diff options
Diffstat (limited to 'sys')
| -rw-r--r-- | sys/amd64/amd64/pmap.c | 1105 | ||||
| -rw-r--r-- | sys/i386/i386/pmap.c | 1105 | ||||
| -rw-r--r-- | sys/vm/pmap.h | 3 | ||||
| -rw-r--r-- | sys/vm/vm_page.c | 7 | ||||
| -rw-r--r-- | sys/vm/vm_pageout.c | 79 |
5 files changed, 1292 insertions, 1007 deletions
diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c index 345b80c..abdb456 100644 --- a/sys/amd64/amd64/pmap.c +++ b/sys/amd64/amd64/pmap.c @@ -39,17 +39,7 @@ * SUCH DAMAGE. * * from: @(#)pmap.c 7.7 (Berkeley) 5/12/91 - * $Id: pmap.c,v 1.103 1996/06/12 05:02:52 gpalmer Exp $ - */ - -/* - * Derived from hp300 version by Mike Hibler, this version by William - * Jolitz uses a recursive map [a pde points to the page directory] to - * map the page tables using the pagetables themselves. This is done to - * reduce the impact on kernel virtual memory for lots of sparse address - * space, and to reduce the cost of memory to each process. - * - * Derived from: hp300/@(#)pmap.c 7.1 (Berkeley) 12/5/90 + * $Id: pmap.c,v 1.104 1996/06/12 06:52:12 dyson Exp $ */ /* @@ -97,6 +87,7 @@ #include <vm/vm_object.h> #include <vm/vm_extern.h> #include <vm/vm_pageout.h> +#include <vm/vm_pager.h> #include <machine/pcb.h> #include <machine/cputypes.h> @@ -172,35 +163,193 @@ static caddr_t CADDR2; static pt_entry_t *msgbufmap; struct msgbuf *msgbufp; +pt_entry_t *PMAP1; +unsigned *PADDR1; + static void free_pv_entry __P((pv_entry_t pv)); static __inline unsigned * get_ptbase __P((pmap_t pmap)); static pv_entry_t get_pv_entry __P((void)); static void i386_protection_init __P((void)); static void pmap_alloc_pv_entry __P((void)); static void pmap_changebit __P((vm_offset_t pa, int bit, boolean_t setem)); -static void pmap_enter_quick __P((pmap_t pmap, vm_offset_t va, - vm_offset_t pa)); + static int pmap_is_managed __P((vm_offset_t pa)); static void pmap_remove_all __P((vm_offset_t pa)); +static void pmap_enter_quick __P((pmap_t pmap, vm_offset_t va, + vm_offset_t pa)); +static int pmap_remove_pte __P((struct pmap *pmap, unsigned *ptq, + vm_offset_t sva)); static void pmap_remove_page __P((struct pmap *pmap, vm_offset_t va)); static __inline int pmap_remove_entry __P((struct pmap *pmap, pv_entry_t *pv, vm_offset_t va)); -static int pmap_remove_pte __P((struct pmap *pmap, unsigned *ptq, - vm_offset_t sva)); -static boolean_t - pmap_testbit __P((vm_offset_t pa, int bit)); +static boolean_t pmap_testbit __P((vm_offset_t pa, int bit)); static __inline void pmap_insert_entry __P((pmap_t pmap, vm_offset_t va, vm_page_t mpte, vm_offset_t pa)); static __inline vm_page_t pmap_allocpte __P((pmap_t pmap, vm_offset_t va)); -static void pmap_remove_pte_mapping __P((vm_offset_t pa)); + static __inline int pmap_release_free_page __P((pmap_t pmap, vm_page_t p)); -static vm_page_t _pmap_allocpte __P((pmap_t pmap, vm_offset_t va, int ptepindex)); +static vm_page_t _pmap_allocpte __P((pmap_t pmap, int ptepindex)); +#define VATRACK 4 #define PDSTACKMAX 16 static vm_offset_t pdstack[PDSTACKMAX]; static int pdstackptr; +/* + * Bootstrap the system enough to run with virtual memory. + * + * On the i386 this is called after mapping has already been enabled + * and just syncs the pmap module with what has already been done. + * [We can't call it easily with mapping off since the kernel is not + * mapped with PA == VA, hence we would have to relocate every address + * from the linked base (virtual) address "KERNBASE" to the actual + * (physical) address starting relative to 0] + */ +void +pmap_bootstrap(firstaddr, loadaddr) + vm_offset_t firstaddr; + vm_offset_t loadaddr; +{ + vm_offset_t va; + pt_entry_t *pte; + + avail_start = firstaddr; + + /* + * XXX The calculation of virtual_avail is wrong. It's NKPT*PAGE_SIZE too + * large. It should instead be correctly calculated in locore.s and + * not based on 'first' (which is a physical address, not a virtual + * address, for the start of unused physical memory). The kernel + * page tables are NOT double mapped and thus should not be included + * in this calculation. + */ + virtual_avail = (vm_offset_t) KERNBASE + firstaddr; + virtual_end = VM_MAX_KERNEL_ADDRESS; + + /* + * Initialize protection array. + */ + i386_protection_init(); + + /* + * The kernel's pmap is statically allocated so we don't have to use + * pmap_create, which is unlikely to work correctly at this part of + * the boot sequence (XXX and which no longer exists). + */ + kernel_pmap = &kernel_pmap_store; + + kernel_pmap->pm_pdir = (pd_entry_t *) (KERNBASE + IdlePTD); + + kernel_pmap->pm_count = 1; + nkpt = NKPT; + + /* + * Reserve some special page table entries/VA space for temporary + * mapping of pages. + */ +#define SYSMAP(c, p, v, n) \ + v = (c)va; va += ((n)*PAGE_SIZE); p = pte; pte += (n); + + va = virtual_avail; + pte = (pt_entry_t *) pmap_pte(kernel_pmap, va); + + /* + * CMAP1/CMAP2 are used for zeroing and copying pages. + */ + SYSMAP(caddr_t, CMAP1, CADDR1, 1) + SYSMAP(caddr_t, CMAP2, CADDR2, 1) + + /* + * ptmmap is used for reading arbitrary physical pages via /dev/mem. + */ + SYSMAP(caddr_t, ptmmap, ptvmmap, 1) + + /* + * msgbufmap is used to map the system message buffer. + */ + SYSMAP(struct msgbuf *, msgbufmap, msgbufp, 1) + + /* + * ptemap is used for pmap_pte_quick + */ + SYSMAP(unsigned *, PMAP1, PADDR1, 1); + + virtual_avail = va; + + *(int *) CMAP1 = *(int *) CMAP2 = *(int *) PTD = 0; + pmap_update(); + +} + +/* + * Initialize the pmap module. + * Called by vm_init, to initialize any structures that the pmap + * system needs to map virtual memory. + * pmap_init has been enhanced to support in a fairly consistant + * way, discontiguous physical memory. + */ +void +pmap_init(phys_start, phys_end) + vm_offset_t phys_start, phys_end; +{ + vm_offset_t addr; + vm_size_t npg, s; + int i; + + /* + * calculate the number of pv_entries needed + */ + vm_first_phys = phys_avail[0]; + for (i = 0; phys_avail[i + 1]; i += 2); + npg = (phys_avail[(i - 2) + 1] - vm_first_phys) / PAGE_SIZE; + + /* + * Allocate memory for random pmap data structures. Includes the + * pv_head_table. + */ + s = (vm_size_t) (sizeof(struct pv_entry *) * npg); + s = round_page(s); + addr = (vm_offset_t) kmem_alloc(kernel_map, s); + pv_table = (pv_entry_t *) addr; + + /* + * init the pv free list + */ + init_pv_entries(npg); + /* + * Now it is safe to enable pv_table recording. + */ + pmap_initialized = TRUE; +} + +/* + * Used to map a range of physical addresses into kernel + * virtual address space. + * + * For now, VM is already on, we only need to map the + * specified memory. + */ +vm_offset_t +pmap_map(virt, start, end, prot) + vm_offset_t virt; + vm_offset_t start; + vm_offset_t end; + int prot; +{ + while (start < end) { + pmap_enter(kernel_pmap, virt, start, prot, FALSE); + virt += PAGE_SIZE; + start += PAGE_SIZE; + } + return (virt); +} + + +/*************************************************** + * Low level helper routines..... + ***************************************************/ + #if defined(PMAP_DIAGNOSTIC) /* @@ -220,6 +369,19 @@ pmap_nw_modified(pt_entry_t ptea) { } #endif + +/* + * this routine defines the region(s) of memory that should + * not be tested for the modified bit. + */ +static __inline int +pmap_track_modified( vm_offset_t va) { + if ((va < clean_sva) || (va >= clean_eva)) + return 1; + else + return 0; +} + /* * The below are finer grained pmap_update routines. These eliminate * the gratuitious tlb flushes on non-i386 architectures. @@ -247,7 +409,7 @@ pmap_update_2pg( vm_offset_t va1, vm_offset_t va2) { } } -static __inline __pure unsigned * +static __pure unsigned * get_ptbase(pmap) pmap_t pmap; { @@ -284,6 +446,31 @@ pmap_pte(pmap, va) } /* + * Super fast pmap_pte routine best used when scanning + * the pv lists. This eliminates many coarse-grained + * pmap_update calls. + */ +__inline unsigned * __pure +pmap_pte_quick(pmap, va) + register pmap_t pmap; + vm_offset_t va; +{ + unsigned pde; + if (pde = (unsigned) pmap->pm_pdir[va >> PDRSHIFT]) { + unsigned frame = (unsigned) pmap->pm_pdir[PTDPTDI] & PG_FRAME; + /* are we current address space or kernel? */ + if (pmap == kernel_pmap || frame == (((unsigned) PTDpde) & PG_FRAME)) { + return (unsigned *) PTmap + i386_btop(va); + } + * (int *) PMAP1 = (pde & PG_FRAME) | PG_V | PG_RW; + pmap_update_1pg((vm_offset_t) PADDR1); + return PADDR1 + ((unsigned) i386_btop(va) & (NPTEPG - 1)); + } + return (0); +} + + +/* * Routine: pmap_extract * Function: * Extract the physical page address associated @@ -304,6 +491,29 @@ pmap_extract(pmap, va) } /* + * determine if a page is managed (memory vs. device) + */ +static __inline __pure int +pmap_is_managed(pa) + vm_offset_t pa; +{ + int i; + + if (!pmap_initialized) + return 0; + + for (i = 0; phys_avail[i + 1]; i += 2) { + if (pa < phys_avail[i + 1] && pa >= phys_avail[i]) + return 1; + } + return 0; +} + +/*************************************************** + * Low level mapping routines..... + ***************************************************/ + +/* * Add a list of wired pages to the kva * this routine is only used for temporary * kernel mappings that do not need to have @@ -386,47 +596,47 @@ pmap_kremove(va) pmap_update_1pg(va); } -/* - * determine if a page is managed (memory vs. device) - */ -static __inline __pure int -pmap_is_managed(pa) - vm_offset_t pa; -{ - int i; - - if (!pmap_initialized) - return 0; - for (i = 0; phys_avail[i + 1]; i += 2) { - if (pa < phys_avail[i + 1] && pa >= phys_avail[i]) - return 1; - } - return 0; -} +/*************************************************** + * Page table page management routines..... + ***************************************************/ /* * This routine unholds page table pages, and if the hold count * drops to zero, then