korg/linux
0
0
Fork 0
mirror of https://mirrors.bfsu.edu.cn/git/linux.git synced 2024-09-21 20:22:13 +08:00
Code Issues Packages Projects Releases Wiki Activity

mm: remove the vma linked list

Browse Source 
Replace any vm_next use with vma_find().

Update free_pgtables(), unmap_vmas(), and zap_page_range() to use the
maple tree.

Use the new free_pgtables() and unmap_vmas() in do_mas_align_munmap().  At
the same time, alter the loop to be more compact.

Now that free_pgtables() and unmap_vmas() take a maple tree as an
argument, rearrange do_mas_align_munmap() to use the new tree to hold the
vmas to remove.

Remove __vma_link_list() and __vma_unlink_list() as they are exclusively
used to update the linked list.

Drop linked list update from __insert_vm_struct().

Rework validation of tree as it was depending on the linked list.

[yang.lee@linux.alibaba.com: fix one kernel-doc comment]
  Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=1949
  Link: https://lkml.kernel.org/r/20220824021918.94116-1-yang.lee@linux.alibaba.comLink: https://lkml.kernel.org/r/20220906194824.2110408-69-Liam.Howlett@oracle.com
Signed-off-by: Liam R. Howlett <Liam.Howlett@Oracle.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Tested-by: Yu Zhao <yuzhao@google.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: SeongJae Park <sj@kernel.org>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
This commit is contained in:
Liam R. Howlett 2022-09-06 19:49:06 +00:00 committed by Andrew Morton
parent 78ba531ff3
commit 763ecb0350
9 changed files with 225 additions and 374 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
include/linux/mm.h
View File

@ -1857,8 +1857,9 @@ void zap_vma_ptes(struct vm_area_struct *vma, unsigned long address,
unsigned long size); unsigned long size);
void zap_page_range(struct vm_area_struct *vma, unsigned long address, void zap_page_range(struct vm_area_struct *vma, unsigned long address,
unsigned long size); unsigned long size);
void unmap_vmas(struct mmu_gather *tlb, struct vm_area_struct *start_vma, void unmap_vmas(struct mmu_gather *tlb, struct maple_tree *mt,
unsigned long start, unsigned long end); struct vm_area_struct *start_vma, unsigned long start,
unsigned long end);
struct mmu_notifier_range; struct mmu_notifier_range;

4
include/linux/mm_types.h
View File

@ -408,8 +408,6 @@ struct vm_area_struct {
unsigned long vm_end; /* The first byte after our end address unsigned long vm_end; /* The first byte after our end address
within vm_mm. */ within vm_mm. */
/* linked list of VM areas per task, sorted by address */
struct vm_area_struct *vm_next, *vm_prev;
struct mm_struct *vm_mm; /* The address space we belong to. */ struct mm_struct *vm_mm; /* The address space we belong to. */
/* /*
@ -473,7 +471,6 @@ struct vm_area_struct {
struct kioctx_table; struct kioctx_table;
struct mm_struct { struct mm_struct {
struct { struct {
struct vm_area_struct *mmap; /* list of VMAs */
struct maple_tree mm_mt; struct maple_tree mm_mt;
#ifdef CONFIG_MMU #ifdef CONFIG_MMU
unsigned long (*get_unmapped_area) (struct file *filp, unsigned long (*get_unmapped_area) (struct file *filp,
@ -488,7 +485,6 @@ struct mm_struct {
unsigned long mmap_compat_legacy_base; unsigned long mmap_compat_legacy_base;
#endif #endif
unsigned long task_size; /* size of task vm space */ unsigned long task_size; /* size of task vm space */
unsigned long highest_vm_end; /* highest vma end address */
pgd_t * pgd; pgd_t * pgd;
#ifdef CONFIG_MEMBARRIER #ifdef CONFIG_MEMBARRIER

19
kernel/fork.c
View File

@ -474,7 +474,6 @@ struct vm_area_struct *vm_area_dup(struct vm_area_struct *orig)
*/ */
*new = data_race(*orig); *new = data_race(*orig);
INIT_LIST_HEAD(&new->anon_vma_chain); INIT_LIST_HEAD(&new->anon_vma_chain);
new->vm_next = new->vm_prev = NULL;
dup_anon_vma_name(orig, new); dup_anon_vma_name(orig, new);
} }
return new; return new;
@ -579,7 +578,7 @@ static void dup_mm_exe_file(struct mm_struct *mm, struct mm_struct *oldmm)
static __latent_entropy int dup_mmap(struct mm_struct *mm, static __latent_entropy int dup_mmap(struct mm_struct *mm,
struct mm_struct *oldmm) struct mm_struct *oldmm)
{ {
struct vm_area_struct *mpnt, *tmp, *prev, **pprev; struct vm_area_struct *mpnt, *tmp;
int retval; int retval;
unsigned long charge = 0; unsigned long charge = 0;
LIST_HEAD(uf); LIST_HEAD(uf);
@ -606,18 +605,11 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
mm->exec_vm = oldmm->exec_vm; mm->exec_vm = oldmm->exec_vm;
mm->stack_vm = oldmm->stack_vm; mm->stack_vm = oldmm->stack_vm;
pprev = &mm->mmap;
retval = ksm_fork(mm, oldmm); retval = ksm_fork(mm, oldmm);
if (retval) if (retval)
goto out; goto out;
khugepaged_fork(mm, oldmm); khugepaged_fork(mm, oldmm);
retval = mas_expected_entries(&mas, oldmm->map_count);
if (retval)
goto out;
prev = NULL;
retval = mas_expected_entries(&mas, oldmm->map_count); retval = mas_expected_entries(&mas, oldmm->map_count);
if (retval) if (retval)
goto out; goto out;
@ -689,14 +681,6 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
if (is_vm_hugetlb_page(tmp)) if (is_vm_hugetlb_page(tmp))
reset_vma_resv_huge_pages(tmp); reset_vma_resv_huge_pages(tmp);
/*
* Link in the new vma and copy the page table entries.
*/
*pprev = tmp;
pprev = &tmp->vm_next;
tmp->vm_prev = prev;
prev = tmp;
/* Link the vma into the MT */ /* Link the vma into the MT */
mas.index = tmp->vm_start; mas.index = tmp->vm_start;
mas.last = tmp->vm_end - 1; mas.last = tmp->vm_end - 1;
@ -1124,7 +1108,6 @@ static void mm_init_uprobes_state(struct mm_struct *mm)
static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p,
struct user_namespace *user_ns) struct user_namespace *user_ns)
{ {
mm->mmap = NULL;
mt_init_flags(&mm->mm_mt, MM_MT_FLAGS); mt_init_flags(&mm->mm_mt, MM_MT_FLAGS);
mt_set_external_lock(&mm->mm_mt, &mm->mmap_lock); mt_set_external_lock(&mm->mm_mt, &mm->mmap_lock);
atomic_set(&mm->mm_users, 1); atomic_set(&mm->mm_users, 1);

