joachim, the fsck kernel parameter set you added in 1e1b3564dda3 is a great addition - but a) it's not documented in cmdline.md yet,
and b) i think the interaction with --enable-fsckfix is not good: if you --enable-fsckfix, those kernel params are ignored.
i think it would be much less surprising and useful if --enable-fsckfix didn't take all choice away,
ie. if --enable-fsckfix only changed the defaults for the new fsck_mode and fsck_repair variables (~line 65 of conf.c) , and
if the fsck invocation in line 259 of finit.c simply used whatever the current mode/repair variables are.
that way a repackager & distributor of finit can pick either an agressive default (--enable-fsckfix) or not, but all is not lost
for the end users.
Good idea. The configure option was added by an external contributor, who also added --enable-fastboot support. So we should take the time to fix that too along the same lines.
joachim, the fsck kernel parameter set you added in 1e1b3564dda3 is a great addition - but a) it's not documented in cmdline.md yet, and b) i think the interaction with --enable-fsckfix is not good: if you --enable-fsckfix, those kernel params are ignored.
i think it would be much less surprising and useful if --enable-fsckfix didn't take all choice away,
ie. if --enable-fsckfix only changed the defaults for the new fsck_mode and fsck_repair variables (~line 65 of conf.c) , and if the fsck invocation in line 259 of finit.c simply used whatever the current mode/repair variables are.
that way a repackager & distributor of finit can pick either an agressive default (--enable-fsckfix) or not, but all is not lost for the end users.