blob: 6d0c2519cf4728044536aae4615a91adf1c4e6c9 [file] [log] [blame]
Yasunori Goto6867c932007-08-10 13:00:59 -07001==============
2Memory Hotplug
3==============
4
Yasunori Goto10020ca2007-10-21 16:41:36 -07005Created: Jul 28 2007
6Add description of notifier of memory hotplug Oct 11 2007
Yasunori Goto6867c932007-08-10 13:00:59 -07007
8This document is about memory hotplug including how-to-use and current status.
9Because Memory Hotplug is still under development, contents of this text will
10be changed often.
11
121. Introduction
13 1.1 purpose of memory hotplug
14 1.2. Phases of memory hotplug
15 1.3. Unit of Memory online/offline operation
162. Kernel Configuration
173. sysfs files for memory hotplug
184. Physical memory hot-add phase
19 4.1 Hardware(Firmware) Support
20 4.2 Notify memory hot-add event by hand
215. Logical Memory hot-add phase
22 5.1. State of memory
23 5.2. How to online memory
246. Logical memory remove
25 6.1 Memory offline and ZONE_MOVABLE
26 6.2. How to offline memory
277. Physical memory remove
Yasunori Goto10020ca2007-10-21 16:41:36 -0700288. Memory hotplug event notifier
299. Future Work List
Yasunori Goto6867c932007-08-10 13:00:59 -070030
31Note(1): x86_64's has special implementation for memory hotplug.
32 This text does not describe it.
33Note(2): This text assumes that sysfs is mounted at /sys.
34
35
36---------------
371. Introduction
38---------------
39
401.1 purpose of memory hotplug
41------------
42Memory Hotplug allows users to increase/decrease the amount of memory.
43Generally, there are two purposes.
44
45(A) For changing the amount of memory.
46 This is to allow a feature like capacity on demand.
47(B) For installing/removing DIMMs or NUMA-nodes physically.
48 This is to exchange DIMMs/NUMA-nodes, reduce power consumption, etc.
49
50(A) is required by highly virtualized environments and (B) is required by
51hardware which supports memory power management.
52
53Linux memory hotplug is designed for both purpose.
54
55
561.2. Phases of memory hotplug
57---------------
58There are 2 phases in Memory Hotplug.
59 1) Physical Memory Hotplug phase
60 2) Logical Memory Hotplug phase.
61
62The First phase is to communicate hardware/firmware and make/erase
63environment for hotplugged memory. Basically, this phase is necessary
64for the purpose (B), but this is good phase for communication between
65highly virtualized environments too.
66
67When memory is hotplugged, the kernel recognizes new memory, makes new memory
68management tables, and makes sysfs files for new memory's operation.
69
70If firmware supports notification of connection of new memory to OS,
71this phase is triggered automatically. ACPI can notify this event. If not,
72"probe" operation by system administration is used instead.
73(see Section 4.).
74
75Logical Memory Hotplug phase is to change memory state into
Matt LaPlante19f59462009-04-27 15:06:31 +020076available/unavailable for users. Amount of memory from user's view is
Yasunori Goto6867c932007-08-10 13:00:59 -070077changed by this phase. The kernel makes all memory in it as free pages
78when a memory range is available.
79
80In this document, this phase is described as online/offline.
81
Matt LaPlante19f59462009-04-27 15:06:31 +020082Logical Memory Hotplug phase is triggered by write of sysfs file by system
Yasunori Goto6867c932007-08-10 13:00:59 -070083administrator. For the hot-add case, it must be executed after Physical Hotplug
84phase by hand.
85(However, if you writes udev's hotplug scripts for memory hotplug, these
86 phases can be execute in seamless way.)
87
88
891.3. Unit of Memory online/offline operation
90------------
91Memory hotplug uses SPARSEMEM memory model. SPARSEMEM divides the whole memory
92into chunks of the same size. The chunk is called a "section". The size of
93a section is architecture dependent. For example, power uses 16MiB, ia64 uses
941GiB. The unit of online/offline operation is "one section". (see Section 3.)
95
96To determine the size of sections, please read this file:
97
98/sys/devices/system/memory/block_size_bytes
99
100This file shows the size of sections in byte.
101
102-----------------------
1032. Kernel Configuration
104-----------------------
105To use memory hotplug feature, kernel must be compiled with following
106config options.
