diff options
author | dyson <dyson@FreeBSD.org> | 1996-05-18 03:38:05 +0000 |
---|---|---|
committer | dyson <dyson@FreeBSD.org> | 1996-05-18 03:38:05 +0000 |
commit | 242e10df110e682d1d2f90d9734a4fd3436f24c1 (patch) | |
tree | fd8af1ab66bdbd2395179d258c5c00c4a695208d /sys/vm/vm_page.c | |
parent | 3c6405efe575ceb7213f2bf914099da9523800cd (diff) | |
download | FreeBSD-src-242e10df110e682d1d2f90d9734a4fd3436f24c1.zip FreeBSD-src-242e10df110e682d1d2f90d9734a4fd3436f24c1.tar.gz |
This set of commits to the VM system does the following, and contain
contributions or ideas from Stephen McKay <syssgm@devetir.qld.gov.au>,
Alan Cox <alc@cs.rice.edu>, David Greenman <davidg@freebsd.org> and me:
More usage of the TAILQ macros. Additional minor fix to queue.h.
Performance enhancements to the pageout daemon.
Addition of a wait in the case that the pageout daemon
has to run immediately.
Slightly modify the pageout algorithm.
Significant revamp of the pmap/fork code:
1) PTE's and UPAGES's are NO LONGER in the process's map.
2) PTE's and UPAGES's reside in their own objects.
3) TOTAL elimination of recursive page table pagefaults.
4) The page directory now resides in the PTE object.
5) Implemented pmap_copy, thereby speeding up fork time.
6) Changed the pv entries so that the head is a pointer
and not an entire entry.
7) Significant cleanup of pmap_protect, and pmap_remove.
8) Removed significant amounts of machine dependent
fork code from vm_glue. Pushed much of that code into
the machine dependent pmap module.
9) Support more completely the reuse of already zeroed
pages (Page table pages and page directories) as being
already zeroed.
Performance and code cleanups in vm_map:
1) Improved and simplified allocation of map entries.
2) Improved vm_map_copy code.
3) Corrected some minor problems in the simplify code.
Implemented splvm (combo of splbio and splimp.) The VM code now
seldom uses splhigh.
Improved the speed of and simplified kmem_malloc.
Minor mod to vm_fault to avoid using pre-zeroed pages in the case
of objects with backing objects along with the already
existant condition of having a vnode. (If there is a backing
object, there will likely be a COW... With a COW, it isn't
necessary to start with a pre-zeroed page.)
Minor reorg of source to perhaps improve locality of ref.
Diffstat (limited to 'sys/vm/vm_page.c')
-rw-r--r-- | sys/vm/vm_page.c | 369 |
1 files changed, 172 insertions, 197 deletions
diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c index 6f10cea..7a95941 100644 --- a/sys/vm/vm_page.c +++ b/sys/vm/vm_page.c @@ -34,7 +34,7 @@ * SUCH DAMAGE. * * from: @(#)vm_page.c 7.4 (Berkeley) 5/7/91 - * $Id: vm_page.c,v 1.49 1996/03/09 06:56:39 dyson Exp $ + * $Id: vm_page.c,v 1.50 1996/03/28 04:53:27 dyson Exp $ */ /* @@ -140,7 +140,6 @@ static u_short vm_page_dev_bsize_chunks[] = { static inline __pure int vm_page_hash __P((vm_object_t object, vm_pindex_t pindex)) __pure2; -static void vm_page_unqueue __P((vm_page_t )); /* * vm_set_page_size: @@ -244,7 +243,7 @@ vm_page_startup(starta, enda, vaddr) vm_page_buckets = (struct pglist *) vaddr; bucket = vm_page_buckets; if (vm_page_bucket_count == 0) { - vm_page_bucket_count = 2; + vm_page_bucket_count = 1; while (vm_page_bucket_count < atop(total)) vm_page_bucket_count <<= 1; } @@ -383,7 +382,7 @@ vm_page_hash(object, pindex) * The