Messages in this thread | | | Date | Fri, 22 Aug 2014 00:51:47 +0300 | From | "Kirill A. Shutemov" <> | Subject | Re: [PATCH] mm: softdirty: write protect PTEs created for read faults after VM_SOFTDIRTY cleared |
| |
On Thu, Aug 21, 2014 at 05:46:01PM -0400, Peter Feiner wrote: > On Fri, Aug 22, 2014 at 12:39:42AM +0300, Kirill A. Shutemov wrote: > > On Fri, Aug 22, 2014 at 12:51:15AM +0400, Cyrill Gorcunov wrote: > > > > Looks good to me. > > > > Would you mind to apply the same pgprot_modify() approach on the > > clear_refs_write(), test and post the patch? > > > > Feel free to use my singed-off-by (or suggested-by if you prefer) once > > it's tested (see merge case below). > > Sure thing :-) > > > One thing: there could be (I haven't checked) complications on > > vma_merge(): since vm_flags are identical it assumes that it can reuse > > vma->vm_page_prot of expanded vma. But VM_SOFTDIRTY is excluded from > > vm_flags compatibility check. What should we do with vm_page_prot there? > > Since the merged VMA will have VM_SOFTDIRTY set, it's OK that it's vm_page_prot > won't be setup for write notifications. For the purpose of process migration, > you'll just get some false positives, which is tolerable.
Right. But should we disable writenotify back to avoid exessive wp-faults if it was enabled due to soft-dirty (the case when expanded vma is soft-dirty)?
-- Kirill A. Shutemov
| |