Re: [PATCH v3] usb: chipidea: Use extcon framework for VBUS and ID detect

From: Peter Chen
Date: Fri Jun 05 2015 - 05:44:25 EST

Next message: Heiko Stübner: "Re: [PATCH 1/3] ARM: rockchip: fix the CPU soft reset"
Previous message: Sudeep Holla: "Re: [PATCH v3 3/5] clk: add support for clocks provided by SCP(System Control Processor)"
In reply to: Ivan T. Ivanov: "Re: [PATCH v3] usb: chipidea: Use extcon framework for VBUS and ID detect"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Fri, Jun 05, 2015 at 10:37:07AM +0300, Ivan T. Ivanov wrote:
>
> On Fri, 2015-06-05 at 15:03 +0800, Peter Chen wrote:
> > On Tue, Jun 02, 2015 at 04:14:33PM +0300, Ivan T. Ivanov wrote:
> > > On recent Qualcomm platforms VBUS and ID lines are not routed to
> > > USB PHY LINK controller. Use extcon framework to receive connect
> > > and disconnect ID and VBUS notification.
> > >
> > > Signed-off-by: Ivan T. Ivanov ivanov@xxxxxxxxxx>
> > > ---
> > > Changes sice v2 [1].
> > >
> > > * Simulate IRQ on extcon event - used to trigger OTG state machine.
> > >
> >
> > No, you may not need to trigger complicate OTG FSM, instead, the
> > role switch based on ID value is ok.
>
> I have tried this also, but I decided that using IRQ simulation
> will handle all use cases.
>

After thinking more, I agree this solution

> >
> > > I have to admit that I couldn't test complete Chipidea OTG state machine,
> > > because my setup is little weird. I am using "qcom,usb-otg-ci" as PHY/OTG
> > > provider, "qcom,ehci-host" as host controller driver and "qcom,ci-hdrc" for
> > > device role.
> > >
> >
> > Then, I am puzzled what kinds of flow you have tested in this patch?
> > extcon is really needed for chipidea core, but I hope we have tested
> > patch set based on it. Rob Herring have tested Marvell's host controller
> > can also use chipidea host driver, I think the same for qualcomm's.
> > The otg flow at phy-msm-usb.c can be removed and use chipidea's instead, and
> > only remain phy part at it.
>
> I have tried this, but there is too many legacy code which is supposed to
> handle old platforms and I can not distinguish, which should stay in and
> what could be taken out. I will take a look again.
>

Yeah, you can test your own patch.

> >
> > Current role switch and vbus flow are based on otgsc register, you can
> > improve it by adding extcon support.
>
> I am not sure that I get this. Isn't exactly what this patch does?
>

Yes, you are right, see my further comments for your patch.

> > >
> > > +static int ci_vbus_notifier(struct notifier_block *nb, unsigned long event,
> > > + void *ptr)
> > > +{
> > > + struct ci_hdrc_cable *vbus = container_of(nb, struct ci_hdrc_cable, nb);
> > > + struct ci_hdrc *ci = vbus->ci;
> > > +
> > > + if (event)
> > > + vbus->state = true;
> > > + else
> > > + vbus->state = false;
> > > +
> > > + vbus->changed = true;
> > > +
> > > + ci_irq(ci->irq, ci);
> > > + return NOTIFY_DONE;
> > > +}
> > > +
> > > +static int ci_id_notifier(struct notifier_block *nb, unsigned long event,
> > > + void *ptr)
> > > +{
> > > + struct ci_hdrc_cable *id = container_of(nb, struct ci_hdrc_cable, nb);
> > > + struct ci_hdrc *ci = id->ci;
> > > +
> > > + if (event)
> > > + id->state = false;
> > > + else
> > > + id->state = true;
> > > +
> > > + id->changed = true;
> > > +

How to know the id value must be changed?
How about using id->changed = (event != id->state) ? true : false?
of cos, it needs to move before if {}.

The same change may need to add to vbus notifier.

> > > + ci_irq(ci->irq, ci);
> > > + return NOTIFY_DONE;
> > > +}
> > > +
> > > static int ci_get_platdata(struct device *dev,
> > > struct ci_hdrc_platform_data *platdata)
> > > {
> > > + struct extcon_dev *ext_vbus, *ext_id;
> > > + struct ci_hdrc_cable *cable;
> > > + int ret;
> > > +
> > > if (!platdata->phy_mode)
> > > platdata->phy_mode = of_usb_get_phy_mode(dev->of_node);
> > >
> > > @@ -591,9 +630,89 @@ static int ci_get_platdata(struct device *dev,
> > > if (of_usb_get_maximum_speed(dev->of_node) == USB_SPEED_FULL)
> > > platdata->flags |= CI_HDRC_FORCE_FULLSPEED;
> > >
> > > + ext_id = ERR_PTR(-ENODEV);
> > > + ext_vbus = ERR_PTR(-ENODEV);
> > > + if (of_property_read_bool(dev->of_node, "extcon")) {
> > > + /* Each one of them is not mandatory */
> > > + ext_vbus = extcon_get_edev_by_phandle(dev, 0);
> > > + if (IS_ERR(ext_vbus) && PTR_ERR(ext_vbus) != -ENODEV)
> > > + return PTR_ERR(ext_vbus);
> > > +
> > > + ext_id = extcon_get_edev_by_phandle(dev, 1);
> > > + if (IS_ERR(ext_id) && PTR_ERR(ext_id) != -ENODEV)
> > > + return PTR_ERR(ext_id);
> > > + }
> > > +
> > > + cable = &platdata->vbus_extcon;
> > > + cable->nb.notifier_call = ci_vbus_notifier;
> > > + cable->edev = ext_vbus;
> > > +
> > > + if (!IS_ERR(ext_vbus)) {
> > > + ret = extcon_get_cable_state(cable->edev, "USB");

