On Mon, 27 Nov 2023 at 19:27, Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote: > Nice changelog. Can you please find more people to cc? Al Viro, > people who have recently worked on this. Also other interested > parties, not necessarily from the kernel side of things - people who > might use this feature. Thanks for the feedback - yes, will do. > Please use tabs to indent, like the surrounding code. Apologies for that, I'll change email client for the next submission. > If the return value is ignored, overmount_rootfs() should return void > type. Or should the return value be used in some fashion? Good point. I wasn't sure about this - at the moment, a failure in overmount_rootfs will result in the kernel continuing to unpack the initramfs "normally" directly to rootfs. This feels nicer than giving up and panicking, but could be unexpected. That said, I'm not sure under what conditions init_mkdir and init_mount (of tmpfs) would fail. I'll get a new patch out shortly, with corrected formatting, returning void, and cc'd to more people. Emily
