/* $NetBSD: uvm_map.c,v 1.362.2.5 2023/04/01 16:22:14 martin Exp $ */ /* * Copyright (c) 1997 Charles D. Cranor and Washington University. * Copyright (c) 1991, 1993, The Regents of the University of California. * * All rights reserved. * * This code is derived from software contributed to Berkeley by * The Mach Operating System project at Carnegie-Mellon University. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)vm_map.c 8.3 (Berkeley) 1/12/94 * from: Id: uvm_map.c,v 1.1.2.27 1998/02/07 01:16:54 chs Exp * * * Copyright (c) 1987, 1990 Carnegie-Mellon University. * All rights reserved. * * Permission to use, copy, modify and distribute this software and * its documentation is hereby granted, provided that both the copyright * notice and this permission notice appear in all copies of the * software, derivative works or modified versions, and any portions * thereof, and that both notices appear in supporting documentation. * * CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS" * CONDITION. CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND * FOR ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE. * * Carnegie Mellon requests users of this software to return to * * Software Distribution Coordinator or Software.Distribution@CS.CMU.EDU * School of Computer Science * Carnegie Mellon University * Pittsburgh PA 15213-3890 * * any improvements or extensions that they make and grant Carnegie the * rights to redistribute these changes. */ /* * uvm_map.c: uvm map operations */ #include __KERNEL_RCSID(0, "$NetBSD: uvm_map.c,v 1.362.2.5 2023/04/01 16:22:14 martin Exp $"); #include "opt_ddb.h" #include "opt_pax.h" #include "opt_uvmhist.h" #include "opt_uvm.h" #include "opt_sysv.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #ifndef __USER_VA0_IS_SAFE #include #include "opt_user_va0_disable_default.h" #endif #include #include #include #if defined(DDB) || defined(DEBUGPRINT) #include #endif #ifdef UVMHIST #ifndef UVMHIST_MAPHIST_SIZE #define UVMHIST_MAPHIST_SIZE 100 #endif #ifndef UVMHIST_PDHIST_SIZE #define UVMHIST_PDHIST_SIZE 100 #endif static struct kern_history_ent maphistbuf[UVMHIST_MAPHIST_SIZE]; UVMHIST_DEFINE(maphist) = UVMHIST_INITIALIZER(maphist, maphistbuf); #endif #if !defined(UVMMAP_COUNTERS) #define UVMMAP_EVCNT_DEFINE(name) /* nothing */ #define UVMMAP_EVCNT_INCR(ev) /* nothing */ #define UVMMAP_EVCNT_DECR(ev) /* nothing */ #else /* defined(UVMMAP_NOCOUNTERS) */ #include #define UVMMAP_EVCNT_DEFINE(name) \ struct evcnt uvmmap_evcnt_##name = EVCNT_INITIALIZER(EVCNT_TYPE_MISC, NULL, \ "uvmmap", #name); \ EVCNT_ATTACH_STATIC(uvmmap_evcnt_##name); #define UVMMAP_EVCNT_INCR(ev) uvmmap_evcnt_##ev.ev_count++ #define UVMMAP_EVCNT_DECR(ev) uvmmap_evcnt_##ev.ev_count-- #endif /* defined(UVMMAP_NOCOUNTERS) */ UVMMAP_EVCNT_DEFINE(ubackmerge) UVMMAP_EVCNT_DEFINE(uforwmerge) UVMMAP_EVCNT_DEFINE(ubimerge) UVMMAP_EVCNT_DEFINE(unomerge) UVMMAP_EVCNT_DEFINE(kbackmerge) UVMMAP_EVCNT_DEFINE(kforwmerge) UVMMAP_EVCNT_DEFINE(kbimerge) UVMMAP_EVCNT_DEFINE(knomerge) UVMMAP_EVCNT_DEFINE(map_call) UVMMAP_EVCNT_DEFINE(mlk_call) UVMMAP_EVCNT_DEFINE(mlk_hint) UVMMAP_EVCNT_DEFINE(mlk_list) UVMMAP_EVCNT_DEFINE(mlk_tree) UVMMAP_EVCNT_DEFINE(mlk_treeloop) UVMMAP_EVCNT_DEFINE(mlk_listloop) const char vmmapbsy[] = "vmmapbsy"; /* * cache for vmspace structures. */ static struct pool_cache uvm_vmspace_cache; /* * cache for dynamically-allocated map entries. */ static struct pool_cache uvm_map_entry_cache; #ifdef PMAP_GROWKERNEL /* * This global represents the end of the kernel virtual address * space. If we want to exceed this, we must grow the kernel * virtual address space dynamically. * * Note, this variable is locked by kernel_map's lock. */ vaddr_t uvm_maxkaddr; #endif #ifndef __USER_VA0_IS_SAFE #ifndef __USER_VA0_DISABLE_DEFAULT #define __USER_VA0_DISABLE_DEFAULT 1 #endif #ifdef USER_VA0_DISABLE_DEFAULT /* kernel config option overrides */ #undef __USER_VA0_DISABLE_DEFAULT #define __USER_VA0_DISABLE_DEFAULT USER_VA0_DISABLE_DEFAULT #endif int user_va0_disable = __USER_VA0_DISABLE_DEFAULT; #endif /* * macros */ /* * uvm_map_align_va: round down or up virtual address */ static __inline void uvm_map_align_va(vaddr_t *vap, vsize_t align, int topdown) { KASSERT(powerof2(align)); if (align != 0 && (*vap & (align - 1)) != 0) { if (topdown) *vap = rounddown2(*vap, align); else *vap = roundup2(*vap, align); } } /* * UVM_ET_ISCOMPATIBLE: check some requirements for map entry merging */ extern struct vm_map *pager_map; #define UVM_ET_ISCOMPATIBLE(ent, type, uobj, meflags, \ prot, maxprot, inh, adv, wire) \ ((ent)->etype == (type) && \ (((ent)->flags ^ (meflags)) & (UVM_MAP_NOMERGE)) == 0 && \ (ent)->object.uvm_obj == (uobj) && \ (ent)->protection == (prot) && \ (ent)->max_protection == (maxprot) && \ (ent)->inheritance == (inh) && \ (ent)->advice == (adv) && \ (ent)->wired_count == (wire)) /* * uvm_map_entry_link: insert entry into a map * * => map must be locked */ #define uvm_map_entry_link(map, after_where, entry) do { \ uvm_mapent_check(entry); \ (map)->nentries++; \ (entry)->prev = (after_where); \ (entry)->next = (after_where)->next; \ (entry)->prev->next = (entry); \ (entry)->next->prev = (entry); \ uvm_rb_insert((map), (entry)); \ } while (/*CONSTCOND*/ 0) /* * uvm_map_entry_unlink: remove entry from a map * * => map must be locked */ #define uvm_map_entry_unlink(map, entry) do { \ KASSERT((entry) != (map)->first_free); \ KASSERT((entry) != (map)->hint); \ uvm_mapent_check(entry); \ (map)->nentries--; \ (entry)->next->prev = (entry)->prev; \ (entry)->prev->next = (entry)->next; \ uvm_rb_remove((map), (entry)); \ } while (/*CONSTCOND*/ 0) /* * SAVE_HINT: saves the specified entry as the hint for future lookups. * * => map need not be locked. */ #define SAVE_HINT(map, check, value) do { \ if ((map)->hint == (check)) \ (map)->hint = (value); \ } while (/*CONSTCOND*/ 0) /* * clear_hints: ensure that hints don't point to the entry. * * => map must be write-locked. */ static void clear_hints(struct vm_map *map, struct vm_map_entry *ent) { SAVE_HINT(map, ent, ent->prev); if (map->first_free == ent) { map->first_free = ent->prev; } } /* * VM_MAP_RANGE_CHECK: check and correct range * * => map must at least be read locked */ #define VM_MAP_RANGE_CHECK(map, start, end) do { \ if (start < vm_map_min(map)) \ start = vm_map_min(map); \ if (end > vm_map_max(map)) \ end = vm_map_max(map); \ if (start > end) \ start = end; \ } while (/*CONSTCOND*/ 0) /* * local prototypes */ static struct vm_map_entry * uvm_mapent_alloc(struct vm_map *, int); static void uvm_mapent_copy(struct vm_map_entry *, struct vm_map_entry *); static void uvm_mapent_free(struct vm_map_entry *); #if defined(DEBUG) static void _uvm_mapent_check(const struct vm_map_entry *, const char *, int); #define uvm_mapent_check(map) _uvm_mapent_check(map, __FILE__, __LINE__) #else /* defined(DEBUG) */ #define uvm_mapent_check(e) /* nothing */ #endif /* defined(DEBUG) */ static void uvm_map_entry_unwire(struct vm_map *, struct vm_map_entry *); static void uvm_map_reference_amap(struct vm_map_entry *, int); static int uvm_map_space_avail(vaddr_t *, vsize_t, voff_t, vsize_t, int, int, struct vm_map_entry *); static void uvm_map_unreference_amap(struct vm_map_entry *, int); int _uvm_map_sanity(struct vm_map *); int _uvm_tree_sanity(struct vm_map *); static vsize_t uvm_rb_maxgap(const struct vm_map_entry *); #define ROOT_ENTRY(map) ((struct vm_map_entry *)(map)->rb_tree.rbt_root) #define LEFT_ENTRY(entry) ((struct vm_map_entry *)(entry)->rb_node.rb_left) #define RIGHT_ENTRY(entry) ((struct vm_map_entry *)(entry)->rb_node.rb_right) #define PARENT_ENTRY(map, entry) \ (ROOT_ENTRY(map) == (entry) \ ? NULL : (struct vm_map_entry *)RB_FATHER(&(entry)->rb_node)) /* * These get filled in if/when SYSVSHM shared memory code is loaded * * We do this with function pointers rather the #ifdef SYSVSHM so the * SYSVSHM code can be loaded and unloaded */ void (*uvm_shmexit)(struct vmspace *) = NULL; void (*uvm_shmfork)(struct vmspace *, struct vmspace *) = NULL; static int uvm_map_compare_nodes(void *ctx, const void *nparent, const void *nkey) { const struct vm_map_entry *eparent = nparent; const struct vm_map_entry *ekey = nkey; KASSERT(eparent->start < ekey->start || eparent->start >= ekey->end); KASSERT(ekey->start < eparent->start || ekey->start >= eparent->end); if (eparent->start < ekey->start) return -1; if (eparent->end >= ekey->start) return 1; return 0; } static int uvm_map_compare_key(void *ctx, const void *nparent, const void *vkey) { const struct vm_map_entry *eparent = nparent; const vaddr_t va = *(const vaddr_t *) vkey; if (eparent->start < va) return -1; if (eparent->end >= va) return 1; return 0; } static const rb_tree_ops_t uvm_map_tree_ops = { .rbto_compare_nodes = uvm_map_compare_nodes, .rbto_compare_key = uvm_map_compare_key, .rbto_node_offset = offsetof(struct vm_map_entry, rb_node), .rbto_context = NULL }; /* * uvm_rb_gap: return the gap size between our entry and next entry. */ static inline vsize_t uvm_rb_gap(const struct vm_map_entry *entry) { KASSERT(entry->next != NULL); return entry->next->start - entry->end; } static vsize_t uvm_rb_maxgap(const struct vm_map_entry *entry) { struct vm_map_entry *child; vsize_t maxgap = entry->gap; /* * We need maxgap to be the largest gap of us or any of our * descendents. Since each of our children's maxgap is the * cached value of their largest gap of themselves or their * descendents, we can just use that value and avoid recursing * down the tree to calculate it. */ if ((child = LEFT_ENTRY(entry)) != NULL && maxgap < child->maxgap) maxgap = child->maxgap; if ((child = RIGHT_ENTRY(entry)) != NULL && maxgap < child->maxgap) maxgap = child->maxgap; return maxgap; } static void uvm_rb_fixup(struct vm_map *map, struct vm_map_entry *entry) { struct vm_map_entry *parent; KASSERT(entry->gap == uvm_rb_gap(entry)); entry->maxgap = uvm_rb_maxgap(entry); while ((parent = PARENT_ENTRY(map, entry)) != NULL) { struct vm_map_entry *brother; vsize_t maxgap = parent->gap; unsigned int which; KDASSERT(parent->gap == uvm_rb_gap(parent)); if (maxgap < entry->maxgap) maxgap = entry->maxgap; /* * Since we work towards the root, we know entry's maxgap * value is OK, but its brothers may now be out-of-date due * to rebalancing. So refresh it. */ which = RB_POSITION(&entry->rb_node) ^ RB_DIR_OTHER; brother = (struct vm_map_entry *)parent->rb_node.rb_nodes[which]; if (brother != NULL) { KDASSERT(brother->gap == uvm_rb_gap(brother)); brother->maxgap = uvm_rb_maxgap(brother); if (maxgap < brother->maxgap) maxgap = brother->maxgap; } parent->maxgap = maxgap; entry = parent; } } static void uvm_rb_insert(struct vm_map *map, struct vm_map_entry *entry) { struct vm_map_entry *ret __diagused; entry->gap = entry->maxgap = uvm_rb_gap(entry); if (entry->prev != &map->header) entry->prev->gap = uvm_rb_gap(entry->prev); ret = rb_tree_insert_node(&map->rb_tree, entry); KASSERTMSG(ret == entry, "uvm_rb_insert: map %p: duplicate entry %p", map, ret); /* * If the previous entry is not our immediate left child, then it's an * ancestor and will be fixed up on the way to the root. We don't * have to check entry->prev against &map->header since &map->header * will never be in the tree. */ uvm_rb_fixup(map, LEFT_ENTRY(entry) == entry->prev ? entry->prev : entry); } static void uvm_rb_remove(struct vm_map *map, struct vm_map_entry *entry) { struct vm_map_entry *prev_parent = NULL, *next_parent = NULL; /* * If we are removing an interior node, then an adjacent node will * be used to replace its position in the tree. Therefore we will * need to fixup the tree starting at the parent of the replacement * node. So record their parents for later use. */ if (entry->prev != &map->header) prev_parent = PARENT_ENTRY(map, entry->prev); if (entry->next != &map->header) next_parent = PARENT_ENTRY(map, entry->next); rb_tree_remove_node(&map->rb_tree, entry); /* * If the previous node has a new parent, fixup the tree starting * at the previous node's old parent. */ if (entry->prev != &map->header) { /* * Update the previous entry's gap due to our absence. */ entry->prev->gap = uvm_rb_gap(entry->prev); uvm_rb_fixup(map, entry->prev); if (prev_parent != NULL && prev_parent != entry && prev_parent != PARENT_ENTRY(map, entry->prev)) uvm_rb_fixup(map, prev_parent); } /* * If the next node has a new parent, fixup the tree starting * at the next node's old parent. */ if (entry->next != &map->header) { uvm_rb_fixup(map, entry->next); if (next_parent != NULL && next_parent != entry && next_parent != PARENT_ENTRY(map, entry->next)) uvm_rb_fixup(map, next_parent); } } #if defined(DEBUG) int uvm_debug_check_map = 0; int uvm_debug_check_rbtree = 0; #define uvm_map_check(map, name) \ _uvm_map_check((map), (name), __FILE__, __LINE__) static void _uvm_map_check(struct vm_map *map, const char *name, const char *file, int line) { if ((uvm_debug_check_map && _uvm_map_sanity(map)) || (uvm_debug_check_rbtree && _uvm_tree_sanity(map))) { panic("uvm_map_check failed: \"%s\" map=%p (%s:%d)", name, map, file, line); } } #else /* defined(DEBUG) */ #define uvm_map_check(map, name) /* nothing */ #endif /* defined(DEBUG) */ #if defined(DEBUG) || defined(DDB) int _uvm_map_sanity(struct vm_map *map) { bool first_free_found = false; bool hint_found = false; const struct vm_map_entry *e; struct vm_map_entry *hint = map->hint; e = &map->header; for (;;) { if (map->first_free == e) { first_free_found = true; } else if (!first_free_found && e->next->start > e->end) { printf("first_free %p should be %p\n", map->first_free, e); return -1; } if (hint == e) { hint_found = true; } e = e->next; if (e == &map->header) { break; } } if (!first_free_found) { printf("stale first_free\n"); return -1; } if (!hint_found) { printf("stale hint\n"); return -1; } return 0; } int _uvm_tree_sanity(struct vm_map *map) { struct vm_map_entry *tmp, *trtmp; int n = 0, i = 1; for (tmp = map->header.next; tmp != &map->header; tmp = tmp->next) { if (tmp->gap != uvm_rb_gap(tmp)) { printf("%d/%d gap %#lx != %#lx %s\n", n + 1, map->nentries, (ulong)tmp->gap, (ulong)uvm_rb_gap(tmp), tmp->next == &map->header ? "(last)" : ""); goto error; } /* * If any entries are out of order, tmp->gap will be unsigned * and will likely exceed the size of the map. */ if (tmp->gap >= vm_map_max(map) - vm_map_min(map)) { printf("too large gap %zu\n", (size_t)tmp->gap); goto error; } n++; } if (n != map->nentries) { printf("nentries: %d vs %d\n", n, map->nentries); goto error; } trtmp = NULL; for (tmp = map->header.next; tmp != &map->header; tmp = tmp->next) { if (tmp->maxgap != uvm_rb_maxgap(tmp)) { printf("maxgap %#lx != %#lx\n", (ulong)tmp->maxgap, (ulong)uvm_rb_maxgap(tmp)); goto error; } if (trtmp != NULL && trtmp->start >= tmp->start) { printf("corrupt: 0x%"PRIxVADDR"x >= 0x%"PRIxVADDR"x\n", trtmp->start, tmp->start); goto error; } trtmp = tmp; } for (tmp = map->header.next; tmp != &map->header; tmp = tmp->next, i++) { trtmp = rb_tree_iterate(&map->rb_tree, tmp, RB_DIR_LEFT); if (trtmp == NULL) trtmp = &map->header; if (tmp->prev != trtmp) { printf("lookup: %d: %p->prev=%p: %p\n", i, tmp, tmp->prev, trtmp); goto error; } trtmp = rb_tree_iterate(&map->rb_tree, tmp, RB_DIR_RIGHT); if (trtmp == NULL) trtmp = &map->header; if (tmp->next != trtmp) { printf("lookup: %d: %p->next=%p: %p\n", i, tmp, tmp->next, trtmp); goto error; } trtmp = rb_tree_find_node(&map->rb_tree, &tmp->start); if (trtmp != tmp) { printf("lookup: %d: %p - %p: %p\n", i, tmp, trtmp, PARENT_ENTRY(map, tmp)); goto error; } } return (0); error: return (-1); } #endif /* defined(DEBUG) || defined(DDB) */ /* * vm_map_lock: acquire an exclusive (write) lock on a map. * * => The locking protocol provides for guaranteed upgrade from shared -> * exclusive by whichever thread currently has the map marked busy. * See "LOCKING PROTOCOL NOTES" in uvm_map.h. This is horrible; among * other problems, it defeats any fairness guarantees provided by RW * locks. */ void vm_map_lock(struct vm_map *map) { for (;;) { rw_enter(&map->lock, RW_WRITER); if (map->busy == NULL || map->busy == curlwp) { break; } mutex_enter(&map->misc_lock); rw_exit(&map->lock); if (map->busy != NULL) { cv_wait(&map->cv, &map->misc_lock); } mutex_exit(&map->misc_lock); } map->timestamp++; } /* * vm_map_lock_try: try to lock a map, failing if it is already locked. */ bool vm_map_lock_try(struct vm_map *map) { if (!rw_tryenter(&map->lock, RW_WRITER)) { return false; } if (map->busy != NULL) { rw_exit(&map->lock); return false; } map->timestamp++; return true; } /* * vm_map_unlock: release an exclusive lock on a map. */ void vm_map_unlock(struct vm_map *map) { KASSERT(rw_write_held(&map->lock)); KASSERT(map->busy == NULL || map->busy == curlwp); rw_exit(&map->lock); } /* * vm_map_unbusy: mark the map as unbusy, and wake any waiters that * want an exclusive lock. */ void vm_map_unbusy(struct vm_map *map) { KASSERT(map->busy == curlwp); /* * Safe to clear 'busy' and 'waiters' with only a read lock held: * * o they can only be set with a write lock held * o writers are blocked out with a read or write hold * o at any time, only one thread owns the set of values */ mutex_enter(&map->misc_lock); map->busy = NULL; cv_broadcast(&map->cv); mutex_exit(&map->misc_lock); } /* * vm_map_lock_read: acquire a shared (read) lock on a map. */ void vm_map_lock_read(struct vm_map *map) { rw_enter(&map->lock, RW_READER); } /* * vm_map_unlock_read: release a shared lock on a map. */ void vm_map_unlock_read(struct vm_map *map) { rw_exit(&map->lock); } /* * vm_map_busy: mark a map as busy. * * => the caller must hold the map write locked */ void vm_map_busy(struct vm_map *map) { KASSERT(rw_write_held(&map->lock)); KASSERT(map->busy == NULL); map->busy = curlwp; } /* * vm_map_locked_p: return true if the map is write locked. * * => only for debug purposes like KASSERTs. * => should not be used to verify that a map is not locked. */ bool vm_map_locked_p(struct vm_map *map) { return rw_write_held(&map->lock); } /* * uvm_mapent_alloc: allocate a map entry */ static struct vm_map_entry * uvm_mapent_alloc(struct vm_map *map, int flags) { struct vm_map_entry *me; int pflags = (flags & UVM_FLAG_NOWAIT) ? PR_NOWAIT : PR_WAITOK; UVMHIST_FUNC("uvm_mapent_alloc"); UVMHIST_CALLED(maphist); me = pool_cache_get(&uvm_map_entry_cache, pflags); if (__predict_false(me == NULL)) { return NULL; } me->flags = 0; UVMHIST_LOG(maphist, "<- new entry=%#jx [kentry=%jd]", (uintptr_t)me, (map == kernel_map), 0, 0); return me; } /* * uvm_mapent_free: free map entry */ static void uvm_mapent_free(struct vm_map_entry *me) { UVMHIST_FUNC("uvm_mapent_free"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"<- freeing map entry=%#jx [flags=%jd]", (uintptr_t)me, me->flags, 0, 0); pool_cache_put(&uvm_map_entry_cache, me); } /* * uvm_mapent_copy: copy a map entry, preserving flags */ static inline void uvm_mapent_copy(struct vm_map_entry *src, struct vm_map_entry *dst) { memcpy(dst, src, ((char *)&src->uvm_map_entry_stop_copy) - ((char *)src)); } #if defined(DEBUG) static void _uvm_mapent_check(const struct vm_map_entry *entry, const char *file, int line) { if (entry->start >= entry->end) { goto bad; } if (UVM_ET_ISOBJ(entry)) { if (entry->object.uvm_obj == NULL) { goto bad; } } else if (UVM_ET_ISSUBMAP(entry)) { if (entry->object.sub_map == NULL) { goto bad; } } else { if (entry->object.uvm_obj != NULL || entry->object.sub_map != NULL) { goto bad; } } if (!UVM_ET_ISOBJ(entry)) { if (entry->offset != 0) { goto bad; } } return; bad: panic("%s: bad entry %p (%s:%d)", __func__, entry, file, line); } #endif /* defined(DEBUG) */ /* * uvm_map_entry_unwire: unwire a map entry * * => map should be locked by caller */ static inline void uvm_map_entry_unwire(struct vm_map *map, struct vm_map_entry *entry) { entry->wired_count = 0; uvm_fault_unwire_locked(map, entry->start, entry->end); } /* * wrapper for calling amap_ref() */ static inline void uvm_map_reference_amap(struct vm_map_entry *entry, int flags) { amap_ref(entry->aref.ar_amap, entry->aref.ar_pageoff, (entry->end - entry->start) >> PAGE_SHIFT, flags); } /* * wrapper for calling amap_unref() */ static inline void uvm_map_unreference_amap(struct vm_map_entry *entry, int flags) { amap_unref(entry->aref.ar_amap, entry->aref.ar_pageoff, (entry->end - entry->start) >> PAGE_SHIFT, flags); } /* * uvm_map_init: init mapping system at boot time. */ void uvm_map_init(void) { #if defined(UVMHIST) static struct kern_history_ent pdhistbuf[UVMHIST_PDHIST_SIZE]; #endif /* * first, init logging system. */ UVMHIST_FUNC("uvm_map_init"); UVMHIST_LINK_STATIC(maphist); UVMHIST_INIT_STATIC(pdhist, pdhistbuf); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"", 0, 0, 0, 0); /* * initialize the global lock for kernel map entry. */ mutex_init(&uvm_kentry_lock, MUTEX_DRIVER, IPL_VM); } /* * uvm_map_init_caches: init mapping system caches. */ void uvm_map_init_caches(void) { /* * initialize caches. */ pool_cache_bootstrap(&uvm_map_entry_cache, sizeof(struct vm_map_entry), 0, 0, 0, "vmmpepl", NULL, IPL_NONE, NULL, NULL, NULL); pool_cache_bootstrap(&uvm_vmspace_cache, sizeof(struct vmspace), 0, 0, 0, "vmsppl", NULL, IPL_NONE, NULL, NULL, NULL); } /* * clippers */ /* * uvm_mapent_splitadj: adjust map entries for splitting, after uvm_mapent_copy. */ static void uvm_mapent_splitadj(struct vm_map_entry *entry1, struct vm_map_entry *entry2, vaddr_t splitat) { vaddr_t adj; KASSERT(entry1->start < splitat); KASSERT(splitat < entry1->end); adj = splitat - entry1->start; entry1->end = entry2->start = splitat; if (entry1->aref.ar_amap) { amap_splitref(&entry1->aref, &entry2->aref, adj); } if (UVM_ET_ISSUBMAP(entry1)) { /* ... unlikely to happen, but play it safe */ uvm_map_reference(entry1->object.sub_map); } else if (UVM_ET_ISOBJ(entry1)) { KASSERT(entry1->object.uvm_obj != NULL); /* suppress coverity */ entry2->offset += adj; if (entry1->object.uvm_obj->pgops && entry1->object.uvm_obj->pgops->pgo_reference) entry1->object.uvm_obj->pgops->pgo_reference( entry1->object.uvm_obj); } } /* * uvm_map_clip_start: ensure that the entry begins at or after * the starting address, if it doesn't we split the entry. * * => caller should use UVM_MAP_CLIP_START macro rather than calling * this directly * => map must be locked by caller */ void uvm_map_clip_start(struct vm_map *map, struct vm_map_entry *entry, vaddr_t start) { struct vm_map_entry *new_entry; /* uvm_map_simplify_entry(map, entry); */ /* XXX */ uvm_map_check(map, "clip_start entry"); uvm_mapent_check(entry); /* * Split off the front portion. note that we must insert the new * entry BEFORE this one, so that this entry has the specified * starting address. */ new_entry = uvm_mapent_alloc(map, 0); uvm_mapent_copy(entry, new_entry); /* entry -> new_entry */ uvm_mapent_splitadj(new_entry, entry, start); uvm_map_entry_link(map, entry->prev, new_entry); uvm_map_check(map, "clip_start leave"); } /* * uvm_map_clip_end: ensure that the entry ends at or before * the ending address, if it does't we split the reference * * => caller should use UVM_MAP_CLIP_END macro rather than calling * this directly * => map must be locked by caller */ void uvm_map_clip_end(struct vm_map *map, struct vm_map_entry *entry, vaddr_t end) { struct vm_map_entry *new_entry; uvm_map_check(map, "clip_end entry"); uvm_mapent_check(entry); /* * Create a new entry and insert it * AFTER the specified entry */ new_entry = uvm_mapent_alloc(map, 0); uvm_mapent_copy(entry, new_entry); /* entry -> new_entry */ uvm_mapent_splitadj(entry, new_entry, end); uvm_map_entry_link(map, entry, new_entry); uvm_map_check(map, "clip_end leave"); } /* * M A P - m a i n e n t r y p o i n t */ /* * uvm_map: establish a valid mapping in a map * * => assume startp is page aligned. * => assume size is a multiple of PAGE_SIZE. * => assume sys_mmap provides enough of a "hint" to have us skip * over text/data/bss area. * => map must be unlocked (we will lock it) * => value meanings (4 cases): * [1] == uoffset is a hint for PMAP_PREFER * [2] == don't PMAP_PREFER * [3] == normal mapping * [4] == uvm_map finds offset based on VA * * case [4] is for kernel mappings where we don't know the offset until * we've found a virtual address. note that kernel object offsets are * always relative to vm_map_min(kernel_map). * * => if `align' is non-zero, we align the virtual address to the specified * alignment. * this is provided as a mechanism for large pages. * * => XXXCDC: need way to map in external amap? */ int uvm_map(struct vm_map *map, vaddr_t *startp /* IN/OUT */, vsize_t size, struct uvm_object *uobj, voff_t uoffset, vsize_t align, uvm_flag_t flags) { struct uvm_map_args args; struct vm_map_entry *new_entry; int error; KASSERT((size & PAGE_MASK) == 0); KASSERT((flags & UVM_FLAG_FIXED) == 0 || align == 0); /* * for pager_map, allocate the new entry first to avoid sleeping * for memory while we have the map locked. */ new_entry = NULL; if (map == pager_map) { new_entry = uvm_mapent_alloc(map, (flags & UVM_FLAG_NOWAIT)); if (__predict_false(new_entry == NULL)) return ENOMEM; } if (map == pager_map) flags |= UVM_FLAG_NOMERGE; error = uvm_map_prepare(map, *startp, size, uobj, uoffset, align, flags, &args); if (!error) { error = uvm_map_enter(map, &args, new_entry); *startp = args.uma_start; } else if (new_entry) { uvm_mapent_free(new_entry); } #if defined(DEBUG) if (!error && VM_MAP_IS_KERNEL(map) && (flags & UVM_FLAG_NOWAIT) == 0) { uvm_km_check_empty(map, *startp, *startp + size); } #endif /* defined(DEBUG) */ return error; } /* * uvm_map_prepare: * * called with map unlocked. * on success, returns the map locked. */ int uvm_map_prepare(struct vm_map *map, vaddr_t start, vsize_t size, struct uvm_object *uobj, voff_t uoffset, vsize_t align, uvm_flag_t flags, struct uvm_map_args *args) { struct vm_map_entry *prev_entry; vm_prot_t prot = UVM_PROTECTION(flags); vm_prot_t maxprot = UVM_MAXPROTECTION(flags); UVMHIST_FUNC("uvm_map_prepare"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist, "(map=%#jx, start=%#jx, size=%ju, flags=%#jx)", (uintptr_t)map, start, size, flags); UVMHIST_LOG(maphist, " uobj/offset %#jx/%jd", (uintptr_t)uobj, uoffset,0,0); /* * detect a popular device driver bug. */ KASSERT(doing_shutdown || curlwp != NULL); /* * zero-sized mapping doesn't make any sense. */ KASSERT(size > 0); KASSERT((~flags & (UVM_FLAG_NOWAIT | UVM_FLAG_WAITVA)) != 0); uvm_map_check(map, "map entry"); /* * check sanity of protection code */ if ((prot & maxprot) != prot) { UVMHIST_LOG(maphist, "<- prot. failure: prot=%#jx, max=%#jx", prot, maxprot,0,0); return EACCES; } /* * figure out where to put new VM range */ retry: if (vm_map_lock_try(map) == false) { if ((flags & UVM_FLAG_TRYLOCK) != 0) { return EAGAIN; } vm_map_lock(map); /* could sleep here */ } if (flags & UVM_FLAG_UNMAP) { KASSERT(flags & UVM_FLAG_FIXED); KASSERT((flags & UVM_FLAG_NOWAIT) == 0); /* * Set prev_entry to what it will need to be after any existing * entries are removed later in uvm_map_enter(). */ if (uvm_map_lookup_entry(map, start, &prev_entry)) { if (start == prev_entry->start) prev_entry = prev_entry->prev; else UVM_MAP_CLIP_END(map, prev_entry, start); SAVE_HINT(map, map->hint, prev_entry); } } else { prev_entry = uvm_map_findspace(map, start, size, &start, uobj, uoffset, align, flags); } if (prev_entry == NULL) { unsigned int timestamp; timestamp = map->timestamp; UVMHIST_LOG(maphist,"waiting va timestamp=%#jx", timestamp,0,0,0); map->flags |= VM_MAP_WANTVA; vm_map_unlock(map); /* * try to reclaim kva and wait until someone does unmap. * fragile locking here, so we awaken every second to * recheck the condition. */ mutex_enter(&map->misc_lock); while ((map->flags & VM_MAP_WANTVA) != 0 && map->timestamp == timestamp) { if ((flags & UVM_FLAG_WAITVA) == 0) { mutex_exit(&map->misc_lock); UVMHIST_LOG(maphist, "<- uvm_map_findspace failed!", 0,0,0,0); return ENOMEM; } else { cv_timedwait(&map->cv, &map->misc_lock, hz); } } mutex_exit(&map->misc_lock); goto retry; } #ifdef PMAP_GROWKERNEL /* * If the kernel pmap can't map the requested space, * then allocate more resources for it. */ if (map == kernel_map && uvm_maxkaddr < (start + size)) uvm_maxkaddr = pmap_growkernel(start + size); #endif UVMMAP_EVCNT_INCR(map_call); /* * if uobj is null, then uoffset is either a VAC hint for PMAP_PREFER * [typically from uvm_map_reserve] or it is UVM_UNKNOWN_OFFSET. in * either case we want to zero it before storing it in the map entry * (because it looks strange and confusing when debugging...) * * if uobj is not null * if uoffset is not UVM_UNKNOWN_OFFSET then we have a normal mapping * and we do not need to change uoffset. * if uoffset is UVM_UNKNOWN_OFFSET then we need to find the offset * now (based on the starting address of the map). this case is * for kernel object mappings where we don't know the offset until * the virtual address is found (with uvm_map_findspace). the * offset is the distance we are from the start of the map. */ if (uobj == NULL) { uoffset = 0; } else { if (uoffset == UVM_UNKNOWN_OFFSET) { KASSERT(UVM_OBJ_IS_KERN_OBJECT(uobj)); uoffset = start - vm_map_min(kernel_map); } } args->uma_flags = flags; args->uma_prev = prev_entry; args->uma_start = start; args->uma_size = size; args->uma_uobj = uobj; args->uma_uoffset = uoffset; UVMHIST_LOG(maphist, "<- done!", 0,0,0,0); return 0; } /* * uvm_map_enter: * * called with map locked. * unlock the map before returning. */ int uvm_map_enter(struct vm_map *map, const struct uvm_map_args *args, struct vm_map_entry *new_entry) { struct vm_map_entry *prev_entry = args->uma_prev; struct vm_map_entry *dead = NULL, *dead_entries = NULL; const uvm_flag_t flags = args->uma_flags; const vm_prot_t prot = UVM_PROTECTION(flags); const vm_prot_t maxprot = UVM_MAXPROTECTION(flags); const vm_inherit_t inherit = UVM_INHERIT(flags); const int amapwaitflag = (flags & UVM_FLAG_NOWAIT) ? AMAP_EXTEND_NOWAIT : 0; const int advice = UVM_ADVICE(flags); vaddr_t start = args->uma_start; vsize_t size = args->uma_size; struct uvm_object *uobj = args->uma_uobj; voff_t uoffset = args->uma_uoffset; const int kmap = (vm_map_pmap(map) == pmap_kernel()); int merged = 0; int error; int newetype; UVMHIST_FUNC("uvm_map_enter"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist, "(map=%#jx, start=%#jx, size=%ju, flags=%#jx)", (uintptr_t)map, start, size, flags); UVMHIST_LOG(maphist, " uobj/offset %#jx/%jd", (uintptr_t)uobj, uoffset,0,0); KASSERT(map->hint == prev_entry); /* bimerge case assumes this */ KASSERT(vm_map_locked_p(map)); KASSERT((flags & (UVM_FLAG_NOWAIT | UVM_FLAG_UNMAP)) != (UVM_FLAG_NOWAIT | UVM_FLAG_UNMAP)); if (uobj) newetype = UVM_ET_OBJ; else newetype = 0; if (flags & UVM_FLAG_COPYONW) { newetype |= UVM_ET_COPYONWRITE; if ((flags & UVM_FLAG_OVERLAY) == 0) newetype |= UVM_ET_NEEDSCOPY; } /* * For mappings with unmap, remove any old entries now. Adding the new * entry cannot fail because that can only happen if UVM_FLAG_NOWAIT * is set, and we do not support nowait and unmap together. */ if (flags & UVM_FLAG_UNMAP) { KASSERT(flags & UVM_FLAG_FIXED); uvm_unmap_remove(map, start, start + size, &dead_entries, 0); #ifdef DEBUG struct vm_map_entry *tmp_entry __diagused; bool rv __diagused; rv = uvm_map_lookup_entry(map, start, &tmp_entry); KASSERT(!rv); KASSERTMSG(prev_entry == tmp_entry, "args %p prev_entry %p tmp_entry %p", args, prev_entry, tmp_entry); #endif SAVE_HINT(map, map->hint, prev_entry); } /* * try and insert in map by extending previous entry, if possible. * XXX: we don't try and pull back the next entry. might be useful * for a stack, but we are currently allocating our stack in advance. */ if (flags & UVM_FLAG_NOMERGE) goto nomerge; if (prev_entry->end == start && prev_entry != &map->header && UVM_ET_ISCOMPATIBLE(prev_entry, newetype, uobj, 0, prot, maxprot, inherit, advice, 0)) { if (uobj && prev_entry->offset + (prev_entry->end - prev_entry->start) != uoffset) goto forwardmerge; /* * can't extend a shared amap. note: no need to lock amap to * look at refs since we don't care about its exact value. * if it is one (i.e. we have only reference) it will stay there */ if (prev_entry->aref.ar_amap && amap_refs(prev_entry->aref.ar_amap) != 1) { goto forwardmerge; } if (prev_entry->aref.ar_amap) { error = amap_extend(prev_entry, size, amapwaitflag | AMAP_EXTEND_FORWARDS); if (error) goto nomerge; } if (kmap) { UVMMAP_EVCNT_INCR(kbackmerge); } else { UVMMAP_EVCNT_INCR(ubackmerge); } UVMHIST_LOG(maphist," starting back merge", 0, 0, 0, 0); /* * drop our reference to uobj since we are extending a reference * that we already have (the ref count can not drop to zero). */ if (uobj && uobj->pgops->pgo_detach) uobj->pgops->pgo_detach(uobj); /* * Now that we've merged the entries, note that we've grown * and our gap has shrunk. Then fix the tree. */ prev_entry->end += size; prev_entry->gap -= size; uvm_rb_fixup(map, prev_entry); uvm_map_check(map, "map backmerged"); UVMHIST_LOG(maphist,"<- done (via backmerge)!", 0, 0, 0, 0); merged++; } forwardmerge: if (prev_entry->next->start == (start + size) && prev_entry->next != &map->header && UVM_ET_ISCOMPATIBLE(prev_entry->next, newetype, uobj, 0, prot, maxprot, inherit, advice, 0)) { if (uobj && prev_entry->next->offset != uoffset + size) goto nomerge; /* * can't extend a shared amap. note: no need to lock amap to * look at refs since we don't care about its exact value. * if it is one (i.e. we have only reference) it will stay there. * * note that we also can't merge two amaps, so if we * merged with the previous entry which has an amap, * and the next entry also has an amap, we give up. * * Interesting cases: * amap, new, amap -> give up second merge (single fwd extend) * amap, new, none -> double forward extend (extend again here) * none, new, amap -> double backward extend (done here) * uobj, new, amap -> single backward extend (done here) * * XXX should we attempt to deal with someone refilling * the deallocated region between two entries that are * backed by the same amap (ie, arefs is 2, "prev" and * "next" refer to it, and adding this allocation will * close the hole, thus restoring arefs to 1 and * deallocating the "next" vm_map_entry)? -- @@@ */ if (prev_entry->next->aref.ar_amap && (amap_refs(prev_entry->next->aref.ar_amap) != 1 || (merged && prev_entry->aref.ar_amap))) { goto nomerge; } if (merged) { /* * Try to extend the amap of the previous entry to * cover the next entry as well. If it doesn't work * just skip on, don't actually give up, since we've * already completed the back merge. */ if (prev_entry->aref.ar_amap) { if (amap_extend(prev_entry, prev_entry->next->end - prev_entry->next->start, amapwaitflag | AMAP_EXTEND_FORWARDS)) goto nomerge; } /* * Try to extend the amap of the *next* entry * back to cover the new allocation *and* the * previous entry as well (the previous merge * didn't have an amap already otherwise we * wouldn't be checking here for an amap). If * it doesn't work just skip on, again, don't * actually give up, since we've already * completed the back merge. */ else if (prev_entry->next->aref.ar_amap) { if (amap_extend(prev_entry->next, prev_entry->end - prev_entry->start, amapwaitflag | AMAP_EXTEND_BACKWARDS)) goto nomerge; } } else { /* * Pull the next entry's amap backwards to cover this * new allocation. */ if (prev_entry->next->aref.ar_amap) { error = amap_extend(prev_entry->next, size, amapwaitflag | AMAP_EXTEND_BACKWARDS); if (error) goto nomerge; } } if (merged) { if (kmap) { UVMMAP_EVCNT_DECR(kbackmerge); UVMMAP_EVCNT_INCR(kbimerge); } else { UVMMAP_EVCNT_DECR(ubackmerge); UVMMAP_EVCNT_INCR(ubimerge); } } else { if (kmap) { UVMMAP_EVCNT_INCR(kforwmerge); } else { UVMMAP_EVCNT_INCR(uforwmerge); } } UVMHIST_LOG(maphist," starting forward merge", 0, 0, 0, 0); /* * drop our reference to uobj since we are extending a reference * that we already have (the ref count can not drop to zero). */ if (uobj && uobj->pgops->pgo_detach) uobj->pgops->pgo_detach(uobj); if (merged) { dead = prev_entry->next; prev_entry->end = dead->end; uvm_map_entry_unlink(map, dead); if (dead->aref.ar_amap != NULL) { prev_entry->aref = dead->aref; dead->aref.ar_amap = NULL; } } else { prev_entry->next->start -= size; if (prev_entry != &map->header) { prev_entry->gap -= size; KASSERT(prev_entry->gap == uvm_rb_gap(prev_entry)); uvm_rb_fixup(map, prev_entry); } if (uobj) prev_entry->next->offset = uoffset; } uvm_map_check(map, "map forwardmerged"); UVMHIST_LOG(maphist,"<- done forwardmerge", 0, 0, 0, 0); merged++; } nomerge: if (!merged) { UVMHIST_LOG(maphist," allocating new map entry", 0, 0, 0, 0); if (kmap) { UVMMAP_EVCNT_INCR(knomerge); } else { UVMMAP_EVCNT_INCR(unomerge); } /* * allocate new entry and link it in. */ if (new_entry == NULL) { new_entry = uvm_mapent_alloc(map, (flags & UVM_FLAG_NOWAIT)); if (__predict_false(new_entry == NULL)) { error = ENOMEM; goto done; } } new_entry->start = start; new_entry->end = new_entry->start + size; new_entry->object.uvm_obj = uobj; new_entry->offset = uoffset; new_entry->etype = newetype; if (flags & UVM_FLAG_NOMERGE) { new_entry->flags |= UVM_MAP_NOMERGE; } new_entry->protection = prot; new_entry->max_protection = maxprot; new_entry->inheritance = inherit; new_entry->wired_count = 0; new_entry->advice = advice; if (flags & UVM_FLAG_OVERLAY) { /* * to_add: for BSS we overallocate a little since we * are likely to extend */ vaddr_t to_add = (flags & UVM_FLAG_AMAPPAD) ? UVM_AMAP_CHUNK << PAGE_SHIFT : 0; struct vm_amap *amap = amap_alloc(size, to_add, (flags & UVM_FLAG_NOWAIT)); if (__predict_false(amap == NULL)) { error = ENOMEM; goto done; } new_entry->aref.ar_pageoff = 0; new_entry->aref.ar_amap = amap; } else { new_entry->aref.ar_pageoff = 0; new_entry->aref.ar_amap = NULL; } uvm_map_entry_link(map, prev_entry, new_entry); /* * Update the free space hint */ if ((map->first_free == prev_entry) && (prev_entry->end >= new_entry->start)) map->first_free = new_entry; new_entry = NULL; } map->size += size; UVMHIST_LOG(maphist,"<- done!", 0, 0, 0, 0); error = 0; done: vm_map_unlock(map); if (new_entry) { uvm_mapent_free(new_entry); } if (dead) { KDASSERT(merged); uvm_mapent_free(dead); } if (dead_entries) uvm_unmap_detach(dead_entries, 0); return error; } /* * uvm_map_lookup_entry_bytree: lookup an entry in tree */ static inline bool uvm_map_lookup_entry_bytree(struct vm_map *map, vaddr_t address, struct vm_map_entry **entry /* OUT */) { struct vm_map_entry *prev = &map->header; struct vm_map_entry *cur = ROOT_ENTRY(map); while (cur) { UVMMAP_EVCNT_INCR(mlk_treeloop); if (address >= cur->start) { if (address < cur->end) { *entry = cur; return true; } prev = cur; cur = RIGHT_ENTRY(cur); } else cur = LEFT_ENTRY(cur); } *entry = prev; return false; } /* * uvm_map_lookup_entry: find map entry at or before an address * * => map must at least be read-locked by caller * => entry is returned in "entry" * => return value is true if address is in the returned entry */ bool uvm_map_lookup_entry(struct vm_map *map, vaddr_t address, struct vm_map_entry **entry /* OUT */) { struct vm_map_entry *cur; bool use_tree = false; UVMHIST_FUNC("uvm_map_lookup_entry"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(map=%#jx,addr=%#jx,ent=%#jx)", (uintptr_t)map, address, (uintptr_t)entry, 0); /* * start looking either from the head of the * list, or from the hint. */ cur = map->hint; if (cur == &map->header) cur = cur->next; UVMMAP_EVCNT_INCR(mlk_call); if (address >= cur->start) { /* * go from hint to end of list. * * but first, make a quick check to see if * we are already looking at the entry we * want (which is usually the case). * note also that we don't need to save the hint * here... it is the same hint (unless we are * at the header, in which case the hint didn't * buy us anything anyway). */ if (cur != &map->header && cur->end > address) { UVMMAP_EVCNT_INCR(mlk_hint); *entry = cur; UVMHIST_LOG(maphist,"<- got it via hint (%#jx)", (uintptr_t)cur, 0, 0, 0); uvm_mapent_check(*entry); return (true); } if (map->nentries > 15) use_tree = true; } else { /* * invalid hint. use tree. */ use_tree = true; } uvm_map_check(map, __func__); if (use_tree) { /* * Simple lookup in the tree. Happens when the hint is * invalid, or nentries reach a threshold. */ UVMMAP_EVCNT_INCR(mlk_tree); if (uvm_map_lookup_entry_bytree(map, address, entry)) { goto got; } else { goto failed; } } /* * search linearly */ UVMMAP_EVCNT_INCR(mlk_list); while (cur != &map->header) { UVMMAP_EVCNT_INCR(mlk_listloop); if (cur->end > address) { if (address >= cur->start) { /* * save this lookup for future * hints, and return */ *entry = cur; got: SAVE_HINT(map, map->hint, *entry); UVMHIST_LOG(maphist,"<- search got it (%#jx)", (uintptr_t)cur, 0, 0, 0); KDASSERT((*entry)->start <= address); KDASSERT(address < (*entry)->end); uvm_mapent_check(*entry); return (true); } break; } cur = cur->next; } *entry = cur->prev; failed: SAVE_HINT(map, map->hint, *entry); UVMHIST_LOG(maphist,"<- failed!",0,0,0,0); KDASSERT((*entry) == &map->header || (*entry)->end <= address); KDASSERT((*entry)->next == &map->header || address < (*entry)->next->start); return (false); } /* * See if the range between start and start + length fits in the gap * entry->next->start and entry->end. Returns 1 if fits, 0 if doesn't * fit, and -1 address wraps around. */ static int uvm_map_space_avail(vaddr_t *start, vsize_t length, voff_t uoffset, vsize_t align, int flags, int topdown, struct vm_map_entry *entry) { vaddr_t end; #ifdef PMAP_PREFER /* * push start address forward as needed to avoid VAC alias problems. * we only do this if a valid offset is specified. */ if (uoffset != UVM_UNKNOWN_OFFSET) PMAP_PREFER(uoffset, start, length, topdown); #endif if ((flags & UVM_FLAG_COLORMATCH) != 0) { KASSERT(align < uvmexp.ncolors); if (uvmexp.ncolors > 1) { const u_int colormask = uvmexp.colormask; const u_int colorsize = colormask + 1; vaddr_t hint = atop(*start); const u_int color = hint & colormask; if (color != align) { hint -= color; /* adjust to color boundary */ KASSERT((hint & colormask) == 0); if (topdown) { if (align > color) hint -= colorsize; } else { if (align < color) hint += colorsize; } *start = ptoa(hint + align); /* adjust to color */ } } } else { KASSERT(powerof2(align)); uvm_map_align_va(start, align, topdown); /* * XXX Should we PMAP_PREFER() here again? * eh...i think we're okay */ } /* * Find the end of the proposed new region. Be sure we didn't * wrap around the address; if so, we lose. Otherwise, if the * proposed new region fits before the next entry, we win. */ end = *start + length; if (end < *start) return (-1); if (entry->next->start >= end && *start >= entry->end) return (1); return (0); } /* * uvm_map_findspace: find "length" sized space in "map". * * => "hint" is a hint about where we want it, unless UVM_FLAG_FIXED is * set in "flags" (in which case we insist on using "hint"). * => "result" is VA returned * => uobj/uoffset are to be used to handle VAC alignment, if required * => if "align" is non-zero, we attempt to align to that value. * => caller must at least have read-locked map * => returns NULL on failure, or pointer to prev. map entry if success * => note this is a cross between the old vm_map_findspace and vm_map_find */ struct vm_map_entry * uvm_map_findspace(struct vm_map *map, vaddr_t hint, vsize_t length, vaddr_t *result /* OUT */, struct uvm_object *uobj, voff_t uoffset, vsize_t align, int flags) { struct vm_map_entry *entry; struct vm_map_entry *child, *prev, *tmp; vaddr_t orig_hint __diagused; const int topdown = map->flags & VM_MAP_TOPDOWN; UVMHIST_FUNC("uvm_map_findspace"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist, "(map=%#jx, hint=%#jx, len=%ju, flags=%#jx)", (uintptr_t)map, hint, length, flags); KASSERT((flags & UVM_FLAG_COLORMATCH) != 0 || powerof2(align)); KASSERT((flags & UVM_FLAG_COLORMATCH) == 0 || align < uvmexp.ncolors); KASSERT((flags & UVM_FLAG_FIXED) == 0 || align == 0); uvm_map_check(map, "map_findspace entry"); /* * Clamp the hint to the VM map's min/max address, and remmeber * the clamped original hint. Remember the original hint, * clamped to the min/max address. If we are aligning, then we * may have to try again with no alignment constraint if we * fail the first time. * * We use the original hint to verify later that the search has * been monotonic -- that is, nonincreasing or nondecreasing, * according to topdown or !topdown respectively. But the * clamping is not monotonic. */ if (hint < vm_map_min(map)) { /* check ranges ... */ if (flags & UVM_FLAG_FIXED) { UVMHIST_LOG(maphist,"<- VA below map range",0,0,0,0); return (NULL); } hint = vm_map_min(map); } if (hint > vm_map_max(map)) { UVMHIST_LOG(maphist,"<- VA %#jx > range [%#jx->%#jx]", hint, vm_map_min(map), vm_map_max(map), 0); return (NULL); } orig_hint = hint; /* * hint may not be aligned properly; we need round up or down it * before proceeding further. */ if ((flags & UVM_FLAG_COLORMATCH) == 0) uvm_map_align_va(&hint, align, topdown); /* * Look for the first possible address; if there's already * something at this address, we have to start after it. */ /* * @@@: there are four, no, eight cases to consider. * * 0: found, fixed, bottom up -> fail * 1: found, fixed, top down -> fail * 2: found, not fixed, bottom up -> start after entry->end, * loop up * 3: found, not fixed, top down -> start before entry->start, * loop down * 4: not found, fixed, bottom up -> check entry->next->start, fail * 5: not found, fixed, top down -> check entry->next->start, fail * 6: not found, not fixed, bottom up -> check entry->next->start, * loop up * 7: not found, not fixed, top down -> check entry->next->start, * loop down * * as you can see, it reduces to roughly five cases, and that * adding top down mapping only adds one unique case (without * it, there would be four cases). */ if ((flags & UVM_FLAG_FIXED) == 0 && hint == (topdown ? vm_map_max(map) : vm_map_min(map))) { /* * The uvm_map_findspace algorithm is monotonic -- for * topdown VM it starts with a high hint and returns a * lower free address; for !topdown VM it starts with a * low hint and returns a higher free address. As an * optimization, start with the first (highest for * topdown, lowest for !topdown) free address. * * XXX This `optimization' probably doesn't actually do * much in practice unless userland explicitly passes * the VM map's minimum or maximum address, which * varies from machine to machine (VM_MAX/MIN_ADDRESS, * e.g. 0x7fbfdfeff000 on amd64 but 0xfffffffff000 on * aarch64) and may vary according to other factors * like sysctl vm.user_va0_disable. In particular, if * the user specifies 0 as a hint to mmap, then mmap * will choose a default address which is usually _not_ * VM_MAX/MIN_ADDRESS but something else instead like * VM_MAX_ADDRESS - stack size - guard page overhead, * in which case this branch is never hit. * * In fact, this branch appears to have been broken for * two decades between when topdown was introduced in * ~2003 and when it was adapted to handle the topdown * case without violating the monotonicity assertion in * 2022. Maybe Someone^TM should either ditch the * optimization or find a better way to do it. */ entry = map->first_free; } else { if (uvm_map_lookup_entry(map, hint, &entry)) { /* "hint" address already in use ... */ if (flags & UVM_FLAG_FIXED) { UVMHIST_LOG(maphist, "<- fixed & VA in use", 0, 0, 0, 0); return (NULL); } if (topdown) /* Start from lower gap. */ entry = entry->prev; } else if (flags & UVM_FLAG_FIXED) { if (entry->next->start >= hint + length && hint + length > hint) goto found; /* "hint" address is gap but too small */ UVMHIST_LOG(maphist, "<- fixed mapping failed", 0, 0, 0, 0); return (NULL); /* only one shot at it ... */ } else { /* * See if given hint fits in this gap. */ switch (uvm_map_space_avail(&hint, length, uoffset, align, flags, topdown, entry)) { case 1: goto found; case -1: goto wraparound; } if (topdown) { /* * Still there is a chance to fit * if hint > entry->end. */ } else { /* Start from higher gap. */ entry = entry->next; if (entry == &map->header) goto notfound; goto nextgap; } } } /* * Note that all UVM_FLAGS_FIXED case is already handled. */ KDASSERT((flags & UVM_FLAG_FIXED) == 0); /* Try to find the space in the red-black tree */ /* Check slot before any entry */ if (topdown) { KASSERTMSG(entry->next->start >= vm_map_min(map), "map=%p entry=%p entry->next=%p" " entry->next->start=0x%"PRIxVADDR" min=0x%"PRIxVADDR, map, entry, entry->next, entry->next->start, vm_map_min(map)); if (length > entry->next->start - vm_map_min(map)) hint = vm_map_min(map); /* XXX goto wraparound? */ else hint = entry->next->start - length; KASSERT(hint >= vm_map_min(map)); } else { hint = entry->end; } switch (uvm_map_space_avail(&hint, length, uoffset, align, flags, topdown, entry)) { case 1: goto found; case -1: goto wraparound; } nextgap: KDASSERT((flags & UVM_FLAG_FIXED) == 0); /* If there is not enough space in the whole tree, we fail */ tmp = ROOT_ENTRY(map); if (tmp == NULL || tmp->maxgap < length) goto notfound; prev = NULL; /* previous candidate */ /* Find an entry close to hint that has enough space */ for (; tmp;) { KASSERT(tmp->next->start == tmp->end + tmp->gap); if (topdown) { if (tmp->next->start < hint + length && (prev == NULL || tmp->end > prev->end)) { if (tmp->gap >= length) prev = tmp; else if ((child = LEFT_ENTRY(tmp)) != NULL && child->maxgap >= length) prev = tmp; } } else { if (tmp->end >= hint && (prev == NULL || tmp->end < prev->end)) { if (tmp->gap >= length) prev = tmp; else if ((child = RIGHT_ENTRY(tmp)) != NULL && child->maxgap >= length) prev = tmp; } } if (tmp->next->start < hint + length) child = RIGHT_ENTRY(tmp); else if (tmp->end > hint) child = LEFT_ENTRY(tmp); else { if (tmp->gap >= length) break; if (topdown) child = LEFT_ENTRY(tmp); else child = RIGHT_ENTRY(tmp); } if (child == NULL || child->maxgap < length) break; tmp = child; } if (tmp != NULL && tmp->start < hint && hint < tmp->next->start) { /* * Check if the entry that we found satifies the * space requirement */ if (topdown) { if (hint > tmp->next->start - length) hint = tmp->next->start - length; } else { if (hint < tmp->end) hint = tmp->end; } switch (uvm_map_space_avail(&hint, length, uoffset, align, flags, topdown, tmp)) { case 1: entry = tmp; goto found; case -1: goto wraparound; } if (tmp->gap >= length) goto listsearch; } if (prev == NULL) goto notfound; if (topdown) { KASSERT(orig_hint >= prev->next->start - length || prev->next->start - length > prev->next->start); hint = prev->next->start - length; } else { KASSERT(orig_hint <= prev->end); hint = prev->end; } switch (uvm_map_space_avail(&hint, length, uoffset, align, flags, topdown, prev)) { case 1: entry = prev; goto found; case -1: goto wraparound; } if (prev->gap >= length) goto listsearch; if (topdown) tmp = LEFT_ENTRY(prev); else tmp = RIGHT_ENTRY(prev); for (;;) { KASSERT(tmp && tmp->maxgap >= length); if (topdown) child = RIGHT_ENTRY(tmp); else child = LEFT_ENTRY(tmp); if (child && child->maxgap >= length) { tmp = child; continue; } if (tmp->gap >= length) break; if (topdown) tmp = LEFT_ENTRY(tmp); else tmp = RIGHT_ENTRY(tmp); } if (topdown) { KASSERT(orig_hint >= tmp->next->start - length || tmp->next->start - length > tmp->next->start); hint = tmp->next->start - length; } else { KASSERT(orig_hint <= tmp->end); hint = tmp->end; } switch (uvm_map_space_avail(&hint, length, uoffset, align, flags, topdown, tmp)) { case 1: entry = tmp; goto found; case -1: goto wraparound; } /* * The tree fails to find an entry because of offset or alignment * restrictions. Search the list instead. */ listsearch: /* * Look through the rest of the map, trying to fit a new region in * the gap between existing regions, or after the very last region. * note: entry->end = base VA of current gap, * entry->next->start = VA of end of current gap */ for (;;) { /* Update hint for current gap. */ hint = topdown ? entry->next->start - length : entry->end; /* See if it fits. */ switch (uvm_map_space_avail(&hint, length, uoffset, align, flags, topdown, entry)) { case 1: goto found; case -1: goto wraparound; } /* Advance to next/previous gap */ if (topdown) { if (entry == &map->header) { UVMHIST_LOG(maphist, "<- failed (off start)", 0,0,0,0); goto notfound; } entry = entry->prev; } else { entry = entry->next; if (entry == &map->header) { UVMHIST_LOG(maphist, "<- failed (off end)", 0,0,0,0); goto notfound; } } } found: SAVE_HINT(map, map->hint, entry); *result = hint; UVMHIST_LOG(maphist,"<- got it! (result=%#jx)", hint, 0,0,0); KASSERTMSG( topdown || hint >= orig_hint, "hint: %jx, orig_hint: %jx", (uintmax_t)hint, (uintmax_t)orig_hint); KASSERTMSG(!topdown || hint <= orig_hint, "hint: %jx, orig_hint: %jx", (uintmax_t)hint, (uintmax_t)orig_hint); KASSERT(entry->end <= hint); KASSERT(hint + length <= entry->next->start); return (entry); wraparound: UVMHIST_LOG(maphist, "<- failed (wrap around)", 0,0,0,0); return (NULL); notfound: UVMHIST_LOG(maphist, "<- failed (notfound)", 0,0,0,0); return (NULL); } /* * U N M A P - m a i n h e l p e r f u n c t i o n s */ /* * uvm_unmap_remove: remove mappings from a vm_map (from "start" up to "stop") * * => caller must check alignment and size * => map must be locked by caller * => we return a list of map entries that we've remove from the map * in "entry_list" */ void uvm_unmap_remove(struct vm_map *map, vaddr_t start, vaddr_t end, struct vm_map_entry **entry_list /* OUT */, int flags) { struct vm_map_entry *entry, *first_entry, *next; vaddr_t len; UVMHIST_FUNC("uvm_unmap_remove"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(map=%#jx, start=%#jx, end=%#jx)", (uintptr_t)map, start, end, 0); VM_MAP_RANGE_CHECK(map, start, end); uvm_map_check(map, "unmap_remove entry"); /* * find first entry */ if (uvm_map_lookup_entry(map, start, &first_entry) == true) { /* clip and go... */ entry = first_entry; UVM_MAP_CLIP_START(map, entry, start); /* critical! prevents stale hint */ SAVE_HINT(map, entry, entry->prev); } else { entry = first_entry->next; } /* * Save the free space hint */ if (map->first_free != &map->header && map->first_free->start >= start) map->first_free = entry->prev; /* * note: we now re-use first_entry for a different task. we remove * a number of map entries from the map and save them in a linked * list headed by "first_entry". once we remove them from the map * the caller should unlock the map and drop the references to the * backing objects [c.f. uvm_unmap_detach]. the object is to * separate unmapping from reference dropping. why? * [1] the map has to be locked for unmapping * [2] the map need not be locked for reference dropping * [3] dropping references may trigger pager I/O, and if we hit * a pager that does synchronous I/O we may have to wait for it. * [4] we would like all waiting for I/O to occur with maps unlocked * so that we don't block other threads. */ first_entry = NULL; *entry_list = NULL; /* * break up the area into map entry sized regions and unmap. note * that all mappings have to be removed before we can even consider * dropping references to amaps or VM objects (otherwise we could end * up with a mapping to a page on the free list which would be very bad) */ while ((entry != &map->header) && (entry->start < end)) { KASSERT((entry->flags & UVM_MAP_STATIC) == 0); UVM_MAP_CLIP_END(map, entry, end); next = entry->next; len = entry->end - entry->start; /* * unwire before removing addresses from the pmap; otherwise * unwiring will put the entries back into the pmap (XXX). */ if (VM_MAPENT_ISWIRED(entry)) { uvm_map_entry_unwire(map, entry); } if (flags & UVM_FLAG_VAONLY) { /* nothing */ } else if ((map->flags & VM_MAP_PAGEABLE) == 0) { /* * if the map is non-pageable, any pages mapped there * must be wired and entered with pmap_kenter_pa(), * and we should free any such pages immediately. * this is mostly used for kmem_map. */ KASSERT(vm_map_pmap(map) == pmap_kernel()); uvm_km_pgremove_intrsafe(map, entry->start, entry->end); } else if (UVM_ET_ISOBJ(entry) && UVM_OBJ_IS_KERN_OBJECT(entry->object.uvm_obj)) { panic("%s: kernel object %p %p\n", __func__, map, entry); } else if (UVM_ET_ISOBJ(entry) || entry->aref.ar_amap) { /* * remove mappings the standard way. lock object * and/or amap to ensure vm_page state does not * change while in pmap_remove(). */ uvm_map_lock_entry(entry); pmap_remove(map->pmap, entry->start, entry->end); uvm_map_unlock_entry(entry); } #if defined(UVMDEBUG) /* * check if there's remaining mapping, * which is a bug in caller. */ vaddr_t va; for (va = entry->start; va < entry->end; va += PAGE_SIZE) { if (pmap_extract(vm_map_pmap(map), va, NULL)) { panic("%s: %#"PRIxVADDR" has mapping", __func__, va); } } if (VM_MAP_IS_KERNEL(map) && (flags & UVM_FLAG_NOWAIT) == 0) { uvm_km_check_empty(map, entry->start, entry->end); } #endif /* defined(UVMDEBUG) */ /* * remove entry from map and put it on our list of entries * that we've nuked. then go to next entry. */ UVMHIST_LOG(maphist, " removed map entry %#jx", (uintptr_t)entry, 0, 0, 0); /* critical! prevents stale hint */ SAVE_HINT(map, entry, entry->prev); uvm_map_entry_unlink(map, entry); KASSERT(map->size >= len); map->size -= len; entry->prev = NULL; entry->next = first_entry; first_entry = entry; entry = next; } /* * Note: if map is dying, leave pmap_update() for pmap_destroy(), * which will be called later. */ if ((map->flags & VM_MAP_DYING) == 0) { pmap_update(vm_map_pmap(map)); } else { KASSERT(vm_map_pmap(map) != pmap_kernel()); } uvm_map_check(map, "unmap_remove leave"); /* * now we've cleaned up the map and are ready for the caller to drop * references to the mapped objects. */ *entry_list = first_entry; UVMHIST_LOG(maphist,"<- done!", 0, 0, 0, 0); if (map->flags & VM_MAP_WANTVA) { mutex_enter(&map->misc_lock); map->flags &= ~VM_MAP_WANTVA; cv_broadcast(&map->cv); mutex_exit(&map->misc_lock); } } /* * uvm_unmap_detach: drop references in a chain of map entries * * => we will free the map entries as we traverse the list. */ void uvm_unmap_detach(struct vm_map_entry *first_entry, int flags) { struct vm_map_entry *next_entry; UVMHIST_FUNC("uvm_unmap_detach"); UVMHIST_CALLED(maphist); while (first_entry) { KASSERT(!VM_MAPENT_ISWIRED(first_entry)); UVMHIST_LOG(maphist, " detach %#jx: amap=%#jx, obj=%#jx, submap?