On Wed, 18 Aug 2010, Vaclav Peroutka wrote: > Alan, I am sorry to disturbing you again. I updated my system to kernel 2.6.35, I think that we went further but still there is something wrong which is silently ignored on MS Windows and not ignored at all on Linux. > > Please have a look into usbmon and dmesg logs... Well, it's doing better than before. But the device has some nasty bugs. It starts out okay. The device tells the computer that it has 64000 sectors of data (that's a little under 32 MB since a sector holds 512 bytes). The computer even reads the first few sectors successfully. Then the computer tries to read 8 sectors starting at sector 63872. This shouldn't cause any problem. But the device chokes, returning an error indication. When the computer sends a REQUEST SENSE command to find out what went wrong, the device just sends another error indication. (In other words, the device doesn't understand REQUEST SENSE -- which is a mandatory command.) The same thing happened when the READ(10) command was repeated several times, and it happened for other reads near the end of the device's storage. It appears that this device is badly broken. Probably the only reason it works with Windows is because Windows doesn't try to read the sectors near the end. I bet it would fail the same way under Windows if you tried to store a 30-MB file on the device and read it back. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html