased the branch on top of benoit's for_3.13/dts merged with v3.12-rc2 tag
45646cd ARM: dts: AM33XX: don't redefine OCP bus and device nodes

Quoted 7 lines

All platforms were to MMC filesystem boot (SD card). u-boot used v2013.10-rc3 on all platforms (mainline u-boot).

OMAP3630: Beagle-XM:
http://pastebin.pandaboard.org/index.php/view/96577127
(needs additional patch
-> https://patchwork.kernel.org/patch/2919661/ )

dependency patch needs rework

Quoted 7 lines

OMAP4460: Panda-ES: http://pastebin.pandaboard.org/index.php/view/85143661

OMAP5432: OMAP5-UEVM:
http://pastebin.pandaboard.org/index.php/view/91361843
(nice to have patch
-> https://patchwork.kernel.org/patch/2907761/ )

Acked,tested, pending merge from maintainers

Quoted 5 lines

AM335x: BeagleBone Black:
http://pastebin.pandaboard.org/index.php/view/42000597
(needs additional patches:
https://patchwork.kernel.org/patch/2902711/

-> merged to Shekar's davinci branch -> pending upstream.

Quoted 1 line

https://patchwork.kernel.org/patch/2902711/

Correction on ^^: should have been:
https://patchwork.kernel.org/patch/2919661/
tested, pending ack, merge.

Quoted 8 lines

)

DRA7: DRA7-EVM: http://pastebin.pandaboard.org/index.php/view/38362805
(needs additional patches:
-> https://patchwork.kernel.org/patch/2849391/
-> https://patchwork.kernel.org/patch/2849601/
-> https://patchwork.kernel.org/patch/2849602/
-> https://patchwork.kernel.org/patch/2907761/

All of the above:
Acked,tested, pending merge from maintainers

Quoted 5 lines

)

So, for the entire series:

Maybe Benoit and Tony could help reduce few of these acked patches pending merge from having to be cherry-picked by folks.

Signature · 2 lines
Regards,
Nishanth Menon
#42 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

In reply to Tero Kristo

Hi,

Just a gentle reminder, anybody have any comments on this series or should we start queuing stuff?

-Tero

Quoted 59 lines · Tero Kristo

Hi all,

Version 7 contains following high level changes:
- Dropped support for basic bindings from Mike Turquette, instead using vendor specific bindings for all clocks
- Mux clock + divider clock vendor specific bindings get rid of use of the bit-masks, instead these are generated runtime based on parent clock / divider min/max info, see individual patches for details
- Added support for dummy_ck nodes, was missing from previous version
- Fixed support for omap3630
- Added support for new clock node type: ti,mux-gate-clock (using composite clock type)
- OMAP4 / OMAP5 only build fixes (compiler flag checks added to some files)
- most of the of_omap_* calls renamed to of_ti_*
- Rebased on top of v3.12-rc2

Some detailed comments about patch level changes (if no mention, no major changes):
- Patch #1:
* removed use of reg-names property, instead registers mapped using compatible string
* removed ti,modes property, instead uses DPLL mode flags now
* separated AM33XX DPLLs under their own compatible strings
- Patch #4 & #5: new patches
- Patch #6: merged basic gate support from Mike to this patch as vendor specific gate clock support
- Patch #9 & #10: new patches
- Patch #11: dummy clocks added, USB / ABE DPLL setup ordering changed
- Patch #14: dummy clocks added
- Patch #20: dropped usage of reg-names property
- Patch #21: dummy clocks added
- Patch #22 & #23: new patches
- Patch #30: AM35XX clock data added to this patch
- Patch #31:
* dummy clocks added
* added missing 3630 clocks

Test branch available here:
https://github.com/t-kristo/linux-pm.git
branch: mainline-3.12-rc2-ti-dt-clks-v7

Testing done:
- am335x-bone : boot only
- omap5-sevm : boot only
- dra7-evm : boot only
- omap3-beagle : boot + suspend/resume (ret + off)
- omap4-panda-es : boot + suspend/resume

Nishanth has also been doing some tests with omap3-beagle-xm with some WIP versions of this branch, but not with this latest one. Nishanth, maybe you can provide test results to this one also?

-Tero

--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo at vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

maintainer
#43 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 4 lines · Tero Kristo

Hi,

Just a gentle reminder, anybody have any comments on this series or should we start queuing stuff?

Well omap4 seems to work for me just fine, and omap3 in legacy mode. But looks like omap3 device tree based booting is breaking after I pulled in your test branch. I get the following on 3630 based dm3730.

