On Sun, 28 Mar 2004, Brian S. Stephan wrote: > Brian S. Stephan wrote: > > More behavioral data: dropped .max_sectors to 32 and that same album took 15 > minutes to cp. There were only three or four instances of the double > command abort behavior, again near the beginning of the transfer; a bit of > data was moved, the fairly successive aborts happened, and then it > recovered and the vast majority of data went smoothly. Note that the time > is still a regression from 2.4 and my laptop's 2.6, both of which did the > cp in about 7 minutes.
David, I have to wonder about this. Is there any possibility at all that the s-g library isn't working right? That's the only significant difference between the 2.4 and 2.6 drivers, aside from optimizations (in HCDs for instance) that might change the transfer speed. Of course, the fact that it is hardware dependent -- works on the laptop but not the desktop -- indicates that it's not a simple software problem. > Another thing I noticed, the 2.6 NEC OHCI controller treats the device as if > I mounted with sync (I'm not) (that is, the cp takes actual time writing to > complete) where the VIA UHCI, either 2.4 or 2.6, does not (cp takes > essentially no time and all writes are done on umount). Does this make > sense? None at all. The actual sequence of writes is independent of the type of controller used. Is it possible that your systems are issuing two differents sets of commands? (About the only way you could tell this _wasn't_ happening would be if both controllers were on the same computer and you simply switched your storage device from one to the other.) > What can I do next to pin this screwy behavior down? It beats me. Alan Stern ------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
