Eric Lee / smarc-ti-linux-kernel

Download as
Browse Code ยป

Commit 92e793495597af4135d94314113bf13eafb0e663

Authored by Glauber Costa
Committed by Linus Torvalds
1 parent 107dab5c92
Exists in master and in 20 other branches dlt-processor-sdk-linux-03.00.00.04, newt-ti-linux-3.12.y, processor-sdk-linux-01.00.00, processor-sdk-linux-02.00.01, smarc-ti-linux-3.12.10, smarc-ti-linux-3.12.y, smarc-ti-linux-3.14.y, smarc-ti-linux-3.15.y, smarc-ti-lsk-linux-4.1.y, smarct3x-processor-sdk-04.01.00.06, smarct3x-processor-sdk-linux-02.00.01, smarct3x-processor-sdk-linux-03.00.00.04, smarct4x-800-processor-sdk-linux-02.00.01, smarct4x-processor-sdk-04.01.00.06, smarct4x-processor-sdk-linux-02.00.01, smarct4x-processor-sdk-linux-03.00.00.04, ti-linux-3.12.y, ti-linux-3.14.y, ti-linux-3.15.y, ti-lsk-linux-4.1.y

kmem: add slab-specific documentation about the kmem controller 

Signed-off-by: Glauber Costa <glommer@parallels.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Frederic Weisbecker <fweisbec@redhat.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: JoonSoo Kim <js1304@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Rik van Riel <riel@redhat.com>
Cc: Suleiman Souhlal <suleiman@google.com>
Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Showing 1 changed file with 7 additions and 0 deletions Inline Diff

Documentation/cgroups/memory.txt
Diff comments   View file @ 92e7934
1 Memory Resource Controller 1 Memory Resource Controller
2 2
3 NOTE: The Memory Resource Controller has generically been referred to as the 3 NOTE: The Memory Resource Controller has generically been referred to as the
4 memory controller in this document. Do not confuse memory controller 4 memory controller in this document. Do not confuse memory controller
5 used here with the memory controller that is used in hardware. 5 used here with the memory controller that is used in hardware.
6 6
7 (For editors) 7 (For editors)
8 In this document: 8 In this document:
9 When we mention a cgroup (cgroupfs's directory) with memory controller, 9 When we mention a cgroup (cgroupfs's directory) with memory controller,
10 we call it "memory cgroup". When you see git-log and source code, you'll 10 we call it "memory cgroup". When you see git-log and source code, you'll
11 see patch's title and function names tend to use "memcg". 11 see patch's title and function names tend to use "memcg".
12 In this document, we avoid using it. 12 In this document, we avoid using it.
13 13
14 Benefits and Purpose of the memory controller 14 Benefits and Purpose of the memory controller
15 15
16 The memory controller isolates the memory behaviour of a group of tasks 16 The memory controller isolates the memory behaviour of a group of tasks
17 from the rest of the system. The article on LWN [12] mentions some probable 17 from the rest of the system. The article on LWN [12] mentions some probable
18 uses of the memory controller. The memory controller can be used to 18 uses of the memory controller. The memory controller can be used to
19 19
20 a. Isolate an application or a group of applications 20 a. Isolate an application or a group of applications
21 Memory-hungry applications can be isolated and limited to a smaller 21 Memory-hungry applications can be isolated and limited to a smaller
22 amount of memory. 22 amount of memory.
23 b. Create a cgroup with a limited amount of memory; this can be used 23 b. Create a cgroup with a limited amount of memory; this can be used
24 as a good alternative to booting with mem=XXXX. 24 as a good alternative to booting with mem=XXXX.
25 c. Virtualization solutions can control the amount of memory they want 25 c. Virtualization solutions can control the amount of memory they want
26 to assign to a virtual machine instance. 26 to assign to a virtual machine instance.
27 d. A CD/DVD burner could control the amount of memory used by the 27 d. A CD/DVD burner could control the amount of memory used by the
28 rest of the system to ensure that burning does not fail due to lack 28 rest of the system to ensure that burning does not fail due to lack
29 of available memory. 29 of available memory.
30 e. There are several other use cases; find one or use the controller just 30 e. There are several other use cases; find one or use the controller just
31 for fun (to learn and hack on the VM subsystem). 31 for fun (to learn and hack on the VM subsystem).
32 32
33 Current Status: linux-2.6.34-mmotm(development version of 2010/April) 33 Current Status: linux-2.6.34-mmotm(development version of 2010/April)
34 34
35 Features: 35 Features:
36 - accounting anonymous pages, file caches, swap caches usage and limiting them. 36 - accounting anonymous pages, file caches, swap caches usage and limiting them.
37 - pages are linked to per-memcg LRU exclusively, and there is no global LRU. 37 - pages are linked to per-memcg LRU exclusively, and there is no global LRU.
38 - optionally, memory+swap usage can be accounted and limited. 38 - optionally, memory+swap usage can be accounted and limited.
39 - hierarchical accounting 39 - hierarchical accounting
40 - soft limit 40 - soft limit
41 - moving (recharging) account at moving a task is selectable. 41 - moving (recharging) account at moving a task is selectable.
42 - usage threshold notifier 42 - usage threshold notifier
43 - oom-killer disable knob and oom-notifier 43 - oom-killer disable knob and oom-notifier
44 - Root cgroup has no limit controls. 44 - Root cgroup has no limit controls.
45 45
46 Kernel memory support is a work in progress, and the current version provides 46 Kernel memory support is a work in progress, and the current version provides
47 basically functionality. (See Section 2.7) 47 basically functionality. (See Section 2.7)
48 48
49 Brief summary of control files. 49 Brief summary of control files.
50 50
51 tasks # attach a task(thread) and show list of threads 51 tasks # attach a task(thread) and show list of threads
52 cgroup.procs # show list of processes 52 cgroup.procs # show list of processes
53 cgroup.event_control # an interface for event_fd() 53 cgroup.event_control # an interface for event_fd()
54 memory.usage_in_bytes # show current res_counter usage for memory 54 memory.usage_in_bytes # show current res_counter usage for memory
55 (See 5.5 for details) 55 (See 5.5 for details)
56 memory.memsw.usage_in_bytes # show current res_counter usage for memory+Swap 56 memory.memsw.usage_in_bytes # show current res_counter usage for memory+Swap
57 (See 5.5 for details) 57 (See 5.5 for details)
58 memory.limit_in_bytes # set/show limit of memory usage 58 memory.limit_in_bytes # set/show limit of memory usage
59 memory.memsw.limit_in_bytes # set/show limit of memory+Swap usage 59 memory.memsw.limit_in_bytes # set/show limit of memory+Swap usage
60 memory.failcnt # show the number of memory usage hits limits 60 memory.failcnt # show the number of memory usage hits limits
61 memory.memsw.failcnt # show the number of memory+Swap hits limits 61 memory.memsw.failcnt # show the number of memory+Swap hits limits
62 memory.max_usage_in_bytes # show max memory usage recorded 62 memory.max_usage_in_bytes # show max memory usage recorded
63 memory.memsw.max_usage_in_bytes # show max memory+Swap usage recorded 63 memory.memsw.max_usage_in_bytes # show max memory+Swap usage recorded
64 memory.soft_limit_in_bytes # set/show soft limit of memory usage 64 memory.soft_limit_in_bytes # set/show soft limit of memory usage
65 memory.stat # show various statistics 65 memory.stat # show various statistics
66 memory.use_hierarchy # set/show hierarchical account enabled 66 memory.use_hierarchy # set/show hierarchical account enabled
67 memory.force_empty # trigger forced move charge to parent 67 memory.force_empty # trigger forced move charge to parent
68 memory.swappiness # set/show swappiness parameter of vmscan 68 memory.swappiness # set/show swappiness parameter of vmscan
69 (See sysctl's vm.swappiness) 69 (See sysctl's vm.swappiness)
70 memory.move_charge_at_immigrate # set/show controls of moving charges 70 memory.move_charge_at_immigrate # set/show controls of moving charges
71 memory.oom_control # set/show oom controls. 71 memory.oom_control # set/show oom controls.
