On Thu, 18 Jul 2002, David Brownell wrote: | > The Linux kernel typically achieves I/O rates of from 10-20 MBytes/sec | > with the 2.4.19 kernel using readily available retail USB 2.0 hard | > drives. These rates currently compare favorably with the speeds that | > other operating systems achieve with the same High Speed host | > controllers and USB disk drives. Future enhancements are expected to | > produce substantial speed increases. These changes will be tested in | > the Linux 2.5 kernel (which is the current development release), and | > will ensure that the USB 2.0 support in the Linux 2.6 stable release | > is both reliable and fast. | | And FYI, those future enhancements aren't quite started yet IMO. | Groundwork is in place as HCD simplifications start to happen, | but we still haven't made the usb-storage driver queue all its | transfers. I count that as a "must have by feature freeze" issue.
I agree that storage bulk queueing is very important and we want to see it in the kernel very soon, but it shouldn't have to be in 2.5 before "feature freeze" IMO. I'd like to see it there sooner instead of later, but it should be permissible to get it in after feature freeze if it takes that long. Feature freeze in no way implies code freeze. | Basically looks good to me; thanks for doing this! Ditto. -- ~Randy ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
