[U-Boot] [PATCH 1/3] ARM: socfpga: clk: Obtain handoff base clock via DM

[Ley Foon Tan]

On Thu, 2018-08-09 at 10:37 +0200, Marek Vasut wrote:
> On 08/09/2018 10:23 AM, Ley Foon Tan wrote:
> > 
> > On Wed, 2018-08-08 at 22:11 +0200, Marek Vasut wrote:
> > > 
> > > Bind fixed clock driver to the base clock instantiated in the
> > > handoff
> > > DT and use DM clock framework to get their clock rate. This
> > > replaces
> > > the ad-hoc DT parsing present thus far.
> > > 
> > > Signed-off-by: Marek Vasut <marex at denx.de>
> > > Cc: Chin Liang See <chin.liang.see at intel.com>
> > > Cc: Dinh Nguyen <dinguyen at kernel.org>
> > > Cc: Ley Foon Tan <ley.foon.tan at intel.com>
> > > ---
> > >  arch/arm/mach-socfpga/Kconfig                 |  2 ++
> > >  arch/arm/mach-socfpga/clock_manager_arria10.c | 41
> > > +++++++++++++++++++--------
> > >  2 files changed, 31 insertions(+), 12 deletions(-)
> > > 
> > > diff --git a/arch/arm/mach-socfpga/Kconfig b/arch/arm/mach-
> > > socfpga/Kconfig
> > > index 5c1df2cf1f..2655289a72 100644
> > > --- a/arch/arm/mach-socfpga/Kconfig
> > > +++ b/arch/arm/mach-socfpga/Kconfig
> > > @@ -11,6 +11,8 @@ config TARGET_SOCFPGA_ARRIA10
> > >  	bool
> > >  	select ALTERA_SDRAM
> > >  	select SPL_BOARD_INIT if SPL
> > > +	select CLK
> > > +	select SPL_CLK if SPL
> > >  
> > >  config TARGET_SOCFPGA_CYCLONE5
> > >  	bool
> > > diff --git a/arch/arm/mach-socfpga/clock_manager_arria10.c
> > > b/arch/arm/mach-socfpga/clock_manager_arria10.c
> > > index defa2f6261..59d3f9d50a 100644
> > > --- a/arch/arm/mach-socfpga/clock_manager_arria10.c
> > > +++ b/arch/arm/mach-socfpga/clock_manager_arria10.c
> > > @@ -7,6 +7,8 @@
> > >  #include <fdtdec.h>
> > >  #include <asm/io.h>
> > >  #include <dm.h>
> > > +#include <clk.h>
> > > +#include <dm/device-internal.h>
> > >  #include <asm/arch/clock_manager.h>
> > >  
> > >  static const struct socfpga_clock_manager *clock_manager_base =
> > > @@ -141,9 +143,9 @@ struct strtopu32 {
> > >  };
> > >  
> > >  const struct strtopu32 dt_to_val[] = {
> > > -	{ "/clocks/altera_arria10_hps_eosc1", &eosc1_hz},
> > > -	{ "/clocks/altera_arria10_hps_cb_intosc_ls",
> > > &cb_intosc_hz},
> > > -	{ "/clocks/altera_arria10_hps_f2h_free", &f2s_free_hz},
> > > +	{ "altera_arria10_hps_eosc1", &eosc1_hz },
> > > +	{ "altera_arria10_hps_cb_intosc_ls", &cb_intosc_hz },
> > > +	{ "altera_arria10_hps_f2h_free", &f2s_free_hz },
> > >  };
> > >  
> > >  static int of_to_struct(const void *blob, int node, const struct
> > > strtou32 *cfg_tab,
> > > @@ -163,28 +165,43 @@ static int of_to_struct(const void *blob,
> > > int
> > > node, const struct strtou32 *cfg_t
> > >  	return 0;
> > >  }
> > >  
> > > -static void of_get_input_clks(const void *blob)
> > > +static int of_get_input_clks(const void *blob)
> > >  {
> > > -	int node, i;
> > > +	struct udevice *dev;
> > > +	struct clk clk;
> > > +	int i, ret;
> > >  
> > >  	for (i = 0; i < ARRAY_SIZE(dt_to_val); i++) {
> > > -		node = fdt_path_offset(blob, dt_to_val[i].str);
> > > +		memset(&clk, 0, sizeof(clk));
> > >  
> > > -		if (node < 0)
> > > -			continue;
> > > +		ret = uclass_get_device_by_name(UCLASS_CLK,
> > > dt_to_val[i].str,
> > > +						&dev);
> > > +		if (ret)
> > > +			return ret;
> > >  
> > > -		fdtdec_get_int_array(blob, node, "clock-
> > > frequency",
> > > -				     dt_to_val[i].p, 1);
> > > +		ret = device_probe(dev);
> > uclass_get_device_by_name() will call to device_probe() function.
> > So,
> > can remove this.
> Are you sure about that ? Why so ?

It will call to device_probe() function eventually. See code in http://
git.denx.de/?p=u-boot.git;a=blob;f=drivers/core/uclass.c

The sequence:
uclass_get_device_by_name() --> uclass_get_device_tail() --
> device_probe()

I used uclass_get_device() function before, it has same implementation
as uclass_get_device_by_name(). We don't need call to device_probe()
manually.


Regards
Ley Foon