On Fri, Feb 06, 2004 at 03:19:01PM -0800, Bradley Chapman wrote: > Mr. Dharm, > > --- Matthew Dharm <[EMAIL PROTECTED]> wrote: > > Your device has some bad firmware. It sees a request to transfer 512 > > bytes, then transfers 512 bytes, then reports that it could only transfer > > everything but the last 512 bytes, which causes an error. > > Then why does it work in 2.6.0 and not in 2.6.1 and beyond? What changed in > the USB storage code between 2.6.0 and 2.6.1 to break my flash disk?
That's a fair question.
Looking at the logs again, I realize that the device isn't actually broken,
just very strange. Technically, what it's doing is valid (I think) -- it's
padding to achieve the requested transfer length, but then indicating that
it's all padding.
But the status comes back as good.... odd. Very strange.
As to why it broke, my guess is that some recent patches to actually pay
attention to the reported residue got merged it. I thought those patches
were older than that, but Greg K-H holds on to them for a while sometimes.
> > This makes me think we need some extra code to screw around with the
> > residue field of the status message....
>
> I will test any patch that you send me.
I'm starting to wonder if this is patchable... it probably is, but I need
to think about this for a while.
Matt
--
Matthew Dharm Home: [EMAIL PROTECTED]
Maintainer, Linux USB Mass Storage Driver
IT KEEPS ASKING ME WHERE I WANT TO GO TODAY! I DONT WANT TO GO ANYWHERE!
-- Greg
User Friendly, 11/28/97
pgp00000.pgp
Description: PGP signature
