Skip to contentSkip to conversation

ARM: socfpga: Enable soft reset

PATCH Superseded dinguyen at altera.com 6 participants 15 messages Mar 19, 2013 – Apr 4, 2013

Subsystems (6)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/2 ARM: socfpga: Enable soft reset

From: Dinh Nguyen <[email protected]>

Enable a cold or warm reset to the HW from userspace.

Also fix a few sparse errors:

warning: symbol 'sys_manager_base_addr' was not declared. Should it be static?
warning: symbol 'rst_manager_base_addr' was not declared. Should it be static?

v2:
- Remove hotplug support while investigating on how implement PSCI

---
 arch/arm/mach-socfpga/core.h    |   11 +++++++++++
 arch/arm/mach-socfpga/platsmp.c |    3 ---
 arch/arm/mach-socfpga/socfpga.c |   10 +++++++++-
 3 files changed, 20 insertions(+), 4 deletions(-)
arch/arm/mach-socfpga/core.h +11 -0
index 315edff..572b8f7 100644
@@ -20,12 +20,23 @@
 #ifndef __MACH_CORE_H
 #define __MACH_CORE_H
 
+#define SOCFPGA_RSTMGR_CTRL	0x04
+#define SOCFPGA_RSTMGR_MODPERRST	0x14
+#define SOCFPGA_RSTMGR_BRGMODRST	0x1c
+
+/* System Manager bits */
+#define RSTMGR_CTRL_SWCOLDRSTREQ	0x1	/* Cold Reset */
+#define RSTMGR_CTRL_SWWARMRSTREQ	0x2	/* Warm Reset */
+
 extern void socfpga_secondary_startup(void);
 extern void __iomem *socfpga_scu_base_addr;
 
 extern void socfpga_init_clocks(void);
 extern void socfpga_sysmgr_init(void);
 
+extern void __iomem *sys_manager_base_addr;
+extern void __iomem *rst_manager_base_addr;
+
 extern struct smp_operations socfpga_smp_ops;
 extern char secondary_trampoline, secondary_trampoline_end;
 
arch/arm/mach-socfpga/platsmp.c +0 -3
index 84c60fa..b907fb9 100644
@@ -30,9 +30,6 @@
 
 #include "core.h"
 
-extern void __iomem *sys_manager_base_addr;
-extern void __iomem *rst_manager_base_addr;
-
 static void __cpuinit socfpga_secondary_init(unsigned int cpu)
 {
 	/*
arch/arm/mach-socfpga/socfpga.c +9 -1
index 1042c02..b41a945 100644
@@ -87,7 +87,15 @@ static void __init socfpga_init_irq(void)
 
 static void socfpga_cyclone5_restart(char mode, const char *cmd)
 {
-	/* TODO: */
+	u32 temp;
+
+	temp = __raw_readl(rst_manager_base_addr + SOCFPGA_RSTMGR_CTRL);
+
+	if (mode == 'h')
+		temp |= RSTMGR_CTRL_SWCOLDRSTREQ;
+	else
+		temp |= RSTMGR_CTRL_SWWARMRSTREQ;
+	__raw_writel(temp, rst_manager_base_addr + SOCFPGA_RSTMGR_CTRL);
 }
 
 static void __init socfpga_cyclone5_init(void)
Signature · 1 line
1.7.9.5
#1 lore raw

PATCH v2 2/2 ARM: socfpga: Add clock entries into device tree

From: Dinh Nguyen <[email protected]>

Adds the main PLL clock groups for SOCFPGA into device tree file so that the clock framework to query the clock and clock rates appropriately.

$cat /sys/kernel/debug/clk/clk_summary
clock enable_cnt prepare_cnt rate
---------------------------------------------------------------------

 osc1                           2           2            25000000
    sdram_pll                   0           0            400000000
       s2f_usr2_clk             0           0            66666666
       ddr_dq_clk               0           0            200000000
       ddr_2x_dqs_clk           0           0            400000000
       ddr_dqs_clk              0           0            200000000
    periph_pll                  2           2            500000000
       s2f_usr1_clk             0           0            50000000
       per_base_clk             4           4            100000000
       per_nand_mmc_clk         0           0            25000000
       per_qsi_clk              0           0            250000000
       emac1_clk                1           1            125000000
       emac0_clk                0           0            125000000
    main_pll                    1           1            1600000000
       cfg_s2f_usr0_clk         0           0            100000000
       main_nand_sdmmc_clk      0           0            100000000
       main_qspi_clk            0           0            400000000
       dbg_base_clk             0           0            400000000
       mainclk                  0           0            400000000
       mpuclk                   1           1            800000000
          smp_twd               1           1            200000000

v2:
- Add copyright from Calxeda for clk.c
- Add device tree "fixed-divider" for clocks with fixed dividers
- Fix space and tab issues