14
mm/debug.c
View File

@ -139,13 +139,11 @@ EXPORT_SYMBOL(dump_page);
void dump_vma(const struct vm_area_struct *vma) void dump_vma(const struct vm_area_struct *vma)
{ {
pr_emerg("vma %px start %px end %px\n" pr_emerg("vma %px start %px end %px mm %px\n"
"next %px prev %px mm %px\n"
"prot %lx anon_vma %px vm_ops %px\n" "prot %lx anon_vma %px vm_ops %px\n"
"pgoff %lx file %px private_data %px\n" "pgoff %lx file %px private_data %px\n"
"flags: %#lx(%pGv)\n", "flags: %#lx(%pGv)\n",
vma, (void *)vma->vm_start, (void *)vma->vm_end, vma->vm_next, vma, (void *)vma->vm_start, (void *)vma->vm_end, vma->vm_mm,
vma->vm_prev, vma->vm_mm,
(unsigned long)pgprot_val(vma->vm_page_prot), (unsigned long)pgprot_val(vma->vm_page_prot),
vma->anon_vma, vma->vm_ops, vma->vm_pgoff, vma->anon_vma, vma->vm_ops, vma->vm_pgoff,
vma->vm_file, vma->vm_private_data, vma->vm_file, vma->vm_private_data,
@ -155,11 +153,11 @@ EXPORT_SYMBOL(dump_vma);
void dump_mm(const struct mm_struct *mm) void dump_mm(const struct mm_struct *mm)
{ {
pr_emerg("mm %px mmap %px task_size %lu\n" pr_emerg("mm %px task_size %lu\n"
#ifdef CONFIG_MMU #ifdef CONFIG_MMU
"get_unmapped_area %px\n" "get_unmapped_area %px\n"
#endif #endif
"mmap_base %lu mmap_legacy_base %lu highest_vm_end %lu\n" "mmap_base %lu mmap_legacy_base %lu\n"
"pgd %px mm_users %d mm_count %d pgtables_bytes %lu map_count %d\n" "pgd %px mm_users %d mm_count %d pgtables_bytes %lu map_count %d\n"
"hiwater_rss %lx hiwater_vm %lx total_vm %lx locked_vm %lx\n" "hiwater_rss %lx hiwater_vm %lx total_vm %lx locked_vm %lx\n"
"pinned_vm %llx data_vm %lx exec_vm %lx stack_vm %lx\n" "pinned_vm %llx data_vm %lx exec_vm %lx stack_vm %lx\n"
@ -183,11 +181,11 @@ void dump_mm(const struct mm_struct *mm)
"tlb_flush_pending %d\n" "tlb_flush_pending %d\n"
"def_flags: %#lx(%pGv)\n", "def_flags: %#lx(%pGv)\n",
mm, mm->mmap, mm->task_size, mm, mm->task_size,
#ifdef CONFIG_MMU #ifdef CONFIG_MMU
mm->get_unmapped_area, mm->get_unmapped_area,
#endif #endif
mm->mmap_base, mm->mmap_legacy_base, mm->highest_vm_end, mm->mmap_base, mm->mmap_legacy_base,
mm->pgd, atomic_read(&mm->mm_users), mm->pgd, atomic_read(&mm->mm_users),
atomic_read(&mm->mm_count), atomic_read(&mm->mm_count),
mm_pgtables_bytes(mm), mm_pgtables_bytes(mm),

8
mm/internal.h
View File

@ -85,8 +85,9 @@ bool __folio_end_writeback(struct folio *folio);
void deactivate_file_folio(struct folio *folio); void deactivate_file_folio(struct folio *folio);
void folio_activate(struct folio *folio); void folio_activate(struct folio *folio);
void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *start_vma, void free_pgtables(struct mmu_gather *tlb, struct maple_tree *mt,
unsigned long floor, unsigned long ceiling); struct vm_area_struct *start_vma, unsigned long floor,
unsigned long ceiling);
void pmd_install(struct mm_struct *mm, pmd_t *pmd, pgtable_t *pte); void pmd_install(struct mm_struct *mm, pmd_t *pmd, pgtable_t *pte);
struct zap_details; struct zap_details;
@ -480,9 +481,6 @@ static inline bool is_data_mapping(vm_flags_t flags)
} }
/* mm/util.c */ /* mm/util.c */
void __vma_link_list(struct mm_struct *mm, struct vm_area_struct *vma,
struct vm_area_struct *prev);
void __vma_unlink_list(struct mm_struct *mm, struct vm_area_struct *vma);
struct anon_vma *folio_anon_vma(struct folio *folio); struct anon_vma *folio_anon_vma(struct folio *folio);
#ifdef CONFIG_MMU #ifdef CONFIG_MMU

34
mm/memory.c
View File

@ -392,12 +392,21 @@ void free_pgd_range(struct mmu_gather *tlb,
} while (pgd++, addr = next, addr != end); } while (pgd++, addr = next, addr != end);
} }
void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *vma, void free_pgtables(struct mmu_gather *tlb, struct maple_tree *mt,
unsigned long floor, unsigned long ceiling) struct vm_area_struct *vma, unsigned long floor,
unsigned long ceiling)
{ {
while (vma) { MA_STATE(mas, mt, vma->vm_end, vma->vm_end);
struct vm_area_struct *next = vma->vm_next;
do {
unsigned long addr = vma->vm_start; unsigned long addr = vma->vm_start;
struct vm_area_struct *next;
/*
* Note: USER_PGTABLES_CEILING may be passed as ceiling and may
* be 0. This will underflow and is okay.
*/
next = mas_find(&mas, ceiling - 1);
/* /*
* Hide vma from rmap and truncate_pagecache before freeing * Hide vma from rmap and truncate_pagecache before freeing
@ -416,7 +425,7 @@ void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *vma,
while (next && next->vm_start <= vma->vm_end + PMD_SIZE while (next && next->vm_start <= vma->vm_end + PMD_SIZE
&& !is_vm_hugetlb_page(next)) { && !is_vm_hugetlb_page(next)) {
vma = next; vma = next;
next = vma->vm_next; next = mas_find(&mas, ceiling - 1);
unlink_anon_vmas(vma); unlink_anon_vmas(vma);
unlink_file_vma(vma); unlink_file_vma(vma);
} }
@ -424,7 +433,7 @@ void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *vma,
floor, next ? next->vm_start : ceiling); floor, next ? next->vm_start : ceiling);
} }
vma = next; vma = next;
} } while (vma);
} }
void pmd_install(struct mm_struct *mm, pmd_t *pmd, pgtable_t *pte) void pmd_install(struct mm_struct *mm, pmd_t *pmd, pgtable_t *pte)
@ -1688,6 +1697,7 @@ static void unmap_single_vma(struct mmu_gather *tlb,
/** /**
* unmap_vmas - unmap a range of memory covered by a list of vma's * unmap_vmas - unmap a range of memory covered by a list of vma's
* @tlb: address of the caller's struct mmu_gather * @tlb: address of the caller's struct mmu_gather
* @mt: the maple tree
* @vma: the starting vma * @vma: the starting vma
* @start_addr: virtual address at which to start unmapping * @start_addr: virtual address at which to start unmapping
* @end_addr: virtual address at which to end unmapping * @end_addr: virtual address at which to end unmapping
@ -1703,7 +1713,7 @@ static void unmap_single_vma(struct mmu_gather *tlb,
* ensure that any thus-far unmapped pages are flushed before unmap_vmas() * ensure that any thus-far unmapped pages are flushed before unmap_vmas()
* drops the lock and schedules. * drops the lock and schedules.
*/ */
void unmap_vmas(struct mmu_gather *tlb, void unmap_vmas(struct mmu_gather *tlb, struct maple_tree *mt,
struct vm_area_struct *vma, unsigned long start_addr, struct vm_area_struct *vma, unsigned long start_addr,
unsigned long end_addr) unsigned long end_addr)
{ {
@ -1713,12 +1723,14 @@ void unmap_vmas(struct mmu_gather *tlb,
/* Careful - we need to zap private pages too! */ /* Careful - we need to zap private pages too! */
.even_cows = true, .even_cows = true,
}; };
MA_STATE(mas, mt, vma->vm_end, vma->vm_end);
mmu_notifier_range_init(&range, MMU_NOTIFY_UNMAP, 0, vma, vma->vm_mm, mmu_notifier_range_init(&range, MMU_NOTIFY_UNMAP, 0, vma, vma->vm_mm,
start_addr, end_addr); start_addr, end_addr);
mmu_notifier_invalidate_range_start(&range); mmu_notifier_invalidate_range_start(&range);
for ( ; vma && vma->vm_start < end_addr; vma = vma->vm_next) do {
unmap_single_vma(tlb, vma, start_addr, end_addr, &details); unmap_single_vma(tlb, vma, start_addr, end_addr, &details);
} while ((vma = mas_find(&mas, end_addr - 1)) != NULL);
mmu_notifier_invalidate_range_end(&range); mmu_notifier_invalidate_range_end(&range);
} }
@ -1733,8 +1745,11 @@ void unmap_vmas(struct mmu_gather *tlb,
void zap_page_range(struct vm_area_struct *vma, unsigned long start, void zap_page_range(struct vm_area_struct *vma, unsigned long start,
unsigned long size) unsigned long size)
{ {
struct maple_tree *mt = &vma->vm_mm->mm_mt;
unsigned long end = start + size;
struct mmu_notifier_range range; struct mmu_notifier_range range;
struct mmu_gather tlb; struct mmu_gather tlb;
MA_STATE(mas, mt, vma->vm_end, vma->vm_end);
lru_add_drain(); lru_add_drain();
mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm, mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, vma->vm_mm,
@ -1742,8 +1757,9 @@ void zap_page_range(struct vm_area_struct *vma, unsigned long start,
tlb_gather_mmu(&tlb, vma->vm_mm); tlb_gather_mmu(&tlb, vma->vm_mm);
update_hiwater_rss(vma->vm_mm); update_hiwater_rss(vma->vm_mm);
mmu_notifier_invalidate_range_start(&range); mmu_notifier_invalidate_range_start(&range);
for ( ; vma && vma->vm_start < range.end; vma = vma->vm_next) do {
unmap_single_vma(&tlb, vma, start, range.end, NULL); unmap_single_vma(&tlb, vma, start, range.end, NULL);
} while ((vma = mas_find(&mas, end - 1)) != NULL);
mmu_notifier_invalidate_range_end(&range); mmu_notifier_invalidate_range_end(&range);
tlb_finish_mmu(&tlb); tlb_finish_mmu(&tlb);
} }