it decrements the wire count. */ static __inline int -pmap_unwire_pte_hold(vm_page_t m) { +pmap_unwire_pte_hold(pmap_t pmap, vm_page_t m) { vm_page_unhold(m); if (m->hold_count == 0) { + vm_offset_t pteva; + /* + * unmap the page table page + */ + pmap->pm_pdir[m->pindex] = 0; + /* + * Do a pmap_update to make the invalidated mapping + * take effect immediately. + */ + pteva = UPT_MIN_ADDRESS + i386_ptob(m->pindex); + pmap_update_1pg(pteva); + /* + * If the page is finally unwired, simply free it. + */ --m->wire_count; if (m->wire_count == 0) { + vm_page_free_zero(m); --cnt.v_wire_count; - m->dirty = 0; - vm_page_deactivate(m); - return 1; } + return 1; } return 0; } -#if !defined(PMAP_DIAGNOSTIC) -__inline -#endif +/* + * After removing a page table entry, this routine is used to + * conditionally free the page, and manage the hold/wire counts. + */ int pmap_unuse_pt(pmap, va, mpte) pmap_t pmap; @@ -438,172 +648,27 @@ pmap_unuse_pt(pmap, va, mpte) if (mpte == NULL) { vm_offset_t ptepa; - ptepa = ((vm_offset_t) *pmap_pde(pmap, va)) /* & PG_FRAME */; + ptepa = ((vm_offset_t) *pmap_pde(pmap, va)); #if defined(PMAP_DIAGNOSTIC) if (!ptepa) panic("pmap_unuse_pt: pagetable page missing, va: 0x%x", va); #endif + if (!ptepa) + return 0; mpte = PHYS_TO_VM_PAGE(ptepa); } #if defined(PMAP_DIAGNOSTIC) + if (mpte->pindex != (va >> PDRSHIFT)) + panic("pmap_unuse_pt: pindex(0x%x) != va(0x%x)", + mpte->pindex, (va >> PDRSHIFT)); + if (mpte->hold_count == 0) { panic("pmap_unuse_pt: hold count < 0, va: 0x%x", va); } #endif -/* - * We don't free page-table-pages anymore because it can have a negative - * impact on perf at times. Now we just deactivate, and it'll get cleaned - * up if needed... Also, if the page ends up getting used, it will be - * brought back into the process address space by pmap_allocpte and be - * reactivated. - */ - return pmap_unwire_pte_hold(mpte); -} - -/* - * Bootstrap the system enough to run with virtual memory. - * - * On the i386 this is called after mapping has already been enabled - * and just syncs the pmap module with what has already been done. - * [We can't call it easily with mapping off since the kernel is not - * mapped with PA == VA, hence we would have to relocate every address - * from the linked base (virtual) address "KERNBASE" to the actual - * (physical) address starting relative to 0] - */ -void -pmap_bootstrap(firstaddr, loadaddr) - vm_offset_t firstaddr; - vm_offset_t loadaddr; -{ - vm_offset_t va; - pt_entry_t *pte; - - avail_start = firstaddr; - - /* - * XXX The calculation of virtual_avail is wrong. It's NKPT*PAGE_SIZE too - * large. It should instead be correctly calculated in locore.s and - * not based on 'first' (which is a physical address, not a virtual - * address, for the start of unused physical memory). The kernel - * page tables are NOT double mapped and thus should not be included - * in this calculation. - */ - virtual_avail = (vm_offset_t) KERNBASE + firstaddr; - virtual_end = VM_MAX_KERNEL_ADDRESS; - - /* - * Initialize protection array. - */ - i386_protection_init(); - - /* - * The kernel's pmap is statically allocated so we don't have to use - * pmap_create, which is unlikely to work correctly at this part of - * the boot sequence (XXX and which no longer exists). - */ - kernel_pmap = &kernel_pmap_store; - - kernel_pmap->pm_pdir = (pd_entry_t *) (KERNBASE + IdlePTD); - - kernel_pmap->pm_count = 1; - nkpt = NKPT; - - /* - * Reserve some special page table entries/VA space for temporary - * mapping of pages. - */ -#define SYSMAP(c, p, v, n) \ - v = (c)va; va += ((n)*PAGE_SIZE); p = pte; pte += (n); - - va = virtual_avail; - pte = (pt_entry_t *) pmap_pte(kernel_pmap, va); - - /* - * CMAP1/CMAP2 are used for zeroing and copying pages. - */ - SYSMAP(caddr_t, CMAP1, CADDR1, 1) - SYSMAP(caddr_t, CMAP2, CADDR2, 1) - - /* - * ptmmap is used for reading arbitrary physical pages via /dev/mem. - */ - SYSMAP(caddr_t, ptmmap, ptvmmap, 1) - - /* - * msgbufmap is used to map the system message buffer. - */ - SYSMAP(struct msgbuf *, msgbufmap, msgbufp, 1) - - virtual_avail = va; - - *(int *) CMAP1 = *(int *) CMAP2 = *(int *) PTD = 0; - pmap_update(); - -} - -/* - * Initialize the pmap module. - * Called by vm_init, to initialize any structures that the pmap - * system needs to map virtual memory. - * pmap_init has been enhanced to support in a fairly consistant - * way, discontiguous physical memory. - */ -void -pmap_init(phys_start, phys_end) - vm_offset_t phys_start, phys_end; -{ - vm_offset_t addr; - vm_size_t npg, s; - int i; - - /* - * calculate the number of pv_entries needed - */ - vm_first_phys = phys_avail[0]; - for (i = 0; phys_avail[i + 1]; i += 2); - npg = (phys_avail[(i - 2) + 1] - vm_first_phys) / PAGE_SIZE; - - /* - * Allocate memory for random pmap data structures. Includes the - * pv_head_table. - */ - s = (vm_size_t) (sizeof(struct pv_entry *) * npg); - s = round_page(s); - addr = (vm_offset_t) kmem_alloc(kernel_map, s); - pv_table = (pv_entry_t *) addr; - - /* - * init the pv free list - */ - init_pv_entries(npg); - /* - * Now it is safe to enable pv_table recording. - */ - pmap_initialized = TRUE; -} - -/* - * Used to map a range of physical addresses into kernel - * virtual address space. - * - * For now, VM is already on, we only need to map the - * specified memory. - */ -vm_offset_t -pmap_map(virt, start, end, prot) - vm_offset_t virt; - vm_offset_t start; - vm_offset_t end; - int prot; -{ - while (start < end) { - pmap_enter(kernel_pmap, virt, start, prot, FALSE); - virt += PAGE_SIZE; - start += PAGE_SIZE; - } - return (virt); + return pmap_unwire_pte_hold(pmap, mpte); } /* @@ -622,8 +687,7 @@ pmap_pinit(pmap) if (pdstackptr > 0) { --pdstackptr; - pmap->pm_pdir = - (pd_entry_t *)pdstack[pdstackptr]; + pmap->pm_pdir = (pd_entry_t *)pdstack[pdstackptr]; } else { pmap->pm_pdir = (pd_entry_t *)kmem_alloc_pageable(kernel_map, PAGE_SIZE); @@ -667,6 +731,7 @@ pmap_release_free_page(pmap, p) vm_page_t p; { int s; + unsigned *pde = (unsigned *) pmap->pm_pdir; /* * This code optimizes the case of freeing non-busy * page-table pages. Those pages are zero now, and @@ -680,12 +745,27 @@ pmap_release_free_page(pmap, p) return 0; } - pmap_remove_pte_mapping(VM_PAGE_TO_PHYS(p)); + /* + * Remove the page table page from the processes address space. + */ + pde[p->pindex] = 0; if (p->hold_count) { + int *kvap; + int i; #if defined(PMAP_DIAGNOSTIC) panic("pmap_release: freeing held page table page"); +#else + printf("pmap_release: freeing held page table page:\n"); #endif + kvap = (int *)vm_pager_map_page(p); + for(i=0;i<NPTEPG;i++) { + if (kvap[i]) { + printf("pte: 0x%x, index: %d\n", kvap[i],i); + } + } + vm_pager_unmap_page((vm_offset_t)kvap); + /* * HACK ALERT!!! * If this failure happens, we must clear the page, because @@ -702,10 +782,8 @@ pmap_release_free_page(pmap, p) * stuff cleared, so they can go into the zero queue also. */ if (p->pindex == PTDPTDI) { - unsigned *pde = (unsigned *) pmap->pm_pdir; bzero(pde + KPTDI, nkpt * PTESIZE); pde[APTDPTDI] = 0; - pde[PTDPTDI] = 0; pmap_kremove((vm_offset_t) pmap->pm_pdir); } @@ -715,6 +793,119 @@ pmap_release_free_page(pmap, p) } /* + * this routine is called if the page table page is not + * mapped correctly. + */ +static vm_page_t +_pmap_allocpte(pmap, ptepindex) + pmap_t pmap; + int ptepindex; +{ + vm_offset_t pteva, ptepa; + vm_page_t m; + + /* + * Find or fabricate a new pagetable page + */ +retry: + m = vm_page_lookup(pmap->pm_pteobj, ptepindex); + if (m == NULL) { + m = vm_page_alloc(pmap->pm_pteobj, ptepindex, VM_ALLOC_ZERO); + if (m == NULL) { + VM_WAIT; + goto retry; + } + if ((m->flags & PG_ZERO) == 0) + pmap_zero_page(VM_PAGE_TO_PHYS(m)); + m->flags &= ~(PG_ZERO|PG_BUSY); + m->valid = VM_PAGE_BITS_ALL; + } else { + if ((m->flags & PG_BUSY) || m->busy) { + m->flags |= PG_WANTED; + tsleep(m, PVM, "ptewai", 0); + goto retry; + } + } + + /* + * mark the object writeable + */ + pmap->pm_pteobj->flags |= OBJ_WRITEABLE; + + if (m->queue != PQ_NONE) { + int s = splvm(); + vm_page_unqueue(m); + splx(s); + } + + if (m->hold_count == 0) { + if (m->wire_count == 0) + ++cnt.v_wire_count; + ++m->wire_count; + } + /* + * Increment the hold count for the page table page + * (denoting a new mapping.) + */ + ++m->hold_count; + + /* + * Map the pagetable page into the process address space, if + * it isn't already there. + */ + + pmap->pm_stats.resident_count++; + + ptepa = VM_PAGE_TO_PHYS(m); + pmap->pm_pdir[ptepindex] = (pd_entry_t) (ptepa | PG_U | PG_RW | PG_V); + + pteva = UPT_MIN_ADDRESS + i386_ptob(ptepindex); + pmap_update_1pg(pteva); + m->flags |= PG_MAPPED; + + return m; +} + +static __inline vm_page_t +pmap_allocpte(pmap, va) + pmap_t pmap; + vm_offset_t va; +{ + int ptepindex; + vm_offset_t ptepa; + vm_page_t m; + + /* + * Calculate pagetable page index + */ + ptepindex = va >> PDRSHIFT; + + /* + * Get the page directory entry + */ + ptepa = (vm_offset_t) pmap->pm_pdir[ptepindex]; + + /* + * If the page table page is mapped, we just increment the + * hold count, and activate it. + */ + if (ptepa) { + m = PHYS_TO_VM_PAGE(ptepa); + ++m->hold_count; + return m; + } + /* + * Here if the pte page isn't mapped, or if it has been deallocated. + */ + return _pmap_allocpte(pmap, ptepindex); +} + + +/*************************************************** +* Pmap allocation/deallocation routines. + ***************************************************/ + +/* * Release any resources held by the given physical map. * Called when a pmap initialized by pmap_pinit is being released. * Should only be called if the map contains no valid mappings. @@ -745,7 +936,7 @@ retry: if (!pmap_release_free_page(pmap, ptdpg)) goto retry; - + vm_object_deallocate(object); if (pdstackptr < PDSTACKMAX) { pdstack[pdstackptr] = (vm_offset_t) pmap->pm_pdir; @@ -758,7 +949,6 @@ retry: /* * grow the number of kernel page table entries, if needed */ - void pmap_growkernel(vm_offset_t addr) { @@ -838,6 +1028,10 @@ pmap_reference(pmap) } } +/*************************************************** +* page management routines. + ***************************************************/ + /* * free the pv_entry back to the free list */ @@ -985,13 +1179,39 @@ pmap_remove_entry(pmap, ppv, va) } /* + * Create a pv entry for page at pa for + * (pmap, va). + */ +static __inline void +pmap_insert_entry(pmap, va, mpte, pa) + pmap_t pmap; + vm_offset_t va; + vm_page_t mpte; + vm_offset_t pa; +{ + + int s; + pv_entry_t *ppv, pv; + + s = splvm(); + pv = get_pv_entry(); + pv->pv_va = va; + pv->pv_pmap = pmap; + pv->pv_ptem = mpte; + + ppv = pa_to_pvh(pa); + if (*ppv) + pv->pv_next = *ppv; + else + pv->pv_next = NULL; + *ppv = pv; + splx(s); +} + +/* * pmap_remove_pte: do the things to unmap a page in a process */ -static -#if !defined(PMAP_DIAGNOSTIC) -__inline -#endif -int +static int pmap_remove_pte(pmap, ptq, va) struct pmap *pmap; unsigned *ptq; @@ -999,7 +1219,6 @@ pmap_remove_pte(pmap, ptq, va) { unsigned oldpte; pv_entry_t *ppv; - int rtval; oldpte = *ptq; *ptq = 0; @@ -1013,20 +1232,11 @@ pmap_remove_pte(pmap, ptq, va) printf("pmap_remove: modified page not writable: va: 0x%lx, pte: 0x%lx\n", va, (int) oldpte); } #endif - - if (va < clean_sva || va >= clean_eva) { - if ((va < UPT_MIN_ADDRESS) || (va >= UPT_MAX_ADDRESS)) - PHYS_TO_VM_PAGE(oldpte)->dirty = VM_PAGE_BITS_ALL; - } + if (pmap_track_modified(va)) + PHYS_TO_VM_PAGE(oldpte)->dirty = VM_PAGE_BITS_ALL; } ppv = pa_to_pvh(oldpte); - rtval = pmap_remove_entry(pmap, ppv, va); -#if defined(notyet) - if (*ppv == NULL) { - PHYS_TO_VM_PAGE(oldpte)->flags &= ~PG_MAPPED; - } -#endif - return rtval; + return pmap_remove_entry(pmap, ppv, va); } else { return pmap_unuse_pt(pmap, va, NULL); } @@ -1037,7 +1247,7 @@ pmap_remove_pte(pmap, ptq, va) /* * Remove a single page from a process address space */ -static __inline void +static void pmap_remove_page(pmap, va) struct pmap *pmap; register vm_offset_t va; @@ -1080,6 +1290,7 @@ pmap_remove(pmap, sva, eva) vm_offset_t sindex, eindex; vm_page_t mpte; int anyvalid; + vm_offset_t vachanged[VATRACK]; if (pmap == NULL) return; @@ -1148,7 +1359,10 @@ pmap_remove(pmap, sva, eva) continue; } va = i386_ptob(sindex); - anyvalid = 1; + + if (anyvalid < VATRACK) + vachanged[anyvalid] = va; + anyvalid++; if (pmap_remove_pte(pmap, ptbase + sindex, va)) break; @@ -1156,43 +1370,15 @@ pmap_remove(pmap, sva, eva) } if (anyvalid) { - pmap_update(); - } -} - -/* - * Remove pte mapping, don't do everything that we would do - * for normal pages because many things aren't necessary (like - * pmap_update())... - */ -void -pmap_remove_pte_mapping(pa) - vm_offset_t pa; -{ - register pv_entry_t pv, *ppv, npv; - register unsigned *pte; - - ppv = pa_to_pvh(pa); - - for (pv = *ppv; pv; pv=pv->pv_next) { - unsigned tpte; - struct pmap *pmap; - - pmap = pv->pv_pmap; - pte = get_ptbase(pmap) + i386_btop(pv->pv_va); - if (tpte = *pte) { - pmap->pm_stats.resident_count--; - *pte = 0; - if (tpte & PG_W) - pmap->pm_stats.wired_count--; + if (anyvalid <= VATRACK) { + int i; + for(i=0;i<anyvalid;i++) + pmap_update_1pg(vachanged[i]); + } else { + pmap_update(); } } - - for (pv = *ppv; pv; pv = npv) { - npv = pv->pv_next; - free_pv_entry(pv); - } - *ppv = NULL; + } /* @@ -1212,8 +1398,7 @@ pmap_remove_all(pa) vm_offset_t pa; { register pv_entry_t pv, *ppv, npv; - register unsigned *pte, *ptbase; - vm_offset_t va; + register unsigned *pte, tpte; vm_page_t m; int s; @@ -1227,39 +1412,28 @@ pmap_remove_all(pa) } #endif - m = PHYS_TO_VM_PAGE(pa); - ppv = pa_to_pvh(pa); - s = splvm(); + m = NULL; + ppv = pa_to_pvh(pa); for (pv = *ppv; pv; pv=pv->pv_next) { - int tpte; - struct pmap *pmap; - - pmap = pv->pv_pmap; - ptbase = get_ptbase(pmap); - va = pv->pv_va; - if (*pmap_pde(pmap, va) == 0) - continue; - pte = ptbase + i386_btop(va); - if (tpte = ((int) *pte)) { - pmap->pm_stats.resident_count--; + pte = pmap_pte_quick(pv->pv_pmap, pv->pv_va); + if (tpte = *pte) { + pv->pv_pmap->pm_stats.resident_count--; *pte = 0; if (tpte & PG_W) - pmap->pm_stats.wired_count--; + pv->pv_pmap->pm_stats.wired_count--; /* * Update the vm_page_t clean and reference bits. */ if (tpte & PG_M) { #if defined(PMAP_DIAGNOSTIC) if (pmap_nw_modified((pt_entry_t) tpte)) { - printf("pmap_remove_all: modified page not writable: va: 0x%lx, pte: 0x%lx\n", va, tpte); + printf("pmap_remove_all: modified page not writable: va: 0x%lx, pte: 0x%lx\n", pv->pv_va, tpte); } #endif - if ((va >= UPT_MIN_ADDRESS) && - (va < UPT_MAX_ADDRESS)) - continue; - - if (va < clean_sva || va >= clean_eva) { + if (pmap_track_modified(pv->pv_va)) { + if (m == NULL) + m = PHYS_TO_VM_PAGE(pa); m->dirty = VM_PAGE_BITS_ALL; } } @@ -1272,7 +1446,6 @@ pmap_remove_all(pa) free_pv_entry(pv); } *ppv = NULL; - splx(s); } @@ -1324,7 +1497,9 @@ pmap_protect(pmap, sva, eva, prot) continue; /* - * Don't look at kernel page table pages + * Skip page ranges, where the page table page isn't wired. + * If the page table page is not wired, there are no page mappings + * there. */ if (sindex < i386_btop(UPT_MIN_ADDRESS)) { mpte = PHYS_TO_VM_PAGE(ptpaddr); @@ -1343,8 +1518,11 @@ pmap_protect(pmap, sva, eva, prot) if (pbits & PG_RW) { if (pbits & PG_M) { - vm_page_t m = PHYS_TO_VM_PAGE(pbits); - m->dirty = VM_PAGE_BITS_ALL; + vm_offset_t sva = i386_ptob(sindex); + if (pmap_track_modified(sva)) { + vm_page_t m = PHYS_TO_VM_PAGE(pbits); + m->dirty = VM_PAGE_BITS_ALL; + } } ptbase[sindex] = pbits & ~(PG_M|PG_RW); anyvalid = 1; @@ -1356,172 +1534,6 @@ pmap_protect(pmap, sva, eva, prot) } /* - * Create a pv entry for page at pa for - * (pmap, va). - */ -static __inline void -pmap_insert_entry(pmap, va, mpte, pa) - pmap_t pmap; - vm_offset_t va; - vm_page_t mpte; - vm_offset_t pa; -{ - - int s; - pv_entry_t *ppv, pv; - - s = splvm(); - pv = get_pv_entry(); - pv->pv_va = va; - pv->pv_pmap = pmap; - pv->pv_ptem = mpte; - - ppv = pa_to_pvh(pa); - if (*ppv) - pv->pv_next = *ppv; - else - pv->pv_next = NULL; - *ppv = pv; - splx(s); -} - -/* - * this routine is called if the page table page is not - * mapped correctly. - */ -static vm_page_t -_pmap_allocpte(pmap, va, ptepindex) - pmap_t pmap; - vm_offset_t va; - int ptepindex; -{ - vm_offset_t pteva, ptepa; - vm_page_t m; - int s; - - /* - * Find or fabricate a new pagetable page - */ -retry: - m = vm_page_lookup(pmap->pm_pteobj, ptepindex); - if (m == NULL) { - m = vm_page_alloc(pmap->pm_pteobj, ptepindex, VM_ALLOC_ZERO); - if (m == NULL) { - VM_WAIT; - goto retry; - } - if ((m->flags & PG_ZERO) == 0) - pmap_zero_page(VM_PAGE_TO_PHYS(m)); - m->flags &= ~(PG_ZERO|PG_BUSY); - m->valid = VM_PAGE_BITS_ALL; - } else { - if ((m->flags & PG_BUSY) || m->busy) { - m->flags |= PG_WANTED; - tsleep(m, PVM, "ptewai", 0); - goto retry; - } - } - - /* - * mark the object writeable - */ - pmap->pm_pteobj->flags |= OBJ_WRITEABLE; - - if (m->hold_count == 0) { - s = splvm(); - vm_page_unqueue(m); - splx(s); - if (m->wire_count == 0) - ++cnt.v_wire_count; - ++m->wire_count; - } - - /* - * Increment the hold count for the page table page - * (denoting a new mapping.) - */ - ++m->hold_count; - - /* - * Map the pagetable page into the process address space, if - * it isn't already there. - */ - pteva = ((vm_offset_t) vtopte(va)) & PG_FRAME; - ptepa = (vm_offset_t) pmap->pm_pdir[ptepindex]; - if (ptepa == 0) { - pv_entry_t pv, *ppv; - - pmap->pm_stats.resident_count++; - - s = splvm(); - pv = get_pv_entry(); - - pv->pv_va = pteva; - pv->pv_pmap = pmap; - pv->pv_next = NULL; - pv->pv_ptem = NULL; - - ptepa = VM_PAGE_TO_PHYS(m); - ppv = pa_to_pvh(ptepa); -#if defined(PMAP_DIAGNOSTIC) - if (*ppv) - panic("pmap_allocpte: page is already mapped"); -#endif - *ppv = pv; - splx(s); - pmap_update_1pg(pteva); - } else { -#if defined(PMAP_DIAGNOSTIC) - if (VM_PAGE_TO_PHYS(m) != (ptepa & PG_FRAME)) - panic("pmap_allocpte: mismatch"); -#endif - } - pmap->pm_pdir[ptepindex] = - (pd_entry_t) (ptepa | PG_U | PG_RW | PG_V | PG_MANAGED); - m->flags |= PG_MAPPED; - return m; -} - -static __inline vm_page_t -pmap_allocpte(pmap, va) - pmap_t pmap; - vm_offset_t va; -{ - int ptepindex; - vm_offset_t ptepa; - vm_page_t m; - - /* - * Calculate pagetable page index - */ - ptepindex = va >> PDRSHIFT; - - /* - * Get the page directory entry - */ - ptepa = (vm_offset_t) pmap->pm_pdir[ptepindex]; - - /* - * If the page table page is mapped, we just increment the - * hold count, and activate it. - */ - if ((ptepa & (PG_RW|PG_U|PG_V)) == (PG_RW|PG_U|PG_V)) { - m = PHYS_TO_VM_PAGE(ptepa); - if (m->hold_count == 0) { - int s = splvm(); - vm_page_unqueue(m); - splx(s); - if (m->wire_count == 0) - ++cnt.v_wire_count; - ++m->wire_count; - } - ++m->hold_count; - return m; - } - return _pmap_allocpte(pmap, va, ptepindex); -} - -/* * Insert the given physical page (p) at * the specified virtual address (v) in the * target physical map with the protection requested. @@ -1550,8 +1562,12 @@ pmap_enter(pmap, va, pa, prot, wired) return; va &= PG_FRAME; +#ifdef PMAP_DIAGNOSTIC if (va > VM_MAX_KERNEL_ADDRESS) panic("pmap_enter: toobig"); + if ((va >= UPT_MIN_ADDRESS) && (va < UPT_MAX_ADDRESS)) + panic("pmap_enter: invalid to pmap_enter page table pages (va: 0x%x)", va); +#endif mpte = NULL; /* @@ -1561,14 +1577,13 @@ pmap_enter(pmap, va, pa, prot, wired) if (va < UPT_MIN_ADDRESS) mpte = pmap_allocpte(pmap, va); - pte = pmap_pte(pmap, va); + pte = pmap_pte_quick(pmap, va); /* * Page Directory table entry not valid, we need a new PT page */ if (pte == NULL) { - printf("kernel page directory invalid