On 06/25/2014 09:18 PM, Jason Cooper wrote: > On Wed, Jun 25, 2014 at 04:08:44PM +0200, Sebastian Hesselbarth wrote: >> This *RFC* adds a driver stub for ChipIdea USB dual role controllers >> found on Marvell MVEBU SoCs. Although, I consider this driver quite >> finished, the corresponding DT binding is not. Also, we currently >> have no corresponding driver in Linux for both MVEBU CI stub and >> USB PHY. >> >> The reason I send it here and now, is to get some input from the >> (Linux) MVEBU guys on the binding and functional tests on Armada XP. > > My first thought is if this chipidea IP is used beyond the mvebu > ecosystem. Shouldn't we name it and it's compatible strings with > 'chipidea' instead of 'marvell'? While the IP core is CI, you remember the mbus upstream registers? That registers are very special to MVEBU SoCs and make the IP itself "marvell". But I agree, that we could add "chipidea", although spec itself names it "ARC" already. ARC aquired CI and was later bought by Synopsys, but at that time it must have been ARC. > Also, is there any other way to differentiate between the two versions > other that manufacturing process resolution? Perhaps the free-electrons > guys could get us some IP revision numbers? I just followed how Marvell's BSP differentiates USB PHYs. In addition to 40nm and 65nm, there will be 90nm and 130nm for the older SoCs. If there is a way for free-electron to get more information about a better naming, that would be nice. Sebastian _______________________________________________ barebox mailing list barebox@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/barebox