72 memory.numa_stat # show the number of memory usage per numa node 72 memory.numa_stat # show the number of memory usage per numa node
73 73
74 memory.kmem.limit_in_bytes # set/show hard limit for kernel memory 74 memory.kmem.limit_in_bytes # set/show hard limit for kernel memory
75 memory.kmem.usage_in_bytes # show current kernel memory allocation 75 memory.kmem.usage_in_bytes # show current kernel memory allocation
76 memory.kmem.failcnt # show the number of kernel memory usage hits limits 76 memory.kmem.failcnt # show the number of kernel memory usage hits limits
77 memory.kmem.max_usage_in_bytes # show max kernel memory usage recorded 77 memory.kmem.max_usage_in_bytes # show max kernel memory usage recorded
78 78
79 memory.kmem.tcp.limit_in_bytes # set/show hard limit for tcp buf memory 79 memory.kmem.tcp.limit_in_bytes # set/show hard limit for tcp buf memory
80 memory.kmem.tcp.usage_in_bytes # show current tcp buf memory allocation 80 memory.kmem.tcp.usage_in_bytes # show current tcp buf memory allocation
81 memory.kmem.tcp.failcnt # show the number of tcp buf memory usage hits limits 81 memory.kmem.tcp.failcnt # show the number of tcp buf memory usage hits limits
82 memory.kmem.tcp.max_usage_in_bytes # show max tcp buf memory usage recorded 82 memory.kmem.tcp.max_usage_in_bytes # show max tcp buf memory usage recorded
83 83
84 1. History 84 1. History
85 85
86 The memory controller has a long history. A request for comments for the memory 86 The memory controller has a long history. A request for comments for the memory
87 controller was posted by Balbir Singh [1]. At the time the RFC was posted 87 controller was posted by Balbir Singh [1]. At the time the RFC was posted
88 there were several implementations for memory control. The goal of the 88 there were several implementations for memory control. The goal of the
89 RFC was to build consensus and agreement for the minimal features required 89 RFC was to build consensus and agreement for the minimal features required
90 for memory control. The first RSS controller was posted by Balbir Singh[2] 90 for memory control. The first RSS controller was posted by Balbir Singh[2]
91 in Feb 2007. Pavel Emelianov [3][4][5] has since posted three versions of the 91 in Feb 2007. Pavel Emelianov [3][4][5] has since posted three versions of the
92 RSS controller. At OLS, at the resource management BoF, everyone suggested 92 RSS controller. At OLS, at the resource management BoF, everyone suggested
93 that we handle both page cache and RSS together. Another request was raised 93 that we handle both page cache and RSS together. Another request was raised
94 to allow user space handling of OOM. The current memory controller is 94 to allow user space handling of OOM. The current memory controller is
95 at version 6; it combines both mapped (RSS) and unmapped Page 95 at version 6; it combines both mapped (RSS) and unmapped Page
96 Cache Control [11]. 96 Cache Control [11].
97 97
98 2. Memory Control 98 2. Memory Control
99 99
100 Memory is a unique resource in the sense that it is present in a limited 100 Memory is a unique resource in the sense that it is present in a limited
101 amount. If a task requires a lot of CPU processing, the task can spread 101 amount. If a task requires a lot of CPU processing, the task can spread
102 its processing over a period of hours, days, months or years, but with 102 its processing over a period of hours, days, months or years, but with
103 memory, the same physical memory needs to be reused to accomplish the task. 103 memory, the same physical memory needs to be reused to accomplish the task.
104 104
105 The memory controller implementation has been divided into phases. These 105 The memory controller implementation has been divided into phases. These
106 are: 106 are:
107 107
108 1. Memory controller 108 1. Memory controller
109 2. mlock(2) controller 109 2. mlock(2) controller
110 3. Kernel user memory accounting and slab control 110 3. Kernel user memory accounting and slab control
111 4. user mappings length controller 111 4. user mappings length controller
112 112
113 The memory controller is the first controller developed. 113 The memory controller is the first controller developed.
114 114
115 2.1. Design 115 2.1. Design
116 116
117 The core of the design is a counter called the res_counter. The res_counter 117 The core of the design is a counter called the res_counter. The res_counter
118 tracks the current memory usage and limit of the group of processes associated 118 tracks the current memory usage and limit of the group of processes associated
119 with the controller. Each cgroup has a memory controller specific data 119 with the controller. Each cgroup has a memory controller specific data
120 structure (mem_cgroup) associated with it. 120 structure (mem_cgroup) associated with it.
121 121
122 2.2. Accounting 122 2.2. Accounting
123 123
124 +--------------------+ 124 +--------------------+
125 | mem_cgroup | 125 | mem_cgroup |
126 | (res_counter) | 126 | (res_counter) |
127 +--------------------+ 127 +--------------------+
128 / ^ \ 128 / ^ \
129 / | \ 129 / | \
130 +---------------+ | +---------------+ 130 +---------------+ | +---------------+
131 | mm_struct | |.... | mm_struct | 131 | mm_struct | |.... | mm_struct |
132 | | | | | 132 | | | | |
133 +---------------+ | +---------------+ 133 +---------------+ | +---------------+
134 | 134 |
135 + --------------+ 135 + --------------+
136 | 136 |
137 +---------------+ +------+--------+ 137 +---------------+ +------+--------+
138 | page +----------> page_cgroup| 138 | page +----------> page_cgroup|
139 | | | | 139 | | | |
140 +---------------+ +---------------+ 140 +---------------+ +---------------+
141 141
142 (Figure 1: Hierarchy of Accounting) 142 (Figure 1: Hierarchy of Accounting)
143 143
144 144
145 Figure 1 shows the important aspects of the controller 145 Figure 1 shows the important aspects of the controller
146 146
147 1. Accounting happens per cgroup 147 1. Accounting happens per cgroup
148 2. Each mm_struct knows about which cgroup it belongs to 148 2. Each mm_struct knows about which cgroup it belongs to
149 3. Each page has a pointer to the page_cgroup, which in turn knows the 149 3. Each page has a pointer to the page_cgroup, which in turn knows the
150 cgroup it belongs to 150 cgroup it belongs to
151 151
152 The accounting is done as follows: mem_cgroup_charge_common() is invoked to 152 The accounting is done as follows: mem_cgroup_charge_common() is invoked to
153 set up the necessary data structures and check if the cgroup that is being 153 set up the necessary data structures and check if the cgroup that is being
154 charged is over its limit. If it is, then reclaim is invoked on the cgroup. 154 charged is over its limit. If it is, then reclaim is invoked on the cgroup.
155 More details can be found in the reclaim section of this document. 155 More details can be found in the reclaim section of this document.
156 If everything goes well, a page meta-data-structure called page_cgroup is 156 If everything goes well, a page meta-data-structure called page_cgroup is
157 updated. page_cgroup has its own LRU on cgroup. 157 updated. page_cgroup has its own LRU on cgroup.
158 (*) page_cgroup structure is allocated at boot/memory-hotplug time. 158 (*) page_cgroup structure is allocated at boot/memory-hotplug time.
159 159
160 2.2.1 Accounting details 160 2.2.1 Accounting details
161 161
162 All mapped anon pages (RSS) and cache pages (Page Cache) are accounted. 162 All mapped anon pages (RSS) and cache pages (Page Cache) are accounted.
163 Some pages which are never reclaimable and will not be on the LRU 163 Some pages which are never reclaimable and will not be on the LRU
164 are not accounted. We just account pages under usual VM management. 164 are not accounted. We just account pages under usual VM management.
165 165
166 RSS pages are accounted at page_fault unless they've already been accounted 166 RSS pages are accounted at page_fault unless they've already been accounted
167 for earlier. A file page will be accounted for as Page Cache when it's 167 for earlier. A file page will be accounted for as Page Cache when it's
168 inserted into inode (radix-tree). While it's mapped into the page tables of 168 inserted into inode (radix-tree). While it's mapped into the page tables of
169 processes, duplicate accounting is carefully avoided. 169 processes, duplicate accounting is carefully avoided.
170 170
171 An RSS page is unaccounted when it's fully unmapped. A PageCache page is 171 An RSS page is unaccounted when it's fully unmapped. A PageCache page is
172 unaccounted when it's removed from radix-tree. Even if RSS pages are fully 172 unaccounted when it's removed from radix-tree. Even if RSS pages are fully
173 unmapped (by kswapd), they may exist as SwapCache in the system until they 173 unmapped (by kswapd), they may exist as SwapCache in the system until they
174 are really freed. Such SwapCaches are also accounted. 174 are really freed. Such SwapCaches are also accounted.
175 A swapped-in page is not accounted until it's mapped. 175 A swapped-in page is not accounted until it's mapped.
176 176
177 Note: The kernel does swapin-readahead and reads multiple swaps at once. 177 Note: The kernel does swapin-readahead and reads multiple swaps at once.
178 This means swapped-in pages may contain pages for other tasks than a task 178 This means swapped-in pages may contain pages for other tasks than a task
179 causing page fault. So, we avoid accounting at swap-in I/O. 179 causing page fault. So, we avoid accounting at swap-in I/O.
