On Thu, 25 Jun 2009, Alek Du wrote: > Alan, > > thanks for review for so long a time. I'd like to use this version - it seems cleaner than previous ones. By "this version", you mean your 1/4 v3 patch? I disagree -- it's a mistake to put a ehci_qh_hw pointer in ehci_shadow. The reason is simple: The shadow pointers are what the driver uses to traverse the software structures, just as hw_next pointers are what the controller uses to traverse the hardware structures. Hence the shadow pointers should always point to software structures. If ehci_itd, ehci_sitd, and ehci_fstn were broken up the same way, I'd say the same thing about them. If you follow my suggestion, you will save one uncached access each time periodic_next_shadow() encounters a QH. And you should also be able to eliminate the backpointer from the hw structure, thereby saving additional precious space in SRAM. Alan Stern -- 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
