mirrors/linux-stable
1
0
Fork 0
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git synced 2024-11-01 00:48:50 +00:00
Code Issues Releases Wiki Activity

mm/munlock: update Documentation/vm/unevictable-lru.rst

Browse source 
Update Documentation/vm/unevictable-lru.rst to reflect the changes made
by the mm/munlock series: keeping an mlock_count instead of page_mlock()
(formerly try_to_munlock()) and munlock_vma_pages_all() etc.  Also make
other little updates or cleanups wherever noticed.

But, I apologize, this is already out of date, in that "folio" appears
nowhere: 5.18 will be in a transitional state from "page" to "folio",
and documenting its current mix of the two does not help to understand
"the Unevictable LRU".  Should be revisited when naming is more settled.

Link: https://lkml.kernel.org/r/3753962-d491-bf60-f59f-51bfe84fd6a0@google.com
Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: David Hildenbrand <david@redhat.com>
Cc: Alistair Popple <apopple@nvidia.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Rik van Riel <riel@surriel.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Yu Zhao <yuzhao@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Shakeel Butt <shakeelb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This commit is contained in:
Hugh Dickins 2022-04-01 11:28:30 -07:00 committed by Linus Torvalds
parent ece369c7e1
commit 577e9846f8
1 changed files with 199 additions and 250 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

449
Documentation/vm/unevictable-lru.rst
View file

@ -52,8 +52,13 @@ The infrastructure may also be able to handle other conditions that make pages
unevictable, either by definition or by circumstance, in the future. unevictable, either by definition or by circumstance, in the future.
The Unevictable Page List The Unevictable LRU Page List
------------------------- -----------------------------
The Unevictable LRU page list is a lie. It was never an LRU-ordered list, but a
companion to the LRU-ordered anonymous and file, active and inactive page lists;
and now it is not even a page list. But following familiar convention, here in
this document and in the source, we often imagine it as a fifth LRU page list.
The Unevictable LRU infrastructure consists of an additional, per-node, LRU list The Unevictable LRU infrastructure consists of an additional, per-node, LRU list
called the "unevictable" list and an associated page flag, PG_unevictable, to called the "unevictable" list and an associated page flag, PG_unevictable, to
@ -63,8 +68,8 @@ The PG_unevictable flag is analogous to, and mutually exclusive with, the
PG_active flag in that it indicates on which LRU list a page resides when PG_active flag in that it indicates on which LRU list a page resides when
PG_lru is set. PG_lru is set.
The Unevictable LRU infrastructure maintains unevictable pages on an additional The Unevictable LRU infrastructure maintains unevictable pages as if they were
LRU list for a few reasons: on an additional LRU list for a few reasons:
(1) We get to "treat unevictable pages just like we treat other pages in the (1) We get to "treat unevictable pages just like we treat other pages in the
system - which means we get to use the same code to manipulate them, the system - which means we get to use the same code to manipulate them, the
@ -72,13 +77,11 @@ LRU list for a few reasons:
of the statistics, etc..." [Rik van Riel] of the statistics, etc..." [Rik van Riel]
(2) We want to be able to migrate unevictable pages between nodes for memory (2) We want to be able to migrate unevictable pages between nodes for memory
defragmentation, workload management and memory hotplug. The linux kernel defragmentation, workload management and memory hotplug. The Linux kernel
can only migrate pages that it can successfully isolate from the LRU can only migrate pages that it can successfully isolate from the LRU
lists. If we were to maintain pages elsewhere than on an LRU-like list, lists (or "Movable" pages: outside of consideration here). If we were to
where they can be found by isolate_lru_page(), we would prevent their maintain pages elsewhere than on an LRU-like list, where they can be
migration, unless we reworked migration code to find the unevictable pages detected by isolate_lru_page(), we would prevent their migration.
itself.
The unevictable list does not differentiate between file-backed and anonymous, The unevictable list does not differentiate between file-backed and anonymous,
swap-backed pages. This differentiation is only important while the pages are, swap-backed pages. This differentiation is only important while the pages are,
@ -92,8 +95,8 @@ Memory Control Group Interaction
-------------------------------- --------------------------------
The unevictable LRU facility interacts with the memory control group [aka The unevictable LRU facility interacts with the memory control group [aka
memory controller; see Documentation/admin-guide/cgroup-v1/memory.rst] by extending the memory controller; see Documentation/admin-guide/cgroup-v1/memory.rst] by
lru_list enum. extending the lru_list enum.
The memory controller data structure automatically gets a per-node unevictable The memory controller data structure automatically gets a per-node unevictable
list as a result of the "arrayification" of the per-node LRU lists (one per list as a result of the "arrayification" of the per-node LRU lists (one per
@ -143,7 +146,6 @@ These are currently used in three places in the kernel:
and this mark remains for the life of the inode. and this mark remains for the life of the inode.
(2) By SYSV SHM to mark SHM_LOCK'd address spaces until SHM_UNLOCK is called. (2) By SYSV SHM to mark SHM_LOCK'd address spaces until SHM_UNLOCK is called.