=%jd", (uintptr_t)first_entry, (uintptr_t)first_entry->aref.ar_amap, (uintptr_t)first_entry->object.uvm_obj, UVM_ET_ISSUBMAP(first_entry)); /* * drop reference to amap, if we've got one */ if (first_entry->aref.ar_amap) uvm_map_unreference_amap(first_entry, flags); /* * drop reference to our backing object, if we've got one */ KASSERT(!UVM_ET_ISSUBMAP(first_entry)); if (UVM_ET_ISOBJ(first_entry) && first_entry->object.uvm_obj->pgops->pgo_detach) { (*first_entry->object.uvm_obj->pgops->pgo_detach) (first_entry->object.uvm_obj); } next_entry = first_entry->next; uvm_mapent_free(first_entry); first_entry = next_entry; } UVMHIST_LOG(maphist, "<- done", 0,0,0,0); } /* * E X T R A C T I O N F U N C T I O N S */ /* * uvm_map_reserve: reserve space in a vm_map for future use. * * => we reserve space in a map by putting a dummy map entry in the * map (dummy means obj=NULL, amap=NULL, prot=VM_PROT_NONE) * => map should be unlocked (we will write lock it) * => we return true if we were able to reserve space * => XXXCDC: should be inline? */ int uvm_map_reserve(struct vm_map *map, vsize_t size, vaddr_t offset /* hint for pmap_prefer */, vsize_t align /* alignment */, vaddr_t *raddr /* IN:hint, OUT: reserved VA */, uvm_flag_t flags /* UVM_FLAG_FIXED or UVM_FLAG_COLORMATCH or 0 */) { UVMHIST_FUNC("uvm_map_reserve"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist, "(map=%#jx, size=%#jx, offset=%#jx, addr=%#jx)", (uintptr_t)map, size, offset, (uintptr_t)raddr); size = round_page(size); /* * reserve some virtual space. */ if (uvm_map(map, raddr, size, NULL, offset, align, UVM_MAPFLAG(UVM_PROT_NONE, UVM_PROT_NONE, UVM_INH_NONE, UVM_ADV_RANDOM, UVM_FLAG_NOMERGE|flags)) != 0) { UVMHIST_LOG(maphist, "<- done (no VM)", 0,0,0,0); return (false); } UVMHIST_LOG(maphist, "<- done (*raddr=%#jx)", *raddr,0,0,0); return (true); } /* * uvm_map_replace: replace a reserved (blank) area of memory with * real mappings. * * => caller must WRITE-LOCK the map * => we return true if replacement was a success * => we expect the newents chain to have nnewents entrys on it and * we expect newents->prev to point to the last entry on the list * => note newents is allowed to be NULL */ static int uvm_map_replace(struct vm_map *map, vaddr_t start, vaddr_t end, struct vm_map_entry *newents, int nnewents, vsize_t nsize, struct vm_map_entry **oldentryp) { struct vm_map_entry *oldent, *last; uvm_map_check(map, "map_replace entry"); /* * first find the blank map entry at the specified address */ if (!uvm_map_lookup_entry(map, start, &oldent)) { return (false); } /* * check to make sure we have a proper blank entry */ if (end < oldent->end) { UVM_MAP_CLIP_END(map, oldent, end); } if (oldent->start != start || oldent->end != end || oldent->object.uvm_obj != NULL || oldent->aref.ar_amap != NULL) { return (false); } #ifdef DIAGNOSTIC /* * sanity check the newents chain */ { struct vm_map_entry *tmpent = newents; int nent = 0; vsize_t sz = 0; vaddr_t cur = start; while (tmpent) { nent++; sz += tmpent->end - tmpent->start; if (tmpent->start < cur) panic("uvm_map_replace1"); if (tmpent->start >= tmpent->end || tmpent->end > end) { panic("uvm_map_replace2: " "tmpent->start=%#"PRIxVADDR ", tmpent->end=%#"PRIxVADDR ", end=%#"PRIxVADDR, tmpent->start, tmpent->end, end); } cur = tmpent->end; if (tmpent->next) { if (tmpent->next->prev != tmpent) panic("uvm_map_replace3"); } else { if (newents->prev != tmpent) panic("uvm_map_replace4"); } tmpent = tmpent->next; } if (nent != nnewents) panic("uvm_map_replace5"); if (sz != nsize) panic("uvm_map_replace6"); } #endif /* * map entry is a valid blank! replace it. (this does all the * work of map entry link/unlink...). */ if (newents) { last = newents->prev; /* critical: flush stale hints out of map */ SAVE_HINT(map, map->hint, newents); if (map->first_free == oldent) map->first_free = last; last->next = oldent->next; last->next->prev = last; /* Fix RB tree */ uvm_rb_remove(map, oldent); newents->prev = oldent->prev; newents->prev->next = newents; map->nentries = map->nentries + (nnewents - 1); /* Fixup the RB tree */ { int i; struct vm_map_entry *tmp; tmp = newents; for (i = 0; i < nnewents && tmp; i++) { uvm_rb_insert(map, tmp); tmp = tmp->next; } } } else { /* NULL list of new entries: just remove the old one */ clear_hints(map, oldent); uvm_map_entry_unlink(map, oldent); } map->size -= end - start - nsize; uvm_map_check(map, "map_replace leave"); /* * now we can free the old blank entry and return. */ *oldentryp = oldent; return (true); } /* * uvm_map_extract: extract a mapping from a map and put it somewhere * (maybe removing the old mapping) * * => maps should be unlocked (we will write lock them) * => returns 0 on success, error code otherwise * => start must be page aligned * => len must be page sized * => flags: * UVM_EXTRACT_REMOVE: remove mappings from srcmap * UVM_EXTRACT_CONTIG: abort if unmapped area (advisory only) * UVM_EXTRACT_QREF: for a temporary extraction do quick obj refs * UVM_EXTRACT_FIXPROT: set prot to maxprot as we go * UVM_EXTRACT_PROT_ALL: set prot to UVM_PROT_ALL as we go * >>>NOTE: if you set REMOVE, you are not allowed to use CONTIG or QREF!<<< * >>>NOTE: QREF's must be unmapped via the QREF path, thus should only * be used from within the kernel in a kernel level map <<< */ int uvm_map_extract(struct vm_map *srcmap, vaddr_t start, vsize_t len, struct vm_map *dstmap, vaddr_t *dstaddrp, int flags) { vaddr_t dstaddr, end, newend, oldoffset, fudge, orig_fudge; struct vm_map_entry *chain, *endchain, *entry, *orig_entry, *newentry, *deadentry, *oldentry; struct vm_map_entry *resentry = NULL; /* a dummy reservation entry */ vsize_t elen __unused; int nchain, error, copy_ok; vsize_t nsize; UVMHIST_FUNC("uvm_map_extract"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(srcmap=%#jx,start=%#jx, len=%#jx", (uintptr_t)srcmap, start, len, 0); UVMHIST_LOG(maphist," ...,dstmap=%#jx, flags=%#jx)", (uintptr_t)dstmap, flags, 0, 0); /* * step 0: sanity check: start must be on a page boundary, length * must be page sized. can't ask for CONTIG/QREF if you asked for * REMOVE. */ KASSERT((start & PAGE_MASK) == 0 && (len & PAGE_MASK) == 0); KASSERT((flags & UVM_EXTRACT_REMOVE) == 0 || (flags & (UVM_EXTRACT_CONTIG|UVM_EXTRACT_QREF)) == 0); /* * step 1: reserve space in the target map for the extracted area */ if ((flags & UVM_EXTRACT_RESERVED) == 0) { dstaddr = vm_map_min(dstmap); if (!uvm_map_reserve(dstmap, len, start, atop(start) & uvmexp.colormask, &dstaddr, UVM_FLAG_COLORMATCH)) return (ENOMEM); KASSERT((atop(start ^ dstaddr) & uvmexp.colormask) == 0); *dstaddrp = dstaddr; /* pass address back to caller */ UVMHIST_LOG(maphist, " dstaddr=%#jx", dstaddr,0,0,0); } else { dstaddr = *dstaddrp; } /* * step 2: setup for the extraction process loop by init'ing the * map entry chain, locking src map, and looking up the first useful * entry in the map. */ end = start + len; newend = dstaddr + len; chain = endchain = NULL; nchain = 0; nsize = 0; vm_map_lock(srcmap); if (uvm_map_lookup_entry(srcmap, start, &entry)) { /* "start" is within an entry */ if (flags & UVM_EXTRACT_QREF) { /* * for quick references we don't clip the entry, so * the entry may map space "before" the starting * virtual address... this is the "fudge" factor * (which can be non-zero only the first time * through the "while" loop in step 3). */ fudge = start - entry->start; } else { /* * normal reference: we clip the map to fit (thus * fudge is zero) */ UVM_MAP_CLIP_START(srcmap, entry, start); SAVE_HINT(srcmap, srcmap->hint, entry->prev); fudge = 0; } } else { /* "start" is not within an entry ... skip to next entry */ if (flags & UVM_EXTRACT_CONTIG) { error = EINVAL; goto bad; /* definite hole here ... */ } entry = entry->next; fudge = 0; } /* save values from srcmap for step 6 */ orig_entry = entry; orig_fudge = fudge; /* * step 3: now start looping through the map entries, extracting * as we go. */ while (entry->start < end && entry != &srcmap->header) { /* if we are not doing a quick reference, clip it */ if ((flags & UVM_EXTRACT_QREF) == 0) UVM_MAP_CLIP_END(srcmap, entry, end); /* clear needs_copy (allow chunking) */ if (UVM_ET_ISNEEDSCOPY(entry)) { amap_copy(srcmap, entry, AMAP_COPY_NOWAIT|AMAP_COPY_NOMERGE, start, end); if (UVM_ET_ISNEEDSCOPY(entry)) { /* failed? */ error = ENOMEM; goto bad; } /* amap_copy could clip (during chunk)! update fudge */ if (fudge) { fudge = start - entry->start; orig_fudge = fudge; } } /* calculate the offset of this from "start" */ oldoffset = (entry->start + fudge) - start; /* allocate a new map entry */ newentry = uvm_mapent_alloc(dstmap, 0); if (newentry == NULL) { error = ENOMEM; goto bad; } /* set up new map entry */ newentry->next = NULL; newentry->prev = endchain; newentry->start = dstaddr + oldoffset; newentry->end = newentry->start + (entry->end - (entry->start + fudge)); if (newentry->end > newend || newentry->end < newentry->start) newentry->end = newend; newentry->object.uvm_obj = entry->object.uvm_obj; if (newentry->object.uvm_obj) { if (newentry->object.uvm_obj->pgops->pgo_reference) newentry->object.uvm_obj->pgops-> pgo_reference(newentry->object.uvm_obj); newentry->offset = entry->offset + fudge; } else { newentry->offset = 0; } newentry->etype = entry->etype; if (flags & UVM_EXTRACT_PROT_ALL) { newentry->protection = newentry->max_protection = UVM_PROT_ALL; } else { newentry->protection = (flags & UVM_EXTRACT_FIXPROT) ? entry->max_protection : entry->protection; newentry->max_protection = entry->max_protection; } newentry->inheritance = entry->inheritance; newentry->wired_count = 0; newentry->aref.ar_amap = entry->aref.ar_amap; if (newentry->aref.ar_amap) { newentry->aref.ar_pageoff = entry->aref.ar_pageoff + (fudge >> PAGE_SHIFT); uvm_map_reference_amap(newentry, AMAP_SHARED | ((flags & UVM_EXTRACT_QREF) ? AMAP_REFALL : 0)); } else { newentry->aref.ar_pageoff = 0; } newentry->advice = entry->advice; if ((flags & UVM_EXTRACT_QREF) != 0) { newentry->flags |= UVM_MAP_NOMERGE; } /* now link it on the chain */ nchain++; nsize += newentry->end - newentry->start; if (endchain == NULL) { chain = endchain = newentry; } else { endchain->next = newentry; endchain = newentry; } /* end of 'while' loop! */ if ((flags & UVM_EXTRACT_CONTIG) && entry->end < end && (entry->next == &srcmap->header || entry->next->start != entry->end)) { error = EINVAL; goto bad; } entry = entry->next; fudge = 0; } /* * step 4: close off chain (in format expected by uvm_map_replace) */ if (chain) chain->prev = endchain; /* * step 5: attempt to lock the dest map so we can pmap_copy. * note usage of copy_ok: * 1 => dstmap locked, pmap_copy ok, and we "replace" here (step 5) * 0 => dstmap unlocked, NO pmap_copy, and we will "replace" in step 7 */ if (srcmap == dstmap || vm_map_lock_try(dstmap) == true) { copy_ok = 1; if (!uvm_map_replace(dstmap, dstaddr, dstaddr+len, chain, nchain, nsize, &resentry)) { if (srcmap != dstmap) vm_map_unlock(dstmap); error = EIO; goto bad; } } else { copy_ok = 0; /* replace defered until step 7 */ } /* * step 6: traverse the srcmap a second time to do the following: * - if we got a lock on the dstmap do pmap_copy * - if UVM_EXTRACT_REMOVE remove the entries * we make use of orig_entry and orig_fudge (saved in step 2) */ if (copy_ok || (flags & UVM_EXTRACT_REMOVE)) { /* purge possible stale hints from srcmap */ if (flags & UVM_EXTRACT_REMOVE) { SAVE_HINT(srcmap, srcmap->hint, orig_entry->prev); if (srcmap->first_free != &srcmap->header && srcmap->first_free->start >= start) srcmap->first_free = orig_entry->prev; } entry = orig_entry; fudge = orig_fudge; deadentry = NULL; /* for UVM_EXTRACT_REMOVE */ while (entry->start < end && entry != &srcmap->header) { if (copy_ok) { oldoffset = (entry->start + fudge) - start; elen = MIN(end, entry->end) - (entry->start + fudge); pmap_copy(dstmap->pmap, srcmap->pmap, dstaddr + oldoffset, elen, entry->start + fudge); } /* we advance "entry" in the following if statement */ if (flags & UVM_EXTRACT_REMOVE) { uvm_map_lock_entry(entry); pmap_remove(srcmap->pmap, entry->start, entry->end); uvm_map_unlock_entry(entry); oldentry = entry; /* save entry */ entry = entry->next; /* advance */ uvm_map_entry_unlink(srcmap, oldentry); /* add to dead list */ oldentry->next = deadentry; deadentry = oldentry; } else { entry = entry->next; /* advance */ } /* end of 'while' loop */ fudge = 0; } pmap_update(srcmap->pmap); /* * unlock dstmap. we will dispose of deadentry in * step 7 if needed */ if (copy_ok && srcmap != dstmap) vm_map_unlock(dstmap); } else { deadentry = NULL; } /* * step 7: we are done with the source map, unlock. if copy_ok * is 0 then we have not replaced the dummy mapping in dstmap yet * and we need to do so now. */ vm_map_unlock(srcmap); if ((flags & UVM_EXTRACT_REMOVE) && deadentry) uvm_unmap_detach(deadentry, 0); /* dispose of old entries */ /* now do the replacement if we didn't do it in step 5 */ if (copy_ok == 0) { vm_map_lock(dstmap); error = uvm_map_replace(dstmap, dstaddr, dstaddr+len, chain, nchain, nsize, &resentry); vm_map_unlock(dstmap); if (error == false) { error = EIO; goto bad2; } } if (resentry != NULL) uvm_mapent_free(resentry); return (0); /* * bad: failure recovery */ bad: vm_map_unlock(srcmap); bad2: /* src already unlocked */ if (chain) uvm_unmap_detach(chain, (flags & UVM_EXTRACT_QREF) ? AMAP_REFALL : 0); if (resentry != NULL) uvm_mapent_free(resentry); if ((flags & UVM_EXTRACT_RESERVED) == 0) { uvm_unmap(dstmap, dstaddr, dstaddr+len); /* ??? */ } return (error); } /* end of extraction functions */ /* * uvm_map_submap: punch down part of a map into a submap * * => only the kernel_map is allowed to be submapped * => the purpose of submapping is to break up the locking granularity * of a larger map * => the range specified must have been mapped previously with a uvm_map() * call [with uobj==NULL] to create a blank map entry in the main map. * [And it had better still be blank!] * => maps which contain submaps should never be copied or forked. * => to remove a submap, use uvm_unmap() on the main map * and then uvm_map_deallocate() the submap. * => main map must be unlocked. * => submap must have been init'd and have a zero reference count. * [need not be locked as we don't actually reference it] */ int uvm_map_submap(struct vm_map *map, vaddr_t start, vaddr_t end, struct vm_map *submap) { struct vm_map_entry *entry; int error; vm_map_lock(map); VM_MAP_RANGE_CHECK(map, start, end); if (uvm_map_lookup_entry(map, start, &entry)) { UVM_MAP_CLIP_START(map, entry, start); UVM_MAP_CLIP_END(map, entry, end); /* to be safe */ } else { entry = NULL; } if (entry != NULL && entry->start == start && entry->end == end && entry->object.uvm_obj == NULL && entry->aref.ar_amap == NULL && !UVM_ET_ISCOPYONWRITE(entry) && !UVM_ET_ISNEEDSCOPY(entry)) { entry->etype |= UVM_ET_SUBMAP; entry->object.sub_map = submap; entry->offset = 0; uvm_map_reference(submap); error = 0; } else { error = EINVAL; } vm_map_unlock(map); return error; } /* * uvm_map_protect_user: change map protection on behalf of the user. * Enforces PAX settings as necessary. */ int uvm_map_protect_user(struct lwp *l, vaddr_t start, vaddr_t end, vm_prot_t new_prot) { int error; if ((error = PAX_MPROTECT_VALIDATE(l, new_prot))) return error; return uvm_map_protect(&l->l_proc->p_vmspace->vm_map, start, end, new_prot, false); } /* * uvm_map_protect: change map protection * * => set_max means set max_protection. * => map must be unlocked. */ #define MASK(entry) (UVM_ET_ISCOPYONWRITE(entry) ? \ ~VM_PROT_WRITE : VM_PROT_ALL) int uvm_map_protect(struct vm_map *map, vaddr_t start, vaddr_t end, vm_prot_t new_prot, bool set_max) { struct vm_map_entry *current, *entry; int error = 0; UVMHIST_FUNC("uvm_map_protect"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(map=%#jx,start=%#jx,end=%#jx,new_prot=%#jx)", (uintptr_t)map, start, end, new_prot); vm_map_lock(map); VM_MAP_RANGE_CHECK(map, start, end); if (uvm_map_lookup_entry(map, start, &entry)) { UVM_MAP_CLIP_START(map, entry, start); } else { entry = entry->next; } /* * make a first pass to check for protection violations. */ current = entry; while ((current != &map->header) && (current->start < end)) { if (UVM_ET_ISSUBMAP(current)) { error = EINVAL; goto out; } if ((new_prot & current->max_protection) != new_prot) { error = EACCES; goto out; } /* * Don't allow VM_PROT_EXECUTE to be set on entries that * point to vnodes that are associated with a NOEXEC file * system. */ if (UVM_ET_ISOBJ(current) && UVM_OBJ_IS_VNODE(current->object.uvm_obj)) { struct vnode *vp = (struct vnode *) current->object.uvm_obj; if ((new_prot & VM_PROT_EXECUTE) != 0 && (vp->v_mount->mnt_flag & MNT_NOEXEC) != 0) { error = EACCES; goto out; } } current = current->next; } /* go back and fix up protections (no need to clip this time). */ current = entry; while ((current != &map->header) && (current->start < end)) { vm_prot_t