Lines Matching refs:page
101 * page of memory exists within exactly one object.
1153 * the page queues lock too long.
1197 * leaves the page structure around for
1222 vm_pageout_cluster(p); /* flush page */
1231 * Add this page to our list of reclaimed pages,
1407 /* Must take page lock for this - using it to protect token queue */
1434 * hogging the page queue lock too long
1938 /* page count then we should interate over pages otherwise */
1939 /* inverse page look-up will be faster */
2160 * lock -- the busy page will be enough.
2164 * Copy the page to the new object.
2197 * top-level placeholder page, if any.
2235 * any page fails [chosen]
2505 * change a page, it may be used read-only by the copy.
2507 * When the original mapping causes a page to be modified,
2508 * it must be copied to a new page that is "pushed" to
2517 * object is *not* marked copy-on-write. A copied page is pushed
2519 * the page.
3421 * Both objects and the cache are locked; the page
3453 * If the parent has a page here, or if
3454 * this page falls outside the parent,
3468 * page from one object to the other in this case.
3469 * We can't decrypt the page here, since we can't drop
3479 * Parent now has no page.
3480 * Move the backing object's page up.
3488 * Parent has an absent page...
3493 * Throw out the absent page...
3495 * page will restart with the new
3506 * Parent object has a real page.
3508 * page.
3744 * Requires that the object be locked and the page queues be unlocked.
4167 * One and two page removals are most popular.
4269 * . paging references (pages might be in page-list)
4322 * physical addresses for each page to be mapped.
4359 /* are initializing the page, its not necessary here */
4360 m->private = TRUE; /* don`t free page */
4621 printf("(off=0x%llX,page=%p)", p->offset, p);
4748 * physical page any longer and the
4752 * physical page. Is it lost ?
4770 base_page++; /* Go to the next physical page */
5161 * On entry the object and page queues are locked, the object must be a
5174 /* release page queues lock every PURGE_LOOP_QUOTA iterations */
5201 * Avoid holding the page queues lock for too long.
5204 * the object's page list doesn't change under us
5217 /* resume with the current page and a new quota */
5222 /* don't discard a wired page */
5227 * This page is no longer "purgeable",
5237 * We can't reclaim a busy page but we can deactivate
5247 /* page is being acted upon, so don't mess with it */
5255 /* we can discard this page */
5257 /* advertize that this page is in a transition state */
5261 /* unmap the page */
5271 /* we saved the cost of cleaning this page ! */
5279 * a page from an "empty" object, so do it explicitly here.
5685 vm_page_t page;
5793 page = (vm_page_t) queue_first(&object2->memq);
5794 vm_page_rename(page, object1, page->offset, FALSE);
5804 page = (vm_page_t) queue_first(&object1->memq);
5805 vm_page_rename(page, object2, page->offset, FALSE);
5812 page = (vm_page_t) queue_first(&object1->memq);
5813 page_offset = page->offset;
5814 vm_page_remove(page);
5815 page->offset = page_offset;
5816 queue_enter(&tmp_object->memq, page, vm_page_t, listq);
5822 page = (vm_page_t) queue_first(&object2->memq);
5823 vm_page_rename(page, object1, page->offset, FALSE);
5828 page = (vm_page_t) queue_first(&tmp_object->memq);
5829 queue_remove(&tmp_object->memq, page,
5831 vm_page_insert(page, object2, page->offset);
5977 * Inputs: *start == offset of page needed
6018 * 1 page, since the original fault must always
6025 * no cluster... just fault the page in
6172 * we know for a fact that the pager can't provide the page
6202 * we know for a fact that the pager can't provide the page
6225 * Allow manipulation of individual page state. This is actually part of
6269 /* someone else is playing with the page, we will */
6304 /* convention and attempts to change page state */
6306 /* because the page may already be busy. However */
6332 * We need to decrypt this encrypted page before the
6334 * But if the caller really wants to access the page's
6335 * contents, they have to keep the page "busy".
6336 * Otherwise, the page could get recycled or re-encrypted
6342 * The page is stable enough to be accessed by
6349 * The page is not busy, so don't bother
6354 * to this page.
6362 * The physical page number will remain valid
6363 * only if the page is kept busy.
6365 * caller access an encrypted page.
6382 * vm_object_page_op for page_op functions which do not require page
6384 * a low-cost alternative to page manipulation via UPLs when only a single
6385 * page was involved. The range_op call establishes the ability in the _op
6386 * family of functions to work on multiple pages where the lack of page level
6426 * page, we will have to wait
6430 * need to relook the page up since it's