[PATCH v2 0/2] MIPS: TXx9: Common Clock Framework Conversion

[PATCH v2 0/2] MIPS: TXx9: Common Clock Framework Conversion

[Geert Uytterhoeven]

	Hi Ralf, Nemoto-san, Wim,

This patch series converts the Toshiba TXx9 platforms from their own
custom minimal clock implementation to the Common Clock Framework.

  - Patch 1 adds missing clock (un)prepare calls to the TXx9 watchdog
    driver,
  - Patch 2 replaces the custom clock implementation by a CCF-based one,
    providing a minimal set of clocks.

This has been tested with the watchdog on RBTX4927.

Changes since v1:
  - Dropped "spi: spi-txx9: Add missing clock (un)prepare calls for
    CCF", which has been accepted in spi/for-next,
  - Protect the TX4938_REV_PCODE() check by #ifdef CONFIG_CPU_TX49XX,
  - Use new clk_hw-centric clock registration API,
  - Add Reviewed-by.

Dependencies:
  - Patch 1 can be applied independently,
  - Patch 2 depends on patch 1 and on "spi: spi-txx9: Add missing clock
    (un)prepare calls for CCF" in spi/for-next,
  - The error path in patch 2 depends on "clkdev: Detect errors in
    clk_hw_register_clkdev() for mass registration", but this is less
    critical.

Wim: Can you please appply patch 1 directly, so we can get at least the
hard dependencies in v4.9?

Thanks!

Geert Uytterhoeven (2):
  watchdog: txx9wdt: Add missing clock (un)prepare calls for CCF
  MIPS: TXx9: Convert to Common Clock Framework

 arch/mips/txx9/Kconfig         |  2 +-
 arch/mips/txx9/generic/setup.c | 70 +++++++++++++++++++++---------------------
 drivers/watchdog/txx9wdt.c     |  6 ++--
 3 files changed, 39 insertions(+), 39 deletions(-)

-- 
1.9.1

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds

[PATCH v2 1/2] watchdog: txx9wdt: Add missing clock (un)prepare calls for CCF

[Geert Uytterhoeven]

While the custom minimal TXx9 clock implementation doesn't need or use
clock (un)prepare calls (they are dummies if !CONFIG_HAVE_CLK_PREPARE),
they are mandatory when using the Common Clock Framework.

Hence add them, to prepare for the advent of CCF.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
---
Tested on RBTX4927.

v2:
  - Add Reviewed-by.
---
 drivers/watchdog/txx9wdt.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/watchdog/txx9wdt.c b/drivers/watchdog/txx9wdt.c
index c2da880292bc2f32..6f7a9deb27d05d25 100644
--- a/drivers/watchdog/txx9wdt.c
+++ b/drivers/watchdog/txx9wdt.c
@@ -112,7 +112,7 @@ static int __init txx9wdt_probe(struct platform_device *dev)
 		txx9_imclk = NULL;
 		goto exit;
 	}
-	ret = clk_enable(txx9_imclk);
+	ret = clk_prepare_enable(txx9_imclk);
 	if (ret) {
 		clk_put(txx9_imclk);
 		txx9_imclk = NULL;
@@ -144,7 +144,7 @@ static int __init txx9wdt_probe(struct platform_device *dev)
 	return 0;
 exit:
 	if (txx9_imclk) {
-		clk_disable(txx9_imclk);
+		clk_disable_unprepare(txx9_imclk);
 		clk_put(txx9_imclk);
 	}
 	return ret;
@@ -153,7 +153,7 @@ exit:
 static int __exit txx9wdt_remove(struct platform_device *dev)
 {
 	watchdog_unregister_device(&txx9wdt);
-	clk_disable(txx9_imclk);
+	clk_disable_unprepare(txx9_imclk);
 	clk_put(txx9_imclk);
 	return 0;
 }
-- 
1.9.1

Re: [PATCH v2 1/2] watchdog: txx9wdt: Add missing clock (un)prepare calls for CCF

[Wim Van Sebroeck]

Hi Geert,

> While the custom minimal TXx9 clock implementation doesn't need or use
> clock (un)prepare calls (they are dummies if !CONFIG_HAVE_CLK_PREPARE),
> they are mandatory when using the Common Clock Framework.
> 
> Hence add them, to prepare for the advent of CCF.
> 
> Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
> Reviewed-by: Guenter Roeck <linux@roeck-us.net>

This patch was added to linux-watchdog-next almost 2 weeks ago.