---
 .../bindings/arm/altera/socfpga-clk-manager.txt    |   11 ++
 .../devicetree/bindings/clock/altr_socfpga.txt     |   18 ++
 arch/arm/boot/dts/socfpga.dtsi                     |  157 +++++++++++++++++
 arch/arm/boot/dts/socfpga_cyclone5.dts             |    8 +
 arch/arm/boot/dts/socfpga_vt.dts                   |    8 +
 arch/arm/mach-socfpga/socfpga.c                    |    6 +
 drivers/clk/socfpga/clk.c                          |  179 +++++++++++++++++---
 7 files changed, 366 insertions(+), 21 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/arm/altera/socfpga-clk-manager.txt
 create mode 100644 Documentation/devicetree/bindings/clock/altr_socfpga.txt
Documentation/devicetree/bindings/arm/altera/socfpga-clk-manager.txt new file +11 -0
new file mode 100644
index 0000000..2c28f1d
@@ -0,0 +1,11 @@
+Altera SOCFPGA Clock Manager
+
+Required properties:
+- compatible : "altr,clk-mgr"
+- reg : Should contain base address and length for Clock Manager
+
+Example:
+	 clkmgr at ffd04000 {
+		compatible = "altr,clk-mgr";
+		reg = <0xffd04000 0x1000>;
+	};
Documentation/devicetree/bindings/clock/altr_socfpga.txt new file +18 -0
new file mode 100644
index 0000000..bd0c841
@@ -0,0 +1,18 @@
+Device Tree Clock bindings for Altera's SoCFPGA platform
+
+This binding uses the common clock binding[1].
+
+[1] Documentation/devicetree/bindings/clock/clock-bindings.txt
+
+Required properties:
+- compatible : shall be one of the following:
+	"altr,socfpga-pll-clock" - for a PLL clock
+	"altr,socfpga-perip-clock" - The peripheral clock divided from the
+		PLL clock.
+- reg : shall be the control register offset from CLOCK_MANAGER's base for the clock.
+- clocks : shall be the input parent clock phandle for the clock. This is
+	either an oscillator or a pll output.
+- #clock-cells : from common clock binding, shall be set to 0.
+
+Optional properties:
+- fixed-divider : If clocks have a fixed divider value, use this property.
arch/arm/boot/dts/socfpga.dtsi +157 -0

163 lines not on this page · raw message

arch/arm/boot/dts/socfpga_cyclone5.dts +8 -0
index 3ae8a83..2495958 100644
@@ -33,6 +33,14 @@
 	};
 
 	soc {
+		clkmgr at ffd04000 {
+			clocks {
+				osc1 {
+					clock-frequency = <25000000>;
+				};
+			};
+		};
+
 		timer0 at ffc08000 {
 			clock-frequency = <100000000>;
 		};
arch/arm/boot/dts/socfpga_vt.dts +8 -0
index 1036eba..0bf035d 100644
@@ -33,6 +33,14 @@
 	};
 
 	soc {
+		clkmgr at ffd04000 {
+			clocks {
+				osc1 {
+					clock-frequency = <10000000>;
+				};
+			};
+		};
+
 		timer0 at ffc08000 {
 			clock-frequency = <7000000>;
 		};
arch/arm/mach-socfpga/socfpga.c +6 -0
index b41a945..856625a 100644
@@ -15,6 +15,7 @@
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 #include <linux/dw_apb_timer.h>
+#include <linux/clk-provider.h>
 #include <linux/irqchip.h>
 #include <linux/of_address.h>
 #include <linux/of_irq.h>
@@ -29,6 +30,7 @@
 void __iomem *socfpga_scu_base_addr = ((void __iomem *)(SOCFPGA_SCU_VIRT_BASE));
 void __iomem *sys_manager_base_addr;
 void __iomem *rst_manager_base_addr;
+void __iomem *clk_mgr_base_addr;
 unsigned long cpu1start_addr;
 
 static struct map_desc scu_io_desc __initdata = {
@@ -77,6 +79,9 @@ void __init socfpga_sysmgr_init(void)
 
 	np = of_find_compatible_node(NULL, NULL, "altr,rst-mgr");
 	rst_manager_base_addr = of_iomap(np, 0);
+
+	np = of_find_compatible_node(NULL, NULL, "altr,clk-mgr");
+	clk_mgr_base_addr = of_iomap(np, 0);
 }
 
 static void __init socfpga_init_irq(void)
@@ -102,6 +107,7 @@ static void __init socfpga_cyclone5_init(void)
 {
 	l2x0_of_init(0, ~0UL);
 	of_platform_populate(NULL, of_default_bus_match_table, NULL, NULL);
+	of_clk_init(NULL);
 	socfpga_init_clocks();
 }
 
drivers/clk/socfpga/clk.c +158 -21

204 lines not on this page · raw message

Signature · 1 line
1.7.9.5
#2 lore raw

PATCH v2 2/2 ARM: socfpga: Add clock entries into device tree

Quoting dinguyen at altera.com (2013-03-19 08:45:36)

Quoted 31 lines

From: Dinh Nguyen <[email protected]>

Adds the main PLL clock groups for SOCFPGA into device tree file so that the clock framework to query the clock and clock rates appropriately.

$cat /sys/kernel/debug/clk/clk_summary
clock enable_cnt prepare_cnt rate
---------------------------------------------------------------------

 osc1                           2           2            25000000
    sdram_pll                   0           0            400000000
       s2f_usr2_clk             0           0            66666666
       ddr_dq_clk               0           0            200000000
       ddr_2x_dqs_clk           0           0            400000000
       ddr_dqs_clk              0           0            200000000
    periph_pll                  2           2            500000000
       s2f_usr1_clk             0           0            50000000
       per_base_clk             4           4            100000000
       per_nand_mmc_clk         0           0            25000000
       per_qsi_clk              0           0            250000000
       emac1_clk                1           1            125000000
       emac0_clk                0           0            125000000
    main_pll                    1           1            1600000000
       cfg_s2f_usr0_clk         0           0            100000000
       main_nand_sdmmc_clk      0           0            100000000
       main_qspi_clk            0           0            400000000
       dbg_base_clk             0           0            400000000
       mainclk                  0           0            400000000
       mpuclk                   1           1            800000000
          smp_twd               1           1            200000000
Quoted 1 line

There is a lot going on in this one patch. I would prefer to see the clock driver broken out separately.