469
mm/mmap.c
View File

@ -75,9 +75,10 @@ int mmap_rnd_compat_bits __read_mostly = CONFIG_ARCH_MMAP_RND_COMPAT_BITS;
static bool ignore_rlimit_data; static bool ignore_rlimit_data;
core_param(ignore_rlimit_data, ignore_rlimit_data, bool, 0644); core_param(ignore_rlimit_data, ignore_rlimit_data, bool, 0644);
static void unmap_region(struct mm_struct *mm, static void unmap_region(struct mm_struct *mm, struct maple_tree *mt,
struct vm_area_struct *vma, struct vm_area_struct *prev, struct vm_area_struct *vma, struct vm_area_struct *prev,
unsigned long start, unsigned long end); struct vm_area_struct *next, unsigned long start,
unsigned long end);
static pgprot_t vm_pgprot_modify(pgprot_t oldprot, unsigned long vm_flags) static pgprot_t vm_pgprot_modify(pgprot_t oldprot, unsigned long vm_flags)
{ {
@ -130,12 +131,10 @@ void unlink_file_vma(struct vm_area_struct *vma)
} }
/* /*
* Close a vm structure and free it, returning the next. * Close a vm structure and free it.
*/ */
static struct vm_area_struct *remove_vma(struct vm_area_struct *vma) static void remove_vma(struct vm_area_struct *vma)
{ {
struct vm_area_struct *next = vma->vm_next;
might_sleep(); might_sleep();
if (vma->vm_ops && vma->vm_ops->close) if (vma->vm_ops && vma->vm_ops->close)
vma->vm_ops->close(vma); vma->vm_ops->close(vma);
@ -143,7 +142,6 @@ static struct vm_area_struct *remove_vma(struct vm_area_struct *vma)
fput(vma->vm_file); fput(vma->vm_file);
mpol_put(vma_policy(vma)); mpol_put(vma_policy(vma));
vm_area_free(vma); vm_area_free(vma);
return next;
} }
/* /*
@ -168,8 +166,7 @@ static int do_brk_munmap(struct ma_state *mas, struct vm_area_struct *vma,
unsigned long newbrk, unsigned long oldbrk, unsigned long newbrk, unsigned long oldbrk,
struct list_head *uf); struct list_head *uf);
static int do_brk_flags(struct ma_state *mas, struct vm_area_struct *brkvma, static int do_brk_flags(struct ma_state *mas, struct vm_area_struct *brkvma,
unsigned long addr, unsigned long request, unsigned long addr, unsigned long request, unsigned long flags);
unsigned long flags);
SYSCALL_DEFINE1(brk, unsigned long, brk) SYSCALL_DEFINE1(brk, unsigned long, brk)
{ {
unsigned long newbrk, oldbrk, origbrk; unsigned long newbrk, oldbrk, origbrk;
@ -238,7 +235,6 @@ SYSCALL_DEFINE1(brk, unsigned long, brk)
* before calling do_brk_munmap(). * before calling do_brk_munmap().
*/ */
mm->brk = brk; mm->brk = brk;
mas.last = oldbrk - 1;
ret = do_brk_munmap(&mas, brkvma, newbrk, oldbrk, &uf); ret = do_brk_munmap(&mas, brkvma, newbrk, oldbrk, &uf);
if (ret == 1) { if (ret == 1) {
downgraded = true; downgraded = true;
@ -293,44 +289,21 @@ extern void mt_dump(const struct maple_tree *mt);
static void validate_mm_mt(struct mm_struct *mm) static void validate_mm_mt(struct mm_struct *mm)
{ {
struct maple_tree *mt = &mm->mm_mt; struct maple_tree *mt = &mm->mm_mt;
struct vm_area_struct *vma_mt, *vma = mm->mmap; struct vm_area_struct *vma_mt;
MA_STATE(mas, mt, 0, 0); MA_STATE(mas, mt, 0, 0);
mt_validate(&mm->mm_mt); mt_validate(&mm->mm_mt);
mas_for_each(&mas, vma_mt, ULONG_MAX) { mas_for_each(&mas, vma_mt, ULONG_MAX) {
if (xa_is_zero(vma_mt)) if ((vma_mt->vm_start != mas.index) ||
continue; (vma_mt->vm_end - 1 != mas.last)) {
if (!vma)
break;
if ((vma != vma_mt) ||
(vma->vm_start != vma_mt->vm_start) ||
(vma->vm_end != vma_mt->vm_end) ||
(vma->vm_start != mas.index) ||
(vma->vm_end - 1 != mas.last)) {
pr_emerg("issue in %s\n", current->comm); pr_emerg("issue in %s\n", current->comm);
dump_stack(); dump_stack();
dump_vma(vma_mt); dump_vma(vma_mt);
pr_emerg("and vm_next\n");
dump_vma(vma->vm_next);
pr_emerg("mt piv: %p %lu - %lu\n", vma_mt, pr_emerg("mt piv: %p %lu - %lu\n", vma_mt,
mas.index, mas.last); mas.index, mas.last);
pr_emerg("mt vma: %p %lu - %lu\n", vma_mt, pr_emerg("mt vma: %p %lu - %lu\n", vma_mt,
vma_mt->vm_start, vma_mt->vm_end); vma_mt->vm_start, vma_mt->vm_end);
if (vma->vm_prev) {
pr_emerg("ll prev: %p %lu - %lu\n",
vma->vm_prev, vma->vm_prev->vm_start,
vma->vm_prev->vm_end);
}
pr_emerg("ll vma: %p %lu - %lu\n", vma,
vma->vm_start, vma->vm_end);
if (vma->vm_next) {
pr_emerg("ll next: %p %lu - %lu\n",
vma->vm_next, vma->vm_next->vm_start,
vma->vm_next->vm_end);
}
mt_dump(mas.tree); mt_dump(mas.tree);
if (vma_mt->vm_end != mas.last + 1) { if (vma_mt->vm_end != mas.last + 1) {
@ -347,23 +320,19 @@ static void validate_mm_mt(struct mm_struct *mm)
} }
VM_BUG_ON_MM(vma_mt->vm_start != mas.index, mm); VM_BUG_ON_MM(vma_mt->vm_start != mas.index, mm);
} }
VM_BUG_ON(vma != vma_mt);
vma = vma->vm_next;
} }
VM_BUG_ON(vma);
} }
static void validate_mm(struct mm_struct *mm) static void validate_mm(struct mm_struct *mm)
{ {
int bug = 0; int bug = 0;
int i = 0; int i = 0;
unsigned long highest_address = 0; struct vm_area_struct *vma;
struct vm_area_struct *vma = mm->mmap; MA_STATE(mas, &mm->mm_mt, 0, 0);
validate_mm_mt(mm); validate_mm_mt(mm);
while (vma) { mas_for_each(&mas, vma, ULONG_MAX) {
#ifdef CONFIG_DEBUG_VM_RB #ifdef CONFIG_DEBUG_VM_RB
struct anon_vma *anon_vma = vma->anon_vma; struct anon_vma *anon_vma = vma->anon_vma;
struct anon_vma_chain *avc; struct anon_vma_chain *avc;
@ -375,18 +344,10 @@ static void validate_mm(struct mm_struct *mm)
anon_vma_unlock_read(anon_vma); anon_vma_unlock_read(anon_vma);
} }
#endif #endif
highest_address = vm_end_gap(vma);
vma = vma->vm_next;
i++; i++;
} }
if (i != mm->map_count) { if (i != mm->map_count) {
pr_emerg("map_count %d vm_next %d\n", mm->map_count, i); pr_emerg("map_count %d mas_for_each %d\n", mm->map_count, i);
bug = 1;
}
if (highest_address != mm->highest_vm_end) {
pr_emerg("mm->highest_vm_end %lx, found %lx\n",
mm->highest_vm_end, highest_address);
bug = 1; bug = 1;
} }
VM_BUG_ON_MM(bug, mm); VM_BUG_ON_MM(bug, mm);
@ -446,29 +407,13 @@ bool range_has_overlap(struct mm_struct *mm, unsigned long start,
struct vm_area_struct *existing; struct vm_area_struct *existing;
MA_STATE(mas, &mm->mm_mt, start, start); MA_STATE(mas, &mm->mm_mt, start, start);
rcu_read_lock();
existing = mas_find(&mas, end - 1); existing = mas_find(&mas, end - 1);
*pprev = mas_prev(&mas, 0); *pprev = mas_prev(&mas, 0);
rcu_read_unlock();
return existing ? true : false; return existing ? true : false;
} }
/*
* __vma_next() - Get the next VMA.
* @mm: The mm_struct.
* @vma: The current vma.
*
* If @vma is NULL, return the first vma in the mm.
*
* Returns: The next VMA after @vma.
*/
static inline struct vm_area_struct *__vma_next(struct mm_struct *mm,
struct vm_area_struct *vma)
{
if (!vma)
return mm->mmap;
return vma->vm_next;
}
static unsigned long count_vma_pages_range(struct mm_struct *mm, static unsigned long count_vma_pages_range(struct mm_struct *mm,
unsigned long addr, unsigned long end) unsigned long addr, unsigned long end)
{ {
@ -553,8 +498,7 @@ static inline void vma_mas_szero(struct ma_state *mas, unsigned long start,
mas_store_prealloc(mas, NULL); mas_store_prealloc(mas, NULL);
} }
static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma, static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma)
struct vm_area_struct *prev)
{ {