Note that SHM_LOCK is not required to page in the locked pages if they're Note that SHM_LOCK is not required to page in the locked pages if they're
swapped out; the application must touch the pages manually if it wants to swapped out; the application must touch the pages manually if it wants to
ensure they're in memory. ensure they're in memory.
@ -156,19 +158,19 @@ These are currently used in three places in the kernel:
Detecting Unevictable Pages Detecting Unevictable Pages
--------------------------- ---------------------------
The function page_evictable() in vmscan.c determines whether a page is The function page_evictable() in mm/internal.h determines whether a page is
evictable or not using the query function outlined above [see section evictable or not using the query function outlined above [see section
:ref:`Marking address spaces unevictable <mark_addr_space_unevict>`] :ref:`Marking address spaces unevictable <mark_addr_space_unevict>`]
to check the AS_UNEVICTABLE flag. to check the AS_UNEVICTABLE flag.
For address spaces that are so marked after being populated (as SHM regions For address spaces that are so marked after being populated (as SHM regions
might be), the lock action (eg: SHM_LOCK) can be lazy, and need not populate might be), the lock action (e.g. SHM_LOCK) can be lazy, and need not populate
the page tables for the region as does, for example, mlock(), nor need it make the page tables for the region as does, for example, mlock(), nor need it make
any special effort to push any pages in the SHM_LOCK'd area to the unevictable any special effort to push any pages in the SHM_LOCK'd area to the unevictable
list. Instead, vmscan will do this if and when it encounters the pages during list. Instead, vmscan will do this if and when it encounters the pages during
a reclamation scan. a reclamation scan.
On an unlock action (such as SHM_UNLOCK), the unlocker (eg: shmctl()) must scan On an unlock action (such as SHM_UNLOCK), the unlocker (e.g. shmctl()) must scan
the pages in the region and "rescue" them from the unevictable list if no other the pages in the region and "rescue" them from the unevictable list if no other
condition is keeping them unevictable. If an unevictable region is destroyed, condition is keeping them unevictable. If an unevictable region is destroyed,
the pages are also "rescued" from the unevictable list in the process of the pages are also "rescued" from the unevictable list in the process of
@ -176,7 +178,7 @@ freeing them.
page_evictable() also checks for mlocked pages by testing an additional page page_evictable() also checks for mlocked pages by testing an additional page
flag, PG_mlocked (as wrapped by PageMlocked()), which is set when a page is flag, PG_mlocked (as wrapped by PageMlocked()), which is set when a page is
faulted into a VM_LOCKED vma, or found in a vma being VM_LOCKED. faulted into a VM_LOCKED VMA, or found in a VMA being VM_LOCKED.
Vmscan's Handling of Unevictable Pages Vmscan's Handling of Unevictable Pages
@ -186,28 +188,23 @@ If unevictable pages are culled in the fault path, or moved to the unevictable
list at mlock() or mmap() time, vmscan will not encounter the pages until they list at mlock() or mmap() time, vmscan will not encounter the pages until they
have become evictable again (via munlock() for example) and have been "rescued" have become evictable again (via munlock() for example) and have been "rescued"
from the unevictable list. However, there may be situations where we decide, from the unevictable list. However, there may be situations where we decide,
for the sake of expediency, to leave a unevictable page on one of the regular for the sake of expediency, to leave an unevictable page on one of the regular
active/inactive LRU lists for vmscan to deal with. vmscan checks for such active/inactive LRU lists for vmscan to deal with. vmscan checks for such
pages in all of the shrink_{active|inactive|page}_list() functions and will pages in all of the shrink_{active|inactive|page}_list() functions and will
"cull" such pages that it encounters: that is, it diverts those pages to the "cull" such pages that it encounters: that is, it diverts those pages to the
unevictable list for the node being scanned. unevictable list for the memory cgroup and node being scanned.
There may be situations where a page is mapped into a VM_LOCKED VMA, but the There may be situations where a page is mapped into a VM_LOCKED VMA, but the
page is not marked as PG_mlocked. Such pages will make it all the way to page is not marked as PG_mlocked. Such pages will make it all the way to
shrink_page_list() where they will be detected when vmscan walks the reverse shrink_active_list() or shrink_page_list() where they will be detected when
map in try_to_unmap(). If try_to_unmap() returns SWAP_MLOCK, vmscan walks the reverse map in page_referenced() or try_to_unmap(). The page
shrink_page_list() will cull the page at that point. is culled to the unevictable list when it is released by the shrinker.
To "cull" an unevictable page, vmscan simply puts the page back on the LRU list To "cull" an unevictable page, vmscan simply puts the page back on the LRU list
using putback_lru_page() - the inverse operation to isolate_lru_page() - after using putback_lru_page() - the inverse operation to isolate_lru_page() - after
dropping the page lock. Because the condition which makes the page unevictable dropping the page lock. Because the condition which makes the page unevictable
may change once the page is unlocked, putback_lru_page() will recheck the may change once the page is unlocked, __pagevec_lru_add_fn() will recheck the
unevictable state of a page that it places on the unevictable list. If the unevictable state of a page before placing it on the unevictable list.
page has become unevictable, putback_lru_page() removes it from the list and
retries, including the page_unevictable() test. Because such a race is a rare
event and movement of pages onto the unevictable list should be rare, these
extra evictabilty checks should not occur in the majority of calls to
putback_lru_page().