180 180
181 At page migration, accounting information is kept. 181 At page migration, accounting information is kept.
182 182
183 Note: we just account pages-on-LRU because our purpose is to control amount 183 Note: we just account pages-on-LRU because our purpose is to control amount
184 of used pages; not-on-LRU pages tend to be out-of-control from VM view. 184 of used pages; not-on-LRU pages tend to be out-of-control from VM view.
185 185
186 2.3 Shared Page Accounting 186 2.3 Shared Page Accounting
187 187
188 Shared pages are accounted on the basis of the first touch approach. The 188 Shared pages are accounted on the basis of the first touch approach. The
189 cgroup that first touches a page is accounted for the page. The principle 189 cgroup that first touches a page is accounted for the page. The principle
190 behind this approach is that a cgroup that aggressively uses a shared 190 behind this approach is that a cgroup that aggressively uses a shared
191 page will eventually get charged for it (once it is uncharged from 191 page will eventually get charged for it (once it is uncharged from
192 the cgroup that brought it in -- this will happen on memory pressure). 192 the cgroup that brought it in -- this will happen on memory pressure).
193 193
194 But see section 8.2: when moving a task to another cgroup, its pages may 194 But see section 8.2: when moving a task to another cgroup, its pages may
195 be recharged to the new cgroup, if move_charge_at_immigrate has been chosen. 195 be recharged to the new cgroup, if move_charge_at_immigrate has been chosen.
196 196
197 Exception: If CONFIG_CGROUP_CGROUP_MEMCG_SWAP is not used. 197 Exception: If CONFIG_CGROUP_CGROUP_MEMCG_SWAP is not used.
198 When you do swapoff and make swapped-out pages of shmem(tmpfs) to 198 When you do swapoff and make swapped-out pages of shmem(tmpfs) to
199 be backed into memory in force, charges for pages are accounted against the 199 be backed into memory in force, charges for pages are accounted against the
200 caller of swapoff rather than the users of shmem. 200 caller of swapoff rather than the users of shmem.
201 201
202 2.4 Swap Extension (CONFIG_MEMCG_SWAP) 202 2.4 Swap Extension (CONFIG_MEMCG_SWAP)
203 203
204 Swap Extension allows you to record charge for swap. A swapped-in page is 204 Swap Extension allows you to record charge for swap. A swapped-in page is
205 charged back to original page allocator if possible. 205 charged back to original page allocator if possible.
206 206
207 When swap is accounted, following files are added. 207 When swap is accounted, following files are added.
208 - memory.memsw.usage_in_bytes. 208 - memory.memsw.usage_in_bytes.
209 - memory.memsw.limit_in_bytes. 209 - memory.memsw.limit_in_bytes.
210 210
211 memsw means memory+swap. Usage of memory+swap is limited by 211 memsw means memory+swap. Usage of memory+swap is limited by
212 memsw.limit_in_bytes. 212 memsw.limit_in_bytes.
213 213
214 Example: Assume a system with 4G of swap. A task which allocates 6G of memory 214 Example: Assume a system with 4G of swap. A task which allocates 6G of memory
215 (by mistake) under 2G memory limitation will use all swap. 215 (by mistake) under 2G memory limitation will use all swap.
216 In this case, setting memsw.limit_in_bytes=3G will prevent bad use of swap. 216 In this case, setting memsw.limit_in_bytes=3G will prevent bad use of swap.
217 By using the memsw limit, you can avoid system OOM which can be caused by swap 217 By using the memsw limit, you can avoid system OOM which can be caused by swap
218 shortage. 218 shortage.
219 219
220 * why 'memory+swap' rather than swap. 220 * why 'memory+swap' rather than swap.
221 The global LRU(kswapd) can swap out arbitrary pages. Swap-out means 221 The global LRU(kswapd) can swap out arbitrary pages. Swap-out means
222 to move account from memory to swap...there is no change in usage of 222 to move account from memory to swap...there is no change in usage of
223 memory+swap. In other words, when we want to limit the usage of swap without 223 memory+swap. In other words, when we want to limit the usage of swap without
224 affecting global LRU, memory+swap limit is better than just limiting swap from 224 affecting global LRU, memory+swap limit is better than just limiting swap from
225 an OS point of view. 225 an OS point of view.
226 226
227 * What happens when a cgroup hits memory.memsw.limit_in_bytes 227 * What happens when a cgroup hits memory.memsw.limit_in_bytes
228 When a cgroup hits memory.memsw.limit_in_bytes, it's useless to do swap-out 228 When a cgroup hits memory.memsw.limit_in_bytes, it's useless to do swap-out
229 in this cgroup. Then, swap-out will not be done by cgroup routine and file 229 in this cgroup. Then, swap-out will not be done by cgroup routine and file
230 caches are dropped. But as mentioned above, global LRU can do swapout memory 230 caches are dropped. But as mentioned above, global LRU can do swapout memory
231 from it for sanity of the system's memory management state. You can't forbid 231 from it for sanity of the system's memory management state. You can't forbid
232 it by cgroup. 232 it by cgroup.
233 233
234 2.5 Reclaim 234 2.5 Reclaim
235 235
236 Each cgroup maintains a per cgroup LRU which has the same structure as 236 Each cgroup maintains a per cgroup LRU which has the same structure as
237 global VM. When a cgroup goes over its limit, we first try 237 global VM. When a cgroup goes over its limit, we first try
238 to reclaim memory from the cgroup so as to make space for the new 238 to reclaim memory from the cgroup so as to make space for the new
239 pages that the cgroup has touched. If the reclaim is unsuccessful, 239 pages that the cgroup has touched. If the reclaim is unsuccessful,
240 an OOM routine is invoked to select and kill the bulkiest task in the 240 an OOM routine is invoked to select and kill the bulkiest task in the
241 cgroup. (See 10. OOM Control below.) 241 cgroup. (See 10. OOM Control below.)
242 242
243 The reclaim algorithm has not been modified for cgroups, except that 243 The reclaim algorithm has not been modified for cgroups, except that
244 pages that are selected for reclaiming come from the per-cgroup LRU 244 pages that are selected for reclaiming come from the per-cgroup LRU
245 list. 245 list.
246 246
247 NOTE: Reclaim does not work for the root cgroup, since we cannot set any 247 NOTE: Reclaim does not work for the root cgroup, since we cannot set any
248 limits on the root cgroup. 248 limits on the root cgroup.
249 249
250 Note2: When panic_on_oom is set to "2", the whole system will panic. 250 Note2: When panic_on_oom is set to "2", the whole system will panic.
251 251
252 When oom event notifier is registered, event will be delivered. 252 When oom event notifier is registered, event will be delivered.
253 (See oom_control section) 253 (See oom_control section)
254 254
255 2.6 Locking 255 2.6 Locking
256 256
257 lock_page_cgroup()/unlock_page_cgroup() should not be called under 257 lock_page_cgroup()/unlock_page_cgroup() should not be called under
258 mapping->tree_lock. 258 mapping->tree_lock.
259 259
260 Other lock order is following: 260 Other lock order is following:
261 PG_locked. 261 PG_locked.
262 mm->page_table_lock 262 mm->page_table_lock
263 zone->lru_lock 263 zone->lru_lock
264 lock_page_cgroup. 264 lock_page_cgroup.
265 In many cases, just lock_page_cgroup() is called. 265 In many cases, just lock_page_cgroup() is called.
266 per-zone-per-cgroup LRU (cgroup's private LRU) is just guarded by 266 per-zone-per-cgroup LRU (cgroup's private LRU) is just guarded by
267 zone->lru_lock, it has no lock of its own. 267 zone->lru_lock, it has no lock of its own.
268 268
269 2.7 Kernel Memory Extension (CONFIG_MEMCG_KMEM) 269 2.7 Kernel Memory Extension (CONFIG_MEMCG_KMEM)
270 270
271 With the Kernel memory extension, the Memory Controller is able to limit 271 With the Kernel memory extension, the Memory Controller is able to limit
272 the amount of kernel memory used by the system. Kernel memory is fundamentally 272 the amount of kernel memory used by the system. Kernel memory is fundamentally
273 different than user memory, since it can't be swapped out, which makes it 273 different than user memory, since it can't be swapped out, which makes it
274 possible to DoS the system by consuming too much of this precious resource. 274 possible to DoS the system by consuming too much of this precious resource.
275 275
276 Kernel memory won't be accounted at all until limit on a group is set. This 276 Kernel memory won't be accounted at all until limit on a group is set. This
277 allows for existing setups to continue working without disruption. The limit 277 allows for existing setups to continue working without disruption. The limit
278 cannot be set if the cgroup have children, or if there are already tasks in the 278 cannot be set if the cgroup have children, or if there are already tasks in the
279 cgroup. Attempting to set the limit under those conditions will return -EBUSY. 279 cgroup. Attempting to set the limit under those conditions will return -EBUSY.
