diff options
Diffstat (limited to 'Documentation/admin-guide/mm')
| -rw-r--r-- | Documentation/admin-guide/mm/cma_debugfs.rst | 25 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/concepts.rst | 223 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/hugetlbpage.rst | 428 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/idle_page_tracking.rst | 121 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/index.rst | 40 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/ksm.rst | 189 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/memory-hotplug.rst | 444 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/nommu-mmap.rst | 283 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/numa_memory_policy.rst | 495 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/numaperf.rst | 178 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/pagemap.rst | 207 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/soft-dirty.rst | 47 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/transhuge.rst | 438 | ||||
| -rw-r--r-- | Documentation/admin-guide/mm/userfaultfd.rst | 293 |
14 files changed, 3411 insertions, 0 deletions
diff --git a/Documentation/admin-guide/mm/cma_debugfs.rst b/Documentation/admin-guide/mm/cma_debugfs.rst new file mode 100644 index 000000000..4e06ffabd --- /dev/null +++ b/Documentation/admin-guide/mm/cma_debugfs.rst @@ -0,0 +1,25 @@ +===================== +CMA Debugfs Interface +===================== + +The CMA debugfs interface is useful to retrieve basic information out of the +different CMA areas and to test allocation/release in each of the areas. + +Each CMA zone represents a directory under <debugfs>/cma/, indexed by the +kernel's CMA index. So the first CMA zone would be: + + <debugfs>/cma/cma-0 + +The structure of the files created under that directory is as follows: + + - [RO] base_pfn: The base PFN (Page Frame Number) of the zone. + - [RO] count: Amount of memory in the CMA area. + - [RO] order_per_bit: Order of pages represented by one bit. + - [RO] bitmap: The bitmap of page states in the zone. + - [WO] alloc: Allocate N pages from that CMA area. For example:: + + echo 5 > <debugfs>/cma/cma-2/alloc + +would try to allocate 5 pages from the cma-2 area. + + - [WO] free: Free N pages from that CMA area, similar to the above. diff --git a/Documentation/admin-guide/mm/concepts.rst b/Documentation/admin-guide/mm/concepts.rst new file mode 100644 index 000000000..fa0974fbe --- /dev/null +++ b/Documentation/admin-guide/mm/concepts.rst @@ -0,0 +1,223 @@ +.. _mm_concepts: + +================= +Concepts overview +================= + +The memory management in Linux is a complex system that evolved over the +years and included more and more functionality to support a variety of +systems from MMU-less microcontrollers to supercomputers. The memory +management for systems without an MMU is called ``nommu`` and it +definitely deserves a dedicated document, which hopefully will be +eventually written. Yet, although some of the concepts are the same, +here we assume that an MMU is available and a CPU can translate a virtual +address to a physical address. + +.. contents:: :local: + +Virtual Memory Primer +===================== + +The physical memory in a computer system is a limited resource and +even for systems that support memory hotplug there is a hard limit on +the amount of memory that can be installed. The physical memory is not +necessarily contiguous; it might be accessible as a set of distinct +address ranges. Besides, different CPU architectures, and even +different implementations of the same architecture have different views +of how these address ranges are defined. + +All this makes dealing directly with physical memory quite complex and +to avoid this complexity a concept of virtual memory was developed. + +The virtual memory abstracts the details of physical memory from the +application software, allows to keep only needed information in the +physical memory (demand paging) and provides a mechanism for the +protection and controlled sharing of data between processes. + +With virtual memory, each and every memory access uses a virtual +address. When the CPU decodes an instruction that reads (or +writes) from (or to) the system memory, it translates the `virtual` +address encoded in that instruction to a `physical` address that the +memory controller can understand. + +The physical system memory is divided into page frames, or pages. The +size of each page is architecture specific. Some architectures allow +selection of the page size from several supported values; this +selection is performed at the kernel build time by setting an +appropriate kernel configuration option. + +Each physical memory page can be mapped as one or more virtual +pages. These mappings are described by page tables that allow +translation from a virtual address used by programs to the physical +memory address. The page tables are organized hierarchically. + +The tables at the lowest level of the hierarchy contain physical +addresses of actual pages used by the software. The tables at higher +levels contain physical addresses of the pages belonging to the lower +levels. The pointer to the top level page table resides in a +register. When the CPU performs the address translation, it uses this +register to access the top level page table. The high bits of the +virtual address are used to index an entry in the top level page +table. That entry is then used to access the next level in the +hierarchy with the next bits of the virtual address as the index to +that level page table. The lowest bits in the virtual address define +the offset inside the actual page. + +Huge Pages +========== + +The address translation requires several memory accesses and memory +accesses are slow relatively to CPU speed. To avoid spending precious +processor cycles on the address translation, CPUs maintain a cache of +such translations called Translation Lookaside Buffer (or +TLB). Usually TLB is pretty scarce resource and applications with +large memory working set will experience performance hit because of +TLB misses. + +Many modern CPU architectures allow mapping of the memory pages +directly by the higher levels in the page table. For instance, on x86, +it is possible to map 2M and even 1G pages using entries in the second +and the third level page tables. In Linux such pages are called +`huge`. Usage of huge pages significantly reduces pressure on TLB, +improves TLB hit-rate and thus improves overall system performance. + +There are two mechanisms in Linux that enable mapping of the physical +memory with the huge pages. The first one is `HugeTLB filesystem`, or +hugetlbfs. It is a pseudo filesystem that uses RAM as its backing +store. For the files created in this filesystem the data resides in +the memory and mapped using huge pages. The hugetlbfs is described at +:ref:`Documentation/admin-guide/mm/hugetlbpage.rst <hugetlbpage>`. + +Another, more recent, mechanism that enables use of the huge pages is +called `Transparent HugePages`, or THP. Unlike the hugetlbfs that +requires users and/or system administrators to configure what parts of +the system memory should and can be mapped by the huge pages, THP +manages such mappings transparently to the user and hence the +name. See +:ref:`Documentation/admin-guide/mm/transhuge.rst <admin_guide_transhuge>` +for more details about THP. + +Zones +===== + +Often hardware poses restrictions on how different physical memory +ranges can be accessed. In some cases, devices cannot perform DMA to +all the addressable memory. In other cases, the size of the physical +memory exceeds the maximal addressable size of virtual memory and +special actions are required to access portions of the memory. Linux +groups memory pages into `zones` according to their possible +usage. For example, ZONE_DMA will contain memory that can be used by +devices for DMA, ZONE_HIGHMEM will contain memory that is not +permanently mapped into kernel's address space and ZONE_NORMAL will +contain normally addressed pages. + +The actual layout of the memory zones is hardware dependent as not all +architectures define all zones, and requirements for DMA are different +for different platforms. + +Nodes +===== + +Many multi-processor machines are NUMA - Non-Uniform Memory Access - +systems. In such systems the memory is arranged into banks that have +different access latency depending on the "distance" from the +processor. Each bank is referred to as a `node` and for each node Linux +constructs an independent memory management subsystem. A node has its +own set of zones, lists of free and used pages and various statistics +counters. You can find more details about NUMA in +:ref:`Documentation/vm/numa.rst <numa>` and in +:ref:`Documentation/admin-guide/mm/numa_memory_policy.rst <numa_memory_policy>`. + +Page cache +========== + +The physical memory is volatile and the common case for getting data +into the memory is to read it from files. Whenever a file is read, the +data is put into the `page cache` to avoid expensive disk access on +the subsequent reads. Similarly, when one writes to a file, the data +is placed in the page cache and eventually gets into the backing +storage device. The written pages are marked as `dirty` and when Linux +decides to reuse them for other purposes, it makes sure to synchronize +the file contents on the device with the updated data. + +Anonymous Memory +================ + +The `anonymous memory` or `anonymous mappings` represent memory that +is not backed by a filesystem. Such mappings are implicitly created +for program's stack and heap or by explicit calls to mmap(2) system +call. Usually, the anonymous mappings only define virtual memory areas +that the program is allowed to access. The read accesses will result +in creation of a page table entry that references a special physical +page filled with zeroes. When the program performs a write, a regular +physical page will be allocated to hold the written data. The page +will be marked dirty and if the kernel decides to repurpose it, +the dirty page will be swapped out. + +Reclaim +======= + +Throughout the system lifetime, a physical page can be used for storing +different types of data. It can be kernel internal data structures, +DMA'able buffers for device drivers use, data read from a filesystem, +memory allocated by user space processes etc. + +Depending on the page usage it is treated differently by the Linux +memory management. The pages that can be freed at any time, either +because they cache the data available elsewhere, for instance, on a +hard disk, or because they can be swapped out, again, to the hard +disk, are called `reclaimable`. The most notable categories of the +reclaimable pages are page cache and anonymous memory. + +In most cases, the pages holding internal kernel data and used as DMA +buffers cannot be repurposed, and they remain pinned until freed by +their user. Such pages are called `unreclaimable`. However, in certain +circumstances, even pages occupied with kernel data structures can be +reclaimed. For instance, in-memory caches of filesystem metadata can +be re-read from the storage device and therefore it is possible to +discard them from the main memory when system is under memory +pressure. + +The process of freeing the reclaimable physical memory pages and +repurposing them is called (surprise!) `reclaim`. Linux can reclaim +pages either asynchronously or synchronously, depending on the state +of the system. When the system is not loaded, most of the memory is free +and allocation requests will be satisfied immediately from the free +pages supply. As the load increases, the amount of the free pages goes +down and when it reaches a certain threshold (high watermark), an +allocation request will awaken the ``kswapd`` daemon. It will +asynchronously scan memory pages and either just free them if the data +they contain is available elsewhere, or evict to the backing storage +device (remember those dirty pages?). As memory usage increases even +more and reaches another threshold - min watermark - an allocation +will trigger `direct reclaim`. In this case allocation is stalled +until enough memory pages are reclaimed to satisfy the request. + +Compaction +========== + +As the system runs, tasks allocate and free the memory and it becomes +fragmented. Although with virtual memory it is possible to present +scattered physical pages as virtually contiguous range, sometimes it is +necessary to allocate large physically contiguous memory areas. Such +need may arise, for instance, when a device driver requires a large +buffer for DMA, or when THP allocates a huge page. Memory `compaction` +addresses the fragmentation issue. This mechanism moves occupied pages +from the lower part of a memory zone to free pages in the upper part +of the zone. When a compaction scan is finished free pages are grouped +together at the beginning of the zone and allocations of large +physically contiguous areas become possible. + +Like reclaim, the compaction may happen asynchronously in the ``kcompactd`` +daemon or synchronously as a result of a memory allocation request. + +OOM killer +========== + +It is possible that on a loaded machine memory will be exhausted and the +kernel will be unable to reclaim enough memory to continue to operate. In +order to save the rest of the system, it invokes the `OOM killer`. + +The `OOM killer` selects a task to sacrifice for the sake of the overall +system health. The selected task is killed in a hope that after it exits +enough memory will be freed to continue normal operation. diff --git a/Documentation/admin-guide/mm/hugetlbpage.rst b/Documentation/admin-guide/mm/hugetlbpage.rst new file mode 100644 index 000000000..f7b1c7462 --- /dev/null +++ b/Documentation/admin-guide/mm/hugetlbpage.rst @@ -0,0 +1,428 @@ +.. _hugetlbpage: + +============= +HugeTLB Pages +============= + +Overview +======== + +The intent of this file is to give a brief summary of hugetlbpage support in +the Linux kernel. This support is built on top of multiple page size support +that is provided by most modern architectures. For example, x86 CPUs normally +support 4K and 2M (1G if architecturally supported) page sizes, ia64 +architecture supports multiple page sizes 4K, 8K, 64K, 256K, 1M, 4M, 16M, +256M and ppc64 supports 4K and 16M. A TLB is a cache of virtual-to-physical +translations. Typically this is a very scarce resource on processor. +Operating systems try to make best use of limited number of TLB resources. +This optimization is more critical now as bigger and bigger physical memories +(several GBs) are more readily available. + +Users can use the huge page support in Linux kernel by either using the mmap +system call or standard SYSV shared memory system calls (shmget, shmat). + +First the Linux kernel needs to be built with the CONFIG_HUGETLBFS +(present under "File systems") and CONFIG_HUGETLB_PAGE (selected +automatically when CONFIG_HUGETLBFS is selected) configuration +options. + +The ``/proc/meminfo`` file provides information about the total number of +persistent hugetlb pages in the kernel's huge page pool. It also displays +default huge page size and information about the number of free, reserved +and surplus huge pages in the pool of huge pages of default size. +The huge page size is needed for generating the proper alignment and +size of the arguments to system calls that map huge page regions. + +The output of ``cat /proc/meminfo`` will include lines like:: + + HugePages_Total: uuu + HugePages_Free: vvv + HugePages_Rsvd: www + HugePages_Surp: xxx + Hugepagesize: yyy kB + Hugetlb: zzz kB + +where: + +HugePages_Total + is the size of the pool of huge pages. +HugePages_Free + is the number of huge pages in the pool that are not yet + allocated. +HugePages_Rsvd + is short for "reserved," and is the number of huge pages for + which a commitment to allocate from the pool has been made, + but no allocation has yet been made. Reserved huge pages + guarantee that an application will be able to allocate a + huge page from the pool of huge pages at fault time. +HugePages_Surp + is short for "surplus," and is the number of huge pages in + the pool above the value in ``/proc/sys/vm/nr_hugepages``. The + maximum number of surplus huge pages is controlled by + ``/proc/sys/vm/nr_overcommit_hugepages``. +Hugepagesize + is the default hugepage size (in Kb). +Hugetlb + is the total amount of memory (in kB), consumed by huge + pages of all sizes. + If huge pages of different sizes are in use, this number + will exceed HugePages_Total \* Hugepagesize. To get more + detailed information, please, refer to + ``/sys/kernel/mm/hugepages`` (described below). + + +``/proc/filesystems`` should also show a filesystem of type "hugetlbfs" +configured in the kernel. + +``/proc/sys/vm/nr_hugepages`` indicates the current number of "persistent" huge +pages in the kernel's huge page pool. "Persistent" huge pages will be +returned to the huge page pool when freed by a task. A user with root +privileges can dynamically allocate more or free some persistent huge pages +by increasing or decreasing the value of ``nr_hugepages``. + +Pages that are used as huge pages are reserved inside the kernel and cannot +be used for other purposes. Huge pages cannot be swapped out under +memory pressure. + +Once a number of huge pages have been pre-allocated to the kernel huge page +pool, a user with appropriate privilege can use either the mmap system call +or shared memory system calls to use the huge pages. See the discussion of +:ref:`Using Huge Pages <using_huge_pages>`, below. + +The administrator can allocate persistent huge pages on the kernel boot +command line by specifying the "hugepages=N" parameter, where 'N' = the +number of huge pages requested. This is the most reliable method of +allocating huge pages as memory has not yet become fragmented. + +Some platforms support multiple huge page sizes. To allocate huge pages +of a specific size, one must precede the huge pages boot command parameters +with a huge page size selection parameter "hugepagesz=<size>". <size> must +be specified in bytes with optional scale suffix [kKmMgG]. The default huge +page size may be selected with the "default_hugepagesz=<size>" boot parameter. + +Hugetlb boot command line parameter semantics + +hugepagesz + Specify a huge page size. Used in conjunction with hugepages + parameter to preallocate a number of huge pages of the specified + size. Hence, hugepagesz and hugepages are typically specified in + pairs such as:: + + hugepagesz=2M hugepages=512 + + hugepagesz can only be specified once on the command line for a + specific huge page size. Valid huge page sizes are architecture + dependent. +hugepages + Specify the number of huge pages to preallocate. This typically + follows a valid hugepagesz or default_hugepagesz parameter. However, + if hugepages is the first or only hugetlb command line parameter it + implicitly specifies the number of huge pages of default size to + allocate. If the number of huge pages of default size is implicitly + specified, it can not be overwritten by a hugepagesz,hugepages + parameter pair for the default size. + + For example, on an architecture with 2M default huge page size:: + + hugepages=256 hugepagesz=2M hugepages=512 + + will result in 256 2M huge pages being allocated and a warning message + indicating that the hugepages=512 parameter is ignored. If a hugepages + parameter is preceded by an invalid hugepagesz parameter, it will + be ignored. +default_hugepagesz + Specify the default huge page size. This parameter can + only be specified once on the command line. default_hugepagesz can + optionally be followed by the hugepages parameter to preallocate a + specific number of huge pages of default size. The number of default + sized huge pages to preallocate can also be implicitly specified as + mentioned in the hugepages section above. Therefore, on an + architecture with 2M default huge page size:: + + hugepages=256 + default_hugepagesz=2M hugepages=256 + hugepages=256 default_hugepagesz=2M + + will all result in 256 2M huge pages being allocated. Valid default + huge page size is architecture dependent. + +When multiple huge page sizes are supported, ``/proc/sys/vm/nr_hugepages`` +indicates the current number of pre-allocated huge pages of the default size. +Thus, one can use the following command to dynamically allocate/deallocate +default sized persistent huge pages:: + + echo 20 > /proc/sys/vm/nr_hugepages + +This command will try to adjust the number of default sized huge pages in the +huge page pool to 20, allocating or freeing huge pages, as required. + +On a NUMA platform, the kernel will attempt to distribute the huge page pool +over all the set of allowed nodes specified by the NUMA memory policy of the +task that modifies ``nr_hugepages``. The default for the allowed nodes--when the +task has default memory policy--is all on-line nodes with memory. Allowed +nodes with insufficient available, contiguous memory for a huge page will be +silently skipped when allocating persistent huge pages. See the +:ref:`discussion below <mem_policy_and_hp_alloc>` +of the interaction of task memory policy, cpusets and per node attributes +with the allocation and freeing of persistent huge pages. + +The success or failure of huge page allocation depends on the amount of +physically contiguous memory that is present in system at the time of the +allocation attempt. If the kernel is unable to allocate huge pages from +some nodes in a NUMA system, it will attempt to make up the difference by +allocating extra pages on other nodes with sufficient available contiguous +memory, if any. + +System administrators may want to put this command in one of the local rc +init files. This will enable the kernel to allocate huge pages early in +the boot process when the possibility of getting physical contiguous pages +is still very high. Administrators can verify the number of huge pages +actually allocated by checking the sysctl or meminfo. To check the per node +distribution of huge pages in a NUMA system, use:: + + cat /sys/devices/system/node/node*/meminfo | fgrep Huge + +``/proc/sys/vm/nr_overcommit_hugepages`` specifies how large the pool of +huge pages can grow, if more huge pages than ``/proc/sys/vm/nr_hugepages`` are +requested by applications. Writing any non-zero value into this file +indicates that the hugetlb subsystem is allowed to try to obtain that +number of "surplus" huge pages from the kernel's normal page pool, when the +persistent huge page pool is exhausted. As these surplus huge pages become +unused, they are freed back to the kernel's normal page pool. + +When increasing the huge page pool size via ``nr_hugepages``, any existing +surplus pages will first be promoted to persistent huge pages. Then, additional +huge pages will be allocated, if necessary and if possible, to fulfill +the new persistent huge page pool size. + +The administrator may shrink the pool of persistent huge pages for +the default huge page size by setting the ``nr_hugepages`` sysctl to a +smaller value. The kernel will attempt to balance the freeing of huge pages +across all nodes in the memory policy of the task modifying ``nr_hugepages``. +Any free huge pages on the selected nodes will be freed back to the kernel's +normal page pool. + +Caveat: Shrinking the persistent huge page pool via ``nr_hugepages`` such that +it becomes less than the number of huge pages in use will convert the balance +of the in-use huge pages to surplus huge pages. This will occur even if +the number of surplus pages would exceed the overcommit value. As long as +this condition holds--that is, until ``nr_hugepages+nr_overcommit_hugepages`` is +increased sufficiently, or the surplus huge pages go out of use and are freed-- +no more surplus huge pages will be allowed to be allocated. + +With support for multiple huge page pools at run-time available, much of +the huge page userspace interface in ``/proc/sys/vm`` has been duplicated in +sysfs. +The ``/proc`` interfaces discussed above have been retained for backwards +compatibility. The root huge page control directory in sysfs is:: + + /sys/kernel/mm/hugepages + +For each huge page size supported by the running kernel, a subdirectory +will exist, of the form:: + + hugepages-${size}kB + +Inside each of these directories, the same set of files will exist:: + + nr_hugepages + nr_hugepages_mempolicy + nr_overcommit_hugepages + free_hugepages + resv_hugepages + surplus_hugepages + +which function as described above for the default huge page-sized case. + +.. _mem_policy_and_hp_alloc: + +Interaction of Task Memory Policy with Huge Page Allocation/Freeing +=================================================================== + +Whether huge pages are allocated and freed via the ``/proc`` interface or +the ``/sysfs`` interface using the ``nr_hugepages_mempolicy`` attribute, the +NUMA nodes from which huge pages are allocated or freed are controlled by the +NUMA memory policy of the task that modifies the ``nr_hugepages_mempolicy`` +sysctl or attribute. When the ``nr_hugepages`` attribute is used, mempolicy +is ignored. + +The recommended method to allocate or free huge pages to/from the kernel +huge page pool, using the ``nr_hugepages`` example above, is:: + + numactl --interleave <node-list> echo 20 \ + >/proc/sys/vm/nr_hugepages_mempolicy + +or, more succinctly:: + + numactl -m <node-list> echo 20 >/proc/sys/vm/nr_hugepages_mempolicy + +This will allocate or free ``abs(20 - nr_hugepages)`` to or from the nodes +specified in <node-list>, depending on whether number of persistent huge pages +is initially less than or greater than 20, respectively. No huge pages will be +allocated nor freed on any node not included in the specified <node-list>. + +When adjusting the persistent hugepage count via ``nr_hugepages_mempolicy``, any +memory policy mode--bind, preferred, local or interleave--may be used. The +resulting effect on persistent huge page allocation is as follows: + +#. Regardless of mempolicy mode [see + :ref:`Documentation/admin-guide/mm/numa_memory_policy.rst <numa_memory_policy>`], + persistent huge pages will be distributed across the node or nodes + specified in the mempolicy as if "interleave" had been specified. + However, if a node in the policy does not contain sufficient contiguous + memory for a huge page, the allocation will not "fallback" to the nearest + neighbor node with sufficient contiguous memory. To do this would cause + undesirable imbalance in the distribution of the huge page pool, or + possibly, allocation of persistent huge pages on nodes not allowed by + the task's memory policy. + +#. One or more nodes may be specified with the bind or interleave policy. + If more than one node is specified with the preferred policy, only the + lowest numeric id will be used. Local policy will select the node where + the task is running at the time the nodes_allowed mask is constructed. + For local policy to be deterministic, the task must be bound to a cpu or + cpus in a single node. Otherwise, the task could be migrated to some + other node at any time after launch and the resulting node will be + indeterminate. Thus, local policy is not very useful for this purpose. + Any of the other mempolicy modes may be used to specify a single node. + +#. The nodes allowed mask will be derived from any non-default task mempolicy, + whether this policy was set explicitly by the task itself or one of its + ancestors, such as numactl. This means that if the task is invoked from a + shell with non-default policy, that policy will be used. One can specify a + node list of "all" with numactl --interleave or --membind [-m] to achieve + interleaving over all nodes in the system or cpuset. + +#. Any task mempolicy specified--e.g., using numactl--will be constrained by + the resource limits of any cpuset in which the task runs. Thus, there will + be no way for a task with non-default policy running in a cpuset with a + subset of the system nodes to allocate huge pages outside the cpuset + without first moving to a cpuset that contains all of the desired nodes. + +#. Boot-time huge page allocation attempts to distribute the requested number + of huge pages over all on-lines nodes with memory. + +Per Node Hugepages Attributes +============================= + +A subset of the contents of the root huge page control directory in sysfs, +described above, will be replicated under each the system device of each +NUMA node with memory in:: + + /sys/devices/system/node/node[0-9]*/hugepages/ + +Under this directory, the subdirectory for each supported huge page size +contains the following attribute files:: + + nr_hugepages + free_hugepages + surplus_hugepages + +The free\_' and surplus\_' attribute files are read-only. They return the number +of free and surplus [overcommitted] huge pages, respectively, on the parent +node. + +The ``nr_hugepages`` attribute returns the total number of huge pages on the +specified node. When this attribute is written, the number of persistent huge +pages on the parent node will be adjusted to the specified value, if sufficient +resources exist, regardless of the task's mempolicy or cpuset constraints. + +Note that the number of overcommit and reserve pages remain global quantities, +as we don't know until fault time, when the faulting task's mempolicy is +applied, from which node the huge page allocation will be attempted. + +.. _using_huge_pages: + +Using Huge Pages +================ + +If the user applications are going to request huge pages using mmap system +call, then it is required that system administrator mount a file system of +type hugetlbfs:: + + mount -t hugetlbfs \ + -o uid=<value>,gid=<value>,mode=<value>,pagesize=<value>,size=<value>,\ + min_size=<value>,nr_inodes=<value> none /mnt/huge + +This command mounts a (pseudo) filesystem of type hugetlbfs on the directory +``/mnt/huge``. Any file created on ``/mnt/huge`` uses huge pages. + +The ``uid`` and ``gid`` options sets the owner and group of the root of the +file system. By default the ``uid`` and ``gid`` of the current process +are taken. + +The ``mode`` option sets the mode of root of file system to value & 01777. +This value is given in octal. By default the value 0755 is picked. + +If the platform supports multiple huge page sizes, the ``pagesize`` option can +be used to specify the huge page size and associated pool. ``pagesize`` +is specified in bytes. If ``pagesize`` is not specified the platform's +default huge page size and associated pool will be used. + +The ``size`` option sets the maximum value of memory (huge pages) allowed +for that filesystem (``/mnt/huge``). The ``size`` option can be specified +in bytes, or as a percentage of the specified huge page pool (``nr_hugepages``). +The size is rounded down to HPAGE_SIZE boundary. + +The ``min_size`` option sets the minimum value of memory (huge pages) allowed +for the filesystem. ``min_size`` can be specified in the same way as ``size``, +either bytes or a percentage of the huge page pool. +At mount time, the number of huge pages specified by ``min_size`` are reserved +for use by the filesystem. +If there are not enough free huge pages available, the mount will fail. +As huge pages are allocated to the filesystem and freed, the reserve count +is adjusted so that the sum of allocated and reserved huge pages is always +at least ``min_size``. + +The option ``nr_inodes`` sets the maximum number of inodes that ``/mnt/huge`` +can use. + +If the ``size``, ``min_size`` or ``nr_inodes`` option is not provided on +command line then no limits are set. + +For ``pagesize``, ``size``, ``min_size`` and ``nr_inodes`` options, you can +use [G|g]/[M|m]/[K|k] to represent giga/mega/kilo. +For example, size=2K has the same meaning as size=2048. + +While read system calls are supported on files that reside on hugetlb +file systems, write system calls are not. + +Regular chown, chgrp, and chmod commands (with right permissions) could be +used to change the file attributes on hugetlbfs. + +Also, it is important to note that no such mount command is required if +applications are going to use only shmat/shmget system calls or mmap with +MAP_HUGETLB. For an example of how to use mmap with MAP_HUGETLB see +:ref:`map_hugetlb <map_hugetlb>` below. + +Users who wish to use hugetlb memory via shared memory segment should be +members of a supplementary group and system admin needs to configure that gid +into ``/proc/sys/vm/hugetlb_shm_group``. It is possible for same or different +applications to use any combination of mmaps and shm* calls, though the mount of +filesystem will be required for using mmap calls without MAP_HUGETLB. + +Syscalls that operate on memory backed by hugetlb pages only have their lengths +aligned to the native page size of the processor; they will normally fail with +errno set to EINVAL or exclude hugetlb pages that extend beyond the length if +not hugepage aligned. For example, munmap(2) will fail if memory is backed by +a hugetlb page and the length is smaller than the hugepage size. + + +Examples +======== + +.. _map_hugetlb: + +``map_hugetlb`` + see tools/testing/selftests/vm/map_hugetlb.c + +``hugepage-shm`` + see tools/testing/selftests/vm/hugepage-shm.c + +``hugepage-mmap`` + see tools/testing/selftests/vm/hugepage-mmap.c + +The `libhugetlbfs`_ library provides a wide range of userspace tools +to help with huge page usability, environment setup, and control. + +.. _libhugetlbfs: https://github.com/libhugetlbfs/libhugetlbfs diff --git a/Documentation/admin-guide/mm/idle_page_tracking.rst b/Documentation/admin-guide/mm/idle_page_tracking.rst new file mode 100644 index 000000000..df9394fb3 --- /dev/null +++ b/Documentation/admin-guide/mm/idle_page_tracking.rst @@ -0,0 +1,121 @@ +.. _idle_page_tracking: + +================== +Idle Page Tracking +================== + +Motivation +========== + +The idle page tracking feature allows to track which memory pages are being +accessed by a workload and which are idle. This information can be useful for +estimating the workload's working set size, which, in turn, can be taken into +account when configuring the workload parameters, setting memory cgroup limits, +or deciding where to place the workload within a compute cluster. + +It is enabled by CONFIG_IDLE_PAGE_TRACKING=y. + +.. _user_api: + +User API +======== + +The idle page tracking API is located at ``/sys/kernel/mm/page_idle``. +Currently, it consists of the only read-write file, +``/sys/kernel/mm/page_idle/bitmap``. + +The file implements a bitmap where each bit corresponds to a memory page. The +bitmap is represented by an array of 8-byte integers, and the page at PFN #i is +mapped to bit #i%64 of array element #i/64, byte order is native. When a bit is +set, the corresponding page is idle. + +A page is considered idle if it has not been accessed since it was marked idle +(for more details on what "accessed" actually means see the :ref:`Implementation +Details <impl_details>` section). +To mark a page idle one has to set the bit corresponding to +the page by writing to the file. A value written to the file is OR-ed with the +current bitmap value. + +Only accesses to user memory pages are tracked. These are pages mapped to a +process address space, page cache and buffer pages, swap cache pages. For other +page types (e.g. SLAB pages) an attempt to mark a page idle is silently ignored, +and hence such pages are never reported idle. + +For huge pages the idle flag is set only on the head page, so one has to read +``/proc/kpageflags`` in order to correctly count idle huge pages. + +Reading from or writing to ``/sys/kernel/mm/page_idle/bitmap`` will return +-EINVAL if you are not starting the read/write on an 8-byte boundary, or +if the size of the read/write is not a multiple of 8 bytes. Writing to +this file beyond max PFN will return -ENXIO. + +That said, in order to estimate the amount of pages that are not used by a +workload one should: + + 1. Mark all the workload's pages as idle by setting corresponding bits in + ``/sys/kernel/mm/page_idle/bitmap``. The pages can be found by reading + ``/proc/pid/pagemap`` if the workload is represented by a process, or by + filtering out alien pages using ``/proc/kpagecgroup`` in case the workload + is placed in a memory cgroup. + + 2. Wait until the workload accesses its working set. + + 3. Read ``/sys/kernel/mm/page_idle/bitmap`` and count the number of bits set. + If one wants to ignore certain types of pages, e.g. mlocked pages since they + are not reclaimable, he or she can filter them out using + ``/proc/kpageflags``. + +The page-types tool in the tools/vm directory can be used to assist in this. +If the tool is run initially with the appropriate option, it will mark all the +queried pages as idle. Subsequent runs of the tool can then show which pages have +their idle flag cleared in the interim. + +See :ref:`Documentation/admin-guide/mm/pagemap.rst <pagemap>` for more +information about ``/proc/pid/pagemap``, ``/proc/kpageflags``, and +``/proc/kpagecgroup``. + +.. _impl_details: + +Implementation Details +====================== + +The kernel internally keeps track of accesses to user memory pages in order to +reclaim unreferenced pages first on memory shortage conditions. A page is +considered referenced if it has been recently accessed via a process address +space, in which case one or more PTEs it is mapped to will have the Accessed bit +set, or marked accessed explicitly by the kernel (see mark_page_accessed()). The +latter happens when: + + - a userspace process reads or writes a page using a system call (e.g. read(2) + or write(2)) + + - a page that is used for storing filesystem buffers is read or written, + because a process needs filesystem metadata stored in it (e.g. lists a + directory tree) + + - a page is accessed by a device driver using get_user_pages() + +When a dirty page is written to swap or disk as a result of memory reclaim or +exceeding the dirty memory limit, it is not marked referenced. + +The idle memory tracking feature adds a new page flag, the Idle flag. This flag +is set manually, by