Lines Matching refs:pages
566 * individually; we guess at most one entry per eight pages in the
568 * or roughly 1900 entries (64K) for a 64M machine with 4K pages.
1092 * out pages (if we implement it
1395 * Force pages from the specified object to be entered into
2136 * old pages here...
2529 vm_page_t m, pages;
2559 &pages, 0, TRUE)) != KERN_SUCCESS)
2572 * Insert pages into object.
2577 m = pages;
2578 pages = NEXT_PAGE(m);
2636 * free up these pages.
2709 panic("vm_allocate_cpm: pages not contig!");
3534 /* We wired what the caller asked for, zero pages */
4045 /* We unwired what the caller asked for: zero pages */
4390 * pmap entries pointing to its pages.
5857 * Unaligned parts of pages have to be physically copied. We use
6067 * Does all the vm_trickery possible for whole pages.
6243 * we could try to push pages into the pmap at this point, BUT
6245 * the pages in the source were currently mapped
6246 * and ALL the pages in the dest were touched, if there were fewer
6247 * than 2/3 of the pages touched, this optimization actually cost more cycles
6701 * map the pages into the destination map.
6729 * copies of all of the pages from
6734 * entry was wired, all of the pages
6736 * (XXX not true if we wire pages for
11003 * VM_SYNC_INVALIDATE - discard pages, only return precious
11004 * pages to manager.
11007 * - discard pages, write dirty or precious
11008 * pages back to memory manager.
11011 * - write dirty or precious pages back to
11203 * only send a m_o_s if we returned pages or if the entry
11204 * is writable (ie dirty pages may have already been sent back)