Regards,

Tony

[    0.187652] omap_hwmod: uart4: cannot clk_get main_clk uart4_fck
[    0.187713] omap_hwmod: uart4: cannot _init_clocks
[    0.187713] ------------[ cut here ]------------
[    0.187774] WARNING: CPU: 0 PID: 1 at arch/arm/mach-omap2/omap_hwmod.c:2434 _init+0x6c/0x80()
[    0.187774] omap_hwmod: uart4: couldn't init clocks
[    0.187774] Modules linked in:
[    0.187805] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.12.0-rc2-00020-gdea1778-dirty #791
[    0.187866] [<c001cd54>] (unwind_backtrace+0x0/0xf4) from [<c0018e80>] (show_stack+0x14/0x1c)
[    0.187896] [<c0018e80>] (show_stack+0x14/0x1c) from [<c0554400>] (dump_stack+0x6c/0x9c)
[    0.187927] [<c0554400>] (dump_stack+0x6c/0x9c) from [<c00488a8>] (warn_slowpath_common+0x64/0x84)
[    0.187957] [<c00488a8>] (warn_slowpath_common+0x64/0x84) from [<c004895c>] (warn_slowpath_fmt+0x30/0x40)
[    0.187988] [<c004895c>] (warn_slowpath_fmt+0x30/0x40) from [<c07a4bec>] (_init+0x6c/0x80)
[    0.188018] [<c07a4bec>] (_init+0x6c/0x80) from [<c002e068>] (omap_hwmod_for_each+0x50/0x64)
[    0.188049] [<c002e068>] (omap_hwmod_for_each+0x50/0x64) from [<c07a5440>] (__omap_hwmod_setup_all+0x34/0x4c)
[    0.188079] [<c07a5440>] (__omap_hwmod_setup_all+0x34/0x4c) from [<c00087ac>] (do_one_initcall+0x2c/0x150)
[    0.188110] [<c00087ac>] (do_one_initcall+0x2c/0x150) from [<c0797508>] (do_basic_setup+0x94/0xd0)
[    0.188140] [<c0797508>] (do_basic_setup+0x94/0xd0) from [<c07975c0>] (kernel_init_freeable+0x7c/0x120)
[    0.188171] [<c07975c0>] (kernel_init_freeable+0x7c/0x120) from [<c05526b4>] (kernel_init+0xc/0x164)
[    0.188201] [<c05526b4>] (kernel_init+0xc/0x164) from [<c00149c8>] (ret_from_fork+0x14/0x2c)
[    0.188415] ---[ end trace 1b75b31a2719ed1c ]---
[    0.810241] ------------[ cut here ]------------
[    0.810302] WARNING: CPU: 0 PID: 1 at arch/arm/mach-omap2/omap_hwmod.c:2126 _enable+0x254/0x280()
[    0.810302] omap_hwmod: timer12: enabled state can only be entered from initialized, idle, or disabled state
[    0.810333] Modules linked in:
[    0.810363] CPU: 0 PID: 1 Comm: swapper/0 Tainted: G        W    3.12.0-rc2-00020-gdea1778-dirty #791
[    0.810394] [<c001cd54>] (unwind_backtrace+0x0/0xf4) from [<c0018e80>] (show_stack+0x14/0x1c)
[    0.810424] [<c0018e80>] (show_stack+0x14/0x1c) from [<c0554400>] (dump_stack+0x6c/0x9c)
[    0.810455] [<c0554400>] (dump_stack+0x6c/0x9c) from [<c00488a8>] (warn_slowpath_common+0x64/0x84)
[    0.810485] [<c00488a8>] (warn_slowpath_common+0x64/0x84) from [<c004895c>] (warn_slowpath_fmt+0x30/0x40)
[    0.810516] [<c004895c>] (warn_slowpath_fmt+0x30/0x40) from [<c003085c>] (_enable+0x254/0x280)
[    0.810546] [<c003085c>] (_enable+0x254/0x280) from [<c00308b0>] (omap_hwmod_enable+0x28/0x40)
[    0.810577] [<c00308b0>] (omap_hwmod_enable+0x28/0x40) from [<c0030f8c>] (omap_device_enable+0x3c/0x80)
[    0.810607] [<c0030f8c>] (omap_device_enable+0x3c/0x80) from [<c0030fe0>] (_od_runtime_resume+0x10/0x1c)
[    0.810638] [<c0030fe0>] (_od_runtime_resume+0x10/0x1c) from [<c036e120>] (__rpm_callback+0x2c/0x80)
[    0.810668] [<c036e120>] (__rpm_callback+0x2c/0x80) from [<c036e198>] (rpm_callback+0x24/0x78)
[    0.810699] [<c036e198>] (rpm_callback+0x24/0x78) from [<c036f444>] (rpm_resume+0x3c0/0x60c)
[    0.810729] [<c036f444>] (rpm_resume+0x3c0/0x60c) from [<c036f970>] (__pm_runtime_resume+0x4c/0x68)
[    0.810760] [<c036f970>] (__pm_runtime_resume+0x4c/0x68) from [<c0044954>] (omap_dm_timer_probe+0x234/0x334)
[    0.810791] [<c0044954>] (omap_dm_timer_probe+0x234/0x334) from [<c03659ac>] (platform_drv_probe+0x1c/0x24)
[    0.810821] [<c03659ac>] (platform_drv_probe+0x1c/0x24) from [<c03645e4>] (really_probe+0x70/0x200)
[    0.810852] [<c03645e4>] (really_probe+0x70/0x200) from [<c03647a8>] (driver_probe_device+0x34/0x50)
[    0.810852] [<c03647a8>] (driver_probe_device+0x34/0x50) from [<c0364858>] (__driver_attach+0x94/0x98)
[    0.810882] [<c0364858>] (__driver_attach+0x94/0x98) from [<c0362cd4>] (bus_for_each_dev+0x74/0x98)
[    0.810913] [<c0362cd4>] (bus_for_each_dev+0x74/0x98) from [<c0363510>] (bus_add_driver+0xe8/0x278)
[    0.810943] [<c0363510>] (bus_add_driver+0xe8/0x278) from [<c0364e30>] (driver_register+0x5c/0xf8)
[    0.810974] [<c0364e30>] (driver_register+0x5c/0xf8) from [<c00087ac>] (do_one_initcall+0x2c/0x150)
[    0.811004] [<c00087ac>] (do_one_initcall+0x2c/0x150) from [<c0797508>] (do_basic_setup+0x94/0xd0)
[    0.811035] [<c0797508>] (do_basic_setup+0x94/0xd0) from [<c07975c0>] (kernel_init_freeable+0x7c/0x120)
[    0.811065] [<c07975c0>] (kernel_init_freeable+0x7c/0x120) from [<c05526b4>] (kernel_init+0xc/0x164)
[    0.811096] [<c05526b4>] (kernel_init+0xc/0x164) from [<c00149c8>] (ret_from_fork+0x14/0x2c)
[    0.811126] ---[ end trace 1b75b31a2719ed1d ]---
[    0.811889] ------------[ cut here ]------------
[    0.811920] WARNING: CPU: 0 PID: 46 at arch/arm/mach-omap2/omap_hwmod.c:2224 _idle+0xd4/0xf8()
[    0.811920] omap_hwmod: timer12: idle state can only be entered from enabled state
[    0.811950] Modules linked in:
[    0.811981] CPU: 0 PID: 46 Comm: kworker/0:1 Tainted: G        W    3.12.0-rc2-00020-gdea1778-dirty #791
[    0.812011] Workqueue: pm pm_runtime_work
[    0.812042] [<c001cd54>] (unwind_backtrace+0x0/0xf4) from [<c0018e80>] (show_stack+0x14/0x1c)
[    0.812072] [<c0018e80>] (show_stack+0x14/0x1c) from [<c0554400>] (dump_stack+0x6c/0x9c)
[    0.812103] [<c0554400>] (dump_stack+0x6c/0x9c) from [<c00488a8>] (warn_slowpath_common+0x64/0x84)
[    0.812133] [<c00488a8>] (warn_slowpath_common+0x64/0x84) from [<c004895c>] (warn_slowpath_fmt+0x30/0x40)
[    0.812164] [<c004895c>] (warn_slowpath_fmt+0x30/0x40) from [<c003099c>] (_idle+0xd4/0xf8)
[    0.812194] [<c003099c>] (_idle+0xd4/0xf8) from [<c0030c9c>] (omap_hwmod_idle+0x28/0x3c)
[    0.812225] [<c0030c9c>] (omap_hwmod_idle+0x28/0x3c) from [<c0030e6c>] (omap_device_idle+0x3c/0x80)
[    0.812225] [<c0030e6c>] (omap_device_idle+0x3c/0x80) from [<c0030ed0>] (_od_runtime_suspend+0x20/0x2c)
[    0.812255] [<c0030ed0>] (_od_runtime_suspend+0x20/0x2c) from [<c036e120>] (__rpm_callback+0x2c/0x80)
[    0.812286] [<c036e120>] (__rpm_callback+0x2c/0x80) from [<c036e198>] (rpm_callback+0x24/0x78)
[    0.812316] [<c036e198>] (rpm_callback+0x24/0x78) from [<c036e5bc>] (rpm_suspend+0xf4/0x5e4)
[    0.812347] [<c036e5bc>] (rpm_suspend+0xf4/0x5e4) from [<c036fb78>] (pm_runtime_work+0x80/0xa4)
[    0.812377] [<c036fb78>] (pm_runtime_work+0x80/0xa4) from [<c0064058>] (process_one_work+0x204/0x594)
[    0.812408] [<c0064058>] (process_one_work+0x204/0x594) from [<c0064530>] (worker_thread+0x148/0x480)
[    0.812438] [<c0064530>] (worker_thread+0x148/0x480) from [<c0069c20>] (kthread+0xa0/0xac)
[    0.812469] [<c0069c20>] (kthread+0xa0/0xac) from [<c00149c8>] (ret_from_fork+0x14/0x2c)
[    0.812499] ---[ end trace 1b75b31a2719ed1e ]---
[    2.595031] ------------[ cut here ]------------
[    2.599945] WARNING: CPU: 0 PID: 1 at arch/arm/mach-omap2/omap_hwmod.c:2126 _enable+0x254/0x280()
[    2.609313] omap_hwmod: uart4: enabled state can only be entered from initialized, idle, or disabled state
[    2.619506] Modules linked in:
[    2.622772] CPU: 0 PID: 1 Comm: swapper/0 Tainted: G        W    3.12.0-rc2-00020-gdea1778-dirty #791
[    2.632537] [<c001cd54>] (unwind_backtrace+0x0/0xf4) from [<c0018e80>] (show_stack+0x14/0x1c)
[    2.641540] [<c0018e80>] (show_stack+0x14/0x1c) from [<c0554400>] (dump_stack+0x6c/0x9c)
[    2.650115] [<c0554400>] (dump_stack+0x6c/0x9c) from [<c00488a8>] (warn_slowpath_common+0x64/0x84)
[    2.659576] [<c00488a8>] (warn_slowpath_common+0x64/0x84) from [<c004895c>] (warn_slowpath_fmt+0x30/0x40)
[    2.669708] [<c004895c>] (warn_slowpath_fmt+0x30/0x40) from [<c003085c>] (_enable+0x254/0x280)
[    2.678802] [<c003085c>] (_enable+0x254/0x280) from [<c00308b0>] (omap_hwmod_enable+0x28/0x40)
[    2.687927] [<c00308b0>] (omap_hwmod_enable+0x28/0x40) from [<c0030f8c>] (omap_device_enable+0x3c/0x80)
[    2.697875] [<c0030f8c>] (omap_device_enable+0x3c/0x80) from [<c0030fe0>] (_od_runtime_resume+0x10/0x1c)
[    2.707916] [<c0030fe0>] (_od_runtime_resume+0x10/0x1c) from [<c036e120>] (__rpm_callback+0x2c/0x80)
[    2.717559] [<c036e120>] (__rpm_callback+0x2c/0x80) from [<c036e198>] (rpm_callback+0x24/0x78)
[    2.726684] [<c036e198>] (rpm_callback+0x24/0x78) from [<c036f444>] (rpm_resume+0x3c0/0x60c)
[    2.735595] [<c036f444>] (rpm_resume+0x3c0/0x60c) from [<c036f970>] (__pm_runtime_resume+0x4c/0x68)
[    2.745178] [<c036f970>] (__pm_runtime_resume+0x4c/0x68) from [<c035ae30>] (serial_omap_probe+0x294/0x4fc)
[    2.755401] [<c035ae30>] (serial_omap_probe+0x294/0x4fc) from [<c03659ac>] (platform_drv_probe+0x1c/0x24)
[    2.765502] [<c03659ac>] (platform_drv_probe+0x1c/0x24) from [<c03645e4>] (really_probe+0x70/0x200)
[    2.775054] [<c03645e4>] (really_probe+0x70/0x200) from [<c03647a8>] (driver_probe_device+0x34/0x50)
[    2.784729] [<c03647a8>] (driver_probe_device+0x34/0x50) from [<c0364858>] (__driver_attach+0x94/0x98)
[    2.794555] [<c0364858>] (__driver_attach+0x94/0x98) from [<c0362cd4>] (bus_for_each_dev+0x74/0x98)
[    2.804138] [<c0362cd4>] (bus_for_each_dev+0x74/0x98) from [<c0363510>] (bus_add_driver+0xe8/0x278)
[    2.813690] [<c0363510>] (bus_add_driver+0xe8/0x278) from [<c0364e30>] (driver_register+0x5c/0xf8)
[    2.823181] [<c0364e30>] (driver_register+0x5c/0xf8) from [<c07d0028>] (serial_omap_init+0x28/0x44)
[    2.832763] [<c07d0028>] (serial_omap_init+0x28/0x44) from [<c00087ac>] (do_one_initcall+0x2c/0x150)
[    2.842407] [<c00087ac>] (do_one_initcall+0x2c/0x150) from [<c0797508>] (do_basic_setup+0x94/0xd0)
[    2.851898] [<c0797508>] (do_basic_setup+0x94/0xd0) from [<c07975c0>] (kernel_init_freeable+0x7c/0x120)
[    2.861846] [<c07975c0>] (kernel_init_freeable+0x7c/0x120) from [<c05526b4>] (kernel_init+0xc/0x164)
[    2.871490] [<c05526b4>] (kernel_init+0xc/0x164) from [<c00149c8>] (ret_from_fork+0x14/0x2c)
[    2.880432] ---[ end trace 1b75b31a2719ed1f ]---
#44 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 10 lines · Tony Lindgren
Quoted 4 lines · Tero Kristo

Hi,

Just a gentle reminder, anybody have any comments on this series or should we start queuing stuff?

Well omap4 seems to work for me just fine, and omap3 in legacy mode. But looks like omap3 device tree based booting is breaking after I pulled in your test branch. I get the following on 3630 based dm3730.

Uart4 breaks because of the issue with basic DT blobs. You need at least this patch:

http://linux-kernel.2935.n7.nabble.com/RFC-PATCH-ARM-OMAP3630-Add-generic-machine-descriptor-td722791.html

... from Nishanth to get it to work right. Otherwise the clock init tries to use omap3430 clock data for omap3630 which leaves uart4 nodes out.

