Currently PTE gets updated in wp_pfn_shared() after dax_pfn_mkwrite() has released corresponding radix tree entry lock. When we want to writeprotect PTE on cache flush, we need PTE modification to happen under radix tree entry lock to ensure consisten updates of PTE and radix tree (standard faults use page lock to ensure this consistency). So move update of PTE bit into dax_pfn_mkwrite().
Signed-off-by: Jan Kara <[email protected]> --- fs/dax.c | 6 ++++++ mm/memory.c | 2 +- 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/fs/dax.c b/fs/dax.c index 513881431be6..e8d61ac3d148 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -1218,6 +1218,12 @@ int dax_pfn_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) if (!entry || !radix_tree_exceptional_entry(entry)) goto out; radix_tree_tag_set(&mapping->page_tree, index, PAGECACHE_TAG_DIRTY); + /* + * If we race with somebody updating the PTE and finish_mkwrite_fault() + * fails, we don't care. We need to return VM_FAULT_NOPAGE and retry + * the fault in either case. + */ + finish_mkwrite_fault(vma, vmf); put_unlocked_mapping_entry(mapping, index, entry); out: spin_unlock_irq(&mapping->tree_lock); diff --git a/mm/memory.c b/mm/memory.c index 30cf7b36df48..47241c2f6178 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -2315,7 +2315,7 @@ static int wp_pfn_shared(struct mm_struct *mm, linear_page_index(vma, address), FAULT_FLAG_WRITE|FAULT_FLAG_MKWRITE, orig_pte); ret = vma->vm_ops->pfn_mkwrite(vma, &vmf); - if (ret & VM_FAULT_ERROR) + if (ret & VM_FAULT_ERROR || ret & VM_FAULT_NOPAGE) return ret; if (finish_mkwrite_fault(vma, &vmf) < 0) return 0; -- 2.6.6 _______________________________________________ Linux-nvdimm mailing list [email protected] https://lists.01.org/mailman/listinfo/linux-nvdimm
