diff options
author | Michel Lespinasse <walken@google.com> | 2012-10-08 16:31:39 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2012-10-09 16:22:41 +0900 |
commit | bf181b9f9d8dfbba58b23441ad60d0bc33806d64 (patch) | |
tree | 7ad0caaf8998f31c5d910dcbb768f5a1d381b5f4 /mm/mmap.c | |
parent | 108d6642ad81bb1d62b401490a334d2c12397517 (diff) |
mm anon rmap: replace same_anon_vma linked list with an interval tree.
When a large VMA (anon or private file mapping) is first touched, which
will populate its anon_vma field, and then split into many regions through
the use of mprotect(), the original anon_vma ends up linking all of the
vmas on a linked list. This can cause rmap to become inefficient, as we
have to walk potentially thousands of irrelevent vmas before finding the
one a given anon page might fall into.
By replacing the same_anon_vma linked list with an interval tree (where
each avc's interval is determined by its vma's start and last pgoffs), we
can make rmap efficient for this use case again.
While the change is large, all of its pieces are fairly simple.
Most places that were walking the same_anon_vma list were looking for a
known pgoff, so they can just use the anon_vma_interval_tree_foreach()
interval tree iterator instead. The exception here is ksm, where the
page's index is not known. It would probably be possible to rework ksm so
that the index would be known, but for now I have decided to keep things
simple and just walk the entirety of the interval tree there.
When updating vma's that already have an anon_vma assigned, we must take
care to re-index the corresponding avc's on their interval tree. This is
done through the use of anon_vma_interval_tree_pre_update_vma() and
anon_vma_interval_tree_post_update_vma(), which remove the avc's from
their interval tree before the update and re-insert them after the update.
The anon_vma stays locked during the update, so there is no chance that
rmap would miss the vmas that are being updated.
Signed-off-by: Michel Lespinasse <walken@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Daniel Santos <daniel.santos@pobox.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'mm/mmap.c')
-rw-r--r-- | mm/mmap.c | 73 |
1 files changed, 55 insertions, 18 deletions
diff --git a/mm/mmap.c b/mm/mmap.c index 66984aab7915..2e580ed79211 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -353,6 +353,38 @@ void validate_mm(struct mm_struct *mm) #define validate_mm(mm) do { } while (0) #endif +/* + * vma has some anon_vma assigned, and is already inserted on that + * anon_vma's interval trees. + * + * Before updating the vma's vm_start / vm_end / vm_pgoff fields, the + * vma must be removed from the anon_vma's interval trees using + * anon_vma_interval_tree_pre_update_vma(). + * + * After the update, the vma will be reinserted using + * anon_vma_interval_tree_post_update_vma(). + * + * The entire update must be protected by exclusive mmap_sem and by + * the root anon_vma's mutex. + */ +static inline void +anon_vma_interval_tree_pre_update_vma(struct vm_area_struct *vma) +{ + struct anon_vma_chain *avc; + + list_for_each_entry(avc, &vma->anon_vma_chain, same_vma) + anon_vma_interval_tree_remove(avc, &avc->anon_vma->rb_root); +} + +static inline void +anon_vma_interval_tree_post_update_vma(struct vm_area_struct *vma) +{ + struct anon_vma_chain *avc; + + list_for_each_entry(avc, &vma->anon_vma_chain, same_vma) + anon_vma_interval_tree_insert(avc, &avc->anon_vma->rb_root); +} + static int find_vma_links(struct mm_struct *mm, unsigned long addr, unsigned long end, struct vm_area_struct **pprev, struct rb_node ***rb_link, struct rb_node **rb_parent) @@ -565,20 +597,17 @@ again: remove_next = 1 + (end > next->vm_end); vma_adjust_trans_huge(vma, start, end, adjust_next); - /* - * When