<snip>

Quoted 4 lines
drivers/clk/socfpga/clk.c +0 -0
index 2c855a6..da6b461 100644

<snip>

Quoted 12 lines
drivers/clk/socfpga/clk.c +12 -0
+static int clk_pll_enable(struct clk_hw *hwclk)
+{
+       struct socfpga_clk *socfpgaclk = to_socfpga_clk(hwclk);
+       u32 reg;
+
+       reg = readl(socfpgaclk->reg);
+       reg |= SOCFPGA_PLL_EXT_ENA;
+       writel(reg, socfpgaclk->reg);
+
+       return 0;
+}
+

drivers/clk/socfpga/clk.c line 76

Quoted 10 lines
drivers/clk/socfpga/clk.c +9 -0
+static void clk_pll_disable(struct clk_hw *hwclk)
 {
+       struct socfpga_clk *socfpgaclk = to_socfpga_clk(hwclk);
+       u32 reg;
+
+       reg = readl(socfpgaclk->reg);
+       reg &= ~SOCFPGA_PLL_EXT_ENA;
+       writel(reg, socfpgaclk->reg);
+}
+

For a simple enable which just sets a bit, you might want to re-use the basic gate clock type. This can be done similar to the composite clock patches (currently on the list) by stuffing a clk_gate structure into your custom socfpga_clk type.

Regards,
Mike

#3 lore raw

PATCH v2 2/2 ARM: socfpga: Add clock entries into device tree

Hi Mike,

Quoted 32 lines · Mike Turquette

Quoting dinguyen at altera.com (2013-03-19 08:45:36)

Quoted 31 lines

From: Dinh Nguyen <[email protected]>

Adds the main PLL clock groups for SOCFPGA into device tree file so that the clock framework to query the clock and clock rates appropriately.

$cat /sys/kernel/debug/clk/clk_summary
clock enable_cnt prepare_cnt rate
---------------------------------------------------------------------

 osc1                           2           2            25000000
    sdram_pll                   0           0            400000000
       s2f_usr2_clk             0           0            66666666
       ddr_dq_clk               0           0            200000000
       ddr_2x_dqs_clk           0           0            400000000
       ddr_dqs_clk              0           0            200000000
    periph_pll                  2           2            500000000
       s2f_usr1_clk             0           0            50000000
       per_base_clk             4           4            100000000
       per_nand_mmc_clk         0           0            25000000
       per_qsi_clk              0           0            250000000
       emac1_clk                1           1            125000000
       emac0_clk                0           0            125000000
    main_pll                    1           1            1600000000
       cfg_s2f_usr0_clk         0           0            100000000
       main_nand_sdmmc_clk      0           0            100000000
       main_qspi_clk            0           0            400000000
       dbg_base_clk             0           0            400000000
       mainclk                  0           0            400000000
       mpuclk                   1           1            800000000
          smp_twd               1           1            200000000
Quoted 4 lines
Quoted 1 line

There is a lot going on in this one patch. I would prefer to see the clock driver broken out separately.

Not sure what you mean by breaking out the clock driver separately. The patch is only touching the clocks for mach-socfpga.

The patch is

7 files changed, 366 insertions(+), 21 deletions(-)

while the patch to enable clk-highbank was:

8 files changed, 463 insertions(+), 64 deletions(-)
Quoted 26 lines

<snip>

Quoted 4 lines
drivers/clk/socfpga/clk.c +0 -0
index 2c855a6..da6b461 100644

<snip>

Quoted 19 lines
drivers/clk/socfpga/clk.c +18 -0
+static int clk_pll_enable(struct clk_hw *hwclk)
+{
+       struct socfpga_clk *socfpgaclk = to_socfpga_clk(hwclk);
+       u32 reg;
+
+       reg = readl(socfpgaclk->reg);
+       reg |= SOCFPGA_PLL_EXT_ENA;
+       writel(reg, socfpgaclk->reg);
+
+       return 0;
+}
+
+static void clk_pll_disable(struct clk_hw *hwclk)
 {
+       struct socfpga_clk *socfpgaclk = to_socfpga_clk(hwclk);
+       u32 reg;
+
+       reg = readl(socfpgaclk->reg);
+       reg &= ~SOCFPGA_PLL_EXT_ENA;
Quoted 8 lines

drivers/clk/socfpga/clk.c line 66

Quoted 3 lines
drivers/clk/socfpga/clk.c +3 -0
+       writel(reg, socfpgaclk->reg);
+}
+

For a simple enable which just sets a bit, you might want to re-use the basic gate clock type. This can be done similar to the composite clock patches (currently on the list) by stuffing a clk_gate structure into your custom socfpga_clk type.

I'll take a look at the list about this.

Thanks for the review.

Dinh

Quoted 4 lines

Regards,
Mike

#4 lore raw

PATCH v2 2/2 ARM: socfpga: Add clock entries into device tree

Quoting Dinh Nguyen (2013-03-19 11:45:50)

Quoted 35 lines

Hi Mike,

Quoted 32 lines · Mike Turquette

Quoting dinguyen at altera.com (2013-03-19 08:45:36)

Quoted 31 lines

From: Dinh Nguyen <[email protected]>

Adds the main PLL clock groups for SOCFPGA into device tree file so that the clock framework to query the clock and clock rates appropriately.

$cat /sys/kernel/debug/clk/clk_summary
clock enable_cnt prepare_cnt rate
---------------------------------------------------------------------

 osc1                           2           2            25000000
    sdram_pll                   0           0            400000000
       s2f_usr2_clk             0           0            66666666
       ddr_dq_clk               0           0            200000000
       ddr_2x_dqs_clk           0           0            400000000
       ddr_dqs_clk              0           0            200000000
    periph_pll                  2           2            500000000
       s2f_usr1_clk             0           0            50000000
       per_base_clk             4           4            100000000
       per_nand_mmc_clk         0           0            25000000
       per_qsi_clk              0           0            250000000
       emac1_clk                1           1            125000000
       emac0_clk                0           0            125000000
    main_pll                    1           1            1600000000
       cfg_s2f_usr0_clk         0           0            100000000
       main_nand_sdmmc_clk      0           0            100000000
       main_qspi_clk            0           0            400000000
       dbg_base_clk             0           0            400000000
       mainclk                  0           0            400000000
       mpuclk                   1           1            800000000
          smp_twd               1           1            200000000
Quoted 8 lines
Quoted 4 lines
Quoted 1 line

There is a lot going on in this one patch. I would prefer to see the clock driver broken out separately.

