On 09/03/2021 23.07, Linus Torvalds wrote: > On Tue, Mar 9, 2021 at 1:17 PM Rasmus Villemoes > <[email protected]> wrote: >> >> So add an initramfs_async= kernel parameter, allowing the main init >> process to proceed to handling device_initcall()s without waiting for >> populate_rootfs() to finish. > > I like this smaller second version of the patch, but am wondering why > we even need the parameter. > > It sounds mostly like a "maybe I didn't think of all cases" thing -
That's exactly what it is. > and one that will mean that this code will not see a lot of actual > test coverage.. Yeah, that's probably true. > And because of the lack of test coverage, I'd rather reverse the > meaning, and have the async case on by default (without even the > Kconfig option), and have the kernel command line purely as a "oops, > it's buggy, easy to ask people to test if this is what ails them". Well, I wasn't bold enough to make it "default y" by myself, but I can certainly do that and nuke the config option. > What *can* happen early boot outside of firmware loading and usermodehelpers? Well, that was what I tried to get people to tell me when I sent the first version as RFC, and also before that (https://lore.kernel.org/lkml/[email protected]/). That you can't think of anything suggests that I have covered the important cases - which does leave random drivers that poke around the filesystem on their own, but (a) it would probably be a good thing to have this flush those out and (b) there's the command line option to make it boot anyway. Rasmus