pdir=%p, va=0x%lx\n", + panic("pmap_enter: invalid page directory, pdir=%p, va=0x%lx\n", pmap->pm_pdir[PTDPTDI], va); - panic("invalid kernel page directory"); } origpte = *(vm_offset_t *)pte; @@ -1603,8 +1618,10 @@ pmap_enter(pmap, va, pa, prot, wired) if (origpte & PG_MANAGED) { vm_page_t m; if (origpte & PG_M) { - m = PHYS_TO_VM_PAGE(pa); - m->dirty = VM_PAGE_BITS_ALL; + if (pmap_track_modified(va)) { + m = PHYS_TO_VM_PAGE(pa); + m->dirty = VM_PAGE_BITS_ALL; + } } pa |= PG_MANAGED; } @@ -1613,13 +1630,17 @@ pmap_enter(pmap, va, pa, prot, wired) --mpte->hold_count; goto validate; - } + } /* * Mapping has changed, invalidate old range and fall through to * handle validating new mapping. */ - if (opa) - (void) pmap_remove_pte(pmap, pte, va); + if (opa) { + int err; + err = pmap_remove_pte(pmap, pte, va); + if (err) + panic("pmap_enter: pte vanished, va: 0x%x", va); + } /* * Enter on the PV list if part of our managed memory Note that we @@ -1688,9 +1709,18 @@ pmap_enter_quick(pmap, va, pa) if (va < UPT_MIN_ADDRESS) mpte = pmap_allocpte(pmap, va); + /* + * This call to vtopte makes the assumption that we are + * entering the page into the current pmap. In order to support + * quick entry into any pmap, one would likely use pmap_pte_quick. + * But that isn't as quick as vtopte. + */ pte = (unsigned *)vtopte(va); - if (*pte) - (void) pmap_remove_pte(pmap, pte, va); + if (*pte) { + if (mpte) + pmap_unwire_pte_hold(pmap, mpte); + return; + } /* * Enter on the PV list if part of our managed memory Note that we @@ -1732,7 +1762,7 @@ pmap_object_init_pt(pmap, addr, object, pindex, size, limit) vm_page_t p; int objpgs; - psize = (size >> PAGE_SHIFT); + psize = i386_btop(size); if (!pmap || (object->type != OBJT_VNODE) || (limit && (psize > MAX_INIT_PT) && @@ -1766,7 +1796,7 @@ pmap_object_init_pt(pmap, addr, object, pindex, size, limit) vm_page_deactivate(p); p->flags |= PG_BUSY; pmap_enter_quick(pmap, - addr + (tmpidx << PAGE_SHIFT), + addr + i386_ptob(tmpidx), VM_PAGE_TO_PHYS(p)); p->flags |= PG_MAPPED; PAGE_WAKEUP(p); @@ -1787,7 +1817,7 @@ pmap_object_init_pt(pmap, addr, object, pindex, size, limit) vm_page_deactivate(p); p->flags |= PG_BUSY; pmap_enter_quick(pmap, - addr + (tmpidx << PAGE_SHIFT), + addr + i386_ptob(tmpidx), VM_PAGE_TO_PHYS(p)); p->flags |= PG_MAPPED; PAGE_WAKEUP(p); @@ -1955,6 +1985,8 @@ pmap_copy(dst_pmap, src_pmap, dst_addr, len, src_addr) vm_page_t dstmpte, srcmpte; vm_offset_t srcptepaddr; + if (addr >= UPT_MIN_ADDRESS) + panic("pmap_copy: invalid to pmap_copy page tables\n"); pdnxt = ((addr + PAGE_SIZE*NPTEPG) & ~(PAGE_SIZE*NPTEPG - 1)); srcptepaddr = (vm_offset_t) src_pmap->pm_pdir[addr >> PDRSHIFT]; if (srcptepaddr == 0) { @@ -1983,7 +2015,7 @@ pmap_copy(dst_pmap, src_pmap, dst_addr, len, src_addr) * block. */ dstmpte = pmap_allocpte(dst_pmap, addr); - if (ptetemp = *src_pte) { + if ((*dst_pte == 0) && (ptetemp = *src_pte)) { /* * Simply clear the modified and accessed (referenced) * bits. @@ -1993,7 +2025,7 @@ pmap_copy(dst_pmap, src_pmap, dst_addr, len, src_addr) pmap_insert_entry(dst_pmap, addr, dstmpte, (ptetemp & PG_FRAME)); } else { - pmap_unwire_pte_hold(dstmpte); + pmap_unwire_pte_hold(dst_pmap, dstmpte); } if (dstmpte->hold_count >= srcmpte->hold_count) break; @@ -2133,38 +2165,36 @@ pmap_testbit(pa, bit) if (!pmap_is_managed(pa)) return FALSE; - s = splvm(); - ppv = pa_to_pvh(pa); + if (*ppv == NULL) + return FALSE; + + s = splvm(); /* * Not found, check current mappings returning immediately if found. */ for (pv = *ppv ;pv; pv = pv->pv_next) { + /* * if the bit being tested is the modified bit, then - * mark UPAGES as always modified, and ptes as never + * mark clean_map and ptes as never * modified. */ if (bit & (PG_A|PG_M)) { - if ((pv->pv_va >= UPT_MIN_ADDRESS) && - (pv->pv_va < UPT_MAX_ADDRESS)) { + if (!pmap_track_modified(pv->pv_va)) continue; - } - if ((pv->pv_va >= clean_sva) && - (pv->pv_va < clean_eva)) { - continue; - } } + if (!pv->pv_pmap) { #if defined(PMAP_DIAGNOSTIC) printf("Null pmap (tb) at va: 0x%lx\n", pv->pv_va); #endif continue; } - pte = pmap_pte(pv->pv_pmap, pv->pv_va); + pte = pmap_pte_quick(pv->pv_pmap, pv->pv_va); if (pte == NULL) continue; - if ((int) *pte & bit) { + if (*pte & bit) { splx(s); return TRUE; } @@ -2192,7 +2222,6 @@ pmap_changebit(pa, bit, setem) return; s = splvm(); - changed = 0; ppv = pa_to_pvh(pa); /* @@ -2216,7 +2245,7 @@ pmap_changebit(pa, bit, setem) continue; } - pte = pmap_pte(pv->pv_pmap, va); + pte = pmap_pte_quick(pv->pv_pmap, va); if (pte == NULL) continue; if (setem) { @@ -2255,9 +2284,9 @@ pmap_page_protect(phys, prot) vm_prot_t prot; { if ((prot & VM_PROT_WRITE) == 0) { - if (prot & (VM_PROT_READ | VM_PROT_EXECUTE)) + if (prot & (VM_PROT_READ | VM_PROT_EXECUTE)) { pmap_changebit(phys, PG_RW, FALSE); - else { + } else { pmap_remove_all(phys); pmap_update(); } @@ -2280,7 +2309,110 @@ pmap_phys_address(ppn) boolean_t pmap_is_referenced(vm_offset_t pa) { - return pmap_testbit((pa), PG_A); + register pv_entry_t *ppv, pv, lpv; + unsigned *pte; + int s; + + if (!pmap_is_managed(pa)) + return FALSE; + + ppv = pa_to_pvh(pa); + + s = splvm(); + /* + * Not found, check current mappings returning immediately if found. + */ + for (lpv = NULL, pv = *ppv ;pv; lpv = pv, pv = pv->pv_next) { + /* + * if the bit being tested is the modified bit, then + * mark clean_map and ptes as never + * modified. + */ + if (!pmap_track_modified(pv->pv_va)) + continue; + if (!pv->pv_pmap) { + continue; + } + pte = pmap_pte_quick(pv->pv_pmap, pv->pv_va); + if (pte == NULL) + continue; + if ((int) *pte & PG_A) { + if (lpv) { + lpv->pv_next = pv->pv_next; + pv->pv_next = *ppv; + *ppv = pv; + } + splx(s); + return TRUE; + } + } + splx(s); + return (FALSE); +} + +/* + * pmap_ts_referenced: + * + * Return the count of reference bits for a page, clearing all of them. + * + */ +int +pmap_ts_referenced(vm_offset_t pa) +{ + register pv_entry_t *ppv, pv; + unsigned *pte; + int s; + int rtval = 0; + vm_offset_t vachanged[VATRACK]; + + if (!pmap_is_managed(pa)) + return FALSE; + + s = splvm(); + + ppv = pa_to_pvh(pa); + + if (*ppv == NULL) { + splx(s); + return 0; + } + + /* + * Not found, check current mappings returning immediately if found. + */ + for (pv = *ppv ;pv; pv = pv->pv_next) { + /* + * if the bit being tested is the modified bit, then + * mark clean_map and ptes as never + * modified. + */ + if (!pmap_track_modified(pv->pv_va)) + continue; + + if (!pv->pv_pmap) { + continue; + } + pte = pmap_pte_quick(pv->pv_pmap, pv->pv_va); + if (pte == NULL) + continue; + if (*pte & PG_A) { + if (rtval < VATRACK) + vachanged[rtval] = pv->pv_va; + rtval++; + *pte &= ~PG_A; + } + } + splx(s); + if (rtval) { + if (rtval <= VATRACK) { + int i; + for(i=0;i<rtval;i++) + pmap_update_1pg(vachanged[i]); + } else { + pmap_update(); + } + } + return (rtval); } /* @@ -2381,6 +2513,9 @@ pmap_mapdev(pa, size) return ((void *) va); } +/* + * perform the pmap work for mincore + */ int pmap_mincore(pmap, addr) pmap_t pmap; @@ -2395,7 +2530,7 @@ pmap_mincore(pmap, addr) return 0; } - if ((pte = *ptep)) { + if (pte = *ptep) { vm_offset_t pa; val = MINCORE_INCORE; pa = pte & PG_FRAME; diff --git a/sys/i386/i386/pmap.c b/sys/i386/i386/pmap.c index 345b80c..abdb456 100644 --- a/sys/i386/i386/pmap.c +++ b/sys/i386/i386/pmap.c @@ -39,17 +39,7 @@ * SUCH DAMAGE. * * from: @(#)pmap.c 7.7 (Berkeley) 5/12/91 - * $Id: pmap.c,v 1.103 1996/06/12 05:02:52 gpalmer Exp $ - */ - -/* - * Derived from hp300 version by Mike Hibler, this version by William - * Jolitz uses a recursive map [a pde points to the page directory] to - * map the page tables using the pagetables themselves. This is done to - * reduce the impact on kernel virtual memory for lots of sparse address - * space, and to reduce the cost of memory to each process. - * - * Derived from: hp300/@(#)pmap.c 7.1 (Berkeley) 12/5/90 + * $Id: pmap.c,v 1.104 1996/06/12 06:52:12 dyson Exp $ */ /* @@ -97,6 +87,7 @@ #include <vm/vm_object.h> #include <vm/vm_extern.h> #include <vm/vm_pageout.h> +#include <vm/vm_pager.h> #include <machine/pcb.h> #include <machine/cputypes.h> @@ -172,35 +163,193 @@ static caddr_t CADDR2; static pt_entry_t *msgbufmap; struct msgbuf *msgbufp; +pt_entry_t *PMAP1; +unsigned *PADDR1; + static void free_pv_entry __P((pv_entry_t pv)); static __inline unsigned * get_ptbase __P((pmap_t pmap)); static pv_entry_t get_pv_entry __P((void)); static void i386_protection_init __P((void)); static void pmap_alloc_pv_entry __P((void)); static void pmap_changebit __P((vm_offset_t pa, int bit, boolean_t setem)); -static void pmap_enter_quick __P((pmap_t pmap, vm_offset_t va, - vm_offset_t pa)); + static int pmap_is_managed __P((vm_offset_t pa)); static void pmap_remove_all __P((vm_offset_t pa)); +static void pmap_enter_quick __P((pmap_t pmap, vm_offset_t va, + vm_offset_t pa)); +static int pmap_remove_pte __P((struct pmap *pmap, unsigned *ptq, + vm_offset_t sva)); static void pmap_remove_page __P((struct pmap *pmap, vm_offset_t va)); static __inline int pmap_remove_entry __P((struct pmap *pmap, pv_entry_t *pv, vm_offset_t va)); -static int pmap_remove_pte __P((struct pmap *pmap, unsigned *ptq, - vm_offset_t sva)); -static boolean_t - pmap_testbit __P((vm_offset_t pa, int bit)); +static boolean_t pmap_testbit __P((vm_offset_t pa, int bit)); static __inline void pmap_insert_entry __P((pmap_t pmap, vm_offset_t va, vm_page_t mpte, vm_offset_t pa)); static __inline vm_page_t pmap_allocpte __P((pmap_t pmap, vm_offset_t va)); -static void pmap_remove_pte_mapping __P((vm_offset_t pa)); + static __inline int pmap_release_free_page __P((pmap_t pmap, vm_page_t p)); -static vm_page_t _pmap_allocpte __P((pmap_t pmap, vm_offset_t va, int ptepindex)); +static vm_page_t _pmap_allocpte __P((pmap_t pmap, int