MA_STATE(mas, &mm->mm_mt, 0, 0); MA_STATE(mas, &mm->mm_mt, 0, 0);
struct address_space *mapping = NULL; struct address_space *mapping = NULL;
@ -568,7 +512,6 @@ static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma,
} }
vma_mas_store(vma, &mas); vma_mas_store(vma, &mas);
__vma_link_list(mm, vma, prev);
__vma_link_file(vma); __vma_link_file(vma);
if (mapping) if (mapping)
@ -579,22 +522,6 @@ static int vma_link(struct mm_struct *mm, struct vm_area_struct *vma,
return 0; return 0;
} }
/*
* Helper for vma_adjust() in the split_vma insert case: insert a vma into the
* mm's list and the mm tree. It has already been inserted into the interval tree.
*/
static void __insert_vm_struct(struct mm_struct *mm, struct ma_state *mas,
struct vm_area_struct *vma, unsigned long location)
{
struct vm_area_struct *prev;
mas_set(mas, location);
prev = mas_prev(mas, 0);
vma_mas_store(vma, mas);
__vma_link_list(mm, vma, prev);
mm->map_count++;
}
/* /*
* vma_expand - Expand an existing VMA * vma_expand - Expand an existing VMA
* *
@ -675,15 +602,8 @@ inline int vma_expand(struct ma_state *mas, struct vm_area_struct *vma,
} }
/* Expanding over the next vma */ /* Expanding over the next vma */
if (remove_next) { if (remove_next && file) {
/* Remove from mm linked list - also updates highest_vm_end */ __remove_shared_vm_struct(next, file, mapping);
__vma_unlink_list(mm, next);
if (file)
__remove_shared_vm_struct(next, file, mapping);
} else if (!next) {
mm->highest_vm_end = vm_end_gap(vma);
} }
if (anon_vma) { if (anon_vma) {
@ -738,7 +658,6 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
int remove_next = 0; int remove_next = 0;
MA_STATE(mas, &mm->mm_mt, 0, 0); MA_STATE(mas, &mm->mm_mt, 0, 0);
struct vm_area_struct *exporter = NULL, *importer = NULL; struct vm_area_struct *exporter = NULL, *importer = NULL;
unsigned long ll_prev = vma->vm_start; /* linked list prev. */
if (next && !insert) { if (next && !insert) {
if (end >= next->vm_end) { if (end >= next->vm_end) {
@ -773,7 +692,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
next_next = find_vma(mm, next->vm_end); next_next = find_vma(mm, next->vm_end);
VM_WARN_ON(remove_next == 2 && VM_WARN_ON(remove_next == 2 &&
end != next->vm_next->vm_end); end != next_next->vm_end);
} }
exporter = next; exporter = next;
@ -784,7 +703,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
* next, if the vma overlaps with it. * next, if the vma overlaps with it.
*/ */
if (remove_next == 2 && !next->anon_vma) if (remove_next == 2 && !next->anon_vma)
exporter = next->vm_next; exporter = next_next;
} else if (end > next->vm_start) { } else if (end > next->vm_start) {
/* /*
@ -879,17 +798,14 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
if (vma->vm_end > end) { if (vma->vm_end > end) {
if (!insert || (insert->vm_start != end)) { if (!insert || (insert->vm_start != end)) {
vma_mas_szero(&mas, end, vma->vm_end); vma_mas_szero(&mas, end, vma->vm_end);
mas_reset(&mas);
VM_WARN_ON(insert && VM_WARN_ON(insert &&
insert->vm_end < vma->vm_end); insert->vm_end < vma->vm_end);
} else if (insert->vm_start == end) {
ll_prev = vma->vm_end;
} }
} else { } else {
vma_changed = true; vma_changed = true;
} }
vma->vm_end = end; vma->vm_end = end;
if (!next)
mm->highest_vm_end = vm_end_gap(vma);
} }
if (vma_changed) if (vma_changed)
@ -909,29 +825,19 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
flush_dcache_mmap_unlock(mapping); flush_dcache_mmap_unlock(mapping);
} }
if (remove_next) { if (remove_next && file) {
/* __remove_shared_vm_struct(next, file, mapping);
* vma_merge has merged next into vma, and needs
* us to remove next before dropping the locks.
* Since we have expanded over this vma, the maple tree will
* have overwritten by storing the value
*/
__vma_unlink_list(mm, next);
if (remove_next == 2) if (remove_next == 2)
__vma_unlink_list(mm, next_next); __remove_shared_vm_struct(next_next, file, mapping);
if (file) {
__remove_shared_vm_struct(next, file, mapping);
if (remove_next == 2)
__remove_shared_vm_struct(next_next, file, mapping);
}
} else if (insert) { } else if (insert) {
/* /*
* split_vma has split insert from vma, and needs * split_vma has split insert from vma, and needs
* us to insert it before dropping the locks * us to insert it before dropping the locks
* (it may either follow vma or precede it). * (it may either follow vma or precede it).
*/ */
__insert_vm_struct(mm, &mas, insert, ll_prev); mas_reset(&mas);
vma_mas_store(insert, &mas);
mm->map_count++;
} }
if (anon_vma) { if (anon_vma) {
@ -965,54 +871,12 @@ again:
/* /*
* In mprotect's case 6 (see comments on vma_merge), * In mprotect's case 6 (see comments on vma_merge),
* we must remove another next too. It would clutter * we must remove next_next too.
* up the code too much to do both in one go.
*/ */
if (remove_next != 3) {
/*
* If "next" was removed and vma->vm_end was
* expanded (up) over it, in turn
* "next->vm_prev->vm_end" changed and the
* "vma->vm_next" gap must be updated.
*/
next = next_next;
} else {
/*
* For the scope of the comment "next" and
* "vma" considered pre-swap(): if "vma" was
* removed, next->vm_start was expanded (down)
* over it and the "next" gap must be updated.
* Because of the swap() the post-swap() "vma"
* actually points to pre-swap() "next"
* (post-swap() "next" as opposed is now a
* dangling pointer).
*/
next = vma;
}
if (remove_next == 2) { if (remove_next == 2) {
remove_next = 1; remove_next = 1;
next = next_next;
goto again; goto again;
} else if (!next) {
/*
* If remove_next == 2 we obviously can't
* reach this path.
*
* If remove_next == 3 we can't reach this
* path because pre-swap() next is always not
* NULL. pre-swap() "next" is not being
* removed and its next->vm_end is not altered
* (and furthermore "end" already matches
* next->vm_end in remove_next == 3).
*
* We reach this only in the remove_next == 1
* case if the "next" vma that was removed was
* the highest vma of the mm. However in such
* case next->vm_end == "end" and the extended
* "vma" has vma->vm_end == next->vm_end so
* mm->highest_vm_end doesn't need any update
* in remove_next == 1 case.
*/
VM_WARN_ON(mm->highest_vm_end != vm_end_gap(vma));
} }
} }
if (insert && file) if (insert && file)
@ -1020,6 +884,7 @@ again:
mas_destroy(&mas); mas_destroy(&mas);
validate_mm(mm); validate_mm(mm);
return 0; return 0;
} }
@ -1179,10 +1044,10 @@ struct vm_area_struct *vma_merge(struct mm_struct *mm,
if (vm_flags & VM_SPECIAL) if (vm_flags & VM_SPECIAL)
return NULL; return NULL;
next = __vma_next(mm, prev); next = find_vma(mm, prev ? prev->vm_end : 0);
area = next; area = next;
if (area && area->vm_end == end) /* cases 6, 7, 8 */ if (area && area->vm_end == end) /* cases 6, 7, 8 */
next = next->vm_next; next = find_vma(mm, next->vm_end);
/* verify some invariant that must be enforced by the caller */ /* verify some invariant that must be enforced by the caller */
VM_WARN_ON(prev && addr <= prev->vm_start); VM_WARN_ON(prev && addr <= prev->vm_start);
@ -1316,18 +1181,24 @@ static struct anon_vma *reusable_anon_vma(struct vm_area_struct *old, struct vm_
