On Mon, Oct 24, 2022 at 03:41:26PM +0800, Wayne Chang wrote:
> From: Sing-Han Chen <singhanc@xxxxxxxxxx>
>
> Add support for the XUSB pad controller found on Tegra234 SoCs. It is
> mostly similar to the same IP found on Tegra194, because most of
> the Tegra234 XUSB PADCTL registers definition and programming sequence
> are the same as Tegra194, Tegra234 XUSB PADCTL can share the same
> driver with Tegra186 and Tegra194 XUSB PADCTL.
>
> Introduce a new feature, USB2 HW tracking, for Tegra234.
> The feature is to enable HW periodical PAD tracking which measure
> and capture the electric parameters of USB2.0 PAD.
>
> Signed-off-by: Sing-Han Chen <singhanc@xxxxxxxxxx>
> Co-developed-by: Wayne Chang <waynec@xxxxxxxxxx>
> Signed-off-by: Wayne Chang <waynec@xxxxxxxxxx>
> ---
> drivers/phy/tegra/Makefile | 1 +
> drivers/phy/tegra/xusb-tegra186.c | 65 +++++++++++++++++++++++++++++--
> drivers/phy/tegra/xusb.c | 6 +++
> drivers/phy/tegra/xusb.h | 23 +++++++++++
> 4 files changed, 92 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/phy/tegra/Makefile b/drivers/phy/tegra/Makefile
> index 89b84067cb4c..eeeea72de117 100644
> --- a/drivers/phy/tegra/Makefile
> +++ b/drivers/phy/tegra/Makefile
> @@ -7,4 +7,5 @@ phy-tegra-xusb-$(CONFIG_ARCH_TEGRA_132_SOC) += xusb-tegra124.o
> phy-tegra-xusb-$(CONFIG_ARCH_TEGRA_210_SOC) += xusb-tegra210.o
> phy-tegra-xusb-$(CONFIG_ARCH_TEGRA_186_SOC) += xusb-tegra186.o
> phy-tegra-xusb-$(CONFIG_ARCH_TEGRA_194_SOC) += xusb-tegra186.o
> +phy-tegra-xusb-$(CONFIG_ARCH_TEGRA_234_SOC) += xusb-tegra186.o
> obj-$(CONFIG_PHY_TEGRA194_P2U) += phy-tegra194-p2u.o
> diff --git a/drivers/phy/tegra/xusb-tegra186.c b/drivers/phy/tegra/xusb-tegra186.c
> index f121b4ffbbfd..cc02cea65a21 100644
> --- a/drivers/phy/tegra/xusb-tegra186.c
> +++ b/drivers/phy/tegra/xusb-tegra186.c
> @@ -89,6 +89,11 @@
> #define USB2_TRK_START_TIMER(x) (((x) & 0x7f) << 12)
> #define USB2_TRK_DONE_RESET_TIMER(x) (((x) & 0x7f) << 19)
> #define USB2_PD_TRK BIT(26)
> +#define USB2_TRK_COMPLETED BIT(31)
> +
> +#define XUSB_PADCTL_USB2_BIAS_PAD_CTL2 0x28c
> +#define USB2_TRK_HW_MODE BIT(0)
> +#define CYA_TRK_CODE_UPDATE_ON_IDLE BIT(31)
>
> #define XUSB_PADCTL_HSIC_PADX_CTL0(x) (0x300 + (x) * 0x20)
> #define HSIC_PD_TX_DATA0 BIT(1)
> @@ -609,9 +614,32 @@ static void tegra186_utmi_bias_pad_power_on(struct tegra_xusb_padctl *padctl)
> value &= ~USB2_PD_TRK;
> padctl_writel(padctl, value, XUSB_PADCTL_USB2_BIAS_PAD_CTL1);
>
> - udelay(100);
> + if (padctl->soc->poll_trk_completed) {
> + err = padctl_readl_poll(padctl, XUSB_PADCTL_USB2_BIAS_PAD_CTL1,
> + USB2_TRK_COMPLETED, USB2_TRK_COMPLETED, 100);
> + if (err) {
> + /* The failure with polling on trk complete will not
> + * cause the failure of powering on the bias pad.