107
108- For all memory hotplug
109 Memory model -> Sparse Memory (CONFIG_SPARSEMEM)
110 Allow for memory hot-add (CONFIG_MEMORY_HOTPLUG)
111
112- To enable memory removal, the followings are also necessary
113 Allow for memory hot remove (CONFIG_MEMORY_HOTREMOVE)
114 Page Migration (CONFIG_MIGRATION)
115
116- For ACPI memory hotplug, the followings are also necessary
117 Memory hotplug (under ACPI Support menu) (CONFIG_ACPI_HOTPLUG_MEMORY)
118 This option can be kernel module.
119
120- As a related configuration, if your box has a feature of NUMA-node hotplug
121 via ACPI, then this option is necessary too.
122 ACPI0004,PNP0A05 and PNP0A06 Container Driver (under ACPI Support menu)
123 (CONFIG_ACPI_CONTAINER).
124 This option can be kernel module too.
125
126--------------------------------
Gary Hadec04fc582009-01-06 14:39:14 -08001274 sysfs files for memory hotplug
Yasunori Goto6867c932007-08-10 13:00:59 -0700128--------------------------------
Nathan Fontenot0c2c99b2011-01-20 10:43:34 -0600129All sections have their device information in sysfs. Each section is part of
130a memory block under /sys/devices/system/memory as
Yasunori Goto6867c932007-08-10 13:00:59 -0700131
132/sys/devices/system/memory/memoryXXX
Nathan Fontenot0c2c99b2011-01-20 10:43:34 -0600133(XXX is the section id.)
Yasunori Goto6867c932007-08-10 13:00:59 -0700134
Nathan Fontenot0c2c99b2011-01-20 10:43:34 -0600135Now, XXX is defined as (start_address_of_section / section_size) of the first
136section contained in the memory block. The files 'phys_index' and
137'end_phys_index' under each directory report the beginning and end section id's
138for the memory block covered by the sysfs directory. It is expected that all
139memory sections in this range are present and no memory holes exist in the
140range. Currently there is no way to determine if there is a memory hole, but
141the existence of one should not affect the hotplug capabilities of the memory
142block.
Yasunori Goto6867c932007-08-10 13:00:59 -0700143
144For example, assume 1GiB section size. A device for a memory starting at
1450x100000000 is /sys/device/system/memory/memory4
146(0x100000000 / 1Gib = 4)
147This device covers address range [0x100000000 ... 0x140000000)
148
Nathan Fontenot0c2c99b2011-01-20 10:43:34 -0600149Under each section, you can see 4 or 5 files, the end_phys_index file being
150a recent addition and not present on older kernels.
Yasunori Goto6867c932007-08-10 13:00:59 -0700151
Nathan Fontenot0c2c99b2011-01-20 10:43:34 -0600152/sys/devices/system/memory/memoryXXX/start_phys_index
153/sys/devices/system/memory/memoryXXX/end_phys_index
Yasunori Goto6867c932007-08-10 13:00:59 -0700154/sys/devices/system/memory/memoryXXX/phys_device
155/sys/devices/system/memory/memoryXXX/state
Gary Hadec04fc582009-01-06 14:39:14 -0800156/sys/devices/system/memory/memoryXXX/removable
Yasunori Goto6867c932007-08-10 13:00:59 -0700157
Nathan Fontenot0c2c99b2011-01-20 10:43:34 -0600158'phys_index' : read-only and contains section id of the first section
159 in the memory block, same as XXX.
160'end_phys_index' : read-only and contains section id of the last section
161 in the memory block.
162'state' : read-write
163 at read: contains online/offline state of memory.
164 at write: user can specify "online", "offline" command
165 which will be performed on al sections in the block.
166'phys_device' : read-only: designed to show the name of physical memory
167 device. This is not well implemented now.
168'removable' : read-only: contains an integer value indicating
169 whether the memory block is removable or not
170 removable. A value of 1 indicates that the memory
171 block is removable and a value of 0 indicates that
172 it is not removable. A memory block is removable only if
173 every section in the block is removable.
Yasunori Goto6867c932007-08-10 13:00:59 -0700174
175NOTE:
176 These directories/files appear after physical memory hotplug phase.
177
Alex Chiangdee5d0d2009-12-14 17:59:05 -0800178If CONFIG_NUMA is enabled the memoryXXX/ directories can also be accessed
179via symbolic links located in the /sys/devices/system/node/node* directories.