*/ */
struct anon_vma *find_mergeable_anon_vma(struct vm_area_struct *vma) struct anon_vma *find_mergeable_anon_vma(struct vm_area_struct *vma)
{ {
MA_STATE(mas, &vma->vm_mm->mm_mt, vma->vm_end, vma->vm_end);
struct anon_vma *anon_vma = NULL; struct anon_vma *anon_vma = NULL;
struct vm_area_struct *prev, *next;
/* Try next first. */ /* Try next first. */
if (vma->vm_next) { next = mas_walk(&mas);
anon_vma = reusable_anon_vma(vma->vm_next, vma, vma->vm_next); if (next) {
anon_vma = reusable_anon_vma(next, vma, next);
if (anon_vma) if (anon_vma)
return anon_vma; return anon_vma;
} }
prev = mas_prev(&mas, 0);
VM_BUG_ON_VMA(prev != vma, vma);
prev = mas_prev(&mas, 0);
/* Try prev next. */ /* Try prev next. */
if (vma->vm_prev) if (prev)
anon_vma = reusable_anon_vma(vma->vm_prev, vma->vm_prev, vma); anon_vma = reusable_anon_vma(prev, prev, vma);
/* /*
* We might reach here with anon_vma == NULL if we can't find * We might reach here with anon_vma == NULL if we can't find
@ -2101,8 +1972,8 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address)
if (gap_addr < address || gap_addr > TASK_SIZE) if (gap_addr < address || gap_addr > TASK_SIZE)
gap_addr = TASK_SIZE; gap_addr = TASK_SIZE;
next = vma->vm_next; next = find_vma_intersection(mm, vma->vm_end, gap_addr);
if (next && next->vm_start < gap_addr && vma_is_accessible(next)) { if (next && vma_is_accessible(next)) {
if (!(next->vm_flags & VM_GROWSUP)) if (!(next->vm_flags & VM_GROWSUP))
return -ENOMEM; return -ENOMEM;
/* Check that both stack segments have the same anon_vma? */ /* Check that both stack segments have the same anon_vma? */
@ -2153,8 +2024,6 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address)
/* Overwrite old entry in mtree. */ /* Overwrite old entry in mtree. */
vma_mas_store(vma, &mas); vma_mas_store(vma, &mas);
anon_vma_interval_tree_post_update_vma(vma); anon_vma_interval_tree_post_update_vma(vma);
if (!vma->vm_next)
mm->highest_vm_end = vm_end_gap(vma);
spin_unlock(&mm->page_table_lock); spin_unlock(&mm->page_table_lock);
perf_event_mmap(vma); perf_event_mmap(vma);
@ -2174,16 +2043,16 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address)
int expand_downwards(struct vm_area_struct *vma, unsigned long address) int expand_downwards(struct vm_area_struct *vma, unsigned long address)
{ {
struct mm_struct *mm = vma->vm_mm; struct mm_struct *mm = vma->vm_mm;
MA_STATE(mas, &mm->mm_mt, vma->vm_start, vma->vm_start);
struct vm_area_struct *prev; struct vm_area_struct *prev;
int error = 0; int error = 0;
MA_STATE(mas, &mm->mm_mt, 0, 0);
address &= PAGE_MASK; address &= PAGE_MASK;
if (address < mmap_min_addr) if (address < mmap_min_addr)
return -EPERM; return -EPERM;
/* Enforce stack_guard_gap */ /* Enforce stack_guard_gap */
prev = vma->vm_prev; prev = mas_prev(&mas, 0);
/* Check that both stack segments have the same anon_vma? */ /* Check that both stack segments have the same anon_vma? */
if (prev && !(prev->vm_flags & VM_GROWSDOWN) && if (prev && !(prev->vm_flags & VM_GROWSDOWN) &&
vma_is_accessible(prev)) { vma_is_accessible(prev)) {
@ -2318,25 +2187,26 @@ find_extend_vma(struct mm_struct *mm, unsigned long addr)
EXPORT_SYMBOL_GPL(find_extend_vma); EXPORT_SYMBOL_GPL(find_extend_vma);
/* /*
* Ok - we have the memory areas we should free on the vma list, * Ok - we have the memory areas we should free on a maple tree so release them,
* so release them, and do the vma updates. * and do the vma updates.
* *
* Called with the mm semaphore held. * Called with the mm semaphore held.
*/ */
static void remove_vma_list(struct mm_struct *mm, struct vm_area_struct *vma) static inline void remove_mt(struct mm_struct *mm, struct ma_state *mas)
{ {
unsigned long nr_accounted = 0; unsigned long nr_accounted = 0;
struct vm_area_struct *vma;
/* Update high watermark before we lower total_vm */ /* Update high watermark before we lower total_vm */
update_hiwater_vm(mm); update_hiwater_vm(mm);
do { mas_for_each(mas, vma, ULONG_MAX) {
long nrpages = vma_pages(vma); long nrpages = vma_pages(vma);
if (vma->vm_flags & VM_ACCOUNT) if (vma->vm_flags & VM_ACCOUNT)
nr_accounted += nrpages; nr_accounted += nrpages;
vm_stat_account(mm, vma->vm_flags, -nrpages); vm_stat_account(mm, vma->vm_flags, -nrpages);
vma = remove_vma(vma); remove_vma(vma);
} while (vma); }
vm_unacct_memory(nr_accounted); vm_unacct_memory(nr_accounted);
validate_mm(mm); validate_mm(mm);
} }
@ -2346,18 +2216,18 @@ static void remove_vma_list(struct mm_struct *mm, struct vm_area_struct *vma)
* *
* Called with the mm semaphore held. * Called with the mm semaphore held.
*/ */
static void unmap_region(struct mm_struct *mm, static void unmap_region(struct mm_struct *mm, struct maple_tree *mt,
struct vm_area_struct *vma, struct vm_area_struct *prev, struct vm_area_struct *vma, struct vm_area_struct *prev,
struct vm_area_struct *next,
unsigned long start, unsigned long end) unsigned long start, unsigned long end)
{ {
struct vm_area_struct *next = __vma_next(mm, prev);
struct mmu_gather tlb; struct mmu_gather tlb;
lru_add_drain(); lru_add_drain();
tlb_gather_mmu(&tlb, mm); tlb_gather_mmu(&tlb, mm);
update_hiwater_rss(mm); update_hiwater_rss(mm);
unmap_vmas(&tlb, vma, start, end); unmap_vmas(&tlb, mt, vma, start, end);
free_pgtables(&tlb, vma, prev ? prev->vm_end : FIRST_USER_ADDRESS, free_pgtables(&tlb, mt, vma, prev ? prev->vm_end : FIRST_USER_ADDRESS,
next ? next->vm_start : USER_PGTABLES_CEILING); next ? next->vm_start : USER_PGTABLES_CEILING);
tlb_finish_mmu(&tlb); tlb_finish_mmu(&tlb);
} }
@ -2444,24 +2314,17 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
return __split_vma(mm, vma, addr, new_below); return __split_vma(mm, vma, addr, new_below);
} }
static inline int static inline int munmap_sidetree(struct vm_area_struct *vma,
unlock_range(struct vm_area_struct *start, struct vm_area_struct **tail, struct ma_state *mas_detach)
unsigned long limit)
{ {
struct mm_struct *mm = start->vm_mm; mas_set_range(mas_detach, vma->vm_start, vma->vm_end - 1);
struct vm_area_struct *tmp = start; if (mas_store_gfp(mas_detach, vma, GFP_KERNEL))
int count = 0; return -ENOMEM;
while (tmp && tmp->vm_start < limit) { if (vma->vm_flags & VM_LOCKED)
*tail = tmp; vma->vm_mm->locked_vm -= vma_pages(vma);
count++;
if (tmp->vm_flags & VM_LOCKED)
mm->locked_vm -= vma_pages(tmp);
tmp = tmp->vm_next; return 0;
}
return count;
} }
/* /*
@ -2481,9 +2344,13 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma,
struct mm_struct *mm, unsigned long start, struct mm_struct *mm, unsigned long start,
unsigned long end, struct list_head *uf, bool downgrade) unsigned long end, struct list_head *uf, bool downgrade)
{ {
struct vm_area_struct *prev, *last; struct vm_area_struct *prev, *next = NULL;
struct maple_tree mt_detach;
int count = 0;
int error = -ENOMEM; int error = -ENOMEM;
/* we have start < vma->vm_end */ MA_STATE(mas_detach, &mt_detach, 0, 0);
mt_init_flags(&mt_detach, MT_FLAGS_LOCK_EXTERN);
mt_set_external_lock(&mt_detach, &mm->mmap_lock);
