On Thu, Sep 21, 2023 at 12:28:13PM -0700, Linus Torvalds wrote: > On Thu, 21 Sept 2023 at 11:51, Jeff Layton <jlayton@xxxxxxxxxx> wrote: > > > > We have many, many inodes though, and 12 bytes per adds up! > > That was my thinking, but honestly, who knows what other alignment > issues might eat up some - or all - of the theoreteical 12 bytes. > > It might be, for example, that the inode is already some aligned size, > and that the allocation alignment means that the size wouldn't > *really* shrink at all. > > So I just want to make clear that I think the 12 bytes isn't > necessarily there. Maybe you'd get it, maybe it would be hidden by > other things. I think all filesystem developers appreciate when struct inode shrinks, it's usually embedded with additional data and the size grows. I'm on a mission to squeeze btrfs_inode under 1024 so it fits better to the slab pages and currently it's about 1100 bytes. 1024 is within reach but it gets harder to find potential space savings.
