On Fri, Jul 14, 2017 at 1:17 PM, Amir Goldstein <amir73il@xxxxxxxxx> wrote: > On Fri, Jul 14, 2017 at 1:53 PM, Miklos Szeredi <miklos@xxxxxxxxxx> wrote: >> On Fri, Jul 14, 2017 at 12:35 PM, Amir Goldstein <amir73il@xxxxxxxxx> wrote: >> >>> What? always warn on boot? on mount? I donno. To me the comment >>> about backward compat in Kconfig seems sufficient and no sysadmin >>> can mistakenly set index=on on mount, but as I said, don't know >>> enough about downstream to say much. >> >> I meant warning in Kconfig. I see that there is a warning there, but >> it could be made shorter and more explicit. No need to explain the >> hows and whys, just warn that enabling this option is not backward >> compatible. > > By all means. I assume you will clean that up. > >> >> And if distros enable it despite this warning, and it burns their >> users, that's their problem. >> > > Hmm... but the issue related to this commit is not backward compatibility > is the sense that is depicted in Kconfig warning. > The warning speaks of mounting new overlay with an old kernel. > The concern discussed in this thread is mounting an old overlay > with a new kernel: > > This used to work with old kernel: > rm -rf /u.new > mkdir /u.new > mount -t overlay d /m -oupperdir=/u.new,lowerdir=/l.old,workdlir=/w.old > > And doesn't work with new kernel with INDEX=y. > > Another big behavior change that me good to warn about in Kconfig > is not being able to copy layers. Right. The fact is: INDEX=y is not compatible with INDEX=n. Despite that, it might well make sense for some to enable it by default, because compatibility of overlay images is not always (rarely?) an issue in practice. Thanks, Miklos -- To unsubscribe from this list: send the line "unsubscribe linux-unionfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
