* handle_pte_fault chooses page fault handler according to an entry
* which was read non-atomically. Before making any commitment, on
* those architectures or configurations (e.g. i386 with PAE) which
- * might give a mix of unmatched parts, do_swap_page and do_file_page
+ * might give a mix of unmatched parts, do_swap_page and do_nonlinear_fault
* must check under lock before unmapping the pte and proceeding
* (but do_wp_page is only called after already making such a check;
- * and do_anonymous_page and do_no_page can safely check later on).
+ * and do_anonymous_page can safely check later on).
*/
static inline int pte_unmap_same(struct mm_struct *mm, pmd_t *pmd,
pte_t *page_table, pte_t orig_pte)
* bit after it clear all dirty ptes, but before a racing
* do_wp_page installs a dirty pte.
*
- * do_no_page is protected similarly.
+ * __do_fault is protected similarly.
*/
if (!page_mkwrite) {
wait_on_page_locked(dirty_page);