object and page must be locked, and must be splhigh. */ -inline void +__inline void vm_page_insert(m, object, pindex) register vm_page_t m; register vm_object_t object; @@ -432,7 +431,7 @@ vm_page_insert(m, object, pindex) * The object and page must be locked, and at splhigh. */ -inline void +__inline void vm_page_remove(m) register vm_page_t m; { @@ -487,14 +486,13 @@ vm_page_lookup(object, pindex) bucket = &vm_page_buckets[vm_page_hash(object, pindex)]; - s = splhigh(); - for (m = bucket->tqh_first; m != NULL; m = m->hashq.tqe_next) { + s = splvm(); + for (m = TAILQ_FIRST(bucket); m != NULL; m = TAILQ_NEXT(m,hashq)) { if ((m->object == object) && (m->pindex == pindex)) { splx(s); return (m); } } - splx(s); return (NULL); } @@ -515,7 +513,7 @@ vm_page_rename(m, new_object, new_pindex) { int s; - s = splhigh(); + s = splvm(); vm_page_remove(m); vm_page_insert(m, new_object, new_pindex); splx(s); @@ -524,7 +522,7 @@ vm_page_rename(m, new_object, new_pindex) /* * vm_page_unqueue must be called at splhigh(); */ -static inline void +__inline void vm_page_unqueue(vm_page_t m) { int queue = m->queue; @@ -575,19 +573,19 @@ vm_page_alloc(object, pindex, page_req) page_req = VM_ALLOC_SYSTEM; }; - s = splhigh(); + s = splvm(); switch (page_req) { case VM_ALLOC_NORMAL: if (cnt.v_free_count >= cnt.v_free_reserved) { - m = vm_page_queue_free.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_free); if (m == NULL) { --vm_page_zero_count; - m = vm_page_queue_zero.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_zero); } } else { - m = vm_page_queue_cache.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_cache); if (m == NULL) { splx(s); pagedaemon_wakeup(); @@ -598,14 +596,14 @@ vm_page_alloc(object, pindex, page_req) case VM_ALLOC_ZERO: if (cnt.v_free_count >= cnt.v_free_reserved) { - m = vm_page_queue_zero.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_zero); if (m) { --vm_page_zero_count; } else { - m = vm_page_queue_free.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_free); } } else { - m = vm_page_queue_cache.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_cache); if (m == NULL) { splx(s); pagedaemon_wakeup(); @@ -618,13 +616,13 @@ vm_page_alloc(object, pindex, page_req) if ((cnt.v_free_count >= cnt.v_free_reserved) || ((cnt.v_cache_count == 0) && (cnt.v_free_count >= cnt.v_interrupt_free_min))) { - m = vm_page_queue_free.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_free); if (m == NULL) { --vm_page_zero_count; - m = vm_page_queue_zero.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_zero); } } else { - m = vm_page_queue_cache.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_cache); if (m == NULL) { splx(s); pagedaemon_wakeup(); @@ -635,10 +633,10 @@ vm_page_alloc(object, pindex, page_req) case VM_ALLOC_INTERRUPT: if (cnt.v_free_count > 0) { - m = vm_page_queue_free.