On Thu, 3 Mar 2016 16:49:21 -0800 Rick Thomas <rbthomas@pobox.com> wrote: > On Wed, 24 Feb 2016 15:06:27 -0800 Vagrant Cascadian <vagrant@debian.org> wrote: > > On 2016-02-24, Vagrant Cascadian wrote: > > > On 2016-02-04, Ben Hutchings wrote: > > >> Oh, so the MODULES=most case is bust and we need to list more host > > >> controller drivers (or include all modules under drivers/usb/host/). > > >> How about MODULES=dep; does that work now? > > > > > > MODULES=dep appears to pull in the necessary drivers on a recent stretch > > > install on a wandboard solo, using initramfs-tools 0.123, but > > > MODULES=most still requires manually including them in > > > /etc/initramfs-tools/modules. > > > > And, FWIW, MODULES=dep appears to also workaround the issue on the wandboard dual > > with initramfs-tools 0.120. > > > > > > live well, > > vagrant > > would it make sense to ensure that MODULES=most is always a superset of MODULES=dep? Could this be done by running MODULES=dep first, then running MODULES=most and delivering the union of the two sets? We could do, but I think it's a useful feature that MODULES=most is hardware-independent and I don't want to change that. I've gradually been moving the behaviour of MODULES=most toward including entire classes of drivers and I'd prefer to handle this the same way. Ben. -- Ben Hutchings Make three consecutive correct guesses and you will be considered an expert.
Attachment:
signature.asc
Description: This is a digitally signed message part