old_prot; UVM_MAP_CLIP_END(map, current, end); old_prot = current->protection; if (set_max) current->protection = (current->max_protection = new_prot) & old_prot; else current->protection = new_prot; /* * update physical map if necessary. worry about copy-on-write * here -- CHECK THIS XXX */ if (current->protection != old_prot) { /* update pmap! */ uvm_map_lock_entry(current); pmap_protect(map->pmap, current->start, current->end, current->protection & MASK(current)); uvm_map_unlock_entry(current); /* * If this entry points at a vnode, and the * protection includes VM_PROT_EXECUTE, mark * the vnode as VEXECMAP. */ if (UVM_ET_ISOBJ(current)) { struct uvm_object *uobj = current->object.uvm_obj; if (UVM_OBJ_IS_VNODE(uobj) && (current->protection & VM_PROT_EXECUTE)) { vn_markexec((struct vnode *) uobj); } } } /* * If the map is configured to lock any future mappings, * wire this entry now if the old protection was VM_PROT_NONE * and the new protection is not VM_PROT_NONE. */ if ((map->flags & VM_MAP_WIREFUTURE) != 0 && VM_MAPENT_ISWIRED(current) == 0 && old_prot == VM_PROT_NONE && new_prot != VM_PROT_NONE) { /* * We must call pmap_update() here because the * pmap_protect() call above might have removed some * pmap entries and uvm_map_pageable() might create * some new pmap entries that rely on the prior * removals being completely finished. */ pmap_update(map->pmap); if (uvm_map_pageable(map, current->start, current->end, false, UVM_LK_ENTER|UVM_LK_EXIT) != 0) { /* * If locking the entry fails, remember the * error if it's the first one. Note we * still continue setting the protection in * the map, but will return the error * condition regardless. * * XXX Ignore what the actual error is, * XXX just call it a resource shortage * XXX so that it doesn't get confused * XXX what uvm_map_protect() itself would * XXX normally return. */ error = ENOMEM; } } current = current->next; } pmap_update(map->pmap); out: vm_map_unlock(map); UVMHIST_LOG(maphist, "<- done, error=%jd",error,0,0,0); return error; } #undef MASK /* * uvm_map_inherit: set inheritance code for range of addrs in map. * * => map must be unlocked * => note that the inherit code is used during a "fork". see fork * code for details. */ int uvm_map_inherit(struct vm_map *map, vaddr_t start, vaddr_t end, vm_inherit_t new_inheritance) { struct vm_map_entry *entry, *temp_entry; UVMHIST_FUNC("uvm_map_inherit"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(map=%#jx,start=%#jx,end=%#jx,new_inh=%#jx)", (uintptr_t)map, start, end, new_inheritance); switch (new_inheritance) { case MAP_INHERIT_NONE: case MAP_INHERIT_COPY: case MAP_INHERIT_SHARE: case MAP_INHERIT_ZERO: break; default: UVMHIST_LOG(maphist,"<- done (INVALID ARG)",0,0,0,0); return EINVAL; } vm_map_lock(map); VM_MAP_RANGE_CHECK(map, start, end); if (uvm_map_lookup_entry(map, start, &temp_entry)) { entry = temp_entry; UVM_MAP_CLIP_START(map, entry, start); } else { entry = temp_entry->next; } while ((entry != &map->header) && (entry->start < end)) { UVM_MAP_CLIP_END(map, entry, end); entry->inheritance = new_inheritance; entry = entry->next; } vm_map_unlock(map); UVMHIST_LOG(maphist,"<- done (OK)",0,0,0,0); return 0; } /* * uvm_map_advice: set advice code for range of addrs in map. * * => map must be unlocked */ int uvm_map_advice(struct vm_map *map, vaddr_t start, vaddr_t end, int new_advice) { struct vm_map_entry *entry, *temp_entry; UVMHIST_FUNC("uvm_map_advice"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(map=%#jx,start=%#jx,end=%#jx,new_adv=%#jx)", (uintptr_t)map, start, end, new_advice); vm_map_lock(map); VM_MAP_RANGE_CHECK(map, start, end); if (uvm_map_lookup_entry(map, start, &temp_entry)) { entry = temp_entry; UVM_MAP_CLIP_START(map, entry, start); } else { entry = temp_entry->next; } /* * XXXJRT: disallow holes? */ while ((entry != &map->header) && (entry->start < end)) { UVM_MAP_CLIP_END(map, entry, end); switch (new_advice) { case MADV_NORMAL: case MADV_RANDOM: case MADV_SEQUENTIAL: /* nothing special here */ break; default: vm_map_unlock(map); UVMHIST_LOG(maphist,"<- done (INVALID ARG)",0,0,0,0); return EINVAL; } entry->advice = new_advice; entry = entry->next; } vm_map_unlock(map); UVMHIST_LOG(maphist,"<- done (OK)",0,0,0,0); return 0; } /* * uvm_map_willneed: apply MADV_WILLNEED */ int uvm_map_willneed(struct vm_map *map, vaddr_t start, vaddr_t end) { struct vm_map_entry *entry; UVMHIST_FUNC("uvm_map_willneed"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(map=%#jx,start=%#jx,end=%#jx)", (uintptr_t)map, start, end, 0); vm_map_lock_read(map); VM_MAP_RANGE_CHECK(map, start, end); if (!uvm_map_lookup_entry(map, start, &entry)) { entry = entry->next; } while (entry->start < end) { struct vm_amap * const amap = entry->aref.ar_amap; struct uvm_object * const uobj = entry->object.uvm_obj; KASSERT(entry != &map->header); KASSERT(start < entry->end); /* * For now, we handle only the easy but commonly-requested case. * ie. start prefetching of backing uobj pages. * * XXX It might be useful to pmap_enter() the already-in-core * pages by inventing a "weak" mode for uvm_fault() which would * only do the PGO_LOCKED pgo_get(). */ if (UVM_ET_ISOBJ(entry) && amap == NULL && uobj != NULL) { off_t offset; off_t size; offset = entry->offset; if (start < entry->start) { offset += entry->start - start; } size = entry->offset + (entry->end - entry->start); if (entry->end < end) { size -= end - entry->end; } uvm_readahead(uobj, offset, size); } entry = entry->next; } vm_map_unlock_read(map); UVMHIST_LOG(maphist,"<- done (OK)",0,0,0,0); return 0; } /* * uvm_map_pageable: sets the pageability of a range in a map. * * => wires map entries. should not be used for transient page locking. * for that, use uvm_fault_wire()/uvm_fault_unwire() (see uvm_vslock()). * => regions specified as not pageable require lock-down (wired) memory * and page tables. * => map must never be read-locked * => if islocked is true, map is already write-locked * => we always unlock the map, since we must downgrade to a read-lock * to call uvm_fault_wire() * => XXXCDC: check this and try and clean it up. */ int uvm_map_pageable(struct vm_map *map, vaddr_t start, vaddr_t end, bool new_pageable, int lockflags) { struct vm_map_entry *entry, *start_entry, *failed_entry; int rv; #ifdef DIAGNOSTIC u_int timestamp_save; #endif UVMHIST_FUNC("uvm_map_pageable"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(map=%#jx,start=%#jx,end=%#jx,new_pageable=%ju)", (uintptr_t)map, start, end, new_pageable); KASSERT(map->flags & VM_MAP_PAGEABLE); if ((lockflags & UVM_LK_ENTER) == 0) vm_map_lock(map); VM_MAP_RANGE_CHECK(map, start, end); /* * only one pageability change may take place at one time, since * uvm_fault_wire assumes it will be called only once for each * wiring/unwiring. therefore, we have to make sure we're actually * changing the pageability for the entire region. we do so before * making any changes. */ if (uvm_map_lookup_entry(map, start, &start_entry) == false) { if ((lockflags & UVM_LK_EXIT) == 0) vm_map_unlock(map); UVMHIST_LOG(maphist,"<- done (fault)",0,0,0,0); return EFAULT; } entry = start_entry; if (start == end) { /* nothing required */ if ((lockflags & UVM_LK_EXIT) == 0) vm_map_unlock(map); UVMHIST_LOG(maphist,"<- done (nothing)",0,0,0,0); return 0; } /* * handle wiring and unwiring separately. */ if (new_pageable) { /* unwire */ UVM_MAP_CLIP_START(map, entry, start); /* * unwiring. first ensure that the range to be unwired is * really wired down and that there are no holes. */ while ((entry != &map->header) && (entry->start < end)) { if (entry->wired_count == 0 || (entry->end < end && (entry->next == &map->header || entry->next->start > entry->end))) { if ((lockflags & UVM_LK_EXIT) == 0) vm_map_unlock(map); UVMHIST_LOG(maphist, "<- done (INVAL)",0,0,0,0); return EINVAL; } entry = entry->next; } /* * POSIX 1003.1b - a single munlock call unlocks a region, * regardless of the number of mlock calls made on that * region. */ entry = start_entry; while ((entry != &map->header) && (entry->start < end)) { UVM_MAP_CLIP_END(map, entry, end); if (VM_MAPENT_ISWIRED(entry)) uvm_map_entry_unwire(map, entry); entry = entry->next; } if ((lockflags & UVM_LK_EXIT) == 0) vm_map_unlock(map); UVMHIST_LOG(maphist,"<- done (OK UNWIRE)",0,0,0,0); return 0; } /* * wire case: in two passes [XXXCDC: ugly block of code here] * * 1: holding the write lock, we create any anonymous maps that need * to be created. then we clip each map entry to the region to * be wired and increment its wiring count. * * 2: we downgrade to a read lock, and call uvm_fault_wire to fault * in the pages for any newly wired area (wired_count == 1). * * downgrading to a read lock for uvm_fault_wire avoids a possible * deadlock with another thread that may have faulted on one of * the pages to be wired (it would mark the page busy, blocking * us, then in turn block on the map lock that we hold). because * of problems in the recursive lock package, we cannot upgrade * to a write lock in vm_map_lookup. thus, any actions that * require the write lock must be done beforehand. because we * keep the read lock on the map, the copy-on-write status of the * entries we modify here cannot change. */ while ((entry != &map->header) && (entry->start < end)) { if (VM_MAPENT_ISWIRED(entry) == 0) { /* not already wired? */ /* * perform actions of vm_map_lookup that need the * write lock on the map: create an anonymous map * for a copy-on-write region, or an anonymous map * for a zero-fill region. (XXXCDC: submap case * ok?) */ if (!UVM_ET_ISSUBMAP(entry)) { /* not submap */ if (UVM_ET_ISNEEDSCOPY(entry) && ((entry->max_protection & VM_PROT_WRITE) || (entry->object.uvm_obj == NULL))) { amap_copy(map, entry, 0, start, end); /* XXXCDC: wait OK? */ } } } UVM_MAP_CLIP_START(map, entry, start); UVM_MAP_CLIP_END(map, entry, end); entry->wired_count++; /* * Check for holes */ if (entry->protection == VM_PROT_NONE || (entry->end < end && (entry->next == &map->header || entry->next->start > entry->end))) { /* * found one. amap creation actions do not need to * be undone, but the wired counts need to be restored. */ while (entry != &map->header && entry->end > start) { entry->wired_count--; entry = entry->prev; } if ((lockflags & UVM_LK_EXIT) == 0) vm_map_unlock(map); UVMHIST_LOG(maphist,"<- done (INVALID WIRE)",0,0,0,0); return EINVAL; } entry = entry->next; } /* * Pass 2. */ #ifdef DIAGNOSTIC timestamp_save = map->timestamp; #endif vm_map_busy(map); vm_map_unlock(map); rv = 0; entry = start_entry; while (entry != &map->header && entry->start < end) { if (entry->wired_count == 1) { rv = uvm_fault_wire(map, entry->start, entry->end, entry->max_protection, 1); if (rv) { /* * wiring failed. break out of the loop. * we'll clean up the map below, once we * have a write lock again. */ break; } } entry = entry->next; } if (rv) { /* failed? */ /* * Get back to an exclusive (write) lock. */ vm_map_lock(map); vm_map_unbusy(map); #ifdef DIAGNOSTIC if (timestamp_save + 1 != map->timestamp) panic("uvm_map_pageable: stale map"); #endif /* * first drop the wiring count on all the entries * which haven't actually been wired yet. */ failed_entry = entry; while (entry != &map->header && entry->start < end) { entry->wired_count--; entry = entry->next; } /* * now, unwire all the entries that were successfully * wired above. */ entry = start_entry; while (entry != failed_entry) { entry->wired_count--; if (VM_MAPENT_ISWIRED(entry) == 0) uvm_map_entry_unwire(map, entry); entry = entry->next; } if ((lockflags & UVM_LK_EXIT) == 0) vm_map_unlock(map); UVMHIST_LOG(maphist, "<- done (RV=%jd)", rv,0,0,0); return (rv); } if ((lockflags & UVM_LK_EXIT) == 0) { vm_map_unbusy(map); } else { /* * Get back to an exclusive (write) lock. */ vm_map_lock(map); vm_map_unbusy(map); } UVMHIST_LOG(maphist,"<- done (OK WIRE)",0,0,0,0); return 0; } /* * uvm_map_pageable_all: special case of uvm_map_pageable - affects * all mapped regions. * * => map must not be locked. * => if no flags are specified, all regions are unwired. * => XXXJRT: has some of the same problems as uvm_map_pageable() above. */ int uvm_map_pageable_all(struct vm_map *map, int flags, vsize_t limit) { struct vm_map_entry *entry, *failed_entry; vsize_t size; int rv; #ifdef DIAGNOSTIC u_int timestamp_save; #endif UVMHIST_FUNC("uvm_map_pageable_all"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(map=%#jx,flags=%#jx)", (uintptr_t)map, flags, 0, 0); KASSERT(map->flags & VM_MAP_PAGEABLE); vm_map_lock(map); /* * handle wiring and unwiring separately. */ if (flags == 0) { /* unwire */ /* * POSIX 1003.1b -- munlockall unlocks all regions, * regardless of how many times mlockall has been called. */ for (entry = map->header.next; entry != &map->header; entry = entry->next) { if (VM_MAPENT_ISWIRED(entry)) uvm_map_entry_unwire(map, entry); } map->flags &= ~VM_MAP_WIREFUTURE; vm_map_unlock(map); UVMHIST_LOG(maphist,"<- done (OK UNWIRE)",0,0,0,0); return 0; } if (flags & MCL_FUTURE) { /* * must wire all future mappings; remember this. */ map->flags |= VM_MAP_WIREFUTURE; } if ((flags & MCL_CURRENT) == 0) { /* * no more work to do! */ UVMHIST_LOG(maphist,"<- done (OK no wire)",0,0,0,0); vm_map_unlock(map); return 0; } /* * wire case: in three passes [XXXCDC: ugly block of code here] * * 1: holding the write lock, count all pages mapped by non-wired * entries. if this would cause us to go over our limit, we fail. * * 2: still holding the write lock, we create any anonymous maps that * need to be created. then we increment its wiring count. * * 3: we downgrade to a read lock, and call uvm_fault_wire to fault * in the pages for any newly wired area (wired_count == 1). * * downgrading to a read lock for uvm_fault_wire avoids a possible * deadlock with another thread that may have faulted on one of * the pages to be wired (it would mark the page busy, blocking * us, then in turn block on the map lock that we hold). because * of problems in the recursive lock package, we cannot upgrade * to a write lock in vm_map_lookup. thus, any actions that * require the write lock must be done beforehand. because we * keep the read lock on the map, the copy-on-write status of the * entries we modify here cannot change. */ for (size = 0, entry = map->header.next; entry != &map->header; entry = entry->next) { if (entry->protection != VM_PROT_NONE && VM_MAPENT_ISWIRED(entry) == 0) { /* not already wired? */ size += entry->end - entry->start; } } if (atop(size) + uvmexp.wired > uvmexp.wiredmax) { vm_map_unlock(map); return ENOMEM; } if (limit != 0 && (size + ptoa(pmap_wired_count(vm_map_pmap(map))) > limit)) { vm_map_unlock(map); return ENOMEM; } /* * Pass 2. */ for (entry = map->header.next; entry != &map->header; entry = entry->next) { if (entry->protection == VM_PROT_NONE) continue; if (VM_MAPENT_ISWIRED(entry) == 0) { /* not already wired? */ /* * perform actions of vm_map_lookup that need the * write lock on the map: create an anonymous map * for a copy-on-write region, or an anonymous map * for a zero-fill region. (XXXCDC: submap case * ok?) */ if (!UVM_ET_ISSUBMAP(entry)) { /* not submap */ if (UVM_ET_ISNEEDSCOPY(entry) && ((entry->max_protection & VM_PROT_WRITE) || (entry->object.uvm_obj == NULL))) { amap_copy(map, entry, 0, entry->start, entry->end); /* XXXCDC: wait OK? */ } } } entry->wired_count++; } /* * Pass 3. */ #ifdef DIAGNOSTIC timestamp_save = map->timestamp; #endif vm_map_busy(map); vm_map_unlock(map); rv = 0; for (entry = map->header.next; entry != &map->header; entry = entry->next) { if (entry->wired_count == 1) { rv = uvm_fault_wire(map, entry->start, entry->end, entry->max_protection, 1); if (rv) { /* * wiring failed. break out of the loop. * we'll clean up the map below, once we * have a write lock again. */ break; } } } if (rv) { /* * Get back an exclusive (write) lock. */ vm_map_lock(map); vm_map_unbusy(map); #ifdef DIAGNOSTIC if (timestamp_save + 1 != map->timestamp) panic("uvm_map_pageable_all: stale map"); #endif /* * first drop the wiring count on all the entries * which haven't actually been wired yet. * * Skip VM_PROT_NONE entries like we did above. */ failed_entry = entry; for (/* nothing */; entry != &map->header; entry = entry->next) { if (entry->protection == VM_PROT_NONE) continue; entry->wired_count--; } /* * now, unwire all the entries that were successfully * wired above. * * Skip VM_PROT_NONE entries like we did above. */ for (entry = map->header.next; entry != failed_entry; entry = entry->next) { if (entry->protection == VM_PROT_NONE) continue; entry->wired_count--; if (VM_MAPENT_ISWIRED(entry)) uvm_map_entry_unwire(map, entry); } vm_map_unlock(map); UVMHIST_LOG(maphist,"<- done (RV=%jd)", rv,0,0,0); return (rv); } vm_map_unbusy(map); UVMHIST_LOG(maphist,"<- done (OK WIRE)",0,0,0,0); return 0; } /* * uvm_map_clean: clean out a map range * * => valid flags: * if (flags & PGO_CLEANIT): dirty pages are cleaned first * if (flags & PGO_SYNCIO): dirty pages are written synchronously * if (flags & PGO_DEACTIVATE): any cached pages are deactivated after clean * if (flags & PGO_FREE): any cached pages are freed after clean * => returns an error if any part of the specified range isn't mapped * => never a need to flush amap layer since the anonymous memory has * no permanent home, but may deactivate pages there * => called from sys_msync() and sys_madvise() * => caller must not write-lock map (read OK). * => we may sleep while cleaning if SYNCIO [with map read-locked] */ int uvm_map_clean(struct vm_map *map, vaddr_t start, vaddr_t end, int flags) { struct vm_map_entry *current, *entry; struct uvm_object *uobj; struct vm_amap *amap; struct vm_anon *anon, *anon_tofree; struct vm_page *pg; vaddr_t offset; vsize_t size; voff_t uoff; int error, refs; UVMHIST_FUNC("uvm_map_clean"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(map=%#jx,start=%#jx,end=%#jx,flags=%#jx)", (uintptr_t)map, start, end, flags); KASSERT((flags & (PGO_FREE|PGO_DEACTIVATE)) != (PGO_FREE|PGO_DEACTIVATE)); vm_map_lock_read(map); VM_MAP_RANGE_CHECK(map, start, end); if (uvm_map_lookup_entry(map, start, &entry) == false) { vm_map_unlock_read(map); return EFAULT; } /* * Make a first pass to check for holes and wiring problems. */ for (current = entry; current->start < end; current = current->next) { if (UVM_ET_ISSUBMAP(current)) { vm_map_unlock_read(map); return EINVAL; } if ((flags & PGO_FREE) != 0 && VM_MAPENT_ISWIRED(entry)) { vm_map_unlock_read(map); return EBUSY; } if (end <= current->end) { break; } if (current->end != current->next->start) { vm_map_unlock_read(map); return EFAULT; } } error = 0; for (current = entry; start < end; current = current->next) { amap = current->aref.ar_amap; /* upper layer */ uobj = current->object.uvm_obj; /* lower layer */ KASSERT(start >= current->start); /* * No amap cleaning necessary if: * * (1) There's no amap. * * (2) We're not deactivating or freeing pages. */ if (amap == NULL || (flags & (PGO_DEACTIVATE|PGO_FREE)) == 0) goto flush_object; offset = start - current->start; size = MIN(end, current->end) - start; anon_tofree = NULL; amap_lock(amap); for ( ; size != 0; size -= PAGE_SIZE, offset += PAGE_SIZE) { anon = amap_lookup(¤t->aref, offset); if (anon == NULL) continue; KASSERT(anon->an_lock == amap->am_lock); pg = anon->an_page; if (pg == NULL) { continue; } if (pg->flags & PG_BUSY) { continue; } switch (flags & (PGO_CLEANIT|PGO_FREE|PGO_DEACTIVATE)) { /* * In these first 3 cases, we just deactivate the page. */ case PGO_CLEANIT|PGO_FREE: case PGO_CLEANIT|PGO_DEACTIVATE: case PGO_DEACTIVATE: deactivate_it: /* * skip the page if it's loaned or wired, * since it shouldn't be on a paging queue * at all in these cases. */ mutex_enter(&uvm_pageqlock); if (pg->loan_count != 0 || pg->wire_count != 0) { mutex_exit(&uvm_pageqlock); continue; } KASSERT(pg->uanon == anon); uvm_pagedeactivate(pg); mutex_exit(&uvm_pageqlock); continue; case PGO_FREE: /* * If there are multiple references to * the amap, just deactivate the page. */ if (amap_refs(amap) > 1) goto deactivate_it; /* skip the page if it's wired */ if (pg->wire_count != 0) { continue; } amap_unadd(¤t->aref, offset); refs = --anon->an_ref; if (refs == 0) { anon->an_link = anon_tofree; anon_tofree = anon; } continue; } } uvm_anon_freelst(amap, anon_tofree); flush_object: /* * flush pages if we've got a valid backing object. * note that we must always clean object pages before * freeing them since otherwise we could reveal stale * data from files. */ uoff = current->offset + (start - current->start); size = MIN(end, current->end) - start; if (uobj != NULL) { mutex_enter(uobj->vmobjlock); if (uobj->pgops->pgo_put != NULL) error = (uobj->pgops->pgo_put)(uobj, uoff, uoff + size, flags | PGO_CLEANIT); else error = 0; } start += size; } vm_map_unlock_read(map); return (error); } /* * uvm_map_checkprot: check protection in map * * => must allow specified protection in a fully allocated region. * => map must be read or write locked by caller. */ bool uvm_map_checkprot(struct vm_map *map, vaddr_t start, vaddr_t end, vm_prot_t protection) { struct vm_map_entry *entry; struct vm_map_entry *tmp_entry; if (!uvm_map_lookup_entry(map, start, &tmp_entry)) { return (false); } entry = tmp_entry; while (start < end) { if (entry == &map->header) { return (false); } /* * no holes allowed */ if (start < entry->start) { return (false); } /* * check protection associated with entry */ if ((entry->protection & protection) != protection) { return (false); } start = entry->end; entry = entry->next; } return (true); } /* * uvmspace_alloc: allocate a vmspace structure. * * - structure includes vm_map and pmap * - XXX: no locking on this structure * - refcnt set to 1, rest must be init'd by caller */ struct vmspace * uvmspace_alloc(vaddr_t vmin, vaddr_t vmax, bool topdown) { struct vmspace *vm; UVMHIST_FUNC("uvmspace_alloc"); UVMHIST_CALLED(maphist); vm = pool_cache_get(&uvm_vmspace_cache, PR_WAITOK); uvmspace_init(vm, NULL, vmin, vmax, topdown); UVMHIST_LOG(maphist,"<- done (vm=%#jx)", (uintptr_t)vm, 0, 0, 0); return (vm); } /* * uvmspace_init: initialize a vmspace structure. * * - XXX: no locking on this structure * - refcnt set to 1, rest must be init'd by caller */ void uvmspace_init(struct vmspace *vm, struct pmap *pmap, vaddr_t vmin, vaddr_t vmax, bool topdown) { UVMHIST_FUNC("uvmspace_init"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist, "(vm=%#jx, pmap=%#jx, vmin=%#jx, vmax=%#jx", (uintptr_t)vm, (uintptr_t)pmap, vmin, vmax); UVMHIST_LOG(maphist, " topdown=%ju)", topdown, 0, 0, 0); memset(vm, 0, sizeof(*vm)); uvm_map_setup(&vm->vm_map, vmin, vmax, VM_MAP_PAGEABLE | (topdown ? VM_MAP_TOPDOWN : 0) ); if (pmap) pmap_reference(pmap); else pmap = pmap_create(); vm->vm_map.pmap = pmap; vm->vm_refcnt = 1; UVMHIST_LOG(maphist,"<- done",0,0,0,0); } /* * uvmspace_share: share a vmspace between two processes * * - used for vfork, threads(?) */ void uvmspace_share(struct proc *p1, struct proc *p2) { uvmspace_addref(p1->p_vmspace); p2->p_vmspace = p1->p_vmspace; } #if 0 /* * uvmspace_unshare: ensure that process "p" has its own, unshared, vmspace * * - XXX: no locking on vmspace */ void uvmspace_unshare(struct lwp *l) { struct proc *p = l->l_proc; struct vmspace *nvm, *ovm = p->p_vmspace; if (ovm->vm_refcnt == 1) /* nothing to do: vmspace isn't shared in the first place */ return; /* make a new vmspace, still holding old one */ nvm = uvmspace_fork(ovm); kpreempt_disable(); pmap_deactivate(l); /* unbind old vmspace */ p->p_vmspace = nvm; pmap_activate(l); /* switch to new vmspace */ kpreempt_enable(); uvmspace_free(ovm); /* drop reference to old vmspace */ } #endif /* * uvmspace_spawn: a new process has been spawned and needs a vmspace */ void uvmspace_spawn(struct lwp *l, vaddr_t start, vaddr_t end, bool topdown) { struct proc *p = l->l_proc; struct vmspace *nvm; #ifdef __HAVE_CPU_VMSPACE_EXEC cpu_vmspace_exec(l, start, end); #endif nvm = uvmspace_alloc(start, end, topdown); kpreempt_disable(); p->p_vmspace = nvm; pmap_activate(l); kpreempt_enable(); } /* * uvmspace_exec: the process wants to exec a new program */ void uvmspace_exec(struct lwp *l, vaddr_t start, vaddr_t end, bool topdown) { struct proc *p = l->l_proc; struct vmspace *nvm, *ovm = p->p_vmspace; struct vm_map *map; KASSERT(ovm != NULL); #ifdef __HAVE_CPU_VMSPACE_EXEC cpu_vmspace_exec(l, start, end); #endif map = &ovm->vm_map; /* * see if more than one process is using this vmspace... */ if (ovm->vm_refcnt == 1 && topdown == ((ovm->vm_map.flags & VM_MAP_TOPDOWN) != 0)) { /* * if p is the only process using its vmspace then we can safely * recycle that vmspace for the program that is being exec'd. * But only if TOPDOWN matches the requested value for the new * vm space! */ /* * SYSV SHM semantics require us to kill all segments on an exec */ if (uvm_shmexit && ovm->vm_shm) (*uvm_shmexit)(ovm); /* * POSIX 1003.1b -- "lock future mappings" is revoked * when a process execs another program image. */ map->flags &= ~VM_MAP_WIREFUTURE; /* * now unmap the old program */ pmap_remove_all(map->pmap); uvm_unmap(map, vm_map_min(map), vm_map_max(map)); KASSERT(map->header.prev == &map->header); KASSERT(map->nentries == 0); /* * resize the map */ vm_map_setmin(map, start); vm_map_setmax(map, end); } else { /* * p's vmspace is being shared, so we can't reuse it for p since * it is still being used for others. allocate a new vmspace * for p */ nvm = uvmspace_alloc(start, end, topdown); /* * install new vmspace and drop our ref to the old one. */ kpreempt_disable(); pmap_deactivate(l); p->p_vmspace = nvm; pmap_activate(l); kpreempt_enable(); uvmspace_free(ovm); } } /* * uvmspace_addref: add a referece to a vmspace. */ void uvmspace_addref(struct vmspace *vm) { struct vm_map *map = &vm->vm_map; KASSERT((map->flags & VM_MAP_DYING) == 0); mutex_enter(&map->misc_lock); KASSERT(vm->vm_refcnt > 0); vm->vm_refcnt++; mutex_exit(&map->misc_lock); } /* * uvmspace_free: free a vmspace data structure */ void uvmspace_free(struct vmspace *vm) { struct vm_map_entry *dead_entries; struct vm_map *map = &vm->vm_map; int n; UVMHIST_FUNC("uvmspace_free"); UVMHIST_CALLED(maphist); UVMHIST_LOG(maphist,"(vm=%#jx) ref=%jd", (uintptr_t)vm, vm->vm_refcnt, 0, 0); mutex_enter(&map->misc_lock); n = --vm->vm_refcnt; mutex_exit(&map->misc_lock); if (n > 0) return; /* * at this point, there should be no other references to the map. * delete all of the mappings, then destroy the pmap. */ map->flags |= VM_MAP_DYING; pmap_remove_all(map->pmap); /* Get rid of any SYSV shared memory segments. */ if (uvm_shmexit && vm->vm_shm != NULL) (*uvm_shmexit)(vm); if (map->nentries) { uvm_unmap_remove(map, vm_map_min(map), vm_map_max(map), &dead_entries, 0); if (dead_entries != NULL) uvm_unmap_detach(dead_entries, 0); } KASSERT(map->nentries == 0); KASSERT(map->size == 0); mutex_destroy(&map->misc_lock); rw_destroy(&map->lock); cv_destroy(&map->cv); pmap_destroy(map->pmap); pool_cache_put(&uvm_vmspace_cache, vm); } static struct vm_map_entry * uvm_mapent_clone(struct vm_map *new_map, struct vm_map_entry *old_entry, int flags) { struct vm_map_entry *new_entry; new_entry = uvm_mapent_alloc(new_map, 0); /* old_entry -> new_entry */ uvm_mapent_copy(old_entry, new_entry); /* new pmap has nothing wired in it */ new_entry->wired_count = 0; /* * gain reference to object backing the map (can't * be a submap, already checked this case). */ if (new_entry->aref.ar_amap) uvm_map_reference_amap(new_entry, flags); if (new_entry->object.uvm_obj && new_entry->object.uvm_obj->pgops->pgo_reference) new_entry->object.uvm_obj->pgops->pgo_reference( new_entry->object.uvm_obj); /* insert entry at end of new_map's entry list */ uvm_map_entry_link(new_map, new_map->header.prev, new_entry); return new_entry; } /* * share the mapping: this means we want the old and * new entries to share amaps and backing objects. */ static void uvm_mapent_forkshared(struct vm_map *new_map, struct vm_map *old_map, struct vm_map_entry *old_entry) { /* * if the old_entry needs a new amap (due to prev fork) * then we need to allocate it now so that we have * something we own to share with the new_entry. [in * other words, we need to clear needs_copy] */ if (UVM_ET_ISNEEDSCOPY(old_entry)) { /* get our own amap, clears needs_copy */ amap_copy(old_map, old_entry, AMAP_COPY_NOCHUNK, 0, 0); /* XXXCDC: WAITOK??? */ } uvm_mapent_clone(new_map, old_entry, AMAP_SHARED); } static void uvm_mapent_forkcopy(struct vm_map *new_map, struct vm_map *old_map, struct vm_map_entry *old_entry) { struct vm_map_entry *new_entry; /* * copy-on-write the mapping (using mmap's * MAP_PRIVATE semantics) * * allocate new_entry, adjust reference counts. * (note that new references are read-only). */ new_entry = uvm_mapent_clone(new_map, old_entry, 0); new_entry->etype |= (UVM_ET_COPYONWRITE|UVM_ET_NEEDSCOPY); /* * the new entry will need an amap. it will either * need to be copied from the old entry or created * from scratch (if the old entry does not have an * amap). can we defer this process until later * (by setting "needs_copy") or do we need to copy * the amap now? * * we must copy the amap now if any of the following * conditions hold: * 1. the old entry has an amap and that amap is * being shared. this means that the old (parent) * process is sharing the amap with another * process. if we do not clear needs_copy here * we will end up in a situation where both the * parent and child process are refering to the * same amap with "needs_copy" set. if the * parent write-faults, the fault routine will * clear "needs_copy" in the parent by allocating * a new amap. this is wrong because the * parent is supposed to be sharing the old amap * and the new amap will break that. * * 2. if the old entry has an amap and a non-zero * wire count then we are going to have to call * amap_cow_now to avoid page faults in the * parent process. since amap_cow_now requires * "needs_copy" to be clear we might as well * clear it here as well. * */ if (old_entry->aref.ar_amap != NULL) { if ((amap_flags(old_entry->aref.ar_amap) & AMAP_SHARED) != 0 || VM_MAPENT_ISWIRED(old_entry)) { amap_copy(new_map, new_entry, AMAP_COPY_NOCHUNK, 