Not sure what you mean by breaking out the clock driver separately. The patch is only touching the clocks for mach-socfpga.

I mean breaking the change to drivers/clk/socfpga/clk.c out into a separate patch.

Quoted 6 lines

The patch is

7 files changed, 366 insertions(+), 21 deletions(-)

while the patch to enable clk-highbank was:

8 files changed, 463 insertions(+), 64 deletions(-)

That is a fair comparison. However I still prefer to see data (e.g. dts changes) separated from logic (clk.c changes). I think it makes for a cleaner git history and makes patches more readable too.

Regards,
Mike

Quoted 44 lines
Quoted 26 lines

<snip>

Quoted 4 lines
drivers/clk/socfpga/clk.c +0 -0
index 2c855a6..da6b461 100644

<snip>

Quoted 19 lines
drivers/clk/socfpga/clk.c +18 -0
+static int clk_pll_enable(struct clk_hw *hwclk)
+{
+       struct socfpga_clk *socfpgaclk = to_socfpga_clk(hwclk);
+       u32 reg;
+
+       reg = readl(socfpgaclk->reg);
+       reg |= SOCFPGA_PLL_EXT_ENA;
+       writel(reg, socfpgaclk->reg);
+
+       return 0;
+}
+
+static void clk_pll_disable(struct clk_hw *hwclk)
 {
+       struct socfpga_clk *socfpgaclk = to_socfpga_clk(hwclk);
+       u32 reg;
+
+       reg = readl(socfpgaclk->reg);
+       reg &= ~SOCFPGA_PLL_EXT_ENA;
Quoted 8 lines

drivers/clk/socfpga/clk.c line 66

Quoted 3 lines
drivers/clk/socfpga/clk.c +3 -0
+       writel(reg, socfpgaclk->reg);
+}
+

For a simple enable which just sets a bit, you might want to re-use the basic gate clock type. This can be done similar to the composite clock patches (currently on the list) by stuffing a clk_gate structure into your custom socfpga_clk type.

I'll take a look at the list about this.

Thanks for the review.

Dinh

Quoted 4 lines

Regards,
Mike

#5 lore raw

PATCH v2 2/2 ARM: socfpga: Add clock entries into device tree

Hi Mike,

Quoted 12 lines · Mike Turquette

Quoting Dinh Nguyen (2013-03-19 11:45:50)

Quoted 10 lines

Hi Mike,

Quoted 7 lines · Mike Turquette

Quoting dinguyen at altera.com (2013-03-19 08:45:36)

Quoted 6 lines

From: Dinh Nguyen <[email protected]>

Adds the main PLL clock groups for SOCFPGA into device tree file so that the clock framework to query the clock and clock rates appropriately.

<snip>

Quoted 14 lines
Quoted 7 lines
Quoted 3 lines

There is a lot going on in this one patch. I would prefer to see the clock driver broken out separately.

Not sure what you mean by breaking out the clock driver separately. The patch is only touching the clocks for mach-socfpga.

I mean breaking the change to drivers/clk/socfpga/clk.c out into a separate patch.

Quoted 3 lines

The patch is

7 files changed, 366 insertions(+), 21 deletions(-)
Quoted 7 lines
Quoted 3 lines

while the patch to enable clk-highbank was:

8 files changed, 463 insertions(+), 64 deletions(-)

That is a fair comparison. However I still prefer to see data (e.g. dts changes) separated from logic (clk.c changes). I think it makes for a cleaner git history and makes patches more readable too.

I agree that it makes things alot cleaner to split DTS and code into separate patches, but at the same time the code is pretty much useless without the DTS entries. I apologize if there has already been a similar discussion on the list about this. I just want to make sure that I know to split up patches in the same manner in the future?

Quoted 40 lines

Regards,
Mike

Quoted 36 lines
Quoted 26 lines

<snip>

Quoted 4 lines
drivers/clk/socfpga/clk.c +0 -0
index 2c855a6..da6b461 100644

<snip>

Quoted 19 lines
drivers/clk/socfpga/clk.c +18 -0
+static int clk_pll_enable(struct clk_hw *hwclk)
+{
+       struct socfpga_clk *socfpgaclk = to_socfpga_clk(hwclk);
+       u32 reg;
+
+       reg = readl(socfpgaclk->reg);
+       reg |= SOCFPGA_PLL_EXT_ENA;
+       writel(reg, socfpgaclk->reg);
+
+       return 0;
+}
+
+static void clk_pll_disable(struct clk_hw *hwclk)
 {
+       struct socfpga_clk *socfpgaclk = to_socfpga_clk(hwclk);
+       u32 reg;
+
+       reg = readl(socfpgaclk->reg);
+       reg &= ~SOCFPGA_PLL_EXT_ENA;
Quoted 8 lines

drivers/clk/socfpga/clk.c line 66

Quoted 3 lines
drivers/clk/socfpga/clk.c +3 -0
+       writel(reg, socfpgaclk->reg);
+}
+

For a simple enable which just sets a bit, you might want to re-use the basic gate clock type. This can be done similar to the composite clock patches (currently on the list) by stuffing a clk_gate structure into your custom socfpga_clk type.

I'll take a look at the list about this.

I think mvebu/clk-gating-ctrl.c is doing the same thing you're suggesting right?

Thanks,
Dinh

Quoted 9 lines
Quoted 8 lines

Thanks for the review.

Dinh

Quoted 4 lines

Regards,
Mike

#6 lore raw

PATCH v2 2/2 ARM: socfpga: Add clock entries into device tree

Quoted 9 lines · Dinh Nguyen
Quoted 3 lines

That is a fair comparison. However I still prefer to see data (e.g. dts changes) separated from logic (clk.c changes). I think it makes for a cleaner git history and makes patches more readable too.

I agree that it makes things alot cleaner to split DTS and code into separate patches, but at the same time the code is pretty much useless without the DTS entries. I apologize if there has already been a similar discussion on the list about this. I just want to make sure that I know to split up patches in the same manner in the future?