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_free); if (m == NULL) { --vm_page_zero_count; - m = vm_page_queue_zero.tqh_first; + m = TAILQ_FIRST(&vm_page_queue_zero); } } else { splx(s); @@ -663,8 +661,8 @@ vm_page_alloc(object, pindex, page_req) m->flags = PG_BUSY; } m->wire_count = 0; - m->hold_count = 0; m->act_count = 0; + m->hold_count = 0; m->busy = 0; m->valid = 0; m->dirty = 0; @@ -688,114 +686,35 @@ vm_page_alloc(object, pindex, page_req) } /* - * This interface is for merging with malloc() someday. - * Even if we never implement compaction so that contiguous allocation - * works after initialization time, malloc()'s data structures are good - * for statistics and for allocations of less than a page. + * vm_page_activate: + * + * Put the specified page on the active list (if appropriate). + * + * The page queues must be locked. */ -void * -contigmalloc(size, type, flags, low, high, alignment, boundary) - unsigned long size; /* should be size_t here and for malloc() */ - int type; - int flags; - unsigned long low; - unsigned long high; - unsigned long alignment; - unsigned long boundary; +void +vm_page_activate(m) + register vm_page_t m; { - int i, s, start; - vm_offset_t addr, phys, tmp_addr; - vm_page_t pga = vm_page_array; - - size = round_page(size); - if (size == 0) - panic("vm_page_alloc_contig: size must not be 0"); - if ((alignment & (alignment - 1)) != 0) - panic("vm_page_alloc_contig: alignment must be a power of 2"); - if ((boundary & (boundary - 1)) != 0) - panic("vm_page_alloc_contig: boundary must be a power of 2"); - - start = 0; - s = splhigh(); -again: - /* - * Find first page in array that is free, within range, aligned, and - * such that the boundary won't be crossed. - */ - for (i = start; i < cnt.v_page_count; i++) { - phys = VM_PAGE_TO_PHYS(&pga[i]); - if ((pga[i].queue == PQ_FREE) && - (phys >= low) && (phys < high) && - ((phys & (alignment - 1)) == 0) && - (((phys ^ (phys + size - 1)) & ~(boundary - 1)) == 0)) - break; - } - - /* - * If the above failed or we will exceed the upper bound, fail. - */ - if ((i == cnt.v_page_count) || - ((VM_PAGE_TO_PHYS(&pga[i]) + size) > high)) { - splx(s); - return (NULL); - } - start = i; + int s; - /* - * Check successive pages for contiguous and free. - */ - for (i = start + 1; i < (start + size / PAGE_SIZE); i++) { - if ((VM_PAGE_TO_PHYS(&pga[i]) != - (VM_PAGE_TO_PHYS(&pga[i - 1]) + PAGE_SIZE)) || - (pga[i].queue != PQ_FREE)) { - start++; - goto again; - } - } + s = splvm(); + if (m->queue == PQ_ACTIVE) + panic("vm_page_activate: already active"); - /* - * We've found a contiguous chunk that meets are requirements. - * Allocate kernel VM, unfree and assign the physical pages to it and - * return kernel VM pointer. - */ - tmp_addr = addr = kmem_alloc_pageable(kernel_map, size); - if (addr == 0) { - splx(s); - return (NULL); - } + if (m->queue == PQ_CACHE) + cnt.v_reactivated++; - for (i = start; i < (start + size / PAGE_SIZE); i++) { - vm_page_t m = &pga[i]; + vm_page_unqueue(m); - TAILQ_REMOVE(&vm_page_queue_free, m, pageq); - cnt.v_free_count--; - m->valid = VM_PAGE_BITS_ALL; - m->flags = 0; - m->dirty = 0; - m->wire_count = 0; - m->act_count = 0; - m->busy = 0; - m->queue = PQ_NONE; - vm_page_insert(m, kernel_object, - OFF_TO_IDX(tmp_addr - VM_MIN_KERNEL_ADDRESS)); - vm_page_wire(m); - pmap_kenter(tmp_addr, VM_PAGE_TO_PHYS(m)); - tmp_addr += PAGE_SIZE; + if (m->wire_count == 0) { + if (m->act_count < 5) + m->act_count = 5; + TAILQ_INSERT_TAIL(&vm_page_queue_active, m, pageq); + m->queue = PQ_ACTIVE; + cnt.v_active_count++; } - splx(s); - return ((void *)addr); -} - -vm_offset_t -vm_page_alloc_contig(size, low, high, alignment) - vm_offset_t size; - vm_offset_t low; - vm_offset_t high; - vm_offset_t alignment; -{ - return ((vm_offset_t)contigmalloc(size, M_DEVBUF, M_NOWAIT, low, high, - alignment, 0ul)); } /* @@ -813,7 +732,7 @@ vm_page_free(m) int s; int flags = m->flags; - s = splhigh(); + s = splvm(); if (m->busy || (flags & PG_BUSY) || (m->queue == PQ_FREE)) { printf("vm_page_free: pindex(%ld), busy(%d), PG_BUSY(%d)\n", m->pindex, m->busy, (flags & PG_BUSY) ? 