On Tue, Mar 6, 2012 at 11:49 AM, Heiko Stübner <heiko@xxxxxxxxx> wrote:
> Am Dienstag 06 März 2012, 19:49:03 schrieb Dima Zavin:
>> On Tue, Mar 6, 2012 at 2:51 AM, Heiko Stübner <heiko@xxxxxxxxx> wrote:
>> > Am Mittwoch, 29. Februar 2012, 23:03:11 schrieb Heiko Stübner:
>> >> Commit 9ad63986c606 (pda_power: Add support for using otg transceiver
>> >> events) converted the pda-power driver to use otg events to determine
>> >> the status of the power supply.
>> >>
>> >> As gpio-vbus didn't use otg events until now, this change breaks setups
>> >> of pda-power with a gpio-vbus transceiver.
>> >>
>> >> This patch adds the necessary otg events and notifiers to gpio-vbus.
>> >>
>> >> Signed-off-by: Heiko Stuebner <heiko@xxxxxxxxx>
>> >
>> > I'm not sure whom to pester more with this, as usb/otg/* does not contain
>> > a MAINTAINERS entry itself. But it would be cool if someone could take a
>> > look at it, as the underlying problem breaks battery indication probably
>> > not only on my machines.
>>
>> Could you describe how you use pda_power? The change only uses OTG
>> events if you did not supply your own is_ac_online/is_usb_online
>> callbacks from the boardfile for pda_power. Were you previously
>> relying on polling?
>>
>> Looking closer at it, I think there may be a bug if you are relying on
>> polling and not supplying your own is_XX_online callbacks. Could you
>> try the attached patch? (untested/uncompiled)
>
> Nope, nothing to do with polling. For reference, my pda_power device can be
> found on [1].
>
> pda_power gets to share the vbus-irq for ac and usb with gpio-vbus. For
> ac_online there is a machine-specific callback registered (not finished
> yet).
>
> By not specifying the usb_is_online callback but enabling OTG_UTILS, the
> pda_power driver uses the otg_is_online function automatically.
>
> Before the change
> static int otg_is_usb_online(void)
> {
> - return (transceiver->state == OTG_STATE_B_PERIPHERAL);
> + return (transceiver->last_event == USB_EVENT_VBUS ||
> + transceiver->last_event == USB_EVENT_ENUMERATED);
> +}
>
> this function simply checked the otg state, which is implemented in gpio-vbus,
> i.e. it sets the state on connect and disconnect.
>
> So, the cause of problem is not in the pda_power change itself, but in
> gpio_vbus which did not create otg events before, which is fixed by my patch.
Ah, right ok.
>
>
> Heiko
>
> [1] http://gitorious.org/oyo-hack/kernel/blobs/topic/board/sg060/arch/arm/mach-s3c2416/es600-common.c#line702
>
>
>
>>
>> --Dima
>>
>> > Thanks
>> > Heiko
>> >
>> >> ---
>> >> To not break the otg rework which also touched gpio-vbus, the patch is
>> >> done against the xceiv branch of
>> >> http://git.kernel.org/?p=linux/kernel/git/balbi/usb.git
>> >>
>> >> drivers/usb/otg/gpio_vbus.c | 15 ++++++++++++++-
>> >> 1 files changed, 14 insertions(+), 1 deletions(-)
>> >>
>> >> diff --git a/drivers/usb/otg/gpio_vbus.c b/drivers/usb/otg/gpio_vbus.c
>> >> index 3ece43a..a0a2178 100644
>> >> --- a/drivers/usb/otg/gpio_vbus.c
>> >> +++ b/drivers/usb/otg/gpio_vbus.c
>> >> @@ -96,7 +96,7 @@ static void gpio_vbus_work(struct work_struct *work)
>> >> struct gpio_vbus_data *gpio_vbus =
>> >> container_of(work, struct gpio_vbus_data, work);
>> >> struct gpio_vbus_mach_info *pdata = gpio_vbus->dev->platform_data;
>> >> - int gpio;
>> >> + int gpio, status;
>> >>
>> >> if (!gpio_vbus->phy.otg->gadget)
>> >> return;
>> >> @@ -108,7 +108,9 @@ static void gpio_vbus_work(struct work_struct *work)
>> >> */
>> >> gpio = pdata->gpio_pullup;
>> >> if (is_vbus_powered(pdata)) {
>> >> + status = USB_EVENT_VBUS;
>> >> gpio_vbus->phy.state = OTG_STATE_B_PERIPHERAL;
>> >> + gpio_vbus->phy.last_event = status;
>> >> usb_gadget_vbus_connect(gpio_vbus->phy.otg->gadget);
>> >>
>> >> /* drawing a "unit load" is *always* OK, except for OTG */
>> >> @@ -117,6 +119,9 @@ static void gpio_vbus_work(struct work_struct *work)
>> >> /* optionally enable D+ pullup */
>> >> if (gpio_is_valid(gpio))
>> >> gpio_set_value(gpio,
>> >> !pdata->gpio_pullup_inverted); +
>> >> + atomic_notifier_call_chain(&gpio_vbus->phy.notifier,
>> >> + status,
>> >> gpio_vbus->phy.otg->gadget); } else {
>> >> /* optionally disable D+ pullup */
>> >> if (gpio_is_valid(gpio))
>> >> @@ -125,7 +130,12 @@ static void gpio_vbus_work(struct work_struct
>> >> *work) set_vbus_draw(gpio_vbus, 0);
>> >>
>> >> usb_gadget_vbus_disconnect(gpio_vbus->phy.otg->gadget);
>> >> + status = USB_EVENT_NONE;
>> >> gpio_vbus->phy.state = OTG_STATE_B_IDLE;
>> >> + gpio_vbus->phy.last_event = status;
>> >> +
>> >> + atomic_notifier_call_chain(&gpio_vbus->phy.notifier,
>> >> + status,
>> >> gpio_vbus->phy.otg->gadget); }
>> >> }
>> >>
>> >> @@ -287,6 +297,9 @@ static int __init gpio_vbus_probe(struct
>> >> platform_device *pdev) irq, err);
>> >> goto err_irq;
>> >> }
>> >> +
>> >> + ATOMIC_INIT_NOTIFIER_HEAD(&gpio_vbus->phy.notifier);
>> >> +
>> >> INIT_WORK(&gpio_vbus->work, gpio_vbus_work);
>> >>
>> >> gpio_vbus->vbus_draw = regulator_get(&pdev->dev, "vbus_draw");
>
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