-Tero

Quoted 117 lines

Regards,

Tony

[    0.187652] omap_hwmod: uart4: cannot clk_get main_clk uart4_fck
[    0.187713] omap_hwmod: uart4: cannot _init_clocks
[    0.187713] ------------[ cut here ]------------
[    0.187774] WARNING: CPU: 0 PID: 1 at arch/arm/mach-omap2/omap_hwmod.c:2434 _init+0x6c/0x80()
[    0.187774] omap_hwmod: uart4: couldn't init clocks
[    0.187774] Modules linked in:
[    0.187805] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.12.0-rc2-00020-gdea1778-dirty #791
[    0.187866] [<c001cd54>] (unwind_backtrace+0x0/0xf4) from [<c0018e80>] (show_stack+0x14/0x1c)
[    0.187896] [<c0018e80>] (show_stack+0x14/0x1c) from [<c0554400>] (dump_stack+0x6c/0x9c)
[    0.187927] [<c0554400>] (dump_stack+0x6c/0x9c) from [<c00488a8>] (warn_slowpath_common+0x64/0x84)
[    0.187957] [<c00488a8>] (warn_slowpath_common+0x64/0x84) from [<c004895c>] (warn_slowpath_fmt+0x30/0x40)
[    0.187988] [<c004895c>] (warn_slowpath_fmt+0x30/0x40) from [<c07a4bec>] (_init+0x6c/0x80)
[    0.188018] [<c07a4bec>] (_init+0x6c/0x80) from [<c002e068>] (omap_hwmod_for_each+0x50/0x64)
[    0.188049] [<c002e068>] (omap_hwmod_for_each+0x50/0x64) from [<c07a5440>] (__omap_hwmod_setup_all+0x34/0x4c)
[    0.188079] [<c07a5440>] (__omap_hwmod_setup_all+0x34/0x4c) from [<c00087ac>] (do_one_initcall+0x2c/0x150)
[    0.188110] [<c00087ac>] (do_one_initcall+0x2c/0x150) from [<c0797508>] (do_basic_setup+0x94/0xd0)
[    0.188140] [<c0797508>] (do_basic_setup+0x94/0xd0) from [<c07975c0>] (kernel_init_freeable+0x7c/0x120)
[    0.188171] [<c07975c0>] (kernel_init_freeable+0x7c/0x120) from [<c05526b4>] (kernel_init+0xc/0x164)
[    0.188201] [<c05526b4>] (kernel_init+0xc/0x164) from [<c00149c8>] (ret_from_fork+0x14/0x2c)
[    0.188415] ---[ end trace 1b75b31a2719ed1c ]---
[    0.810241] ------------[ cut here ]------------
[    0.810302] WARNING: CPU: 0 PID: 1 at arch/arm/mach-omap2/omap_hwmod.c:2126 _enable+0x254/0x280()
[    0.810302] omap_hwmod: timer12: enabled state can only be entered from initialized, idle, or disabled state
[    0.810333] Modules linked in:
[    0.810363] CPU: 0 PID: 1 Comm: swapper/0 Tainted: G        W    3.12.0-rc2-00020-gdea1778-dirty #791
[    0.810394] [<c001cd54>] (unwind_backtrace+0x0/0xf4) from [<c0018e80>] (show_stack+0x14/0x1c)
[    0.810424] [<c0018e80>] (show_stack+0x14/0x1c) from [<c0554400>] (dump_stack+0x6c/0x9c)
[    0.810455] [<c0554400>] (dump_stack+0x6c/0x9c) from [<c00488a8>] (warn_slowpath_common+0x64/0x84)
[    0.810485] [<c00488a8>] (warn_slowpath_common+0x64/0x84) from [<c004895c>] (warn_slowpath_fmt+0x30/0x40)
[    0.810516] [<c004895c>] (warn_slowpath_fmt+0x30/0x40) from [<c003085c>] (_enable+0x254/0x280)
[    0.810546] [<c003085c>] (_enable+0x254/0x280) from [<c00308b0>] (omap_hwmod_enable+0x28/0x40)
[    0.810577] [<c00308b0>] (omap_hwmod_enable+0x28/0x40) from [<c0030f8c>] (omap_device_enable+0x3c/0x80)
[    0.810607] [<c0030f8c>] (omap_device_enable+0x3c/0x80) from [<c0030fe0>] (_od_runtime_resume+0x10/0x1c)
[    0.810638] [<c0030fe0>] (_od_runtime_resume+0x10/0x1c) from [<c036e120>] (__rpm_callback+0x2c/0x80)
[    0.810668] [<c036e120>] (__rpm_callback+0x2c/0x80) from [<c036e198>] (rpm_callback+0x24/0x78)
[    0.810699] [<c036e198>] (rpm_callback+0x24/0x78) from [<c036f444>] (rpm_resume+0x3c0/0x60c)
[    0.810729] [<c036f444>] (rpm_resume+0x3c0/0x60c) from [<c036f970>] (__pm_runtime_resume+0x4c/0x68)
[    0.810760] [<c036f970>] (__pm_runtime_resume+0x4c/0x68) from [<c0044954>] (omap_dm_timer_probe+0x234/0x334)
[    0.810791] [<c0044954>] (omap_dm_timer_probe+0x234/0x334) from [<c03659ac>] (platform_drv_probe+0x1c/0x24)
[    0.810821] [<c03659ac>] (platform_drv_probe+0x1c/0x24) from [<c03645e4>] (really_probe+0x70/0x200)
[    0.810852] [<c03645e4>] (really_probe+0x70/0x200) from [<c03647a8>] (driver_probe_device+0x34/0x50)
[    0.810852] [<c03647a8>] (driver_probe_device+0x34/0x50) from [<c0364858>] (__driver_attach+0x94/0x98)
[    0.810882] [<c0364858>] (__driver_attach+0x94/0x98) from [<c0362cd4>] (bus_for_each_dev+0x74/0x98)
[    0.810913] [<c0362cd4>] (bus_for_each_dev+0x74/0x98) from [<c0363510>] (bus_add_driver+0xe8/0x278)
[    0.810943] [<c0363510>] (bus_add_driver+0xe8/0x278) from [<c0364e30>] (driver_register+0x5c/0xf8)
[    0.810974] [<c0364e30>] (driver_register+0x5c/0xf8) from [<c00087ac>] (do_one_initcall+0x2c/0x150)
[    0.811004] [<c00087ac>] (do_one_initcall+0x2c/0x150) from [<c0797508>] (do_basic_setup+0x94/0xd0)
[    0.811035] [<c0797508>] (do_basic_setup+0x94/0xd0) from [<c07975c0>] (kernel_init_freeable+0x7c/0x120)
[    0.811065] [<c07975c0>] (kernel_init_freeable+0x7c/0x120) from [<c05526b4>] (kernel_init+0xc/0x164)
[    0.811096] [<c05526b4>] (kernel_init+0xc/0x164) from [<c00149c8>] (ret_from_fork+0x14/0x2c)
[    0.811126] ---[ end trace 1b75b31a2719ed1d ]---
[    0.811889] ------------[ cut here ]------------
[    0.811920] WARNING: CPU: 0 PID: 46 at arch/arm/mach-omap2/omap_hwmod.c:2224 _idle+0xd4/0xf8()
[    0.811920] omap_hwmod: timer12: idle state can only be entered from enabled state
[    0.811950] Modules linked in:
[    0.811981] CPU: 0 PID: 46 Comm: kworker/0:1 Tainted: G        W    3.12.0-rc2-00020-gdea1778-dirty #791
[    0.812011] Workqueue: pm pm_runtime_work
[    0.812042] [<c001cd54>] (unwind_backtrace+0x0/0xf4) from [<c0018e80>] (show_stack+0x14/0x1c)
[    0.812072] [<c0018e80>] (show_stack+0x14/0x1c) from [<c0554400>] (dump_stack+0x6c/0x9c)
[    0.812103] [<c0554400>] (dump_stack+0x6c/0x9c) from [<c00488a8>] (warn_slowpath_common+0x64/0x84)
[    0.812133] [<c00488a8>] (warn_slowpath_common+0x64/0x84) from [<c004895c>] (warn_slowpath_fmt+0x30/0x40)
[    0.812164] [<c004895c>] (warn_slowpath_fmt+0x30/0x40) from [<c003099c>] (_idle+0xd4/0xf8)
[    0.812194] [<c003099c>] (_idle+0xd4/0xf8) from [<c0030c9c>] (omap_hwmod_idle+0x28/0x3c)
[    0.812225] [<c0030c9c>] (omap_hwmod_idle+0x28/0x3c) from [<c0030e6c>] (omap_device_idle+0x3c/0x80)
[    0.812225] [<c0030e6c>] (omap_device_idle+0x3c/0x80) from [<c0030ed0>] (_od_runtime_suspend+0x20/0x2c)
[    0.812255] [<c0030ed0>] (_od_runtime_suspend+0x20/0x2c) from [<c036e120>] (__rpm_callback+0x2c/0x80)
[    0.812286] [<c036e120>] (__rpm_callback+0x2c/0x80) from [<c036e198>] (rpm_callback+0x24/0x78)
[    0.812316] [<c036e198>] (rpm_callback+0x24/0x78) from [<c036e5bc>] (rpm_suspend+0xf4/0x5e4)
[    0.812347] [<c036e5bc>] (rpm_suspend+0xf4/0x5e4) from [<c036fb78>] (pm_runtime_work+0x80/0xa4)
[    0.812377] [<c036fb78>] (pm_runtime_work+0x80/0xa4) from [<c0064058>] (process_one_work+0x204/0x594)
[    0.812408] [<c0064058>] (process_one_work+0x204/0x594) from [<c0064530>] (worker_thread+0x148/0x480)
[    0.812438] [<c0064530>] (worker_thread+0x148/0x480) from [<c0069c20>] (kthread+0xa0/0xac)
[    0.812469] [<c0069c20>] (kthread+0xa0/0xac) from [<c00149c8>] (ret_from_fork+0x14/0x2c)
[    0.812499] ---[ end trace 1b75b31a2719ed1e ]---
[    2.595031] ------------[ cut here ]------------
[    2.599945] WARNING: CPU: 0 PID: 1 at arch/arm/mach-omap2/omap_hwmod.c:2126 _enable+0x254/0x280()
[    2.609313] omap_hwmod: uart4: enabled state can only be entered from initialized, idle, or disabled state
[    2.619506] Modules linked in:
[    2.622772] CPU: 0 PID: 1 Comm: swapper/0 Tainted: G        W    3.12.0-rc2-00020-gdea1778-dirty #791
[    2.632537] [<c001cd54>] (unwind_backtrace+0x0/0xf4) from [<c0018e80>] (show_stack+0x14/0x1c)
[    2.641540] [<c0018e80>] (show_stack+0x14/0x1c) from [<c0554400>] (dump_stack+0x6c/0x9c)
[    2.650115] [<c0554400>] (dump_stack+0x6c/0x9c) from [<c00488a8>] (warn_slowpath_common+0x64/0x84)
[    2.659576] [<c00488a8>] (warn_slowpath_common+0x64/0x84) from [<c004895c>] (warn_slowpath_fmt+0x30/0x40)
[    2.669708] [<c004895c>] (warn_slowpath_fmt+0x30/0x40) from [<c003085c>] (_enable+0x254/0x280)
[    2.678802] [<c003085c>] (_enable+0x254/0x280) from [<c00308b0>] (omap_hwmod_enable+0x28/0x40)
[    2.687927] [<c00308b0>] (omap_hwmod_enable+0x28/0x40) from [<c0030f8c>] (omap_device_enable+0x3c/0x80)
[    2.697875] [<c0030f8c>] (omap_device_enable+0x3c/0x80) from [<c0030fe0>] (_od_runtime_resume+0x10/0x1c)
[    2.707916] [<c0030fe0>] (_od_runtime_resume+0x10/0x1c) from [<c036e120>] (__rpm_callback+0x2c/0x80)
[    2.717559] [<c036e120>] (__rpm_callback+0x2c/0x80) from [<c036e198>] (rpm_callback+0x24/0x78)
[    2.726684] [<c036e198>] (rpm_callback+0x24/0x78) from [<c036f444>] (rpm_resume+0x3c0/0x60c)
[    2.735595] [<c036f444>] (rpm_resume+0x3c0/0x60c) from [<c036f970>] (__pm_runtime_resume+0x4c/0x68)
[    2.745178] [<c036f970>] (__pm_runtime_resume+0x4c/0x68) from [<c035ae30>] (serial_omap_probe+0x294/0x4fc)
[    2.755401] [<c035ae30>] (serial_omap_probe+0x294/0x4fc) from [<c03659ac>] (platform_drv_probe+0x1c/0x24)
[    2.765502] [<c03659ac>] (platform_drv_probe+0x1c/0x24) from [<c03645e4>] (really_probe+0x70/0x200)
[    2.775054] [<c03645e4>] (really_probe+0x70/0x200) from [<c03647a8>] (driver_probe_device+0x34/0x50)
[    2.784729] [<c03647a8>] (driver_probe_device+0x34/0x50) from [<c0364858>] (__driver_attach+0x94/0x98)
[    2.794555] [<c0364858>] (__driver_attach+0x94/0x98) from [<c0362cd4>] (bus_for_each_dev+0x74/0x98)
[    2.804138] [<c0362cd4>] (bus_for_each_dev+0x74/0x98) from [<c0363510>] (bus_add_driver+0xe8/0x278)
[    2.813690] [<c0363510>] (bus_add_driver+0xe8/0x278) from [<c0364e30>] (driver_register+0x5c/0xf8)
[    2.823181] [<c0364e30>] (driver_register+0x5c/0xf8) from [<c07d0028>] (serial_omap_init+0x28/0x44)
[    2.832763] [<c07d0028>] (serial_omap_init+0x28/0x44) from [<c00087ac>] (do_one_initcall+0x2c/0x150)
[    2.842407] [<c00087ac>] (do_one_initcall+0x2c/0x150) from [<c0797508>] (do_basic_setup+0x94/0xd0)
[    2.851898] [<c0797508>] (do_basic_setup+0x94/0xd0) from [<c07975c0>] (kernel_init_freeable+0x7c/0x120)
[    2.861846] [<c07975c0>] (kernel_init_freeable+0x7c/0x120) from [<c05526b4>] (kernel_init+0xc/0x164)
[    2.871490] [<c05526b4>] (kernel_init+0xc/0x164) from [<c00149c8>] (ret_from_fork+0x14/0x2c)
[    2.880432] ---[ end trace 1b75b31a2719ed1f ]---
maintainer
#45 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 12 lines · Tero Kristo
Quoted 10 lines · Tony Lindgren
Quoted 4 lines · Tero Kristo

Hi,

Just a gentle reminder, anybody have any comments on this series or should we start queuing stuff?