180
181For example:
Gary Hadec04fc582009-01-06 14:39:14 -0800182/sys/devices/system/node/node0/memory9 -> ../../memory/memory9
Yasunori Goto6867c932007-08-10 13:00:59 -0700183
Alex Chiangdee5d0d2009-12-14 17:59:05 -0800184A backlink will also be created:
185/sys/devices/system/memory/memory9/node0 -> ../../node/node0
186
Yasunori Goto6867c932007-08-10 13:00:59 -0700187--------------------------------
1884. Physical memory hot-add phase
189--------------------------------
190
1914.1 Hardware(Firmware) Support
192------------
193On x86_64/ia64 platform, memory hotplug by ACPI is supported.
194
195In general, the firmware (ACPI) which supports memory hotplug defines
196memory class object of _HID "PNP0C80". When a notify is asserted to PNP0C80,
197Linux's ACPI handler does hot-add memory to the system and calls a hotplug udev
198script. This will be done automatically.
199
200But scripts for memory hotplug are not contained in generic udev package(now).
201You may have to write it by yourself or online/offline memory by hand.
202Please see "How to online memory", "How to offline memory" in this text.
203
204If firmware supports NUMA-node hotplug, and defines an object _HID "ACPI0004",
205"PNP0A05", or "PNP0A06", notification is asserted to it, and ACPI handler
206calls hotplug code for all of objects which are defined in it.
207If memory device is found, memory hotplug code will be called.
208
209
2104.2 Notify memory hot-add event by hand
211------------
212In some environments, especially virtualized environment, firmware will not
213notify memory hotplug event to the kernel. For such environment, "probe"
214interface is supported. This interface depends on CONFIG_ARCH_MEMORY_PROBE.
215
216Now, CONFIG_ARCH_MEMORY_PROBE is supported only by powerpc but it does not
217contain highly architecture codes. Please add config if you need "probe"
218interface.
219
220Probe interface is located at
221/sys/devices/system/memory/probe
222
223You can tell the physical address of new memory to the kernel by
224
225% echo start_address_of_new_memory > /sys/devices/system/memory/probe
226
227Then, [start_address_of_new_memory, start_address_of_new_memory + section_size)
228memory range is hot-added. In this case, hotplug script is not called (in
229current implementation). You'll have to online memory by yourself.
230Please see "How to online memory" in this text.
231
232
233
234------------------------------
2355. Logical Memory hot-add phase
236------------------------------
237
2385.1. State of memory
239------------
240To see (online/offline) state of memory section, read 'state' file.
241
242% cat /sys/device/system/memory/memoryXXX/state
243
244
245If the memory section is online, you'll read "online".
246If the memory section is offline, you'll read "offline".
247
248
2495.2. How to online memory
250------------
251Even if the memory is hot-added, it is not at ready-to-use state.
252For using newly added memory, you have to "online" the memory section.
253
254For onlining, you have to write "online" to the section's state file as:
255
256% echo online > /sys/devices/system/memory/memoryXXX/state
257
258After this, section memoryXXX's state will be 'online' and the amount of
259available memory will be increased.
260
261Currently, newly added memory is added as ZONE_NORMAL (for powerpc, ZONE_DMA).
262This may be changed in future.
263
264
265
266------------------------
2676. Logical memory remove
268------------------------
269
2706.1 Memory offline and ZONE_MOVABLE
271------------
272Memory offlining is more complicated than memory online. Because memory offline
273has to make the whole memory section be unused, memory offline can fail if
274the section includes memory which cannot be freed.
275
276In general, memory offline can use 2 techniques.
277
278(1) reclaim and free all memory in the section.
279(2) migrate all pages in the section.
280
281In the current implementation, Linux's memory offline uses method (2), freeing
282all pages in the section by page migration. But not all pages are
283migratable. Under current Linux, migratable pages are anonymous pages and
284page caches. For offlining a section by migration, the kernel has to guarantee
285that the section contains only migratable pages.
286
287Now, a boot option for making a section which consists of migratable pages is
288supported. By specifying "kernelcore=" or "movablecore=" boot option, you can
289create ZONE_MOVABLE...a zone which is just used for movable pages.
290(See also Documentation/kernel-parameters.txt)
291
292Assume the system has "TOTAL" amount of memory at boot time, this boot option
293creates ZONE_MOVABLE as following.
294
2951) When kernelcore=YYYY boot option is used,
296 Size of memory not for movable pages (not for offline) is YYYY.
297 Size of memory for movable pages (for offline) is TOTAL-YYYY.
298
2992) When movablecore=ZZZZ boot option is used,
300 Size of memory not for movable pages (not for offline) is TOTAL - ZZZZ.
301 Size of memory for movable pages (for offline) is ZZZZ.
302
303
304Note) Unfortunately, there is no information to show which section belongs
305to ZONE_MOVABLE. This is TBD.