if (mas_preallocate(mas, vma, GFP_KERNEL)) if (mas_preallocate(mas, vma, GFP_KERNEL))
return -ENOMEM; return -ENOMEM;
@ -2496,6 +2363,8 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma,
* unmapped vm_area_struct will remain in use: so lower split_vma * unmapped vm_area_struct will remain in use: so lower split_vma
* places tmp vma above, and higher split_vma places tmp vma below. * places tmp vma above, and higher split_vma places tmp vma below.
*/ */
/* Does it split the first one? */
if (start > vma->vm_start) { if (start > vma->vm_start) {
/* /*
@ -2506,35 +2375,60 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma,
if (end < vma->vm_end && mm->map_count >= sysctl_max_map_count) if (end < vma->vm_end && mm->map_count >= sysctl_max_map_count)
goto map_count_exceeded; goto map_count_exceeded;
/*
* mas_pause() is not needed since mas->index needs to be set
* differently than vma->vm_end anyways.
*/
error = __split_vma(mm, vma, start, 0); error = __split_vma(mm, vma, start, 0);
if (error) if (error)
goto split_failed; goto start_split_failed;
prev = vma; mas_set(mas, start);
vma = __vma_next(mm, prev); vma = mas_walk(mas);
mas->index = start;
mas_reset(mas);
} else {
prev = vma->vm_prev;
} }
if (vma->vm_end >= end) prev = mas_prev(mas, 0);
last = vma; if (unlikely((!prev)))
else mas_set(mas, start);
last = find_vma_intersection(mm, end - 1, end);
/* Does it split the last one? */ /*
if (last && end < last->vm_end) { * Detach a range of VMAs from the mm. Using next as a temp variable as
error = __split_vma(mm, last, end, 1); * it is always overwritten.
*/
mas_for_each(mas, next, end - 1) {
/* Does it split the end? */
if (next->vm_end > end) {
struct vm_area_struct *split;
error = __split_vma(mm, next, end, 1);
if (error)
goto end_split_failed;
mas_set(mas, end);
split = mas_prev(mas, 0);
error = munmap_sidetree(split, &mas_detach);
if (error)
goto munmap_sidetree_failed;
count++;
if (vma == next)
vma = split;
break;
}
error = munmap_sidetree(next, &mas_detach);
if (error) if (error)
goto split_failed; goto munmap_sidetree_failed;
if (vma == last) count++;
vma = __vma_next(mm, prev); #ifdef CONFIG_DEBUG_VM_MAPLE_TREE
mas_reset(mas); BUG_ON(next->vm_start < start);
BUG_ON(next->vm_start > end);
#endif
} }
if (!next)
next = mas_next(mas, ULONG_MAX);
if (unlikely(uf)) { if (unlikely(uf)) {
/* /*
* If userfaultfd_unmap_prep returns an error the vmas * If userfaultfd_unmap_prep returns an error the vmas
@ -2551,35 +2445,36 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma,
goto userfaultfd_error; goto userfaultfd_error;
} }
/* /* Point of no return */
* unlock any mlock()ed ranges before detaching vmas, count the number mas_set_range(mas, start, end - 1);
* of VMAs to be dropped, and return the tail entry of the affected #if defined(CONFIG_DEBUG_VM_MAPLE_TREE)
* area. /* Make sure no VMAs are about to be lost. */
*/ {
mm->map_count -= unlock_range(vma, &last, end); MA_STATE(test, &mt_detach, start, end - 1);
/* Drop removed area from the tree */ struct vm_area_struct *vma_mas, *vma_test;
int test_count = 0;
rcu_read_lock();
vma_test = mas_find(&test, end - 1);
mas_for_each(mas, vma_mas, end - 1) {
BUG_ON(vma_mas != vma_test);
test_count++;
vma_test = mas_next(&test, end - 1);
}
rcu_read_unlock();
BUG_ON(count != test_count);
mas_set_range(mas, start, end - 1);
}
#endif
mas_store_prealloc(mas, NULL); mas_store_prealloc(mas, NULL);
mm->map_count -= count;
/* Detach vmas from the MM linked list */
vma->vm_prev = NULL;
if (prev)
prev->vm_next = last->vm_next;
else
mm->mmap = last->vm_next;
if (last->vm_next) {
last->vm_next->vm_prev = prev;
last->vm_next = NULL;
} else
mm->highest_vm_end = prev ? vm_end_gap(prev) : 0;
/* /*
* Do not downgrade mmap_lock if we are next to VM_GROWSDOWN or * Do not downgrade mmap_lock if we are next to VM_GROWSDOWN or
* VM_GROWSUP VMA. Such VMAs can change their size under * VM_GROWSUP VMA. Such VMAs can change their size under
* down_read(mmap_lock) and collide with the VMA we are about to unmap. * down_read(mmap_lock) and collide with the VMA we are about to unmap.
*/ */
if (downgrade) { if (downgrade) {
if (last && (last->vm_flags & VM_GROWSDOWN)) if (next && (next->vm_flags & VM_GROWSDOWN))
downgrade = false; downgrade = false;
else if (prev && (prev->vm_flags & VM_GROWSUP)) else if (prev && (prev->vm_flags & VM_GROWSUP))
downgrade = false; downgrade = false;
@ -2587,18 +2482,22 @@ do_mas_align_munmap(struct ma_state *mas, struct vm_area_struct *vma,
mmap_write_downgrade(mm); mmap_write_downgrade(mm);
} }
unmap_region(mm, vma, prev, start, end); unmap_region(mm, &mt_detach, vma, prev, next, start, end);
/* Statistics and freeing VMAs */
/* Fix up all other VM information */ mas_set(&mas_detach, start);
remove_vma_list(mm, vma); remove_mt(mm, &mas_detach);
__mt_destroy(&mt_detach);
validate_mm(mm); validate_mm(mm);
return downgrade ? 1 : 0; return downgrade ? 1 : 0;
map_count_exceeded:
split_failed:
userfaultfd_error: userfaultfd_error:
munmap_sidetree_failed:
end_split_failed:
__mt_destroy(&mt_detach);
start_split_failed:
map_count_exceeded:
mas_destroy(mas); mas_destroy(mas);
return error; return error;
} }
@ -2833,7 +2732,6 @@ cannot_expand:
i_mmap_lock_write(vma->vm_file->f_mapping); i_mmap_lock_write(vma->vm_file->f_mapping);
vma_mas_store(vma, &mas); vma_mas_store(vma, &mas);
__vma_link_list(mm, vma, prev);
mm->map_count++; mm->map_count++;
if (vma->vm_file) { if (vma->vm_file) {
if (vma->vm_flags & VM_SHARED) if (vma->vm_flags & VM_SHARED)
@ -2891,7 +2789,7 @@ unmap_and_free_vma:
vma->vm_file = NULL; vma->vm_file = NULL;
/* Undo any partial mapping done by a device driver. */ /* Undo any partial mapping done by a device driver. */
unmap_region(mm, vma, prev, vma->vm_start, vma->vm_end); unmap_region(mm, mas.tree, vma, prev, next, vma->vm_start, vma->vm_end);
if (vm_flags & VM_SHARED) if (vm_flags & VM_SHARED)
mapping_unmap_writable(file->f_mapping); mapping_unmap_writable(file->f_mapping);
free_vma: free_vma:
@ -2979,11 +2877,12 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size,
goto out; goto out;
if (start + size > vma->vm_end) { if (start + size > vma->vm_end) {
struct vm_area_struct *next; VMA_ITERATOR(vmi, mm, vma->vm_end);
struct vm_area_struct *next, *prev = vma;
for (next = vma->vm_next; next; next = next->vm_next) { for_each_vma_range(vmi, next, start + size) {
/* hole between vmas ? */ /* hole between vmas ? */
if (next->vm_start != next->vm_prev->vm_end) if (next->vm_start != prev->vm_end)
goto out; goto out;
if (next->vm_file != vma->vm_file) if (next->vm_file != vma->vm_file)
@ -2992,8 +2891,7 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size,
if (next->vm_flags != vma->vm_flags) if (next->vm_flags != vma->vm_flags)
goto out; goto out;
if (start + size <= next->vm_end) prev = next;
break;
} }
if (!next) if (!next)
@ -3060,11 +2958,9 @@ static int do_brk_munmap(struct ma_state *mas, struct vm_area_struct *vma,
* do some brk-specific accounting here. * do some brk-specific accounting here.
*/ */