Well omap4 seems to work for me just fine, and omap3 in legacy mode. But looks like omap3 device tree based booting is breaking after I pulled in your test branch. I get the following on 3630 based dm3730.

Quoted 8 lines

Uart4 breaks because of the issue with basic DT blobs. You need at least this patch:

http://linux-kernel.2935.n7.nabble.com/RFC-PATCH-ARM-OMAP3630-Add-generic-machine-descriptor-td722791.html

... from Nishanth to get it to work right. Otherwise the clock init tries to use omap3430 clock data for omap3630 which leaves uart4 nodes out.

OK thanks yes that solves it, I was missing the compatible flag for my not quite ready test boards (3730-evm and zoom3). I'll apply Nishant's fix with some changes.

Then for this series, once you have the necessary acks, can you please split into following two separate immutable branches against v3.12-rc3:

1. clock code changes

2. dts changes

Then those can be pulled in as needed by Benoit, Mike and I.

For the whole series, I think this quite nicely removes some nasty kernel data dependencies for omaps, so feel free to add:

Regards,

Tony

#46 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 18 lines · Tony Lindgren
Quoted 17 lines · Tero Kristo
Quoted 10 lines · Tony Lindgren
Quoted 4 lines · Tero Kristo

Hi,

Just a gentle reminder, anybody have any comments on this series or should we start queuing stuff?

Well omap4 seems to work for me just fine, and omap3 in legacy mode. But looks like omap3 device tree based booting is breaking after I pulled in your test branch. I get the following on 3630 based dm3730.

Uart4 breaks because of the issue with basic DT blobs. You need at least this patch:

http://linux-kernel.2935.n7.nabble.com/RFC-PATCH-ARM-OMAP3630-Add-generic-machine-descriptor-td722791.html

Quoted 8 lines
Quoted 3 lines

... from Nishanth to get it to work right. Otherwise the clock init tries to use omap3430 clock data for omap3630 which leaves uart4 nodes out.

OK thanks yes that solves it, I was missing the compatible flag for my not quite ready test boards (3730-evm and zoom3). I'll apply Nishant's fix with some changes.

I will post out an RFC based on Olof's suggestion in a few mins. will be nice to see any deltas needed.

Signature · 2 lines
Regards,
Nishanth Menon
maintainer
#47 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

In reply to Tony Lindgren
Quoted 30 lines · Tony Lindgren
Quoted 12 lines · Tero Kristo
Quoted 10 lines · Tony Lindgren
Quoted 4 lines · Tero Kristo

Hi,

Just a gentle reminder, anybody have any comments on this series or should we start queuing stuff?

Well omap4 seems to work for me just fine, and omap3 in legacy mode. But looks like omap3 device tree based booting is breaking after I pulled in your test branch. I get the following on 3630 based dm3730.

Quoted 8 lines

Uart4 breaks because of the issue with basic DT blobs. You need at least this patch:

http://linux-kernel.2935.n7.nabble.com/RFC-PATCH-ARM-OMAP3630-Add-generic-machine-descriptor-td722791.html

... from Nishanth to get it to work right. Otherwise the clock init tries to use omap3430 clock data for omap3630 which leaves uart4 nodes out.

OK thanks yes that solves it, I was missing the compatible flag for my not quite ready test boards (3730-evm and zoom3). I'll apply Nishant's fix with some changes.

Then for this series, once you have the necessary acks, can you please split into following two separate immutable branches against v3.12-rc3:

Quoted 5 lines

1. clock code changes

2. dts changes

Then those can be pulled in as needed by Benoit, Mike and I.

And Paul too natuarlly if these patches conflict with stuff that's he's merging. And assuming Paul is OK with these patches in general.

Quoted 12 lines

For the whole series, I think this quite nicely removes some nasty kernel data dependencies for omaps, so feel free to add:

Regards,

Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo at vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

maintainer
#48 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 1 line · Tony Lindgren

And assuming Paul is OK with these patches in general.

Actually, I have several concerns with this series, as you and I discussed. Some of us have been talking them over with other folks for the past few months to try to figure out what to do. Most of the concerns have fairly easy technical solutions, but we shouldn't merge these patches until they're resolved.

The issues are:

1. whether the clock data should go into DT
2. what the right place for DT clock data is in the DT
3. whether it makes sense to merge "experimental" DT bindings in this case
4. where clockdomain data belongs

The first issue - and the one I'm least concerned about - is that, in my view, it still does not make technical sense to move this data into DT. This is chip-level hardware data that never needs to be changed by board designers or end users. Adding it to DT is going to result in a boot-time cost (due to DT parse overhead) and additional memory consumption for very little benefit, compared to an implementation that places this data into a dynamically loadable module. For some users, the boot-time bloat is a big
deal: the example that's been mentioned to me recently is an automotive back-up camera that needs to cold-boot to complete functionality in a few hundred microseconds.

However, according to some other upstream maintainers, Linus's goal is to move most of the device-specific static data out of the kernel tree, into DT files (in the ARM case). If that non-technical constraint is indeed the dominant consideration, then I agree that moving this data to DT is the only viable course of action.

...

The second issue concerns where the SoC clock nodes should go in the DT. In these patches, the clock data has been encoded in a large "clocks" node off the top level. This is almost certainly not the right thing to do from a device hardware point of view. These clocks don't exist as standalone devices with their own address spaces decoded on the interconnect. In all of the SoC cases that I'm aware of, clock control registers are part of larger IP blocks.

For example, in the OMAP case, most of the system integration clock control registers are part of the OMAP-specific PRCM block, PRM block, or CM block. Then there are other device-specific clocks, like DSS PLLs or UART dividers. The control registers for these are generally located in the subsystem or device IP block itself, and are inaccessible when the subsystem or IP block is disabled. These device-specific clocks belong to the IP block, not the SoC integration. So, for example, if two SoCs use the same IP block, then the clock registers, and their offsets from the IP block address space base, are likely to be identical.

So in my view, the right things to do here are to:

1. associate SoC DT clock data with the device node that contains the clock control registers

2. specify relative offsets for clock registers from the start of the IP block address range, rather than absolute addresses for clock registers

3. place the responsibility for registering clocks into the IP block drivers themselves

This naturally separates clocks into per-IP block DT files. It also provides the CCF with an easy way to ensure that the device that encloses the clock is enabled and accessible by the CPU core, before trying to access registers inside.

Similarly, non-SoC off-chip clock data (such as for dedicated I2C PLLs) should also be associated with their I2C device node.

Making these changes to Tero's existing patches should be relatively straightforward, based on what I've seen.

...

Regarding the third issue: let's postulate for the moment that the clock binding issues that I mention in #2 above are ignored (ha!), and that the existing DT clock data is merged. These bindings are currently marked as "Unstable - ABI compatibility may be broken in the future". What happens if, when we meet to discuss clock bindings at the kernel summit, we decide that significant changes are needed? We could easily wind up with kernels that won't boot at all when used with newer DT data.

Not to mention, merging such a large amount of code and data before the bindings are stable will increase the risk of massive churn as the bindings evolve towards stability.

So IMHO, the way to fix this is to consider the clock data to be IP-block specific, as mentioned in #2. Then there's no need for global clock bindings for the SoC clock cases.

Otherwise, it seems prudent to at least wait until the global clock bindings are no longer marked as unstable. The whole DT migration was predicated on the ideas of reducing churn in the Linux codebase, and preserving forward compatibility for DT data. We shouldn't discard these goals just to merge something a little sooner.

...

The fourth issue is where the clockdomain data should go. The basic issue here is that the notion of "clockdomains" here is completely OMAP-specific. OMAP clockdomains demarcate a group of clocks or IP blocks that share the same automatic idle behavior, controlled by their enclosing IP block (e.g., PRCM, PRM, CM, etc.). Clockdomains are also used in the OMAP kernel code to link the clocks to their enclosing power domains, voltage rail control, etc.

So since these are OMAP PRCM/CM/PRM-specific constructs, the right place for OMAP clockdomain data is underneath the OMAP-specific CM/PRCM DT nodes. Again this should be an easy change to Tero's existing patches.

...

So hey, if you've made it this far, thanks for reading. My sense is that implementing #2 and #4 are relatively straightforward.

Also, since this seems to have been a problem for some folks in the past, I want to make clear that I think both Mike and Tero have been doing good jobs on the CCF and OMAP clock patches so far.

regards

- Paul

maintainer
#49 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Eh, one correction:

Quoted 2 lines · Paul Walmsley

We could easily wind up with kernels that won't boot at all when used with newer DT data.

This is a misstatement of the issue: the concern here is that newer kernels may not boot at all with older DT data - which could easily be in locked areas of the flash or firmware.

- Paul

maintainer
#50 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 10 lines · Paul Walmsley

Eh, one correction:

Quoted 2 lines · Paul Walmsley

We could easily wind up with kernels that won't boot at all when used with newer DT data.

This is a misstatement of the issue: the concern here is that newer kernels may not boot at all with older DT data - which could easily be in locked areas of the flash or firmware.

That's easy to avoid by allowing any combination of DT defined clocks, loadable module clocks and /lib/firmware clocks depending on the use case.

The older kernels have the clocks defined in DT, the newer kernels just need to parse them.

Regards,

Tony

#51 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

In reply to Paul Walmsley
Quoted 10 lines · Paul Walmsley

Eh, one correction:

Quoted 2 lines · Paul Walmsley

We could easily wind up with kernels that won't boot at all when used with newer DT data.

This is a misstatement of the issue: the concern here is that newer kernels may not boot at all with older DT data - which could easily be in locked areas of the flash or firmware.

I wonder who would be crazy enough to put DT data into a locked area, and to what purpose. If you can update the kernel, there is no point locking down DT data, this will just cause you unnecessary misery.

-Tero

maintainer
#52 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 8 lines · Tero Kristo
Quoted 7 lines · Paul Walmsley

Eh, one correction:

Quoted 2 lines · Paul Walmsley

We could easily wind up with kernels that won't boot at all when used with newer DT data.