changing only vma->vm_end, we don't really need anon_vma - * lock. This is a fairly rare case by itself, but the anon_vma - * lock may be shared between many sibling processes. Skipping - * the lock for brk adjustments makes a difference sometimes. - */ - if (vma->anon_vma && (importer || start != vma->vm_start)) { - anon_vma = vma->anon_vma; + anon_vma = vma->anon_vma; + if (!anon_vma && adjust_next) + anon_vma = next->anon_vma; + if (anon_vma) { VM_BUG_ON(adjust_next && next->anon_vma && anon_vma != next->anon_vma); - } else if (adjust_next && next->anon_vma) - anon_vma = next->anon_vma; - if (anon_vma) anon_vma_lock(anon_vma); + anon_vma_interval_tree_pre_update_vma(vma); + if (adjust_next) + anon_vma_interval_tree_pre_update_vma(next); + } if (root) { flush_dcache_mmap_lock(mapping); @@ -619,8 +648,12 @@ again: remove_next = 1 + (end > next->vm_end); __insert_vm_struct(mm, insert); } - if (anon_vma) + if (anon_vma) { + anon_vma_interval_tree_post_update_vma(vma); + if (adjust_next) + anon_vma_interval_tree_post_update_vma(next); anon_vma_unlock(anon_vma); + } if (mapping) mutex_unlock(&mapping->i_mmap_mutex); @@ -1748,7 +1781,9 @@ int expand_upwards(struct vm_area_struct *vma, unsigned long address) if (vma->vm_pgoff + (size >> PAGE_SHIFT) >= vma->vm_pgoff) { error = acct_stack_growth(vma, size, grow); if (!error) { + anon_vma_interval_tree_pre_update_vma(vma); vma->vm_end = address; + anon_vma_interval_tree_post_update_vma(vma); perf_event_mmap(vma); } } @@ -1798,8 +1833,10 @@ int expand_downwards(struct vm_area_struct *vma, if (grow <= vma->vm_pgoff) { error = acct_stack_growth(vma, size, grow); if (!error) { + anon_vma_interval_tree_pre_update_vma(vma); vma->vm_start = address; vma->vm_pgoff -= grow; + anon_vma_interval_tree_post_update_vma(vma); perf_event_mmap(vma); } } @@ -2515,7 +2552,7 @@ static DEFINE_MUTEX(mm_all_locks_mutex); static void vm_lock_anon_vma(struct mm_struct *mm, struct anon_vma *anon_vma) { - if (!test_bit(0, (unsigned long *) &anon_vma->root->head.next)) { + if (!test_bit(0, (unsigned long *) &anon_vma->root->rb_root.rb_node)) { /* * The LSB of head.next can't change from under us * because we hold the mm_all_locks_mutex. @@ -2531,7 +2568,7 @@ static void vm_lock_anon_vma(struct mm_struct *mm, struct anon_vma *anon_vma) * anon_vma->root->mutex. */ if (__test_and_set_bit(0, (unsigned long *) - &anon_vma->root->head.next)) + &anon_vma->root->rb_root.rb_node)) BUG(); } } @@ -2572,7 +2609,7 @@ static void vm_lock_mapping(struct mm_struct *mm, struct address_space *mapping) * A single task can't take more than one mm_take_all_locks() in a row * or it would deadlock. * - * The LSB in anon_vma->head.next and the AS_MM_ALL_LOCKS bitflag in + * The LSB in anon_vma->rb_root.rb_node and the AS_MM_ALL_LOCKS bitflag in * mapping->flags avoid to take the same lock twice, if more than one * vma in this mm is backed by the same anon_vma or address_space. * @@ -2619,13 +2656,13 @@ out_unlock: static void vm_unlock_anon_vma(struct anon_vma *anon_vma) { - if (test_bit(0, (unsigned long *) &anon_vma->root->head.next)) { + if (test_bit(0, (unsigned long *) &anon_vma->root->rb_root.rb_node)) { /* * The LSB of head.next can't change to 0 from under * us because we hold the mm_all_locks_mutex. * * We must however clear the bitflag before unlocking - * the vma so the users using the anon_vma->head will + * the vma so the users using the anon_vma->rb_root will * never see our bitflag. * * No need of atomic instructions here, head.next @@ -2633,7 +2670,7 @@ static void vm_unlock_anon_vma(struct anon_vma *anon_vma) * anon_vma->root->mutex. */ if (!__test_and_clear_bit(0, (unsigned long *) - &anon_vma->root->head.next)) + &anon_vma->root->rb_root.rb_node)) BUG(); anon_vma_unlock(anon_vma); } |