280 When use_hierarchy == 1 and a group is accounted, its children will 280 When use_hierarchy == 1 and a group is accounted, its children will
281 automatically be accounted regardless of their limit value. 281 automatically be accounted regardless of their limit value.
282 282
283 After a group is first limited, it will be kept being accounted until it 283 After a group is first limited, it will be kept being accounted until it
284 is removed. The memory limitation itself, can of course be removed by writing 284 is removed. The memory limitation itself, can of course be removed by writing
285 -1 to memory.kmem.limit_in_bytes. In this case, kmem will be accounted, but not 285 -1 to memory.kmem.limit_in_bytes. In this case, kmem will be accounted, but not
286 limited. 286 limited.
287 287
288 Kernel memory limits are not imposed for the root cgroup. Usage for the root 288 Kernel memory limits are not imposed for the root cgroup. Usage for the root
289 cgroup may or may not be accounted. The memory used is accumulated into 289 cgroup may or may not be accounted. The memory used is accumulated into
290 memory.kmem.usage_in_bytes, or in a separate counter when it makes sense. 290 memory.kmem.usage_in_bytes, or in a separate counter when it makes sense.
291 (currently only for tcp). 291 (currently only for tcp).
292 The main "kmem" counter is fed into the main counter, so kmem charges will 292 The main "kmem" counter is fed into the main counter, so kmem charges will
293 also be visible from the user counter. 293 also be visible from the user counter.
294 294
295 Currently no soft limit is implemented for kernel memory. It is future work 295 Currently no soft limit is implemented for kernel memory. It is future work
296 to trigger slab reclaim when those limits are reached. 296 to trigger slab reclaim when those limits are reached.
297 297
298 2.7.1 Current Kernel Memory resources accounted 298 2.7.1 Current Kernel Memory resources accounted
299 299
300 * stack pages: every process consumes some stack pages. By accounting into 300 * stack pages: every process consumes some stack pages. By accounting into
301 kernel memory, we prevent new processes from being created when the kernel 301 kernel memory, we prevent new processes from being created when the kernel
302 memory usage is too high. 302 memory usage is too high.
303 303
304 * slab pages: pages allocated by the SLAB or SLUB allocator are tracked. A copy
305 of each kmem_cache is created everytime the cache is touched by the first time
306 from inside the memcg. The creation is done lazily, so some objects can still be
307 skipped while the cache is being created. All objects in a slab page should
308 belong to the same memcg. This only fails to hold when a task is migrated to a
309 different memcg during the page allocation by the cache.
310
304 * sockets memory pressure: some sockets protocols have memory pressure 311 * sockets memory pressure: some sockets protocols have memory pressure
305 thresholds. The Memory Controller allows them to be controlled individually 312 thresholds. The Memory Controller allows them to be controlled individually
306 per cgroup, instead of globally. 313 per cgroup, instead of globally.
307 314
308 * tcp memory pressure: sockets memory pressure for the tcp protocol. 315 * tcp memory pressure: sockets memory pressure for the tcp protocol.
309 316
310 2.7.3 Common use cases 317 2.7.3 Common use cases
311 318
312 Because the "kmem" counter is fed to the main user counter, kernel memory can 319 Because the "kmem" counter is fed to the main user counter, kernel memory can
313 never be limited completely independently of user memory. Say "U" is the user 320 never be limited completely independently of user memory. Say "U" is the user
314 limit, and "K" the kernel limit. There are three possible ways limits can be 321 limit, and "K" the kernel limit. There are three possible ways limits can be
315 set: 322 set:
316 323
317 U != 0, K = unlimited: 324 U != 0, K = unlimited:
318 This is the standard memcg limitation mechanism already present before kmem 325 This is the standard memcg limitation mechanism already present before kmem
319 accounting. Kernel memory is completely ignored. 326 accounting. Kernel memory is completely ignored.
320 327
321 U != 0, K < U: 328 U != 0, K < U:
322 Kernel memory is a subset of the user memory. This setup is useful in 329 Kernel memory is a subset of the user memory. This setup is useful in
323 deployments where the total amount of memory per-cgroup is overcommited. 330 deployments where the total amount of memory per-cgroup is overcommited.
324 Overcommiting kernel memory limits is definitely not recommended, since the 331 Overcommiting kernel memory limits is definitely not recommended, since the
325 box can still run out of non-reclaimable memory. 332 box can still run out of non-reclaimable memory.
326 In this case, the admin could set up K so that the sum of all groups is 333 In this case, the admin could set up K so that the sum of all groups is
327 never greater than the total memory, and freely set U at the cost of his 334 never greater than the total memory, and freely set U at the cost of his
328 QoS. 335 QoS.
329 336
330 U != 0, K >= U: 337 U != 0, K >= U:
331 Since kmem charges will also be fed to the user counter and reclaim will be 338 Since kmem charges will also be fed to the user counter and reclaim will be
332 triggered for the cgroup for both kinds of memory. This setup gives the 339 triggered for the cgroup for both kinds of memory. This setup gives the
333 admin a unified view of memory, and it is also useful for people who just 340 admin a unified view of memory, and it is also useful for people who just
334 want to track kernel memory usage. 341 want to track kernel memory usage.
335 342
336 3. User Interface 343 3. User Interface
337 344
338 0. Configuration 345 0. Configuration
339 346
340 a. Enable CONFIG_CGROUPS 347 a. Enable CONFIG_CGROUPS
341 b. Enable CONFIG_RESOURCE_COUNTERS 348 b. Enable CONFIG_RESOURCE_COUNTERS
342 c. Enable CONFIG_MEMCG 349 c. Enable CONFIG_MEMCG
343 d. Enable CONFIG_MEMCG_SWAP (to use swap extension) 350 d. Enable CONFIG_MEMCG_SWAP (to use swap extension)
344 d. Enable CONFIG_MEMCG_KMEM (to use kmem extension) 351 d. Enable CONFIG_MEMCG_KMEM (to use kmem extension)
345 352
346 1. Prepare the cgroups (see cgroups.txt, Why are cgroups needed?) 353 1. Prepare the cgroups (see cgroups.txt, Why are cgroups needed?)
347 # mount -t tmpfs none /sys/fs/cgroup 354 # mount -t tmpfs none /sys/fs/cgroup
348 # mkdir /sys/fs/cgroup/memory 355 # mkdir /sys/fs/cgroup/memory
349 # mount -t cgroup none /sys/fs/cgroup/memory -o memory 356 # mount -t cgroup none /sys/fs/cgroup/memory -o memory
350 357
351 2. Make the new group and move bash into it 358 2. Make the new group and move bash into it
352 # mkdir /sys/fs/cgroup/memory/0 359 # mkdir /sys/fs/cgroup/memory/0
353 # echo $$ > /sys/fs/cgroup/memory/0/tasks 360 # echo $$ > /sys/fs/cgroup/memory/0/tasks
354 361
355 Since now we're in the 0 cgroup, we can alter the memory limit: 362 Since now we're in the 0 cgroup, we can alter the memory limit:
356 # echo 4M > /sys/fs/cgroup/memory/0/memory.limit_in_bytes 363 # echo 4M > /sys/fs/cgroup/memory/0/memory.limit_in_bytes
357 364
358 NOTE: We can use a suffix (k, K, m, M, g or G) to indicate values in kilo, 365 NOTE: We can use a suffix (k, K, m, M, g or G) to indicate values in kilo,
359 mega or gigabytes. (Here, Kilo, Mega, Giga are Kibibytes, Mebibytes, Gibibytes.) 366 mega or gigabytes. (Here, Kilo, Mega, Giga are Kibibytes, Mebibytes, Gibibytes.)
360 367
361 NOTE: We can write "-1" to reset the *.limit_in_bytes(unlimited). 368 NOTE: We can write "-1" to reset the *.limit_in_bytes(unlimited).
362 NOTE: We cannot set limits on the root cgroup any more. 369 NOTE: We cannot set limits on the root cgroup any more.
363 370
364 # cat /sys/fs/cgroup/memory/0/memory.limit_in_bytes 371 # cat /sys/fs/cgroup/memory/0/memory.limit_in_bytes
365 4194304 372 4194304
366 373
367 We can check the usage: 374 We can check the usage:
368 # cat /sys/fs/cgroup/memory/0/memory.usage_in_bytes 375 # cat /sys/fs/cgroup/memory/0/memory.usage_in_bytes
369 1216512 376 1216512
370 377
371 A successful write to this file does not guarantee a successful setting of 378 A successful write to this file does not guarantee a successful setting of
372 this limit to the value written into the file. This can be due to a 379 this limit to the value written into the file. This can be due to a
373 number of factors, such as rounding up to page boundaries or the total 380 number of factors, such as rounding up to page boundaries or the total
374 availability of memory on the system. The user is required to re-read 381 availability of memory on the system. The user is required to re-read
375 this file after a write to guarantee the value committed by the kernel. 382 this file after a write to guarantee the value committed by the kernel.
