On Sun, Jun 08, 2003 at 07:44:32PM +0100, Ralph Corderoy wrote: > After soon realising that fdformat wouldn't work I tried sformat having > set up CONFIG_CHR_DEV_SG but, although that initially detects the drive, > it soon locks the laptop requiring a reboot.
That's interesting... (the lockup, that is). > I've found usbmass-ufi10.pdf and see that there's a Format Unit command > so it's possible from the USB point of view. Does this need explicit > kernel support or is it possible that a user-space program could use > this as things stand in 2.4.18? libdsk is an excellent set of programs > for writing and detecting a myriad of floppy formats with various device > backends and it would seem a USB UFI one for Linux is an obvious > addition. > > http://www.seasip.demon.co.uk/Unix/LibDsk/ The purpose of FORMAT UNIT isn't well defined. It may well do what you want. A userspace program can easily issue this command via the SG interface without any kernel modification. > Finally, I see there's a Read Format Capacity command to `Report current > media capacity and formattable capacities supported by media'. This > would seem to suggest that unless the hardware manufacturer knows about > all these old floppy formats, e.g. the Acorn one above, they aren't > available. Have I got that right? Is it possible to circumvent the UFI > and talk to the FDC behind it directly? You're pretty much right there -- the device needs to be able to identify the blocks on the device itself. The interface the device presents is basically LBA. It _may_ be possible to circumvent UFI, but you would need help from the manufacturer to do it. There's nothing in the spec to allow you that sort of raw access. Matt -- Matthew Dharm Home: [EMAIL PROTECTED] Maintainer, Linux USB Mass Storage Driver S: Another stupid question? G: There's no such thing as a stupid question, only stupid people. -- Stef and Greg User Friendly, 7/15/1998
pgp00000.pgp
Description: PGP signature