Kind regards,
Wim.

[PATCH v2 2/2] MIPS: TXx9: Convert to Common Clock Framework

[Geert Uytterhoeven]

Replace the custom minimal clock implementation for Toshiba TXx9 by a
basic implementation using the Common Clock Framework.

The only clocks that are provided are those needed by TXx9-specific
drivers ("imbus" and "spi" (TX4938 only)), and their common parent
clock "gbus". Other clocks can be added when needed.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
---
Tested on RBTX4927.

v2:
  - Protect the TX4938_REV_PCODE() check by #ifdef CONFIG_CPU_TX49XX,
  - Use new clk_hw-centric clock registration API.
---
 arch/mips/txx9/Kconfig         |  2 +-
 arch/mips/txx9/generic/setup.c | 70 +++++++++++++++++++++---------------------
 2 files changed, 36 insertions(+), 36 deletions(-)

diff --git a/arch/mips/txx9/Kconfig b/arch/mips/txx9/Kconfig
index 8c337d60f790db9f..42923478d45ca363 100644
--- a/arch/mips/txx9/Kconfig
+++ b/arch/mips/txx9/Kconfig
@@ -20,7 +20,7 @@ config MACH_TXX9
 	select SYS_SUPPORTS_32BIT_KERNEL
 	select SYS_SUPPORTS_LITTLE_ENDIAN
 	select SYS_SUPPORTS_BIG_ENDIAN
-	select HAVE_CLK
+	select COMMON_CLK
 
 config TOSHIBA_JMR3927
 	bool "Toshiba JMR-TX3927 board"
diff --git a/arch/mips/txx9/generic/setup.c b/arch/mips/txx9/generic/setup.c
index ada92db92f87d91a..a1d98b5c8fd67576 100644
--- a/arch/mips/txx9/generic/setup.c
+++ b/arch/mips/txx9/generic/setup.c
@@ -15,7 +15,8 @@
 #include <linux/interrupt.h>
 #include <linux/string.h>
 #include <linux/module.h>
-#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#include <linux/clkdev.h>
 #include <linux/err.h>
 #include <linux/gpio/driver.h>
 #include <linux/platform_device.h>
@@ -83,40 +84,6 @@ int txx9_ccfg_toeon __initdata;
 int txx9_ccfg_toeon __initdata = 1;
 #endif
 
-/* Minimum CLK support */
-
-struct clk *clk_get(struct device *dev, const char *id)
-{
-	if (!strcmp(id, "spi-baseclk"))
-		return (struct clk *)((unsigned long)txx9_gbus_clock / 2 / 2);
-	if (!strcmp(id, "imbus_clk"))
-		return (struct clk *)((unsigned long)txx9_gbus_clock / 2);
-	return ERR_PTR(-ENOENT);
-}
-EXPORT_SYMBOL(clk_get);
-
-int clk_enable(struct clk *clk)
-{
-	return 0;
-}
-EXPORT_SYMBOL(clk_enable);
-
-void clk_disable(struct clk *clk)
-{
-}
-EXPORT_SYMBOL(clk_disable);
-
-unsigned long clk_get_rate(struct clk *clk)
-{
-	return (unsigned long)clk;
-}
-EXPORT_SYMBOL(clk_get_rate);
-
-void clk_put(struct clk *clk)
-{
-}
-EXPORT_SYMBOL(clk_put);
-
 #define BOARD_VEC(board)	extern struct txx9_board_vec board;
 #include <asm/txx9/boards.h>
 #undef BOARD_VEC
@@ -560,8 +527,41 @@ void __init plat_time_init(void)
 	txx9_board_vec->time_init();
 }
 
