Hi, Bill Davidsen: > My assumption, as I posted earlier, is that it is trying to mount media of > any type it understands. To guess the media type the first sector is read. > This requires a seek to sector zero and a read, leaving the burner at a > location which has already been written.
The mere seek aspect is not to blame. Each WRITE command carries its target block address. We have to increment it in sync with the written data amount or else we get an error. Nevertheless, the sequential write models of CD and DVD-R[W] get disturbed by unexpected other commands. One may inquire the drive state but not much more. It might be that the problem gets triggered by the Linux device driver and not directly by hald. Me and Eduard Bloch once made experiments where our own programs were able to spoil burn runs by simply opening the drive device file O_RDONLY. One needed about a dozen tries for one failure. My opinion: A sequential burn run on CD or DVD-minus needs the drive exclusively. Anything else is not safe. This exclusivity must be negotiated without successfully opening the device file. > I think growisofs uses O_EXCL and I don't recall any > behavior after burning which made me unhappy. You need DVD-R or unformatted DVD-RW to have a chance to see the problem with growisofs. You probably need several tries to get bitten - if ever. With DVD+R there seems to be no such problem. For cost reasons i never tried to spoil a BD-R burn run. But they are very similar to DVD+R. DVD+RW, formatted DVD-RW, DVD-RAM, BD-RE are surely safe from hald. (Formatted CD-RW too, i would assume.) Have a nice day :) Thomas -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
