On Tue, Mar 14 2000, Matthew Dharm wrote:
> > > Unfortunately, for USB Mass Storage, getting an accurate idea of how much
> > > data there is to transfer (especially for commands that can only transfer
> > > a fixed amount of information) is very important. So far, this is the
> > > only case I've found where it's wrong, and I'm afraid I don't have the
> > > time to discover where is is going wrong before 2.4.0
> >
> > This must be a loose requirement, I hope.
>
> "loose" is an odd word... for my driver, it's pretty loose. For some
> devices, it's not. That is, the protocol has the capacity to handle
> requests for more data than is there. However, some devices do not
> respond well to this condition, but most do. In general, the device
> firmware code paths have been tested for command which you expect to
> return a different amount of information (MODE_SENSE, for example), but
> not for commands which expect a fixed amount of information
> (READ_CAPACITY, for example). This is a device firmware problem, but
> really it's being caused by the SCSI layer doing the wrong thing.
>
> For my driver, I've tested code paths to deal with both cases very
> intensively. But I generally rely on the SCSI command to tell me how much
> data to expect, and I have a CD-ROM that freaks out if that's wrong.
Ok, this makes some sense at least. And as a matter of principle of course
we should set the known size when it is known.
> > I assume that you meant CD-ROM then?
>
> This is correct. Has this patch been sent to Linus?
Not yet, I just investigated the issue when I saw your mail. I'll
bundle it with my other stuff and send it.
--
* Jens Axboe <[EMAIL PROTECTED]>
* Linux CD-ROM Maintainer
* http://www.kernel.dk
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [EMAIL PROTECTED]