MLOCKED Pages MLOCKED Pages
@ -227,16 +224,25 @@ Nick posted his patch as an alternative to a patch posted by Christoph Lameter
to achieve the same objective: hiding mlocked pages from vmscan. to achieve the same objective: hiding mlocked pages from vmscan.
In Nick's patch, he used one of the struct page LRU list link fields as a count In Nick's patch, he used one of the struct page LRU list link fields as a count
of VM_LOCKED VMAs that map the page. This use of the link field for a count of VM_LOCKED VMAs that map the page (Rik van Riel had the same idea three years
prevented the management of the pages on an LRU list, and thus mlocked pages earlier). But this use of the link field for a count prevented the management
were not migratable as isolate_lru_page() could not find them, and the LRU list of the pages on an LRU list, and thus mlocked pages were not migratable as
link field was not available to the migration subsystem. isolate_lru_page() could not detect them, and the LRU list link field was not
available to the migration subsystem.
Nick resolved this by putting mlocked pages back on the lru list before Nick resolved this by putting mlocked pages back on the LRU list before
attempting to isolate them, thus abandoning the count of VM_LOCKED VMAs. When attempting to isolate them, thus abandoning the count of VM_LOCKED VMAs. When
Nick's patch was integrated with the Unevictable LRU work, the count was Nick's patch was integrated with the Unevictable LRU work, the count was
replaced by walking the reverse map to determine whether any VM_LOCKED VMAs replaced by walking the reverse map when munlocking, to determine whether any
mapped the page. More on this below. other VM_LOCKED VMAs still mapped the page.
However, walking the reverse map for each page when munlocking was ugly and
inefficient, and could lead to catastrophic contention on a file's rmap lock,
when many processes which had it mlocked were trying to exit. In 5.18, the
idea of keeping mlock_count in Unevictable LRU list link field was revived and
put to work, without preventing the migration of mlocked pages. This is why
the "Unevictable LRU list" cannot be a linked list of pages now; but there was
no use for that linked list anyway - though its size is maintained for meminfo.
Basic Management Basic Management
@ -250,22 +256,18 @@ PageMlocked() functions.
A PG_mlocked page will be placed on the unevictable list when it is added to A PG_mlocked page will be placed on the unevictable list when it is added to
the LRU. Such pages can be "noticed" by memory management in several places: the LRU. Such pages can be "noticed" by memory management in several places:
(1) in the mlock()/mlockall() system call handlers; (1) in the mlock()/mlock2()/mlockall() system call handlers;
(2) in the mmap() system call handler when mmapping a region with the (2) in the mmap() system call handler when mmapping a region with the
MAP_LOCKED flag; MAP_LOCKED flag;
(3) mmapping a region in a task that has called mlockall() with the MCL_FUTURE (3) mmapping a region in a task that has called mlockall() with the MCL_FUTURE
flag flag;
(4) in the fault path, if mlocked pages are "culled" in the fault path, (4) in the fault path and when a VM_LOCKED stack segment is expanded; or
and when a VM_LOCKED stack segment is expanded; or
(5) as mentioned above, in vmscan:shrink_page_list() when attempting to (5) as mentioned above, in vmscan:shrink_page_list() when attempting to
reclaim a page in a VM_LOCKED VMA via try_to_unmap() reclaim a page in a VM_LOCKED VMA by page_referenced() or try_to_unmap().
all of which result in the VM_LOCKED flag being set for the VMA if it doesn't
already have it set.
mlocked pages become unlocked and rescued from the unevictable list when: mlocked pages become unlocked and rescued from the unevictable list when:
@ -280,51 +282,53 @@ mlocked pages become unlocked and rescued from the unevictable list when:
(4) before a page is COW'd in a VM_LOCKED VMA. (4) before a page is COW'd in a VM_LOCKED VMA.
mlock()/mlockall() System Call Handling mlock()/mlock2()/mlockall() System Call Handling
--------------------------------------- ------------------------------------------------
Both [do\_]mlock() and [do\_]mlockall() system call handlers call mlock_fixup() mlock(), mlock2() and mlockall() system call handlers proceed to mlock_fixup()
for each VMA in the range specified by the call. In the case of mlockall(), for each VMA in the range specified by the call. In the case of mlockall(),
this is the entire active address space of the task. Note that mlock_fixup() this is the entire active address space of the task. Note that mlock_fixup()
is used for both mlocking and munlocking a range of memory. A call to mlock() is used for both mlocking and munlocking a range of memory. A call to mlock()
an already VM_LOCKED VMA, or to munlock() a VMA that is not VM_LOCKED is an already VM_LOCKED VMA, or to munlock() a VMA that is not VM_LOCKED, is
treated as a no-op, and mlock_fixup() simply returns. treated as a no-op and mlock_fixup() simply returns.
If the VMA passes some filtering as described in "Filtering Special Vmas" If the VMA passes some filtering as described in "Filtering Special VMAs"
below, mlock_fixup() will attempt to merge the VMA with its neighbors or split below, mlock_fixup() will attempt to merge the VMA with its neighbors or split
off a subset of the VMA if the range does not cover the entire VMA. Once the off a subset of the VMA if the range does not cover the entire VMA. Any pages
VMA has been merged or split or neither, mlock_fixup() will call already present in the VMA are then marked as mlocked by mlock_page() via
populate_vma_page_range() to fault in the pages via get_user_pages() and to mlock_pte_range() via walk_page_range() via mlock_vma_pages_range().
mark the pages as mlocked via mlock_vma_page().
Before returning from the system call, do_mlock() or mlockall() will call
__mm_populate() to fault in the remaining pages via get_user_pages() and to
mark those pages as mlocked as they are faulted.
Note that the VMA being mlocked might be mapped with PROT_NONE. In this case, Note that the VMA being mlocked might be mapped with PROT_NONE. In this case,
get_user_pages() will be unable to fault in the pages. That's okay. If pages get_user_pages() will be unable to fault in the pages. That's okay. If pages
do end up getting faulted into this VM_LOCKED VMA, we'll handle them in the do end up getting faulted into this VM_LOCKED VMA, they will be handled in the
fault path or in vmscan. fault path - which is also how mlock2()'s MLOCK_ONFAULT areas are handled.
Also note that a page returned by get_user_pages() could be truncated or For each PTE (or PMD) being faulted into a VMA, the page add rmap function
migrated out from under us, while we're trying to mlock it. To detect this, calls mlock_vma_page(), which calls mlock_page() when the VMA is VM_LOCKED
populate_vma_page_range() checks page_mapping() after acquiring the page lock. (unless it is a PTE mapping of a part of a transparent huge page). Or when
If the page is still associated with its mapping, we'll go ahead and call it is a newly allocated anonymous page, lru_cache_add_inactive_or_unevictable()
mlock_vma_page(). If the mapping is gone, we just unlock the page and move on. calls mlock_new_page() instead: similar to mlock_page(), but can make better
In the worst case, this will result in a page mapped in a VM_LOCKED VMA judgments, since this page is held exclusively and known not to be on LRU yet.
remaining on a normal LRU list without being PageMlocked(). Again, vmscan will
detect and cull such pages.
mlock_vma_page() will call TestSetPageMlocked() for each page returned by mlock_page() sets PageMlocked immediately, then places the page on the CPU's
get_user_pages(). We use TestSetPageMlocked() because the page might already mlock pagevec, to batch up the rest of the work to be done under lru_lock by
be mlocked by another task/VMA and we don't want to do extra work. We __mlock_page(). __mlock_page() sets PageUnevictable, initializes mlock_count
especially do not want to count an mlocked page more than once in the and moves the page to unevictable state ("the unevictable LRU", but with
statistics. If the page was already mlocked, mlock_vma_page() need do nothing mlock_count in place of LRU threading). Or if the page was already PageLRU
more. and PageUnevictable and PageMlocked, it simply increments the mlock_count.
If the page was NOT already mlocked, mlock_vma_page() attempts to isolate the But in practice that may not work ideally: the page may not yet be on an LRU, or
page from the LRU, as it is likely on the appropriate active or inactive list it may have been temporarily isolated from LRU. In such cases the mlock_count
at that time. If the isolate_lru_page() succeeds, mlock_vma_page() will put field cannot be touched, but will be set to 0 later when __pagevec_lru_add_fn()
back the page - by calling putback_lru_page() - which will notice that the page returns the page to "LRU". Races prohibit mlock_count from being set to 1 then:
is now mlocked and divert the page to the node's unevictable list. If rather than risk stranding a page indefinitely as unevictable, always err with
mlock_vma_page() is unable to isolate the page from the LRU, vmscan will handle mlock_count on the low side, so that when munlocked the page will be rescued to
it later if and when it attempts to reclaim the page. an evictable LRU, then perhaps be mlocked again later if vmscan finds it in a
VM_LOCKED VMA.
Filtering Special VMAs Filtering Special VMAs
@ -339,68 +343,48 @@ mlock_fixup() filters several classes of "special" VMAs:
so there is no sense in attempting to visit them. so there is no sense in attempting to visit them.
2) VMAs mapping hugetlbfs page are already effectively pinned into memory. We 2) VMAs mapping hugetlbfs page are already effectively pinned into memory. We
neither need nor want to mlock() these pages. However, to preserve the neither need nor want to mlock() these pages. But __mm_populate() includes
prior behavior of mlock() - before the unevictable/mlock changes - hugetlbfs ranges, allocating the huge pages and populating the PTEs.
mlock_fixup() will call make_pages_present() in the hugetlbfs VMA range to
allocate the huge pages and populate the ptes.
3) VMAs with VM_DONTEXPAND are generally userspace mappings of kernel pages, 3) VMAs with VM_DONTEXPAND are generally userspace mappings of kernel pages,
such as the VDSO page, relay channel pages, etc. These pages such as the VDSO page, relay channel pages, etc. These pages are inherently
are inherently unevictable and are not managed on the LRU lists. unevictable and are not managed on the LRU lists. __mm_populate() includes
mlock_fixup() treats these VMAs the same as hugetlbfs VMAs. It calls these ranges, populating the PTEs if not already populated.
make_pages_present() to populate the ptes.
