lkml.org 
[lkml]   [2014]   [Nov]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: [PATCHv4 2/6] phy: improved lookup method
    Hi,

    On Tuesday 18 November 2014 09:06 PM, Heikki Krogerus wrote:
    > On Tue, Nov 18, 2014 at 10:49:18AM +0530, Kishon Vijay Abraham I wrote:
    >> On Monday 17 November 2014 09:10 PM, Heikki Krogerus wrote:
    >>> On Thu, Nov 13, 2014 at 07:03:01PM +0530, Vivek Gautam wrote:
    >>>> How about adding the change in attached patch [1] on top of this patch.
    >>>> Just introduced the phy pointer in "phy_lookup" structure, and
    >>>> modified phy_find() accordingly.
    >>>
    >>> I would be fine if we used the phy pointer to match, but if we have
    >>> the pointer to the phy in the lookup, then I would say we need to
    >>> remove the phy_name member.
    >>>
    >>> Otherwise we would have to support two ways of finding the lookups
    >>> (one for the "static" lookups where we match to the phy_name and other
    >>> where we match to the pointer). That means we will also not be able to
    >>> create "static" lookups, but those would be only needed if we wanted to
    >>> create them in board files. I don't think there will ever be need to
    >>> create the lookups in board files, so I'm more then happy to remove
    >>> the static way of creating the lookups.
    >>
    >> Just using the phy pointer sounds good. But interested to know where the phy
    >> pointer will be added to the lookup table.
    >
    > I'm making assumption that there will not be any (new) platforms where
    > the bindings are not provided in HW descriptions like dt or ACPI
    > tables. That leaves the lookups to be useful only in cases where a
    > driver has to, for whatever reason, pass it's phy's forward, like
    > dwc3 host. ULPI will be an other case where we will need to use them.

    right.
    >
    > Now the only user for the lookups is twl4030-usb, but since it's
    > actually just ULPI type PHY in cases where we have platform data to
    > deal with (please correct me if I'm wrong), I think we could later

    you mean passing consumer information using platform data?
    > remove it's need for platform data completely with the help of the
    > ULPI bus I'm working with.

    Sure.

    Thanks
    Kishon


    \
     
     \ /
      Last update: 2014-11-19 07:01    [W:4.105 / U:0.008 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site