Conversation
|
Rebased on latest |
|
Because the MDIO bus can support multiple PHYs it is often shared between PHYs (this is the case on all the colorlight boards as far as I know). Thus I think it is unwise to bundle it in with the RGMII signals as you can only request a given pin once. Adding it to the RGMIIResource for each PHY will result in only being able to use one of them at a time. |
|
I agree with @TechnoMancer that MDIO poses a problem if it is shared, however the current resource definition system is already extremely prone to this problem and solutions for it will require more of the redesign and rethink on how potentially shared resources are requested. In the meantime I think that defining a separate For future designs we should think about how we might want to indicate that one resource is tangentially related to another (say, this MDIO applies to these PHYs), mostly for a sort of documentary function. |
| return Resource.family(*args, default_name="ps2", ios=ios) | ||
|
|
||
|
|
||
| def RGMIIResource(*args, txc, txd, tx_ctl, rxc, rxd, rx_ctl, mdc, mdio, attrs=None, conn=None): |
There was a problem hiding this comment.
- Should
mdcandmdiobe optional pins? - Should there be an optional
rst_npin?
|
Closing due to missing interest. |
|
This is not a lack of interest but a lack of resources (and also the resource system being not very well designed). I apologize for the lack of attention; I would still like to merge this improvement when I'm able to. |
Preparation for PR #148