On Sat, 29 Jul 2006, James Bottomley wrote: > > But not for every CD burner ... that's the point. The heuristics we > have now work for a large subset. For the rest, we get a stream of "my > CD won't burn as a user like it's supposed to" bug reports. That's not _our_ problem. That's likely the problem of "cdrecord" _thinking_ that it needs to do something smart, when it really really doesn't need to. I have yet to hear anything that implies anything else than "cdrecord is just buggy". > There are certain CDs that just require vendor specific magic to > work ... even cdrecord has no choice but to do this. IF that is actually true (and if you were told this by Joerg, I would double- and triple-check it from some other source), then in the end you do end up having to have a per-device translation table, and it probably shouldn't be in the kernel. It probably _should_ be in a suid-root thing that validates the commands. People used to make cdrecord suid, but that thing really isn't written to be done that way. That doesn't mean that something else couldn't be done properly. There's no way the kernel should do this. In a hotplug environment in particular (which is not _that_ common with CD-ROM's but certainly not unheard of), you don't want to have the kernel know about every quirk of every device anyway. Linus - : send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