writing to ``/sys/kernel/mm/page_idle/bitmap`` (see the +:ref:`User API <user_api>` +section), and cleared automatically whenever a page is referenced as defined +above. + +When a page is marked idle, the Accessed bit must be cleared in all PTEs it is +mapped to, otherwise we will not be able to detect accesses to the page coming +from a process address space. To avoid interference with the reclaimer, which, +as noted above, uses the Accessed bit to promote actively referenced pages, one +more page flag is introduced, the Young flag. When the PTE Accessed bit is +cleared as a result of setting or updating a page's Idle flag, the Young flag +is set on the page. The reclaimer treats the Young flag as an extra PTE +Accessed bit and therefore will consider such a page as referenced. + +Since the idle memory tracking feature is based on the memory reclaimer logic, +it only works with pages that are on an LRU list, other pages are silently +ignored. That means it will ignore a user memory page if it is isolated, but +since there are usually not many of them, it should not affect the overall +result noticeably. In order not to stall scanning of the idle page bitmap, +locked pages may be skipped too. diff --git a/Documentation/admin-guide/mm/index.rst b/Documentation/admin-guide/mm/index.rst new file mode 100644 index 000000000..cd727cfc1 --- /dev/null +++ b/Documentation/admin-guide/mm/index.rst @@ -0,0 +1,40 @@ +================= +Memory Management +================= + +Linux memory management subsystem is responsible, as the name implies, +for managing the memory in the system. This includes implemnetation of +virtual memory and demand paging, memory allocation both for kernel +internal structures and user space programms, mapping of files into +processes address space and many other cool things. + +Linux memory management is a complex system with many configurable +settings. Most of these settings are available via ``/proc`` +filesystem and can be quired and adjusted using ``sysctl``. These APIs +are described in Documentation/admin-guide/sysctl/vm.rst and in `man 5 proc`_. + +.. _man 5 proc: http://man7.org/linux/man-pages/man5/proc.5.html + +Linux memory management has its own jargon and if you are not yet +familiar with it, consider reading +:ref:`Documentation/admin-guide/mm/concepts.rst <mm_concepts>`. + +Here we document in detail how to interact with various mechanisms in +the Linux memory management. + +.. toctree:: + :maxdepth: 1 + + concepts + cma_debugfs + hugetlbpage + idle_page_tracking + ksm + memory-hotplug + nommu-mmap + numa_memory_policy + numaperf + pagemap + soft-dirty + transhuge + userfaultfd diff --git a/Documentation/admin-guide/mm/ksm.rst b/Documentation/admin-guide/mm/ksm.rst new file mode 100644 index 000000000..97d816791 --- /dev/null +++ b/Documentation/admin-guide/mm/ksm.rst @@ -0,0 +1,189 @@ +.. _admin_guide_ksm: + +======================= +Kernel Samepage Merging +======================= + +Overview +======== + +KSM is a memory-saving de-duplication feature, enabled by CONFIG_KSM=y, +added to the Linux kernel in 2.6.32. See ``mm/ksm.c`` for its implementation, +and http://lwn.net/Articles/306704/ and https://lwn.net/Articles/330589/ + +KSM was originally developed for use with KVM (where it was known as +Kernel Shared Memory), to fit more virtual machines into physical memory, +by sharing the data common between them. But it can be useful to any +application which generates many instances of the same data. + +The KSM daemon ksmd periodically scans those areas of user memory +which have been registered with it, looking for pages of identical +content which can be replaced by a single write-protected page (which +is automatically copied if a process later wants to update its +content). The amount of pages that KSM daemon scans in a single pass +and the time between the passes are configured using :ref:`sysfs +intraface <ksm_sysfs>` + +KSM only merges anonymous (private) pages, never pagecache (file) pages. +KSM's merged pages were originally locked into kernel memory, but can now +be swapped out just like other user pages (but sharing is broken when they +are swapped back in: ksmd must rediscover their identity and merge again). + +Controlling KSM with madvise +============================ + +KSM only operates on those areas of address space which an application +has advised to be likely candidates for merging, by using the madvise(2) +system call:: + + int madvise(addr, length, MADV_MERGEABLE) + +The app may call + +:: + + int madvise(addr, length, MADV_UNMERGEABLE) + +to cancel that advice and restore unshared pages: whereupon KSM +unmerges whatever it merged in that range. Note: this unmerging call +may suddenly require more memory than is available - possibly failing +with EAGAIN, but more probably arousing the Out-Of-Memory killer. + +If KSM is not configured into the running kernel, madvise MADV_MERGEABLE +and MADV_UNMERGEABLE simply fail with EINVAL. If the running kernel was +built with CONFIG_KSM=y, those calls will normally succeed: even if the +KSM daemon is not currently running, MADV_MERGEABLE still registers +the range for whenever the KSM daemon is started; even if the range +cannot contain any pages which KSM could actually merge; even if +MADV_UNMERGEABLE is applied to a range which was never MADV_MERGEABLE. + +If a region of memory must be split into at least one new MADV_MERGEABLE +or MADV_UNMERGEABLE region, the madvise may return ENOMEM if the process +will exceed ``vm.max_map_count`` (see Documentation/admin-guide/sysctl/vm.rst). + +Like other madvise calls, they are intended for use on mapped areas of +the user address space: they will report ENOMEM if the specified range +includes unmapped gaps (though working on the intervening mapped areas), +and might fail with EAGAIN if not enough memory for internal structures. + +Applications should be considerate in their use of MADV_MERGEABLE, +restricting its use to areas likely to benefit. KSM's scans may use a lot +of processing power: some installations will disable KSM for that reason. + +.. _ksm_sysfs: + +KSM daemon sysfs interface +========================== + +The KSM daemon is controlled by sysfs files in ``/sys/kernel/mm/ksm/``, +readable by all but writable only by root: + +pages_to_scan + how many pages to scan before ksmd goes to sleep + e.g. ``echo 100 > /sys/kernel/mm/ksm/pages_to_scan``. + + Default: 100 (chosen for demonstration purposes) + +sleep_millisecs + how many milliseconds ksmd should sleep before next scan + e.g. ``echo 20 > /sys/kernel/mm/ksm/sleep_millisecs`` + + Default: 20 (chosen for demonstration purposes) + +merge_across_nodes + specifies if pages from different NUMA nodes can be merged. + When set to 0, ksm merges only pages which physically reside + in the memory area of same NUMA node. That brings lower + latency to access of shared pages. Systems with more nodes, at + significant NUMA distances, are likely to benefit from the + lower latency of setting 0. Smaller systems, which need to + minimize memory usage, are likely to benefit from the greater + sharing of setting 1 (default). You may wish to compare how + your system performs under each setting, before deciding on + which to use. ``merge_across_nodes`` setting can be changed only + when there are no ksm shared pages in the system: set run 2 to + unmerge pages first, then to 1 after changing + ``merge_across_nodes``, to remerge according to the new setting. + + Default: 1 (merging across nodes as in earlier releases) + +run + * set to 0 to stop ksmd from running but keep merged pages, + * set to 1 to run ksmd e.g. ``echo 1 > /sys/kernel/mm/ksm/run``, + * set to 2 to stop ksmd and unmerge all pages currently merged, but + leave mergeable areas registered for next run. + + Default: 0 (must be changed to 1 to activate KSM, except if + CONFIG_SYSFS is disabled) + +use_zero_pages + specifies whether empty pages (i.e. allocated pages that only + contain zeroes) should be treated specially. When set to 1, + empty pages are merged with the kernel zero page(s) instead of + with each other as it would happen normally. This can improve + the performance on architectures with coloured zero pages, + depending on the workload. Care should be taken when enabling + this setting, as it can potentially degrade the performance of + KSM for some workloads, for example if the checksums of pages + candidate for merging match the checksum of an empty + page. This setting can be changed at any time, it is only + effective for pages merged after the change. + + Default: 0 (normal KSM behaviour as in earlier releases) + +max_page_sharing + Maximum sharing allowed for each KSM page. This enforces a + deduplication limit to avoid high latency for virtual memory + operations that involve traversal of the virtual mappings that + share the KSM page. The minimum value is 2 as a newly created + KSM page will have at least two sharers. The higher this value + the faster KSM will merge the memory and the higher the + deduplication factor will be, but the slower the worst case + virtual mappings traversal could be for any given KSM + page. Slowing down this traversal means there will be higher + latency for certain virtual memory operations happening during + swapping, compaction, NUMA balancing and page migration, in + turn decreasing responsiveness for the caller of those virtual + memory operations. The scheduler latency of other tasks not + involved with the VM operations doing the virtual mappings + traversal is not affected by this parameter as these + traversals are always schedule friendly themselves. + +stable_node_chains_prune_millisecs + specifies how frequently KSM checks the metadata of the pages + that hit the deduplication limit for stale information. + Smaller milllisecs values will free up the KSM metadata with + lower latency, but they will make ksmd use more CPU during the + scan. It's a noop if not a single KSM page hit the + ``max_page_sharing`` yet. + +The effectiveness of KSM and MADV_MERGEABLE is shown in ``/sys/kernel/mm/ksm/``: + +pages_shared + how many shared pages are being used +pages_sharing + how many more sites are sharing them i.e. how much saved +pages_unshared + how many pages unique but repeatedly checked for merging +pages_volatile + how many pages changing too fast to be placed in a tree +full_scans + how many times all mergeable areas have been scanned +stable_node_chains + the number of KSM pages that hit the ``max_page_sharing`` limit +stable_node_dups + number of duplicated KSM pages + +A high ratio of ``pages_sharing`` to ``pages_shared`` indicates good +sharing, but a high ratio of ``pages_unshared`` to ``pages_sharing`` +indicates wasted effort. ``pages_volatile`` embraces several +different kinds of activity, but a high proportion there would also +indicate poor use of madvise MADV_MERGEABLE. + +The maximum possible ``pages_sharing/pages_shared`` ratio is limited by the +``max_page_sharing`` tunable. To increase the ratio ``max_page_sharing`` must +be increased accordingly. + +-- +Izik Eidus, +Hugh Dickins, 17 Nov 2009 diff --git a/Documentation/admin-guide/mm/memory-hotplug.rst b/Documentation/admin-guide/mm/memory-hotplug.rst new file mode 100644 index 000000000..245739f55 --- /dev/null +++ b/Documentation/admin-guide/mm/memory-hotplug.rst @@ -0,0 +1,444 @@ +.. _admin_guide_memory_hotplug: + +============== +Memory Hotplug +============== + +:Created: Jul 28 2007 +:Updated: Add some details about locking internals: Aug 20 2018 + +This document is about memory hotplug including how-to-use and current status. +Because Memory Hotplug is still under development, contents of this text will +be changed often. + +.. contents:: :local: + +.. note:: + + (1) x86_64's has special implementation for memory hotplug. + This text does not describe it. + (2) This text assumes that sysfs is mounted at ``/sys``. + + +Introduction +============ + +Purpose of memory hotplug +------------------------- + +Memory Hotplug allows users to increase/decrease the amount of memory. +Generally, there are two purposes. + +(A) For changing the amount of memory. + This is to allow a feature like capacity on demand. +(B) For installing/removing DIMMs or NUMA-nodes physically. + This is to exchange DIMMs/NUMA-nodes, reduce power consumption, etc. + +(A) is required by highly virtualized environments and (B) is required by +hardware which supports memory power management. + +Linux memory hotplug is designed for both purpose. + +Phases of memory hotplug +------------------------ + +There are 2 phases in Memory Hotplug: + + 1) Physical Memory Hotplug phase + 2) Logical Memory Hotplug phase. + +The First phase is to communicate hardware/firmware and make/erase +environment for hotplugged memory. Basically, this phase is necessary +for the purpose (B), but this is good phase for communication between +highly virtualized environments too. + +When memory is hotplugged, the kernel recognizes new memory, makes new memory +management tables, and makes sysfs files for new memory's operation. + +If firmware supports notification of connection of new memory to OS, +this phase is triggered automatically. ACPI can notify this event. If not, +"probe" operation by system administration is used instead. +(see :ref:`memory_hotplug_physical_mem`). + +Logical Memory Hotplug phase is to change memory state into +available/unavailable for users. Amount of memory from user's view is +changed by this phase. The kernel makes all memory in it as free pages +when a memory range is available. + +In this document, this phase is described as online/offline. + +Logical Memory Hotplug phase is triggered by write of sysfs file by system +administrator. For the hot-add case, it must be executed after Physical Hotplug +phase by hand. +(However, if you writes udev's hotplug scripts for memory hotplug, these +phases can be execute in seamless way.) + +Unit of Memory online/offline operation +--------------------------------------- + +Memory hotplug uses SPARSEMEM memory model which allows memory to be divided +into chunks of the same size. These chunks are called "sections". The size of +a memory section is architecture dependent. For example, power uses 16MiB, ia64 +uses 1GiB. + +Memory sections are combined into chunks referred to as "memory blocks". The +size of a memory block is architecture dependent and represents the logical +unit upon which memory online/offline operations are to be performed. The +default size of a memory block is the same as memory section size unless an +architecture specifies otherwise. (see :ref:`memory_hotplug_sysfs_files`.) + +To determine the size (in bytes) of a memory block please read this file:: + + /sys/devices/system/memory/block_size_bytes + +Kernel Configuration +==================== + +To use memory hotplug feature, kernel must be compiled with following +config options. + +- For all memory hotplug: + - Memory model -> Sparse Memory (``CONFIG_SPARSEMEM``) + - Allow for memory hot-add (``CONFIG_MEMORY_HOTPLUG``) + +- To enable memory removal, the following are also necessary: + - Allow for memory hot remove (``CONFIG_MEMORY_HOTREMOVE``) + - Page Migration (``CONFIG_MIGRATION``) + +- For ACPI memory hotplug, the following are also necessary: + - Memory hotplug (under ACPI Support menu) (``CONFIG_ACPI_HOTPLUG_MEMORY``) + - This option can be kernel module. + +- As a related configuration, if your box has a feature of NUMA-node hotplug + via ACPI, then this option is necessary too. + + - ACPI0004,PNP0A05 and PNP0A06 Container Driver (under ACPI Support menu) + (``CONFIG_ACPI_CONTAINER``). + + This option can be kernel module too. + + +.. _memory_hotplug_sysfs_files: + +sysfs files for memory hotplug +============================== + +All memory blocks have their device information in sysfs. Each memory block +is described under ``/sys/devices/system/memory`` as:: + + /sys/devices/system/memory/memoryXXX + +where XXX is the memory block id. + +For the memory block covered by the sysfs directory. It is expected that all +memory sections in this range are present and no memory holes exist in the +range. Currently there is no way to determine if there is a memory hole, but +the existence of one should not affect the hotplug capabilities of the memory +block. + +For example, assume 1GiB memory block size. A device for a memory starting at +0x100000000 is ``/sys/device/system/memory/memory4``:: + + (0x100000000 / 1Gib = 4) + +This device covers address range [0x100000000 ... 