On Fri, 9 Jan 2004, Alan Stern wrote:

> One approach -- actually, a regression -- I've been considering is to put
> an option in usb-storage to not use the scatter-gather library.  Instead
> just queue bulk transfer requests one at a time, like we used to do.  
> (For all I know, that's how Windows works...)  That should slow things
> down!  The max_sectors stuff would then become unimportant.  I'll try
> writing this over the weekend.

On further thought I realized this would have less of an effect than
simply reducing max_sectors even more.  Try decreasing it to 64, or even
32.  You could go as low as 1 if you want, although I would be surprised 
if anything below 8 would make any difference.

Note: This _will_ reduce your throughput.  But if it keeps the device 
working, then it's worth it.

Alan Stern



-------------------------------------------------------
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to