Current mainline doesn't support USB hosts 2 and 3 which only use
HSIC mode and I was wondering how this would need to be implemented.
The topic has been discussed before: [1]
And there is some implementation in the vendor kernel: [2]
It seems like two things need to be done:
1. Switch the pinmux of the strobe signal to use a pullup after
the core has been initialized.
2. Enable HSIC mode and HSIC clock
This patch only implements these basics in a minimal approach.
You need to have an additional pinmux setting "active" in the dt,
that sets the pullup.
It was tested with the SMSC LAN9730 USB Ethernet adapter on the
iMXceet Solo S board.
[1] https://patchwork.kernel.org/patch/3541771/
[2] http://git.freescale.com/git/cgit.cgi/imx/linux-imx.git/commit/?id=cf2d3ff6b217ef41f0e594daa9615e2
Signed-off-by: Frieder Schrempf <[email protected]>
---
drivers/usb/chipidea/ci_hdrc_imx.c | 10 ++++++++++
drivers/usb/chipidea/usbmisc_imx.c | 22 ++++++++++++++++++++++
2 files changed, 32 insertions(+)
diff --git a/drivers/usb/chipidea/ci_hdrc_imx.c b/drivers/usb/chipidea/ci_hdrc_imx.c
index 19f5f5f..fef8bda 100644
--- a/drivers/usb/chipidea/ci_hdrc_imx.c
+++ b/drivers/usb/chipidea/ci_hdrc_imx.c
@@ -256,6 +256,7 @@ static int ci_hdrc_imx_probe(struct platform_device *pdev)
const struct of_device_id *of_id;
const struct ci_hdrc_imx_platform_flag *imx_platform_flag;
struct device_node *np = pdev->dev.of_node;
+ struct pinctrl *pinctrl;
of_id = of_match_device(ci_hdrc_imx_dt_ids, &pdev->dev);
if (!of_id)
@@ -331,6 +332,15 @@ static int ci_hdrc_imx_probe(struct platform_device *pdev)
pm_runtime_enable(&pdev->dev);
}
+ pinctrl = devm_pinctrl_get(&pdev->dev);
+ if (!IS_ERR(pinctrl)) {
+ struct pinctrl_state *state;
+
+ state = pinctrl_lookup_state(pinctrl, "active");
+ if (!IS_ERR(state))
+ pinctrl_select_state(pinctrl, state);
+ }
+
device_set_wakeup_capable(&pdev->dev, true);
return 0;
diff --git a/drivers/usb/chipidea/usbmisc_imx.c b/drivers/usb/chipidea/usbmisc_imx.c
index 34ad5bf..a6556c8 100644
--- a/drivers/usb/chipidea/usbmisc_imx.c
+++ b/drivers/usb/chipidea/usbmisc_imx.c
@@ -64,10 +64,16 @@
#define MX6_BM_OVER_CUR_DIS BIT(7)
#define MX6_BM_OVER_CUR_POLARITY BIT(8)
#define MX6_BM_WAKEUP_ENABLE BIT(10)
+#define MX6_BM_UTMI_ON_CLOCK BIT(13)
#define MX6_BM_ID_WAKEUP BIT(16)
#define MX6_BM_VBUS_WAKEUP BIT(17)
#define MX6SX_BM_DPDM_WAKEUP_EN BIT(29)
#define MX6_BM_WAKEUP_INTR BIT(31)
+
+#define MX6_USB_HSIC_CTRL_OFFSET 0x10
+#define MX6_BM_HSIC_CLK_ON BIT(11)
+#define MX6_BM_HSIC_EN BIT(12)
+
#define MX6_USB_OTG1_PHY_CTRL 0x18
/* For imx6dql, it is host-only controller, for later imx6, it is otg's */
#define MX6_USB_OTG2_PHY_CTRL 0x1c
@@ -351,6 +357,22 @@ static int usbmisc_imx6q_init(struct imx_usbmisc_data *data)
writel(reg | MX6_BM_NON_BURST_SETTING,
usbmisc->base + data->index * 4);
+ /*
+ * Core 2 and 3 are host only and HSIC only,
+ * so we enable HSIC by default to make them usable
+ */
+ if (data->index == 2 || data->index == 3) {
+ reg = readl(usbmisc->base + data->index * 4);
+ writel(reg | MX6_BM_UTMI_ON_CLOCK,
+ usbmisc->base + data->index * 4);
+
+ reg = readl(usbmisc->base + MX6_USB_HSIC_CTRL_OFFSET +
+ (data->index - 2) * 4);
+ reg |= MX6_BM_HSIC_EN | MX6_BM_HSIC_CLK_ON;
+ writel(reg, usbmisc->base + MX6_USB_HSIC_CTRL_OFFSET +
+ (data->index - 2) * 4);
+ }
+
spin_unlock_irqrestore(&usbmisc->lock, flags);
usbmisc_imx6q_set_wakeup(data, false);
--
2.7.4
>
> The topic has been discussed before: [1] And there is some implementation in the
> vendor kernel: [2]
>
> It seems like two things need to be done:
>
> 1. Switch the pinmux of the strobe signal to use a pullup after
> the core has been initialized.
