On Wednesday 28 January 2009, David Brownell wrote: > On Tuesday 27 January 2009, Inaky Perez-Gonzalez wrote: > > > That sounds to me like "<this implementation> sucks". > > > > A lot. But is what the spec says :( And actually the code sucks > > even more. Trust me, I wrote it. > > So to partition discussion a bit: for WUSB there's > an issue with block I/O performance, where three > layers don't mesh well: > > (a) block, which mostly works in n*512 byte units; > (b) DMA, which mostly works in page (4K/8K) units; > (c) WUSB, which works in oddball sizes (3.5K, variable), > and *also* wants to splice in headers all over. > > The WUSB spec imposes the oddball sizes; I presume > those headers are a hardware issue; and that conceptual > mess inevitably caused some code suckage too, which has > not yet been fixed. ;) > > That's distinct from any issues USB 3.0 has, from > what stuck in my brain after a quick read: its 1KB > packets for bulk I/o work nicely with the block and > DMA layers. > > Is that a good summary? I'd say so (warning: have been away from the WUSB code for a while and I am not too filled into the USB3.0 details). David and Sarah's say should be authoritative. -- Inaky -- 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