In particular, it would be helpful to merge the dts changes through the arm-soc tree and the clk changes through Mike's clock tree. Ideally those two should be completely independent of one another, although you obviously need both to get the new feature.

Arnd

#7 lore raw

PATCH v2 2/2 ARM: socfpga: Add clock entries into device tree

In reply to dinguyen at altera.com

Hi!

Quoted 3 lines

Adds the main PLL clock groups for SOCFPGA into device tree file so that the clock framework to query the clock and clock rates appropriately.

Is there an easy way to test it?

I was keeping modified clk.c with hardcoded clock to have working ethernet; I reverted that patch and applied 2/2, but now NFS root refuses to mount (similar symptoms to missing clk.c modifications).

Also I need following patch to get it to compile:

commit f85232eec5330e64984facd300ef864c53a326f5

Author: Pavel Machek <[email protected]>
Date:   Wed Mar 20 14:41:59 2013 +0100

Patch needed to get clk.c to compile.

drivers/clk/socfpga/clk.c +1 -0
index da6b461..3504dbf 100644
@@ -49,6 +49,7 @@ struct socfpga_clk {
 	void __iomem *reg;
 	char *parent_name;
 	char *clk_name;
+	int fixed_div;
 };
 #define to_socfpga_clk(p) container_of(p, struct socfpga_clk, hw)
 

Thanks,
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#8 lore raw

PATCH v2 2/2 ARM: socfpga: Add clock entries into device tree

Hi!

Quoted 5 lines
Quoted 3 lines

Adds the main PLL clock groups for SOCFPGA into device tree file so that the clock framework to query the clock and clock rates appropriately.

Is there an easy way to test it?

Please consider cleanups below. If compiles and NFS root fails the same way it failed below.

Thanks,
Pavel

commit 82158c4cf015e3b8c3db1b5ba84722fe1a480dce

Author: Pavel Machek <[email protected]>
Date:   Wed Mar 20 14:56:56 2013 +0100

Convert clock_manager_base_addr to structure, to get type checking and nicer code.

Improve constants alignment a bit.

arch/arm/mach-socfpga/socfpga.c +1 -1
index 856625a..b88b992 100644
@@ -30,7 +30,7 @@
 void __iomem *socfpga_scu_base_addr = ((void __iomem *)(SOCFPGA_SCU_VIRT_BASE));
 void __iomem *sys_manager_base_addr;
 void __iomem *rst_manager_base_addr;
-void __iomem *clk_mgr_base_addr;
+struct clock_manager_hw __iomem *clk_mgr_base_addr;
 unsigned long cpu1start_addr;
 
 static struct map_desc scu_io_desc __initdata = {
drivers/clk/socfpga/clk.c +10 -8
index 3504dbf..0239ff5 100644
@@ -17,16 +17,13 @@
  * You should have received a copy of the GNU General Public License
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
+
 #include <linux/clk.h>
 #include <linux/clkdev.h>
 #include <linux/clk-provider.h>
 #include <linux/io.h>
 #include <linux/of.h>
 
-/* Clock Manager offsets */
-#define CLKMGR_CTRL    0x0
-#define CLKMGR_BYPASS 0x4
-
 /* Clock bypass bits */
 #define MAINPLL_BYPASS (1<<0)
 #define SDRAMPLL_BYPASS (1<<1)
@@ -37,12 +34,17 @@
 #define SOCFPGA_PLL_BG_PWRDWN  0x00000001
 #define SOCFPGA_PLL_EXT_ENA    0x00000002
 #define SOCFPGA_PLL_PWR_DOWN   0x00000004
-#define SOCFPGA_PLL_DIVF_MASK 0x0000FFF8
+#define SOCFPGA_PLL_DIVF_MASK  0x0000FFF8
 #define SOCFPGA_PLL_DIVF_SHIFT 3
-#define SOCFPGA_PLL_DIVQ_MASK 0x003F0000
+#define SOCFPGA_PLL_DIVQ_MASK  0x003F0000
 #define SOCFPGA_PLL_DIVQ_SHIFT 15
 
-extern void __iomem *clk_mgr_base_addr;
+struct clock_manager_hw {
+	u32 ctrl;
+	u32 bypass;
+};
+
+extern struct clock_manager_hw __iomem *clk_mgr_base_addr;
 
 struct socfpga_clk {
 	struct clk_hw hw;
@@ -83,7 +85,7 @@ static unsigned long clk_pll_recalc_rate(struct clk_hw *hwclk,
 	unsigned long bypass;
 
 	reg = readl(socfpgaclk->reg);
-	bypass = readl(clk_mgr_base_addr + CLKMGR_BYPASS);
+	bypass = readl(&clk_mgr_base_addr->bypass);
 	if (bypass & MAINPLL_BYPASS)
 		return parent_rate;
 
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#9 lore raw

PATCH v2 1/2 ARM: socfpga: Enable soft reset

In reply to dinguyen at altera.com

Hi!

Quoted 10 lines

From: Dinh Nguyen <[email protected]>

Enable a cold or warm reset to the HW from userspace.

Also fix a few sparse errors:

warning: symbol 'sys_manager_base_addr' was not declared. Should it be static?
warning: symbol 'rst_manager_base_addr' was not declared. Should it be static?

Would it make sense to apply something like this? Struct looks cleaner than offset defines...

Thanks,
Pavel

Switch reset manager to using struct (not defines), cleanups.

Convert SMP code to use the struct instead of open-coded numbers.

