On Thu, Jul 28, 2016 at 6:19 AM, Gregory CLEMENT <gregory.clement@xxxxxxxxxxxxxxxxxx> wrote: > Hi Rob, > > On lun., juil. 25 2016, Thomas Petazzoni <thomas.petazzoni@xxxxxxxxxxxxxxxxxx> wrote: > >> Hello, >> >> On Mon, 25 Jul 2016 10:12:43 -0500, Rob Herring wrote: >> >>> Yes, I get that, but that is only meaningful if you want to run an OS >>> that is only aware of 395 on a 398 SoC/board (though I'd guess the 390 >>> compat is enough for that). Otherwise, that property is not really >>> meaningful as the additional nodes are enough to handle what is the >>> superset. >>> >>> I would agree both are fine if both chips are in fact the same die, >>> just fused or packaged differently. I've seen a lot of chips that are >>> supposed to be sub/supersets of each other, but have different errata >>> lists because they are different die. >> >> Unfortunately HW vendors are rarely willing to publicly indicate whether >> the different chips in their families are actually the same die fused >> differently, or really different dies. Then it is safest to assume they are different. > So do you want that we keep both "marvell,armada398" and > "marvell,armada395" or do you xant we use only "marvell,armada398" ? For the 398 based boards, I think it should only have "marvell,armada398" and don't add "marvell,armada395". Rob -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html