1.. SPDX-License-Identifier: GPL-2.0
2
3=============
4Multi-Gen LRU
5=============
6The multi-gen LRU is an alternative LRU implementation that optimizes
7page reclaim and improves performance under memory pressure. Page
8reclaim decides the kernel's caching policy and ability to overcommit
9memory. It directly impacts the kswapd CPU usage and RAM efficiency.
10
11Design overview
12===============
13Objectives
14----------
15The design objectives are:
16
17* Good representation of access recency
18* Try to profit from spatial locality
19* Fast paths to make obvious choices
20* Simple self-correcting heuristics
21
22The representation of access recency is at the core of all LRU
23implementations. In the multi-gen LRU, each generation represents a
24group of pages with similar access recency. Generations establish a
25(time-based) common frame of reference and therefore help make better
26choices, e.g., between different memcgs on a computer or different
27computers in a data center (for job scheduling).
28
29Exploiting spatial locality improves efficiency when gathering the
30accessed bit. A rmap walk targets a single page and does not try to
31profit from discovering a young PTE. A page table walk can sweep all
32the young PTEs in an address space, but the address space can be too
33sparse to make a profit. The key is to optimize both methods and use
34them in combination.
35
36Fast paths reduce code complexity and runtime overhead. Unmapped pages
37do not require TLB flushes; clean pages do not require writeback.
38These facts are only helpful when other conditions, e.g., access
39recency, are similar. With generations as a common frame of reference,
40additional factors stand out. But obvious choices might not be good
41choices; thus self-correction is necessary.
42
43The benefits of simple self-correcting heuristics are self-evident.
44Again, with generations as a common frame of reference, this becomes
45attainable. Specifically, pages in the same generation can be
46categorized based on additional factors, and a feedback loop can
47statistically compare the refault percentages across those categories
48and infer which of them are better choices.
49
50Assumptions
51-----------
52The protection of hot pages and the selection of cold pages are based
53on page access channels and patterns. There are two access channels:
54
55* Accesses through page tables
56* Accesses through file descriptors
57
58The protection of the former channel is by design stronger because:
59
601. The uncertainty in determining the access patterns of the former
61   channel is higher due to the approximation of the accessed bit.
622. The cost of evicting the former channel is higher due to the TLB
63   flushes required and the likelihood of encountering the dirty bit.
643. The penalty of underprotecting the former channel is higher because
65   applications usually do not prepare themselves for major page
66   faults like they do for blocked I/O. E.g., GUI applications
67   commonly use dedicated I/O threads to avoid blocking rendering
68   threads.
69
70There are also two access patterns:
71
72* Accesses exhibiting temporal locality
73* Accesses not exhibiting temporal locality
74
75For the reasons listed above, the former channel is assumed to follow
76the former pattern unless ``VM_SEQ_READ`` or ``VM_RAND_READ`` is
77present, and the latter channel is assumed to follow the latter
78pattern unless outlying refaults have been observed.
79
80Workflow overview
81=================
82Evictable pages are divided into multiple generations for each
83``lruvec``. The youngest generation number is stored in
84``lrugen->max_seq`` for both anon and file types as they are aged on
85an equal footing. The oldest generation numbers are stored in
86``lrugen->min_seq[]`` separately for anon and file types as clean file
87pages can be evicted regardless of swap constraints. These three
88variables are monotonically increasing.
89
90Generation numbers are truncated into ``order_base_2(MAX_NR_GENS+1)``
91bits in order to fit into the gen counter in ``folio->flags``. Each
92truncated generation number is an index to ``lrugen->folios[]``. The
93sliding window technique is used to track at least ``MIN_NR_GENS`` and
94at most ``MAX_NR_GENS`` generations. The gen counter stores a value
95within ``[1, MAX_NR_GENS]`` while a page is on one of
96``lrugen->folios[]``; otherwise it stores zero.
97
98Each generation is divided into multiple tiers. A page accessed ``N``
99times through file descriptors is in tier ``order_base_2(N)``. Unlike
100generations, tiers do not have dedicated ``lrugen->folios[]``. In
101contrast to moving across generations, which requires the LRU lock,
102moving across tiers only involves atomic operations on
103``folio->flags`` and therefore has a negligible cost. A feedback loop
104modeled after the PID controller monitors refaults over all the tiers
105from anon and file types and decides which tiers from which types to
106evict or protect.
107
108There are two conceptually independent procedures: the aging and the
109eviction. They form a closed-loop system, i.e., the page reclaim.
110
111Aging
112-----
113The aging produces young generations. Given an ``lruvec``, it
114increments ``max_seq`` when ``max_seq-min_seq+1`` approaches
115``MIN_NR_GENS``. The aging promotes hot pages to the youngest
116generation when it finds them accessed through page tables; the
117demotion of cold pages happens consequently when it increments
118``max_seq``. The aging uses page table walks and rmap walks to find
119young PTEs. For the former, it iterates ``lruvec_memcg()->mm_list``
120and calls ``walk_page_range()`` with each ``mm_struct`` on this list
121to scan PTEs, and after each iteration, it increments ``max_seq``. For
122the latter, when the eviction walks the rmap and finds a young PTE,
123the aging scans the adjacent PTEs. For both, on finding a young PTE,
124the aging clears the accessed bit and updates the gen counter of the
125page mapped by this PTE to ``(max_seq%MAX_NR_GENS)+1``.