Quoted 7 lines
Quoted 3 lines

This is a misstatement of the issue: the concern here is that newer kernels may not boot at all with older DT data - which could easily be in locked areas of the flash or firmware.

I wonder who would be crazy enough to put DT data into a locked area, and to what purpose. If you can update the kernel, there is no point locking down DT data, this will just cause you unnecessary misery.

The DT data will be used by bootloaders also :-(

In situations where the bootloaders are signed and locked, the security people are also insisting that the DT data be signed and locked.

- Paul

#53 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 14 lines · Paul Walmsley
Quoted 12 lines · Tero Kristo
Quoted 10 lines · Paul Walmsley

Eh, one correction:

Quoted 2 lines · Paul Walmsley

We could easily wind up with kernels that won't boot at all when used with newer DT data.

This is a misstatement of the issue: the concern here is that newer kernels may not boot at all with older DT data - which could easily be in locked areas of the flash or firmware.

Quoted 8 lines
Quoted 3 lines

I wonder who would be crazy enough to put DT data into a locked area, and to what purpose. If you can update the kernel, there is no point locking down DT data, this will just cause you unnecessary misery.

The DT data will be used by bootloaders also :-(

In situations where the bootloaders are signed and locked, the security people are also insisting that the DT data be signed and locked.

Well, even if you sign something, you can still update it. Writing any software to true OTP memory is one way to commit suicide IMO. How many nasty bugs have you seen with ROM code? Also, if people want to make their custom security solutions which are not supported by the kernel, why should the kernel care about it? We don't know the details, and can't influence the design, so we can't prepare for it anyway.

-Tero

maintainer
#54 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 6 lines · Tero Kristo

Well, even if you sign something, you can still update it. Writing any software to true OTP memory is one way to commit suicide IMO. How many nasty bugs have you seen with ROM code? Also, if people want to make their custom security solutions which are not supported by the kernel, why should the kernel care about it? We don't know the details, and can't influence the design, so we can't prepare for it anyway.

The point is that newer kernels need to be compatible with older DT data:
that's the "stable ABI" criterion that's been discussed quite a bit. The worst-case scenario if the bindings are at a top-level node and not fully cooked is that it can make it very difficult, if not impossible, to have this forward compatibility for the DT data without having extremely nasty hacks.

- Paul

#55 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

In reply to Paul Walmsley
Quoted 14 lines · Paul Walmsley
Quoted 12 lines · Tero Kristo
Quoted 10 lines · Paul Walmsley

Eh, one correction:

Quoted 2 lines · Paul Walmsley

We could easily wind up with kernels that won't boot at all when used with newer DT data.

This is a misstatement of the issue: the concern here is that newer kernels may not boot at all with older DT data - which could easily be in locked areas of the flash or firmware.

Quoted 8 lines
Quoted 3 lines

I wonder who would be crazy enough to put DT data into a locked area, and to what purpose. If you can update the kernel, there is no point locking down DT data, this will just cause you unnecessary misery.

The DT data will be used by bootloaders also :-(

In situations where the bootloaders are signed and locked, the security people are also insisting that the DT data be signed and locked.

Oh yea, one additional note you probably have missed. Mike asked us to fall back to vendor specific bindings at around v6 or so of this set. Take a look at v8, we have dropped the use of generic bindings, we are not trying to declare those with this set. This set introduces vendor specific bindings only, and even these are claimed 'unstable', which should be enough to discourage people from burning those to OTP type memory.

-Tero

maintainer
#56 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 4 lines · Tero Kristo

Oh yea, one additional note you probably have missed. Mike asked us to fall back to vendor specific bindings at around v6 or so of this set. Take a look at v8, we have dropped the use of generic bindings, we are not trying to declare those with this set.

No, I didn't miss it.

Quoted 3 lines

This set introduces vendor specific bindings only, and even these are claimed 'unstable', which should be enough to discourage people from burning those to OTP type memory.

Better to just avoid merging unstable bindings in the first place.

Please keep this in mind: any change that anyone makes to the DT data needs to be supportable by the kernel into the indefinite future, once it makes it into arch/arm/boot/dts or the DT binding documentation. Also, any changes need to work for other OSes, i.e., the changes should not be Linux-specific.

Those have been stated goals for the Linux DT project since day one. Some folks haven't been monitoring those goals very closely and that's unfortunate. We could have avoided some pretty big messes.

- Paul

#57 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 13 lines · Paul Walmsley
Quoted 4 lines · Tero Kristo

Oh yea, one additional note you probably have missed. Mike asked us to fall back to vendor specific bindings at around v6 or so of this set. Take a look at v8, we have dropped the use of generic bindings, we are not trying to declare those with this set.

No, I didn't miss it.

Quoted 3 lines

This set introduces vendor specific bindings only, and even these are claimed 'unstable', which should be enough to discourage people from burning those to OTP type memory.

Quoted 7 lines

Better to just avoid merging unstable bindings in the first place.

Please keep this in mind: any change that anyone makes to the DT data needs to be supportable by the kernel into the indefinite future, once it makes it into arch/arm/boot/dts or the DT binding documentation. Also, any changes need to work for other OSes, i.e., the changes should not be Linux-specific.

As pointed out earlier, the unstable claim just points to the fact that we might have some generic clock bindings in distant future at which point we _might_ evolve that way. Personally I have no problems carrying these binding to the far future, as they are vendor specific, and pretty much also device specific (am3-dpll-* omap4-dpll-* etc.)

Later on, we can add also support for clock firmware or whatever if we want to optimize stuff.

If we do not merge the bindings now, what is your proposal then? We have a couple of new SoC:s in queue and getting them boot mainline currently we only have this option to get them to work. Hardware is not going to wait for two years so we can have generic clock bindings in place. I would rather have this resolved as soon as possible, and it would have been much better if you came forward with your claims at version 1 or 2 of this set, rather than at v8 and saying all should be forfeit.

-Tero

Quoted 8 lines

Those have been stated goals for the Linux DT project since day one. Some folks haven't been monitoring those goals very closely and that's unfortunate. We could have avoided some pretty big messes.

- Paul

maintainer
#58 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Here you go:

1. Create DT nodes for the IP blocks that contain the clock control registers, underneath the appropriate bus node. So for example OMAP4 would have:

/ocp {

    prm at 4a306000 { 
	compatible = xxx;
	regs = <0x4a306000 xxxx>;
    };

    cm1 at 4a004000 { 
	compatible = xxx;
	regs = <0x4a004000 xxxx>;
    };

    cm2 at 4a008000 { 
	compatible = xxx;
	regs = <0x4a008000 xxxx>;
    };

};

This is pure OS-independent hardware description (with the unfortunate exception of the "ocp" part).

2. Create the clock data underneath the corresponding PRCM/CM/PRM device nodes that their control registers exist in. So for example, dpll_per_ck would belong to the "cm1" device, since that's where its control registers are located. Something like this:

    cm1 at 4a004000 {
        ...
        clocks { 
            dpll_per_ck at .... {
		<clock data goes here>
            };
        };
    };

The DT data is intended to portray addressable devices from the perspective of the CPUs that are booted with that data, organized by bus structure. From this point of view, clocks that are controlled via a particular IP block should have their data associated with that block.

3. Place clock register accesses in the low-level PRCM/CM/PRM device driver. If other kernel code needs some resource that's provided by the IP block, it should either come from common Linux framework routines (like the clock code), or use functions exported from the low-level driver that don't expose register addresses.

Otherwise, if MMIO accesses to that device are allowed from all over the codebase, it creates an undebuggable mess:

A. It could result in IP blocks being partially programmed before they are probed by their drivers. If their drivers (or the integration code) reset the IP blocks, the non-driver code has no way of knowing that it needs to reprogram the underlying IP block.

B. If any preparation is needed before the clock registers can be accessed, like runtime PM calls or device_enable()-type calls, this should be coordinated by the low-level IP block itself. We don't expect this to be the case for PRCM/CM/PRM, but we do expect it to be the case for UART, DSS, some audio clocks, etc.

C. The low-level IP block drivers may want to implement some caching layer like regmap. If some register writes bypass the low-level driver, then someone is likely to get a big unpleasant surprise.

D. It impairs the common debugging technique of adding code to the IP block MMIO read/write functions to log register accesses.

4. Use relative register offsets from the top of the containing IP block's base address, rather than absolute addresses.

   cm1 at 4a004000 {
        ...
        clocks {
            dpll_per_ck at 4140 {
                <clock data goes here>
            };
        };
    };

This makes it possible to reuse the same DT clock data in cases where the same IP block is used, but at a different base address. This is probably not a big issue with the system integration IP blocks, but quite possible with the non-SoC-integration IP blocks. It also makes it obvious if someone tries to sneak clocks into the IP block data that aren't controlled by that IP block.

5. Register the clocks from the low-level IP block drivers, rather than from external code. That way there's no need to export low-level register manipulation functions off to other kernel code. This registration can be done when the PRCM/CM/PRM driver probes.

6. Move the OMAP clockdomain data underneath the DT node for the low-level IP block that contains them:

   cm1 at 4a004000 {
        ...
        clocks {
	       ...
        };

        clockdomains {
               l3_init_clkdm: l3_init_clkdm at ... {
                   ... 
               };
        };
    };

For non-OMAP folks reading this thread, OMAP clockdomains have control registers associated with them, located in the PRCM/CM/PRM IP block address space, so that's where they belong.

7. drivers/clk/ti is probably the wrong place for most of the low-level drivers for IP blocks like the PRM/PRCM/CM. Most of these IP blocks do more than just control clocks: they also control other system entities like reset lines, OMAP powerdomains, AVS, or OMAP "clockdomains" (which in some cases may have nothing to do with clocks).

drivers/clk/ti is almost defensible for CM, if it weren't for the OMAP "clockdomain" registers that are there. However, for the other OMAP integration IP blocks, drivers/clk/ti definitely seems out of place. And since some OMAP chips like OMAP2420 or AM43XX don't have separate CM and PRM blocks -- they combine them both into a single PRCM module -- it seems best to place CM, PRM, and other related code into a single common directory, so they can easily share code across chips.

I'd suggest putting them in drivers/sysctrl/omap/. This would be intended for low-level SoC IP block drivers that control system integration. Otherwise drivers/clk/XXX is going to turn into another drivers/mfd.

8. A few random comments about the individual clock binding data formats themselves, based on a quick look:

A. It seems pretty odd and unnecessarily obscure to do something like this:

dpll_usb_ck: dpll_usb_ck at ... {
       ...
       reg = <0x4a008180 0x4>, <0x4a008184 0x4>, <0x4a008188 0x4>, <0x4a00818c 0x4>;
       ...
};

It's at least self-documenting to do something like this:

dpll_usb_ck: dpll_usb_ck at ... {     
       ...
       control-reg = < ... >;
       idlest-reg = < ... >;
       .. etc. ..
};

Which itself might not even be needed, depending on how the DPLL control code is implemented. For example, if the relative offsets are always the same for all OMAP4-family devices, maybe there's not even a need to explicitly encode that into the DT data.

B. Seems like you can remove the "ti," prefix on the properties, since they have no pretentions at genericity: they are specific to the PRCM/CM/PRM IP block data, and registered by those drivers.

