Lines Matching refs:the
13 This document describes the Linux memory manager's "Unevictable LRU"
14 infrastructure and the use of this to manage several types of "unevictable"
17 The document attempts to provide the overall rationale behind this mechanism
18 and the rationale for some of the design decisions that drove the
19 implementation. The latter design rationale is discussed in the context of an
20 implementation description. Admittedly, one can obtain the implementation
21 details - the "what does it do?" - by reading the code. One hopes that the
22 descriptions below add value by provide the answer to "why does it do that?".
38 will spend a lot of time scanning the LRU lists looking for the small fraction
40 spending 100% of their time in vmscan for hours or days on end, with the system
43 The unevictable list addresses the following classes of unevictable pages:
52 unevictable, either by definition or by circumstance, in the future.
59 called the "unevictable" list and an associated page flag, PG_unevictable, to
60 indicate that the page is being managed on the unevictable list.
62 The PG_unevictable flag is analogous to, and mutually exclusive with, the
69 (1) We get to "treat unevictable pages just like we treat other pages in the
70 system - which means we get to use the same code to manipulate them, the
71 same code to isolate them (for migrate, etc.), the same code to keep track
72 of the statistics, etc..." [Rik van Riel]
76 can only migrate pages that it can successfully isolate from the LRU
79 migration, unless we reworked migration code to find the unevictable pages
84 swap-backed pages. This differentiation is only important while the pages are,
87 The unevictable list benefits from the "arrayification" of the per-zone LRU
90 The unevictable list does not use the LRU pagevec mechanism. Rather,
91 unevictable pages are placed directly on the page's zone's unevictable list
92 under the zone lru_lock. This allows us to prevent the stranding of pages on
93 the unevictable list when one task has the page isolated from the LRU and other
94 tasks are changing the "evictability" state of the page.
100 The unevictable LRU facility interacts with the memory control group [aka
101 memory controller; see Documentation/cgroup-v1/memory.txt] by extending the
105 list as a result of the "arrayification" of the per-zone LRU lists (one per
106 lru_list enum element). The memory controller tracks the movement of pages to
107 and from the unevictable list.
109 When a memory control group comes under memory pressure, the controller will
110 not attempt to reclaim pages on the unevictable list. This has a couple of
113 (1) Because the pages are "hidden" from reclaim on the unevictable list, the
117 (2) On the other hand, if too many of the pages charged to the control group
118 are unevictable, the evictable portion of the working set of the tasks in
119 the control group may not fit into the available memory. This can cause
120 the control group to thrash or to OOM-kill tasks.
128 For facilities such as ramfs none of the pages attached to the address space
129 may be evicted. To prevent eviction of any such pages, the AS_UNEVICTABLE
135 Mark the address space as being completely unevictable.
139 Mark the address space as being evictable.
143 Query the address space, and return true if it is completely
146 These are currently used in two places in the kernel:
148 (1) By ramfs to mark the address spaces of its inodes when they are created,
149 and this mark remains for the life of the inode.
153 Note that SHM_LOCK is not required to page in the locked pages if they're
154 swapped out; the application must touch the pages manually if it wants to
162 evictable or not using the query function outlined above [see section
164 to check the AS_UNEVICTABLE flag.
167 might be), the lock action (eg: SHM_LOCK) can be lazy, and need not populate
168 the page tables for the region as does, for example, mlock(), nor need it make
169 any special effort to push any pages in the SHM_LOCK'd area to the unevictable
170 list. Instead, vmscan will do this if and when it encounters the pages during
173 On an unlock action (such as SHM_UNLOCK), the unlocker (eg: shmctl()) must scan
174 the pages in the region and "rescue" them from the unevictable list if no other
176 the pages are also "rescued" from the unevictable list in the process of
187 If unevictable pages are culled in the fault path, or moved to the unevictable
188 list at mlock() or mmap() time, vmscan will not encounter the pages until they
190 from the unevictable list. However, there may be situations where we decide,
191 for the sake of expediency, to leave a unevictable page on one of the regular
193 pages in all of the shrink_{active|inactive|page}_list() functions and will
194 "cull" such pages that it encounters: that is, it diverts those pages to the
195 unevictable list for the zone being scanned.
197 There may be situations where a page is mapped into a VM_LOCKED VMA, but the
198 page is not marked as PG_mlocked. Such pages will make it all the way to
199 shrink_page_list() where they will be detected when vmscan walks the reverse
201 shrink_page_list() will cull the page at that point.
