On Tue, Dec 1, 2015 at 10:19 PM, Chen-Yu Tsai [off-list ref] wrote:

On Tue, Dec 1, 2015 at 8:50 PM, Maxime Ripard
[off-list ref] wrote:

quoted

On Tue, Dec 01, 2015 at 07:54:06PM +0800, Chen-Yu Tsai wrote:

quoted

On Tue, Dec 1, 2015 at 6:04 PM, Maxime Ripard
[off-list ref] wrote:

quoted

Hi,

On Sun, Nov 29, 2015 at 11:03:06AM +0800, Chen-Yu Tsai wrote:

quoted

The APBS clock on sun9i is the same as the APB0 clock on sun8i. With
sun9i we are supporting the PRCM clocks by using CLK_OF_DECLARE,
instead of through a PRCM mfd device and subdevices for each clock
and reset control. As such we need a CLK_OF_DECLARE version of
the sun8i-a23-apb0-clk driver.

Also, build it for sun9i/A80, and not just for configurations with
MFD_SUN6I_PRCM enabled.

Signed-off-by: Chen-Yu Tsai <redacted>
---
 drivers/clk/sunxi/Makefile         |  5 +--
 drivers/clk/sunxi/clk-sun8i-apb0.c | 71 +++++++++++++++++++++++++++++++-------
 2 files changed, 62 insertions(+), 14 deletions(-)

diff --git a/drivers/clk/sunxi/Makefile b/drivers/clk/sunxi/Makefile
index cb4c299214ce..c55d5cd1c0e5 100644
--- a/drivers/clk/sunxi/Makefile
+++ b/drivers/clk/sunxi/Makefile

@@ -15,6 +15,7 @@ obj-y += clk-sun9i-core.o
 obj-y += clk-sun9i-mmc.o
 obj-y += clk-usb.o

+obj-$(CONFIG_MACH_SUN9I) += clk-sun8i-apb0.o
+

So sun8i doesn't use it?

Shit... I messed up. clk-sun8i-apb0.o should also be under
CONFIG_MFD_SUN6I_PRCM.

I'll send a new version of this patch. No need to keep spamming
people with the whole series.

Ok.

quoted

quoted

quoted

 obj-$(CONFIG_MFD_SUN6I_PRCM) += \
-     clk-sun6i-ar100.o clk-sun6i-apb0.o clk-sun6i-apb0-gates.o \
-     clk-sun8i-apb0.o
+     clk-sun6i-ar100.o clk-sun6i-apb0.o clk-sun6i-apb0-gates.o

diff --git a/drivers/clk/sunxi/clk-sun8i-apb0.c b/drivers/clk/sunxi/clk-sun8i-apb0.c
index 7ae5d2c2cde1..c1e2ac8f4b0d 100644
--- a/drivers/clk/sunxi/clk-sun8i-apb0.c
+++ b/drivers/clk/sunxi/clk-sun8i-apb0.c

@@ -17,13 +17,68 @@
 #include <linux/clk-provider.h>
 #include <linux/module.h>
 #include <linux/of.h>
+#include <linux/of_address.h>
 #include <linux/platform_device.h>

+static struct clk *sun8i_a23_apb0_register(struct device_node *node,
+                                        void __iomem *reg)
+{
+     const char *clk_name = node->name;
+     const char *clk_parent;
+     struct clk *clk;
+     int ret;
+
+     clk_parent = of_clk_get_parent_name(node, 0);
+     if (!clk_parent)
+             return ERR_PTR(-EINVAL);
+
+     of_property_read_string(node, "clock-output-names", &clk_name);
+
+     /* The A23 APB0 clock is a standard 2 bit wide divider clock */
+     clk = clk_register_divider(NULL, clk_name, clk_parent, 0, reg,
+                                0, 2, CLK_DIVIDER_POWER_OF_TWO, NULL);
+     if (IS_ERR(clk))
+             return clk;
+
+     ret = of_clk_add_provider(node, of_clk_src_simple_get, clk);
+     if (ret)
+             goto err_unregister;
+
+     return clk;
+
+err_unregister:
+     clk_unregister_divider(clk);
+
+     return ERR_PTR(ret);
+}
+
+static void sun8i_a23_apb0_setup(struct device_node *node)
+{
+     void __iomem *reg;
+     struct resource res;
+     struct clk *clk;
+
+     reg = of_io_request_and_map(node, 0, of_node_full_name(node));
+     if (IS_ERR(reg))
+             return;
+
+     clk = sun8i_a23_apb0_register(node, reg);
+     if (IS_ERR(clk))
+             goto err_unmap;
+
+     return;
+
+err_unmap:
+     iounmap(reg);
+     of_address_to_resource(node, 0, &res);
+     release_mem_region(res.start, resource_size(&res));
+}
+CLK_OF_DECLARE(sun8i_a23_apb0, "allwinner,sun8i-a23-apb0-clk",
+            sun8i_a23_apb0_setup);
+
 static int sun8i_a23_apb0_clk_probe(struct platform_device *pdev)
 {
      struct device_node *np = pdev->dev.of_node;
-     const char *clk_name = np->name;
-     const char *clk_parent;
      struct resource *r;
      void __iomem *reg;
      struct clk *clk;

@@ -33,19 +88,11 @@ static int sun8i_a23_apb0_clk_probe(struct platform_device *pdev)
      if (IS_ERR(reg))
              return PTR_ERR(reg);

-     clk_parent = of_clk_get_parent_name(np, 0);
-     if (!clk_parent)
-             return -EINVAL;
-
-     of_property_read_string(np, "clock-output-names", &clk_name);
-
-     /* The A23 APB0 clock is a standard 2 bit wide divider clock */
-     clk = clk_register_divider(&pdev->dev, clk_name, clk_parent, 0, reg,
-                                0, 2, CLK_DIVIDER_POWER_OF_TWO, NULL);
+     clk = sun8i_a23_apb0_register(np, reg);
      if (IS_ERR(clk))
              return PTR_ERR(clk);

-     return of_clk_add_provider(np, of_clk_src_simple_get, clk);
+     return 0;
 }

Won't this probe twice now? First the CLK_OF_DECLARE will register a
clock, and then the device model will call probe a second time.

AFAIK it will.

quoted

I guess then request_mem_region will catch it, but then you return an
error code, which is probably an error success, since the clock is
registered :)

Indeed. clk-mod0.c actually has a comment block explaining this.
I didn't want to repeat the whole thing though.

I guess, we can simply return 0 if of_io_request_and_map returns
EBUSY, since it would simply mean that we already probed.

I think CLK_OF_DECLARE gets probed first, then platform devices.
of_io_request_and_map fails for mfd cells, which have no DT reg
properties. So we should be ignoring -EINVAL instead of -EBUSY.

Right... I misread your comment. You're talking about the platform device
part. About ignoring the error here: normally we use CLK_OF_DECLARE with
clocks under the "clocks" node, which don't get registered as platform
device. (Or is this not so common?) So if one does get probed twice, I
think we shouldn't ignore the error. It's likely the DT has a problem
and the probe error will let the user know.

What do you think?

ChenYu

But yeah, I'll have it return 0 in this case, and add a brief
comment.