Re: [RFC PATCH] net: stmmac: Add OXNAS Glue Driver

[Neil Armstrong]

On 10/20/2016 06:26 PM, Joachim Eastwood wrote:
> Hi Neil,
> 
> On 20 October 2016 at 17:54, Neil Armstrong  wrote:
>> Add Synopsys Designware MAC Glue layer for the Oxford Semiconductor OX820.
>>
>> Signed-off-by: Neil Armstrong 
>> ---
>>  .../devicetree/bindings/net/oxnas-dwmac.txt|  44 ++
>>  drivers/net/ethernet/stmicro/stmmac/Kconfig|  11 ++
>>  drivers/net/ethernet/stmicro/stmmac/Makefile   |   1 +
>>  drivers/net/ethernet/stmicro/stmmac/dwmac-oxnas.c  | 173 
>> +
>>  4 files changed, 229 insertions(+)
>>  create mode 100644 Documentation/devicetree/bindings/net/oxnas-dwmac.txt
>>  create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-oxnas.c
> 
>> +
>> +static int oxnas_dwmac_probe(struct platform_device *pdev)
>> +{
>> +   struct plat_stmmacenet_data *plat_dat;
>> +   struct stmmac_resources stmmac_res;
>> +   struct device_node *sysctrl;
>> +   struct oxnas_dwmac *dwmac;
>> +   int ret;
>> +
>> +   sysctrl = of_parse_phandle(pdev->dev.of_node, "oxsemi,sys-ctrl", 0);
>> +   if (!sysctrl) {
>> +   dev_err(>dev, "failed to get sys-ctrl node\n");
>> +   return -EINVAL;
>> +   }
>> +
>> +   ret = stmmac_get_platform_resources(pdev, _res);
>> +   if (ret)
>> +   return ret;
>> +
>> +   plat_dat = stmmac_probe_config_dt(pdev, _res.mac);
>> +   if (IS_ERR(plat_dat))
>> +   return PTR_ERR(plat_dat);
>> +
>> +   dwmac = devm_kzalloc(>dev, sizeof(*dwmac), GFP_KERNEL);
>> +   if (!dwmac)
>> +   return -ENOMEM;
>> +
>> +   dwmac->regmap = syscon_node_to_regmap(sysctrl);
>> +   if (IS_ERR(dwmac->regmap)) {
>> +   dev_err(>dev, "failed to have sysctrl regmap\n");
>> +   return PTR_ERR(dwmac->regmap);
>> +   }
>> +
>> +   dwmac->clk = devm_clk_get(>dev, "gmac");
>> +   if (IS_ERR(dwmac->clk))
>> +   return PTR_ERR(dwmac->clk);
>> +
>> +   plat_dat->bsp_priv = dwmac;
>> +   plat_dat->init = oxnas_dwmac_init;
>> +   plat_dat->exit = oxnas_dwmac_exit;
> 
> Please do not use the init/exit callbacks. Implement proper driver
> callbacks instead. I.e: PM resume/suspend and driver remove.
> 
> Shouldn't you call oxnas_dwmac_init() from probe as well?
> As it is now it will only be called during PM resume and that can't be right.
> 
> 
>> +
>> +   return stmmac_dvr_probe(>dev, plat_dat, _res);
> 
> If stmmac_dvr_probe() fails you should disable your clocks.
> 
> 
> regards,
> Joachim Eastwood
> 

Hi Joachim,

Thanks for the hints, stm32 glue does that and I will sync on it.

Neil

Re: [RFC PATCH] net: stmmac: Add OXNAS Glue Driver

[Neil Armstrong]