0x140000000) + +Under each memory block, you can see 5 files: + +- ``/sys/devices/system/memory/memoryXXX/phys_index`` +- ``/sys/devices/system/memory/memoryXXX/phys_device`` +- ``/sys/devices/system/memory/memoryXXX/state`` +- ``/sys/devices/system/memory/memoryXXX/removable`` +- ``/sys/devices/system/memory/memoryXXX/valid_zones`` + +=================== ============================================================ +``phys_index`` read-only and contains memory block id, same as XXX. +``state`` read-write + + - at read: contains online/offline state of memory. + - at write: user can specify "online_kernel", + + "online_movable", "online", "offline" command + which will be performed on all sections in the block. +``phys_device`` read-only: legacy interface only ever used on s390x to + expose the covered storage increment. +``removable`` read-only: contains an integer value indicating + whether the memory block is removable or not + removable. A value of 1 indicates that the memory + block is removable and a value of 0 indicates that + it is not removable. A memory block is removable only if + every section in the block is removable. +``valid_zones`` read-only: designed to show which zones this memory block + can be onlined to. + + The first column shows it`s default zone. + + "memory6/valid_zones: Normal Movable" shows this memoryblock + can be onlined to ZONE_NORMAL by default and to ZONE_MOVABLE + by online_movable. + + "memory7/valid_zones: Movable Normal" shows this memoryblock + can be onlined to ZONE_MOVABLE by default and to ZONE_NORMAL + by online_kernel. +=================== ============================================================ + +.. note:: + + These directories/files appear after physical memory hotplug phase. + +If CONFIG_NUMA is enabled the memoryXXX/ directories can also be accessed +via symbolic links located in the ``/sys/devices/system/node/node*`` directories. + +For example:: + + /sys/devices/system/node/node0/memory9 -> ../../memory/memory9 + +A backlink will also be created:: + + /sys/devices/system/memory/memory9/node0 -> ../../node/node0 + +.. _memory_hotplug_physical_mem: + +Physical memory hot-add phase +============================= + +Hardware(Firmware) Support +-------------------------- + +On x86_64/ia64 platform, memory hotplug by ACPI is supported. + +In general, the firmware (ACPI) which supports memory hotplug defines +memory class object of _HID "PNP0C80". When a notify is asserted to PNP0C80, +Linux's ACPI handler does hot-add memory to the system and calls a hotplug udev +script. This will be done automatically. + +But scripts for memory hotplug are not contained in generic udev package(now). +You may have to write it by yourself or online/offline memory by hand. +Please see :ref:`memory_hotplug_how_to_online_memory` and +:ref:`memory_hotplug_how_to_offline_memory`. + +If firmware supports NUMA-node hotplug, and defines an object _HID "ACPI0004", +"PNP0A05", or "PNP0A06", notification is asserted to it, and ACPI handler +calls hotplug code for all of objects which are defined in it. +If memory device is found, memory hotplug code will be called. + +Notify memory hot-add event by hand +----------------------------------- + +On some architectures, the firmware may not notify the kernel of a memory +hotplug event. Therefore, the memory "probe" interface is supported to +explicitly notify the kernel. This interface depends on +CONFIG_ARCH_MEMORY_PROBE and can be configured on powerpc, sh, and x86 +if hotplug is supported, although for x86 this should be handled by ACPI +notification. + +Probe interface is located at:: + + /sys/devices/system/memory/probe + +You can tell the physical address of new memory to the kernel by:: + + % echo start_address_of_new_memory > /sys/devices/system/memory/probe + +Then, [start_address_of_new_memory, start_address_of_new_memory + +memory_block_size] memory range is hot-added. In this case, hotplug script is +not called (in current implementation). You'll have to online memory by +yourself. Please see :ref:`memory_hotplug_how_to_online_memory`. + +Logical Memory hot-add phase +============================ + +State of memory +--------------- + +To see (online/offline) state of a memory block, read 'state' file:: + + % cat /sys/device/system/memory/memoryXXX/state + + +- If the memory block is online, you'll read "online". +- If the memory block is offline, you'll read "offline". + + +.. _memory_hotplug_how_to_online_memory: + +How to online memory +-------------------- + +When the memory is hot-added, the kernel decides whether or not to "online" +it according to the policy which can be read from "auto_online_blocks" file:: + + % cat /sys/devices/system/memory/auto_online_blocks + +The default depends on the CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config +option. If it is disabled the default is "offline" which means the newly added +memory is not in a ready-to-use state and you have to "online" the newly added +memory blocks manually. Automatic onlining can be requested by writing "online" +to "auto_online_blocks" file:: + + % echo online > /sys/devices/system/memory/auto_online_blocks + +This sets a global policy and impacts all memory blocks that will subsequently +be hotplugged. Currently offline blocks keep their state. It is possible, under +certain circumstances, that some memory blocks will be added but will fail to +online. User space tools can check their "state" files +(``/sys/devices/system/memory/memoryXXX/state``) and try to online them manually. + +If the automatic onlining wasn't requested, failed, or some memory block was +offlined it is possible to change the individual block's state by writing to the +"state" file:: + + % echo online > /sys/devices/system/memory/memoryXXX/state + +This onlining will not change the ZONE type of the target memory block, +If the memory block doesn't belong to any zone an appropriate kernel zone +(usually ZONE_NORMAL) will be used unless movable_node kernel command line +option is specified when ZONE_MOVABLE will be used. + +You can explicitly request to associate it with ZONE_MOVABLE by:: + + % echo online_movable > /sys/devices/system/memory/memoryXXX/state + +.. note:: current limit: this memory block must be adjacent to ZONE_MOVABLE + +Or you can explicitly request a kernel zone (usually ZONE_NORMAL) by:: + + % echo online_kernel > /sys/devices/system/memory/memoryXXX/state + +.. note:: current limit: this memory block must be adjacent to ZONE_NORMAL + +An explicit zone onlining can fail (e.g. when the range is already within +and existing and incompatible zone already). + +After this, memory block XXX's state will be 'online' and the amount of +available memory will be increased. + +This may be changed in future. + +Logical memory remove +===================== + +Memory offline and ZONE_MOVABLE +------------------------------- + +Memory offlining is more complicated than memory online. Because memory offline +has to make the whole memory block be unused, memory offline can fail if +the memory block includes memory which cannot be freed. + +In general, memory offline can use 2 techniques. + +(1) reclaim and free all memory in the memory block. +(2) migrate all pages in the memory block. + +In the current implementation, Linux's memory offline uses method (2), freeing +all pages in the memory block by page migration. But not all pages are +migratable. Under current Linux, migratable pages are anonymous pages and +page caches. For offlining a memory block by migration, the kernel has to +guarantee that the memory block contains only migratable pages. + +Now, a boot option for making a memory block which consists of migratable pages +is supported. By specifying "kernelcore=" or "movablecore=" boot option, you can +create ZONE_MOVABLE...a zone which is just used for movable pages. +(See also Documentation/admin-guide/kernel-parameters.rst) + +Assume the system has "TOTAL" amount of memory at boot time, this boot option +creates ZONE_MOVABLE as following. + +1) When kernelcore=YYYY boot option is used, + Size of memory not for movable pages (not for offline) is YYYY. + Size of memory for movable pages (for offline) is TOTAL-YYYY. + +2) When movablecore=ZZZZ boot option is used, + Size of memory not for movable pages (not for offline) is TOTAL - ZZZZ. + Size of memory for movable pages (for offline) is ZZZZ. + +.. note:: + + Unfortunately, there is no information to show which memory block belongs + to ZONE_MOVABLE. This is TBD. + +.. _memory_hotplug_how_to_offline_memory: + +How to offline memory +--------------------- + +You can offline a memory block by using the same sysfs interface that was used +in memory onlining:: + + % echo offline > /sys/devices/system/memory/memoryXXX/state + +If offline succeeds, the state of the memory block is changed to be "offline". +If it fails, some error core (like -EBUSY) will be returned by the kernel. +Even if a memory block does not belong to ZONE_MOVABLE, you can try to offline +it. If it doesn't contain 'unmovable' memory, you'll get success. + +A memory block under ZONE_MOVABLE is considered to be able to be offlined +easily. But under some busy state, it may return -EBUSY. Even if a memory +block cannot be offlined due to -EBUSY, you can retry offlining it and may be +able to offline it (or not). (For example, a page is referred to by some kernel +internal call and released soon.) + +Consideration: + Memory hotplug's design direction is to make the possibility of memory + offlining higher and to guarantee unplugging memory under any situation. But + it needs more work. Returning -EBUSY under some situation may be good because + the user can decide to retry more or not by himself. Currently, memory + offlining code does some amount of retry with 120 seconds timeout. + +Physical memory remove +====================== + +Need more implementation yet.... + - Notification completion of remove works by OS to firmware. + - Guard from remove if not yet. + + +Locking Internals +================= + +When adding/removing memory that uses memory block devices (i.e. ordinary RAM), +the device_hotplug_lock should be held to: + +- synchronize against online/offline requests (e.g. via sysfs). This way, memory + block devices can only be accessed (.online/.state attributes) by user + space once memory has been fully added. And when removing memory, we + know nobody is in critical sections. +- synchronize against CPU hotplug and similar (e.g. relevant for ACPI and PPC) + +Especially, there is a possible lock inversion that is avoided using +device_hotplug_lock when adding memory and user space tries to online that +memory faster than expected: + +- device_online() will first take the device_lock(), followed by + mem_hotplug_lock +- add_memory_resource() will first take the mem_hotplug_lock, followed by + the device_lock() (while creating the devices, during bus_add_device()). + +As the device is visible to user space before taking the device_lock(), this +can result in a lock inversion. + +onlining/offlining of memory should be done via device_online()/ +device_offline() - to make sure it is properly synchronized to actions +via sysfs. Holding device_hotplug_lock is advised (to e.g. protect online_type) + +When adding/removing/onlining/offlining memory or adding/removing +heterogeneous/device memory, we should always hold the mem_hotplug_lock in +write mode to serialise memory hotplug (e.g. access to global/zone +variables). + +In addition, mem_hotplug_lock (in contrast to device_hotplug_lock) in read +mode allows for a quite efficient get_online_mems/put_online_mems +implementation, so code accessing memory can protect from that memory +vanishing. + + +Future Work +=========== + + - allowing memory hot-add to ZONE_MOVABLE. maybe we need some switch like + sysctl or new control file. + - showing memory block and physical device relationship. + - test and make it better memory offlining. + - support HugeTLB page migration and offlining. + - memmap removing at memory offline. + - physical remove memory. diff --git a/Documentation/admin-guide/mm/nommu-mmap.rst b/Documentation/admin-guide/mm/nommu-mmap.rst new file mode 100644 index 000000000..530fed08d --- /dev/null +++ b/Documentation/admin-guide/mm/nommu-mmap.rst @@ -0,0 +1,283 @@ +============================= +No-MMU memory mapping support +============================= + +The kernel has limited support for memory mapping under no-MMU conditions, such +as are used in uClinux environments. From the userspace point of view, memory +mapping is made use of in conjunction with the mmap() system call, the shmat() +call and the execve() system call. From the kernel's point of view, execve() +mapping is actually performed by the binfmt drivers, which call back into the +mmap() routines to do the actual work. + +Memory mapping behaviour also involves the way fork(), vfork(), clone() and +ptrace() work. Under uClinux there is no fork(), and clone() must be supplied +the CLONE_VM flag. + +The behaviour is similar between the MMU and no-MMU cases, but not identical; +and it's also much more restricted in the latter case: + + (#) Anonymous mapping, MAP_PRIVATE + + In the MMU case: VM regions backed by arbitrary pages; copy-on-write + across fork. + + In the no-MMU case: VM regions backed by arbitrary contiguous runs of + pages. + + (#) Anonymous mapping, MAP_SHARED + + These behave very much like private mappings, except that they're + shared across fork() or clone() without CLONE_VM in the MMU case. Since + the no-MMU case doesn't support these, behaviour is identical to + MAP_PRIVATE there. + + (#) File, MAP_PRIVATE, PROT_READ / PROT_EXEC, !PROT_WRITE + + In the MMU case: VM regions backed by pages read from file; changes to + the underlying file are reflected in the mapping; copied across fork. + + In the no-MMU case: + + - If one exists, the kernel will re-use an existing mapping to the + same segment of the same file if that has compatible permissions, + even if this was created by another process. + + - If possible, the file mapping will be directly on the backing device + if the backing device has the NOMMU_MAP_DIRECT capability and + appropriate mapping protection capabilities. Ramfs, romfs, cramfs + and mtd might all permit this. + + - If the backing device can't or won't permit direct sharing, + but does have the NOMMU_MAP_COPY capability, then a copy of the + appropriate bit of the file will be read into a contiguous bit of + memory and any extraneous space beyond the EOF will be cleared + + - Writes to the file do not affect the mapping; writes to the mapping + are visible in other processes (no MMU protection), but should not + happen. + + (#) File, MAP_PRIVATE, PROT_READ / PROT_EXEC, PROT_WRITE + + In the MMU case: like the non-PROT_WRITE case, except that the pages in + question get copied before the write actually happens. From that point + on writes to the file underneath that page no longer get reflected into + the mapping's backing pages. The page is then backed by swap instead. + + In the no-MMU case: works much like the non-PROT_WRITE case, except + that a copy is always taken and never shared. + + (#) Regular file / blockdev, MAP_SHARED, PROT_READ / PROT_EXEC / PROT_WRITE + + In the MMU case: VM regions backed by pages read from file; changes to + pages written back to file; writes to file reflected into pages backing + mapping; shared across fork. + + In the no-MMU case: not supported. + + (#) Memory backed regular file, MAP_SHARED, PROT_READ / PROT_EXEC / PROT_WRITE + + In the MMU case: As for ordinary regular files. + + In the no-MMU case: The filesystem providing the memory-backed file + (such as ramfs or tmpfs) may choose to honour an open, truncate, mmap + sequence by providing a contiguous sequence of pages to map. In that + case, a shared-writable memory mapping will be possible. It will work + as for the MMU case. If the filesystem does not provide any such + support, then the mapping request will be denied. + + (#) Memory backed blockdev, MAP_SHARED, PROT_READ / PROT_EXEC / PROT_WRITE + + In the MMU case: As for ordinary regular files. + + In the no-MMU case: As for memory backed regular files, but the + blockdev must be able to provide a contiguous run of pages without + truncate being called. The ramdisk driver could do this if it allocated + all its memory as a contiguous array upfront. + + (#) Memory backed chardev, MAP_SHARED, PROT_READ / PROT_EXEC / PROT_WRITE + + In the MMU case: As for ordinary regular files. + + In the no-MMU case: The character device driver may choose to honour + the mmap() by providing direct access to the underlying device if it + provides memory or quasi-memory that can be accessed directly. Examples + of such are frame buffers and flash devices. If the driver does not + provide any such support, then the mapping request will be denied. + + +Further notes on no-MMU MMAP +============================ + + (#) A request for a private mapping of a file may return a buffer that is not + page-aligned. This is because XIP may take place, and the data may not be + paged aligned in the backing store. + + (#) A request for an anonymous mapping will always be page aligned. If + possible the size of the request should be a power of two otherwise some + of the space may be wasted as the kernel must allocate a power-of-2 + granule but will only discard the excess if appropriately configured as + this has an effect on fragmentation. + + (#) The memory allocated by a request for an anonymous mapping will normally + be cleared by the kernel before being returned in accordance with the + Linux man pages (ver 2.22 or later). + + In the MMU case this can be achieved with reasonable performance as + regions are backed by virtual pages, with the contents only being mapped + to cleared physical pages when a write happens on that specific page + (prior to which, the pages are effectively mapped to the global zero page + from which reads can take place). This spreads out the time it takes to + initialize the contents of a page - depending on the write-usage of the + mapping. + + In the no-MMU case, however, anonymous mappings are backed by physical + pages, and the entire map is cleared at allocation time. This can cause + significant delays during a userspace malloc() as the C library does an + anonymous mapping and the kernel then does a memset for the entire map. + + However, for memory that isn't required to be precleared - such as that + returned by malloc() - mmap() can take a MAP_UNINITIALIZED flag to + indicate to the kernel that it shouldn't bother clearing the memory before + returning it. Note that CONFIG_MMAP_ALLOW_UNINITIALIZED must be enabled + to permit this, otherwise the flag will be ignored. + + uClibc uses this to speed up malloc(), and the ELF-FDPIC binfmt uses this + to allocate the brk and stack region. + + (#) A list of all the private copy and anonymous mappings on the system is + visible through /proc/maps in no-MMU mode. + + (#) A list of all the mappings in use by a process is visible through + /proc/<pid>/maps in no-MMU mode. + + (#) Supplying MAP_FIXED or a requesting a particular mapping address will + result in an error. + + (#) Files mapped privately usually have to have a read method provided by the + driver or filesystem so that the contents can be read into the memory + allocated if mmap() chooses not to map the backing device directly. An + error will result if they don't. This is most likely to be encountered + with character device files, pipes, fifos and sockets. + + +Interprocess shared memory +========================== + +Both SYSV IPC SHM shared memory and POSIX shared memory is supported in NOMMU +mode. The former through the usual mechanism, the latter through files created +on ramfs or tmpfs mounts. + + +Futexes +======= + +Futexes are supported in NOMMU mode if the arch supports them. An error will +be given if an address passed to the futex system call lies outside the +mappings made by a process or if the mapping in which the address lies does not +support futexes (such as an I/O chardev mapping). + + +No-MMU mremap +============= + +The mremap() function is partially supported. It may change the size of a +mapping, and may move it [#]_ if MREMAP_MAYMOVE is specified and if the new size +of the mapping exceeds the size of the slab object currently occupied by the +memory to which the mapping refers, or if a smaller slab object could be used. + +MREMAP_FIXED is not supported, though it is ignored if there's no change of +address and the object does not need to be moved. + +Shared mappings may not be moved. Shareable mappings may not be moved either, +even if they are not currently shared. + +The mremap() function must be given an exact match for base address and size of +a previously mapped object. It may not be used to create holes in existing +mappings, move parts of existing mappings or resize parts of mappings. It must +act on a complete mapping. + +.. [#] Not currently supported. + + +Providing shareable character device support +============================================ + +To provide shareable character device support, a driver must provide a +file->f_op->get_unmapped_area() operation. The mmap() routines will call this +to get a proposed address for the mapping. This may return an error if it +doesn't wish to honour the mapping because it's too long, at a weird offset, +under some unsupported combination of flags or whatever. + +The driver should also provide backing device information with capabilities set +to indicate the permitted types of mapping on such devices. The default is +assumed to be readable and writable, not executable, and only shareable +directly (can't be copied). + +The file->f_op->mmap() operation will be called to actually inaugurate the +mapping. It can be rejected at that point. Returning the ENOSYS error will +cause the mapping to be copied instead if NOMMU_MAP_COPY is specified. + +The vm_ops->close() routine will be invoked when the last mapping on a chardev +is removed. An existing mapping will be shared, partially or not, if possible +without notifying the driver. + +It is permitted also for the file->f_op->get_unmapped_area() operation to +return -ENOSYS. This will be taken to mean that this operation just doesn't +want to handle it, despite the fact it's got an operation. For instance, it +might try directing the call to a secondary driver which turns out not to +implement it. Such is the case for the framebuffer driver which attempts to +direct the call to the device-specific driver. Under such circumstances, the +mapping request will be rejected if NOMMU_MAP_COPY is not specified, and a +copy mapped otherwise. + +.. important:: + + Some types of device may present a different appearance to anyone + looking at them in certain modes. Flash chips can be like this; for + instance if they're in programming or erase mode, you might see the + status reflected in the mapping, instead of the data. + + In such a case, care must be taken lest userspace see a shared or a + private mapping showing such information when the driver is busy + controlling the device. Remember especially: private executable + mappings may still be mapped directly off the device under some + circumstances! + + +Providing shareable memory-backed file support +============================================== + +Provision of shared mappings on memory backed files is similar to the provision +of support for shared mapped character devices. The main difference is that the +filesystem providing the service will probably allocate a contiguous collection +of pages and permit mappings to be made on that. + +It is recommended that a truncate operation applied to such a file that +increases the file size, if that file is empty, be taken as a request to gather +enough pages to honour a mapping. This is required to support POSIX shared +memory. + +Memory backed devices are indicated by the mapping's backing device info having +the memory_backed flag set. + + +Providing shareable block device support +======================================== + +Provision of shared mappings on block device files is exactly the same as for +character devices. If there isn't a real device underneath, then the driver +should allocate sufficient contiguous memory to honour any supported mapping. + + +Adjusting page trimming behaviour +================================= + +NOMMU mmap automatically rounds up to the nearest power-of-2 number of pages +when performing an allocation. This can have adverse effects on memory +fragmentation, and as such, is left configurable. The default behaviour is to +aggressively trim allocations and discard any excess pages back in to the page +allocator. In order to retain finer-grained control over fragmentation, this +behaviour can either be disabled completely, or bumped up to a higher page +watermark where trimming begins. + +Page trimming behaviour is configurable via the sysctl ``vm.nr_trim_pages``. diff --git a/Documentation/admin-guide/mm/numa_memory_policy.rst b/Documentation/admin-guide/mm/numa_memory_policy.rst new file mode 100644 index 000000000..067a90a14 --- /dev/null +++ b/Documentation/admin-guide/mm/numa_memory_policy.rst @@ -0,0 +1,495 @@ +.. _numa_memory_policy: + +================== +NUMA Memory Policy +================== + +What is NUMA Memory Policy? +============================ + +In the Linux kernel, "memory policy" determines from which node the kernel will +allocate memory in a NUMA system or in an emulated NUMA system. Linux has +supported platforms with Non-Uniform Memory Access architectures since 2.4.?. +The current memory policy support was added to Linux 2.6 around May 2004. This +document attempts to describe the concepts and APIs of the 2.6 memory policy +support. + +Memory policies should not be confused with cpusets +(``Documentation/admin-guide/cgroup-v1/cpusets.rst``) +which is an administrative mechanism for restricting the nodes from which +memory may be allocated by a set of processes. Memory policies are a +programming interface that a NUMA-aware application can take advantage of. When +both cpusets and policies are applied to a task, the restrictions of the cpuset +takes priority. See :ref:`Memory Policies and cpusets <mem_pol_and_cpusets>` +below for more details. + +Memory Policy Concepts +====================== + +Scope of Memory Policies +------------------------ + +The Linux kernel supports _scopes_ of memory policy, described here from +most general to most specific: + +System Default Policy + this policy is "hard coded" into the kernel. It is the policy + that governs all page allocations that aren't controlled by + one of the more specific policy scopes discussed below. When + the system is "up and running", the system default policy will + use "local allocation" described below. However, during boot + up, the system default policy will be set to interleave + allocations across all nodes with "sufficient" memory, so as + not to overload the initial boot node with boot-time + allocations. + +Task/Process Policy + this is an optional, per-task policy. When defined for a + specific task, this policy controls all page allocations made + by or on behalf of the task that aren't controlled by a more + specific scope. If a task does not define a task policy, then + all page allocations that would have been controlled by the + task policy "fall back" to the System Default Policy. + + The task policy applies to the entire address space of a task. Thus, + it is inheritable, and indeed is inherited, across both fork() + [clone() w/o the CLONE_VM flag] and exec*(). This allows a parent task + to establish the task policy for a child task exec()'d from an + executable image that has no awareness of memory policy. See the + :ref:`Memory Policy APIs <memory_policy_apis>` section, + below, for an overview of the system call + that a task may use to set/change its task/process policy. + + In a multi-threaded task, task policies apply only to the thread + [Linux kernel task] that installs the policy and any threads + subsequently created by that thread. Any sibling threads existing + at the time a new task policy is installed retain their current + policy. + + A task policy applies only to pages allocated after the policy is + installed. Any pages already faulted in by the task when the task + changes its task policy remain where they were allocated based on + the policy at the time they were allocated. + +.. _vma_policy: + +VMA Policy + A "VMA" or "Virtual Memory Area" refers to a range of a task's + virtual address space. A task may define a specific policy for a range + of its virtual address space. See the + :ref:`Memory Policy APIs <memory_policy_apis>` section, + below, for an overview of the mbind() system call used to set a VMA + policy. + + A VMA policy will govern the allocation of pages that back + this region of the address space. Any regions of the task's + address space that don't have an explicit VMA policy will fall + back to the task policy, which may itself fall back to the + System Default Policy. + + VMA policies have a few complicating details: + + * VMA policy applies ONLY to anonymous pages. These include + pages allocated for anonymous segments, such as the task + stack and heap, and any regions of the address space + mmap()ed with the MAP_ANONYMOUS flag. If a VMA policy is + applied to a file mapping, it will be ignored if the mapping + used the MAP_SHARED flag. If the file mapping used the + MAP_PRIVATE flag, the VMA policy will only be applied when + an anonymous page is allocated on an attempt to write to the + mapping-- i.e., at Copy-On-Write. + + * VMA policies are shared between all tasks that share a + virtual address space--a.k.a. threads--independent of when + the policy is installed; and they are inherited across + fork(). However, because VMA policies refer to a specific + region of a task's address space, and because the address + space is discarded and recreated on exec*(), VMA policies + are NOT inheritable across exec(). Thus, only NUMA-aware + applications may use VMA policies. + + * A task may install a new VMA policy on a sub-range of a + previously mmap()ed region. When this happens, Linux splits + the existing virtual memory area into 2 or 3 VMAs, each with + it's own policy. + + * By default, VMA policy applies only to pages allocated after + the policy is installed. Any pages already faulted into the + VMA range remain where they were allocated based on the + policy at the time they were allocated. However, since + 2.6.16, Linux supports page migration via the mbind() system + call, so that page contents can be moved to match a newly + installed policy. + +Shared Policy + Conceptually, shared policies apply to "memory objects" mapped + shared into one or more tasks' distinct address spaces. An + application installs shared policies the same way as VMA + policies--using the mbind() system call specifying a range of + virtual addresses that map the shared object. However, unlike + VMA policies, which can be considered to be an attribute of a + range of a task's address space, shared policies apply + directly to the shared object. Thus, all tasks that attach to + the object share the policy, and all pages allocated for the + shared object, by any task, will obey the shared policy. + + As of 2.6.22, only shared memory segments, created by shmget() or + mmap(MAP_ANONYMOUS|MAP_SHARED), support shared policy. When shared + policy support was added to Linux, the associated data structures were + added to hugetlbfs shmem segments. At the time, hugetlbfs did not + support allocation at fault time--a.k.a lazy allocation--so hugetlbfs + shmem segments were never "hooked up" to the shared policy support. + Although hugetlbfs segments now support lazy allocation, their support + for shared policy has not been completed. + + As mentioned above in :ref:`VMA policies <vma_policy>` section, + allocations of page cache pages for regular files mmap()ed + with MAP_SHARED ignore any VMA policy installed on the virtual + address range backed by the shared file mapping. Rather, + shared page cache pages, including pages backing private + mappings that have not yet been written by the task, follow + task policy, if any, else System Default Policy. + + The shared policy infrastructure supports different policies on subset + ranges of the shared object. However, Linux still splits the VMA of + the task that installs the policy for each range of distinct policy. + Thus, different tasks that attach to a shared memory segment can have + different VMA configurations mapping that one shared object. This + can be seen by examining the /proc/<pid>/numa_maps of tasks sharing + a shared memory region, when one task has installed shared policy on + one or more ranges of the region. + +Components of Memory Policies +----------------------------- + +A NUMA memory policy consists of a "mode", optional mode flags, and +an optional set of nodes. The mode determines the behavior of the +policy, the optional mode flags determine the behavior of the mode, +and the optional set of nodes can be viewed as the arguments to the +policy behavior. + +Internally, memory policies are implemented by a reference counted +structure, struct mempolicy. Details of this structure will be +discussed in context, below, as required to explain the behavior. + +NUMA memory policy supports the following 4 behavioral modes: + +Default Mode--MPOL_DEFAULT + This mode is only used in the memory policy APIs. Internally, + MPOL_DEFAULT is converted to the NULL memory policy in all + policy scopes. Any existing non-default policy will simply be + removed when MPOL_DEFAULT is specified. As a result, + MPOL_DEFAULT means "fall back to the next most specific policy + scope." + + For example, a NULL or default task policy will fall back to the + system default policy. A NULL or default vma policy will fall + back to the task policy. + + When specified in one of the memory policy APIs, the Default mode + does not use the optional set of nodes. + + It is an error for the set of nodes specified for this policy to + be non-empty. + +MPOL_BIND + This mode specifies that memory must come from the set of + nodes specified by the policy. Memory will be allocated from + the node in the set with sufficient free memory that is + closest to the node where the allocation takes place. + +MPOL_PREFERRED + This mode specifies that the allocation should be attempted + from the single node specified in the policy. If that + allocation fails, the kernel will search other nodes, in order + of increasing distance from the preferred node based on + information provided by the platform firmware. + + Internally, the Preferred policy uses a single node--the + preferred_node member of struct mempolicy. When the internal + mode flag MPOL_F_LOCAL is set, the preferred_node is ignored + and the policy is interpreted as local allocation. "Local" + allocation policy can be viewed as a Preferred policy that + starts at the node containing the cpu where the allocation + takes place. + + It is possible for the user to specify that local allocation + is always preferred by passing an empty nodemask with this + mode. If an empty nodemask is passed, the policy cannot use + the MPOL_F_STATIC_NODES or MPOL_F_RELATIVE_NODES flags + described below. + +MPOL_INTERLEAVED + This mode specifies that page allocations be interleaved, on a + page granularity, across the nodes specified in the policy. + This mode also behaves slightly differently, based on the + context where it is used: + + For allocation of anonymous pages and shared memory pages, + Interleave mode indexes the set of nodes specified by the + policy using the page offset of the faulting address into the + segment [VMA] containing the address modulo the number of + nodes specified by the policy. It then attempts to allocate a + page, starting at the selected node, as if the node had been + specified by a Preferred policy or had been selected by a + local allocation. That is, allocation will follow the per + node zonelist. + + For allocation of page cache pages, Interleave mode indexes + the set of nodes specified by the policy using a node counter + maintained per task. This counter wraps around to the lowest + specified node after it reaches the highest specified node. + This will tend to spread the pages out over the nodes + specified by the policy based on the order in which they are + allocated, rather than based on any page offset into an + address range or file. During system boot up, the temporary + interleaved system default policy works in this mode. + +NUMA memory policy supports the following optional mode flags: + +MPOL_F_STATIC_NODES + This flag specifies that the nodemask passed by + the user should not be remapped if the task or VMA's set of allowed + nodes changes after the memory policy has been defined. + + Without this flag, any time a mempolicy is rebound because of a + change in the set of allowed nodes, the node (Preferred) or + nodemask (Bind, Interleave) is remapped to the new set of + allowed nodes. This may result in nodes being used that were + previously undesired. + + With this flag, if the user-specified nodes overlap with the + nodes allowed by the task's cpuset, then the memory policy is + applied to their intersection. If the two sets of nodes do not + overlap, the Default policy is used. + + For example, consider a task that is attached to a cpuset with + mems 1-3 that sets an Interleave policy over the same set. If + the cpuset's mems change to 3-5, the Interleave will now occur + over nodes 3, 4, and 5. With this flag, however, since only node + 3 is allowed from the user's nodemask, the "interleave" only + occurs over that node. If no nodes from the user's nodemask are + now allowed, the Default behavior is used. + + MPOL_F_STATIC_NODES cannot be combined with the + MPOL_F_RELATIVE_NODES flag. It also cannot be used for + MPOL_PREFERRED policies that were created with an empty nodemask + (local allocation). + +MPOL_F_RELATIVE_NODES + This flag specifies that the nodemask passed + by the user will be mapped relative to the set of the task or VMA's + set of allowed nodes. The kernel stores the user-passed nodemask, + and if the allowed nodes changes, then that original nodemask will + be remapped relative to the new set of allowed nodes. + + Without this flag (and without MPOL_F_STATIC_NODES), anytime a + mempolicy is rebound because of a change in the set of allowed + nodes, the node (Preferred) or nodemask (Bind, Interleave) is + remapped to the new set of allowed nodes. That remap may not + preserve the relative nature of the user's passed nodemask to its + set of allowed nodes upon successive rebinds: a nodemask of + 1,3,5 may be remapped to 7-9 and then to 1-3 if the set of + allowed nodes is restored to its original state. + + With this flag, the remap is done so that the node numbers from + the user's passed nodemask are relative to the set of allowed + nodes. In other words, if nodes 0, 2, and 4 are set in the user's + nodemask, the policy will be effected over the first (and in the + Bind or Interleave case, the third and fifth) nodes in the set of + allowed nodes. The nodemask passed by the user represents nodes + relative to task or VMA's set of allowed nodes. + + If the user's nodemask includes nodes that are outside the range + of the new set of allowed nodes (for example, node 5 is set in + the user's nodemask when the set of allowed nodes is only 0-3), + then the remap wraps around to the beginning of the nodemask and, + if not already set, sets the node in the mempolicy nodemask. + + For example, consider a task that is attached to a cpuset with + mems 2-5 that sets an Interleave policy over the same set with + MPOL_F_RELATIVE_NODES. If the cpuset's mems change to 3-7, the + interleave now occurs over nodes 3,5-7. If the cpuset's mems + then change to 0,2-3,5, then the interleave occurs over nodes + 0,2-3,5. + + Thanks to the consistent remapping, applications preparing + nodemasks to specify memory policies using this flag should + disregard their current, actual cpuset imposed memory placement + and prepare the nodemask as if they were always located on + memory nodes 0 to N-1, where N is the number of memory nodes the + policy is intended to manage. Let the kernel then remap to the + set of memory nodes allowed by the task's cpuset, as that may + change over time. + + MPOL_F_RELATIVE_NODES cannot be combined with the + MPOL_F_STATIC_NODES flag. It also cannot be used for + MPOL_PREFERRED policies that were created with an empty nodemask + (local allocation). + +Memory Policy Reference Counting +================================ + +To resolve use/free races, struct mempolicy contains