> 2. Enable HSIC mode and HSIC clock
>
> This patch only implements these basics in a minimal approach.
> You need to have an additional pinmux setting "active" in the dt, that sets the pullup.
>
> It was tested with the SMSC LAN9730 USB Ethernet adapter on the iMXceet Solo S
> board.
>
Thanks, I should do it earlier, I could not find a suitable board with HSIC support at mainline kernel.
It is very kind that you could help on it and test function at real board. My suggestion is
follow all flows in Link [2] since we need to cover suspend/resume and remote wakeup.
Peter
> [1]
> https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fpatchwork.
> kernel.org%2Fpatch%2F3541771%2F&data=02%7C01%7CPeter.Chen%40nx
> p.com%7Cea362518eae9440028aa08d62d22246b%7C686ea1d3bc2b4c6fa92cd99
> c5c301635%7C0%7C0%7C636746022015840824&sdata=sHE2yZha9%2FhN
> SSBjXogDMtdflhNc2n9mNlUe9t8E1kY%3D&reserved=0
> [2]
> https://emea01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fgit.freescale
> .com%2Fgit%2Fcgit.cgi%2Fimx%2Flinux-
> imx.git%2Fcommit%2F%3Fid%3Dcf2d3ff6b217ef41f0e594daa9615e2&data=0
> 2%7C01%7CPeter.Chen%40nxp.com%7Cea362518eae9440028aa08d62d22246b
> %7C686ea1d3bc2b4c6fa92cd99c5c301635%7C0%7C0%7C636746022015840824
> &sdata=iDRlnA9liWua7UMHBcVdUlMkBP6MIG7KUpHIAwQbDA0%3D&re
> served=0
>
> Signed-off-by: Frieder Schrempf <[email protected]>
> ---
> drivers/usb/chipidea/ci_hdrc_imx.c | 10 ++++++++++
> drivers/usb/chipidea/usbmisc_imx.c | 22 ++++++++++++++++++++++
> 2 files changed, 32 insertions(+)
>
> diff --git a/drivers/usb/chipidea/ci_hdrc_imx.c b/drivers/usb/chipidea/ci_hdrc_imx.c
> index 19f5f5f..fef8bda 100644
> --- a/drivers/usb/chipidea/ci_hdrc_imx.c
> +++ b/drivers/usb/chipidea/ci_hdrc_imx.c
> @@ -256,6 +256,7 @@ static int ci_hdrc_imx_probe(struct platform_device *pdev)
> const struct of_device_id *of_id;
> const struct ci_hdrc_imx_platform_flag *imx_platform_flag;
> struct device_node *np = pdev->dev.of_node;
> + struct pinctrl *pinctrl;
>
> of_id = of_match_device(ci_hdrc_imx_dt_ids, &pdev->dev);
> if (!of_id)
> @@ -331,6 +332,15 @@ static int ci_hdrc_imx_probe(struct platform_device *pdev)
> pm_runtime_enable(&pdev->dev);
> }
>
> + pinctrl = devm_pinctrl_get(&pdev->dev);
> + if (!IS_ERR(pinctrl)) {
> + struct pinctrl_state *state;
> +
> + state = pinctrl_lookup_state(pinctrl, "active");
> + if (!IS_ERR(state))
> + pinctrl_select_state(pinctrl, state);
> + }
> +
> device_set_wakeup_capable(&pdev->dev, true);
>
> return 0;
> diff --git a/drivers/usb/chipidea/usbmisc_imx.c b/drivers/usb/chipidea/usbmisc_imx.c
> index 34ad5bf..a6556c8 100644
> --- a/drivers/usb/chipidea/usbmisc_imx.c
> +++ b/drivers/usb/chipidea/usbmisc_imx.c
> @@ -64,10 +64,16 @@
> #define MX6_BM_OVER_CUR_DIS BIT(7)
> #define MX6_BM_OVER_CUR_POLARITY BIT(8)
> #define MX6_BM_WAKEUP_ENABLE BIT(10)
> +#define MX6_BM_UTMI_ON_CLOCK BIT(13)
> #define MX6_BM_ID_WAKEUP BIT(16)
> #define MX6_BM_VBUS_WAKEUP BIT(17)
> #define MX6SX_BM_DPDM_WAKEUP_EN BIT(29)
> #define MX6_BM_WAKEUP_INTR BIT(31)
> +
> +#define MX6_USB_HSIC_CTRL_OFFSET 0x10
> +#define MX6_BM_HSIC_CLK_ON BIT(11)
> +#define MX6_BM_HSIC_EN BIT(12)
> +
> #define MX6_USB_OTG1_PHY_CTRL 0x18
> /* For imx6dql, it is host-only controller, for later imx6, it is otg's */
> #define MX6_USB_OTG2_PHY_CTRL 0x1c
> @@ -351,6 +357,22 @@ static int usbmisc_imx6q_init(struct imx_usbmisc_data
> *data)
> writel(reg | MX6_BM_NON_BURST_SETTING,
> usbmisc->base + data->index * 4);
>
> + /*
> + * Core 2 and 3 are host only and HSIC only,
> + * so we enable HSIC by default to make them usable