On 10/20/2016 06:26 PM, Joachim Eastwood wrote:
> Hi Neil,
> 
> On 20 October 2016 at 17:54, Neil Armstrong  wrote:
>> Add Synopsys Designware MAC Glue layer for the Oxford Semiconductor OX820.
>>
>> Signed-off-by: Neil Armstrong 
>> ---
>>  .../devicetree/bindings/net/oxnas-dwmac.txt|  44 ++
>>  drivers/net/ethernet/stmicro/stmmac/Kconfig|  11 ++
>>  drivers/net/ethernet/stmicro/stmmac/Makefile   |   1 +
>>  drivers/net/ethernet/stmicro/stmmac/dwmac-oxnas.c  | 173 
>> +
>>  4 files changed, 229 insertions(+)
>>  create mode 100644 Documentation/devicetree/bindings/net/oxnas-dwmac.txt
>>  create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-oxnas.c
> 
>> +
>> +static int oxnas_dwmac_probe(struct platform_device *pdev)
>> +{
>> +   struct plat_stmmacenet_data *plat_dat;
>> +   struct stmmac_resources stmmac_res;
>> +   struct device_node *sysctrl;
>> +   struct oxnas_dwmac *dwmac;
>> +   int ret;
>> +
>> +   sysctrl = of_parse_phandle(pdev->dev.of_node, "oxsemi,sys-ctrl", 0);
>> +   if (!sysctrl) {
>> +   dev_err(>dev, "failed to get sys-ctrl node\n");
>> +   return -EINVAL;
>> +   }
>> +
>> +   ret = stmmac_get_platform_resources(pdev, _res);
>> +   if (ret)
>> +   return ret;
>> +
>> +   plat_dat = stmmac_probe_config_dt(pdev, _res.mac);
>> +   if (IS_ERR(plat_dat))
>> +   return PTR_ERR(plat_dat);
>> +
>> +   dwmac = devm_kzalloc(>dev, sizeof(*dwmac), GFP_KERNEL);
>> +   if (!dwmac)
>> +   return -ENOMEM;
>> +
>> +   dwmac->regmap = syscon_node_to_regmap(sysctrl);
>> +   if (IS_ERR(dwmac->regmap)) {
>> +   dev_err(>dev, "failed to have sysctrl regmap\n");
>> +   return PTR_ERR(dwmac->regmap);
>> +   }
>> +
>> +   dwmac->clk = devm_clk_get(>dev, "gmac");
>> +   if (IS_ERR(dwmac->clk))
>> +   return PTR_ERR(dwmac->clk);
>> +
>> +   plat_dat->bsp_priv = dwmac;
>> +   plat_dat->init = oxnas_dwmac_init;
>> +   plat_dat->exit = oxnas_dwmac_exit;
> 
> Please do not use the init/exit callbacks. Implement proper driver
> callbacks instead. I.e: PM resume/suspend and driver remove.
> 
> Shouldn't you call oxnas_dwmac_init() from probe as well?
> As it is now it will only be called during PM resume and that can't be right.
> 
> 
>> +
>> +   return stmmac_dvr_probe(>dev, plat_dat, _res);
> 
> If stmmac_dvr_probe() fails you should disable your clocks.
> 
> 
> regards,
> Joachim Eastwood
> 

Hi Joachim,

Thanks for the hints, stm32 glue does that and I will sync on it.

Neil

Re: [RFC PATCH] net: stmmac: Add OXNAS Glue Driver

[Joachim Eastwood]

Hi Neil,

On 20 October 2016 at 17:54, Neil Armstrong  wrote:
> Add Synopsys Designware MAC Glue layer for the Oxford Semiconductor OX820.
>
> Signed-off-by: Neil Armstrong 
> ---
>  .../devicetree/bindings/net/oxnas-dwmac.txt|  44 ++
>  drivers/net/ethernet/stmicro/stmmac/Kconfig|  11 ++
>  drivers/net/ethernet/stmicro/stmmac/Makefile   |   1 +
>  drivers/net/ethernet/stmicro/stmmac/dwmac-oxnas.c  | 173 
> +
>  4 files changed, 229 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/net/oxnas-dwmac.txt
>  create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-oxnas.c

> +
> +static int oxnas_dwmac_probe(struct platform_device *pdev)
> +{
> +   struct plat_stmmacenet_data *plat_dat;
> +   struct stmmac_resources stmmac_res;
> +   struct device_node *sysctrl;
> +   struct oxnas_dwmac *dwmac;
> +   int ret;
> +
> +   sysctrl = of_parse_phandle(pdev->dev.of_node, "oxsemi,sys-ctrl", 0);
> +   if (!sysctrl) {
> +   dev_err(>dev, "failed to get sys-ctrl node\n");
> +   return -EINVAL;
> +   }
> +
> +   ret = stmmac_get_platform_resources(pdev, _res);
> +   if (ret)
> +   return ret;
> +
> +   plat_dat = stmmac_probe_config_dt(pdev, _res.mac);
> +   if (IS_ERR(plat_dat))
> +   return PTR_ERR(plat_dat);
> +
> +   dwmac = devm_kzalloc(>dev, sizeof(*dwmac), GFP_KERNEL);
> +   if (!dwmac)
> +   return -ENOMEM;
> +
> +   dwmac->regmap = syscon_node_to_regmap(sysctrl);
> +   if (IS_ERR(dwmac->regmap)) {
> +   dev_err(>dev, "failed to have sysctrl regmap\n");
> +   return PTR_ERR(dwmac->regmap);
> +   }
> +
> +   dwmac->clk = devm_clk_get(>dev, "gmac");
> +   if (IS_ERR(dwmac->clk))
> +   return PTR_ERR(dwmac->clk);
> +
> +   plat_dat->bsp_priv = dwmac;
> +   plat_dat->init = oxnas_dwmac_init;
> +   plat_dat->exit = oxnas_dwmac_exit;