static int do_brk_flags(struct ma_state *mas, struct vm_area_struct *vma, static int do_brk_flags(struct ma_state *mas, struct vm_area_struct *vma,
unsigned long addr, unsigned long len, unsigned long addr, unsigned long len, unsigned long flags)
unsigned long flags)
{ {
struct mm_struct *mm = current->mm; struct mm_struct *mm = current->mm;
struct vm_area_struct *prev = NULL;
validate_mm_mt(mm); validate_mm_mt(mm);
/* /*
@ -3107,7 +3003,6 @@ static int do_brk_flags(struct ma_state *mas, struct vm_area_struct *vma,
khugepaged_enter_vma(vma, flags); khugepaged_enter_vma(vma, flags);
goto out; goto out;
} }
prev = vma;
/* create a vma struct for an anonymous mapping */ /* create a vma struct for an anonymous mapping */
vma = vm_area_alloc(mm); vma = vm_area_alloc(mm);
@ -3124,10 +3019,6 @@ static int do_brk_flags(struct ma_state *mas, struct vm_area_struct *vma,
if (mas_store_gfp(mas, vma, GFP_KERNEL)) if (mas_store_gfp(mas, vma, GFP_KERNEL))
goto mas_store_fail; goto mas_store_fail;
if (!prev)
prev = mas_prev(mas, 0);
__vma_link_list(mm, vma, prev);
mm->map_count++; mm->map_count++;
out: out:
perf_event_mmap(vma); perf_event_mmap(vma);
@ -3136,7 +3027,7 @@ out:
if (flags & VM_LOCKED) if (flags & VM_LOCKED)
mm->locked_vm += (len >> PAGE_SHIFT); mm->locked_vm += (len >> PAGE_SHIFT);
vma->vm_flags |= VM_SOFTDIRTY; vma->vm_flags |= VM_SOFTDIRTY;
validate_mm_mt(mm); validate_mm(mm);
return 0; return 0;
mas_store_fail: mas_store_fail:
@ -3217,6 +3108,8 @@ void exit_mmap(struct mm_struct *mm)
struct mmu_gather tlb; struct mmu_gather tlb;
struct vm_area_struct *vma; struct vm_area_struct *vma;
unsigned long nr_accounted = 0; unsigned long nr_accounted = 0;
MA_STATE(mas, &mm->mm_mt, 0, 0);
int count = 0;
/* mm's last user has gone, and its about to be pulled down */ /* mm's last user has gone, and its about to be pulled down */
mmu_notifier_release(mm); mmu_notifier_release(mm);
@ -3241,7 +3134,7 @@ void exit_mmap(struct mm_struct *mm)
mmap_write_lock(mm); mmap_write_lock(mm);
arch_exit_mmap(mm); arch_exit_mmap(mm);
vma = mm->mmap; vma = mas_find(&mas, ULONG_MAX);
if (!vma) { if (!vma) {
/* Can happen if dup_mmap() received an OOM */ /* Can happen if dup_mmap() received an OOM */
mmap_write_unlock(mm); mmap_write_unlock(mm);
@ -3252,22 +3145,29 @@ void exit_mmap(struct mm_struct *mm)
flush_cache_mm(mm); flush_cache_mm(mm);
tlb_gather_mmu_fullmm(&tlb, mm); tlb_gather_mmu_fullmm(&tlb, mm);
/* update_hiwater_rss(mm) here? but nobody should be looking */ /* update_hiwater_rss(mm) here? but nobody should be looking */
/* Use -1 here to ensure all VMAs in the mm are unmapped */ /* Use ULONG_MAX here to ensure all VMAs in the mm are unmapped */
unmap_vmas(&tlb, vma, 0, -1); unmap_vmas(&tlb, &mm->mm_mt, vma, 0, ULONG_MAX);
free_pgtables(&tlb, vma, FIRST_USER_ADDRESS, USER_PGTABLES_CEILING); free_pgtables(&tlb, &mm->mm_mt, vma, FIRST_USER_ADDRESS,
USER_PGTABLES_CEILING);
tlb_finish_mmu(&tlb); tlb_finish_mmu(&tlb);
/* Walk the list again, actually closing and freeing it. */ /*
while (vma) { * Walk the list again, actually closing and freeing it, with preemption
* enabled, without holding any MM locks besides the unreachable
* mmap_write_lock.
*/
do {
if (vma->vm_flags & VM_ACCOUNT) if (vma->vm_flags & VM_ACCOUNT)
nr_accounted += vma_pages(vma); nr_accounted += vma_pages(vma);
vma = remove_vma(vma); remove_vma(vma);
count++;
cond_resched(); cond_resched();
} } while ((vma = mas_find(&mas, ULONG_MAX)) != NULL);
BUG_ON(count != mm->map_count);
trace_exit_mmap(mm); trace_exit_mmap(mm);
__mt_destroy(&mm->mm_mt); __mt_destroy(&mm->mm_mt);
mm->mmap = NULL;
mmap_write_unlock(mm); mmap_write_unlock(mm);
vm_unacct_memory(nr_accounted); vm_unacct_memory(nr_accounted);
} }
@ -3306,7 +3206,7 @@ int insert_vm_struct(struct mm_struct *mm, struct vm_area_struct *vma)
vma->vm_pgoff = vma->vm_start >> PAGE_SHIFT; vma->vm_pgoff = vma->vm_start >> PAGE_SHIFT;
} }
if (vma_link(mm, vma, prev)) { if (vma_link(mm, vma)) {
vm_unacct_memory(charged); vm_unacct_memory(charged);
return -ENOMEM; return -ENOMEM;
} }
@ -3338,7 +3238,8 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap,
faulted_in_anon_vma = false; faulted_in_anon_vma = false;
} }
if (range_has_overlap(mm, addr, addr + len, &prev)) new_vma = find_vma_prev(mm, addr, &prev);
if (new_vma && new_vma->vm_start < addr + len)
return NULL; /* should never get here */ return NULL; /* should never get here */
new_vma = vma_merge(mm, prev, addr, addr + len, vma->vm_flags, new_vma = vma_merge(mm, prev, addr, addr + len, vma->vm_flags,
@ -3381,7 +3282,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap,
get_file(new_vma->vm_file); get_file(new_vma->vm_file);
if (new_vma->vm_ops && new_vma->vm_ops->open) if (new_vma->vm_ops && new_vma->vm_ops->open)
new_vma->vm_ops->open(new_vma); new_vma->vm_ops->open(new_vma);
if (vma_link(mm, new_vma, prev)) if (vma_link(mm, new_vma))
goto out_vma_link; goto out_vma_link;
*need_rmap_locks = false; *need_rmap_locks = false;
} }
@ -3686,12 +3587,13 @@ int mm_take_all_locks(struct mm_struct *mm)
{ {
struct vm_area_struct *vma; struct vm_area_struct *vma;
struct anon_vma_chain *avc; struct anon_vma_chain *avc;
MA_STATE(mas, &mm->mm_mt, 0, 0);
mmap_assert_write_locked(mm); mmap_assert_write_locked(mm);
mutex_lock(&mm_all_locks_mutex); mutex_lock(&mm_all_locks_mutex);
for (vma = mm->mmap; vma; vma = vma->vm_next) { mas_for_each(&mas, vma, ULONG_MAX) {
if (signal_pending(current)) if (signal_pending(current))
goto out_unlock; goto out_unlock;
if (vma->vm_file && vma->vm_file->f_mapping && if (vma->vm_file && vma->vm_file->f_mapping &&
@ -3699,7 +3601,8 @@ int mm_take_all_locks(struct mm_struct *mm)
vm_lock_mapping(mm, vma->vm_file->f_mapping); vm_lock_mapping(mm, vma->vm_file->f_mapping);
} }
for (vma = mm->mmap; vma; vma = vma->vm_next) { mas_set(&mas, 0);
mas_for_each(&mas, vma, ULONG_MAX) {
if (signal_pending(current)) if (signal_pending(current))
goto out_unlock; goto out_unlock;
if (vma->vm_file && vma->vm_file->f_mapping && if (vma->vm_file && vma->vm_file->f_mapping &&
@ -3707,7 +3610,8 @@ int mm_take_all_locks(struct mm_struct *mm)
vm_lock_mapping(mm, vma->vm_file->f_mapping); vm_lock_mapping(mm, vma->vm_file->f_mapping);
} }
for (vma = mm->mmap; vma; vma = vma->vm_next) { mas_set(&mas, 0);
mas_for_each(&mas, vma, ULONG_MAX) {
if (signal_pending(current)) if (signal_pending(current))
goto out_unlock; goto out_unlock;
if (vma->anon_vma) if (vma->anon_vma)
@ -3766,11 +3670,12 @@ void mm_drop_all_locks(struct mm_struct *mm)
{ {
struct vm_area_struct *vma; struct vm_area_struct *vma;
struct anon_vma_chain *avc; struct anon_vma_chain *avc;
MA_STATE(mas, &mm->mm_mt, 0, 0);
mmap_assert_write_locked(mm); mmap_assert_write_locked(mm);
BUG_ON(!mutex_is_locked(&mm_all_locks_mutex)); BUG_ON(!mutex_is_locked(&mm_all_locks_mutex));
for (vma = mm->mmap; vma; vma = vma->vm_next) { mas_for_each(&mas, vma, ULONG_MAX) {
if (vma->anon_vma) if (vma->anon_vma)
list_for_each_entry(avc, &vma->anon_vma_chain, same_vma) list_for_each_entry(avc, &vma->anon_vma_chain, same_vma)
vm_unlock_anon_vma(avc->anon_vma); vm_unlock_anon_vma(avc->anon_vma);

