On Mon, Nov 17, 2014 at 9:46 PM, Sergei Shtylyov
[off-list ref] wrote:

Hello.

On 11/17/2014 9:36 AM, Vivek Gautam wrote:

quoted

quoted

quoted

The host controller by itself may sometimes need to handle PHY
and re-initialize it to re-configure some of the PHY parameters
to get full support out of the PHY controller.
Therefore, facilitate getting the two possible PHYs, viz.
USB 2.0 type (UTMI+) and USB 3.0 type (PIPE3), and initialize them.

quoted

quoted

quoted

Signed-off-by: Vivek Gautam <redacted>
---
   drivers/usb/host/xhci-plat.c |   74
++++++++++++++++++++++++++++++++++++++++++
   1 file changed, 74 insertions(+)

quoted

quoted

quoted

diff --git a/drivers/usb/host/xhci-plat.c b/drivers/usb/host/xhci-plat.c
index 3d78b0c..5207d5b 100644
--- a/drivers/usb/host/xhci-plat.c
+++ b/drivers/usb/host/xhci-plat.c

[...]

quoted

quoted

quoted

@@ -204,6 +271,8 @@ static int xhci_plat_suspend(struct device *dev)
         struct usb_hcd  *hcd = dev_get_drvdata(dev);
         struct xhci_hcd *xhci = hcd_to_xhci(hcd);

+       phy_exit(hcd->phy);

quoted

quoted

    Hm, in the suspend() method?

quoted

phy_exit() should eventually be suspending the PHY and put it to
low power state.

   I thought it's a role that the power_off() mothod should play,
considering that the power_on() method gets called after the init()
method....

phy_power_off() should be cutting the clocks and power from PHY
completely, no ?
In that case one may not be able to wakeup the system from USB.

So phy_exit() gets the responsibility to put the PHY into low power state.

Ccing Kishon also to get his opinion on actual role of the two callbacks -
phy_init/exit() and phy_power_on/off().

quoted

phy_init() in resume() will then take up the task of activating the
PHY again.

quoted

phy_power_on() and phy_power_off() are called at xhci_probe() and remove()
time.

   Of course.

quoted

Does this makes sense ?

   Not much, really.

WBR, Sergei

-- 
Best Regards
Vivek Gautam
Samsung R&D Institute, Bangalore
India