an atomic reference +count field. Internal interfaces, mpol_get()/mpol_put() increment and +decrement this reference count, respectively. mpol_put() will only free +the structure back to the mempolicy kmem cache when the reference count +goes to zero. + +When a new memory policy is allocated, its reference count is initialized +to '1', representing the reference held by the task that is installing the +new policy. When a pointer to a memory policy structure is stored in another +structure, another reference is added, as the task's reference will be dropped +on completion of the policy installation. + +During run-time "usage" of the policy, we attempt to minimize atomic operations +on the reference count, as this can lead to cache lines bouncing between cpus +and NUMA nodes. "Usage" here means one of the following: + +1) querying of the policy, either by the task itself [using the get_mempolicy() + API discussed below] or by another task using the /proc/<pid>/numa_maps + interface. + +2) examination of the policy to determine the policy mode and associated node + or node lists, if any, for page allocation. This is considered a "hot + path". Note that for MPOL_BIND, the "usage" extends across the entire + allocation process, which may sleep during page reclaimation, because the + BIND policy nodemask is used, by reference, to filter ineligible nodes. + +We can avoid taking an extra reference during the usages listed above as +follows: + +1) we never need to get/free the system default policy as this is never + changed nor freed, once the system is up and running. + +2) for querying the policy, we do not need to take an extra reference on the + target task's task policy nor vma policies because we always acquire the + task's mm's mmap_lock for read during the query. The set_mempolicy() and + mbind() APIs [see below] always acquire the mmap_lock for write when + installing or replacing task or vma policies. Thus, there is no possibility + of a task or thread freeing a policy while another task or thread is + querying it. + +3) Page allocation usage of task or vma policy occurs in the fault path where + we hold them mmap_lock for read. Again, because replacing the task or vma + policy requires that the mmap_lock be held for write, the policy can't be + freed out from under us while we're using it for page allocation. + +4) Shared policies require special consideration. One task can replace a + shared memory policy while another task, with a distinct mmap_lock, is + querying or allocating a page based on the policy. To resolve this + potential race, the shared policy infrastructure adds an extra reference + to the shared policy during lookup while holding a spin lock on the shared + policy management structure. This requires that we drop this extra + reference when we're finished "using" the policy. We must drop the + extra reference on shared policies in the same query/allocation paths + used for non-shared policies. For this reason, shared policies are marked + as such, and the extra reference is dropped "conditionally"--i.e., only + for shared policies. + + Because of this extra reference counting, and because we must lookup + shared policies in a tree structure under spinlock, shared policies are + more expensive to use in the page allocation path. This is especially + true for shared policies on shared memory regions shared by tasks running + on different NUMA nodes. This extra overhead can be avoided by always + falling back to task or system default policy for shared memory regions, + or by prefaulting the entire shared memory region into memory and locking + it down. However, this might not be appropriate for all applications. + +.. _memory_policy_apis: + +Memory Policy APIs +================== + +Linux supports 3 system calls for controlling memory policy. These APIS +always affect only the calling task, the calling task's address space, or +some shared object mapped into the calling task's address space. + +.. note:: + the headers that define these APIs and the parameter data types for + user space applications reside in a package that is not part of the + Linux kernel. The kernel system call interfaces, with the 'sys\_' + prefix, are defined in <linux/syscalls.h>; the mode and flag + definitions are defined in <linux/mempolicy.h>. + +Set [Task] Memory Policy:: + + long set_mempolicy(int mode, const unsigned long *nmask, + unsigned long maxnode); + +Set's the calling task's "task/process memory policy" to mode +specified by the 'mode' argument and the set of nodes defined by +'nmask'. 'nmask' points to a bit mask of node ids containing at least +'maxnode' ids. Optional mode flags may be passed by combining the +'mode' argument with the flag (for example: MPOL_INTERLEAVE | +MPOL_F_STATIC_NODES). + +See the set_mempolicy(2) man page for more details + + +Get [Task] Memory Policy or Related Information:: + + long get_mempolicy(int *mode, + const unsigned long *nmask, unsigned long maxnode, + void *addr, int flags); + +Queries the "task/process memory policy" of the calling task, or the +policy or location of a specified virtual address, depending on the +'flags' argument. + +See the get_mempolicy(2) man page for more details + + +Install VMA/Shared Policy for a Range of Task's Address Space:: + + long mbind(void *start, unsigned long len, int mode, + const unsigned long *nmask, unsigned long maxnode, + unsigned flags); + +mbind() installs the policy specified by (mode, nmask, maxnodes) as a +VMA policy for the range of the calling task's address space specified +by the 'start' and 'len' arguments. Additional actions may be +requested via the 'flags' argument. + +See the mbind(2) man page for more details. + +Memory Policy Command Line Interface +==================================== + +Although not strictly part of the Linux implementation of memory policy, +a command line tool, numactl(8), exists that allows one to: + ++ set the task policy for a specified program via set_mempolicy(2), fork(2) and + exec(2) + ++ set the shared policy for a shared memory segment via mbind(2) + +The numactl(8) tool is packaged with the run-time version of the library +containing the memory policy system call wrappers. Some distributions +package the headers and compile-time libraries in a separate development +package. + +.. _mem_pol_and_cpusets: + +Memory Policies and cpusets +=========================== + +Memory policies work within cpusets as described above. For memory policies +that require a node or set of nodes, the nodes are restricted to the set of +nodes whose memories are allowed by the cpuset constraints. If the nodemask +specified for the policy contains nodes that are not allowed by the cpuset and +MPOL_F_RELATIVE_NODES is not used, the intersection of the set of nodes +specified for the policy and the set of nodes with memory is used. If the +result is the empty set, the policy is considered invalid and cannot be +installed. If MPOL_F_RELATIVE_NODES is used, the policy's nodes are mapped +onto and folded into the task's set of allowed nodes as previously described. + +The interaction of memory policies and cpusets can be problematic when tasks +in two cpusets share access to a memory region, such as shared memory segments +created by shmget() of mmap() with the MAP_ANONYMOUS and MAP_SHARED flags, and +any of the tasks install shared policy on the region, only nodes whose +memories are allowed in both cpusets may be used in the policies. Obtaining +this information requires "stepping outside" the memory policy APIs to use the +cpuset information and requires that one know in what cpusets other task might +be attaching to the shared region. Furthermore, if the cpusets' allowed +memory sets are disjoint, "local" allocation is the only valid policy. diff --git a/Documentation/admin-guide/mm/numaperf.rst b/Documentation/admin-guide/mm/numaperf.rst new file mode 100644 index 000000000..86f2a3c4b --- /dev/null +++ b/Documentation/admin-guide/mm/numaperf.rst @@ -0,0 +1,178 @@ +.. _numaperf: + +============= +NUMA Locality +============= + +Some platforms may have multiple types of memory attached to a compute +node. These disparate memory ranges may share some characteristics, such +as CPU cache coherence, but may have different performance. For example, +different media types and buses affect bandwidth and latency. + +A system supports such heterogeneous memory by grouping each memory type +under different domains, or "nodes", based on locality and performance +characteristics. Some memory may share the same node as a CPU, and others +are provided as memory only nodes. While memory only nodes do not provide +CPUs, they may still be local to one or more compute nodes relative to +other nodes. The following diagram shows one such example of two compute +nodes with local memory and a memory only node for each of compute node:: + + +------------------+ +------------------+ + | Compute Node 0 +-----+ Compute Node 1 | + | Local Node0 Mem | | Local Node1 Mem | + +--------+---------+ +--------+---------+ + | | + +--------+---------+ +--------+---------+ + | Slower Node2 Mem | | Slower Node3 Mem | + +------------------+ +--------+---------+ + +A "memory initiator" is a node containing one or more devices such as +CPUs or separate memory I/O devices that can initiate memory requests. +A "memory target" is a node containing one or more physical address +ranges accessible from one or more memory initiators. + +When multiple memory initiators exist, they may not all have the same +performance when accessing a given memory target. Each initiator-target +pair may be organized into different ranked access classes to represent +this relationship. The highest performing initiator to a given target +is considered to be one of that target's local initiators, and given +the highest access class, 0. Any given target may have one or more +local initiators, and any given initiator may have multiple local +memory targets. + +To aid applications matching memory targets with their initiators, the +kernel provides symlinks to each other. The following example lists the +relationship for the access class "0" memory initiators and targets:: + + # symlinks -v /sys/devices/system/node/nodeX/access0/targets/ + relative: /sys/devices/system/node/nodeX/access0/targets/nodeY -> ../../nodeY + + # symlinks -v /sys/devices/system/node/nodeY/access0/initiators/ + relative: /sys/devices/system/node/nodeY/access0/initiators/nodeX -> ../../nodeX + +A memory initiator may have multiple memory targets in the same access +class. The target memory's initiators in a given class indicate the +nodes' access characteristics share the same performance relative to other +linked initiator nodes. Each target within an initiator's access class, +though, do not necessarily perform the same as each other. + +The access class "1" is used to allow differentiation between initiators +that are CPUs and hence suitable for generic task scheduling, and +IO initiators such as GPUs and NICs. Unlike access class 0, only +nodes containing CPUs are considered. + +================ +NUMA Performance +================ + +Applications may wish to consider which node they want their memory to +be allocated from based on the node's performance characteristics. If +the system provides these attributes, the kernel exports them under the +node sysfs hierarchy by appending the attributes directory under the +memory node's access class 0 initiators as follows:: + + /sys/devices/system/node/nodeY/access0/initiators/ + +These attributes apply only when accessed from nodes that have the +are linked under the this access's inititiators. + +The performance characteristics the kernel provides for the local initiators +are exported are as follows:: + + # tree -P "read*|write*" /sys/devices/system/node/nodeY/access0/initiators/ + /sys/devices/system/node/nodeY/access0/initiators/ + |-- read_bandwidth + |-- read_latency + |-- write_bandwidth + `-- write_latency + +The bandwidth attributes are provided in MiB/second. + +The latency attributes are provided in nanoseconds. + +The values reported here correspond to the rated latency and bandwidth +for the platform. + +Access class 1 takes the same form but only includes values for CPU to +memory activity. + +========== +NUMA Cache +========== + +System memory may be constructed in a hierarchy of elements with various +performance characteristics in order to provide large address space of +slower performing memory cached by a smaller higher performing memory. The +system physical addresses memory initiators are aware of are provided +by the last memory level in the hierarchy. The system meanwhile uses +higher performing memory to transparently cache access to progressively +slower levels. + +The term "far memory" is used to denote the last level memory in the +hierarchy. Each increasing cache level provides higher performing +initiator access, and the term "near memory" represents the fastest +cache provided by the system. + +This numbering is different than CPU caches where the cache level (ex: +L1, L2, L3) uses the CPU-side view where each increased level is lower +performing. In contrast, the memory cache level is centric to the last +level memory, so the higher numbered cache level corresponds to memory +nearer to the CPU, and further from far memory. + +The memory-side caches are not directly addressable by software. When +software accesses a system address, the system will return it from the +near memory cache if it is present. If it is not present, the system +accesses the next level of memory until there is either a hit in that +cache level, or it reaches far memory. + +An application does not need to know about caching attributes in order +to use the system. Software may optionally query the memory cache +attributes in order to maximize the performance out of such a setup. +If the system provides a way for the kernel to discover this information, +for example with ACPI HMAT (Heterogeneous Memory Attribute Table), +the kernel will append these attributes to the NUMA node memory target. + +When the kernel first registers a memory cache with a node, the kernel +will create the following directory:: + + /sys/devices/system/node/nodeX/memory_side_cache/ + +If that directory is not present, the system either does not provide +a memory-side cache, or that information is not accessible to the kernel. + +The attributes for each level of cache is provided under its cache +level index:: + + /sys/devices/system/node/nodeX/memory_side_cache/indexA/ + /sys/devices/system/node/nodeX/memory_side_cache/indexB/ + /sys/devices/system/node/nodeX/memory_side_cache/indexC/ + +Each cache level's directory provides its attributes. For example, the +following shows a single cache level and the attributes available for +software to query:: + + # tree sys/devices/system/node/node0/memory_side_cache/ + /sys/devices/system/node/node0/memory_side_cache/ + |-- index1 + | |-- indexing + | |-- line_size + | |-- size + | `-- write_policy + +The "indexing" will be 0 if it is a direct-mapped cache, and non-zero +for any other indexed based, multi-way associativity. + +The "line_size" is the number of bytes accessed from the next cache +level on a miss. + +The "size" is the number of bytes provided by this cache level. + +The "write_policy" will be 0 for write-back, and non-zero for +write-through caching. + +======== +See Also +======== + +[1] https://www.uefi.org/sites/default/files/resources/ACPI_6_2.pdf +- Section 5.2.27 diff --git a/Documentation/admin-guide/mm/pagemap.rst b/Documentation/admin-guide/mm/pagemap.rst new file mode 100644 index 000000000..340a5aee9 --- /dev/null +++ b/Documentation/admin-guide/mm/pagemap.rst @@ -0,0 +1,207 @@ +.. _pagemap: + +============================= +Examining Process Page Tables +============================= + +pagemap is a new (as of 2.6.25) set of interfaces in the kernel that allow +userspace programs to examine the page tables and related information by +reading files in ``/proc``. + +There are four components to pagemap: + + * ``/proc/pid/pagemap``. This file lets a userspace process find out which + physical frame each virtual page is mapped to. It contains one 64-bit + value for each virtual page, containing the following data (from + ``fs/proc/task_mmu.c``, above pagemap_read): + + * Bits 0-54 page frame number (PFN) if present + * Bits 0-4 swap type if swapped + * Bits 5-54 swap offset if swapped + * Bit 55 pte is soft-dirty (see + :ref:`Documentation/admin-guide/mm/soft-dirty.rst <soft_dirty>`) + * Bit 56 page exclusively mapped (since 4.2) + * Bits 57-60 zero + * Bit 61 page is file-page or shared-anon (since 3.5) + * Bit 62 page swapped + * Bit 63 page present + + Since Linux 4.0 only users with the CAP_SYS_ADMIN capability can get PFNs. + In 4.0 and 4.1 opens by unprivileged fail with -EPERM. Starting from + 4.2 the PFN field is zeroed if the user does not have CAP_SYS_ADMIN. + Reason: information about PFNs helps in exploiting Rowhammer vulnerability. + + If the page is not present but in swap, then the PFN contains an + encoding of the swap file number and the page's offset into the + swap. Unmapped pages return a null PFN. This allows determining + precisely which pages are mapped (or in swap) and comparing mapped + pages between processes. + + Efficient users of this interface will use ``/proc/pid/maps`` to + determine which areas of memory are actually mapped and llseek to + skip over unmapped regions. + + * ``/proc/kpagecount``. This file contains a 64-bit count of the number of + times each page is mapped, indexed by PFN. + +The page-types tool in the tools/vm directory can be used to query the +number of times a page is mapped. + + * ``/proc/kpageflags``. This file contains a 64-bit set of flags for each + page, indexed by PFN. + + The flags are (from ``fs/proc/page.c``, above kpageflags_read): + + 0. LOCKED + 1. ERROR + 2. REFERENCED + 3. UPTODATE + 4. DIRTY + 5. LRU + 6. ACTIVE + 7. SLAB + 8. WRITEBACK + 9. RECLAIM + 10. BUDDY + 11. MMAP + 12. ANON + 13. SWAPCACHE + 14. SWAPBACKED + 15. COMPOUND_HEAD + 16. COMPOUND_TAIL + 17. HUGE + 18. UNEVICTABLE + 19. HWPOISON + 20. NOPAGE + 21. KSM + 22. THP + 23. OFFLINE + 24. ZERO_PAGE + 25. IDLE + 26. PGTABLE + + * ``/proc/kpagecgroup``. This file contains a 64-bit inode number of the + memory cgroup each page is charged to, indexed by PFN. Only available when + CONFIG_MEMCG is set. + +Short descriptions to the page flags +==================================== + +0 - LOCKED + page is being locked for exclusive access, e.g. by undergoing read/write IO +7 - SLAB + page is managed by the SLAB/SLOB/SLUB/SLQB kernel memory allocator + When compound page is used, SLUB/SLQB will only set this flag on the head + page; SLOB will not flag it at all. +10 - BUDDY + a free memory block managed by the buddy system allocator + The buddy system organizes free memory in blocks of various orders. + An order N block has 2^N physically contiguous pages, with the BUDDY flag + set for and _only_ for the first page. +15 - COMPOUND_HEAD + A compound page with order N consists of 