> + */
> + dev_warn(dev, "failed to poll USB2 trk completed: %d\n",
> + err);
> + }
>
> - clk_disable_unprepare(priv->usb2_trk_clk);
> + value = padctl_readl(padctl, XUSB_PADCTL_USB2_BIAS_PAD_CTL1);
> + value |= USB2_TRK_COMPLETED;
> + padctl_writel(padctl, value, XUSB_PADCTL_USB2_BIAS_PAD_CTL1);
> + } else {
> + udelay(100);
> + }
> +
> + if (padctl->soc->trk_hw_mode) {
> + value = padctl_readl(padctl, XUSB_PADCTL_USB2_BIAS_PAD_CTL2);
> + value |= USB2_TRK_HW_MODE;
> + value &= ~CYA_TRK_CODE_UPDATE_ON_IDLE;
> + padctl_writel(padctl, value, XUSB_PADCTL_USB2_BIAS_PAD_CTL2);
> + } else {
> + clk_disable_unprepare(priv->usb2_trk_clk);
> + }
>
> mutex_unlock(&padctl->lock);
> }
> @@ -637,6 +665,13 @@ static void tegra186_utmi_bias_pad_power_off(struct tegra_xusb_padctl *padctl)
> value |= USB2_PD_TRK;
> padctl_writel(padctl, value, XUSB_PADCTL_USB2_BIAS_PAD_CTL1);
>
> + if (padctl->soc->trk_hw_mode) {
> + value = padctl_readl(padctl, XUSB_PADCTL_USB2_BIAS_PAD_CTL2);
> + value &= ~USB2_TRK_HW_MODE;
> + padctl_writel(padctl, value, XUSB_PADCTL_USB2_BIAS_PAD_CTL2);
> + clk_disable_unprepare(priv->usb2_trk_clk);
> + }
> +
> mutex_unlock(&padctl->lock);
> }
>
> @@ -1560,7 +1595,8 @@ const struct tegra_xusb_padctl_soc tegra186_xusb_padctl_soc = {
> EXPORT_SYMBOL_GPL(tegra186_xusb_padctl_soc);
> #endif
>
> -#if IS_ENABLED(CONFIG_ARCH_TEGRA_194_SOC)
> +#if IS_ENABLED(CONFIG_ARCH_TEGRA_194_SOC) || \
> + IS_ENABLED(CONFIG_ARCH_TEGRA_234_SOC)
> static const char * const tegra194_xusb_padctl_supply_names[] = {
> "avdd-usb",
> "vclamp-usb",
> @@ -1616,8 +1652,31 @@ const struct tegra_xusb_padctl_soc tegra194_xusb_padctl_soc = {
> .supply_names = tegra194_xusb_padctl_supply_names,
> .num_supplies = ARRAY_SIZE(tegra194_xusb_padctl_supply_names),
> .supports_gen2 = true,
> + .poll_trk_completed = true,
> };
> EXPORT_SYMBOL_GPL(tegra194_xusb_padctl_soc);
> +
> +const struct tegra_xusb_padctl_soc tegra234_xusb_padctl_soc = {
> + .num_pads = ARRAY_SIZE(tegra194_pads),
> + .pads = tegra194_pads,
> + .ports = {
> + .usb2 = {
> + .ops = &tegra186_usb2_port_ops,
> + .count = 4,
> + },
> + .usb3 = {
> + .ops = &tegra186_usb3_port_ops,
> + .count = 4,
> + },
> + },
> + .ops = &tegra186_xusb_padctl_ops,
> + .supply_names = tegra194_xusb_padctl_supply_names,
> + .num_supplies = ARRAY_SIZE(tegra194_xusb_padctl_supply_names),
> + .supports_gen2 = true,
> + .poll_trk_completed = true,
> + .trk_hw_mode = true,
> +};
> +EXPORT_SYMBOL_GPL(tegra234_xusb_padctl_soc);
I'm beginning to wonder if we perhaps went a bit overboard with this.