376 383
377 # echo 1 > memory.limit_in_bytes 384 # echo 1 > memory.limit_in_bytes
378 # cat memory.limit_in_bytes 385 # cat memory.limit_in_bytes
379 4096 386 4096
380 387
381 The memory.failcnt field gives the number of times that the cgroup limit was 388 The memory.failcnt field gives the number of times that the cgroup limit was
382 exceeded. 389 exceeded.
383 390
384 The memory.stat file gives accounting information. Now, the number of 391 The memory.stat file gives accounting information. Now, the number of
385 caches, RSS and Active pages/Inactive pages are shown. 392 caches, RSS and Active pages/Inactive pages are shown.
386 393
387 4. Testing 394 4. Testing
388 395
389 For testing features and implementation, see memcg_test.txt. 396 For testing features and implementation, see memcg_test.txt.
390 397
391 Performance test is also important. To see pure memory controller's overhead, 398 Performance test is also important. To see pure memory controller's overhead,
392 testing on tmpfs will give you good numbers of small overheads. 399 testing on tmpfs will give you good numbers of small overheads.
393 Example: do kernel make on tmpfs. 400 Example: do kernel make on tmpfs.
394 401
395 Page-fault scalability is also important. At measuring parallel 402 Page-fault scalability is also important. At measuring parallel
396 page fault test, multi-process test may be better than multi-thread 403 page fault test, multi-process test may be better than multi-thread
397 test because it has noise of shared objects/status. 404 test because it has noise of shared objects/status.
398 405
399 But the above two are testing extreme situations. 406 But the above two are testing extreme situations.
400 Trying usual test under memory controller is always helpful. 407 Trying usual test under memory controller is always helpful.
401 408
402 4.1 Troubleshooting 409 4.1 Troubleshooting
403 410
404 Sometimes a user might find that the application under a cgroup is 411 Sometimes a user might find that the application under a cgroup is
405 terminated by the OOM killer. There are several causes for this: 412 terminated by the OOM killer. There are several causes for this:
406 413
407 1. The cgroup limit is too low (just too low to do anything useful) 414 1. The cgroup limit is too low (just too low to do anything useful)
408 2. The user is using anonymous memory and swap is turned off or too low 415 2. The user is using anonymous memory and swap is turned off or too low
409 416
410 A sync followed by echo 1 > /proc/sys/vm/drop_caches will help get rid of 417 A sync followed by echo 1 > /proc/sys/vm/drop_caches will help get rid of
411 some of the pages cached in the cgroup (page cache pages). 418 some of the pages cached in the cgroup (page cache pages).
412 419
413 To know what happens, disabling OOM_Kill as per "10. OOM Control" (below) and 420 To know what happens, disabling OOM_Kill as per "10. OOM Control" (below) and
414 seeing what happens will be helpful. 421 seeing what happens will be helpful.
415 422
416 4.2 Task migration 423 4.2 Task migration
417 424
418 When a task migrates from one cgroup to another, its charge is not 425 When a task migrates from one cgroup to another, its charge is not
419 carried forward by default. The pages allocated from the original cgroup still 426 carried forward by default. The pages allocated from the original cgroup still
420 remain charged to it, the charge is dropped when the page is freed or 427 remain charged to it, the charge is dropped when the page is freed or
421 reclaimed. 428 reclaimed.
422 429
423 You can move charges of a task along with task migration. 430 You can move charges of a task along with task migration.
424 See 8. "Move charges at task migration" 431 See 8. "Move charges at task migration"
425 432
426 4.3 Removing a cgroup 433 4.3 Removing a cgroup
427 434
428 A cgroup can be removed by rmdir, but as discussed in sections 4.1 and 4.2, a 435 A cgroup can be removed by rmdir, but as discussed in sections 4.1 and 4.2, a
429 cgroup might have some charge associated with it, even though all 436 cgroup might have some charge associated with it, even though all
430 tasks have migrated away from it. (because we charge against pages, not 437 tasks have migrated away from it. (because we charge against pages, not
431 against tasks.) 438 against tasks.)
432 439
433 We move the stats to root (if use_hierarchy==0) or parent (if 440 We move the stats to root (if use_hierarchy==0) or parent (if
434 use_hierarchy==1), and no change on the charge except uncharging 441 use_hierarchy==1), and no change on the charge except uncharging
435 from the child. 442 from the child.
436 443
437 Charges recorded in swap information is not updated at removal of cgroup. 444 Charges recorded in swap information is not updated at removal of cgroup.
438 Recorded information is discarded and a cgroup which uses swap (swapcache) 445 Recorded information is discarded and a cgroup which uses swap (swapcache)
439 will be charged as a new owner of it. 446 will be charged as a new owner of it.
440 447
441 About use_hierarchy, see Section 6. 448 About use_hierarchy, see Section 6.
442 449
443 5. Misc. interfaces. 450 5. Misc. interfaces.
444 451
445 5.1 force_empty 452 5.1 force_empty
446 memory.force_empty interface is provided to make cgroup's memory usage empty. 453 memory.force_empty interface is provided to make cgroup's memory usage empty.
447 You can use this interface only when the cgroup has no tasks. 454 You can use this interface only when the cgroup has no tasks.
448 When writing anything to this 455 When writing anything to this
449 456
450 # echo 0 > memory.force_empty 457 # echo 0 > memory.force_empty
451 458
452 Almost all pages tracked by this memory cgroup will be unmapped and freed. 459 Almost all pages tracked by this memory cgroup will be unmapped and freed.
453 Some pages cannot be freed because they are locked or in-use. Such pages are 460 Some pages cannot be freed because they are locked or in-use. Such pages are
454 moved to parent (if use_hierarchy==1) or root (if use_hierarchy==0) and this 461 moved to parent (if use_hierarchy==1) or root (if use_hierarchy==0) and this
455 cgroup will be empty. 462 cgroup will be empty.
456 463
457 The typical use case for this interface is before calling rmdir(). 464 The typical use case for this interface is before calling rmdir().
458 Because rmdir() moves all pages to parent, some out-of-use page caches can be 465 Because rmdir() moves all pages to parent, some out-of-use page caches can be
459 moved to the parent. If you want to avoid that, force_empty will be useful. 466 moved to the parent. If you want to avoid that, force_empty will be useful.
460 467
461 Also, note that when memory.kmem.limit_in_bytes is set the charges due to 468 Also, note that when memory.kmem.limit_in_bytes is set the charges due to
462 kernel pages will still be seen. This is not considered a failure and the 469 kernel pages will still be seen. This is not considered a failure and the
463 write will still return success. In this case, it is expected that 470 write will still return success. In this case, it is expected that
464 memory.kmem.usage_in_bytes == memory.usage_in_bytes. 471 memory.kmem.usage_in_bytes == memory.usage_in_bytes.
465 472
466 About use_hierarchy, see Section 6. 473 About use_hierarchy, see Section 6.
467 474
468 5.2 stat file 475 5.2 stat file
469 476
470 memory.stat file includes following statistics 477 memory.stat file includes following statistics
471 478
472 # per-memory cgroup local status 479 # per-memory cgroup local status
473 cache - # of bytes of page cache memory. 480 cache - # of bytes of page cache memory.
474 rss - # of bytes of anonymous and swap cache memory. 481 rss - # of bytes of anonymous and swap cache memory.
475 mapped_file - # of bytes of mapped file (includes tmpfs/shmem) 482 mapped_file - # of bytes of mapped file (includes tmpfs/shmem)
476 pgpgin - # of charging events to the memory cgroup. The charging 483 pgpgin - # of charging events to the memory cgroup. The charging
477 event happens each time a page is accounted as either mapped 484 event happens each time a page is accounted as either mapped
478 anon page(RSS) or cache page(Page Cache) to the cgroup. 485 anon page(RSS) or cache page(Page Cache) to the cgroup.
479 pgpgout - # of uncharging events to the memory cgroup. The uncharging 486 pgpgout - # of uncharging events to the memory cgroup. The uncharging
480 event happens each time a page is unaccounted from the cgroup. 487 event happens each time a page is unaccounted from the cgroup.
481 swap - # of bytes of swap usage 488 swap - # of bytes of swap usage
482 inactive_anon - # of bytes of anonymous memory and swap cache memory on 489 inactive_anon - # of bytes of anonymous memory and swap cache memory on
483 LRU list. 490 LRU list.
484 active_anon - # of bytes of anonymous and swap cache memory on active 491 active_anon - # of bytes of anonymous and swap cache memory on active
485 inactive LRU list. 492 inactive LRU list.