2^N physically contiguous pages. + A compound page with order 2 takes the form of "HTTT", where H donates its + head page and T donates its tail page(s). The major consumers of compound + pages are hugeTLB pages + (:ref:`Documentation/admin-guide/mm/hugetlbpage.rst <hugetlbpage>`), + the SLUB etc. memory allocators and various device drivers. + However in this interface, only huge/giga pages are made visible + to end users. +16 - COMPOUND_TAIL + A compound page tail (see description above). +17 - HUGE + this is an integral part of a HugeTLB page +19 - HWPOISON + hardware detected memory corruption on this page: don't touch the data! +20 - NOPAGE + no page frame exists at the requested address +21 - KSM + identical memory pages dynamically shared between one or more processes +22 - THP + contiguous pages which construct transparent hugepages +23 - OFFLINE + page is logically offline +24 - ZERO_PAGE + zero page for pfn_zero or huge_zero page +25 - IDLE + page has not been accessed since it was marked idle (see + :ref:`Documentation/admin-guide/mm/idle_page_tracking.rst <idle_page_tracking>`). + Note that this flag may be stale in case the page was accessed via + a PTE. To make sure the flag is up-to-date one has to read + ``/sys/kernel/mm/page_idle/bitmap`` first. +26 - PGTABLE + page is in use as a page table + +IO related page flags +--------------------- + +1 - ERROR + IO error occurred +3 - UPTODATE + page has up-to-date data + ie. for file backed page: (in-memory data revision >= on-disk one) +4 - DIRTY + page has been written to, hence contains new data + i.e. for file backed page: (in-memory data revision > on-disk one) +8 - WRITEBACK + page is being synced to disk + +LRU related page flags +---------------------- + +5 - LRU + page is in one of the LRU lists +6 - ACTIVE + page is in the active LRU list +18 - UNEVICTABLE + page is in the unevictable (non-)LRU list It is somehow pinned and + not a candidate for LRU page reclaims, e.g. ramfs pages, + shmctl(SHM_LOCK) and mlock() memory segments +2 - REFERENCED + page has been referenced since last LRU list enqueue/requeue +9 - RECLAIM + page will be reclaimed soon after its pageout IO completed +11 - MMAP + a memory mapped page +12 - ANON + a memory mapped page that is not part of a file +13 - SWAPCACHE + page is mapped to swap space, i.e. has an associated swap entry +14 - SWAPBACKED + page is backed by swap/RAM + +The page-types tool in the tools/vm directory can be used to query the +above flags. + +Using pagemap to do something useful +==================================== + +The general procedure for using pagemap to find out about a process' memory +usage goes like this: + + 1. Read ``/proc/pid/maps`` to determine which parts of the memory space are + mapped to what. + 2. Select the maps you are interested in -- all of them, or a particular + library, or the stack or the heap, etc. + 3. Open ``/proc/pid/pagemap`` and seek to the pages you would like to examine. + 4. Read a u64 for each page from pagemap. + 5. Open ``/proc/kpagecount`` and/or ``/proc/kpageflags``. For each PFN you + just read, seek to that entry in the file, and read the data you want. + +For example, to find the "unique set size" (USS), which is the amount of +memory that a process is using that is not shared with any other process, +you can go through every map in the process, find the PFNs, look those up +in kpagecount, and tally up the number of pages that are only referenced +once. + +Other notes +=========== + +Reading from any of the files will return -EINVAL if you are not starting +the read on an 8-byte boundary (e.g., if you sought an odd number of bytes +into the file), or if the size of the read is not a multiple of 8 bytes. + +Before Linux 3.11 pagemap bits 55-60 were used for "page-shift" (which is +always 12 at most architectures). Since Linux 3.11 their meaning changes +after first clear of soft-dirty bits. Since Linux 4.2 they are used for +flags unconditionally. diff --git a/Documentation/admin-guide/mm/soft-dirty.rst b/Documentation/admin-guide/mm/soft-dirty.rst new file mode 100644 index 000000000..cb0cfd667 --- /dev/null +++ b/Documentation/admin-guide/mm/soft-dirty.rst @@ -0,0 +1,47 @@ +.. _soft_dirty: + +=============== +Soft-Dirty PTEs +=============== + +The soft-dirty is a bit on a PTE which helps to track which pages a task +writes to. In order to do this tracking one should + + 1. Clear soft-dirty bits from the task's PTEs. + + This is done by writing "4" into the ``/proc/PID/clear_refs`` file of the + task in question. + + 2. Wait some time. + + 3. Read soft-dirty bits from the PTEs. + + This is done by reading from the ``/proc/PID/pagemap``. The bit 55 of the + 64-bit qword is the soft-dirty one. If set, the respective PTE was + written to since step 1. + + +Internally, to do this tracking, the writable bit is cleared from PTEs +when the soft-dirty bit is cleared. So, after this, when the task tries to +modify a page at some virtual address the #PF occurs and the kernel sets +the soft-dirty bit on the respective PTE. + +Note, that although all the task's address space is marked as r/o after the +soft-dirty bits clear, the #PF-s that occur after that are processed fast. +This is so, since the pages are still mapped to physical memory, and thus all +the kernel does is finds this fact out and puts both writable and soft-dirty +bits on the PTE. + +While in most cases tracking memory changes by #PF-s is more than enough +there is still a scenario when we can lose soft dirty bits -- a task +unmaps a previously mapped memory region and then maps a new one at exactly +the same place. When unmap is called, the kernel internally clears PTE values +including soft dirty bits. To notify user space application about such +memory region renewal the kernel always marks new memory regions (and +expanded regions) as soft dirty. + +This feature is actively used by the checkpoint-restore project. You +can find more details about it on http://criu.org + + +-- Pavel Emelyanov, Apr 9, 2013 diff --git a/Documentation/admin-guide/mm/transhuge.rst b/Documentation/admin-guide/mm/transhuge.rst new file mode 100644 index 000000000..b2acd0d39 --- /dev/null +++ b/Documentation/admin-guide/mm/transhuge.rst @@ -0,0 +1,438 @@ +.. _admin_guide_transhuge: + +============================ +Transparent Hugepage Support +============================ + +Objective +========= + +Performance critical computing applications dealing with large memory +working sets are already running on top of libhugetlbfs and in turn +hugetlbfs. Transparent HugePage Support (THP) is an alternative mean of +using huge pages for the backing of virtual memory with huge pages +that supports the automatic promotion and demotion of page sizes and +without the shortcomings of hugetlbfs. + +Currently THP only works for anonymous memory mappings and tmpfs/shmem. +But in the future it can expand to other filesystems. + +.. note:: + in the examples below we presume that the basic page size is 4K and + the huge page size is 2M, although the actual numbers may vary + depending on the CPU architecture. + +The reason applications are running faster is because of two +factors. The first factor is almost completely irrelevant and it's not +of significant interest because it'll also have the downside of +requiring larger clear-page copy-page in page faults which is a +potentially negative effect. The first factor consists in taking a +single page fault for each 2M virtual region touched by userland (so +reducing the enter/exit kernel frequency by a 512 times factor). This +only matters the first time the memory is accessed for the lifetime of +a memory mapping. The second long lasting and much more important +factor will affect all subsequent accesses to the memory for the whole +runtime of the application. The second factor consist of two +components: + +1) the TLB miss will run faster (especially with virtualization using + nested pagetables but almost always also on bare metal without + virtualization) + +2) a single TLB entry will be mapping a much larger amount of virtual + memory in turn reducing the number of TLB misses. With + virtualization and nested pagetables the TLB can be mapped of + larger size only if both KVM and the Linux guest are using + hugepages but a significant speedup already happens if only one of + the two is using hugepages just because of the fact the TLB miss is + going to run faster. + +THP can be enabled system wide or restricted to certain tasks or even +memory ranges inside task's address space. Unless THP is completely +disabled, there is ``khugepaged`` daemon that scans memory and +collapses sequences of basic pages into huge pages. + +The THP behaviour is controlled via :ref:`sysfs <thp_sysfs>` +interface and using madvise(2) and prctl(2) system calls. + +Transparent Hugepage Support maximizes the usefulness of free memory +if compared to the reservation approach of hugetlbfs by allowing all +unused memory to be used as cache or other movable (or even unmovable +entities). It doesn't require reservation to prevent hugepage +allocation failures to be noticeable from userland. It allows paging +and all other advanced VM features to be available on the +hugepages. It requires no modifications for applications to take +advantage of it. + +Applications however can be further optimized to take advantage of +this feature, like for example they've been optimized before to avoid +a flood of mmap system calls for every malloc(4k). Optimizing userland +is by far not mandatory and khugepaged already can take care of long +lived page allocations even for hugepage unaware applications that +deals with large amounts of memory. + +In certain cases when hugepages are enabled system wide, application +may end up allocating more memory resources. An application may mmap a +large region but only touch 1 byte of it, in that case a 2M page might +be allocated instead of a 4k page for no good. This is why it's +possible to disable hugepages system-wide and to only have them inside +MADV_HUGEPAGE madvise regions. + +Embedded systems should enable hugepages only inside madvise regions +to eliminate any risk of wasting any precious byte of memory and to +only run faster. + +Applications that gets a lot of benefit from hugepages and that don't +risk to lose memory by using hugepages, should use +madvise(MADV_HUGEPAGE) on their critical mmapped regions. + +.. _thp_sysfs: + +sysfs +===== + +Global THP controls +------------------- + +Transparent Hugepage Support for anonymous memory can be entirely disabled +(mostly for debugging purposes) or only enabled inside MADV_HUGEPAGE +regions (to avoid the risk of consuming more memory resources) or enabled +system wide. This can be achieved with one of:: + + echo always >/sys/kernel/mm/transparent_hugepage/enabled + echo madvise >/sys/kernel/mm/transparent_hugepage/enabled + echo never >/sys/kernel/mm/transparent_hugepage/enabled + +It's also possible to limit defrag efforts in the VM to generate +anonymous hugepages in case they're not immediately free to madvise +regions or to never try to defrag memory and simply fallback to regular +pages unless hugepages are immediately available. Clearly if we spend CPU +time to defrag memory, we would expect to gain even more by the fact we +use hugepages later instead of regular pages. This isn't always +guaranteed, but it may be more likely in case the allocation is for a +MADV_HUGEPAGE region. + +:: + + echo always >/sys/kernel/mm/transparent_hugepage/defrag + echo defer >/sys/kernel/mm/transparent_hugepage/defrag + echo defer+madvise >/sys/kernel/mm/transparent_hugepage/defrag + echo madvise >/sys/kernel/mm/transparent_hugepage/defrag + echo never >/sys/kernel/mm/transparent_hugepage/defrag + +always + means that an application requesting THP will stall on + allocation failure and directly reclaim pages and compact + memory in an effort to allocate a THP immediately. This may be + desirable for virtual machines that benefit heavily from THP + use and are willing to delay the VM start to utilise them. + +defer + means that an application will wake kswapd in the background + to reclaim pages and wake kcompactd to compact memory so that + THP is available in the near future. It's the responsibility + of khugepaged to then install the THP pages later. + +defer+madvise + will enter direct reclaim and compaction like ``always``, but + only for regions that have used madvise(MADV_HUGEPAGE); all + other regions will wake kswapd in the background to reclaim + pages and wake kcompactd to compact memory so that THP is + available in the near future. + +madvise + will enter direct reclaim like ``always`` but only for regions + that are have used madvise(MADV_HUGEPAGE). This is the default + behaviour. + +never + should be self-explanatory. + +By default kernel tries to use huge zero page on read page fault to +anonymous mapping. It's possible to disable huge zero page by writing 0 +or enable it back by writing 1:: + + echo 0 >/sys/kernel/mm/transparent_hugepage/use_zero_page + echo 1 >/sys/kernel/mm/transparent_hugepage/use_zero_page + +Some userspace (such as a test program, or an optimized memory allocation +library) may want to know the size (in bytes) of a transparent hugepage:: + + cat /sys/kernel/mm/transparent_hugepage/hpage_pmd_size + +khugepaged will be automatically started when +transparent_hugepage/enabled is set to "always" or "madvise, and it'll +be automatically shutdown if it's set to "never". + +Khugepaged controls +------------------- + +khugepaged runs usually at low frequency so while one may not want to +invoke defrag algorithms synchronously during the page faults, it +should be worth invoking defrag at least in khugepaged. However it's +also possible to disable defrag in khugepaged by writing 0 or enable +defrag in khugepaged by writing 1:: + + echo 0 >/sys/kernel/mm/transparent_hugepage/khugepaged/defrag + echo 1 >/sys/kernel/mm/transparent_hugepage/khugepaged/defrag + +You can also control how many pages khugepaged should scan at each +pass:: + + /sys/kernel/mm/transparent_hugepage/khugepaged/pages_to_scan + +and how many milliseconds to wait in khugepaged between each pass (you +can set this to 0 to run khugepaged at 100% utilization of one core):: + + /sys/kernel/mm/transparent_hugepage/khugepaged/scan_sleep_millisecs + +and how many milliseconds to wait in khugepaged if there's an hugepage +allocation failure to throttle the next allocation attempt:: + + /sys/kernel/mm/transparent_hugepage/khugepaged/alloc_sleep_millisecs + +The khugepaged progress can be seen in the number of pages collapsed:: + + /sys/kernel/mm/transparent_hugepage/khugepaged/pages_collapsed + +for each pass:: + + /sys/kernel/mm/transparent_hugepage/khugepaged/full_scans + +``max_ptes_none`` specifies how many extra small pages (that are +not already mapped) can be allocated when collapsing a group +of small pages into one large page:: + + /sys/kernel/mm/transparent_hugepage/khugepaged/max_ptes_none + +A higher value leads to use additional memory for programs. +A lower value leads to gain less thp performance. Value of +max_ptes_none can waste cpu time very little, you can +ignore it. + +``max_ptes_swap`` specifies how many pages can be brought in from +swap when collapsing a group of pages into a transparent huge page:: + + /sys/kernel/mm/transparent_hugepage/khugepaged/max_ptes_swap + +A higher value can cause excessive swap IO and waste +memory. A lower value can prevent THPs from being +collapsed, resulting fewer pages being collapsed into +THPs, and lower memory access performance. + +``max_ptes_shared`` specifies how many pages can be shared across multiple +processes. Exceeding the number would block the collapse:: + + /sys/kernel/mm/transparent_hugepage/khugepaged/max_ptes_shared + +A higher value may increase memory footprint for some workloads. + +Boot parameter +============== + +You can change the sysfs boot time defaults of Transparent Hugepage +Support by passing the parameter ``transparent_hugepage=always`` or +``transparent_hugepage=madvise`` or ``transparent_hugepage=never`` +to the kernel command line. + +Hugepages in tmpfs/shmem +======================== + +You can control hugepage allocation policy in tmpfs with mount option +``huge=``. It can have following values: + +always + Attempt to allocate huge pages every time we need a new page; + +never + Do not allocate huge pages; + +within_size + Only allocate huge page if it will be fully within i_size. + Also respect fadvise()/madvise() hints; + +advise + Only allocate huge pages if requested with fadvise()/madvise(); + +The default policy is ``never``. + +``mount -o remount,huge= /mountpoint`` works fine after mount: remounting +``huge=never`` will not attempt to break up huge pages at all, just stop more +from being allocated. + +There's also sysfs knob to control hugepage allocation policy for internal +shmem mount: /sys/kernel/mm/transparent_hugepage/shmem_enabled. The mount +is used for SysV SHM, memfds, shared anonymous mmaps (of /dev/zero or +MAP_ANONYMOUS), GPU drivers' DRM objects, Ashmem. + +In addition to policies listed above, shmem_enabled allows two further +values: + +deny + For use in emergencies, to force the huge option off from + all mounts; +force + Force the huge option on for all - very useful for testing; + +Need of application restart +=========================== + +The transparent_hugepage/enabled values and tmpfs mount option only affect +future behavior. So to make them effective you need to restart any +application that could have been using hugepages. This also applies to the +regions registered in khugepaged. + +Monitoring usage +================ + +The number of anonymous transparent huge pages currently used by the +system is available by reading the AnonHugePages field in ``/proc/meminfo``. +To identify what applications are using anonymous transparent huge pages, +it is necessary to read ``/proc/PID/smaps`` and count the AnonHugePages fields +for each mapping. + +The number of file transparent huge pages mapped to userspace is available +by reading ShmemPmdMapped and ShmemHugePages fields in ``/proc/meminfo``. +To identify what applications are mapping file transparent huge pages, it +is necessary to read ``/proc/PID/smaps`` and count the FileHugeMapped fields +for each mapping. + +Note that reading the smaps file is expensive and reading it +frequently will incur overhead. + +There are a number of counters in ``/proc/vmstat`` that may be used to +monitor how successfully the system is providing huge pages for use. + +thp_fault_alloc + is incremented every time a huge page is successfully + allocated to handle a page fault. + +thp_collapse_alloc + is incremented by khugepaged when it has found + a range of pages to collapse into one huge page and has + successfully allocated a new huge page to store the data. + +thp_fault_fallback + is incremented if a page fault fails to allocate + a huge page and instead falls back to using small pages. + +thp_fault_fallback_charge + is incremented if a page fault fails to charge a huge page and + instead falls back to using small pages even though the + allocation was successful. + +thp_collapse_alloc_failed + is incremented if khugepaged found a range + of pages that should be collapsed into one huge page but failed + the allocation. + +thp_file_alloc + is incremented every time a file huge page is successfully + allocated. + +thp_file_fallback + is incremented if a file huge page is attempted to be allocated + but fails and instead falls back to using small pages. + +thp_file_fallback_charge + is incremented if a file huge page cannot be charged and instead + falls back to using small pages even though the allocation was + successful. + +thp_file_mapped + is incremented every time a file huge page is mapped into + user address space. + +thp_split_page + is incremented every time a huge page is split into base + pages. This can happen for a variety of reasons but a common + reason is that a huge page is old and is being reclaimed. + This action implies splitting all PMD the page mapped with. + +thp_split_page_failed + is incremented if kernel fails to split huge + page. This can happen if the page was pinned by somebody. + +thp_deferred_split_page + is incremented when a huge page is put onto split + queue. This happens when a huge page is partially unmapped and + splitting it would free up some memory. Pages on split queue are + going to be split under memory pressure. + +thp_split_pmd + is incremented every time a PMD split into table of PTEs. + This can happen, for instance, when application calls mprotect() or + munmap() on part of huge page. It doesn't split huge page, only + page table entry. + +thp_zero_page_alloc + is incremented every time a huge zero page is + successfully allocated. It includes allocations which where + dropped due race with other allocation. Note, it doesn't count + every map of the huge zero page, only its allocation. + +thp_zero_page_alloc_failed + is incremented if kernel fails to allocate + huge zero page and falls back to using small pages. + +thp_swpout + is incremented every time a huge page is swapout in one + piece without splitting. + +thp_swpout_fallback + is incremented if a huge page has to be split before swapout. + Usually because failed to allocate some continuous swap space + for the huge page. + +As the system ages, allocating huge pages may be expensive as the +system uses memory compaction to copy data around memory to free a +huge page for use. There are some counters in ``/proc/vmstat`` to help +monitor this overhead. + +compact_stall + is incremented every time a process stalls to run + memory compaction so that a huge page is free for use. + +compact_success + is incremented if the system compacted memory and + freed a huge page for use. + +compact_fail + is incremented if the system tries to compact memory + but failed. + +compact_pages_moved + is incremented each time a page is moved. If + this value is increasing rapidly, it implies that the system + is copying a lot of data to satisfy the huge page allocation. + It is possible that the cost of copying exceeds any savings + from reduced TLB misses. + +compact_pagemigrate_failed + is incremented when the underlying mechanism + for moving a page failed. + +compact_blocks_moved + is incremented each time memory compaction examines + a huge page aligned range of pages. + +It is possible to establish how long the stalls were using the function +tracer to record how long was spent in __alloc_pages_nodemask and +using the mm_page_alloc tracepoint to identify which allocations were +for huge pages. + +Optimizing the applications +=========================== + +To be guaranteed that the kernel will map a 2M page immediately in any +memory region, the mmap region has to be hugepage naturally +aligned. posix_memalign() can provide that guarantee. + +Hugetlbfs +========= + +You can use hugetlbfs on a kernel that has transparent hugepage +support enabled just fine as always. No difference can be noted in +hugetlbfs other than there will be less overall fragmentation. All +usual features belonging to hugetlbfs are preserved and +unaffected. libhugetlbfs will also work fine as usual. diff --git a/Documentation/admin-guide/mm/userfaultfd.rst b/Documentation/admin-guide/mm/userfaultfd.rst new file mode 100644 index 000000000..1dc2d5f82 --- /dev/null +++ b/Documentation/admin-guide/mm/userfaultfd.rst @@ -0,0 +1,293 @@ +.. _userfaultfd: + +=========== +Userfaultfd +=========== + +Objective +========= + +Userfaults allow the implementation of on-demand paging from userland +and more generally they allow userland to take control of various +memory page faults, something otherwise only the kernel code could do. + +For example userfaults allows a proper and more optimal implementation +of the ``PROT_NONE+SIGSEGV`` trick. + +Design +====== + +Userfaults are delivered and resolved through the ``userfaultfd`` syscall. + +The ``userfaultfd`` (aside from registering and unregistering virtual +memory ranges) provides two primary functionalities: + +1) ``read/POLLIN`` protocol to notify a userland thread of the faults + happening + +2) various ``UFFDIO_*`` ioctls that can manage the virtual memory regions + registered in the ``userfaultfd`` that allows userland to efficiently + resolve the userfaults it receives via 1) or to manage the virtual + memory in the background + +The real advantage of userfaults if compared to regular virtual memory +management of mremap/mprotect is that the userfaults in all their +operations never involve heavyweight structures like vmas (in fact the +``userfaultfd`` runtime load never takes the mmap_lock for writing). + +Vmas are not suitable for page- (or hugepage) granular fault tracking +when dealing with virtual address spaces that could span +Terabytes. Too many vmas would be needed for that. + +The ``userfaultfd`` once opened by invoking the syscall, can also be +passed using unix domain sockets to a manager process, so the same +manager process could handle the userfaults of a multitude of +different processes without them being aware about what is going on +(well of course unless they later try to use the ``userfaultfd`` +themselves on the same region the manager is already tracking, which +is a corner case that would currently return ``-EBUSY``). + +API +=== + +When first opened the ``userfaultfd`` must be enabled invoking the +``UFFDIO_API`` ioctl specifying a ``uffdio_api.api`` value set to ``UFFD_API`` (or +a later API version) which will specify the ``read/POLLIN`` protocol +userland intends to speak on the ``UFFD`` and the ``uffdio_api.features`` +userland requires. The ``UFFDIO_API`` ioctl if successful (i.e. if the +requested ``uffdio_api.api`` is spoken also by the running kernel and the +requested features are going to be enabled) will return into +``uffdio_api.features`` and ``uffdio_api.ioctls`` two 64bit bitmasks of +respectively all the available features of the read(2) protocol and +the generic ioctl available. + +The ``uffdio_api.features`` bitmask returned by the ``UFFDIO_API`` ioctl +defines what memory types are supported by the ``userfaultfd`` and what +events, except page fault notifications, may be generated. + +If the kernel supports registering ``userfaultfd`` ranges on hugetlbfs +virtual memory areas, ``UFFD_FEATURE_MISSING_HUGETLBFS`` will be set in +``uffdio_api.features``. Similarly, ``UFFD_FEATURE_MISSING_SHMEM`` will be +set if the kernel supports registering ``userfaultfd`` ranges on shared +memory (covering all shmem APIs, i.e. tmpfs, ``IPCSHM``, ``/dev/zero``, +``MAP_SHARED``, ``memfd_create``, etc). + +The userland application that wants to use ``userfaultfd`` with hugetlbfs +or shared memory need to set the corresponding flag in +``uffdio_api.features`` to enable those features. + +If the userland desires to receive notifications for events other than +page faults, it has to verify that ``uffdio_api.features`` has appropriate +``UFFD_FEATURE_EVENT_*`` bits set. These events are described in more +detail below in `Non-cooperative userfaultfd`_ section. + +Once the ``userfaultfd`` has been enabled the ``UFFDIO_REGISTER`` ioctl should +be invoked (if present in the returned ``uffdio_api.ioctls`` bitmask) to +register a memory range in the ``userfaultfd`` by setting the +uffdio_register structure accordingly. The ``uffdio_register.mode`` +bitmask will specify to the kernel which kind of faults to track for +the range (``UFFDIO_REGISTER_MODE_MISSING`` would track missing +pages). The ``UFFDIO_REGISTER`` ioctl will return the +``uffdio_register.ioctls`` bitmask of ioctls that are suitable to resolve +userfaults on the range registered. Not all ioctls will necessarily be +supported for all memory types depending on the underlying virtual +memory backend (anonymous memory vs tmpfs vs real filebacked +mappings). + +Userland can use the ``uffdio_register.ioctls`` to manage the virtual +address space in the background (to add or potentially also remove +memory from the ``userfaultfd`` registered range). This means a userfault +could be triggering just before userland maps in the background the +user-faulted page. + +The primary ioctl to resolve userfaults is ``UFFDIO_COPY``. That +atomically copies a page into the userfault registered range and wakes +up the blocked userfaults +(unless ``uffdio_copy.mode & UFFDIO_COPY_MODE_DONTWAKE`` is set). +Other ioctl works similarly to ``UFFDIO_COPY``. They're atomic as in +guaranteeing that nothing can see an half copied page since it'll +keep userfaulting until the copy has finished. + +Notes: + +- If you requested ``UFFDIO_REGISTER_MODE_MISSING`` when registering then + you must provide some kind of page in your thread after reading from + the uffd. You must provide either ``UFFDIO_COPY`` or ``UFFDIO_ZEROPAGE``. + The normal behavior of the OS automatically providing a zero page on + an annonymous mmaping is not in place. + +- None of the page-delivering ioctls default to the range that you + registered with. You must fill in all fields for the appropriate + ioctl struct including the range. + +- You get the address of the access that triggered the missing page + event out of a struct uffd_msg that you read in the thread from the + uffd. You can supply as many pages as you want with ``UFFDIO_COPY`` or + ``UFFDIO_ZEROPAGE``. Keep in mind that unless you used DONTWAKE then + the first of any of those IOCTLs wakes up the faulting thread. + +- Be sure to test for all errors including + (``pollfd[0].revents & POLLERR``). This can happen, e.g. when ranges + supplied were incorrect. + +Write Protect Notifications +--------------------------- + +This is equivalent to (but faster than) using mprotect and a SIGSEGV +signal handler. + +Firstly you need to register a range with ``UFFDIO_REGISTER_MODE_WP``. +Instead of using mprotect(2) you use +``ioctl(uffd, UFFDIO_WRITEPROTECT, struct *uffdio_writeprotect)`` +while ``mode = UFFDIO_WRITEPROTECT_MODE_WP`` +in the struct passed in. The range does not default to and does not +have to be identical to the range you registered with. You can write +protect as many ranges as you like (inside the registered range). +Then, in the thread reading from uffd the struct will have +``msg.arg.pagefault.flags & UFFD_PAGEFAULT_FLAG_WP`` set. Now you send +``ioctl(uffd, UFFDIO_WRITEPROTECT, struct *uffdio_writeprotect)`` +again while ``pagefault.mode`` does not have ``UFFDIO_WRITEPROTECT_MODE_WP`` +set. This wakes up the thread which will continue to run with writes. This +allows you to do the bookkeeping about the write in the uffd reading +thread before the ioctl. + +If you registered with both ``UFFDIO_REGISTER_MODE_MISSING`` and +``UFFDIO_REGISTER_MODE_WP`` then you need to think about the sequence in +which you supply a page and undo write protect. Note that there is a +difference between writes into a WP area and into a !WP area. The +former will have ``UFFD_PAGEFAULT_FLAG_WP`` set, the latter +``UFFD_PAGEFAULT_FLAG_WRITE``. The latter did not fail on protection but +you still need to supply a page when ``UFFDIO_REGISTER_MODE_MISSING`` was +used. + +QEMU/KVM +======== + +QEMU/KVM is using the ``userfaultfd`` syscall to implement postcopy live +migration. Postcopy live migration is one form of memory +externalization consisting of a virtual machine running with part or +all of its memory residing on a different node in the cloud. The +``userfaultfd`` abstraction is generic enough that not a single line of +KVM kernel code had to be modified in order to add postcopy live +migration to QEMU. + +Guest async page faults, ``FOLL_NOWAIT`` and all other ``GUP*`` features work +just fine in combination with userfaults. Userfaults trigger async +page faults in the guest scheduler so those guest processes that +aren't waiting for userfaults (i.e. network bound) can keep running in +the guest vcpus. + +It is generally beneficial to run one pass of precopy live migration +just before starting postcopy live migration, in order to avoid +generating userfaults for readonly guest regions. + +The implementation of postcopy live migration currently uses one +single bidirectional socket but in the future two different sockets +will be used (to reduce the latency of the userfaults to the minimum +possible without having to decrease ``/proc/sys/net/ipv4/tcp_wmem``). + +The QEMU in the source node writes all pages that it knows are missing +in the destination node, into the socket, and the migration thread of +the QEMU running in the destination node runs ``UFFDIO_COPY|ZEROPAGE`` +ioctls on the ``userfaultfd`` in order to map the received pages into the +guest (``UFFDIO_ZEROCOPY`` is used if the source page was a zero page). + +A different postcopy thread in the destination node listens with +poll() to the ``userfaultfd`` in parallel. When a ``POLLIN`` event is +generated after a userfault triggers, the postcopy thread read() from +the ``userfaultfd`` and receives the fault address (or ``-EAGAIN`` in case the +userfault was already resolved and waken by a ``UFFDIO_COPY|ZEROPAGE`` run +by the parallel QEMU migration thread). + +After the QEMU postcopy thread (running in the destination node) gets +the userfault address it writes the information about the missing page +into the socket. The QEMU source node receives the information and +roughly "seeks" to that page address and continues sending all +remaining missing pages from that new page offset. Soon after that +(just the time to flush the tcp_wmem queue through the network) the +migration thread in the QEMU running in the destination node will +receive the page that triggered the userfault and it'll map it as +usual with the ``UFFDIO_COPY|ZEROPAGE`` (without actually knowing if it +was spontaneously sent by the source or if it was an urgent page +requested through a userfault). + +By the time the userfaults start, the QEMU in the destination node +doesn't need to keep any per-page state bitmap relative to the live +migration around and a single per-page bitmap has to be maintained in +the QEMU running in the source node to know which pages are still +missing in the destination node. The bitmap in the source node is +checked to find which missing pages to send in round robin and we seek +over it when receiving incoming userfaults. After sending each page of +course the bitmap is updated accordingly. It's also useful to avoid +sending the same page twice (in case the userfault is read by the +postcopy thread just before ``UFFDIO_COPY|ZEROPAGE`` runs in the migration +thread). + +Non-cooperative userfaultfd +=========================== + +When the ``userfaultfd`` is monitored by an external manager, the manager +must be able to track changes in the process virtual memory +layout. Userfaultfd can notify the manager about such changes using +the same read(2) protocol as for the page fault notifications. The +manager has to explicitly enable these events by setting appropriate +bits in ``uffdio_api.features`` passed to ``UFFDIO_API`` ioctl: + +``UFFD_FEATURE_EVENT_FORK`` + enable ``userfaultfd`` hooks for fork(). When this feature is + enabled, the ``userfaultfd`` context of the parent process is + duplicated into the newly created process. The manager + receives ``UFFD_EVENT_FORK`` with file descriptor of the new + ``userfaultfd`` context in the ``uffd_msg.fork``. + +``UFFD_FEATURE_EVENT_REMAP`` + enable notifications about mremap() calls. When the + non-cooperative process moves a virtual memory area to a + different location, the manager will receive + ``UFFD_EVENT_REMAP``. The ``uffd_msg.remap`` will contain the old and + new addresses of the area and its original length. + +``UFFD_FEATURE_EVENT_REMOVE`` + enable notifications about madvise(MADV_REMOVE) and + madvise(MADV_DONTNEED) calls. The event ``UFFD_EVENT_REMOVE`` will + be generated upon these calls to madvise(). The ``uffd_msg.remove`` + will contain start and end addresses of the removed area. + +``UFFD_FEATURE_EVENT_UNMAP`` + enable notifications about memory unmapping. The manager will + get ``UFFD_EVENT_UNMAP`` with ``uffd_msg.remove`` containing start and + end addresses of the unmapped area. + +Although the ``UFFD_FEATURE_EVENT_REMOVE`` and ``UFFD_FEATURE_EVENT_UNMAP`` +are pretty similar, they quite differ in the action expected from the +``userfaultfd`` manager. In the former case, the virtual memory is +removed, but the area is not, the area remains monitored by the +``userfaultfd``, and if a page fault occurs in that area it will be +delivered to the manager. The proper resolution for such page fault is +to zeromap the faulting address. However, in the latter case, when an +area is unmapped, either explicitly (with munmap() system call), or +implicitly (e.g. during mremap()), the area is removed and in turn the +``userfaultfd`` context for such area disappears too and the manager will +not get further userland page faults from the removed area. Still, the +notification is required in order to prevent manager from using +``UFFDIO_COPY`` on the unmapped area. + +Unlike userland page faults which have to be synchronous and require +explicit or implicit wakeup, all the events are delivered +asynchronously and the non-cooperative process resumes execution as +soon as manager executes read(). The ``userfaultfd`` manager should +carefully synchronize calls to ``UFFDIO_COPY`` with the events +processing. To aid the synchronization, the ``UFFDIO_COPY`` ioctl will +return ``-ENOSPC`` when the monitored process exits at the time of +``UFFDIO_COPY``, and ``-ENOENT``, when the non-cooperative process has changed +its virtual memory layout simultaneously with outstanding ``UFFDIO_COPY`` +operation. + +The current asynchronous model of the event delivery is optimal for +single threaded non-cooperative ``userfaultfd`` manager implementations. A +synchronous event delivery model can be added later as a new +``userfaultfd`` feature to facilitate multithreading enhancements of the +non cooperative manager, for example to allow ``UFFDIO_COPY`` ioctls to +run in parallel to the event reception. Single threaded +implementations should continue to use the current async event +delivery model instead. |