126
127Eviction
128--------
129The eviction consumes old generations. Given an ``lruvec``, it
130increments ``min_seq`` when ``lrugen->folios[]`` indexed by
131``min_seq%MAX_NR_GENS`` becomes empty. To select a type and a tier to
132evict from, it first compares ``min_seq[]`` to select the older type.
133If both types are equally old, it selects the one whose first tier has
134a lower refault percentage. The first tier contains single-use
135unmapped clean pages, which are the best bet. The eviction sorts a
136page according to its gen counter if the aging has found this page
137accessed through page tables and updated its gen counter. It also
138moves a page to the next generation, i.e., ``min_seq+1``, if this page
139was accessed multiple times through file descriptors and the feedback
140loop has detected outlying refaults from the tier this page is in. To
141this end, the feedback loop uses the first tier as the baseline, for
142the reason stated earlier.
143
144Working set protection
145----------------------
146Each generation is timestamped at birth. If ``lru_gen_min_ttl`` is
147set, an ``lruvec`` is protected from the eviction when its oldest
148generation was born within ``lru_gen_min_ttl`` milliseconds. In other
149words, it prevents the working set of ``lru_gen_min_ttl`` milliseconds
150from getting evicted. The OOM killer is triggered if this working set
151cannot be kept in memory.
152
153This time-based approach has the following advantages:
154
1551. It is easier to configure because it is agnostic to applications
156   and memory sizes.
1572. It is more reliable because it is directly wired to the OOM killer.
158
159Rmap/PT walk feedback
160---------------------
161Searching the rmap for PTEs mapping each page on an LRU list (to test
162and clear the accessed bit) can be expensive because pages from
163different VMAs (PA space) are not cache friendly to the rmap (VA
164space). For workloads mostly using mapped pages, searching the rmap
165can incur the highest CPU cost in the reclaim path.
166
167``lru_gen_look_around()`` exploits spatial locality to reduce the
168trips into the rmap. It scans the adjacent PTEs of a young PTE and
169promotes hot pages. If the scan was done cacheline efficiently, it
170adds the PMD entry pointing to the PTE table to the Bloom filter. This
171forms a feedback loop between the eviction and the aging.
172
173Bloom Filters
174-------------
175Bloom filters are a space and memory efficient data structure for set
176membership test, i.e., test if an element is not in the set or may be
177in the set.
178
179In the eviction path, specifically, in ``lru_gen_look_around()``, if a
180PMD has a sufficient number of hot pages, its address is placed in the
181filter. In the aging path, set membership means that the PTE range
182will be scanned for young pages.
183
184Note that Bloom filters are probabilistic on set membership. If a test
185is false positive, the cost is an additional scan of a range of PTEs,
186which may yield hot pages anyway. Parameters of the filter itself can
187control the false positive rate in the limit.
188
189Memcg LRU
190---------
191An memcg LRU is a per-node LRU of memcgs. It is also an LRU of LRUs,
192since each node and memcg combination has an LRU of folios (see
193``mem_cgroup_lruvec()``). Its goal is to improve the scalability of
194global reclaim, which is critical to system-wide memory overcommit in
195data centers. Note that memcg LRU only applies to global reclaim.
196
197The basic structure of an memcg LRU can be understood by an analogy to
198the active/inactive LRU (of folios):
199
2001. It has the young and the old (generations), i.e., the counterparts
201   to the active and the inactive;
2022. The increment of ``max_seq`` triggers promotion, i.e., the
203   counterpart to activation;
2043. Other events trigger similar operations, e.g., offlining an memcg
205   triggers demotion, i.e., the counterpart to deactivation.
206
207In terms of global reclaim, it has two distinct features:
208
2091. Sharding, which allows each thread to start at a random memcg (in
210   the old generation) and improves parallelism;
2112. Eventual fairness, which allows direct reclaim to bail out at will
212   and reduces latency without affecting fairness over some time.
213
214In terms of traversing memcgs during global reclaim, it improves the
215best-case complexity from O(n) to O(1) and does not affect the
216worst-case complexity O(n). Therefore, on average, it has a sublinear
217complexity.
218
219Summary
220-------
221The multi-gen LRU (of folios) can be disassembled into the following
222parts:
223
224* Generations
225* Rmap walks
226* Page table walks
227* Bloom filters
228* PID controller
229
230The aging and the eviction form a producer-consumer model;
231specifically, the latter drives the former by the sliding window over
232generations. Within the aging, rmap walks drive page table walks by
233inserting hot densely populated page tables to the Bloom filters.
234Within the eviction, the PID controller uses refaults as the feedback
235to select types to evict and tiers to protect.
236