I have not read extcon framework too much, but seems you should only
can get cable state after register it (through ci_extcon_register)?
ci_get_platdata is called before ci core probe.

> > > + if (ret)
> > > + cable->state = true;
> > > + else
> > > + cable->state = false;
> > > + }
> > > +
> > > + cable = &platdata->id_extcon;
> > > + cable->nb.notifier_call = ci_id_notifier;
> > > + cable->edev = ext_id;
> > > +
> > > + if (!IS_ERR(ext_id)) {
> > > + ret = extcon_get_cable_state(cable->edev, "USB-HOST");
> > > + if (ret)
> > > + cable->state = false;
> > > + else
> > > + cable->state = true;
> > > + }
> > > + return 0;
> > > +}
> > > +
> > > +static int ci_extcon_register(struct ci_hdrc *ci)
> > > +{
> > > + struct ci_hdrc_cable *id, *vbus;
> > > + int ret;
> > > +
> > > + id = &ci->platdata->id_extcon;
> > > + id->ci = ci;
> > > + if (!IS_ERR(id->edev)) {
> > > + ret = extcon_register_interest(&id->conn, id->edev->name,
> > > + "USB-HOST", &id->nb);
> > > + if (ret < 0) {
> > > + dev_err(ci->dev, "register ID failed\n");
> > > + return ret;
> > > + }
> > > + }
> > > +
> > > + vbus = &ci->platdata->vbus_extcon;
> > > + vbus->ci = ci;
> > > + if (!IS_ERR(vbus->edev)) {
> > > + ret = extcon_register_interest(&vbus->conn, vbus->edev->name,
> > > + "USB", &vbus->nb);
> > > + if (ret < 0) {
> > > + extcon_unregister_interest(&id->conn);
> > > + dev_err(ci->dev, "register VBUS failed\n");
> > > + return ret;
> > > + }
> > > + }
> > > +
> > > return 0;
> > > }
> > >
> > > +static void ci_extcon_unregister(struct ci_hdrc *ci)
> > > +{
> > > + struct ci_hdrc_cable *cable;
> > > +
> > > + cable = &ci->platdata->id_extcon;
> > > + if (!IS_ERR(cable->edev))
> > > + extcon_unregister_interest(&cable->conn);
> > > +
> > > + cable = &ci->platdata->vbus_extcon;
> > > + if (!IS_ERR(cable->edev))
> > > + extcon_unregister_interest(&cable->conn);
> > > +}
> > > +
> > > static DEFINE_IDA(ci_ida);
> > >
> > > struct platform_device *ci_hdrc_add_device(struct device *dev,
> > > @@ -817,6 +936,10 @@ static int ci_hdrc_probe(struct platform_device *pdev)
> > > if (ret)
> > > goto stop;
> > >
> > > + ret = ci_extcon_register(ci);
> > > + if (ret)
> > > + goto stop;
> > > +
> > > if (ci->supports_runtime_pm) {
> > > pm_runtime_set_active(&pdev->dev);
> > > pm_runtime_enable(&pdev->dev);
> > > @@ -834,6 +957,7 @@ static int ci_hdrc_probe(struct platform_device *pdev)
> > > if (!ret)
> > > return 0;
> > >

There is dbg_create_files between these code which may also fail.

> > > + ci_extcon_unregister(ci);
> > > stop:
> > > ci_role_destroy(ci);
> > > deinit_phy:
> > > @@ -853,6 +977,7 @@ static int ci_hdrc_remove(struct platform_device *pdev)
> > > }
> > >
> > > dbg_remove_files(ci);
> > > + ci_extcon_unregister(ci);
> > > ci_role_destroy(ci);
> > > ci_hdrc_enter_lpm(ci, true);
> > > ci_usb_phy_exit(ci);
> > > diff --git a/drivers/usb/chipidea/otg.c b/drivers/usb/chipidea/otg.c
> > > index ad6c87a..77fe2df 100644
> > > --- a/drivers/usb/chipidea/otg.c
> > > +++ b/drivers/usb/chipidea/otg.c
> > > @@ -30,7 +30,41 @@
> > > */
> > > u32 hw_read_otgsc(struct ci_hdrc *ci, u32 mask)
> > > {
> > > - return hw_read(ci, OP_OTGSC, mask);
> > > + struct ci_hdrc_cable *cable;
> > > + u32 val = hw_read(ci, OP_OTGSC, mask);
> > > +
> > > + /*
> > > + * If using extcon framework for VBUS and/or ID signal
> > > + * detection overwrite OTGSC resiter value
> > > + */

