On Fri, Oct 01, 2021 at 08:10:35PM +0200, Martin Pieuchot wrote:
> Diff below turns i386's pmap_collect() into a noop like it is on
> amd64/arm64/powerpc64... This is part of the UVM unlocking diff and
> might no longer be necessary now that pmap_extract() has been fixed.
>
> So I'd like to know if we want to align i386's behavior with other
> archs, which should help us debug MI issues or if I drop this diff.
>
> Since sthen@ tested it, I'd be in favor of putting it in. ok?
It survived a full i386 regress run.
bluhm
> Index: arch/i386/i386/pmap.c
> ===================================================================
> RCS file: /cvs/src/sys/arch/i386/i386/pmap.c,v
> retrieving revision 1.217
> diff -u -p -r1.217 pmap.c
> --- arch/i386/i386/pmap.c 11 Sep 2021 18:08:32 -0000 1.217
> +++ arch/i386/i386/pmap.c 1 Oct 2021 17:04:20 -0000
> @@ -2259,13 +2259,6 @@ pmap_unwire_86(struct pmap *pmap, vaddr_
> void
> pmap_collect(struct pmap *pmap)
> {
> - /*
> - * free all of the pt pages by removing the physical mappings
> - * for its entire address space.
> - */
> -
> - pmap_do_remove(pmap, VM_MIN_ADDRESS, VM_MAX_ADDRESS,
> - PMAP_REMOVE_SKIPWIRED);
> }
>
> /*