6
mm/nommu.c
View File

@ -584,17 +584,12 @@ static void setup_vma_to_mm(struct vm_area_struct *vma, struct mm_struct *mm)
static void mas_add_vma_to_mm(struct ma_state *mas, struct mm_struct *mm, static void mas_add_vma_to_mm(struct ma_state *mas, struct mm_struct *mm,
struct vm_area_struct *vma) struct vm_area_struct *vma)
{ {
struct vm_area_struct *prev;
BUG_ON(!vma->vm_region); BUG_ON(!vma->vm_region);
setup_vma_to_mm(vma, mm); setup_vma_to_mm(vma, mm);
prev = mas_prev(mas, 0);
mas_reset(mas);
/* add the VMA to the tree */ /* add the VMA to the tree */
vma_mas_store(vma, mas); vma_mas_store(vma, mas);
__vma_link_list(mm, vma, prev);
} }
/* /*
@ -647,7 +642,6 @@ static int delete_vma_from_mm(struct vm_area_struct *vma)
/* remove from the MM's tree and list */ /* remove from the MM's tree and list */
vma_mas_remove(vma, &mas); vma_mas_remove(vma, &mas);
__vma_unlink_list(vma->vm_mm, vma);
return 0; return 0;
} }

40
mm/util.c
View File

@ -272,46 +272,6 @@ void *memdup_user_nul(const void __user *src, size_t len)
} }
EXPORT_SYMBOL(memdup_user_nul); EXPORT_SYMBOL(memdup_user_nul);
void __vma_link_list(struct mm_struct *mm, struct vm_area_struct *vma,
struct vm_area_struct *prev)
{
struct vm_area_struct *next;
vma->vm_prev = prev;
if (prev) {
next = prev->vm_next;
prev->vm_next = vma;
} else {
next = mm->mmap;
mm->mmap = vma;
}
vma->vm_next = next;
if (next)
next->vm_prev = vma;
else
mm->highest_vm_end = vm_end_gap(vma);
}
void __vma_unlink_list(struct mm_struct *mm, struct vm_area_struct *vma)
{
struct vm_area_struct *prev, *next;
next = vma->vm_next;
prev = vma->vm_prev;
if (prev)
prev->vm_next = next;
else
mm->mmap = next;
if (next) {
next->vm_prev = prev;
} else {
if (prev)
mm->highest_vm_end = vm_end_gap(prev);
else
mm->highest_vm_end = 0;
}
}
/* Check if the vma is being used as a stack by this task */ /* Check if the vma is being used as a stack by this task */
int vma_is_stack_for_current(struct vm_area_struct *vma) int vma_is_stack_for_current(struct vm_area_struct *vma)
{ {