4) VMAs with VM_MIXEDMAP set are not marked VM_LOCKED, but __mm_populate()
includes these ranges, populating the PTEs if not already populated.
Note that for all of these special VMAs, mlock_fixup() does not set the Note that for all of these special VMAs, mlock_fixup() does not set the
VM_LOCKED flag. Therefore, we won't have to deal with them later during VM_LOCKED flag. Therefore, we won't have to deal with them later during
munlock(), munmap() or task exit. Neither does mlock_fixup() account these munlock(), munmap() or task exit. Neither does mlock_fixup() account these
VMAs against the task's "locked_vm". VMAs against the task's "locked_vm".
.. _munlock_munlockall_handling:
munlock()/munlockall() System Call Handling munlock()/munlockall() System Call Handling
------------------------------------------- -------------------------------------------
The munlock() and munlockall() system calls are handled by the same functions - The munlock() and munlockall() system calls are handled by the same
do_mlock[all]() - as the mlock() and mlockall() system calls with the unlock vs mlock_fixup() function as mlock(), mlock2() and mlockall() system calls are.
lock operation indicated by an argument. So, these system calls are also If called to munlock an already munlocked VMA, mlock_fixup() simply returns.
handled by mlock_fixup(). Again, if called for an already munlocked VMA, Because of the VMA filtering discussed above, VM_LOCKED will not be set in
mlock_fixup() simply returns. Because of the VMA filtering discussed above, any "special" VMAs. So, those VMAs will be ignored for munlock.
VM_LOCKED will not be set in any "special" VMAs. So, these VMAs will be
ignored for munlock.
If the VMA is VM_LOCKED, mlock_fixup() again attempts to merge or split off the If the VMA is VM_LOCKED, mlock_fixup() again attempts to merge or split off the
specified range. The range is then munlocked via the function specified range. All pages in the VMA are then munlocked by munlock_page() via
populate_vma_page_range() - the same function used to mlock a VMA range - mlock_pte_range() via walk_page_range() via mlock_vma_pages_range() - the same
passing a flag to indicate that munlock() is being performed. function used when mlocking a VMA range, with new flags for the VMA indicating
that it is munlock() being performed.
Because the VMA access protections could have been changed to PROT_NONE after munlock_page() uses the mlock pagevec to batch up work to be done under
faulting in and mlocking pages, get_user_pages() was unreliable for visiting lru_lock by __munlock_page(). __munlock_page() decrements the page's
these pages for munlocking. Because we don't want to leave pages mlocked, mlock_count, and when that reaches 0 it clears PageMlocked and clears
get_user_pages() was enhanced to accept a flag to ignore the permissions when PageUnevictable, moving the page from unevictable state to inactive LRU.
fetching the pages - all of which should be resident as a result of previous
mlocking.
For munlock(), populate_vma_page_range() unlocks individual pages by calling But in practice that may not work ideally: the page may not yet have reached
munlock_vma_page(). munlock_vma_page() unconditionally clears the PG_mlocked "the unevictable LRU", or it may have been temporarily isolated from it. In
flag using TestClearPageMlocked(). As with mlock_vma_page(), those cases its mlock_count field is unusable and must be assumed to be 0: so
munlock_vma_page() use the Test*PageMlocked() function to handle the case where that the page will be rescued to an evictable LRU, then perhaps be mlocked
the page might have already been unlocked by another task. If the page was again later if vmscan finds it in a VM_LOCKED VMA.
mlocked, munlock_vma_page() updates that zone statistics for the number of
mlocked pages. Note, however, that at this point we haven't checked whether
the page is mapped by other VM_LOCKED VMAs.
We can't call page_mlock(), the function that walks the reverse map to
check for other VM_LOCKED VMAs, without first isolating the page from the LRU.
page_mlock() is a variant of try_to_unmap() and thus requires that the page
not be on an LRU list [more on these below]. However, the call to
isolate_lru_page() could fail, in which case we can't call page_mlock(). So,
we go ahead and clear PG_mlocked up front, as this might be the only chance we
have. If we can successfully isolate the page, we go ahead and call
page_mlock(), which will restore the PG_mlocked flag and update the zone
page statistics if it finds another VMA holding the page mlocked. If we fail
to isolate the page, we'll have left a potentially mlocked page on the LRU.
This is fine, because we'll catch it later if and if vmscan tries to reclaim
the page. This should be relatively rare.