C. Looks like the patches use the property "autoidle-low" to indicate that the autoidle bit should be inverted. "Low" seems like the wrong expression here - it invokes the actual voltage logic level of a hardware signal, and we have no idea whether the hardware signal is using a low voltage or a high voltage to express this condition. Would suggest something like 'invert-autoidle-bit' instead.

D. Regarding "ti,index-starts-at-one", it seems best to explicitly state which index starts at one. The code mentions a "mux index" so please consider renaming this something like "mux-index-starts-at-one" or "one-based-mux-index"

- Paul

#59 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Hi Paul,

I've been looking at this stuff today, and this will cause pretty huge changes to the set. This probably means a few weeks delay in implementation at least.

Some comments on the technical perspectives below.

Quoted 92 lines · Paul Walmsley

Here you go:

1. Create DT nodes for the IP blocks that contain the clock control registers, underneath the appropriate bus node. So for example OMAP4 would have:

/ocp {

     prm at 4a306000 {
	compatible = xxx;
	regs = <0x4a306000 xxxx>;
     };

     cm1 at 4a004000 {
	compatible = xxx;
	regs = <0x4a004000 xxxx>;
     };

     cm2 at 4a008000 {
	compatible = xxx;
	regs = <0x4a008000 xxxx>;
     };

};

This is pure OS-independent hardware description (with the unfortunate exception of the "ocp" part).

2. Create the clock data underneath the corresponding PRCM/CM/PRM device nodes that their control registers exist in. So for example, dpll_per_ck would belong to the "cm1" device, since that's where its control registers are located. Something like this:

     cm1 at 4a004000 {
         ...
         clocks {
             dpll_per_ck at .... {
		<clock data goes here>
             };
         };
     };

The DT data is intended to portray addressable devices from the perspective of the CPUs that are booted with that data, organized by bus structure. From this point of view, clocks that are controlled via a particular IP block should have their data associated with that block.

3. Place clock register accesses in the low-level PRCM/CM/PRM device driver. If other kernel code needs some resource that's provided by the IP block, it should either come from common Linux framework routines (like the clock code), or use functions exported from the low-level driver that don't expose register addresses.

Otherwise, if MMIO accesses to that device are allowed from all over the codebase, it creates an undebuggable mess:

A. It could result in IP blocks being partially programmed before they are probed by their drivers. If their drivers (or the integration code) reset the IP blocks, the non-driver code has no way of knowing that it needs to reprogram the underlying IP block.

B. If any preparation is needed before the clock registers can be accessed, like runtime PM calls or device_enable()-type calls, this should be coordinated by the low-level IP block itself. We don't expect this to be the case for PRCM/CM/PRM, but we do expect it to be the case for UART, DSS, some audio clocks, etc.

C. The low-level IP block drivers may want to implement some caching layer like regmap. If some register writes bypass the low-level driver, then someone is likely to get a big unpleasant surprise.

D. It impairs the common debugging technique of adding code to the IP block MMIO read/write functions to log register accesses.

4. Use relative register offsets from the top of the containing IP block's base address, rather than absolute addresses.

    cm1 at 4a004000 {
         ...
         clocks {
             dpll_per_ck at 4140 {
                 <clock data goes here>
             };
         };
     };
Quoted 6 lines

This makes it possible to reuse the same DT clock data in cases where the same IP block is used, but at a different base address. This is probably not a big issue with the system integration IP blocks, but quite possible with the non-SoC-integration IP blocks. It also makes it obvious if someone tries to sneak clocks into the IP block data that aren't controlled by that IP block.

Relative addressing basically means I need to copy paste the code for clk_divider + clk_mux from drivers/clk to drivers/clk/ti. Or alternatively add a way to provide register read/write ops to the basic clock.

Quoted 10 lines

5. Register the clocks from the low-level IP block drivers, rather than from external code. That way there's no need to export low-level register manipulation functions off to other kernel code. This registration can be done when the PRCM/CM/PRM driver probes.

6. Move the OMAP clockdomain data underneath the DT node for the low-level IP block that contains them:

Quoted 12 lines
    cm1 at 4a004000 {
         ...
         clocks {
	       ...
         };

         clockdomains {
                l3_init_clkdm: l3_init_clkdm at ... {
                    ...
                };
         };
     };

I think clocks + clockdomains grouping nodes are unnecessary in this case. Same can be accomplished with simply using the compatible string for mapping to corresponding types. (This is probably valid for current code also.)

Quoted 19 lines

For non-OMAP folks reading this thread, OMAP clockdomains have control registers associated with them, located in the PRCM/CM/PRM IP block address space, so that's where they belong.

7. drivers/clk/ti is probably the wrong place for most of the low-level drivers for IP blocks like the PRM/PRCM/CM. Most of these IP blocks do more than just control clocks: they also control other system entities like reset lines, OMAP powerdomains, AVS, or OMAP "clockdomains" (which in some cases may have nothing to do with clocks).

drivers/clk/ti is almost defensible for CM, if it weren't for the OMAP "clockdomain" registers that are there. However, for the other OMAP integration IP blocks, drivers/clk/ti definitely seems out of place. And since some OMAP chips like OMAP2420 or AM43XX don't have separate CM and PRM blocks -- they combine them both into a single PRCM module -- it seems best to place CM, PRM, and other related code into a single common directory, so they can easily share code across chips.

Quoted 3 lines

I'd suggest putting them in drivers/sysctrl/omap/. This would be intended for low-level SoC IP block drivers that control system integration. Otherwise drivers/clk/XXX is going to turn into another drivers/mfd.

I think initially I would place this stuff under mach-omap2. Much better than to try to introduce multiple new drivers with a single set.

Quoted 12 lines

8. A few random comments about the individual clock binding data formats themselves, based on a quick look:

A. It seems pretty odd and unnecessarily obscure to do something like this:

dpll_usb_ck: dpll_usb_ck at ... {
        ...
        reg = <0x4a008180 0x4>, <0x4a008184 0x4>, <0x4a008188 0x4>, <0x4a00818c 0x4>;
        ...
};
Quoted 8 lines

It's at least self-documenting to do something like this:

dpll_usb_ck: dpll_usb_ck at ... {
        ...
        control-reg = < ... >;
        idlest-reg = < ... >;
        .. etc. ..
};

Some earlier version had something along these lines, but it was turned down. I had also reg-names as documentation purposes along, but this was unnecessary and was dropped also.

Quoted 4 lines

Which itself might not even be needed, depending on how the DPLL control code is implemented. For example, if the relative offsets are always the same for all OMAP4-family devices, maybe there's not even a need to explicitly encode that into the DT data.

If I want to get rid of these, I need to add extra compatible strings for the dpll types. There are several weird register offsets for omap3/am3 devices. omap4/5/dra7/am4 behave more sanely.

Quoted 3 lines

B. Seems like you can remove the "ti," prefix on the properties, since they have no pretentions at genericity: they are specific to the PRCM/CM/PRM IP block data, and registered by those drivers.

Can or should? It seems existing bindings use "ti," prefix even on non-generic bindings, meaning if I look at any other data in DT.

Quoted 11 lines

C. Looks like the patches use the property "autoidle-low" to indicate that the autoidle bit should be inverted. "Low" seems like the wrong expression here - it invokes the actual voltage logic level of a hardware signal, and we have no idea whether the hardware signal is using a low voltage or a high voltage to express this condition. Would suggest something like 'invert-autoidle-bit' instead.

D. Regarding "ti,index-starts-at-one", it seems best to explicitly state which index starts at one. The code mentions a "mux index" so please consider renaming this something like "mux-index-starts-at-one" or "one-based-mux-index"

The index is explicitly defined by the clock node where this is present. If it is a mux-clock, then it is for mux-index. If for divider clock, it is index for the divider.

-Tero

maintainer
#60 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 37 lines · Tero Kristo
Quoted 13 lines · Paul Walmsley

8. A few random comments about the individual clock binding data formats themselves, based on a quick look:

A. It seems pretty odd and unnecessarily obscure to do something like this:

dpll_usb_ck: dpll_usb_ck at ... {
        ...
        reg = <0x4a008180 0x4>, <0x4a008184 0x4>, <0x4a008188 0x4>,
<0x4a00818c 0x4>;
        ...
};
Quoted 8 lines

It's at least self-documenting to do something like this:

dpll_usb_ck: dpll_usb_ck at ... {
        ...
        control-reg = < ... >;
        idlest-reg = < ... >;
        .. etc. ..
};

Some earlier version had something along these lines, but it was turned down. I had also reg-names as documentation purposes along, but this was unnecessary and was dropped also.

Quoted 4 lines

Which itself might not even be needed, depending on how the DPLL control code is implemented. For example, if the relative offsets are always the same for all OMAP4-family devices, maybe there's not even a need to explicitly encode that into the DT data.

If I want to get rid of these, I need to add extra compatible strings for the dpll types. There are several weird register offsets for omap3/am3 devices. omap4/5/dra7/am4 behave more sanely.

Quoted 6 lines
Quoted 3 lines

B. Seems like you can remove the "ti," prefix on the properties, since they have no pretentions at genericity: they are specific to the PRCM/CM/PRM IP block data, and registered by those drivers.

Can or should? It seems existing bindings use "ti," prefix even on non-generic bindings, meaning if I look at any other data in DT.

My comments in #8 are just regarding minor issues that don't seem right. I don't have a significant objection to staying with the existing property names here if you think that they are important.

Quoted 7 lines
Quoted 7 lines

C. Looks like the patches use the property "autoidle-low" to indicate that the autoidle bit should be inverted. "Low" seems like the wrong expression here - it invokes the actual voltage logic level of a hardware signal, and we have no idea whether the hardware signal is using a low voltage or a high voltage to express this condition. Would suggest something like 'invert-autoidle-bit' instead.

Quoted 8 lines
Quoted 4 lines

D. Regarding "ti,index-starts-at-one", it seems best to explicitly state which index starts at one. The code mentions a "mux index" so please consider renaming this something like "mux-index-starts-at-one" or "one-based-mux-index"

The index is explicitly defined by the clock node where this is present. If it is a mux-clock, then it is for mux-index. If for divider clock, it is index for the divider.

- Paul

#61 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 43 lines · Paul Walmsley
Quoted 41 lines · Tero Kristo
Quoted 13 lines · Paul Walmsley

8. A few random comments about the individual clock binding data formats themselves, based on a quick look:

A. It seems pretty odd and unnecessarily obscure to do something like this:

dpll_usb_ck: dpll_usb_ck at ... {
         ...
         reg = <0x4a008180 0x4>, <0x4a008184 0x4>, <0x4a008188 0x4>,
<0x4a00818c 0x4>;
         ...
};
Quoted 8 lines

It's at least self-documenting to do something like this:

dpll_usb_ck: dpll_usb_ck at ... {
         ...
         control-reg = < ... >;
         idlest-reg = < ... >;
         .. etc. ..
};

Some earlier version had something along these lines, but it was turned down. I had also reg-names as documentation purposes along, but this was unnecessary and was dropped also.

Quoted 4 lines

Which itself might not even be needed, depending on how the DPLL control code is implemented. For example, if the relative offsets are always the same for all OMAP4-family devices, maybe there's not even a need to explicitly encode that into the DT data.

