On 07/15/2013 02:27 AM, Matthijs Kooijman wrote: > Hi Stephen, > >>> I'm not sure how many of the below driver parameters are actually >>> required. I think as least the three fifo size parameters, plus >>> max_transfer_size and max_packet_count, are needed. >> >> I'm curious why any of them are required; why aren't the values embedded >> in the HW registers correct? > > As far as I've understood, there are some values (mostly PHY related) > which cannot be detected from the HW registers. My interpretation is > that the HW registers only store parameters used to synthesize the core, > (i.e., I think they store if the core supports UTMI+ and/or ULPI PHYs, > but not what kind of PHY is actually attached). > > Also IIUC, the FIFO size registers store the maximum values per FIFO and > a global maximum, but it could be that the sum of the individual maximum > FIFO sizes is more than the global maximum. This is something that could > be fixed in the driver though, by somehow "fairly" (or otherwise > intelligently) distributing the available FIFO memory over the various > FIFOs. > > These are just the suspicions I have encountered while going through the > code, perhaps Paul can confirm them. Ah, that makes sense. It wasn't clear to me from the config structure that some of the variables didn't correspond directly to config fields/registers in the HW. It'd be nice if the DT bindings (when we create them) and/or platform device wrapper only explicitly set values in the config structure that fall into this category; anything that can be queried from HW probably should be. Perhaps the config structure could even be split up into two sub-structures: One for "must specify" values which can't be probed from HW, and another to "override" values where the HW value is wrong for some reason, but usually shouldn't need to be specified. -- 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