Migrating MLOCKED Pages Migrating MLOCKED Pages
@ -410,33 +394,38 @@ A page that is being migrated has been isolated from the LRU lists and is held
locked across unmapping of the page, updating the page's address space entry locked across unmapping of the page, updating the page's address space entry
and copying the contents and state, until the page table entry has been and copying the contents and state, until the page table entry has been
replaced with an entry that refers to the new page. Linux supports migration replaced with an entry that refers to the new page. Linux supports migration
of mlocked pages and other unevictable pages. This involves simply moving the of mlocked pages and other unevictable pages. PG_mlocked is cleared from the
PG_mlocked and PG_unevictable states from the old page to the new page. the old page when it is unmapped from the last VM_LOCKED VMA, and set when the
new page is mapped in place of migration entry in a VM_LOCKED VMA. If the page
was unevictable because mlocked, PG_unevictable follows PG_mlocked; but if the
page was unevictable for other reasons, PG_unevictable is copied explicitly.
Note that page migration can race with mlocking or munlocking of the same page. Note that page migration can race with mlocking or munlocking of the same page.
This has been discussed from the mlock/munlock perspective in the respective There is mostly no problem since page migration requires unmapping all PTEs of
sections above. Both processes (migration and m[un]locking) hold the page the old page (including munlock where VM_LOCKED), then mapping in the new page
locked. This provides the first level of synchronization. Page migration (including mlock where VM_LOCKED). The page table locks provide sufficient
zeros out the page_mapping of the old page before unlocking it, so m[un]lock synchronization.
can skip these pages by testing the page mapping under page lock.
To complete page migration, we place the new and old pages back onto the LRU However, since mlock_vma_pages_range() starts by setting VM_LOCKED on a VMA,
after dropping the page lock. The "unneeded" page - old page on success, new before mlocking any pages already present, if one of those pages were migrated
page on failure - will be freed when the reference count held by the migration before mlock_pte_range() reached it, it would get counted twice in mlock_count.
process is released. To ensure that we don't strand pages on the unevictable To prevent that, mlock_vma_pages_range() temporarily marks the VMA as VM_IO,
list because of a race between munlock and migration, page migration uses the so that mlock_vma_page() will skip it.
putback_lru_page() function to add migrated pages back to the LRU.
To complete page migration, we place the old and new pages back onto the LRU
afterwards. The "unneeded" page - old page on success, new page on failure -
is freed when the reference count held by the migration process is released.
Compacting MLOCKED Pages Compacting MLOCKED Pages
------------------------ ------------------------
The unevictable LRU can be scanned for compactable regions and the default The memory map can be scanned for compactable regions and the default behavior
behavior is to do so. /proc/sys/vm/compact_unevictable_allowed controls is to let unevictable pages be moved. /proc/sys/vm/compact_unevictable_allowed
this behavior (see Documentation/admin-guide/sysctl/vm.rst). Once scanning of the controls this behavior (see Documentation/admin-guide/sysctl/vm.rst). The work
unevictable LRU is enabled, the work of compaction is mostly handled by of compaction is mostly handled by the page migration code and the same work
the page migration code and the same work flow as described in MIGRATING flow as described in Migrating MLOCKED Pages will apply.
MLOCKED PAGES will apply.
MLOCKING Transparent Huge Pages MLOCKING Transparent Huge Pages
------------------------------- -------------------------------
@ -445,51 +434,44 @@ A transparent huge page is represented by a single entry on an LRU list.
Therefore, we can only make unevictable an entire compound page, not Therefore, we can only make unevictable an entire compound page, not
individual subpages. individual subpages.
If a user tries to mlock() part of a huge page, we want the rest of the If a user tries to mlock() part of a huge page, and no user mlock()s the
page to be reclaimable. whole of the huge page, we want the rest of the page to be reclaimable.
We cannot just split the page on partial mlock() as split_huge_page() can We cannot just split the page on partial mlock() as split_huge_page() can
fail and new intermittent failure mode for the syscall is undesirable. fail and a new intermittent failure mode for the syscall is undesirable.
We handle this by keeping PTE-mapped huge pages on normal LRU lists: the We handle this by keeping PTE-mlocked huge pages on evictable LRU lists:
PMD on border of VM_LOCKED VMA will be split into PTE table. the PMD on the border of a VM_LOCKED VMA will be split into a PTE table.
This way the huge page is accessible for vmscan. Under memory pressure the This way the huge page is accessible for vmscan. Under memory pressure the
page will be split, subpages which belong to VM_LOCKED VMAs will be moved page will be split, subpages which belong to VM_LOCKED VMAs will be moved
to unevictable LRU and the rest can be reclaimed. to the unevictable LRU and the rest can be reclaimed.
/proc/meminfo's Unevictable and Mlocked amounts do not include those parts
of a transparent huge page which are mapped only by PTEs in VM_LOCKED VMAs.
See also comment in follow_trans_huge_pmd().
mmap(MAP_LOCKED) System Call Handling mmap(MAP_LOCKED) System Call Handling
------------------------------------- -------------------------------------
In addition the mlock()/mlockall() system calls, an application can request In addition to the mlock(), mlock2() and mlockall() system calls, an application
that a region of memory be mlocked supplying the MAP_LOCKED flag to the mmap() can request that a region of memory be mlocked by supplying the MAP_LOCKED flag
call. There is one important and subtle difference here, though. mmap() + mlock() to the mmap() call. There is one important and subtle difference here, though.
will fail if the range cannot be faulted in (e.g. because mm_populate fails) mmap() + mlock() will fail if the range cannot be faulted in (e.g. because
and returns with ENOMEM while mmap(MAP_LOCKED) will not fail. The mmaped mm_populate fails) and returns with ENOMEM while mmap(MAP_LOCKED) will not fail.
area will still have properties of the locked area - aka. pages will not get The mmaped area will still have properties of the locked area - pages will not
swapped out - but major page faults to fault memory in might still happen. get swapped out - but major page faults to fault memory in might still happen.