203 To "cull" an unevictable page, vmscan simply puts the page back on the LRU list
204 using putback_lru_page() - the inverse operation to isolate_lru_page() - after
205 dropping the page lock. Because the condition which makes the page unevictable
206 may change once the page is unlocked, putback_lru_page() will recheck the
207 unevictable state of a page that it places on the unevictable list. If the
208 page has become unevictable, putback_lru_page() removes it from the list and
209 retries, including the page_unevictable() test. Because such a race is a rare
210 event and movement of pages onto the unevictable list should be rare, these
211 extra evictabilty checks should not occur in the majority of calls to
229 to achieve the same objective: hiding mlocked pages from vmscan.
231 In Nick's patch, he used one of the struct page LRU list link fields as a count
232 of VM_LOCKED VMAs that map the page. This use of the link field for a count
233 prevented the management of the pages on an LRU list, and thus mlocked pages
234 were not migratable as isolate_lru_page() could not find them, and the LRU list
235 link field was not available to the migration subsystem.
237 Nick resolved this by putting mlocked pages back on the lru list before
238 attempting to isolate them, thus abandoning the count of VM_LOCKED VMAs. When
239 Nick's patch was integrated with the Unevictable LRU work, the count was
240 replaced by walking the reverse map to determine whether any VM_LOCKED VMAs
241 mapped the page. More on this below.
248 pages. When such a page has been "noticed" by the memory management subsystem,
249 the page is marked with the PG_mlocked flag. This can be manipulated using the
252 A PG_mlocked page will be placed on the unevictable list when it is added to
253 the LRU. Such pages can be "noticed" by memory management in several places:
255 (1) in the mlock()/mlockall() system call handlers;
257 (2) in the mmap() system call handler when mmapping a region with the
260 (3) mmapping a region in a task that has called mlockall() with the MCL_FUTURE
263 (4) in the fault path, if mlocked pages are "culled" in the fault path,
269 all of which result in the VM_LOCKED flag being set for the VMA if it doesn't
272 mlocked pages become unlocked and rescued from the unevictable list when:
274 (1) mapped in a range unlocked via the munlock()/munlockall() system calls;
276 (2) munmap()'d out of the last VM_LOCKED VMA that maps the page, including
279 (3) when the page is truncated from the last VM_LOCKED VMA of an mmapped file;
289 for each VMA in the range specified by the call. In the case of mlockall(),
290 this is the entire active address space of the task. Note that mlock_fixup()
295 If the VMA passes some filtering as described in "Filtering Special Vmas"
296 below, mlock_fixup() will attempt to merge the VMA with its neighbors or split
297 off a subset of the VMA if the range does not cover the entire VMA. Once the
299 populate_vma_page_range() to fault in the pages via get_user_pages() and to
300 mark the pages as mlocked via mlock_vma_page().
302 Note that the VMA being mlocked might be mapped with PROT_NONE. In this case,
303 get_user_pages() will be unable to fault in the pages. That's okay. If pages
304 do end up getting faulted into this VM_LOCKED VMA, we'll handle them in the
309 populate_vma_page_range() checks page_mapping() after acquiring the page lock.
310 If the page is still associated with its mapping, we'll go ahead and call
311 mlock_vma_page(). If the mapping is gone, we just unlock the page and move on.
312 In the worst case, this will result in a page mapped in a VM_LOCKED VMA
317 get_user_pages(). We use TestSetPageMlocked() because the page might already
319 especially do not want to count an mlocked page more than once in the
320 statistics. If the page was already mlocked, mlock_vma_page() need do nothing
323 If the page was NOT already mlocked, mlock_vma_page() attempts to isolate the
324 page from the LRU, as it is likely on the appropriate active or inactive list
325 at that time. If the isolate_lru_page() succeeds, mlock_vma_page() will put
326 back the page - by calling putback_lru_page() - which will notice that the page
327 is now mlocked and divert the page to the zone's unevictable list. If
328 mlock_vma_page() is unable to isolate the page from the LRU, vmscan will handle
329 it later if and when it attempts to reclaim the page.