0, 0); /* XXXCDC: M_WAITOK ... ok? */ } } /* * if the parent's entry is wired down, then the * parent process does not want page faults on * access to that memory. this means that we * cannot do copy-on-write because we can't write * protect the old entry. in this case we * resolve all copy-on-write faults now, using * amap_cow_now. note that we have already * allocated any needed amap (above). */ if (VM_MAPENT_ISWIRED(old_entry)) { /* * resolve all copy-on-write faults now * (note that there is nothing to do if * the old mapping does not have an amap). */ if (old_entry->aref.ar_amap) amap_cow_now(new_map, new_entry); } else { /* * setup mappings to trigger copy-on-write faults * we must write-protect the parent if it has * an amap and it is not already "needs_copy"... * if it is already "needs_copy" then the parent * has already been write-protected by a previous * fork operation. */ if (old_entry->aref.ar_amap && !UVM_ET_ISNEEDSCOPY(old_entry)) { if (old_entry->max_protection & VM_PROT_WRITE) { uvm_map_lock_entry(old_entry); pmap_protect(old_map->pmap, old_entry->start, old_entry->end, old_entry->protection & ~VM_PROT_WRITE); uvm_map_unlock_entry(old_entry); } old_entry->etype |= UVM_ET_NEEDSCOPY; } } } /* * zero the mapping: the new entry will be zero initialized */ static void uvm_mapent_forkzero(struct vm_map *new_map, struct vm_map *old_map, struct vm_map_entry *old_entry) { struct vm_map_entry *new_entry; new_entry = uvm_mapent_clone(new_map, old_entry, 0); new_entry->etype |= (UVM_ET_COPYONWRITE|UVM_ET_NEEDSCOPY); if (new_entry->aref.ar_amap) { uvm_map_unreference_amap(new_entry, 0); new_entry->aref.ar_pageoff = 0; new_entry->aref.ar_amap = NULL; } if (UVM_ET_ISOBJ(new_entry)) { if (new_entry->object.uvm_obj->pgops->pgo_detach) new_entry->object.uvm_obj->pgops->pgo_detach( new_entry->object.uvm_obj); new_entry->object.uvm_obj = NULL; new_entry->etype &= ~UVM_ET_OBJ; } } /* * F O R K - m a i n e n t r y p o i n t */ /* * uvmspace_fork: fork a process' main map * * => create a new vmspace for child process from parent. * => parent's map must not be locked. */ struct vmspace * uvmspace_fork(struct vmspace *vm1) { struct vmspace *vm2; struct vm_map *old_map = &vm1->vm_map; struct vm_map *new_map; struct vm_map_entry *old_entry; UVMHIST_FUNC("uvmspace_fork"); UVMHIST_CALLED(maphist); vm_map_lock(old_map); vm2 = uvmspace_alloc(vm_map_min(old_map), vm_map_max(old_map), vm1->vm_map.flags & VM_MAP_TOPDOWN); memcpy(&vm2->vm_startcopy, &vm1->vm_startcopy, (char *) (vm1 + 1) - (char *) &vm1->vm_startcopy); new_map = &vm2->vm_map; /* XXX */ old_entry = old_map->header.next; new_map->size = old_map->size; /* * go entry-by-entry */ while (old_entry != &old_map->header) { /* * first, some sanity checks on the old entry */ KASSERT(!UVM_ET_ISSUBMAP(old_entry)); KASSERT(UVM_ET_ISCOPYONWRITE(old_entry) || !UVM_ET_ISNEEDSCOPY(old_entry)); switch (old_entry->inheritance) { case MAP_INHERIT_NONE: /* * drop the mapping, modify size */ new_map->size -= old_entry->end - old_entry->start; break; case MAP_INHERIT_SHARE: uvm_mapent_forkshared(new_map, old_map, old_entry); break; case MAP_INHERIT_COPY: uvm_mapent_forkcopy(new_map, old_map, old_entry); break; case MAP_INHERIT_ZERO: uvm_mapent_forkzero(new_map, old_map, old_entry); break; default: KASSERT(0); break; } old_entry = old_entry->next; } pmap_update(old_map->pmap); vm_map_unlock(old_map); if (uvm_shmfork && vm1->vm_shm) (*uvm_shmfork)(vm1, vm2); #ifdef PMAP_FORK pmap_fork(vm1->vm_map.pmap, vm2->vm_map.pmap); #endif UVMHIST_LOG(maphist,"<- done",0,0,0,0); return (vm2); } /* * uvm_mapent_trymerge: try to merge an entry with its neighbors. * * => called with map locked. * => return non zero if successfully merged. */ int uvm_mapent_trymerge(struct vm_map *map, struct vm_map_entry *entry, int flags) { struct uvm_object *uobj; struct vm_map_entry *next; struct vm_map_entry *prev; vsize_t size; int merged = 0; bool copying; int newetype; if (entry->aref.ar_amap != NULL) { return 0; } if ((entry->flags & UVM_MAP_NOMERGE) != 0) { return 0; } uobj = entry->object.uvm_obj; size = entry->end - entry->start; copying = (flags & UVM_MERGE_COPYING) != 0; newetype = copying ? (entry->etype & ~UVM_ET_NEEDSCOPY) : entry->etype; next = entry->next; if (next != &map->header && next->start == entry->end && ((copying && next->aref.ar_amap != NULL && amap_refs(next->aref.ar_amap) == 1) || (!copying && next->aref.ar_amap == NULL)) && UVM_ET_ISCOMPATIBLE(next, newetype, uobj, entry->flags, entry->protection, entry->max_protection, entry->inheritance, entry->advice, entry->wired_count) && (uobj == NULL || entry->offset + size == next->offset)) { int error; if (copying) { error = amap_extend(next, size, AMAP_EXTEND_NOWAIT|AMAP_EXTEND_BACKWARDS); } else { error = 0; } if (error == 0) { if (uobj) { if (uobj->pgops->pgo_detach) { uobj->pgops->pgo_detach(uobj); } } entry->end = next->end; clear_hints(map, next); uvm_map_entry_unlink(map, next); if (copying) { entry->aref = next->aref; entry->etype &= ~UVM_ET_NEEDSCOPY; } uvm_map_check(map, "trymerge forwardmerge"); uvm_mapent_free(next); merged++; } } prev = entry->prev; if (prev != &map->header && prev->end == entry->start && ((copying && !merged && prev->aref.ar_amap != NULL && amap_refs(prev->aref.ar_amap) == 1) || (!copying && prev->aref.ar_amap == NULL)) && UVM_ET_ISCOMPATIBLE(prev, newetype, uobj, entry->flags, entry->protection, entry->max_protection, entry->inheritance, entry->advice, entry->wired_count) && (uobj == NULL || prev->offset + prev->end - prev->start == entry->offset)) { int error; if (copying) { error = amap_extend(prev, size, AMAP_EXTEND_NOWAIT|AMAP_EXTEND_FORWARDS); } else { error = 0; } if (error == 0) { if (uobj) { if (uobj->pgops->pgo_detach) { uobj->pgops->pgo_detach(uobj); } entry->offset = prev->offset; } entry->start = prev->start; clear_hints(map, prev); uvm_map_entry_unlink(map, prev); if (copying) { entry->aref = prev->aref; entry->etype &= ~UVM_ET_NEEDSCOPY; } uvm_map_check(map, "trymerge backmerge"); uvm_mapent_free(prev); merged++; } } return merged; } /* * uvm_map_setup: init map * * => map must not be in service yet. */ void uvm_map_setup(struct vm_map *map, vaddr_t vmin, vaddr_t vmax, int flags) { rb_tree_init(&map->rb_tree, &uvm_map_tree_ops); map->header.next = map->header.prev = &map->header; map->nentries = 0; map->size = 0; map->ref_count = 1; vm_map_setmin(map, vmin); vm_map_setmax(map, vmax); map->flags = flags; map->first_free = &map->header; map->hint = &map->header; map->timestamp = 0; map->busy = NULL; rw_init(&map->lock); cv_init(&map->cv, "vm_map"); mutex_init(&map->misc_lock, MUTEX_DRIVER, IPL_NONE); } /* * U N M A P - m a i n e n t r y p o i n t */ /* * uvm_unmap1: remove mappings from a vm_map (from "start" up to "stop") * * => caller must check alignment and size * => map must be unlocked (we will lock it) * => flags is UVM_FLAG_QUANTUM or 0. */ void uvm_unmap1(struct vm_map *map, vaddr_t start, vaddr_t end, int flags) { struct vm_map_entry *dead_entries; UVMHIST_FUNC("uvm_unmap"); UVMHIST_CALLED(maphist); KASSERT(start < end); UVMHIST_LOG(maphist, " (map=%#jx, start=%#jx, end=%#jx)", (uintptr_t)map, start, end, 0); if (map == kernel_map) { LOCKDEBUG_MEM_CHECK((void *)start, end - start); } /* * work now done by helper functions. wipe the pmap's and then * detach from the dead entries... */ vm_map_lock(map); uvm_unmap_remove(map, start, end, &dead_entries, flags); vm_map_unlock(map); if (dead_entries != NULL) uvm_unmap_detach(dead_entries, 0); UVMHIST_LOG(maphist, "<- done", 0,0,0,0); } /* * uvm_map_reference: add reference to a map * * => map need not be locked (we use misc_lock). */ void uvm_map_reference(struct vm_map *map) { mutex_enter(&map->misc_lock); map->ref_count++; mutex_exit(&map->misc_lock); } bool vm_map_starved_p(struct vm_map *map) { if ((map->flags & VM_MAP_WANTVA) != 0) { return true; } /* XXX */ if ((vm_map_max(map) - vm_map_min(map)) / 16 * 15 < map->size) { return true; } return false; } void uvm_map_lock_entry(struct vm_map_entry *entry) { if (entry->aref.ar_amap != NULL) { amap_lock(entry->aref.ar_amap); } if (UVM_ET_ISOBJ(entry)) { mutex_enter(entry->object.uvm_obj->vmobjlock); } } void uvm_map_unlock_entry(struct vm_map_entry *entry) { if (UVM_ET_ISOBJ(entry)) { mutex_exit(entry->object.uvm_obj->vmobjlock); } if (entry->aref.ar_amap != NULL) { amap_unlock(entry->aref.ar_amap); } } #if defined(DDB) || defined(DEBUGPRINT) /* * uvm_map_printit: actually prints the map */ void uvm_map_printit(struct vm_map *map, bool full, void (*pr)(const char *, ...)) { struct vm_map_entry *entry; (*pr)("MAP %p: [%#lx->%#lx]\n", map, vm_map_min(map), vm_map_max(map)); (*pr)("\t#ent=%d, sz=%d, ref=%d, version=%d, flags=%#x\n", map->nentries, map->size, map->ref_count, map->timestamp, map->flags); (*pr)("\tpmap=%p(resident=%ld, wired=%ld)\n", map->pmap, pmap_resident_count(map->pmap), pmap_wired_count(map->pmap)); if (!full) return; for (entry = map->header.next; entry != &map->header; entry = entry->next) { (*pr)(" - %p: %#lx->%#lx: obj=%p/%#llx, amap=%p/%d\n", entry, entry->start, entry->end, entry->object.uvm_obj, (long long)entry->offset, entry->aref.ar_amap, entry->aref.ar_pageoff); (*pr)( "\tsubmap=%c, cow=%c, nc=%c, prot(max)=%d/%d, inh=%d, " "wc=%d, adv=%d\n", (entry->etype & UVM_ET_SUBMAP) ? 'T' : 'F', (entry->etype & UVM_ET_COPYONWRITE) ? 'T' : 'F', (entry->etype & UVM_ET_NEEDSCOPY) ? 'T' : 'F', entry->protection, entry->max_protection, entry->inheritance, entry->wired_count, entry->advice); } } void uvm_whatis(uintptr_t addr, void (*pr)(const char *, ...)) { struct vm_map *map; for (map = kernel_map;;) { struct vm_map_entry *entry; if (!uvm_map_lookup_entry_bytree(map, (vaddr_t)addr, &entry)) { break; } (*pr)("%p is %p+%zu from VMMAP %p\n", (void *)addr, (void *)entry->start, (size_t)(addr - (uintptr_t)entry->start), map); if (!UVM_ET_ISSUBMAP(entry)) { break; } map = entry->object.sub_map; } } #endif /* DDB || DEBUGPRINT */ #ifndef __USER_VA0_IS_SAFE static int sysctl_user_va0_disable(SYSCTLFN_ARGS) { struct sysctlnode node; int t, error; node = *rnode; node.sysctl_data = &t; t = user_va0_disable; error = sysctl_lookup(SYSCTLFN_CALL(&node)); if (error || newp == NULL) return (error); if (!t && user_va0_disable && kauth_authorize_system(l->l_cred, KAUTH_SYSTEM_MAP_VA_ZERO, 0, NULL, NULL, NULL)) return EPERM; user_va0_disable = !!t; return 0; } #endif static int fill_vmentry(struct lwp *l, struct proc *p, struct kinfo_vmentry *kve, struct vm_map *m, struct vm_map_entry *e) { #ifndef _RUMPKERNEL int error; memset(kve, 0, sizeof(*kve)); KASSERT(e != NULL); if (UVM_ET_ISOBJ(e)) { struct uvm_object *uobj = e->object.uvm_obj; KASSERT(uobj != NULL); kve->kve_ref_count = uobj->uo_refs; kve->kve_count = uobj->uo_npages; if (UVM_OBJ_IS_VNODE(uobj)) { struct vattr va; struct vnode *vp = (struct vnode *)uobj; vn_lock(vp, LK_SHARED | LK_RETRY); error = VOP_GETATTR(vp, &va, l->l_cred); VOP_UNLOCK(vp); kve->kve_type = KVME_TYPE_VNODE; if (error == 0) { kve->kve_vn_size = vp->v_size; kve->kve_vn_type = (int)vp->v_type; kve->kve_vn_mode = va.va_mode; kve->kve_vn_rdev = va.va_rdev; kve->kve_vn_fileid = va.va_fileid; kve->kve_vn_fsid = va.va_fsid; error = vnode_to_path(kve->kve_path, sizeof(kve->kve_path) / 2, vp, l, p); #ifdef DIAGNOSTIC if (error) printf("%s: vp %p error %d\n", __func__, vp, error); #endif } } else if (UVM_OBJ_IS_KERN_OBJECT(uobj)) { kve->kve_type = KVME_TYPE_KERN; } else if (UVM_OBJ_IS_DEVICE(uobj)) { kve->kve_type = KVME_TYPE_DEVICE; } else if (UVM_OBJ_IS_AOBJ(uobj)) { kve->kve_type = KVME_TYPE_ANON; } else { kve->kve_type = KVME_TYPE_OBJECT; } } else if (UVM_ET_ISSUBMAP(e)) { struct vm_map *map = e->object.sub_map; KASSERT(map != NULL); kve->kve_ref_count = map->ref_count; kve->kve_count = map->nentries; kve->kve_type = KVME_TYPE_SUBMAP; } else kve->kve_type = KVME_TYPE_UNKNOWN; kve->kve_start = e->start; kve->kve_end = e->end; kve->kve_offset = e->offset; kve->kve_wired_count = e->wired_count; kve->kve_inheritance = e->inheritance; kve->kve_attributes = 0; /* e->map_attrib */ kve->kve_advice = e->advice; #define PROT(p) (((p) & VM_PROT_READ) ? KVME_PROT_READ : 0) | \ (((p) & VM_PROT_WRITE) ? KVME_PROT_WRITE : 0) | \ (((p) & VM_PROT_EXECUTE) ? KVME_PROT_EXEC : 0) kve->kve_protection = PROT(e->protection); kve->kve_max_protection = PROT(e->max_protection); kve->kve_flags |= (e->etype & UVM_ET_COPYONWRITE) ? KVME_FLAG_COW : 0; kve->kve_flags |= (e->etype & UVM_ET_NEEDSCOPY) ? KVME_FLAG_NEEDS_COPY : 0; kve->kve_flags |= (m->flags & VM_MAP_TOPDOWN) ? KVME_FLAG_GROWS_DOWN : KVME_FLAG_GROWS_UP; kve->kve_flags |= (m->flags & VM_MAP_PAGEABLE) ? KVME_FLAG_PAGEABLE : 0; #endif return 0; } static int fill_vmentries(struct lwp *l, pid_t pid, u_int elem_size, void *oldp, size_t *oldlenp) { int error; struct proc *p; struct kinfo_vmentry *vme; struct vmspace *vm; struct vm_map *map; struct vm_map_entry *entry; char *dp; size_t count, vmesize; if (elem_size == 0 || elem_size > 2 * sizeof(*vme)) return EINVAL; if (oldp) { if (*oldlenp > 10UL * 1024UL * 1024UL) return E2BIG; count = *oldlenp / elem_size; if (count == 0) return ENOMEM; vmesize = count * sizeof(*vme); } else vmesize = 0; if ((error = proc_find_locked(l, &p, pid)) != 0) return error; vme = NULL; count = 0; if ((error = proc_vmspace_getref(p, &vm)) != 0) goto out; map = &vm->vm_map; vm_map_lock_read(map); dp = oldp; if (oldp) vme = kmem_alloc(vmesize, KM_SLEEP); for (entry = map->header.next; entry != &map->header; entry = entry->next) { if (oldp && (dp - (char *)oldp) < vmesize) { error = fill_vmentry(l, p, &vme[count], map, entry); if (error) goto out; dp += elem_size; } count++; } vm_map_unlock_read(map); uvmspace_free(vm); out: if (pid != -1) mutex_exit(p->p_lock); if (error == 0) { const u_int esize = uimin(sizeof(*vme), elem_size); dp = oldp; for (size_t i = 0; i < count; i++) { if (oldp && (dp - (char *)oldp) < vmesize) { error = sysctl_copyout(l, &vme[i], dp, esize); if (error) break; dp += elem_size; } else break; } count *= elem_size; if (oldp != NULL && *oldlenp < count) error = ENOSPC; *oldlenp = count; } if (vme) kmem_free(vme, vmesize); return error; } static int sysctl_vmproc(SYSCTLFN_ARGS) { int error; if (namelen == 1 && name[0] == CTL_QUERY) return (sysctl_query(SYSCTLFN_CALL(rnode))); if (namelen == 0) return EINVAL; switch (name[0]) { case VM_PROC_MAP: if (namelen != 3) return EINVAL; sysctl_unlock(); error = fill_vmentries(l, name[1], name[2], oldp, oldlenp); sysctl_relock(); return error; default: return EINVAL; } } SYSCTL_SETUP(sysctl_uvmmap_setup, "sysctl uvmmap setup") { sysctl_createv(clog, 0, NULL, NULL, CTLFLAG_PERMANENT, CTLTYPE_STRUCT, "proc", SYSCTL_DESCR("Process vm information"), sysctl_vmproc, 0, NULL, 0, CTL_VM, VM_PROC, CTL_EOL); #ifndef __USER_VA0_IS_SAFE sysctl_createv(clog, 0, NULL, NULL, CTLFLAG_PERMANENT|CTLFLAG_READWRITE, CTLTYPE_INT, "user_va0_disable", SYSCTL_DESCR("Disable VA 0"), sysctl_user_va0_disable, 0, &user_va0_disable, 0, CTL_VM, CTL_CREATE, CTL_EOL); #endif }