If I want to get rid of these, I need to add extra compatible strings for the dpll types. There are several weird register offsets for omap3/am3 devices. omap4/5/dra7/am4 behave more sanely.

Quoted 3 lines

B. Seems like you can remove the "ti," prefix on the properties, since they have no pretentions at genericity: they are specific to the PRCM/CM/PRM IP block data, and registered by those drivers.

Quoted 6 lines
Quoted 2 lines

Can or should? It seems existing bindings use "ti," prefix even on non-generic bindings, meaning if I look at any other data in DT.

My comments in #8 are just regarding minor issues that don't seem right. I don't have a significant objection to staying with the existing property names here if you think that they are important.

Ok thanks, I'll be reworking most of the other items and will hopefully have something ready soonish. Taking care of the register mappings is rather nasty thing to do.

-Tero

Quoted 20 lines
Quoted 7 lines
Quoted 7 lines

C. Looks like the patches use the property "autoidle-low" to indicate that the autoidle bit should be inverted. "Low" seems like the wrong expression here - it invokes the actual voltage logic level of a hardware signal, and we have no idea whether the hardware signal is using a low voltage or a high voltage to express this condition. Would suggest something like 'invert-autoidle-bit' instead.

Quoted 8 lines
Quoted 4 lines

D. Regarding "ti,index-starts-at-one", it seems best to explicitly state which index starts at one. The code mentions a "mux index" so please consider renaming this something like "mux-index-starts-at-one" or "one-based-mux-index"

The index is explicitly defined by the clock node where this is present. If it is a mux-clock, then it is for mux-index. If for divider clock, it is index for the divider.

- Paul

maintainer
#62 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

In reply to Paul Walmsley
Quoted 28 lines · Paul Walmsley
Quoted 1 line · Tony Lindgren

And assuming Paul is OK with these patches in general.

Actually, I have several concerns with this series, as you and I discussed. Some of us have been talking them over with other folks for the past few months to try to figure out what to do. Most of the concerns have fairly easy technical solutions, but we shouldn't merge these patches until they're resolved.

The issues are:

1. whether the clock data should go into DT
2. what the right place for DT clock data is in the DT
3. whether it makes sense to merge "experimental" DT bindings in this case
4. where clockdomain data belongs

The first issue - and the one I'm least concerned about - is that, in my view, it still does not make technical sense to move this data into DT. This is chip-level hardware data that never needs to be changed by board designers or end users. Adding it to DT is going to result in a boot-time cost (due to DT parse overhead) and additional memory consumption for very little benefit, compared to an implementation that places this data into a dynamically loadable module. For some users, the boot-time bloat is a big
deal: the example that's been mentioned to me recently is an automotive back-up camera that needs to cold-boot to complete functionality in a few hundred microseconds.

Quoted 5 lines

However, according to some other upstream maintainers, Linus's goal is to move most of the device-specific static data out of the kernel tree, into DT files (in the ARM case). If that non-technical constraint is indeed the dominant consideration, then I agree that moving this data to DT is the only viable course of action.

I've been advocating that any combination of dt defined clocks, loadable module clocks and clocks loaded via /lib/firmware should be allowed.

That allows mapping just the bootime clocks in DT, or all the clocks in DT depending what suits the use case.

And yes, based on the flames I've received over the years, Linus clearly wants the SoC specific data out of the kernel. And there's already a lot of the SoC specific data that's now building up in various linux generic frameworks..

Quoted 35 lines

The second issue concerns where the SoC clock nodes should go in the DT. In these patches, the clock data has been encoded in a large "clocks" node off the top level. This is almost certainly not the right thing to do from a device hardware point of view. These clocks don't exist as standalone devices with their own address spaces decoded on the interconnect. In all of the SoC cases that I'm aware of, clock control registers are part of larger IP blocks.

For example, in the OMAP case, most of the system integration clock control registers are part of the OMAP-specific PRCM block, PRM block, or CM block. Then there are other device-specific clocks, like DSS PLLs or UART dividers. The control registers for these are generally located in the subsystem or device IP block itself, and are inaccessible when the subsystem or IP block is disabled. These device-specific clocks belong to the IP block, not the SoC integration. So, for example, if two SoCs use the same IP block, then the clock registers, and their offsets from the IP block address space base, are likely to be identical.

So in my view, the right things to do here are to:

1. associate SoC DT clock data with the device node that contains the clock control registers

2. specify relative offsets for clock registers from the start of the IP block address range, rather than absolute addresses for clock registers

3. place the responsibility for registering clocks into the IP block drivers themselves

This naturally separates clocks into per-IP block DT files. It also provides the CCF with an easy way to ensure that the device that encloses the clock is enabled and accessible by the CPU core, before trying to access registers inside.

Quoted 5 lines

Similarly, non-SoC off-chip clock data (such as for dedicated I2C PLLs) should also be associated with their I2C device node.

Making these changes to Tero's existing patches should be relatively straightforward, based on what I've seen.

Yes that's a valid concern and makes sense to me. It also it seems that moving clocks around with this approach does not break the binding ABI.

Quoted 16 lines

Regarding the third issue: let's postulate for the moment that the clock binding issues that I mention in #2 above are ignored (ha!), and that the existing DT clock data is merged. These bindings are currently marked as "Unstable - ABI compatibility may be broken in the future". What happens if, when we meet to discuss clock bindings at the kernel summit, we decide that significant changes are needed? We could easily wind up with kernels that won't boot at all when used with newer DT data.

Not to mention, merging such a large amount of code and data before the bindings are stable will increase the risk of massive churn as the bindings evolve towards stability.

So IMHO, the way to fix this is to consider the clock data to be IP-block specific, as mentioned in #2. Then there's no need for global clock bindings for the SoC clock cases.

Quoted 5 lines

Otherwise, it seems prudent to at least wait until the global clock bindings are no longer marked as unstable. The whole DT migration was predicated on the ideas of reducing churn in the Linux codebase, and preserving forward compatibility for DT data. We shouldn't discard these goals just to merge something a little sooner.

Mike suggested using ti specific bindings as the generic bindings may not be ready for quite some time. In any case, these bindings seem like they can be easily supported in the long run no matter what happens with the Linux generic clock bindings.

Quoted 11 lines

The fourth issue is where the clockdomain data should go. The basic issue here is that the notion of "clockdomains" here is completely OMAP-specific. OMAP clockdomains demarcate a group of clocks or IP blocks that share the same automatic idle behavior, controlled by their enclosing IP block (e.g., PRCM, PRM, CM, etc.). Clockdomains are also used in the OMAP kernel code to link the clocks to their enclosing power domains, voltage rail control, etc.

So since these are OMAP PRCM/CM/PRM-specific constructs, the right place for OMAP clockdomain data is underneath the OMAP-specific CM/PRCM DT nodes. Again this should be an easy change to Tero's existing patches.

I believe Mike has made it a requirement that the clockdomains will be moved out of the clock code in follow up patches.

Quoted 6 lines

So hey, if you've made it this far, thanks for reading. My sense is that implementing #2 and #4 are relatively straightforward.

Also, since this seems to have been a problem for some folks in the past, I want to make clear that I think both Mike and Tero have been doing good jobs on the CCF and OMAP clock patches so far.

Yes I agree.

Tony

#63 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

In reply to Paul Walmsley

Hey Paul,

My dibs on this below.

Quoted 17 lines · Paul Walmsley
Quoted 1 line · Tony Lindgren

And assuming Paul is OK with these patches in general.

Actually, I have several concerns with this series, as you and I discussed. Some of us have been talking them over with other folks for the past few months to try to figure out what to do. Most of the concerns have fairly easy technical solutions, but we shouldn't merge these patches until they're resolved.

The issues are:

1. whether the clock data should go into DT
2. what the right place for DT clock data is in the DT
3. whether it makes sense to merge "experimental" DT bindings in this case
4. where clockdomain data belongs

Quoted 10 lines

The first issue - and the one I'm least concerned about - is that, in my view, it still does not make technical sense to move this data into DT. This is chip-level hardware data that never needs to be changed by board designers or end users. Adding it to DT is going to result in a boot-time cost (due to DT parse overhead) and additional memory consumption for very little benefit, compared to an implementation that places this data into a dynamically loadable module. For some users, the boot-time bloat is a big
deal: the example that's been mentioned to me recently is an automotive back-up camera that needs to cold-boot to complete functionality in a few hundred microseconds.

Personally I share the concern, I don't see much point in using the DT for any kind of purpose... it is just another binary compatibility breaker in the picture in addition to boot-loader, and it basically does not solve any _real_ problems either.

Boot time issues can be solved by adding alternative clock data sources like Tony pointed out, granted, we don't have support for those yet, but we need to start somewhere.

Quoted 5 lines

However, according to some other upstream maintainers, Linus's goal is to move most of the device-specific static data out of the kernel tree, into DT files (in the ARM case). If that non-technical constraint is indeed the dominant consideration, then I agree that moving this data to DT is the only viable course of action.

Yeah... Personally I can't see any other way forward right now either as I was basically given the option to use DT for this work or not do it at all...

Quoted 11 lines

...

The second issue concerns where the SoC clock nodes should go in the DT. In these patches, the clock data has been encoded in a large "clocks" node off the top level. This is almost certainly not the right thing to do from a device hardware point of view. These clocks don't exist as standalone devices with their own address spaces decoded on the interconnect. In all of the SoC cases that I'm aware of, clock control registers are part of larger IP blocks.

Quoted 9 lines

For example, in the OMAP case, most of the system integration clock control registers are part of the OMAP-specific PRCM block, PRM block, or CM block. Then there are other device-specific clocks, like DSS PLLs or UART dividers. The control registers for these are generally located in the subsystem or device IP block itself, and are inaccessible when the subsystem or IP block is disabled. These device-specific clocks belong to the IP block, not the SoC integration. So, for example, if two SoCs use the same IP block, then the clock registers, and their offsets from the IP block address space base, are likely to be identical.

None of the clock registers defined in this set reside outside PRCM / control module, so they are always accessible. IP block internal dividers for UART and like, are defined and used only internally by the device drivers.

Quoted 5 lines

So in my view, the right things to do here are to:

1. associate SoC DT clock data with the device node that contains the clock control registers

So, either "cm", "prcm", and maybe "control_module" instead of current "clocks". How much do we benefit from this? This would get rid of need to call of_iomap() for each register basically.

Quoted 13 lines

2. specify relative offsets for clock registers from the start of the IP block address range, rather than absolute addresses for clock registers

3. place the responsibility for registering clocks into the IP block drivers themselves

This naturally separates clocks into per-IP block DT files. It also provides the CCF with an easy way to ensure that the device that encloses the clock is enabled and accessible by the CPU core, before trying to access registers inside.

Quoted 5 lines

Similarly, non-SoC off-chip clock data (such as for dedicated I2C PLLs) should also be associated with their I2C device node.

Making these changes to Tero's existing patches should be relatively straightforward, based on what I've seen.

For the set, it doesn't matter where the clock nodes reside, if someone can point me out where to put them, they can be easily moved around. Separating the individual clocks based on their IP mapping can probably just be done by checking their register address.

