Jeff Garzik wrote: > > True enough :/ > > It's been feature-complete for a while, but the reports from testers > in the field have made me too nervous to push it into the upstream > kernel. > > I might push it upstream, but disable it by default, which would allow > for a wider test audience. Could you specify what tests and reports would be useful and what risks are involved? Eg If I have 2 SATA drives then could (OK, of course it *could* - but is it likely) I break sda whilst testing with sdb? I can live with crashes and hangs and I can mitigate data loss, but I may think twice if it'll toast the drive. Nb: I often think that if people bemoaning the lack of testers put a bit of effort into saying what tests would be useful then more people would run them. "Here run this and just say if it crashes" is one approach. "Try these options, use smartd, turn on debugging like this and send this part of the output if you have a problem. Previously reported problems include: <blah, blah blah>. Oh, it's only ever going to affect the drive you specify and the worst case scenario is a low-level format using the vendor's download (which may or may not be available)" - makes me more aware of what I'm getting into. David -- - : 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
