diff options
author | Vlastimil Babka <vbabka@suse.cz> | 2016-01-14 15:19:20 -0800 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2016-01-14 16:00:49 -0800 |
commit | 6a15a37097c7e02390bb08d83dac433c9f10144f (patch) | |
tree | 6571e0fd8b882842d89f36897706977936c75c1d /mm | |
parent | c261e7d94f0dd33a34b6cf98686e8b9699b62340 (diff) | |
download | op-kernel-dev-6a15a37097c7e02390bb08d83dac433c9f10144f.zip op-kernel-dev-6a15a37097c7e02390bb08d83dac433c9f10144f.tar.gz |
mm, proc: reduce cost of /proc/pid/smaps for shmem mappings
The previous patch has improved swap accounting for shmem mapping, which
however made /proc/pid/smaps more expensive for shmem mappings, as we
consult the radix tree for each pte_none entry, so the overal complexity
is O(n*log(n)).
We can reduce this significantly for mappings that cannot contain COWed
pages, because then we can either use the statistics tha shmem object
itself tracks (if the mapping contains the whole object, or the swap
usage of the whole object is zero), or use the radix tree iterator,
which is much more effective than repeated find_get_entry() calls.
This patch therefore introduces a function shmem_swap_usage(vma) and
makes /proc/pid/smaps use it when possible. Only for writable private
mappings of shmem objects (i.e. tmpfs files) with the shmem object
itself (partially) swapped outwe have to resort to the find_get_entry()
approach.
Hopefully such mappings are relatively uncommon.
To demonstrate the diference, I have measured this on a process that
creates a 2GB mapping and dirties single pages with a stride of 2MB, and
time how long does it take to cat /proc/pid/smaps of this process 100
times.
Private writable mapping of a /dev/shm/file (the most complex case):
real 0m3.831s
user 0m0.180s
sys 0m3.212s
Shared mapping of an almost full mapping of a partially swapped /dev/shm/file
(which needs to employ the radix tree iterator).
real 0m1.351s
user 0m0.096s
sys 0m0.768s
Same, but with /dev/shm/file not swapped (so no radix tree walk needed)
real 0m0.935s
user 0m0.128s
sys 0m0.344s
Private anonymous mapping:
real 0m0.949s
user 0m0.116s
sys 0m0.348s
The cost is now much closer to the private anonymous mapping case, unless
the shmem mapping is private and writable.
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Hugh Dickins <hughd@google.com>
Cc: Jerome Marchand <jmarchan@redhat.com>
Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'mm')
-rw-r--r-- | mm/shmem.c | 70 |
1 files changed, 70 insertions, 0 deletions
@@ -360,6 +360,76 @@ static int shmem_free_swap(struct address_space *mapping, } /* + * Determine (in bytes) how many of the shmem object's pages mapped by the + * given vma is swapped out. + * + * This is safe to call without i_mutex or mapping->tree_lock thanks to RCU, + * as long as the inode doesn't go away and racy results are not a problem. + */ +unsigned long shmem_swap_usage(struct vm_area_struct *vma) +{ + struct inode *inode = file_inode(vma->vm_file); + struct shmem_inode_info *info = SHMEM_I(inode); + struct address_space *mapping = inode->i_mapping; + unsigned long swapped; + pgoff_t start, end; + struct radix_tree_iter iter; + void **slot; + struct page *page; + + /* Be careful as we don't hold info->lock */ + swapped = READ_ONCE(info->swapped); + + /* + * The easier cases are when the shmem object has nothing in swap, or + * the vma maps it whole. Then we can simply use the stats that we + * already track. + */ + if (!swapped) + return 0; + + if (!vma->vm_pgoff && vma->vm_end - vma->vm_start >= inode->i_size) + return swapped << PAGE_SHIFT; + + swapped = 0; + + /* Here comes the more involved part */ + start = linear_page_index(vma, vma->vm_start); + end = linear_page_index(vma, vma->vm_end); + + rcu_read_lock(); + +restart: + radix_tree_for_each_slot(slot, &mapping->page_tree, &iter, start) { + if (iter.index >= end) + break; + + page = radix_tree_deref_slot(slot); + + /* + * This should only be possible to happen at index 0, so we + * don't need to reset the counter, nor do we risk infinite + * restarts. + */ + if (radix_tree_deref_retry(page)) + goto restart; + + if (radix_tree_exceptional_entry(page)) + swapped++; + + if (need_resched()) { + cond_resched_rcu(); + start = iter.index + 1; + goto restart; + } + } + + rcu_read_unlock(); + + return swapped << PAGE_SHIFT; +} + +/* * SysV IPC SHM_UNLOCK restore Unevictable pages to their evictable lists. */ void shmem_unlock_mapping(struct address_space *mapping) |