These symbols are used exclusively by drivers/phy/tegra/xusb.c, which
ends up in the same link unit as xusb-tegra186.c, so the export should
not be necessary.
Not necessarily something that needs fixing right now, but certainly
something to circle back to eventually.
> #endif
>
> MODULE_AUTHOR("JC Kuo <jckuo@xxxxxxxxxx>");
> diff --git a/drivers/phy/tegra/xusb.c b/drivers/phy/tegra/xusb.c
> index 95091876c422..23d179b1a5b5 100644
> --- a/drivers/phy/tegra/xusb.c
> +++ b/drivers/phy/tegra/xusb.c
> @@ -71,6 +71,12 @@ static const struct of_device_id tegra_xusb_padctl_of_match[] = {
> .compatible = "nvidia,tegra194-xusb-padctl",
> .data = &tegra194_xusb_padctl_soc,
> },
> +#endif
> +#if defined(CONFIG_ARCH_TEGRA_234_SOC)
> + {
> + .compatible = "nvidia,tegra234-xusb-padctl",
> + .data = &tegra234_xusb_padctl_soc,
> + },
> #endif
> { }
> };
> diff --git a/drivers/phy/tegra/xusb.h b/drivers/phy/tegra/xusb.h
> index 8cfbbdbd6e0c..ec0b5b023ad1 100644
> --- a/drivers/phy/tegra/xusb.h
> +++ b/drivers/phy/tegra/xusb.h
> @@ -8,6 +8,7 @@
> #define __PHY_TEGRA_XUSB_H
>
> #include <linux/io.h>
> +#include <linux/iopoll.h>
> #include <linux/mutex.h>
> #include <linux/workqueue.h>
>
> @@ -433,6 +434,8 @@ struct tegra_xusb_padctl_soc {
> unsigned int num_supplies;
> bool supports_gen2;
> bool need_fake_usb3_port;
> + bool poll_trk_completed;
> + bool trk_hw_mode;
> };
>
> struct tegra_xusb_padctl {
> @@ -475,6 +478,23 @@ static inline u32 padctl_readl(struct tegra_xusb_padctl *padctl,
> return value;
> }
>
> +static inline u32 padctl_readl_poll(struct tegra_xusb_padctl *padctl,
> + unsigned long offset, u32 val, u32 mask, int us)
> +{
> + u32 regval;
> + int err;
> +
> + err = readl_poll_timeout_atomic(padctl->regs + offset, regval,
> + (regval & mask) == val, 1, us);
Do we really need the atomic variant here? The function that calls this
already uses a mutex for protection, so it can already sleep anyway.
Also, do we really need the helper here? We use this exactly once and
this doesn't make the invocation more readable, either.
Thierry
> + dev_dbg(padctl->dev, "%08lx poll > %08x\n", offset, regval);
> + if (err) {
> + dev_err(padctl->dev, "%08lx poll timeout > %08x\n", offset,
> + regval);
> + }
> +
> + return err;
> +}
> +
> struct tegra_xusb_lane *tegra_xusb_find_lane(struct tegra_xusb_padctl *padctl,
> const char *name,
> unsigned int index);
> @@ -491,5 +511,8 @@ extern const struct tegra_xusb_padctl_soc tegra186_xusb_padctl_soc;
> #if defined(CONFIG_ARCH_TEGRA_194_SOC)
> extern const struct tegra_xusb_padctl_soc tegra194_xusb_padctl_soc;
> #endif
> +#if defined(CONFIG_ARCH_TEGRA_234_SOC)
> +extern const struct tegra_xusb_padctl_soc tegra234_xusb_padctl_soc;
> +#endif
>
> #endif /* __PHY_TEGRA_XUSB_H */
> --
> 2.25.1
>
Attachment:
signature.asc
Description: PGP signature