Please do not use the init/exit callbacks. Implement proper driver
callbacks instead. I.e: PM resume/suspend and driver remove.

Shouldn't you call oxnas_dwmac_init() from probe as well?
As it is now it will only be called during PM resume and that can't be right.


> +
> +   return stmmac_dvr_probe(>dev, plat_dat, _res);

If stmmac_dvr_probe() fails you should disable your clocks.


regards,
Joachim Eastwood

Re: [RFC PATCH] net: stmmac: Add OXNAS Glue Driver

[Joachim Eastwood]

Hi Neil,

On 20 October 2016 at 17:54, Neil Armstrong  wrote:
> Add Synopsys Designware MAC Glue layer for the Oxford Semiconductor OX820.
>
> Signed-off-by: Neil Armstrong 
> ---
>  .../devicetree/bindings/net/oxnas-dwmac.txt|  44 ++
>  drivers/net/ethernet/stmicro/stmmac/Kconfig|  11 ++
>  drivers/net/ethernet/stmicro/stmmac/Makefile   |   1 +
>  drivers/net/ethernet/stmicro/stmmac/dwmac-oxnas.c  | 173 
> +
>  4 files changed, 229 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/net/oxnas-dwmac.txt
>  create mode 100644 drivers/net/ethernet/stmicro/stmmac/dwmac-oxnas.c

> +
> +static int oxnas_dwmac_probe(struct platform_device *pdev)
> +{
> +   struct plat_stmmacenet_data *plat_dat;
> +   struct stmmac_resources stmmac_res;
> +   struct device_node *sysctrl;
> +   struct oxnas_dwmac *dwmac;
> +   int ret;
> +
> +   sysctrl = of_parse_phandle(pdev->dev.of_node, "oxsemi,sys-ctrl", 0);
> +   if (!sysctrl) {
> +   dev_err(>dev, "failed to get sys-ctrl node\n");
> +   return -EINVAL;
> +   }
> +
> +   ret = stmmac_get_platform_resources(pdev, _res);
> +   if (ret)
> +   return ret;
> +
> +   plat_dat = stmmac_probe_config_dt(pdev, _res.mac);
> +   if (IS_ERR(plat_dat))
> +   return PTR_ERR(plat_dat);
> +
> +   dwmac = devm_kzalloc(>dev, sizeof(*dwmac), GFP_KERNEL);
> +   if (!dwmac)
> +   return -ENOMEM;
> +
> +   dwmac->regmap = syscon_node_to_regmap(sysctrl);
> +   if (IS_ERR(dwmac->regmap)) {
> +   dev_err(>dev, "failed to have sysctrl regmap\n");
> +   return PTR_ERR(dwmac->regmap);
> +   }
> +
> +   dwmac->clk = devm_clk_get(>dev, "gmac");
> +   if (IS_ERR(dwmac->clk))
> +   return PTR_ERR(dwmac->clk);
> +
> +   plat_dat->bsp_priv = dwmac;
> +   plat_dat->init = oxnas_dwmac_init;
> +   plat_dat->exit = oxnas_dwmac_exit;

Please do not use the init/exit callbacks. Implement proper driver
callbacks instead. I.e: PM resume/suspend and driver remove.

Shouldn't you call oxnas_dwmac_init() from probe as well?
As it is now it will only be called during PM resume and that can't be right.


> +
> +   return stmmac_dvr_probe(>dev, plat_dat, _res);

If stmmac_dvr_probe() fails you should disable your clocks.


regards,
Joachim Eastwood