%s/resiter/register

> > > + cable = &ci->platdata->vbus_extcon;
> > > + if (!IS_ERR(cable->edev)) {
> > > + if (cable->changed)
> > > + val |= OTGSC_BSVIS;
> > > + else
> > > + val &= ~OTGSC_BSVIS;
> > > +
> > > + if (cable->state)
> > > + val |= OTGSC_BSV;
> > > + else
> > > + val &= ~OTGSC_BSV;
> > > + }

You can clear cable->changed after that, then no change for hw_write_otgsc
is needed.

> > > +
> > > + cable = &ci->platdata->id_extcon;
> > > + if (!IS_ERR(cable->edev)) {
> > > + if (cable->changed)
> > > + val |= OTGSC_IDIS;
> > > + else
> > > + val &= ~OTGSC_IDIS;
> > > +
> > > + if (cable->state)
> > > + val |= OTGSC_ID;
> > > + else
> > > + val &= ~OTGSC_ID;
> > > + }
> > > +
> > > + val &= mask;

It seems you do not need to do &mask again.

> > > + return val;
> > > }
> > >
> > > /**
> > > @@ -40,7 +74,24 @@ u32 hw_read_otgsc(struct ci_hdrc *ci, u32 mask)
> > > */
> > > void hw_write_otgsc(struct ci_hdrc *ci, u32 mask, u32 data)
> > > {
> > > - hw_write(ci, OP_OTGSC, mask | OTGSC_INT_STATUS_BITS, data);
> > > + struct ci_hdrc_cable *cable;
> > > +
> > > + mask |= OTGSC_INT_STATUS_BITS;
> > > +
> > > + /* clear artificial bits */
> > > + cable = &ci->platdata->vbus_extcon;
> > > + if (!IS_ERR(cable->edev)) {
> > > + if (data & OTGSC_BSVIS)
> > > + cable->changed = false;
> > > + }
> > > +
> > > + cable = &ci->platdata->id_extcon;
> > > + if (!IS_ERR(cable->edev)) {
> > > + if (data & OTGSC_IDIS)
> > > + cable->changed = false;
> > > + }
> > > +
> > > + hw_write(ci, OP_OTGSC, mask, data);
> > > }
> > >
> > > /**
> > > diff --git a/include/linux/usb/chipidea.h b/include/linux/usb/chipidea.h
> > > index ab94f78..05915cb1 100644
> > > --- a/include/linux/usb/chipidea.h
> > > +++ b/include/linux/usb/chipidea.h
> > > @@ -5,9 +5,29 @@
> > > #ifndef __LINUX_USB_CHIPIDEA_H
> > > #define __LINUX_USB_CHIPIDEA_H
> > >
> > > +#include <linux/extcon.h>
> > > #include <linux/usb/otg.h>
> > >
> > > struct ci_hdrc;
> > > +
> > > +/**
> > > + * struct ci_hdrc_cable - structure for external connector cable state tracking
> > > + * @state: current state of the line
> > > + * @changed: set to true when extcon event happen
> > > + * @edev: device which generate events
> > > + * @ci: driver state of the chipidea device
> > > + * @nb: hold event notification callback
> > > + * @conn: used for notification registration
> > > + */
> > > +struct ci_hdrc_cable {
> > > + bool state;
> > > + bool changed;
> > > + struct extcon_dev*edev;
> > > + struct ci_hdrc*ci;
> > > + struct notifier_blocknb;
> > > + struct extcon_specific_cable_nb conn;
> > > +};
> > > +
> > > struct ci_hdrc_platform_data {
> > > const char*name;
> > > /* offset of the capability registers */
> > > @@ -35,6 +55,10 @@ struct ci_hdrc_platform_data {
> > > void (*notify_event) (struct ci_hdrc *ci, unsigned event);
> > > struct regulator*reg_vbus;
> > > bool tpl_support;
> > > +
> > > + /* VBUS and ID signal state tracking, using extcon framework */
> > > + struct ci_hdrc_cablevbus_extcon;
> > > + struct ci_hdrc_cableid_extcon;
> > > };
> > >
> > > /* Default offset of capability registers */
> > > --
> > > 1.9.1
> > >
> >
> >

--

Best Regards,
Peter Chen
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Heiko Stübner: "Re: [PATCH 1/3] ARM: rockchip: fix the CPU soft reset"
Previous message: Sudeep Holla: "Re: [PATCH v3 3/5] clk: add support for clocks provided by SCP(System Control Processor)"
In reply to: Ivan T. Ivanov: "Re: [PATCH v3] usb: chipidea: Use extcon framework for VBUS and ID detect"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]