On Mon, 7 Jun 2004, Ingo van Lil wrote:

> Oh, I did that already: The corrupted blocks seem to be scattered
> randomly. Even when I'm copying from /dev/zero the result is always
> different. The number of bytes to be written before any I/O errors
> appear isn't constant either.
> 
> To make things even more strange: Yesterday, after excessive plugging
> and unplugging, rebooting, loading and unloading modules and card
> swapping the bloody thing did eventually begin to work as expected.
> Well, almost: from time to time the data transfer would stop for about
> 30 seconds, but I could still fill the entire card with uncorrupted
> files. After a further reboot the errors were back.

Those 30-second delays were probably device resets, indicating that
something else was going wrong.  This really feels like a hardware
problem.  Software problems are generally (but not always!) more
consistent.

In spite of the fact that transfers under Windows worked okay, I wonder if
there isn't something wrong with your device.  It might be that Windows
and Linux stress the hardware in different ways, because of different
timings for example.  If you can test transfers to a totally different
USB mass-storage device maybe you'll find that they work okay.

I'll also try writing 7 MB worth of zeros on my system, to see how it 
works out.

> There's another striking coincidence: The additional byte's value is
> always 0x40 == 64.

Probably nothing more than a coincidence, but who knows?

By the way, I just noticed in your original posting that the ASCII data in 
your hexdump doesn't match the hex values.  What happened?

Alan Stern



-------------------------------------------------------
This SF.Net email is sponsored by: GNOME Foundation
Hackers Unite!  GUADEC: The world's #1 Open Source Desktop Event.
GNOME Users and Developers European Conference, 28-30th June in Norway
http://2004/guadec.org
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to