On Mon, Nov 13, 2023 at 10:48:39PM +0000, Matthew Wilcox wrote: > On Mon, Nov 13, 2023 at 11:48:57AM -0800, Christoph Lameter wrote: > > On Fri, 10 Nov 2023, Matthew Wilcox wrote: > > > > > > Maybe Christoph is playing with min_slab_order or something, so we're > > > > getting 8 pages per slab. That's still only 2496 bytes. Why are we > > > > calling into the large kmalloc path? What's really going on here? > > > > > > Christoph? > > > > Sorry I thought I already answered that. > > > > This was a boot with slub_min_order=5 that was inadvertently left in from a > > performance test. > > Ah. So do you think we need to fix this? I'd leave the fix, so that we don't have to look into this "problem" next time. But I'm not inclined to work on a proper fix: fixing the slab accounting for this non-trivial setup. Maybe we should add a note into a doc saying that raising slub_min_order might affect the slab accounting precision?
