linux-stable.git/drivers/usb/otg, branch linux-3.5.y

usb: otg: twl6030-usb: Fix twl writes

2012-06-22T10:12:06+00:00

There were mistakes in writing to few twl
registers. There was interchange in the
parameters being passed to twl6030_writeb().

Signed-off-by: Moiz Sonasath 
Signed-off-by: Ruslan Bilovol 
Signed-off-by: Felipe Balbi

USB: gpio_vbus: wakeup support on GPIO VBUS interrupts

2012-05-17T18:20:34+00:00

We'd like to see the system waking up from the system-wide suspend
when it gets plugged-in, or the USB cable is pulled out.

Also makes it configurable via platform data 'wakeup'.

Signed-off-by: Shinya Kuribayashi 
Signed-off-by: Greg Kroah-Hartman

USB: gpio_vbus: a missing cancellation of workqueue in remove() function

2012-05-17T18:20:34+00:00

Signed-off-by: Shinya Kuribayashi 
Signed-off-by: Greg Kroah-Hartman

USB: gpio_vbus: handle IRQ flags properly

2012-05-17T18:20:34+00:00

Currently, 'res->flags' handlings are wrong in three respects:
* the driver _modifies_ the contents of platform data
* res->flags is set up, but not used anywhere in the driver
* request_irq() always takes VBUS_IRQ_FLAGS, regardless of refs->flags

This patch tries to fix this with a policy: If a platform IRQ resource
is available, give preference to its IRQ flag(s) over a default one
(VBUS_IRQ_FLAGS).

Signed-off-by: Shinya Kuribayashi 
Signed-off-by: Greg Kroah-Hartman

USB: gpio_vbus: remove IRQF_SAMPLE_RANDOM use

2012-05-17T18:20:34+00:00

IRQF_SAMPLE_RANDOM has been scheduled for removal for years (it was
scheduled by July 2009, but not yet remvoed).

I'm not sure when it's going to take place, but would be better to
remove it now.  Thanks for scripts/checkpatch secretary.

Signed-off-by: Shinya Kuribayashi 
Signed-off-by: Greg Kroah-Hartman

USB: gpio_vbus: use cached IRQ number for consistency with the probed one

2012-05-17T18:20:34+00:00

gpio_vbus is designed to be able to get an IRQ number for VBUS change
interrupt either (1) through platform_get_resource(IORESOURCE_IRQ) or
(2) by processing gpio_to_irq(pdata->gpio_vbus), in probe() function.

On the other hand, gpio_vbus_set_peripheral() and gpio_vbus_remove()
are always doing gpio_to_irq(pdata->gpio_vbus) to get an IRQ number.

This is not just inconsistent, but also broken.  There is no guarantee
that an IRQ number obtained by platform_get_resource() is equal to
gpio_to_irq(pdata->gpio_vbus).

Cache an IRQ number in probe() function, and use it where necessary.

Signed-off-by: Shinya Kuribayashi 
Signed-off-by: Greg Kroah-Hartman

USB: twl6030-usb: pass IRQF_ONESHOT to request_threaded_irq

2012-05-17T16:49:23+00:00

The flag of IRQF_ONESHOT should be passed to request_threaded_irq,
otherwise the following failure message will be dumped because
hardware handler is defined as NULL:

[    2.271148] genirq: Threaded irq requested with handler=NULL and
!ONESHOT for irq 356
[    2.279541] twl6030_usb twl6030_usb: can't get IRQ 356, err -22
[    2.285919] twl6030_usb: probe of twl6030_usb failed with error -22

The patch fixes the twl6030-usb probe failure.

Signed-off-by: Ming Lei 
Signed-off-by: Greg Kroah-Hartman

USB: gpio_vbus: avoid consecutive vbus_session calls with the same "is_active"

2012-05-14T16:02:23+00:00

Basically, ->vbus_session() calls should be served when VBUS session
starts and ends (it's not whenever transciever drivers detect VBUS
_changes_).  Otherwise, if UDC gadget drivers don't want for some
reason ->vbus_session() calls with the same "is_active" value, either
OTG or UDC drivers need to have some protection handlings.

Also, on platforms using this 'gpio_vbus' driver, the driver is only
allowed to check whether VBUS is applied.  There is no kernel-standard
way prepared for UDC gadget drivers to do that.

With this in mind, gpio_vbus should try to prevent unnecessary
consecutive vbus_session calls being served with the same "in_active"
value.

Signed-off-by: Shinya Kuribayashi 
Signed-off-by: Greg Kroah-Hartman

USB: gpio_vbus: put a missing regulator_put() on error

2012-05-14T16:02:23+00:00

Note that regulator_put() doesn't care about whether ->vbus_draw is
valid or not.

Signed-off-by: Shinya Kuribayashi 
Signed-off-by: Greg Kroah-Hartman

USB: gpio_vbus: provide an appropriate debounce interval

2012-05-14T16:01:50+00:00

In commit c2344f13b59e007d782a3e591ebc551bc583a8b7 (USB: gpio_vbus:
add delayed vbus_session calls, 2009-01-24), usb_gadget_vbus_connect()
and ...disconnect() were extracted from the interrupt handler, so to
allow vbus_session handlers to deal with msleep() calls.

This patch takes the approach one step further.

USB2.0 specification (7.1.7.3 Connect and Disconnect Signaling) says
that the USB system software (shall) provide a debounce interval with
a minimum duration of 100 ms, which ensures that the electrical and
mechanical connection is stable before software attempts to reset
the attached device.

Signed-off-by: Shinya Kuribayashi 
Signed-off-by: Greg Kroah-Hartman