Hi Peter,
On Fri, Jan 26, 2018 at 10:06 AM, Peter Chen <peter.chen@xxxxxxx> wrote:
>
>>
>> Now that usb_add_hcd parses all generic PHYs anyways the code which skips
>> initialization of a single PHY will go away.
>> Remove the code which sets struct usb_hcd's phy field from the chipidea driver as
>> this field will go away soon.
>>
>> Signed-off-by: Martin Blumenstingl <martin.blumenstingl@xxxxxxxxxxxxxx>
>> ---
>> drivers/usb/chipidea/host.c | 4 +---
>> 1 file changed, 1 insertion(+), 3 deletions(-)
>>
>> diff --git a/drivers/usb/chipidea/host.c b/drivers/usb/chipidea/host.c index
>> 19d60ed7e41f..fc324767cb0f 100644
>> --- a/drivers/usb/chipidea/host.c
>> +++ b/drivers/usb/chipidea/host.c
>> @@ -124,9 +124,7 @@ static int host_start(struct ci_hdrc *ci)
>>
>> hcd->power_budget = ci->platdata->power_budget;
>> hcd->tpl_support = ci->platdata->tpl_support;
>> - if (ci->phy)
>> - hcd->phy = ci->phy;
>> - else
>> + if (!ci->phy)
>> hcd->usb_phy = ci->usb_phy;
>>
>
> The reason hcd->phy is initialized by chipidea core is we do not need HCD core to
> touch PHY, and PHY operation is shared for both device and host mode for chipidea.
Chunfeng wanted me to drop the mtu3 patch because I forgot about
device mode in the mtu3 driver.
however, the chipidea driver seems to be different because I'm not
dropping the whole phy_{init,power_on,power_off,exit} code from it,
but only a "flag" that tells the HCD core to skip managing the USB PHY
> If I understand correct, your HCD core PHY wrapper patch set will do PHY operation if
> there is a "phy" node under controller's? If it is correct, you may supply one way to let
> the HCD core bypass phy operations for some USB controllers, eg dual-role controllers.
> Thanks.
could you please explain why the HCD core should not manage the the
PHYs when the chipidea driver is used?
my understanding is that all phy_{init,power_on,power_off,exit}
operations are ref-counted internally in the PHY framework
this means that if the chipidea driver calls phy_{init,power_on} first
then the same functions called from within the HCD core are no-ops
(except for the ref-counting)
so I think it should not change anything - however, I have no hardware
to actually prove that.
it would be great if you could explain the issue behind this (and
thereby answer the question: why we would not want the HCD core to
manage the PHY states)!
Regards
Martin
--
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