ptepindex)); +#define VATRACK 4 #define PDSTACKMAX 16 static vm_offset_t pdstack[PDSTACKMAX]; static int pdstackptr; +/* + * Bootstrap the system enough to run with virtual memory. + * + * On the i386 this is called after mapping has already been enabled + * and just syncs the pmap module with what has already been done. + * [We can't call it easily with mapping off since the kernel is not + * mapped with PA == VA, hence we would have to relocate every address + * from the linked base (virtual) address "KERNBASE" to the actual + * (physical) address starting relative to 0] + */ +void +pmap_bootstrap(firstaddr, loadaddr) + vm_offset_t firstaddr; + vm_offset_t loadaddr; +{ + vm_offset_t va; + pt_entry_t *pte; + + avail_start = firstaddr; + + /* + * XXX The calculation of virtual_avail is wrong. It's NKPT*PAGE_SIZE too + * large. It should instead be correctly calculated in locore.s and + * not based on 'first' (which is a physical address, not a virtual + * address, for the start of unused physical memory). The kernel + * page tables are NOT double mapped and thus should not be included + * in this calculation. + */ + virtual_avail = (vm_offset_t) KERNBASE + firstaddr; + virtual_end = VM_MAX_KERNEL_ADDRESS; + + /* + * Initialize protection array. + */ + i386_protection_init(); + + /* + * The kernel's pmap is statically allocated so we don't have to use + * pmap_create, which is unlikely to work correctly at this part of + * the boot sequence (XXX and which no longer exists). + */ + kernel_pmap = &kernel_pmap_store; + + kernel_pmap->pm_pdir = (pd_entry_t *) (KERNBASE + IdlePTD); + + kernel_pmap->pm_count = 1; + nkpt = NKPT; + + /* + * Reserve some special page table entries/VA space for temporary + * mapping of pages. + */ +#define SYSMAP(c, p, v, n) \ + v = (c)va; va += ((n)*PAGE_SIZE); p = pte; pte += (n); + + va = virtual_avail; + pte = (pt_entry_t *) pmap_pte(kernel_pmap, va); + + /* + * CMAP1/CMAP2 are used for zeroing and copying pages. + */ + SYSMAP(caddr_t, CMAP1, CADDR1, 1) + SYSMAP(caddr_t, CMAP2, CADDR2, 1) + + /* + * ptmmap is used for reading arbitrary physical pages via /dev/mem. + */ + SYSMAP(caddr_t, ptmmap, ptvmmap, 1) + + /* + * msgbufmap is used to map the system message buffer. + */ + SYSMAP(struct msgbuf *, msgbufmap, msgbufp, 1) + + /* + * ptemap is used for pmap_pte_quick + */ + SYSMAP(unsigned *, PMAP1, PADDR1, 1); + + virtual_avail = va; + + *(int *) CMAP1 = *(int *) CMAP2 = *(int *) PTD = 0; + pmap_update(); + +} + +/* + * Initialize the pmap module. + * Called by vm_init, to initialize any structures that the pmap + * system needs to map virtual memory. + * pmap_init has been enhanced to support in a fairly consistant + * way, discontiguous physical memory. + */ +void +pmap_init(phys_start, phys_end) + vm_offset_t phys_start, phys_end; +{ + vm_offset_t addr; + vm_size_t npg, s; + int i; + + /* + * calculate the number of pv_entries needed + */ + vm_first_phys = phys_avail[0]; + for (i = 0; phys_avail[i + 1]; i += 2); + npg = (phys_avail[(i - 2) + 1] - vm_first_phys) / PAGE_SIZE; + + /* + * Allocate memory for random pmap data structures. Includes the + * pv_head_table. + */ + s = (vm_size_t) (sizeof(struct pv_entry *) * npg); + s = round_page(s); + addr = (vm_offset_t) kmem_alloc(kernel_map, s); + pv_table = (pv_entry_t *) addr; + + /* + * init the pv free list + */ + init_pv_entries(npg); + /* + * Now it is safe to enable pv_table recording. + */ + pmap_initialized = TRUE; +} + +/* + * Used to map a range of physical addresses into kernel + * virtual address space. + * + * For now, VM is already on, we only need to map the + * specified memory. + */ +vm_offset_t +pmap_map(virt, start, end, prot) + vm_offset_t virt; + vm_offset_t start; + vm_offset_t end; + int prot; +{ + while (start < end) { + pmap_enter(kernel_pmap, virt, start, prot, FALSE); + virt += PAGE_SIZE; + start += PAGE_SIZE; + } + return (virt); +} + + +/*************************************************** + * Low level helper routines..... + ***************************************************/ + #if defined(PMAP_DIAGNOSTIC) /* @@ -220,6 +369,19 @@ pmap_nw_modified(pt_entry_t ptea) { } #endif + +/* + * this routine defines the region(s) of memory that should + * not be tested for the modified bit. + */ +static __inline int +pmap_track_modified( vm_offset_t va) { + if ((va < clean_sva) || (va >= clean_eva)) + return 1; + else + return 0; +} + /* * The below are finer grained pmap_update routines. These eliminate * the gratuitious tlb flushes on non-i386 architectures. @@ -247,7 +409,7 @@ pmap_update_2pg( vm_offset_t va1, vm_offset_t va2) { } } -static __inline __pure unsigned * +static __pure unsigned * get_ptbase(pmap) pmap_t pmap; { @@ -284,6 +446,31 @@ pmap_pte(pmap, va) } /* + * Super fast pmap_pte routine best used when scanning + * the pv lists. This eliminates many coarse-grained + * pmap_update calls. + */ +__inline unsigned * __pure +pmap_pte_quick(pmap, va) + register pmap_t pmap; + vm_offset_t va; +{ + unsigned pde; + if (pde = (unsigned) pmap->pm_pdir[va >> PDRSHIFT]) { + unsigned frame = (unsigned) pmap->pm_pdir[PTDPTDI] & PG_FRAME; + /* are we current address space or kernel? */ + if (pmap == kernel_pmap || frame == (((unsigned) PTDpde) & PG_FRAME)) { + return (unsigned *) PTmap + i386_btop(va); + } + * (int *) PMAP1 = (pde & PG_FRAME) | PG_V | PG_RW; + pmap_update_1pg((vm_offset_t) PADDR1); + return PADDR1 + ((unsigned) i386_btop(va) & (NPTEPG - 1)); + } + return (0); +} + + +/* * Routine: pmap_extract * Function: * Extract the physical page address associated @@ -304,6 +491,29 @@ pmap_extract(pmap, va) } /* + * determine if a page is managed (memory vs. device) + */ +static __inline __pure int +pmap_is_managed(pa) + vm_offset_t pa; +{ + int i; + + if (!pmap_initialized) + return 0; + + for (i = 0; phys_avail[i + 1]; i += 2) { + if (pa < phys_avail[i + 1] && pa >= phys_avail[i]) + return 1; + } + return 0; +} + +/*************************************************** + * Low level mapping routines..... + ***************************************************/ + +/* * Add a list of wired pages to the kva * this routine is only used for temporary * kernel mappings that do not need to have @@ -386,47 +596,47 @@ pmap_kremove(va) pmap_update_1pg(va); } -/* - * determine if a page is managed (memory vs. device) - */ -static __inline __pure int -pmap_is_managed(pa) - vm_offset_t pa; -{ - int i; - - if (!pmap_initialized) - return 0; - for (i = 0; phys_avail[i + 1]; i += 2) { - if (pa < phys_avail[i + 1] && pa >= phys_avail[i]) - return 1; - } - return 0; -} +/*************************************************** + * Page table page management routines..... + ***************************************************/ /* * This routine unholds page table pages, and if the hold count * drops to zero, then it decrements the wire count. */ static __inline int -pmap_unwire_pte_hold(vm_page_t m) { +pmap_unwire_pte_hold(pmap_t pmap, vm_page_t m) { vm_page_unhold(m); if (m->hold_count == 0) { + vm_offset_t pteva; + /* + * unmap the page table page + */ + pmap->pm_pdir[m->pindex] = 0; + /* + * Do a pmap_update to make the invalidated mapping + * take effect immediately. + */ + pteva = UPT_MIN_ADDRESS + i386_ptob(m->pindex); + pmap_update_1pg(pteva); + /* + * If the page is finally unwired, simply free it. + */ --m->wire_count; if (m->wire_count == 0) { + vm_page_free_zero(m); --cnt.v_wire_count; - m->dirty = 0; - vm_page_deactivate(m); - return 1; } + return 1; } return 0; } -#if !defined(PMAP_DIAGNOSTIC) -__inline -#endif +/* + * After removing a page table entry, this routine is used to + * conditionally free the page, and manage the hold/wire counts. + */ int pmap_unuse_pt(pmap, va, mpte) pmap_t pmap; @@ -438,172 +648,27 @@ pmap_unuse_pt(pmap, va, mpte) if (mpte == NULL) { vm_offset_t ptepa; - ptepa = ((vm_offset_t) *pmap_pde(pmap, va)) /* & PG_FRAME */; + ptepa = ((vm_offset_t) *pmap_pde(pmap, va)); #if defined(PMAP_DIAGNOSTIC) if (!ptepa) panic("pmap_unuse_pt: pagetable page missing, va: 0x%x", va); #endif + if (!ptepa) + return 0; mpte = PHYS_TO_VM_PAGE(ptepa); } #if defined(PMAP_DIAGNOSTIC) + if (mpte->pindex != (va >> PDRSHIFT)) + panic("pmap_unuse_pt: pindex(0x%x) != va(0x%x)", + mpte->pindex, (va >> PDRSHIFT)); + if (mpte->hold_count == 0) { panic("pmap_unuse_pt: hold count < 0, va: 0x%x", va); } #endif -/* - * We don't free page-table-pages anymore because it can have a negative - * impact on perf at times. Now we just deactivate, and it'll get cleaned - * up if needed... Also, if the page ends up getting used, it will be - * brought back into the process address space by pmap_allocpte and be - * reactivated. - */ - return pmap_unwire_pte_hold(mpte); -} - -/* - * Bootstrap the system enough to run with virtual memory. - * - * On the i386 this is called after mapping has already been enabled - * and just syncs the pmap module with what has already been done. - * [We can't call it easily with mapping off since the kernel is not - * mapped with PA == VA, hence we would have to relocate every address - * from the linked base (virtual) address "KERNBASE" to the actual - * (physical) address starting relative to 0] - */ -void -pmap_bootstrap(firstaddr, loadaddr) - vm_offset_t firstaddr; - vm_offset_t loadaddr; -{ - vm_offset_t va; - pt_entry_t *pte; - - avail_start = firstaddr; - - /* - * XXX The calculation of virtual_avail is wrong. It's NKPT*PAGE_SIZE too - * large. It should instead be correctly calculated in locore.s and - * not based on 'first' (which is a physical address, not a virtual - * address, for the start of unused physical memory). The kernel - * page tables are NOT double mapped and thus should not be included - * in this calculation. - */ - virtual_avail = (vm_offset_t) KERNBASE + firstaddr; - virtual_end = VM_MAX_KERNEL_ADDRESS; - - /* - * Initialize protection array. - */ - i386_protection_init(); - - /* - * The kernel's pmap is statically allocated so we don't have to use - * pmap_create, which is unlikely to work correctly at this part of - * the boot sequence (XXX and which no longer exists). - */ - kernel_pmap = &kernel_pmap_store; - - kernel_pmap->pm_pdir = (pd_entry_t *) (KERNBASE + IdlePTD); - - kernel_pmap->pm_count = 1; - nkpt = NKPT; - - /* - * Reserve some special page table entries/VA space for temporary - * mapping of pages. - */ -#define SYSMAP(c, p, v, n) \ - v = (c)va; va += ((n)*PAGE_SIZE); p = pte; pte += (n); - - va = virtual_avail; - pte = (pt_entry_t *) pmap_pte(kernel_pmap, va); - - /* - * CMAP1/CMAP2 are used for zeroing and copying pages. - */ - SYSMAP(caddr_t, CMAP1, CADDR1, 1) - SYSMAP(caddr_t, CMAP2, CADDR2, 1) - - /* - * ptmmap is used for reading arbitrary physical pages via /dev/mem. - */ - SYSMAP(caddr_t, ptmmap, ptvmmap, 1) - - /* - * msgbufmap is used to map the system message buffer. - */ - SYSMAP(struct msgbuf *, msgbufmap, msgbufp, 1) - - virtual_avail = va; - - *(int *) CMAP1 = *(int *) CMAP2 = *(int *) PTD = 0; - pmap_update(); - -} - -/* - * Initialize the pmap module. - * Called by vm_init, to initialize any structures that the pmap - * system needs to map virtual memory. - * pmap_init has been enhanced to support in a fairly consistant - * way, discontiguous physical memory. - */ -void -pmap_init(phys_start, phys_end) - vm_offset_t phys_start, phys_end; -{ - vm_offset_t addr; - vm_size_t npg, s; - int i; - - /* - * calculate the number of pv_entries needed - */ - vm_first_phys = phys_avail[0]; - for (i = 0; phys_avail[i + 1]; i += 2); - npg = (phys_avail[(i - 2) + 1] - vm_first_phys) / PAGE_SIZE; - - /* - * Allocate memory for random pmap data structures. Includes the - * pv_head_table. - */ - s = (vm_size_t) (sizeof(struct pv_entry *) * npg); - s = round_page(s); - addr = (vm_offset_t) kmem_alloc(kernel_map, s); - pv_table = (pv_entry_t *) addr; - - /* - * init the pv free list - */ - init_pv_entries(npg); - /* - * Now it is safe to enable pv_table recording. - */ - pmap_initialized = TRUE; -} - -/* - * Used to map a range of physical addresses into kernel - * virtual address space. - * - * For now, VM is already on, we only need to map the - * specified memory. - */ -vm_offset_t -pmap_map(virt, start, end, prot) - vm_offset_t virt; - vm_offset_t start; - vm_offset_t end; - int prot; -{ - while (start < end) { - pmap_enter(kernel_pmap, virt, start, prot, FALSE); - virt += PAGE_SIZE; - start += PAGE_SIZE; - } - return (virt); + return pmap_unwire_pte_hold(pmap, mpte); } /* @@ -622,8 +687,7 @@ pmap_pinit(pmap) if (pdstackptr > 0) { --pdstackptr; - pmap->pm_pdir = - (pd_entry_t *)pdstack[pdstackptr]; + pmap->pm_pdir = (pd_entry_t *)pdstack[pdstackptr]; } else { pmap->pm_pdir = (pd_entry_t *)kmem_alloc_pageable(kernel_map, PAGE_SIZE); @@ -667,6 +731,7 @@ pmap_release_free_page(pmap, p) vm_page_t p; { int s; + unsigned *pde = (unsigned *) pmap->pm_pdir; /* * This code optimizes the case of freeing non-busy * page-table pages. Those pages are zero now, and @@ -680,12 +745,27 @@ pmap_release_free_page(pmap, p) return 0; } - pmap_remove_pte_mapping(VM_PAGE_TO_PHYS(p)); + /* + * Remove the page table page from the processes address space. + */ + pde[p->pindex] = 0; if (p->hold_count) { + int *kvap; + int i; #if defined(PMAP_DIAGNOSTIC) panic("pmap_release: freeing held page table page"); +#else + printf("pmap_release: freeing held page table page:\n"); #endif + kvap = (int *)vm_pager_map_page(p); + for(i=0;i<NPTEPG;i++) { + if (kvap[i]) { + printf("pte: 0x%x, index: %d\n", kvap[i],i); + } + } + vm_pager_unmap_page((vm_offset_t)kvap); + /* * HACK ALERT!!! * If this failure happens, we must clear the page, because @@ -702,10 +782,8 @@ pmap_release_free_page(pmap, p) * stuff cleared, so they can go into the zero queue also. */ if (p->pindex == PTDPTDI) { - unsigned *pde = (unsigned *) pmap->pm_pdir; bzero(pde + KPTDI, nkpt * PTESIZE); pde[APTDPTDI] = 0; - pde[PTDPTDI] = 0; pmap_kremove((vm_offset_t) pmap->pm_pdir); } @@ -715,6 +793,119 @@ pmap_release_free_page(pmap, p) } /* + * this routine is called if the page table page is not + * mapped correctly. + */ +static vm_page_t +_pmap_allocpte(pmap, ptepindex) + pmap_t pmap; + int ptepindex; +{ + vm_offset_t pteva, ptepa; + vm_page_t m; + + /* + * Find or fabricate a new pagetable page + */ +retry: + m = vm_page_lookup(pmap->pm_pteobj, ptepindex); + if (m == NULL) { + m = vm_page_alloc(pmap->pm_pteobj, ptepindex, VM_ALLOC_ZERO); + if (m == NULL) { + VM_WAIT; + goto retry; + } + if ((m->flags & PG_ZERO) == 0) + pmap_zero_page(VM_PAGE_TO_PHYS(m)); + m->flags &= ~(PG_ZERO|PG_BUSY); + m->valid = VM_PAGE_BITS_ALL; + } else { + if ((m->flags & PG_BUSY) || m->busy) { + m->flags |= PG_WANTED; + tsleep(m, PVM, "ptewai", 0); + goto retry; + } + } + + /* + * mark the object writeable + */ + pmap->pm_pteobj->flags |= OBJ_WRITEABLE; + + if (m->queue != PQ_NONE) { + int s = splvm(); + vm_page_unqueue(m); + splx(s); + } + + if (m->hold_count == 0) { + if (m->wire_count == 0) + ++cnt.v_wire_count; + ++m->wire_count; + } + /* + * Increment the hold count for the page table page + * (denoting a new mapping.) + */ + ++m->hold_count; + + /* + * Map the pagetable page into the process address space, if + * it isn't already there. + */ + + pmap->pm_stats.resident_count++; + + ptepa = VM_PAGE_TO_PHYS(m); + pmap->pm_pdir[ptepindex] = (pd_entry_t) (ptepa | PG_U | PG_RW | PG_V); + + pteva = UPT_MIN_ADDRESS + i386_ptob(ptepindex); + pmap_update_1pg(pteva); + m->flags |= PG_MAPPED; + + return m; +} + +static __inline vm_page_t +pmap_allocpte(pmap, va) + pmap_t pmap; + vm_offset_t va; +{ + int ptepindex; + vm_offset_t ptepa; + vm_page_t m; + + /* + * Calculate pagetable page index + */ + ptepindex = va >> PDRSHIFT; + + /* + * Get the page directory entry + */ + ptepa = (vm_offset_t) pmap->pm_pdir[ptepindex]; + + /* + * If the page table page is mapped, we just increment the + * hold count, and activate it. + */ + if (ptepa) { + m = PHYS_TO_VM_PAGE(ptepa); + ++m->hold_count; + return m; + } + /* + * Here if the pte page isn't mapped, or if it has been deallocated. + */ + return _pmap_allocpte(pmap, ptepindex); +} + + +/*************************************************** +* Pmap allocation/deallocation routines. + ***************************************************/ + +/* * Release any resources held by the given physical map. * Called when a pmap initialized by pmap_pinit is being released. * Should only be called if the map contains no valid mappings. @@ -745,7 +936,7 @@ retry: if (!pmap_release_free_page(pmap, ptdpg)) goto retry; - + vm_object_deallocate(object); if (pdstackptr < PDSTACKMAX) { pdstack[pdstackptr] = (vm_offset_t) pmap->pm_pdir; @@ -758,7 +949,6 @@ retry: /* * grow the number of kernel page table entries, if needed */ - void pmap_growkernel(vm_offset_t addr) { @@ -838,6 +1028,10 @@ pmap_reference(pmap) } } +/*************************************************** +* page management routines. + ***************************************************/ + /* * free the pv_entry back to the free list */ @@ -985,13 +1179,39 @@ pmap_remove_entry(pmap, ppv, va) } /* + * Create a pv entry for page at pa for + * (pmap, va). + */ +static __inline void +pmap_insert_entry(pmap, va, mpte, pa) + pmap_t pmap; + vm_offset_t va; + vm_page_t mpte; + vm_offset_t pa; +{ + + int s; + pv_entry_t *ppv, pv; + + s = splvm(); + pv = get_pv_entry(); + pv->pv_va = va; + pv->pv_pmap = pmap; + pv->pv_ptem = mpte; + + ppv = pa_to_pvh(pa); + if (*ppv) + pv->pv_next = *ppv; + else + pv->pv_next = NULL; + *ppv = pv; + splx(s); +} + +/* * pmap_remove_pte: do the things to unmap a page in a process */ -static -#if !defined(PMAP_DIAGNOSTIC) -__inline -#endif -int +static int pmap_remove_pte(pmap, ptq, va) struct pmap *pmap; unsigned *ptq; @@ -999,7 +1219,6 @@ pmap_remove_pte(pmap, ptq, va) { unsigned oldpte; pv_entry_t *ppv; - int rtval; oldpte = *ptq; *ptq = 0; @@ -1013,20 +1232,11 @@ pmap_remove_pte(pmap, ptq, va) printf("pmap_remove: modified page not writable: va: 0x%lx, pte: 0x%lx\n", va, (int) oldpte); } #endif - - if (va < clean_sva || va >= clean_eva) { - if ((va < UPT_MIN_ADDRESS) || (va >= UPT_MAX_ADDRESS)) - PHYS_TO_VM_PAGE(oldpte)->dirty = VM_PAGE_BITS_ALL; - } + if (pmap_track_modified(va)) + PHYS_TO_VM_PAGE(oldpte)->dirty = VM_PAGE_BITS_ALL; } ppv = pa_to_pvh(oldpte); - rtval = pmap_remove_entry(pmap, ppv, va); -#if defined(notyet) - if (*ppv == NULL) { - PHYS_TO_VM_PAGE(oldpte)->flags &= ~PG_MAPPED; - } -#endif - return rtval; + return pmap_remove_entry(pmap, ppv, va); } else { return pmap_unuse_pt(pmap, va, NULL); } @@ -1037,7 +1247,7 @@ pmap_remove_pte(pmap, ptq, va) /* * Remove a single page from a process address space */ -static __inline void +static void pmap_remove_page(pmap, va) struct pmap *pmap; register vm_offset_t va; @@ -1080,6 +1290,7 @@ pmap_remove(pmap, sva, eva) vm_offset_t sindex, eindex; vm_page_t mpte; int anyvalid; + vm_offset_t vachanged[VATRACK]; if (pmap == NULL) return; @@ -1148,7 +1359,10 @@ pmap_remove(pmap, sva, eva) continue; } va = i386_ptob(sindex); - anyvalid = 1; + + if (anyvalid < VATRACK) + vachanged[anyvalid] = va; + anyvalid++; if (pmap_remove_pte(pmap, ptbase + sindex, va)) break; @@ -1156,43 +1370,15 @@ pmap_remove(pmap, sva, eva) } if (anyvalid) { - pmap_update(); - } -} - -/* - * Remove pte mapping, don't do everything that we would do - * for normal pages because many things aren't necessary (like - * pmap_update())... - */ -void -pmap_remove_pte_mapping(pa) - vm_offset_t pa; -{ - register pv_entry_t pv, *ppv, npv; - register