339 mlocked. In any case, most of the pages have no struct page in which to so
340 mark the page. Because of this, get_user_pages() will fail for these VMAs,
344 neither need nor want to mlock() these pages. However, to preserve the
345 prior behavior of mlock() - before the unevictable/mlock changes -
346 mlock_fixup() will call make_pages_present() in the hugetlbfs VMA range to
347 allocate the huge pages and populate the ptes.
350 such as the VDSO page, relay channel pages, etc. These pages
351 are inherently unevictable and are not managed on the LRU lists.
352 mlock_fixup() treats these VMAs the same as hugetlbfs VMAs. It calls
353 make_pages_present() to populate the ptes.
355 Note that for all of these special VMAs, mlock_fixup() does not set the
358 VMAs against the task's "locked_vm".
365 The munlock() and munlockall() system calls are handled by the same functions -
366 do_mlock[all]() - as the mlock() and mlockall() system calls with the unlock vs
369 mlock_fixup() simply returns. Because of the VMA filtering discussed above,
373 If the VMA is VM_LOCKED, mlock_fixup() again attempts to merge or split off the
374 specified range. The range is then munlocked via the function
375 populate_vma_page_range() - the same function used to mlock a VMA range -
378 Because the VMA access protections could have been changed to PROT_NONE after
381 get_user_pages() was enhanced to accept a flag to ignore the permissions when
382 fetching the pages - all of which should be resident as a result of previous
386 munlock_vma_page(). munlock_vma_page() unconditionally clears the PG_mlocked
388 munlock_vma_page() use the Test*PageMlocked() function to handle the case where
389 the page might have already been unlocked by another task. If the page was
390 mlocked, munlock_vma_page() updates that zone statistics for the number of
392 the page is mapped by other VM_LOCKED VMAs.
394 We can't call try_to_munlock(), the function that walks the reverse map to
395 check for other VM_LOCKED VMAs, without first isolating the page from the LRU.
396 try_to_munlock() is a variant of try_to_unmap() and thus requires that the page
397 not be on an LRU list [more on these below]. However, the call to
399 we go ahead and clear PG_mlocked up front, as this might be the only chance we
400 have. If we can successfully isolate the page, we go ahead and
401 try_to_munlock(), which will restore the PG_mlocked flag and update the zone
402 page statistics if it finds another VMA holding the page mlocked. If we fail
403 to isolate the page, we'll have left a potentially mlocked page on the LRU.
405 the page. This should be relatively rare.
411 A page that is being migrated has been isolated from the LRU lists and is held
412 locked across unmapping of the page, updating the page's address space entry
413 and copying the contents and state, until the page table entry has been
414 replaced with an entry that refers to the new page. Linux supports migration
415 of mlocked pages and other unevictable pages. This involves simply moving the
416 PG_mlocked and PG_unevictable states from the old page to the new page.
418 Note that page migration can race with mlocking or munlocking of the same page.
419 This has been discussed from the mlock/munlock perspective in the respective
420 sections above. Both processes (migration and m[un]locking) hold the page
421 locked. This provides the first level of synchronization. Page migration
422 zeros out the page_mapping of the old page before unlocking it, so m[un]lock
423 can skip these pages by testing the page mapping under page lock.
425 To complete page migration, we place the new and old pages back onto the LRU
426 after dropping the page lock. The "unneeded" page - old page on success, new
427 page on failure - will be freed when the reference count held by the migration
428 process is released. To ensure that we don't strand pages on the unevictable
429 list because of a race between munlock and migration, page migration uses the
430 putback_lru_page() function to add migrated pages back to the LRU.
436 The unevictable LRU can be scanned for compactable regions and the default
438 this behavior (see Documentation/sysctl/vm.txt). Once scanning of the
439 unevictable LRU is enabled, the work of compaction is mostly handled by
440 the page migration code and the same work flow as described in MIGRATING
450 If a user tries to mlock() part of a huge page, we want the rest of the
453 We cannot just split the page on partial mlock() as split_huge_page() can
454 fail and new intermittent failure mode for the syscall is undesirable.
456 We handle this by keeping PTE-mapped huge pages on normal LRU lists: the
459 This way the huge page is accessible for vmscan. Under memory pressure the
461 to unevictable LRU and the rest can be reclaimed.