Furthermore, any mmap() call or brk() call that expands the heap by a Furthermore, any mmap() call or brk() call that expands the heap by a task
task that has previously called mlockall() with the MCL_FUTURE flag will result that has previously called mlockall() with the MCL_FUTURE flag will result
in the newly mapped memory being mlocked. Before the unevictable/mlock in the newly mapped memory being mlocked. Before the unevictable/mlock
changes, the kernel simply called make_pages_present() to allocate pages and changes, the kernel simply called make_pages_present() to allocate pages
populate the page table. and populate the page table.
To mlock a range of memory under the unevictable/mlock infrastructure, the To mlock a range of memory under the unevictable/mlock infrastructure,
mmap() handler and task address space expansion functions call the mmap() handler and task address space expansion functions call
populate_vma_page_range() specifying the vma and the address range to mlock. populate_vma_page_range() specifying the vma and the address range to mlock.
The callers of populate_vma_page_range() will have already added the memory range
to be mlocked to the task's "locked_vm". To account for filtered VMAs,
populate_vma_page_range() returns the number of pages NOT mlocked. All of the
callers then subtract a non-negative return value from the task's locked_vm. A
negative return value represent an error - for example, from get_user_pages()
attempting to fault in a VMA with PROT_NONE access. In this case, we leave the
memory range accounted as locked_vm, as the protections could be changed later
and pages allocated into that region.
munmap()/exit()/exec() System Call Handling munmap()/exit()/exec() System Call Handling
------------------------------------------- -------------------------------------------
@ -500,81 +482,53 @@ munlock the pages if we're removing the last VM_LOCKED VMA that maps the pages.
Before the unevictable/mlock changes, mlocking did not mark the pages in any Before the unevictable/mlock changes, mlocking did not mark the pages in any
way, so unmapping them required no processing. way, so unmapping them required no processing.
To munlock a range of memory under the unevictable/mlock infrastructure, the For each PTE (or PMD) being unmapped from a VMA, page_remove_rmap() calls
munmap() handler and task address space call tear down function munlock_vma_page(), which calls munlock_page() when the VMA is VM_LOCKED
munlock_vma_pages_all(). The name reflects the observation that one always (unless it was a PTE mapping of a part of a transparent huge page).
specifies the entire VMA range when munlock()ing during unmap of a region.
Because of the VMA filtering when mlocking() regions, only "normal" VMAs that
actually contain mlocked pages will be passed to munlock_vma_pages_all().
munlock_vma_pages_all() clears the VM_LOCKED VMA flag and, like mlock_fixup() munlock_page() uses the mlock pagevec to batch up work to be done under
for the munlock case, calls __munlock_vma_pages_range() to walk the page table lru_lock by __munlock_page(). __munlock_page() decrements the page's
for the VMA's memory range and munlock_vma_page() each resident page mapped by mlock_count, and when that reaches 0 it clears PageMlocked and clears
the VMA. This effectively munlocks the page, only if this is the last PageUnevictable, moving the page from unevictable state to inactive LRU.
VM_LOCKED VMA that maps the page.
But in practice that may not work ideally: the page may not yet have reached
"the unevictable LRU", or it may have been temporarily isolated from it. In
those cases its mlock_count field is unusable and must be assumed to be 0: so
that the page will be rescued to an evictable LRU, then perhaps be mlocked
again later if vmscan finds it in a VM_LOCKED VMA.
try_to_unmap() Truncating MLOCKED Pages
-------------- ------------------------
Pages can, of course, be mapped into multiple VMAs. Some of these VMAs may File truncation or hole punching forcibly unmaps the deleted pages from
have VM_LOCKED flag set. It is possible for a page mapped into one or more userspace; truncation even unmaps and deletes any private anonymous pages
VM_LOCKED VMAs not to have the PG_mlocked flag set and therefore reside on one which had been Copied-On-Write from the file pages now being truncated.
of the active or inactive LRU lists. This could happen if, for example, a task
in the process of munlocking the page could not isolate the page from the LRU.
As a result, vmscan/shrink_page_list() might encounter such a page as described
in section "vmscan's handling of unevictable pages". To handle this situation,
try_to_unmap() checks for VM_LOCKED VMAs while it is walking a page's reverse
map.
try_to_unmap() is always called, by either vmscan for reclaim or for page Mlocked pages can be munlocked and deleted in this way: like with munmap(),
migration, with the argument page locked and isolated from the LRU. Separate for each PTE (or PMD) being unmapped from a VMA, page_remove_rmap() calls
functions handle anonymous and mapped file and KSM pages, as these types of munlock_vma_page(), which calls munlock_page() when the VMA is VM_LOCKED
pages have different reverse map lookup mechanisms, with different locking. (unless it was a PTE mapping of a part of a transparent huge page).
