On Wed, Apr 10, 2024 at 03:23:10PM +0100, Matthew Wilcox wrote: > On Wed, Apr 10, 2024 at 12:21:07PM +0200, Oscar Salvador wrote: > > On Mon, Apr 08, 2024 at 05:34:29PM -0700, Jane Chu wrote: > > > At this stage, 'p' is not expected to be a large page since a _refcount has > > > been taken, right? So is it reasonable to replace the "if (retry)" block > > > with a VM_BUG_ON warning? otherwise, if 'p' became part of a different > > > large folio, how to recover from here ? > > > > We might have split the THP (if it was part of), but AFAICS > > nothing stops the kernel to coallesce this page again into a new THP via e.g: > > madvise(MADV_HUGEPAGE) ? > > Won't that allocate a _new_ THP rather than coalescing in place? Sorry, yes, I think it will. I guess I meant MADV_COLLAPSE. My point is that somehow this page can be used to form a new THP, or is there anything stoping the page to become part of a compound page again? -- Oscar Salvador SUSE Labs