Also none of the code is time-critical, so it does not make sense to use __raw_writel variants.

arch/arm/mach-socfpga/core.h +16 -14
index d2a251f..f4b6048 100644
@@ -20,19 +20,21 @@
 #ifndef __MACH_CORE_H
 #define __MACH_CORE_H
 
-#define SOCFPGA_RSTMGR_CTRL	0x04
-#define SOCFPGA_RSTMGR_MODPERRST	0x14
-#define SOCFPGA_RSTMGR_BRGMODRST	0x1c
-
-/* System Manager bits */
-#define RSTMGR_CTRL_SWCOLDRSTREQ	0x1	/* Cold Reset */
-#define RSTMGR_CTRL_SWWARMRSTREQ	0x2	/* Warm Reset */
-/*MPU Module Reset Register */
-#define RSTMGR_MPUMODRST_CPU0	0x1	/*CPU0 Reset*/
-#define RSTMGR_MPUMODRST_CPU1	0x2	/*CPU1 Reset*/
-#define RSTMGR_MPUMODRST_WDS		0x4	/*Watchdog Reset*/
-#define RSTMGR_MPUMODRST_SCUPER	0x8	/*SCU and periphs reset*/
-#define RSTMGR_MPUMODRST_L2		0x10	/*L2 Cache reset*/
+struct socfpga_rstmgr_hw {
+	u32 unk;
+	u32 ctrl;		/* 0x04 */
+	u32 unk2, unk3;
+/* MPU Module Reset Register */
+#define RSTMGR_MPUMODRST_CPU0	0x1	/* CPU0 Reset */
+#define RSTMGR_MPUMODRST_CPU1	0x2	/* CPU1 Reset */
+#define RSTMGR_MPUMODRST_WDS	0x4	/* Watchdog Reset */
+#define RSTMGR_MPUMODRST_SCUPER	0x8	/* SCU and periphs reset */
+#define RSTMGR_MPUMODRST_L2	0x10	/* L2 Cache reset */
+	u32 mpumodrst; 		/* 0x10 */
+	u32 modperrst;		/* 0x14 */
+	u32 unk5;
+	u32 bgrmodrst;		/* 0x1c */
+};
 
 extern void socfpga_secondary_startup(void);
 extern void __iomem *socfpga_scu_base_addr;
@@ -41,7 +43,7 @@ extern void socfpga_init_clocks(void);
 extern void socfpga_sysmgr_init(void);
 
 extern void __iomem *sys_manager_base_addr;
-extern void __iomem *rst_manager_base_addr;
+extern struct socfpga_rstmgr_hw __iomem *rst_manager_base_addr;
 
 extern struct smp_operations socfpga_smp_ops;
 extern char secondary_trampoline, secondary_trampoline_end;
arch/arm/mach-socfpga/socfpga.c +5 -5
index b41a945..81b8f1e 100644
@@ -28,7 +28,7 @@
 
 void __iomem *socfpga_scu_base_addr = ((void __iomem *)(SOCFPGA_SCU_VIRT_BASE));
 void __iomem *sys_manager_base_addr;
-void __iomem *rst_manager_base_addr;
+struct socfpga_rstmgr_hw __iomem *rst_manager_base_addr;
 unsigned long cpu1start_addr;
 
 static struct map_desc scu_io_desc __initdata = {
@@ -89,13 +89,13 @@ static void socfpga_cyclone5_restart(char mode, const char *cmd)
 {
 	u32 temp;
 
-	temp = __raw_readl(rst_manager_base_addr + SOCFPGA_RSTMGR_CTRL);
+	temp = readl(&rst_manager_base_addr->ctrl);
 
 	if (mode == 'h')
-		temp |= RSTMGR_CTRL_SWCOLDRSTREQ;
+		temp |= 1; /* RSTMGR_CTRL_SWCOLDRSTREQ, cold reset */
 	else
-		temp |= RSTMGR_CTRL_SWWARMRSTREQ;
-	__raw_writel(temp, rst_manager_base_addr + SOCFPGA_RSTMGR_CTRL);
+		temp |= 2; /* RSTMGR_CTRL_SWWARMRSTREQ, warm reset */
+	writel(temp, &rst_manager_base_addr->ctrl);
 }
 
 static void __init socfpga_cyclone5_init(void)
arch/arm/mach-socfpga/platsmp.c +3 -3
index c75c33d..822a93e 100644
@@ -47,7 +47,7 @@ static int __cpuinit socfpga_boot_secondary(unsigned int cpu, struct task_struct
 	if (cpu1start_addr) {
 		memcpy(phys_to_virt(0), &secondary_trampoline, trampoline_size);
 
-		__raw_writel(virt_to_phys(socfpga_secondary_startup),
+		writel(virt_to_phys(socfpga_secondary_startup),
 			(sys_manager_base_addr + (cpu1start_addr & 0x000000ff)));
 
 		flush_cache_all();
@@ -55,7 +55,7 @@ static int __cpuinit socfpga_boot_secondary(unsigned int cpu, struct task_struct
 		outer_clean_range(0, trampoline_size);
 
 		/* This will release CPU #1 out of reset.*/
-		__raw_writel(0, rst_manager_base_addr + 0x10);
+		writel(0, &rst_manager_base_addr->mpumodrst);
 	}
 
 	return 0;
@@ -101,7 +101,7 @@ static void socfpga_cpu_die(unsigned int cpu)
 	flush_cache_all();
 
 	/* This will put CPU1 into reset.*/
-	__raw_writel(RSTMGR_MPUMODRST_CPU1, rst_manager_base_addr + 0x10);
+	writel(RSTMGR_MPUMODRST_CPU1, &rst_manager_base_addr->mpumodrst);
 
 	cpu_do_idle();
 
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#10 lore raw

PATCH v2 1/2 ARM: socfpga: Enable soft reset

Quoted 11 lines · Pavel Machek

Hi!

Quoted 9 lines

From: Dinh Nguyen <[email protected]>

Enable a cold or warm reset to the HW from userspace.

Also fix a few sparse errors:

warning: symbol 'sys_manager_base_addr' was not declared. Should it be static?
warning: symbol 'rst_manager_base_addr' was not declared. Should it be static?

Quoted 6 lines
Quoted 1 line

Would it make sense to apply something like this? Struct looks cleaner than offset defines...