In each case, whether rmap_walk_anon() or rmap_walk_file() or rmap_walk_ksm(),
it will call try_to_unmap_one() for every VMA which might contain the page.
When trying to reclaim, if try_to_unmap_one() finds the page in a VM_LOCKED However, if there is a racing munlock(), since mlock_vma_pages_range() starts
VMA, it will then mlock the page via mlock_vma_page() instead of unmapping it, munlocking by clearing VM_LOCKED from a VMA, before munlocking all the pages
and return SWAP_MLOCK to indicate that the page is unevictable: and the scan present, if one of those pages were unmapped by truncation or hole punch before
stops there. mlock_pte_range() reached it, it would not be recognized as mlocked by this VMA,
and would not be counted out of mlock_count. In this rare case, a page may
mlock_vma_page() is called while holding the page table's lock (in addition still appear as PageMlocked after it has been fully unmapped: and it is left to
to the page lock, and the rmap lock): to serialize against concurrent mlock or release_pages() (or __page_cache_release()) to clear it and update statistics
munlock or munmap system calls, mm teardown (munlock_vma_pages_all), reclaim, before freeing (this event is counted in /proc/vmstat unevictable_pgs_cleared,
holepunching, and truncation of file pages and their anonymous COWed pages. which is usually 0).
page_mlock() Reverse Map Scan
---------------------------------
When munlock_vma_page() [see section :ref:`munlock()/munlockall() System Call
Handling <munlock_munlockall_handling>` above] tries to munlock a
page, it needs to determine whether or not the page is mapped by any
VM_LOCKED VMA without actually attempting to unmap all PTEs from the
page. For this purpose, the unevictable/mlock infrastructure
introduced a variant of try_to_unmap() called page_mlock().
page_mlock() walks the respective reverse maps looking for VM_LOCKED VMAs. When
such a VMA is found the page is mlocked via mlock_vma_page(). This undoes the
pre-clearing of the page's PG_mlocked done by munlock_vma_page.
Note that page_mlock()'s reverse map walk must visit every VMA in a page's
reverse map to determine that a page is NOT mapped into any VM_LOCKED VMA.
However, the scan can terminate when it encounters a VM_LOCKED VMA.
Although page_mlock() might be called a great many times when munlocking a
large region or tearing down a large address space that has been mlocked via
mlockall(), overall this is a fairly rare event.
Page Reclaim in shrink_*_list() Page Reclaim in shrink_*_list()
------------------------------- -------------------------------
shrink_active_list() culls any obviously unevictable pages - i.e. vmscan's shrink_active_list() culls any obviously unevictable pages -
!page_evictable(page) - diverting these to the unevictable list. i.e. !page_evictable(page) pages - diverting those to the unevictable list.
However, shrink_active_list() only sees unevictable pages that made it onto the However, shrink_active_list() only sees unevictable pages that made it onto the
active/inactive lru lists. Note that these pages do not have PageUnevictable active/inactive LRU lists. Note that these pages do not have PageUnevictable
set - otherwise they would be on the unevictable list and shrink_active_list set - otherwise they would be on the unevictable list and shrink_active_list()
would never see them. would never see them.
Some examples of these unevictable pages on the LRU lists are: Some examples of these unevictable pages on the LRU lists are:
@ -586,20 +540,15 @@ Some examples of these unevictable pages on the LRU lists are:
when an application accesses the page the first time after SHM_LOCK'ing when an application accesses the page the first time after SHM_LOCK'ing
the segment. the segment.
(3) mlocked pages that could not be isolated from the LRU and moved to the (3) pages still mapped into VM_LOCKED VMAs, which should be marked mlocked,
unevictable list in mlock_vma_page(). but events left mlock_count too low, so they were munlocked too early.
shrink_inactive_list() also diverts any unevictable pages that it finds on the vmscan's shrink_inactive_list() and shrink_page_list() also divert obviously
inactive lists to the appropriate node's unevictable list. unevictable pages found on the inactive lists to the appropriate memory cgroup
and node unevictable list.
shrink_inactive_list() should only see SHM_LOCK'd pages that became SHM_LOCK'd rmap's page_referenced_one(), called via vmscan's shrink_active_list() or
after shrink_active_list() had moved them to the inactive list, or pages mapped shrink_page_list(), and rmap's try_to_unmap_one() called via shrink_page_list(),
into VM_LOCKED VMAs that munlock_vma_page() couldn't isolate from the LRU to check for (3) pages still mapped into VM_LOCKED VMAs, and call mlock_vma_page()
recheck via page_mlock(). shrink_inactive_list() won't notice the latter, to correct them. Such pages are culled to the unevictable list when released
but will pass on to shrink_page_list(). by the shrinker.
shrink_page_list() again culls obviously unevictable pages that it could
encounter for similar reason to shrink_inactive_list(). Pages mapped into
VM_LOCKED VMAs but without PG_mlocked set will make it all the way to
try_to_unmap(). shrink_page_list() will divert them to the unevictable list
when try_to_unmap() returns SWAP_MLOCK, as discussed above.