Hi Torsten, I have no experience with the wlan-ng-driver. Just speculating.
> The isp116x-driver enumerates the device. The isp116x-driver also > handles packages shorter than one ptd-buffer-len (I patched wlan-ng for > testing). > But there is one packet-constellation, the driver can't handle. > There are 2 packages. The first is 3000 Byte (Bulk, Recv) and has to be By 'Recv' you nean IN in usb terms? > chopped up by the isp116x-driver. The second is 64 Byte (Command, Send). > I think, the second package "confuses" the driver. As I understand you, these two TD's address different endpoints. As far as I know, it should be perfectly legal to address different endpoints at the same usb function simultaneously. > I've seen at the low performance block driver, that buf_curlen gets > smaller when chopping. This doesn't happen with wlan. The first part of > the 3000 Byte package is handled, then the 64 Byte packed is handled. > After that the isp116x-driver starts again with the full 3000 Byte > package and the driver loops, sending and reading to/from the device. I cannot speak for your current version of the driver, but from this behavior it seems that the device NAK-ed the requests for the first portion of the 3000-byte TD data (however large that was). I.e. the device was not ready to handle this request. I think it was NAK-ing, otherwise the TD would have been finished with error (like DeviceNotResponding, for example). Then the 64-byte packet was processed (it is not clear from your text, whether this happened successfully or not). Then the 3000-byte one was retried. It was retried from the beginning, because not a single byte was successfully transferred at the first trial. I have no idea, what may have caused the device to NAK the requests. Olav ------------------------------------------------------- The SF.Net email is sponsored by: Beat the post-holiday blues Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek. It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