> + */
> + if (data->index == 2 || data->index == 3) {
> + reg = readl(usbmisc->base + data->index * 4);
> + writel(reg | MX6_BM_UTMI_ON_CLOCK,
> + usbmisc->base + data->index * 4);
> +
> + reg = readl(usbmisc->base + MX6_USB_HSIC_CTRL_OFFSET +
> + (data->index - 2) * 4);
> + reg |= MX6_BM_HSIC_EN | MX6_BM_HSIC_CLK_ON;
> + writel(reg, usbmisc->base + MX6_USB_HSIC_CTRL_OFFSET +
> + (data->index - 2) * 4);
> + }
> +
> spin_unlock_irqrestore(&usbmisc->lock, flags);
>
> usbmisc_imx6q_set_wakeup(data, false);
> --
> 2.7.4
Hi Peter,
On 09.10.2018 03:24, Peter Chen wrote:
>
>>
>> The topic has been discussed before: [1] And there is some implementation in the
>> vendor kernel: [2]
>>
>> It seems like two things need to be done:
>>
>> 1. Switch the pinmux of the strobe signal to use a pullup after
>> the core has been initialized.
>> 2. Enable HSIC mode and HSIC clock
>>
>> This patch only implements these basics in a minimal approach.
>> You need to have an additional pinmux setting "active" in the dt, that sets the pullup.
>>
>> It was tested with the SMSC LAN9730 USB Ethernet adapter on the iMXceet Solo S
>> board.
>>
>
> Thanks, I should do it earlier, I could not find a suitable board with HSIC support at mainline kernel.
> It is very kind that you could help on it and test function at real board. My suggestion is
> follow all flows in Link [2] since we need to cover suspend/resume and remote wakeup.
It seems like power management and other things in the linux-imx kernel
are not fully implemented in mainline so there are quite a few differences.
My knowledge of USB, etc. is very limited, so I guess this is currently
beyond my skills. Can you or someone else provide a patch to get
started? I could help with testing, reviewing, etc.
Or maybe you can point out what needs to be done in more detail?
Thanks,
Frieder
> > Thanks, I should do it earlier, I could not find a suitable board with HSIC support
> at mainline kernel.
> > It is very kind that you could help on it and test function at real
> > board. My suggestion is follow all flows in Link [2] since we need to cover
> suspend/resume and remote wakeup.
>
> It seems like power management and other things in the linux-imx kernel are not
> fully implemented in mainline so there are quite a few differences.
>
> My knowledge of USB, etc. is very limited, so I guess this is currently beyond my
> skills. Can you or someone else provide a patch to get started? I could help with
> testing, reviewing, etc.
>
Ok, I will supply the patch, and you could add your dts changes as well as testing.
Thanks.
Peter
On 12.10.2018 03:19, Peter Chen wrote:
>
>>> Thanks, I should do it earlier, I could not find a suitable board with HSIC support
>> at mainline kernel.
>>> It is very kind that you could help on it and test function at real
>>> board. My suggestion is follow all flows in Link [2] since we need to cover
>> suspend/resume and remote wakeup.
>>
>> It seems like power management and other things in the linux-imx kernel are not
>> fully implemented in mainline so there are quite a few differences.
>>
>> My knowledge of USB, etc. is very limited, so I guess this is currently beyond my
>> skills. Can you or someone else provide a patch to get started? I could help with
>> testing, reviewing, etc.
>>
>
> Ok, I will supply the patch, and you could add your dts changes as well as testing.
> Thanks.
Ok, great. Just let me know when you have something that is ready for
testing and I will give it a try on my setup.
Thanks,
Frieder