Structs used to define offsets is technically dodgy according to the C standards, so I personally really do not like this practice, and I will not use it ever myself where offsets matter.

Your use is fine though, but only because we know how the compiler behaves on ARM - and that's a very important point with doing this. We're reliant on that compiler behaviour if we go down this path.

#11 lore raw

PATCH v2 1/2 ARM: socfpga: Enable soft reset

In reply to Pavel Machek
Quoted 56 lines · Pavel Machek

Hi!

Quoted 10 lines

From: Dinh Nguyen <[email protected]>

Enable a cold or warm reset to the HW from userspace.

Also fix a few sparse errors:

warning: symbol 'sys_manager_base_addr' was not declared. Should it be static?
warning: symbol 'rst_manager_base_addr' was not declared. Should it be static?

Would it make sense to apply something like this? Struct looks cleaner than offset defines...

Thanks,
Pavel

Switch reset manager to using struct (not defines), cleanups.

Convert SMP code to use the struct instead of open-coded numbers.

Also none of the code is time-critical, so it does not make sense to use __raw_writel variants.

arch/arm/mach-socfpga/core.h +5 -13
index d2a251f..f4b6048 100644
@@ -20,19 +20,21 @@
 #ifndef __MACH_CORE_H
 #define __MACH_CORE_H
 
-#define SOCFPGA_RSTMGR_CTRL	0x04
-#define SOCFPGA_RSTMGR_MODPERRST	0x14
-#define SOCFPGA_RSTMGR_BRGMODRST	0x1c
-
-/* System Manager bits */
-#define RSTMGR_CTRL_SWCOLDRSTREQ	0x1	/* Cold Reset */
-#define RSTMGR_CTRL_SWWARMRSTREQ	0x2	/* Warm Reset */
-/*MPU Module Reset Register */
-#define RSTMGR_MPUMODRST_CPU0	0x1	/*CPU0 Reset*/
-#define RSTMGR_MPUMODRST_CPU1	0x2	/*CPU1 Reset*/
-#define RSTMGR_MPUMODRST_WDS		0x4	/*Watchdog Reset*/
-#define RSTMGR_MPUMODRST_SCUPER	0x8	/*SCU and periphs reset*/
-#define RSTMGR_MPUMODRST_L2		0x10	/*L2 Cache reset*/
+struct socfpga_rstmgr_hw {
+	u32 unk;
+	u32 ctrl;		/* 0x04 */
+	u32 unk2, unk3;
+/* MPU Module Reset Register */

arch/arm/mach-socfpga/core.h line 37

Quoted 10 lines
arch/arm/mach-socfpga/core.h +10 -0
+#define RSTMGR_MPUMODRST_CPU0	0x1	/* CPU0 Reset */
+#define RSTMGR_MPUMODRST_CPU1	0x2	/* CPU1 Reset */
+#define RSTMGR_MPUMODRST_WDS	0x4	/* Watchdog Reset */
+#define RSTMGR_MPUMODRST_SCUPER	0x8	/* SCU and periphs reset */
+#define RSTMGR_MPUMODRST_L2	0x10	/* L2 Cache reset */
+	u32 mpumodrst; 		/* 0x10 */
+	u32 modperrst;		/* 0x14 */
+	u32 unk5;
+	u32 bgrmodrst;		/* 0x1c */
+};

As Russell already replied, struct used to represent register layout is normally frowned upon in drivers.

If you want to tighten up the code in this case, make a local helper function that just takes the register offset instead, i.e.:

arch/arm/mach-socfpga/core.h line 38

Quoted 2 lines
arch/arm/mach-socfpga/core.h +1 -1
-	temp = __raw_readl(rst_manager_base_addr + SOCFPGA_RSTMGR_CTRL);
+	temp = readl(&rst_manager_base_addr->ctrl);
	temp = rst_readl(SOCFPGA_RSTMGR_CTRL);

(and then have rst_readl/rst_writel do the math based on base address).

-Olof

#12 lore raw

PATCH v2 1/2 ARM: socfpga: Enable soft reset

Hi!

Quoted 10 lines
Quoted 10 lines
arch/arm/mach-socfpga/core.h +10 -0
+struct socfpga_rstmgr_hw {
+	u32 unk;
+	u32 ctrl;		/* 0x04 */
+	u32 unk2, unk3;
+/* MPU Module Reset Register */
+#define RSTMGR_MPUMODRST_CPU0	0x1	/* CPU0 Reset */
+#define RSTMGR_MPUMODRST_CPU1	0x2	/* CPU1 Reset */
+#define RSTMGR_MPUMODRST_WDS	0x4	/* Watchdog Reset */
+#define RSTMGR_MPUMODRST_SCUPER	0x8	/* SCU and periphs reset */
+#define RSTMGR_MPUMODRST_L2	0x10	/* L2 Cache reset */
Quoted 8 lines

arch/arm/mach-socfpga/core.h line 37

Quoted 5 lines
arch/arm/mach-socfpga/core.h +5 -0
+	u32 mpumodrst; 		/* 0x10 */
+	u32 modperrst;		/* 0x14 */
+	u32 unk5;
+	u32 bgrmodrst;		/* 0x1c */
+};

As Russell already replied, struct used to represent register layout is normally frowned upon in drivers.

Well... as Russell also said, this is arch-specific code, so it is actually ok.

Quoted 10 lines

If you want to tighten up the code in this case, make a local helper function that just takes the register offset instead, i.e.:

arch/arm/mach-socfpga/socfpga.c static void socfpga_cyclone5_restart(char mode, const char *cmd) · line 92

Quoted 2 lines
arch/arm/mach-socfpga/socfpga.c +1 -1
-	temp = __raw_readl(rst_manager_base_addr + SOCFPGA_RSTMGR_CTRL);
+	temp = readl(&rst_manager_base_addr->ctrl);
	temp = rst_readl(SOCFPGA_RSTMGR_CTRL);

(and then have rst_readl/rst_writel do the math based on base address).