486 inactive_file - # of bytes of file-backed memory on inactive LRU list. 493 inactive_file - # of bytes of file-backed memory on inactive LRU list.
487 active_file - # of bytes of file-backed memory on active LRU list. 494 active_file - # of bytes of file-backed memory on active LRU list.
488 unevictable - # of bytes of memory that cannot be reclaimed (mlocked etc). 495 unevictable - # of bytes of memory that cannot be reclaimed (mlocked etc).
489 496
490 # status considering hierarchy (see memory.use_hierarchy settings) 497 # status considering hierarchy (see memory.use_hierarchy settings)
491 498
492 hierarchical_memory_limit - # of bytes of memory limit with regard to hierarchy 499 hierarchical_memory_limit - # of bytes of memory limit with regard to hierarchy
493 under which the memory cgroup is 500 under which the memory cgroup is
494 hierarchical_memsw_limit - # of bytes of memory+swap limit with regard to 501 hierarchical_memsw_limit - # of bytes of memory+swap limit with regard to
495 hierarchy under which memory cgroup is. 502 hierarchy under which memory cgroup is.
496 503
497 total_<counter> - # hierarchical version of <counter>, which in 504 total_<counter> - # hierarchical version of <counter>, which in
498 addition to the cgroup's own value includes the 505 addition to the cgroup's own value includes the
499 sum of all hierarchical children's values of 506 sum of all hierarchical children's values of
500 <counter>, i.e. total_cache 507 <counter>, i.e. total_cache
501 508
502 # The following additional stats are dependent on CONFIG_DEBUG_VM. 509 # The following additional stats are dependent on CONFIG_DEBUG_VM.
503 510
504 recent_rotated_anon - VM internal parameter. (see mm/vmscan.c) 511 recent_rotated_anon - VM internal parameter. (see mm/vmscan.c)
505 recent_rotated_file - VM internal parameter. (see mm/vmscan.c) 512 recent_rotated_file - VM internal parameter. (see mm/vmscan.c)
506 recent_scanned_anon - VM internal parameter. (see mm/vmscan.c) 513 recent_scanned_anon - VM internal parameter. (see mm/vmscan.c)
507 recent_scanned_file - VM internal parameter. (see mm/vmscan.c) 514 recent_scanned_file - VM internal parameter. (see mm/vmscan.c)
508 515
509 Memo: 516 Memo:
510 recent_rotated means recent frequency of LRU rotation. 517 recent_rotated means recent frequency of LRU rotation.
511 recent_scanned means recent # of scans to LRU. 518 recent_scanned means recent # of scans to LRU.
512 showing for better debug please see the code for meanings. 519 showing for better debug please see the code for meanings.
513 520
514 Note: 521 Note:
515 Only anonymous and swap cache memory is listed as part of 'rss' stat. 522 Only anonymous and swap cache memory is listed as part of 'rss' stat.
516 This should not be confused with the true 'resident set size' or the 523 This should not be confused with the true 'resident set size' or the
517 amount of physical memory used by the cgroup. 524 amount of physical memory used by the cgroup.
518 'rss + file_mapped" will give you resident set size of cgroup. 525 'rss + file_mapped" will give you resident set size of cgroup.
519 (Note: file and shmem may be shared among other cgroups. In that case, 526 (Note: file and shmem may be shared among other cgroups. In that case,
520 file_mapped is accounted only when the memory cgroup is owner of page 527 file_mapped is accounted only when the memory cgroup is owner of page
521 cache.) 528 cache.)
522 529
523 5.3 swappiness 530 5.3 swappiness
524 531
525 Similar to /proc/sys/vm/swappiness, but affecting a hierarchy of groups only. 532 Similar to /proc/sys/vm/swappiness, but affecting a hierarchy of groups only.
526 Please note that unlike the global swappiness, memcg knob set to 0 533 Please note that unlike the global swappiness, memcg knob set to 0
527 really prevents from any swapping even if there is a swap storage 534 really prevents from any swapping even if there is a swap storage
528 available. This might lead to memcg OOM killer if there are no file 535 available. This might lead to memcg OOM killer if there are no file
529 pages to reclaim. 536 pages to reclaim.
530 537
531 Following cgroups' swappiness can't be changed. 538 Following cgroups' swappiness can't be changed.
532 - root cgroup (uses /proc/sys/vm/swappiness). 539 - root cgroup (uses /proc/sys/vm/swappiness).
533 - a cgroup which uses hierarchy and it has other cgroup(s) below it. 540 - a cgroup which uses hierarchy and it has other cgroup(s) below it.
534 - a cgroup which uses hierarchy and not the root of hierarchy. 541 - a cgroup which uses hierarchy and not the root of hierarchy.
535 542
536 5.4 failcnt 543 5.4 failcnt
537 544
538 A memory cgroup provides memory.failcnt and memory.memsw.failcnt files. 545 A memory cgroup provides memory.failcnt and memory.memsw.failcnt files.
539 This failcnt(== failure count) shows the number of times that a usage counter 546 This failcnt(== failure count) shows the number of times that a usage counter
540 hit its limit. When a memory cgroup hits a limit, failcnt increases and 547 hit its limit. When a memory cgroup hits a limit, failcnt increases and
541 memory under it will be reclaimed. 548 memory under it will be reclaimed.
542 549
543 You can reset failcnt by writing 0 to failcnt file. 550 You can reset failcnt by writing 0 to failcnt file.
544 # echo 0 > .../memory.failcnt 551 # echo 0 > .../memory.failcnt
545 552
546 5.5 usage_in_bytes 553 5.5 usage_in_bytes
547 554
548 For efficiency, as other kernel components, memory cgroup uses some optimization 555 For efficiency, as other kernel components, memory cgroup uses some optimization
549 to avoid unnecessary cacheline false sharing. usage_in_bytes is affected by the 556 to avoid unnecessary cacheline false sharing. usage_in_bytes is affected by the
550 method and doesn't show 'exact' value of memory (and swap) usage, it's a fuzz 557 method and doesn't show 'exact' value of memory (and swap) usage, it's a fuzz
551 value for efficient access. (Of course, when necessary, it's synchronized.) 558 value for efficient access. (Of course, when necessary, it's synchronized.)
552 If you want to know more exact memory usage, you should use RSS+CACHE(+SWAP) 559 If you want to know more exact memory usage, you should use RSS+CACHE(+SWAP)
553 value in memory.stat(see 5.2). 560 value in memory.stat(see 5.2).
554 561
555 5.6 numa_stat 562 5.6 numa_stat
556 563
557 This is similar to numa_maps but operates on a per-memcg basis. This is 564 This is similar to numa_maps but operates on a per-memcg basis. This is
558 useful for providing visibility into the numa locality information within 565 useful for providing visibility into the numa locality information within
559 an memcg since the pages are allowed to be allocated from any physical 566 an memcg since the pages are allowed to be allocated from any physical
560 node. One of the use cases is evaluating application performance by 567 node. One of the use cases is evaluating application performance by
561 combining this information with the application's CPU allocation. 568 combining this information with the application's CPU allocation.
562 569
563 We export "total", "file", "anon" and "unevictable" pages per-node for 570 We export "total", "file", "anon" and "unevictable" pages per-node for
564 each memcg. The ouput format of memory.numa_stat is: 571 each memcg. The ouput format of memory.numa_stat is:
565 572
566 total=<total pages> N0=<node 0 pages> N1=<node 1 pages> ... 573 total=<total pages> N0=<node 0 pages> N1=<node 1 pages> ...
567 file=<total file pages> N0=<node 0 pages> N1=<node 1 pages> ... 574 file=<total file pages> N0=<node 0 pages> N1=<node 1 pages> ...
568 anon=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ... 575 anon=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ...
569 unevictable=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ... 576 unevictable=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ...
570 577
571 And we have total = file + anon + unevictable. 578 And we have total = file + anon + unevictable.
572 579
573 6. Hierarchy support 580 6. Hierarchy support
574 581
575 The memory controller supports a deep hierarchy and hierarchical accounting. 582 The memory controller supports a deep hierarchy and hierarchical accounting.