468 In addition the mlock()/mlockall() system calls, an application can request
469 that a region of memory be mlocked supplying the MAP_LOCKED flag to the mmap()
471 will fail if the range cannot be faulted in (e.g. because mm_populate fails)
473 area will still have properties of the locked area - aka. pages will not get
476 Furthermore, any mmap() call or brk() call that expands the heap by a
477 task that has previously called mlockall() with the MCL_FUTURE flag will result
478 in the newly mapped memory being mlocked. Before the unevictable/mlock
479 changes, the kernel simply called make_pages_present() to allocate pages and
480 populate the page table.
482 To mlock a range of memory under the unevictable/mlock infrastructure, the
484 populate_vma_page_range() specifying the vma and the address range to mlock.
486 The callers of populate_vma_page_range() will have already added the memory range
487 to be mlocked to the task's "locked_vm". To account for filtered VMAs,
488 populate_vma_page_range() returns the number of pages NOT mlocked. All of the
489 callers then subtract a non-negative return value from the task's locked_vm. A
491 attempting to fault in a VMA with PROT_NONE access. In this case, we leave the
492 memory range accounted as locked_vm, as the protections could be changed later
501 munlock the pages if we're removing the last VM_LOCKED VMA that maps the pages.
502 Before the unevictable/mlock changes, mlocking did not mark the pages in any
505 To munlock a range of memory under the unevictable/mlock infrastructure, the
507 munlock_vma_pages_all(). The name reflects the observation that one always
508 specifies the entire VMA range when munlock()ing during unmap of a region.
509 Because of the VMA filtering when mlocking() regions, only "normal" VMAs that
512 munlock_vma_pages_all() clears the VM_LOCKED VMA flag and, like mlock_fixup()
513 for the munlock case, calls __munlock_vma_pages_range() to walk the page table
514 for the VMA's memory range and munlock_vma_page() each resident page mapped by
515 the VMA. This effectively munlocks the page, only if this is the last
516 VM_LOCKED VMA that maps the page.
524 VM_LOCKED VMAs not to have the PG_mlocked flag set and therefore reside on one
525 of the active or inactive LRU lists. This could happen if, for example, a task
526 in the process of munlocking the page could not isolate the page from the LRU.
533 migration, with the argument page locked and isolated from the LRU. Separate
537 it will call try_to_unmap_one() for every VMA which might contain the page.
539 When trying to reclaim, if try_to_unmap_one() finds the page in a VM_LOCKED
540 VMA, it will then mlock the page via mlock_vma_page() instead of unmapping it,
541 and return SWAP_MLOCK to indicate that the page is unevictable: and the scan
544 mlock_vma_page() is called while holding the page table's lock (in addition
545 to the page lock, and the rmap lock): to serialize against concurrent mlock or
554 [!] TODO/FIXME: a better name might be page_mlocked() - analogous to the
559 page, it needs to determine whether or not the page is mapped by any
560 VM_LOCKED VMA without actually attempting to unmap all PTEs from the
561 page. For this purpose, the unevictable/mlock infrastructure
564 try_to_munlock() calls the same functions as try_to_unmap() for anonymous and
566 processing. Again, these functions walk the respective reverse maps looking
567 for VM_LOCKED VMAs. When such a VMA is found, as in the try_to_unmap() case,
568 the functions mlock the page via mlock_vma_page() and return SWAP_MLOCK. This
569 undoes the pre-clearing of the page's PG_mlocked done by munlock_vma_page.
573 However, the scan can terminate when it encounters a VM_LOCKED VMA.
583 !page_evictable(page) - diverting these to the unevictable list.
584 However, shrink_active_list() only sees unevictable pages that made it onto the
586 set - otherwise they would be on the unevictable list and shrink_active_list
589 Some examples of these unevictable pages on the LRU lists are:
591 (1) ramfs pages that have been placed on the LRU lists when first allocated.
594 allocate or fault in the pages in the shared memory region. This happens
595 when an application accesses the page the first time after SHM_LOCK'ing
596 the segment.
598 (3) mlocked pages that could not be isolated from the LRU and moved to the
601 shrink_inactive_list() also diverts any unevictable pages that it finds on the
602 inactive lists to the appropriate zone's unevictable list.
605 after shrink_active_list() had moved them to the inactive list, or pages mapped
606 into VM_LOCKED VMAs that munlock_vma_page() couldn't isolate from the LRU to
607 recheck via try_to_munlock(). shrink_inactive_list() won't notice the latter,
612 VM_LOCKED VMAs but without PG_mlocked set will make it all the way to
613 try_to_unmap(). shrink_page_list() will divert them to the unevictable list