1 : 0); @@ -824,7 +743,8 @@ vm_page_free(m) } if (m->hold_count) { - panic("freeing held page, count=%d", m->hold_count); + panic("freeing held page, count=%d, pindex=%d(0x%x)", + m->hold_count, m->pindex, m->pindex); } vm_page_remove(m); @@ -840,7 +760,19 @@ vm_page_free(m) m->wire_count = 0; } m->queue = PQ_FREE; - TAILQ_INSERT_TAIL(&vm_page_queue_free, m, pageq); + + /* + * If the pageout process is grabbing the page, it is likely + * that the page is NOT in the cache. It is more likely that + * the page will be partially in the cache if it is being + * explicitly freed. + */ + if (curproc == pageproc) { + TAILQ_INSERT_TAIL(&vm_page_queue_free, m, pageq); + } else { + TAILQ_INSERT_HEAD(&vm_page_queue_free, m, pageq); + } + splx(s); /* * if pageout daemon needs pages, then tell it that there are @@ -859,7 +791,6 @@ vm_page_free(m) */ if ((cnt.v_free_count + cnt.v_cache_count) == cnt.v_free_min) { wakeup(&cnt.v_free_count); - wakeup(&proc0); } } else { splx(s); @@ -884,7 +815,7 @@ vm_page_wire(m) int s; if (m->wire_count == 0) { - s = splhigh(); + s = splvm(); vm_page_unqueue(m); splx(s); cnt.v_wire_count++; @@ -907,7 +838,7 @@ vm_page_unwire(m) { int s; - s = splhigh(); + s = splvm(); if (m->wire_count > 0) m->wire_count--; @@ -916,47 +847,14 @@ vm_page_unwire(m) cnt.v_wire_count--; TAILQ_INSERT_TAIL(&vm_page_queue_active, m, pageq); m->queue = PQ_ACTIVE; - if( m->act_count < ACT_MAX) - m->act_count += 1; - cnt.v_active_count++; - } - splx(s); -} - -/* - * vm_page_activate: - * - * Put the specified page on the active list (if appropriate). - * - * The page queues must be locked. - */ -void -vm_page_activate(m) - register vm_page_t m; -{ - int s; - - s = splhigh(); - if (m->queue == PQ_ACTIVE) - panic("vm_page_activate: already active"); - - if (m->queue == PQ_CACHE) - cnt.v_reactivated++; - - vm_page_unqueue(m); - - if (m->wire_count == 0) { - TAILQ_INSERT_TAIL(&vm_page_queue_active, m, pageq); - m->queue = PQ_ACTIVE; if (m->act_count < 5) m->act_count = 5; - else if( m->act_count < ACT_MAX) - m->act_count += 1; cnt.v_active_count++; } splx(s); } + /* * vm_page_deactivate: * @@ -982,7 +880,7 @@ vm_page_deactivate(m) if (m->queue == PQ_INACTIVE) return; - spl = splhigh(); + spl = splvm(); if (m->wire_count == 0 && m->hold_count == 0) { if (m->queue == PQ_CACHE) cnt.v_reactivated++; @@ -990,7 +888,6 @@ vm_page_deactivate(m) TAILQ_INSERT_TAIL(&vm_page_queue_inactive, m, pageq); m->queue = PQ_INACTIVE; cnt.v_inactive_count++; - m->act_count = 0; } splx(spl); } @@ -1014,7 +911,7 @@ vm_page_cache(m) return; vm_page_protect(m, VM_PROT_NONE); - s = splhigh(); + s = splvm(); vm_page_unqueue(m); TAILQ_INSERT_TAIL(&vm_page_queue_cache, m, pageq); m->queue = PQ_CACHE; @@ -1030,35 +927,6 @@ vm_page_cache(m) splx(s); } -/* - * vm_page_zero_fill: - * - * Zero-fill the specified page. - * Written as a standard pagein routine, to - * be used by the zero-fill object. - */ -boolean_t -vm_page_zero_fill(m) - vm_page_t m; -{ - pmap_zero_page(VM_PAGE_TO_PHYS(m)); - return (TRUE); -} - -/* - * vm_page_copy: - * - * Copy one page to another - */ -void -vm_page_copy(src_m, dest_m) - vm_page_t src_m; - vm_page_t dest_m; -{ - pmap_copy_page(VM_PAGE_TO_PHYS(src_m), VM_PAGE_TO_PHYS(dest_m)); - dest_m->valid = VM_PAGE_BITS_ALL; -} - /* * mapping function for valid bits or for dirty bits in @@ -1126,8 +994,6 @@ vm_page_is_valid(m, base, size) return 0; } - - void vm_page_test_dirty(m) vm_page_t m; @@ -1138,6 +1004,115 @@ vm_page_test_dirty(m) } } +/* + * This interface is for merging with malloc() someday. + * Even if we never implement compaction so that contiguous allocation + * works after initialization time, malloc()'s data structures are good + * for statistics and for allocations of less than a page. + */ +void * +contigmalloc(size, type, flags, low, high, alignment, boundary) + unsigned long size; /* should be size_t here and for malloc() */ + int type; + int flags; + unsigned long low; + unsigned long high; + unsigned long alignment; + unsigned long boundary; +{ + int i, s, start; + vm_offset_t addr, phys, tmp_addr; + vm_page_t pga = vm_page_array; + + size = round_page(size); + if (size == 0) + panic("vm_page_alloc_contig: size must not be 0"); + if ((alignment & (alignment - 1)) != 0) + panic("vm_page_alloc_contig: alignment must be a power of 2"); + if ((boundary & (boundary - 1)) != 0) + panic("vm_page_alloc_contig: boundary must be a power of 2"); + + start = 0; + s = splvm(); +again: + /* + * Find first page in array that is free, within range, aligned, and + * such that the boundary won't be crossed. + */ + for (i = start; i < cnt.v_page_count; i++) { + phys = VM_PAGE_TO_PHYS(&pga[i]); + if ((pga[i].queue == PQ_FREE) && + (phys >= low) && (phys < high) && + ((phys & (alignment - 1)) == 0) && + (((phys ^ (phys + size - 1)) & ~(boundary - 1)) == 0)) + break; + } + + /* + * If the above failed or we will exceed the upper bound, fail. + */ + if ((i == cnt.v_page_count) || + ((VM_PAGE_TO_PHYS(&pga[i]) + size) > high)) { + splx(s); + return (NULL); + } + start = i; + + /* + * Check successive pages for contiguous and free. + */ + for (i = start + 1; i < (start + size / PAGE_SIZE); i++) { + if ((VM_PAGE_TO_PHYS(&pga[i]) != + (VM_PAGE_TO_PHYS(&pga[i - 1]) + PAGE_SIZE)) || + (pga[i].queue != PQ_FREE)) { + start++; + goto again; + } + } + + /* + * We've found a contiguous chunk that meets are requirements. + * Allocate kernel VM, unfree and assign the physical pages to it and + * return kernel VM pointer. + */ + tmp_addr = addr = kmem_alloc_pageable(kernel_map, size); + if (addr == 0) { + splx(s); + return (NULL); + } + + for (i = start; i < (start + size / PAGE_SIZE); i++) { + vm_page_t m = &pga[i]; + + TAILQ_REMOVE(&vm_page_queue_free, m, pageq); + cnt.v_free_count--; + m->valid = VM_PAGE_BITS_ALL; + m->flags = 0; + m->dirty = 0; + m->wire_count = 0; + m->busy = 0; + m->queue = PQ_NONE; + vm_page_insert(m, kernel_object, + OFF_TO_IDX(tmp_addr - VM_MIN_KERNEL_ADDRESS)); + vm_page_wire(m); + pmap_kenter(tmp_addr, VM_PAGE_TO_PHYS(m)); + tmp_addr += PAGE_SIZE; + } + + splx(s); + return ((void *)addr); +} + +vm_offset_t +vm_page_alloc_contig(size, low, high, alignment) + vm_offset_t size; + vm_offset_t low; + vm_offset_t high; + vm_offset_t alignment; +{ + return ((vm_offset_t)contigmalloc(size, M_DEVBUF, M_NOWAIT, low, high, + alignment, 0ul)); +} #ifdef DDB void DDB_print_page_info(void) |