unsigned *pte; - - ppv = pa_to_pvh(pa); - - for (pv = *ppv; pv; pv=pv->pv_next) { - unsigned tpte; - struct pmap *pmap; - - pmap = pv->pv_pmap; - pte = get_ptbase(pmap) + i386_btop(pv->pv_va); - if (tpte = *pte) { - pmap->pm_stats.resident_count--; - *pte = 0; - if (tpte & PG_W) - pmap->pm_stats.wired_count--; + if (anyvalid <= VATRACK) { + int i; + for(i=0;i<anyvalid;i++) + pmap_update_1pg(vachanged[i]); + } else { + pmap_update(); } } - - for (pv = *ppv; pv; pv = npv) { - npv = pv->pv_next; - free_pv_entry(pv); - } - *ppv = NULL; + } /* @@ -1212,8 +1398,7 @@ pmap_remove_all(pa) vm_offset_t pa; { register pv_entry_t pv, *ppv, npv; - register unsigned *pte, *ptbase; - vm_offset_t va; + register unsigned *pte, tpte; vm_page_t m; int s; @@ -1227,39 +1412,28 @@ pmap_remove_all(pa) } #endif - m = PHYS_TO_VM_PAGE(pa); - ppv = pa_to_pvh(pa); - s = splvm(); + m = NULL; + ppv = pa_to_pvh(pa); for (pv = *ppv; pv; pv=pv->pv_next) { - int tpte; - struct pmap *pmap; - - pmap = pv->pv_pmap; - ptbase = get_ptbase(pmap); - va = pv->pv_va; - if (*pmap_pde(pmap, va) == 0) - continue; - pte = ptbase + i386_btop(va); - if (tpte = ((int) *pte)) { - pmap->pm_stats.resident_count--; + pte = pmap_pte_quick(pv->pv_pmap, pv->pv_va); + if (tpte = *pte) { + pv->pv_pmap->pm_stats.resident_count--; *pte = 0; if (tpte & PG_W) - pmap->pm_stats.wired_count--; + pv->pv_pmap->pm_stats.wired_count--; /* * Update the vm_page_t clean and reference bits. */ if (tpte & PG_M) { #if defined(PMAP_DIAGNOSTIC) if (pmap_nw_modified((pt_entry_t) tpte)) { - printf("pmap_remove_all: modified page not writable: va: 0x%lx, pte: 0x%lx\n", va, tpte); + printf("pmap_remove_all: modified page not writable: va: 0x%lx, pte: 0x%lx\n", pv->pv_va, tpte); } #endif - if ((va >= UPT_MIN_ADDRESS) && - (va < UPT_MAX_ADDRESS)) - continue; - - if (va < clean_sva || va >= clean_eva) { + if (pmap_track_modified(pv->pv_va)) { + if (m == NULL) + m = PHYS_TO_VM_PAGE(pa); m->dirty = VM_PAGE_BITS_ALL; } } @@ -1272,7 +1446,6 @@ pmap_remove_all(pa) free_pv_entry(pv); } *ppv = NULL; - splx(s); } @@ -1324,7 +1497,9 @@ pmap_protect(pmap, sva, eva, prot) continue; /* - * Don't look at kernel page table pages + * Skip page ranges, where the page table page isn't wired. + * If the page table page is not wired, there are no page mappings + * there. */ if (sindex < i386_btop(UPT_MIN_ADDRESS)) { mpte = PHYS_TO_VM_PAGE(ptpaddr); @@ -1343,8 +1518,11 @@ pmap_protect(pmap, sva, eva, prot) if (pbits & PG_RW) { if (pbits & PG_M) { - vm_page_t m = PHYS_TO_VM_PAGE(pbits); - m->dirty = VM_PAGE_BITS_ALL; + vm_offset_t sva = i386_ptob(sindex); + if (pmap_track_modified(sva)) { + vm_page_t m = PHYS_TO_VM_PAGE(pbits); + m->dirty = VM_PAGE_BITS_ALL; + } } ptbase[sindex] = pbits & ~(PG_M|PG_RW); anyvalid = 1; @@ -1356,172 +1534,6 @@ pmap_protect(pmap, sva, eva, prot) } /* - * Create a pv entry for page at pa for - * (pmap, va). - */ -static __inline void -pmap_insert_entry(pmap, va, mpte, pa) - pmap_t pmap; - vm_offset_t va; - vm_page_t mpte; - vm_offset_t pa; -{ - - int s; - pv_entry_t *ppv, pv; - - s = splvm(); - pv = get_pv_entry(); - pv->pv_va = va; - pv->pv_pmap = pmap; - pv->pv_ptem = mpte; - - ppv = pa_to_pvh(pa); - if (*ppv) - pv->pv_next = *ppv; - else - pv->pv_next = NULL; - *ppv = pv; - splx(s); -} - -/* - * this routine is called if the page table page is not - * mapped correctly. - */ -static vm_page_t -_pmap_allocpte(pmap, va, ptepindex) - pmap_t pmap; - vm_offset_t va; - int ptepindex; -{ - vm_offset_t pteva, ptepa; - vm_page_t m; - int s; - - /* - * Find or fabricate a new pagetable page - */ -retry: - m = vm_page_lookup(pmap->pm_pteobj, ptepindex); - if (m == NULL) { - m = vm_page_alloc(pmap->pm_pteobj, ptepindex, VM_ALLOC_ZERO); - if (m == NULL) { - VM_WAIT; - goto retry; - } - if ((m->flags & PG_ZERO) == 0) - pmap_zero_page(VM_PAGE_TO_PHYS(m)); - m->flags &= ~(PG_ZERO|PG_BUSY); - m->valid = VM_PAGE_BITS_ALL; - } else { - if ((m->flags & PG_BUSY) || m->busy) { - m->flags |= PG_WANTED; - tsleep(m, PVM, "ptewai", 0); - goto retry; - } - } - - /* - * mark the object writeable - */ - pmap->pm_pteobj->flags |= OBJ_WRITEABLE; - - if (m->hold_count == 0) { - s = splvm(); - vm_page_unqueue(m); - splx(s); - if (m->wire_count == 0) - ++cnt.v_wire_count; - ++m->wire_count; - } - - /* - * Increment the hold count for the page table page - * (denoting a new mapping.) - */ - ++m->hold_count; - - /* - * Map the pagetable page into the process address space, if - * it isn't already there. - */ - pteva = ((vm_offset_t) vtopte(va)) & PG_FRAME; - ptepa = (vm_offset_t) pmap->pm_pdir[ptepindex]; - if (ptepa == 0) { - pv_entry_t pv, *ppv; - - pmap->pm_stats.resident_count++; - - s = splvm(); - pv = get_pv_entry(); - - pv->pv_va = pteva; - pv->pv_pmap = pmap; - pv->pv_next = NULL; - pv->pv_ptem = NULL; - - ptepa = VM_PAGE_TO_PHYS(m); - ppv = pa_to_pvh(ptepa); -#if defined(PMAP_DIAGNOSTIC) - if (*ppv) - panic("pmap_allocpte: page is already mapped"); -#endif - *ppv = pv; - splx(s); - pmap_update_1pg(pteva); - } else { -#if defined(PMAP_DIAGNOSTIC) - if (VM_PAGE_TO_PHYS(m) != (ptepa & PG_FRAME)) - panic("pmap_allocpte: mismatch"); -#endif - } - pmap->pm_pdir[ptepindex] = - (pd_entry_t) (ptepa | PG_U | PG_RW | PG_V | PG_MANAGED); - m->flags |= PG_MAPPED; - return m; -} - -static __inline vm_page_t -pmap_allocpte(pmap, va) - pmap_t pmap; - vm_offset_t va; -{ - int ptepindex; - vm_offset_t ptepa; - vm_page_t m; - - /* - * Calculate pagetable page index - */ - ptepindex = va >> PDRSHIFT; - - /* - * Get the page directory entry - */ - ptepa = (vm_offset_t) pmap->pm_pdir[ptepindex]; - - /* - * If the page table page is mapped, we just increment the - * hold count, and activate it. - */ - if ((ptepa & (PG_RW|PG_U|PG_V)) == (PG_RW|PG_U|PG_V)) { - m = PHYS_TO_VM_PAGE(ptepa); - if (m->hold_count == 0) { - int s = splvm(); - vm_page_unqueue(m); - splx(s); - if (m->wire_count == 0) - ++cnt.v_wire_count; - ++m->wire_count; - } - ++m->hold_count; - return m; - } - return _pmap_allocpte(pmap, va, ptepindex); -} - -/* * Insert the given physical page (p) at * the specified virtual address (v) in the * target physical map with the protection requested. @@ -1550,8 +1562,12 @@ pmap_enter(pmap, va, pa, prot, wired) return; va &= PG_FRAME; +#ifdef PMAP_DIAGNOSTIC if (va > VM_MAX_KERNEL_ADDRESS) panic("pmap_enter: toobig"); + if ((va >= UPT_MIN_ADDRESS) && (va < UPT_MAX_ADDRESS)) + panic("pmap_enter: invalid to pmap_enter page table pages (va: 0x%x)", va); +#endif mpte = NULL; /* @@ -1561,14 +1577,13 @@ pmap_enter(pmap, va, pa, prot, wired) if (va < UPT_MIN_ADDRESS) mpte = pmap_allocpte(pmap, va); - pte = pmap_pte(pmap, va); + pte = pmap_pte_quick(pmap, va); /* * Page Directory table entry not valid, we need a new PT page */ if (pte == NULL) { - printf("kernel page directory invalid pdir=%p, va=0x%lx\n", + panic("pmap_enter: invalid page directory, pdir=%p, va=0x%lx\n", pmap->pm_pdir[PTDPTDI], va); - panic("invalid kernel page directory"); } origpte = *(vm_offset_t *)pte; @@ -1603,8 +1618,10 @@ pmap_enter(pmap, va, pa, prot, wired) if (origpte & PG_MANAGED) { vm_page_t m; if (origpte & PG_M) { - m = PHYS_TO_VM_PAGE(pa); - m->dirty = VM_PAGE_BITS_ALL; + if (pmap_track_modified(va)) { + m = PHYS_TO_VM_PAGE(pa); + m->dirty = VM_PAGE_BITS_ALL; + } } pa |= PG_MANAGED; } @@ -1613,13 +1630,17 @@ pmap_enter(pmap, va, pa, prot, wired) --mpte->hold_count; goto validate; - } + } /* * Mapping has changed, invalidate old range and fall through to * handle validating new mapping. */ - if (opa) - (void) pmap_remove_pte(pmap, pte, va); + if (opa) { + int err; + err = pmap_remove_pte(pmap, pte, va); + if (err) + panic("pmap_enter: pte vanished, va: 0x%x", va); + } /* * Enter on the PV list if part of our managed memory Note that we @@ -1688,9 +1709,18 @@ pmap_enter_quick(pmap, va, pa) if (va < UPT_MIN_ADDRESS) mpte = pmap_allocpte(pmap, va); + /* + * This call to vtopte makes the assumption that we are + * entering the page into the current pmap. In order to support + * quick entry into any pmap, one would likely use pmap_pte_quick. + * But that isn't as quick as vtopte. + */ pte = (unsigned *)vtopte(va); - if (*pte) - (void) pmap_remove_pte(pmap, pte, va); + if (*pte) { + if (mpte) + pmap_unwire_pte_hold(pmap, mpte); + return; + } /* * Enter on the PV list if part of our managed memory Note that we @@ -1732,7 +1762,7 @@ pmap_object_init_pt(pmap, addr, object, pindex, size, limit) vm_page_t p; int objpgs; - psize = (size >> PAGE_SHIFT); + psize = i386_btop(size); if (!pmap || (object->type != OBJT_VNODE) || (limit && (psize > MAX_INIT_PT) && @@ -1766,7 +1796,7 @@ pmap_object_init_pt(pmap, addr, object, pindex, size, limit) vm_page_deactivate(p); p->flags |= PG_BUSY; pmap_enter_quick(pmap, - addr + (tmpidx << PAGE_SHIFT), + addr + i386_ptob(tmpidx), VM_PAGE_TO_PHYS(p)); p->flags |= PG_MAPPED; PAGE_WAKEUP(p); @@ -1787,7 +1817,7 @@ pmap_object_init_pt(pmap, addr, object, pindex, size, limit) vm_page_deactivate(p); p->flags |= PG_BUSY; pmap_enter_quick(pmap, - addr + (tmpidx << PAGE_SHIFT), + addr + i386_ptob(tmpidx), VM_PAGE_TO_PHYS(p)); p->flags |= PG_MAPPED; PAGE_WAKEUP(p); @@ -1955,6 +1985,8 @@ pmap_copy(dst_pmap, src_pmap, dst_addr, len, src_addr) vm_page_t dstmpte, srcmpte; vm_offset_t srcptepaddr; + if (addr >= UPT_MIN_ADDRESS) + panic("pmap_copy: invalid to pmap_copy page tables\n"); pdnxt = ((addr + PAGE_SIZE*NPTEPG) & ~(PAGE_SIZE*NPTEPG - 1)); srcptepaddr = (vm_offset_t) src_pmap->pm_pdir[addr >> PDRSHIFT]; if (srcptepaddr == 0) { @@ -1983,7 +2015,7 @@ pmap_copy(dst_pmap, src_pmap, dst_addr, len, src_addr) * block. */ dstmpte = pmap_allocpte(dst_pmap, addr); - if (ptetemp = *src_pte) { + if ((*dst_pte == 0) && (ptetemp = *src_pte)) { /* * Simply clear the modified and accessed (referenced) * bits. @@ -1993,7 +2025,7 @@ pmap_copy(dst_pmap, src_pmap, dst_addr, len, src_addr) pmap_insert_entry(dst_pmap, addr, dstmpte, (ptetemp & PG_FRAME)); } else { - pmap_unwire_pte_hold(dstmpte); + pmap_unwire_pte_hold(dst_pmap, dstmpte); } if (dstmpte->hold_count >= srcmpte->hold_count) break; @@ -2133,38 +2165,36 @@ pmap_testbit(pa, bit) if (!pmap_is_managed(pa)) return FALSE; - s = splvm(); - ppv = pa_to_pvh(pa); + if (*ppv == NULL) + return FALSE; + + s = splvm(); /* * Not found, check current mappings returning immediately if found. */ for (pv = *ppv ;pv; pv = pv->pv_next) { + /* * if the bit being tested is the modified bit, then - * mark UPAGES as always modified, and ptes as never + * mark clean_map and ptes as never * modified. */ if (bit & (PG_A|PG_M)) { - if ((pv->pv_va >= UPT_MIN_ADDRESS) && - (pv->pv_va < UPT_MAX_ADDRESS)) { + if (!pmap_track_modified(pv->pv_va)) continue; - } - if ((pv->pv_va >= clean_sva) && - (pv->pv_va < clean_eva)) { - continue; - } } + if (!pv->pv_pmap) { #if defined(PMAP_DIAGNOSTIC) printf("Null pmap (tb) at va: 0x%lx\n", pv->pv_va); #endif continue; } - pte = pmap_pte(pv->pv_pmap, pv->pv_va); + pte = pmap_pte_quick(pv->pv_pmap, pv->pv_va); if (pte == NULL) continue; - if ((int) *pte & bit) { + if (*pte & bit) { splx(s); return TRUE; } @@ -2192,7 +2222,6 @@ pmap_changebit(pa, bit, setem) return; s = splvm(); - changed = 0; ppv = pa_to_pvh(pa); /* @@ -2216,7 +2245,7 @@ pmap_changebit(pa, bit, setem) continue; } - pte = pmap_pte(pv->pv_pmap, va); + pte = pmap_pte_quick(pv->pv_pmap, va); if (pte == NULL) continue; if (setem) { @@ -2255,9 +2284,9 @@ pmap_page_protect(phys, prot) vm_prot_t prot; { if ((prot & VM_PROT_WRITE) == 0) { - if (prot & (VM_PROT_READ | VM_PROT_EXECUTE)) + if (prot & (VM_PROT_READ | VM_PROT_EXECUTE)) { pmap_changebit(phys, PG_RW, FALSE); - else { + } else { pmap_remove_all(phys); pmap_update(); } @@ -2280,7 +2309,110 @@ pmap_phys_address(ppn) boolean_t pmap_is_referenced(vm_offset_t pa) { - return pmap_testbit((pa), PG_A); + register pv_entry_t *ppv, pv, lpv; + unsigned *pte; + int s; + + if (!pmap_is_managed(pa)) + return FALSE; + + ppv = pa_to_pvh(pa); + + s = splvm(); + /* + * Not found, check current mappings returning immediately if found. + */ + for (lpv = NULL, pv = *ppv ;pv; lpv = pv, pv = pv->pv_next) { + /* + * if the bit being tested is the modified bit, then + * mark clean_map and ptes as never + * modified. + */ + if (!pmap_track_modified(pv->pv_va)) + continue; + if (!pv->pv_pmap) { + continue; + } + pte = pmap_pte_quick(pv->pv_pmap, pv->pv_va); + if (pte == NULL) + continue; + if ((int) *pte & PG_A) { + if (lpv) { + lpv->pv_next = pv->pv_next; + pv->pv_next = *ppv; + *ppv = pv; + } + splx(s); + return TRUE; + } + } + splx(s); + return (FALSE); +} + +/* + * pmap_ts_referenced: + * + * Return the count of reference bits for a page, clearing all of them. + * + */ +int +pmap_ts_referenced(vm_offset_t pa) +{ + register pv_entry_t *ppv, pv; + unsigned *pte; + int s; + int rtval = 0; + vm_offset_t vachanged[VATRACK]; + + if (!pmap_is_managed(pa)) + return FALSE; + + s = splvm(); + + ppv = pa_to_pvh(pa); + + if (*ppv == NULL) { + splx(s); + return 0; + } + + /* + * Not found, check current mappings returning immediately if found. + */ + for (pv = *ppv ;pv; pv = pv->pv_next) { + /* + * if the bit being tested is the modified bit, then + * mark clean_map and ptes as never + * modified. + */ + if (!pmap_track_modified(pv->pv_va)) + continue; + + if (!pv->pv_pmap) { + continue; + } + pte = pmap_pte_quick(pv->pv_pmap, pv->pv_va); + if (pte == NULL) + continue; + if (*pte & PG_A) { + if (rtval < VATRACK) + vachanged[rtval] = pv->pv_va; + rtval++; + *pte &= ~PG_A; + } + } + splx(s); + if (rtval) { + if (rtval <= VATRACK) { + int i; + for(i=0;i<rtval;i++) + pmap_update_1pg(vachanged[i]); + } else { + pmap_update(); + } + } + return (rtval); } /* @@ -2381,6 +2513,9 @@ pmap_mapdev(pa, size) return ((void *) va); } +/* + * perform the pmap work for mincore + */ int pmap_mincore(pmap, addr) pmap_t pmap; @@ -2395,7 +2530,7 @@ pmap_mincore(pmap, addr) return 0; } - if ((pte = *ptep)) { + if (pte = *ptep) { vm_offset_t pa; val = MINCORE_INCORE; pa = pte & PG_FRAME; diff --git a/sys/vm/pmap.h b/sys/vm/pmap.h index 78c59a5..e123b90 100644 --- a/sys/vm/pmap.h +++ b/sys/vm/pmap.h @@ -61,7 +61,7 @@ * any improvements or extensions that they make and grant Carnegie the * rights to redistribute these changes. * - * $Id: pmap.h,v 1.9 1996/03/28 04:54:50 dyson Exp $ + * $Id: pmap.h,v 1.10 1996/05/19 07:36:44 dyson Exp $ */ /* @@ -101,6 +101,7 @@ void pmap_growkernel __P((vm_offset_t)); void pmap_init __P((vm_offset_t, vm_offset_t)); boolean_t pmap_is_modified __P((vm_offset_t pa)); boolean_t pmap_is_referenced __P((vm_offset_t pa)); +boolean_t pmap_ts_referenced __P((vm_offset_t pa)); void pmap_kenter __P((vm_offset_t, vm_offset_t)); void pmap_kremove __P((vm_offset_t)); vm_offset_t pmap_map __P((vm_offset_t, vm_offset_t, vm_offset_t, int)); diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c index d33495f..445c207 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 1.56 1996/06/12 06:52:06 dyson Exp $ + * $Id: vm_page.c,v 1.57 1996/06/16 20:37:31 dyson Exp $ */ /* @@ -802,12 +802,13 @@ vm_page_free_wakeup() * high water mark. And wakeup scheduler process if we have * lots of memory. this process will swapin processes. */ - if ((cnt.v_free_count + cnt.v_cache_count) == cnt.v_free_min) { + if (vm_pages_needed && + ((cnt.v_free_count + cnt.v_cache_count) >= cnt.v_free_min)) { wakeup(&cnt.v_free_count); + vm_pages_needed = 0; } } - /* * vm_page_free: * diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c index 3799f3a..035f3d9 100644 --- a/sys/vm/vm_pageout.c +++ b/sys/vm/vm_pageout.c @@ -65,7 +65,7 @@ * any improvements or extensions that they make and grant Carnegie the * rights to redistribute these changes. * - * $Id: vm_pageout.c,v 1.75 1996/05/29 06:33:30 dyson Exp $ + * $Id: vm_pageout.c,v 1.76 1996/05/31 00:38:04 dyson Exp $ */ /* @@ -416,6 +416,7 @@ vm_pageout_object_deactivate_pages(map, object, count, map_remove_only) rcount = object->resident_page_count; p = TAILQ_FIRST(&object->memq); while (p && (rcount-- > 0)) { + int refcount; next = TAILQ_NEXT(p, listq); cnt.v_pdpages++; if (p->wire_count != 0 || @@ -426,13 +427,27 @@ vm_pageout_object_deactivate_pages(map, object, count, map_remove_only) p = next; continue; } + + refcount = 0; + if ((p->flags & PG_REFERENCED) == 0) { + refcount = pmap_ts_referenced(VM_PAGE_TO_PHYS(p)); + if (refcount) { + p->flags |= PG_REFERENCED; + } + } else { + pmap_clear_reference(VM_PAGE_TO_PHYS(p)); + } + + if ((p->queue != PQ_ACTIVE) && (p->flags & PG_REFERENCED)) { + vm_page_activate(p); + } + /* * if a page is active, not wired and is in the processes * pmap, then deactivate the page. */ if (p->queue == PQ_ACTIVE) { - if (!pmap_is_referenced(VM_PAGE_TO_PHYS(p)) && - (p->flags & PG_REFERENCED) == 0) { + if ((p->flags & PG_REFERENCED) == 0) { vm_page_protect(p, VM_PROT_NONE); if (!map_remove_only) vm_page_deactivate(p); @@ -448,12 +463,7 @@ vm_pageout_object_deactivate_pages(map, object, count, map_remove_only) } } } else { - /* - * Move the page to the bottom of the queue. - */ - pmap_clear_reference(VM_PAGE_TO_PHYS(p)); p->flags &= ~PG_REFERENCED; - s = splvm(); TAILQ_REMOVE(&vm_page_queue_active, p, pageq); TAILQ_INSERT_TAIL(&vm_page_queue_active, p, pageq); @@ -576,14 +586,15 @@ rescan0: continue; } - if (((m->flags & PG_REFERENCED) == 0) && - pmap_is_referenced(VM_PAGE_TO_PHYS(m))) { - m->flags |= PG_REFERENCED; - } if (m->object->ref_count == 0) { m->flags &= ~PG_REFERENCED; pmap_clear_reference(VM_PAGE_TO_PHYS(m)); + } else if (((m->flags & PG_REFERENCED) == 0) && + pmap_ts_referenced(VM_PAGE_TO_PHYS(m))) { + vm_page_activate(m); + continue; } + if ((m->flags & PG_REFERENCED) != 0) { m->flags &= ~PG_REFERENCED; pmap_clear_reference(VM_PAGE_TO_PHYS(m)); @@ -596,6 +607,7 @@ rescan0: } else if (m->dirty != 0) { m->dirty = VM_PAGE_BITS_ALL; } + if (m->valid == 0) { vm_page_protect(m, VM_PROT_NONE); vm_page_free(m); @@ -706,9 +718,12 @@ rescan0: page_shortage += addl_page_shortage; } + +rescan1: pcount = cnt.v_active_count; m = TAILQ_FIRST(&vm_page_queue_active); while ((m != NULL) && (pcount-- > 0) && (page_shortage > 0)) { + int refcount; if (m->queue != PQ_ACTIVE) { #if defined(DIAGNOSTIC) @@ -721,7 +736,7 @@ rescan0: else printf("object type: %d\n", m->object->type); #endif - break; + goto rescan1; } next = TAILQ_NEXT(m, pageq); @@ -745,33 +760,31 @@ rescan0: * page for eligbility... */ cnt.v_pdpages++; - if ((m->flags & PG_REFERENCED) == 0) { - if (pmap_is_referenced(VM_PAGE_TO_PHYS(m))) { - pmap_clear_reference(VM_PAGE_TO_PHYS(m)); - m->flags |= PG_REFERENCED; + + refcount = 0; + if (m->object->ref_count != 0) { + if (m->flags & PG_REFERENCED) { + refcount += 1; } - } else { - pmap_clear_reference(VM_PAGE_TO_PHYS(m)); + refcount += pmap_ts_referenced(VM_PAGE_TO_PHYS(m)); } - if ( (m->object->ref_count != 0) && - (m->flags & PG_REFERENCED) ) { - m->flags &= ~PG_REFERENCED; + + m->flags &= ~PG_REFERENCED; + + if (refcount && m->object->ref_count != 0) { s = splvm(); TAILQ_REMOVE(&vm_page_queue_active, m, pageq); TAILQ_INSERT_TAIL(&vm_page_queue_active, m, pageq); splx(s); } else { - m->flags &= ~PG_REFERENCED; - if (page_shortage > 0) { - --page_shortage; - if (m->dirty == 0) - vm_page_test_dirty(m); - if (m->dirty == 0) { - vm_page_cache(m); - } else { - vm_page_protect(m, VM_PROT_NONE); - vm_page_deactivate(m); - } + --page_shortage; + vm_page_protect(m, VM_PROT_NONE); + if (m->dirty == 0) + vm_page_test_dirty(m); + if ((m->object->ref_count == 0) && (m->dirty == 0)) { + vm_page_cache(m); + } else { + vm_page_deactivate(m); } } m = next; |