576 The hierarchy is created by creating the appropriate cgroups in the 583 The hierarchy is created by creating the appropriate cgroups in the
577 cgroup filesystem. Consider for example, the following cgroup filesystem 584 cgroup filesystem. Consider for example, the following cgroup filesystem
578 hierarchy 585 hierarchy
579 586
580 root 587 root
581 / | \ 588 / | \
582 / | \ 589 / | \
583 a b c 590 a b c
584 | \ 591 | \
585 | \ 592 | \
586 d e 593 d e
587 594
588 In the diagram above, with hierarchical accounting enabled, all memory 595 In the diagram above, with hierarchical accounting enabled, all memory
589 usage of e, is accounted to its ancestors up until the root (i.e, c and root), 596 usage of e, is accounted to its ancestors up until the root (i.e, c and root),
590 that has memory.use_hierarchy enabled. If one of the ancestors goes over its 597 that has memory.use_hierarchy enabled. If one of the ancestors goes over its
591 limit, the reclaim algorithm reclaims from the tasks in the ancestor and the 598 limit, the reclaim algorithm reclaims from the tasks in the ancestor and the
592 children of the ancestor. 599 children of the ancestor.
593 600
594 6.1 Enabling hierarchical accounting and reclaim 601 6.1 Enabling hierarchical accounting and reclaim
595 602
596 A memory cgroup by default disables the hierarchy feature. Support 603 A memory cgroup by default disables the hierarchy feature. Support
597 can be enabled by writing 1 to memory.use_hierarchy file of the root cgroup 604 can be enabled by writing 1 to memory.use_hierarchy file of the root cgroup
598 605
599 # echo 1 > memory.use_hierarchy 606 # echo 1 > memory.use_hierarchy
600 607
601 The feature can be disabled by 608 The feature can be disabled by
602 609
603 # echo 0 > memory.use_hierarchy 610 # echo 0 > memory.use_hierarchy
604 611
605 NOTE1: Enabling/disabling will fail if either the cgroup already has other 612 NOTE1: Enabling/disabling will fail if either the cgroup already has other
606 cgroups created below it, or if the parent cgroup has use_hierarchy 613 cgroups created below it, or if the parent cgroup has use_hierarchy
607 enabled. 614 enabled.
608 615
609 NOTE2: When panic_on_oom is set to "2", the whole system will panic in 616 NOTE2: When panic_on_oom is set to "2", the whole system will panic in
610 case of an OOM event in any cgroup. 617 case of an OOM event in any cgroup.
611 618
612 7. Soft limits 619 7. Soft limits
613 620
614 Soft limits allow for greater sharing of memory. The idea behind soft limits 621 Soft limits allow for greater sharing of memory. The idea behind soft limits
615 is to allow control groups to use as much of the memory as needed, provided 622 is to allow control groups to use as much of the memory as needed, provided
616 623
617 a. There is no memory contention 624 a. There is no memory contention
618 b. They do not exceed their hard limit 625 b. They do not exceed their hard limit
619 626
620 When the system detects memory contention or low memory, control groups 627 When the system detects memory contention or low memory, control groups
621 are pushed back to their soft limits. If the soft limit of each control 628 are pushed back to their soft limits. If the soft limit of each control
622 group is very high, they are pushed back as much as possible to make 629 group is very high, they are pushed back as much as possible to make
623 sure that one control group does not starve the others of memory. 630 sure that one control group does not starve the others of memory.
624 631
625 Please note that soft limits is a best-effort feature; it comes with 632 Please note that soft limits is a best-effort feature; it comes with
626 no guarantees, but it does its best to make sure that when memory is 633 no guarantees, but it does its best to make sure that when memory is
627 heavily contended for, memory is allocated based on the soft limit 634 heavily contended for, memory is allocated based on the soft limit
628 hints/setup. Currently soft limit based reclaim is set up such that 635 hints/setup. Currently soft limit based reclaim is set up such that
629 it gets invoked from balance_pgdat (kswapd). 636 it gets invoked from balance_pgdat (kswapd).
630 637
631 7.1 Interface 638 7.1 Interface
632 639
633 Soft limits can be setup by using the following commands (in this example we 640 Soft limits can be setup by using the following commands (in this example we
634 assume a soft limit of 256 MiB) 641 assume a soft limit of 256 MiB)
635 642
636 # echo 256M > memory.soft_limit_in_bytes 643 # echo 256M > memory.soft_limit_in_bytes
637 644
638 If we want to change this to 1G, we can at any time use 645 If we want to change this to 1G, we can at any time use
639 646
640 # echo 1G > memory.soft_limit_in_bytes 647 # echo 1G > memory.soft_limit_in_bytes
641 648
642 NOTE1: Soft limits take effect over a long period of time, since they involve 649 NOTE1: Soft limits take effect over a long period of time, since they involve
643 reclaiming memory for balancing between memory cgroups 650 reclaiming memory for balancing between memory cgroups
644 NOTE2: It is recommended to set the soft limit always below the hard limit, 651 NOTE2: It is recommended to set the soft limit always below the hard limit,
645 otherwise the hard limit will take precedence. 652 otherwise the hard limit will take precedence.
646 653
647 8. Move charges at task migration 654 8. Move charges at task migration
648 655
649 Users can move charges associated with a task along with task migration, that 656 Users can move charges associated with a task along with task migration, that
650 is, uncharge task's pages from the old cgroup and charge them to the new cgroup. 657 is, uncharge task's pages from the old cgroup and charge them to the new cgroup.
651 This feature is not supported in !CONFIG_MMU environments because of lack of 658 This feature is not supported in !CONFIG_MMU environments because of lack of
652 page tables. 659 page tables.
653 660
654 8.1 Interface 661 8.1 Interface
655 662
656 This feature is disabled by default. It can be enabledi (and disabled again) by 663 This feature is disabled by default. It can be enabledi (and disabled again) by
657 writing to memory.move_charge_at_immigrate of the destination cgroup. 664 writing to memory.move_charge_at_immigrate of the destination cgroup.
658 665
659 If you want to enable it: 666 If you want to enable it:
660 667
661 # echo (some positive value) > memory.move_charge_at_immigrate 668 # echo (some positive value) > memory.move_charge_at_immigrate
662 669
663 Note: Each bits of move_charge_at_immigrate has its own meaning about what type 670 Note: Each bits of move_charge_at_immigrate has its own meaning about what type
664 of charges should be moved. See 8.2 for details. 671 of charges should be moved. See 8.2 for details.
665 Note: Charges are moved only when you move mm->owner, in other words, 672 Note: Charges are moved only when you move mm->owner, in other words,
666 a leader of a thread group. 673 a leader of a thread group.
667 Note: If we cannot find enough space for the task in the destination cgroup, we 674 Note: If we cannot find enough space for the task in the destination cgroup, we
668 try to make space by reclaiming memory. Task migration may fail if we 675 try to make space by reclaiming memory. Task migration may fail if we
669 cannot make enough space. 676 cannot make enough space.
670 Note: It can take several seconds if you move charges much. 677 Note: It can take several seconds if you move charges much.
671 678
672 And if you want disable it again: 679 And if you want disable it again:
673 680
674 # echo 0 > memory.move_charge_at_immigrate 681 # echo 0 > memory.move_charge_at_immigrate
675 682
676 8.2 Type of charges which can be moved 683 8.2 Type of charges which can be moved
677 684
678 Each bit in move_charge_at_immigrate has its own meaning about what type of 685 Each bit in move_charge_at_immigrate has its own meaning about what type of
679 charges should be moved. But in any case, it must be noted that an account of 686 charges should be moved. But in any case, it must be noted that an account of
680 a page or a swap can be moved only when it is charged to the task's current 687 a page or a swap can be moved only when it is charged to the task's current
681 (old) memory cgroup. 688 (old) memory cgroup.
682 689
683 bit | what type of charges would be moved ? 690 bit | what type of charges would be moved ?
684 -----+------------------------------------------------------------------------ 691 -----+------------------------------------------------------------------------
685 0 | A charge of an anonymous page (or swap of it) used by the target task. 692 0 | A charge of an anonymous page (or swap of it) used by the target task.
686 | You must enable Swap Extension (see 2.4) to enable move of swap charges. 693 | You must enable Swap Extension (see 2.4) to enable move of swap charges.
687 -----+------------------------------------------------------------------------ 694 -----+------------------------------------------------------------------------
688 1 | A charge of file pages (normal file, tmpfs file (e.g. ipc shared memory) 695 1 | A charge of file pages (normal file, tmpfs file (e.g. ipc shared memory)
689 | and swaps of tmpfs file) mmapped by the target task. Unlike the case of 696 | and swaps of tmpfs file) mmapped by the target task. Unlike the case of
690 | anonymous pages, file pages (and swaps) in the range mmapped by the task 697 | anonymous pages, file pages (and swaps) in the range mmapped by the task
691 | will be moved even if the task hasn't done page fault, i.e. they might 698 | will be moved even if the task hasn't done page fault, i.e. they might
692 | not be the task's "RSS", but other task's "RSS" that maps the same file. 699 | not be the task's "RSS", but other task's "RSS" that maps the same file.
