[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#550584: More info

14:59 < bwh> tbm: Are you going to put those hooks back in flash-kernel
  some time soon?
15:02 < tbm> bwh: I asked joeyh and Sledge for help but nobody suggested
  a solution.  If someone sends a patch, I'd be happy to upload.
15:06 < bwh> tbm: The solution to what?
15:12 < tbm> bwh: see #550584
15:12 < tbm> bwh: the problem with just a hook script (afaiu) is that it
  would run flash-kernel several times when you have multiple kernels installed
15:12 < tbm> so it probably needs a trigger
15:12 < tbm> but I don't have time to learn about this stuff so I'm waiting
  for someone to send a patch.
15:13 < bwh> Ah yes, and so you rely on initramfs-tools using triggers
15:13 < bwh> although, of course, it won't do that any more
15:13 < bwh> not for multiple kernel updates, anyway
15:16 < tbm> bwh: so is there any solution that would run flash-kernel only once?
15:17 < bwh> Yeah, check whether the kernel being installed/updated is the one
  the user wants
15:19 < bwh> But you can use your own trigger (the policy only forbids the initramfs
  creator from doing this, so boot loader hooks work properly)
15:20 < bwh> Sorry: to be clear, the kernel hook script for the initramfs creator
  must do its work synchronously
15:20 < bwh> The kernel hook script for the boot loader can use a trigger to
  do the work asynchronously (though it is not advisable for e.g. lilo)

Martin Michlmayr

Reply to: