Martin Michlmayr wrote: > I asked Joey Hess whether the postinst hook for flash-kernel can > safely be removed (maybe there are corner cases where it's still > needed) but didn't get a reply. Joey?
I don't know if I've thought of every corner case, but in the general case, I don't think we need the postinst hook. * Current linux-kernel postinst runs initramfs-tools -c, which does not involve triggers, and also does not run flash-kernel. So it generates an initramfs that never gets used, but the old one in flash should keep working. * Current initramfs-tools runs flash-kernel when triggered. The git version passes a kernel version too, but even w/o that, flash-kernel should flash a coherent kernel+initramfs pair at the end of the apt run. -- see shy jo
signature.asc
Description: Digital signature
