Hin-Tak, I want to refer you to this thread which started with
http://marc.theaimsgroup.com/?l=linux-usb-users&m=111633747808908&w=2
This looks like exactly the same sort of problem that you had, although with a different kind of drive.
Thanks for letting me know. Indeed, it looks like the same problem.
On Fri, 20 May 2005, Garnet Ulrich wrote:
Alan, some more fun with plscsi. Here is some output. Seems any READ_BUFFER with length > 0x8000 is bad with this drive (except for coincidentally the 0xfbfe and maybe other values I haven't stumbled upon) you chose (randomly?) earlier.
When I try > 0x8000, plscsi hangs and usb is reset etc.. I power cycle the drive to reset for the next test. Also of note is that I was able to burn a ~260M ISO to a DVD using growisofs and it seemed to be OK. Burnt at a steady ~2x and didn't spew any scsi errors.
I'm learning a lot and I may be getting somewhere. Thank you. Anything else you can suggest I mess with (especially if it helps anyone else)?
All that springs to mind at the moment is to try some more tests to see if 0x8000 really is the boundary. My guess is that this represents the point at which some bug in the USB interface firmware shows up, not a problem with the drive itself.
At some point maybe I'll be able to find the location in cdrecord where it
decides the transfer length to use with the READ BUFFER commands. Reducing it to 32768 ought to be enough to get your device working.
As I saw it in a quick look, it seems to be with different hardware?
RH9 ships cdrecord 2.00.3 or earlier, and FC3 ships 2.01 (I encounted
the problem first in that form on my work box when I upgraded from RH9 to FC2 then FC3, and I just assumed it was a 2.6/SG_IO change until I upgraded
cdrecord on my 2.4 box at home without upgrading the kernel).
The difference should be possible to spot by diff'ing cdrecord 2.00.3 and 2.01; although I haven't found the time to do it yet.
yes, please keep me posted, and if appropriate, please subscribe me to any usb/cdwriting list on my behalf.
Regards, Hin-Tak
------------------------------------------------------- This SF.Net email is sponsored by Oracle Space Sweepstakes Want to be the first software developer in space? Enter now for the Oracle Space Sweepstakes! http://ads.osdn.com/?ad_id=7412&alloc_id=16344&op=click _______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users
