On Thu, 28 Sep 2017 14:09:04 +0300 Felipe Balbi <felipe.balbi@xxxxxxxxxxxxxxx> wrote: > > > Actually, I think we should go the other way around: > > psic = read_psic_from_register(); > > if (psic) > setup_supported_speeds_based_on_psic(xhci, psic); > else > setup_supported_speeds_using_default_mappings(xhci); > > This should be enough to sort everything out (assuming no-quirky HW, of > course). Even if a new spec is released defining a new default mapping > for 10Gbps and/or adding 0x31 to SBRN, we really don't need to rely on > that. > > Well, maybe we can rely on SBRN to append a new default > mapping. Something like (also considering possible quirky hosts): > I also agree with this. Meanwhile, the USB 3.2 specification has been published. While much of it was already known, I was curious to see what names will be used fothe additional interface speeds. I really hoped that there will be no inventions of new words, such as HyperEnhancedSuperSpeedExtraPlus. Fortunately, they gave up on such names and they have chosen a more rational approach. Even if the following words are not officially deprecated, we should really stop using them, because in the new specification they have become ambiguous and their use conveys no useful information: "SuperSpeed", "Enhanced SuperSpeed", "SuperSpeedPlus". The correct names used in the new specification for the 4 speeds that can be supported by a USB 3 interface are: "Gen 1x1", "Gen 2x1", "Gen 1x2" and "Gen 2x2". In my opinion, when the speed related code from the xHCI driver will be revised now, this opportunity should also be used to purge the junk names "SuperSpeed", "Enhanced SuperSpeed" and "SuperSpeedPlus" and replace them with the 4 precise names for the 4 speeds supported by USB 3. -- 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