693 | And mapcount of the page is ignored (the page can be moved even if 700 | And mapcount of the page is ignored (the page can be moved even if
694 | page_mapcount(page) > 1). You must enable Swap Extension (see 2.4) to 701 | page_mapcount(page) > 1). You must enable Swap Extension (see 2.4) to
695 | enable move of swap charges. 702 | enable move of swap charges.
696 703
697 8.3 TODO 704 8.3 TODO
698 705
699 - All of moving charge operations are done under cgroup_mutex. It's not good 706 - All of moving charge operations are done under cgroup_mutex. It's not good
700 behavior to hold the mutex too long, so we may need some trick. 707 behavior to hold the mutex too long, so we may need some trick.
701 708
702 9. Memory thresholds 709 9. Memory thresholds
703 710
704 Memory cgroup implements memory thresholds using the cgroups notification 711 Memory cgroup implements memory thresholds using the cgroups notification
705 API (see cgroups.txt). It allows to register multiple memory and memsw 712 API (see cgroups.txt). It allows to register multiple memory and memsw
706 thresholds and gets notifications when it crosses. 713 thresholds and gets notifications when it crosses.
707 714
708 To register a threshold, an application must: 715 To register a threshold, an application must:
709 - create an eventfd using eventfd(2); 716 - create an eventfd using eventfd(2);
710 - open memory.usage_in_bytes or memory.memsw.usage_in_bytes; 717 - open memory.usage_in_bytes or memory.memsw.usage_in_bytes;
711 - write string like "<event_fd> <fd of memory.usage_in_bytes> <threshold>" to 718 - write string like "<event_fd> <fd of memory.usage_in_bytes> <threshold>" to
712 cgroup.event_control. 719 cgroup.event_control.
713 720
714 Application will be notified through eventfd when memory usage crosses 721 Application will be notified through eventfd when memory usage crosses
715 threshold in any direction. 722 threshold in any direction.
716 723
717 It's applicable for root and non-root cgroup. 724 It's applicable for root and non-root cgroup.
718 725
719 10. OOM Control 726 10. OOM Control
720 727
721 memory.oom_control file is for OOM notification and other controls. 728 memory.oom_control file is for OOM notification and other controls.
722 729
723 Memory cgroup implements OOM notifier using the cgroup notification 730 Memory cgroup implements OOM notifier using the cgroup notification
724 API (See cgroups.txt). It allows to register multiple OOM notification 731 API (See cgroups.txt). It allows to register multiple OOM notification
725 delivery and gets notification when OOM happens. 732 delivery and gets notification when OOM happens.
726 733
727 To register a notifier, an application must: 734 To register a notifier, an application must:
728 - create an eventfd using eventfd(2) 735 - create an eventfd using eventfd(2)
729 - open memory.oom_control file 736 - open memory.oom_control file
730 - write string like "<event_fd> <fd of memory.oom_control>" to 737 - write string like "<event_fd> <fd of memory.oom_control>" to
731 cgroup.event_control 738 cgroup.event_control
732 739
733 The application will be notified through eventfd when OOM happens. 740 The application will be notified through eventfd when OOM happens.
734 OOM notification doesn't work for the root cgroup. 741 OOM notification doesn't work for the root cgroup.
735 742
736 You can disable the OOM-killer by writing "1" to memory.oom_control file, as: 743 You can disable the OOM-killer by writing "1" to memory.oom_control file, as:
737 744
738 #echo 1 > memory.oom_control 745 #echo 1 > memory.oom_control
739 746
740 This operation is only allowed to the top cgroup of a sub-hierarchy. 747 This operation is only allowed to the top cgroup of a sub-hierarchy.
741 If OOM-killer is disabled, tasks under cgroup will hang/sleep 748 If OOM-killer is disabled, tasks under cgroup will hang/sleep
742 in memory cgroup's OOM-waitqueue when they request accountable memory. 749 in memory cgroup's OOM-waitqueue when they request accountable memory.
743 750
744 For running them, you have to relax the memory cgroup's OOM status by 751 For running them, you have to relax the memory cgroup's OOM status by
745 * enlarge limit or reduce usage. 752 * enlarge limit or reduce usage.
746 To reduce usage, 753 To reduce usage,
747 * kill some tasks. 754 * kill some tasks.
748 * move some tasks to other group with account migration. 755 * move some tasks to other group with account migration.
749 * remove some files (on tmpfs?) 756 * remove some files (on tmpfs?)
750 757
751 Then, stopped tasks will work again. 758 Then, stopped tasks will work again.
752 759
753 At reading, current status of OOM is shown. 760 At reading, current status of OOM is shown.
754 oom_kill_disable 0 or 1 (if 1, oom-killer is disabled) 761 oom_kill_disable 0 or 1 (if 1, oom-killer is disabled)
755 under_oom 0 or 1 (if 1, the memory cgroup is under OOM, tasks may 762 under_oom 0 or 1 (if 1, the memory cgroup is under OOM, tasks may
756 be stopped.) 763 be stopped.)
757 764
758 11. TODO 765 11. TODO
759 766
760 1. Add support for accounting huge pages (as a separate controller) 767 1. Add support for accounting huge pages (as a separate controller)
761 2. Make per-cgroup scanner reclaim not-shared pages first 768 2. Make per-cgroup scanner reclaim not-shared pages first
762 3. Teach controller to account for shared-pages 769 3. Teach controller to account for shared-pages
763 4. Start reclamation in the background when the limit is 770 4. Start reclamation in the background when the limit is
764 not yet hit but the usage is getting closer 771 not yet hit but the usage is getting closer
765 772
766 Summary 773 Summary
767 774
768 Overall, the memory controller has been a stable controller and has been 775 Overall, the memory controller has been a stable controller and has been
769 commented and discussed quite extensively in the community. 776 commented and discussed quite extensively in the community.
770 777
771 References 778 References
772 779
773 1. Singh, Balbir. RFC: Memory Controller, http://lwn.net/Articles/206697/ 780 1. Singh, Balbir. RFC: Memory Controller, http://lwn.net/Articles/206697/
774 2. Singh, Balbir. Memory Controller (RSS Control), 781 2. Singh, Balbir. Memory Controller (RSS Control),
775 http://lwn.net/Articles/222762/ 782 http://lwn.net/Articles/222762/
776 3. Emelianov, Pavel. Resource controllers based on process cgroups 783 3. Emelianov, Pavel. Resource controllers based on process cgroups
777 http://lkml.org/lkml/2007/3/6/198 784 http://lkml.org/lkml/2007/3/6/198
778 4. Emelianov, Pavel. RSS controller based on process cgroups (v2) 785 4. Emelianov, Pavel. RSS controller based on process cgroups (v2)
779 http://lkml.org/lkml/2007/4/9/78 786 http://lkml.org/lkml/2007/4/9/78
780 5. Emelianov, Pavel. RSS controller based on process cgroups (v3) 787 5. Emelianov, Pavel. RSS controller based on process cgroups (v3)
781 http://lkml.org/lkml/2007/5/30/244 788 http://lkml.org/lkml/2007/5/30/244
782 6. Menage, Paul. Control Groups v10, http://lwn.net/Articles/236032/ 789 6. Menage, Paul. Control Groups v10, http://lwn.net/Articles/236032/
783 7. Vaidyanathan, Srinivasan, Control Groups: Pagecache accounting and control 790 7. Vaidyanathan, Srinivasan, Control Groups: Pagecache accounting and control
784 subsystem (v3), http://lwn.net/Articles/235534/ 791 subsystem (v3), http://lwn.net/Articles/235534/
785 8. Singh, Balbir. RSS controller v2 test results (lmbench), 792 8. Singh, Balbir. RSS controller v2 test results (lmbench),
786 http://lkml.org/lkml/2007/5/17/232 793 http://lkml.org/lkml/2007/5/17/232
787 9. Singh, Balbir. RSS controller v2 AIM9 results 794 9. Singh, Balbir. RSS controller v2 AIM9 results
788 http://lkml.org/lkml/2007/5/18/1 795 http://lkml.org/lkml/2007/5/18/1
789 10. Singh, Balbir. Memory controller v6 test results, 796 10. Singh, Balbir. Memory controller v6 test results,
790 http://lkml.org/lkml/2007/8/19/36 797 http://lkml.org/lkml/2007/8/19/36
791 11. Singh, Balbir. Memory controller introduction (v6), 798 11. Singh, Balbir. Memory controller introduction (v6),
792 http://lkml.org/lkml/2007/8/17/69 799 http://lkml.org/lkml/2007/8/17/69
793 12. Corbet, Jonathan, Controlling memory use in cgroups, 800 12. Corbet, Jonathan, Controlling memory use in cgroups,
794 http://lwn.net/Articles/243795/ 801 http://lwn.net/Articles/243795/
795 802