306
307
3086.2. How to offline memory
309------------
310You can offline a section by using the same sysfs interface that was used in
311memory onlining.
312
313% echo offline > /sys/devices/system/memory/memoryXXX/state
314
315If offline succeeds, the state of the memory section is changed to be "offline".
316If it fails, some error core (like -EBUSY) will be returned by the kernel.
317Even if a section does not belong to ZONE_MOVABLE, you can try to offline it.
318If it doesn't contain 'unmovable' memory, you'll get success.
319
320A section under ZONE_MOVABLE is considered to be able to be offlined easily.
321But under some busy state, it may return -EBUSY. Even if a memory section
322cannot be offlined due to -EBUSY, you can retry offlining it and may be able to
323offline it (or not).
324(For example, a page is referred to by some kernel internal call and released
325 soon.)
326
327Consideration:
328Memory hotplug's design direction is to make the possibility of memory offlining
329higher and to guarantee unplugging memory under any situation. But it needs
330more work. Returning -EBUSY under some situation may be good because the user
331can decide to retry more or not by himself. Currently, memory offlining code
332does some amount of retry with 120 seconds timeout.
333
334-------------------------
3357. Physical memory remove
336-------------------------
337Need more implementation yet....
338 - Notification completion of remove works by OS to firmware.
339 - Guard from remove if not yet.
340
Yasunori Goto10020ca2007-10-21 16:41:36 -0700341--------------------------------
3428. Memory hotplug event notifier
343--------------------------------
Masanari Iidac94bed8e2012-04-10 00:22:13 +0900344Memory hotplug has event notifier. There are 6 types of notification.
Yasunori Goto10020ca2007-10-21 16:41:36 -0700345
346MEMORY_GOING_ONLINE
347 Generated before new memory becomes available in order to be able to
348 prepare subsystems to handle memory. The page allocator is still unable
349 to allocate from the new memory.
350
351MEMORY_CANCEL_ONLINE
352 Generated if MEMORY_GOING_ONLINE fails.
353
354MEMORY_ONLINE
Matt LaPlante19f59462009-04-27 15:06:31 +0200355 Generated when memory has successfully brought online. The callback may
Yasunori Goto10020ca2007-10-21 16:41:36 -0700356 allocate pages from the new memory.
357
358MEMORY_GOING_OFFLINE
359 Generated to begin the process of offlining memory. Allocations are no
360 longer possible from the memory but some of the memory to be offlined
361 is still in use. The callback can be used to free memory known to a
362 subsystem from the indicated memory section.
363
364MEMORY_CANCEL_OFFLINE
365 Generated if MEMORY_GOING_OFFLINE fails. Memory is available again from
366 the section that we attempted to offline.
367
368MEMORY_OFFLINE
369 Generated after offlining memory is complete.
370
371A callback routine can be registered by
372 hotplug_memory_notifier(callback_func, priority)
373
374The second argument of callback function (action) is event types of above.
375The third argument is passed by pointer of struct memory_notify.
376
377struct memory_notify {
378 unsigned long start_pfn;
379 unsigned long nr_pages;
Matt LaPlante19f59462009-04-27 15:06:31 +0200380 int status_change_nid;
Yasunori Goto10020ca2007-10-21 16:41:36 -0700381}
382
383start_pfn is start_pfn of online/offline memory.
384nr_pages is # of pages of online/offline memory.
385status_change_nid is set node id when N_HIGH_MEMORY of nodemask is (will be)
386set/clear. It means a new(memoryless) node gets new memory by online and a
387node loses all memory. If this is -1, then nodemask status is not changed.
388If status_changed_nid >= 0, callback should create/discard structures for the
389node if necessary.
390
Yasunori Goto6867c932007-08-10 13:00:59 -0700391--------------
Yasunori Goto10020ca2007-10-21 16:41:36 -07003929. Future Work
Yasunori Goto6867c932007-08-10 13:00:59 -0700393--------------
394 - allowing memory hot-add to ZONE_MOVABLE. maybe we need some switch like
395 sysctl or new control file.
396 - showing memory section and physical device relationship.
Yasunori Goto6867c932007-08-10 13:00:59 -0700397 - showing memory section is under ZONE_MOVABLE or not
398 - test and make it better memory offlining.
399 - support HugeTLB page migration and offlining.
400 - memmap removing at memory offline.
401 - physical remove memory.
402