David Brownell wrote: >>> Looking at the usb_stor_transfer_partial() code, it looks to me >>> like it expected to know it was in the middle of an abort, but >>> didn't ... and so it gave a catch-all message and bypassed the >>> abort processing, rather than knowing that the only reason it'd >>> ever see async unlinks is because the request was aborted. >>> >>> This particular usb-storage abort problem has been seen before. >> >> >> And is fixed in the 2.5 tree. But it doesn't really address the question >> of why the command lingered for so long... > > > Not fixed in 2.5.25, fwiw.
Or 2.5.26 ... maybe it knows better that it was aborting, but that's still an (error prone) indirect path, not triggered by seeing -ECONNRESET status. - Dave ------------------------------------------------------- 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