+static void txx9_clk_init(void)
+{
+	struct clk_hw *hw;
+	int error;
+
+	hw = clk_hw_register_fixed_rate(NULL, "gbus", NULL, 0, txx9_gbus_clock);
+	if (IS_ERR(hw)) {
+		error = PTR_ERR(hw);
+		goto fail;
+	}
+
+	hw = clk_hw_register_fixed_factor(NULL, "imbus", "gbus", 0, 1, 2);
+	error = clk_hw_register_clkdev(hw, "imbus_clk", NULL);
+	if (error)
+		goto fail;
+
+#ifdef CONFIG_CPU_TX49XX
+	if (TX4938_REV_PCODE() == 0x4938) {
+		hw = clk_hw_register_fixed_factor(NULL, "spi", "gbus", 0, 1, 4);
+		error = clk_hw_register_clkdev(hw, "spi-baseclk", NULL);
+		if (error)
+			goto fail;
+	}
+#endif
+
+	return;
+
+fail:
+	pr_err("Failed to register clocks: %d\n", error);
+}
+
 static int __init _txx9_arch_init(void)
 {
+	txx9_clk_init();
+
 	if (txx9_board_vec->arch_init)
 		txx9_board_vec->arch_init();
 	return 0;
-- 
1.9.1

Re: [PATCH v2 2/2] MIPS: TXx9: Convert to Common Clock Framework

[Stephen Boyd]

On 09/11, Geert Uytterhoeven wrote:
> Replace the custom minimal clock implementation for Toshiba TXx9 by a
> basic implementation using the Common Clock Framework.
> 
> The only clocks that are provided are those needed by TXx9-specific
> drivers ("imbus" and "spi" (TX4938 only)), and their common parent
> clock "gbus". Other clocks can be added when needed.
> 
> Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
> ---

Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>

-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project

Re: [PATCH v2 2/2] MIPS: TXx9: Convert to Common Clock Framework

[Atsushi Nemoto]

On Sun, 11 Sep 2016 10:59:58 +0200, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> Replace the custom minimal clock implementation for Toshiba TXx9 by a
> basic implementation using the Common Clock Framework.
> 
> The only clocks that are provided are those needed by TXx9-specific
> drivers ("imbus" and "spi" (TX4938 only)), and their common parent
> clock "gbus". Other clocks can be added when needed.
> 
> Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
> ---
> Tested on RBTX4927.
> 
> v2:
>   - Protect the TX4938_REV_PCODE() check by #ifdef CONFIG_CPU_TX49XX,
>   - Use new clk_hw-centric clock registration API.

Thank you for updated patch.

Reviewed-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>

---
Atsushi Nemoto

Re: [PATCH v2 0/2] MIPS: TXx9: Common Clock Framework Conversion

[Guenter Roeck]

On 09/11/2016 01:59 AM, Geert Uytterhoeven wrote:
> 	Hi Ralf, Nemoto-san, Wim,
>
> This patch series converts the Toshiba TXx9 platforms from their own
> custom minimal clock implementation to the Common Clock Framework.
>
>   - Patch 1 adds missing clock (un)prepare calls to the TXx9 watchdog
>     driver,
>   - Patch 2 replaces the custom clock implementation by a CCF-based one,
>     providing a minimal set of clocks.
>
> This has been tested with the watchdog on RBTX4927.
>
> Changes since v1:
>   - Dropped "spi: spi-txx9: Add missing clock (un)prepare calls for
>     CCF", which has been accepted in spi/for-next,
>   - Protect the TX4938_REV_PCODE() check by #ifdef CONFIG_CPU_TX49XX,
>   - Use new clk_hw-centric clock registration API,
>   - Add Reviewed-by.
>
> Dependencies:
>   - Patch 1 can be applied independently,
>   - Patch 2 depends on patch 1 and on "spi: spi-txx9: Add missing clock
>     (un)prepare calls for CCF" in spi/for-next,
>   - The error path in patch 2 depends on "clkdev: Detect errors in
>     clk_hw_register_clkdev() for mass registration", but this is less
>     critical.
>
> Wim: Can you please appply patch 1 directly, so we can get at least the
> hard dependencies in v4.9?
>

It is in my watchdog-next branch, and it has my Reviewed-by:, which
normally means that Wim will apply it.

Thanks,
Guenter