That does not prevent mistake such as rst_readl(SOCFPGA_SYSMGR_CTRL)... and with number of different subsystems on socfpga (each with separate base address), I fear that might be an issue.

Unfortunately, C does not check type of enums, so they can't be used to solve that.

Any other ideas?
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#13 lore raw

PATCH v2 1/2 ARM: socfpga: Enable soft reset

Quoted 21 lines · Pavel Machek

Hi!

Quoted 10 lines
Quoted 10 lines
arch/arm/mach-socfpga/core.h +10 -0
+struct socfpga_rstmgr_hw {
+   u32 unk;
+   u32 ctrl;               /* 0x04 */
+   u32 unk2, unk3;
+/* MPU Module Reset Register */
+#define RSTMGR_MPUMODRST_CPU0      0x1     /* CPU0 Reset */
+#define RSTMGR_MPUMODRST_CPU1      0x2     /* CPU1 Reset */
+#define RSTMGR_MPUMODRST_WDS       0x4     /* Watchdog Reset */
+#define RSTMGR_MPUMODRST_SCUPER    0x8     /* SCU and periphs reset */
+#define RSTMGR_MPUMODRST_L2        0x10    /* L2 Cache reset */
Quoted 8 lines

arch/arm/mach-socfpga/core.h line 37

Quoted 5 lines
arch/arm/mach-socfpga/core.h +5 -0
+   u32 mpumodrst;          /* 0x10 */
+   u32 modperrst;          /* 0x14 */
+   u32 unk5;
+   u32 bgrmodrst;          /* 0x1c */
+};

As Russell already replied, struct used to represent register layout is normally frowned upon in drivers.

Quoted 2 lines

Well... as Russell also said, this is arch-specific code, so it is actually ok.

Which is why I said "normally frowned upon" instead of "always frowned upon".

Quoted 11 lines
Quoted 10 lines

If you want to tighten up the code in this case, make a local helper function that just takes the register offset instead, i.e.:

arch/arm/mach-socfpga/socfpga.c static void socfpga_cyclone5_restart(char mode, const char *cmd) · line 92

Quoted 2 lines
arch/arm/mach-socfpga/socfpga.c +1 -1
-   temp = __raw_readl(rst_manager_base_addr + SOCFPGA_RSTMGR_CTRL);
+   temp = readl(&rst_manager_base_addr->ctrl);
      temp = rst_readl(SOCFPGA_RSTMGR_CTRL);

(and then have rst_readl/rst_writel do the math based on base address).

Quoted 7 lines

That does not prevent mistake such as rst_readl(SOCFPGA_SYSMGR_CTRL)... and with number of different subsystems on socfpga (each with separate base address), I fear that might be an issue.

Unfortunately, C does not check type of enums, so they can't be used to solve that.

That's no different from accidentally doing
readl(&sys_manager_base_addr->ctrl) instead of
rst_manager_base_addr->ctrl.

-Olof

#14 lore raw

PATCH v2 1/2 ARM: socfpga: Enable soft reset

Hi Pavel,

Quoted 43 lines · Olof Johansson
Quoted 21 lines · Pavel Machek

Hi!

Quoted 10 lines
Quoted 10 lines
arch/arm/mach-socfpga/core.h +10 -0
+struct socfpga_rstmgr_hw {
+   u32 unk;
+   u32 ctrl;               /* 0x04 */
+   u32 unk2, unk3;
+/* MPU Module Reset Register */
+#define RSTMGR_MPUMODRST_CPU0      0x1     /* CPU0 Reset */
+#define RSTMGR_MPUMODRST_CPU1      0x2     /* CPU1 Reset */
+#define RSTMGR_MPUMODRST_WDS       0x4     /* Watchdog Reset */
+#define RSTMGR_MPUMODRST_SCUPER    0x8     /* SCU and periphs reset */
+#define RSTMGR_MPUMODRST_L2        0x10    /* L2 Cache reset */
Quoted 8 lines

arch/arm/mach-socfpga/core.h line 37

Quoted 5 lines
arch/arm/mach-socfpga/core.h +5 -0
+   u32 mpumodrst;          /* 0x10 */
+   u32 modperrst;          /* 0x14 */
+   u32 unk5;
+   u32 bgrmodrst;          /* 0x1c */
+};

As Russell already replied, struct used to represent register layout is normally frowned upon in drivers.

Quoted 2 lines

Well... as Russell also said, this is arch-specific code, so it is actually ok.

Which is why I said "normally frowned upon" instead of "always frowned upon".

Quoted 16 lines
Quoted 10 lines

If you want to tighten up the code in this case, make a local helper function that just takes the register offset instead, i.e.:

arch/arm/mach-socfpga/socfpga.c static void socfpga_cyclone5_restart(char mode, const char *cmd) · line 92

Quoted 2 lines
arch/arm/mach-socfpga/socfpga.c +1 -1
-   temp = __raw_readl(rst_manager_base_addr + SOCFPGA_RSTMGR_CTRL);
+   temp = readl(&rst_manager_base_addr->ctrl);
      temp = rst_readl(SOCFPGA_RSTMGR_CTRL);

(and then have rst_readl/rst_writel do the math based on base address).

That does not prevent mistake such as rst_readl(SOCFPGA_SYSMGR_CTRL)... and with number of different subsystems on socfpga (each with separate base address), I fear that might be an issue.

Quoted 7 lines
Quoted 2 lines

Unfortunately, C does not check type of enums, so they can't be used to solve that.

That's no different from accidentally doing
readl(&sys_manager_base_addr->ctrl) instead of
rst_manager_base_addr->ctrl.

I also don't really like to use structs for register offsets. So unless, there's some standard to use them in machine code, I'm not going to use them in mach-socfpga.

I apologize for not getting to a v3 for this patch series with Mike's recommended change for using composite clocks as quickly as I would have like. I hope to have a v3 by next week.

Dinh

Quoted 3 lines

-Olof