On Sat, 29 Apr 2006 [EMAIL PROTECTED] wrote: > > We keep hitting these, so I'm wondering if it would be wise > > to create a unified flag with only 32KB limit. I understand that > > this would limit the performance more than the 64KB limit which > > Benjamin used. But perhaps it's something acceptable... Something > > to think about. > > > > I have tested my device with max_sectors set to 64 and without > IGNORE_RESIDUE, > but it doesn't work. Thus it would profit by the 64 KB limit instead of a > 32 KB limit. But it doesn't matter for me. What would be a solution to > allow both values ? Initializers ? Or could this be a problem > for some devices which already needs a special initializer. If yes, it > could still be changed there. But I think it would be less clean.
I think there wouldn't be much objection to having both flags. Of course, it's up to Matt. I don't have detailed records to check, but I'm pretty sure there have been logs posted of both kinds: the residue was correct whenever the transfer length was < 32 KB, and the residue was wrong no matter what. I speculate that 32 KB is a magic number because some devices store various internal values as signed 16-bit numbers instead of unsigned (or instead of 32-bit). As long as everything is below 32 KB it works fine; once you get above, the conversion to 32 bits gives garbage. Alan Stern ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