Anyway, I would like to have a concensus if something like this should be done or not before starting to do the changes, as this is going to change the layout of the clock data completely, and will require a complete new testing + debugging round to be done, in addition to script changes.

Quoted 18 lines

...

Regarding the third issue: let's postulate for the moment that the clock binding issues that I mention in #2 above are ignored (ha!), and that the existing DT clock data is merged. These bindings are currently marked as "Unstable - ABI compatibility may be broken in the future". What happens if, when we meet to discuss clock bindings at the kernel summit, we decide that significant changes are needed? We could easily wind up with kernels that won't boot at all when used with newer DT data.

Not to mention, merging such a large amount of code and data before the bindings are stable will increase the risk of massive churn as the bindings evolve towards stability.

So IMHO, the way to fix this is to consider the clock data to be IP-block specific, as mentioned in #2. Then there's no need for global clock bindings for the SoC clock cases.

Quoted 5 lines

Otherwise, it seems prudent to at least wait until the global clock bindings are no longer marked as unstable. The whole DT migration was predicated on the ideas of reducing churn in the Linux codebase, and preserving forward compatibility for DT data. We shouldn't discard these goals just to merge something a little sooner.

Well, we can keep support for the vendor specific clock bindings in the kernel and if generic bindings will eventually go forward, we can evolve towards that. A pressing problem for this work to proceed is that we need to get boot support for some new SoCs in rather sooner than again wait two years and work with TI internal trees. And getting the boot support quickly in, it seems the only way to get it done is via vendor specific bindings, which are unstable in a sense that _if_ the generic bindings go forward, we should start using those instead.

Quoted 10 lines

...

The fourth issue is where the clockdomain data should go. The basic issue here is that the notion of "clockdomains" here is completely OMAP-specific. OMAP clockdomains demarcate a group of clocks or IP blocks that share the same automatic idle behavior, controlled by their enclosing IP block (e.g., PRCM, PRM, CM, etc.). Clockdomains are also used in the OMAP kernel code to link the clocks to their enclosing power domains, voltage rail control, etc.

Quoted 3 lines

So since these are OMAP PRCM/CM/PRM-specific constructs, the right place for OMAP clockdomain data is underneath the OMAP-specific CM/PRCM DT nodes. Again this should be an easy change to Tero's existing patches.

Yea, does not matter where in the dts files these reside. Clockdomain bindings / code implementation can be moved to wherever the CM code is going to move, and this should not cause any changes to the clockdomain binding definitions. Anyway, again I would like to have a clear concensus if the bindings should move someplace else before doing the changes.

-Tero

Quoted 14 lines

...

So hey, if you've made it this far, thanks for reading. My sense is that implementing #2 and #4 are relatively straightforward.

Also, since this seems to have been a problem for some folks in the past, I want to make clear that I think both Mike and Tero have been doing good jobs on the CCF and OMAP clock patches so far.

regards

- Paul

maintainer
#64 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 9 lines · Tero Kristo
Quoted 4 lines · Paul Walmsley

So in my view, the right things to do here are to:

1. associate SoC DT clock data with the device node that contains the clock control registers

So, either "cm", "prcm", and maybe "control_module" instead of current "clocks". How much do we benefit from this? This would get rid of need to call of_iomap() for each register basically.

It seems that all you need to do for that is to set up two or three clock driver instances and have the clocks registered under them.

Then later on those can be children of cm, prcm and scm core drivers as needed without much of an issue.

Regards,

Tony

#65 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

In reply to Tero Kristo

Quoting Tero Kristo (2013-09-25 01:48:06)

Quoted 15 lines

Hi all,

Version 7 contains following high level changes:
- Dropped support for basic bindings from Mike Turquette, instead using vendor specific bindings for all clocks
- Mux clock + divider clock vendor specific bindings get rid of use of the bit-masks, instead these are generated runtime based on parent clock / divider min/max info, see individual patches for details
- Added support for dummy_ck nodes, was missing from previous version
- Fixed support for omap3630
- Added support for new clock node type: ti,mux-gate-clock (using composite clock type)
- OMAP4 / OMAP5 only build fixes (compiler flag checks added to some files)
- most of the of_omap_* calls renamed to of_ti_*
- Rebased on top of v3.12-rc2

Hi Tero,

I had one really minor request for one of the patches, otherwise I'm pretty happy with this series and I can take in the clock parts for 3.13. A few general comments:

1) I think I mentioned some time back that it would be wise to put a disclaimer at the top of each DT binding description stating something to the effect: "Binding status: Unstable - ABI compatibility may be broken in the future". Are you OK merging these bindings as-is without that kind of a disclaimer?

2) I hope to see the clockdomain stuff go away in the future. I'm OK to take it for now to aid in this DT transition but I would like some commitment that it will not stay in drivers/clk/omap forever. I guess for clockdomain you'll need to figure out how to handle those in a generic driver way.

3) Same concern as above but for the DT clkdev alias stuff. I guess you'll need to make all of the dependent drivers play nice with DT. Do you plan to remove it within a reasonable timeframe? It would be a nice reduction in LoC and more importantly it would mean that drivers were using clkdev in a more-correct fashion.

Regards,
Mike

Quoted 38 lines

Some detailed comments about patch level changes (if no mention, no major changes):
- Patch #1:
* removed use of reg-names property, instead registers mapped using compatible string
* removed ti,modes property, instead uses DPLL mode flags now
* separated AM33XX DPLLs under their own compatible strings
- Patch #4 & #5: new patches
- Patch #6: merged basic gate support from Mike to this patch as vendor specific gate clock support
- Patch #9 & #10: new patches
- Patch #11: dummy clocks added, USB / ABE DPLL setup ordering changed
- Patch #14: dummy clocks added
- Patch #20: dropped usage of reg-names property
- Patch #21: dummy clocks added
- Patch #22 & #23: new patches
- Patch #30: AM35XX clock data added to this patch
- Patch #31:
* dummy clocks added
* added missing 3630 clocks

Test branch available here:
https://github.com/t-kristo/linux-pm.git
branch: mainline-3.12-rc2-ti-dt-clks-v7

Testing done:
- am335x-bone : boot only
- omap5-sevm : boot only
- dra7-evm : boot only
- omap3-beagle : boot + suspend/resume (ret + off)
- omap4-panda-es : boot + suspend/resume

Nishanth has also been doing some tests with omap3-beagle-xm with some WIP versions of this branch, but not with this latest one. Nishanth, maybe you can provide test results to this one also?

-Tero

#66 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 17 lines · Mike Turquette

Quoting Tero Kristo (2013-09-25 01:48:06)

Quoted 15 lines

Hi all,

Version 7 contains following high level changes:
- Dropped support for basic bindings from Mike Turquette, instead using vendor specific bindings for all clocks
- Mux clock + divider clock vendor specific bindings get rid of use of the bit-masks, instead these are generated runtime based on parent clock / divider min/max info, see individual patches for details
- Added support for dummy_ck nodes, was missing from previous version
- Fixed support for omap3630
- Added support for new clock node type: ti,mux-gate-clock (using composite clock type)
- OMAP4 / OMAP5 only build fixes (compiler flag checks added to some files)
- most of the of_omap_* calls renamed to of_ti_*
- Rebased on top of v3.12-rc2

Quoted 5 lines

Hi Tero,

I had one really minor request for one of the patches, otherwise I'm pretty happy with this series and I can take in the clock parts for 3.13. A few general comments:

Yea, I'll add the is_enabled check to the APLL and repost.

Quoted 6 lines

1) I think I mentioned some time back that it would be wise to put a disclaimer at the top of each DT binding description stating something to the effect: "Binding status: Unstable - ABI compatibility may be broken in the future". Are you OK merging these bindings as-is without that kind of a disclaimer?

Well, personally I don't mind, but if you think this would be good, I'll add this to next rev.

Quoted 5 lines

2) I hope to see the clockdomain stuff go away in the future. I'm OK to take it for now to aid in this DT transition but I would like some commitment that it will not stay in drivers/clk/omap forever. I guess for clockdomain you'll need to figure out how to handle those in a generic driver way.

I think the clockdomain stuff should be possible to move to wherever CM driver is going to move. drivers/power/omap/... or something maybe. I have been starting some initial work on the CM driver cleanup for this purpose.

Quoted 5 lines

3) Same concern as above but for the DT clkdev alias stuff. I guess you'll need to make all of the dependent drivers play nice with DT. Do you plan to remove it within a reasonable timeframe? It would be a nice reduction in LoC and more importantly it would mean that drivers were using clkdev in a more-correct fashion.

This is probably harder, I think Tony is better to comment here, but my idea is that in the worst case we can just break non-conforming drivers by dropping the clkdev tables if nothing else helps. :) Most of the tables are currently needed by hwmod, and there are already patches around for adding DT clock support for hwmods, so once these go in, we can at least reduce the table sizes considerably already.

-Tero

Quoted 42 lines

Regards,
Mike

Quoted 38 lines

Some detailed comments about patch level changes (if no mention, no major changes):
- Patch #1:
* removed use of reg-names property, instead registers mapped using compatible string
* removed ti,modes property, instead uses DPLL mode flags now
* separated AM33XX DPLLs under their own compatible strings
- Patch #4 & #5: new patches
- Patch #6: merged basic gate support from Mike to this patch as vendor specific gate clock support
- Patch #9 & #10: new patches
- Patch #11: dummy clocks added, USB / ABE DPLL setup ordering changed
- Patch #14: dummy clocks added
- Patch #20: dropped usage of reg-names property
- Patch #21: dummy clocks added
- Patch #22 & #23: new patches
- Patch #30: AM35XX clock data added to this patch
- Patch #31:
* dummy clocks added
* added missing 3630 clocks

Test branch available here:
https://github.com/t-kristo/linux-pm.git
branch: mainline-3.12-rc2-ti-dt-clks-v7

Testing done:
- am335x-bone : boot only
- omap5-sevm : boot only
- dra7-evm : boot only
- omap3-beagle : boot + suspend/resume (ret + off)
- omap4-panda-es : boot + suspend/resume

Nishanth has also been doing some tests with omap3-beagle-xm with some WIP versions of this branch, but not with this latest one. Nishanth, maybe you can provide test results to this one also?

-Tero

maintainer
#67 lore raw

v7 cover ARM: OMAP: clock data conversion to DT

Quoted 8 lines · Tero Kristo
Quoted 5 lines · Mike Turquette

3) Same concern as above but for the DT clkdev alias stuff. I guess you'll need to make all of the dependent drivers play nice with DT. Do you plan to remove it within a reasonable timeframe? It would be a nice reduction in LoC and more importantly it would mean that drivers were using clkdev in a more-correct fashion.

Quoted 6 lines

This is probably harder, I think Tony is better to comment here, but my idea is that in the worst case we can just break non-conforming drivers by dropping the clkdev tables if nothing else helps. :) Most of the tables are currently needed by hwmod, and there are already patches around for adding DT clock support for hwmods, so once these go in, we can at least reduce the table sizes considerably already.

It should be possible to remove these as we go as things become DT only. Probably the biggest blocker right now is making omap3 DT only.

But let's not start breaking things intentionally!

Tony