| /Linux-v5.15/Documentation/vm/ |
| D | page_migration.rst | 4 Page migration
7 Page migration allows moving the physical location of pages between
15 The main intent of page migration is to reduce the latency of memory accesses
19 Page migration allows a process to manually relocate the node on which its
25 Page migration functions are provided by the numactl package by Andi Kleen
29 migration. cat ``/proc/<pid>/numa_maps`` allows an easy review of where the
33 Manual migration is useful if for example the scheduler has relocated
37 manual page migration support. Automatic page migration may be implemented
54 Page migration allows the preservation of the relative location of pages
55 within a group of nodes for all migration techniques which will preserve a
[all …]
|
| D | hmm.rst | 25 Finally, the last section presents a new migration helper that allows
106 Shared address space and migration
126 main memory to device memory using existing migration mechanisms and everything
130 memory for the device memory and second to perform migration. Policy decisions
133 Note that any CPU access to a device page triggers a page fault and a migration
136 address A triggers a page fault and initiates a migration back to main memory.
271 Migration to and from device memory
306 The overall migration steps are similar to migrating NUMA pages within system
307 memory (see :ref:`Page migration <page_migration>`) but the steps are split
314 be held for the duration of the migration.
[all …]
|
| D | unevictable-lru.rst | 79 migration, unless we reworked migration code to find the unevictable pages
233 link field was not available to the migration subsystem.
412 replaced with an entry that refers to the new page. Linux supports migration
416 Note that page migration can race with mlocking or munlocking of the same page.
418 sections above. Both processes (migration and m[un]locking) hold the page
419 locked. This provides the first level of synchronization. Page migration
423 To complete page migration, we place the new and old pages back onto the LRU
425 page on failure - will be freed when the reference count held by the migration
427 list because of a race between munlock and migration, page migration uses the
438 the page migration code and the same work flow as described in MIGRATING
[all …]
|
| /Linux-v5.15/tools/testing/selftests/kvm/ |
| D | rseq_test.c | 36 * Any bug related to task migration is likely to be timing-dependent; perform
97 * that a migration may have occurred in between rseq and sched in migration_worker()
98 * CPU ID reads. An odd sequence count indicates a migration in migration_worker()
100 * a migration occurred since the count was last read. in migration_worker()
122 * 1. To allow sched_setaffinity() to prompt migration before in migration_worker()
139 * before the next round of migration. The test's check on in migration_worker()
140 * the rseq CPU ID must wait for migration to complete in in migration_worker()
142 * will be missed if the next migration starts before the in migration_worker()
220 print_skip("Only one usable CPU, task migration not possible"); in main()
242 * Verify rseq's CPU matches sched's CPU. Ensure migration in main()
[all …]
|
| /Linux-v5.15/mm/ |
| D | migrate.c | 3 * Memory Migration functionality - linux/mm/migrate.c
7 * Page migration was first developed in the context of the memory hotplug
8 * project. The main authors of the migration code are:
86 * compaction threads can race against page migration functions in isolate_movable_page()
90 * being (wrongly) re-isolated while it is under migration, in isolate_movable_page()
133 * from where they were once taken off for compaction/migration.
173 * Restore a potential migration pte to a working pte entry
197 /* PMD-mapped THP migration entry */ in remove_migration_pte()
211 * Recheck VMA as permissions can change since migration started in remove_migration_pte()
268 * Get rid of all migration entries and replace them by
[all …]
|
| D | compaction.c | 6 * this heavily depends upon page migration to do all the real heavy
144 * With it, VM migration doesn't try to put it back. in __ClearPageMovable()
306 * is necessary for the block to be a migration source/target. in __reset_isolation_pfn()
610 * so it is correct to skip the suitable migration target in isolate_freepages_block()
821 /* async migration should just abort */ in isolate_migratepages_block()
838 /* Time to isolate some pages for migration */ in isolate_migratepages_block()
843 * We have isolated all migration candidates in the in isolate_migratepages_block()
995 * Migration will fail if an anonymous page is pinned in memory, in isolate_migratepages_block()
1146 * scanning in the near future or migration failed in which case the in isolate_migratepages_block()
1238 /* Returns true if the page is within a block suitable for migration to */
[all …]
|
| D | Kconfig | 157 depends on MIGRATION
194 bool "Allow for balloon memory compaction/migration"
202 by the guest workload. Allowing the compaction & migration for memory
211 select MIGRATION
235 # support for page migration
237 config MIGRATION config
238 bool "Page migration"
242 Allows the migration of the physical location of pages of processes
246 pages as migration can relocate pages to satisfy a huge page
473 select MIGRATION
|
| /Linux-v5.15/include/linux/ |
| D | balloon_compaction.h | 7 * Balloon page migration makes use of the general non-lru movable page
11 * page->mapping is used in context of non-lru page migration to reference
12 * the address space operations for page isolation/migration/compaction.
17 * and safely perform balloon's page compaction and migration we must, always,
33 * compaction / migration.
51 * have to cope for page compaction / migration, as well as it serves the
55 unsigned long isolated_pages; /* # of isolated pages for migration */
|
| /Linux-v5.15/Documentation/admin-guide/mm/ |
| D | userfaultfd.rst | 193 migration. Postcopy live migration is one form of memory
198 migration to QEMU.
206 It is generally beneficial to run one pass of precopy live migration
207 just before starting postcopy live migration, in order to avoid
210 The implementation of postcopy live migration currently uses one
216 in the destination node, into the socket, and the migration thread of
226 by the parallel QEMU migration thread).
234 migration thread in the QEMU running in the destination node will
242 migration around and a single per-page bitmap has to be maintained in
249 postcopy thread just before ``UFFDIO_COPY|ZEROPAGE`` runs in the migration
|
| /Linux-v5.15/Documentation/admin-guide/cgroup-v1/ |
| D | memcg_test.rst | 102 The logic is very clear. (About migration, see below)
129 7. Page Migration
162 9.3 Migration
165 For NUMA, migration is an another special case. To do easy test, cpuset
166 is useful. Following is a sample script to do migration::
179 In above set, when you moves a task from 01 to 02, page migration to
211 This is an easy way to test page migration, too.
298 9.9 Move charges at task migration
301 Charges associated with a task can be moved along with task migration.
|
| /Linux-v5.15/Documentation/ABI/testing/ |
| D | sysfs-kernel-mm-numa | 11 Page migration during reclaim is intended for systems
16 Allowing page migration during reclaim enables these
18 when the fast tier is under pressure. This migration
|
| /Linux-v5.15/kernel/cgroup/ |
| D | cgroup-internal.h | 88 /* used to track tasks and csets during migration */
103 * Before migration is committed, the target migration tasks are on
106 * or ->dst_csets depending on whether migration is committed.
116 /* migration context also tracks preloading */
120 * atomic success or failure on actual migration.
128 /* subsystems affected by migration */
|
| /Linux-v5.15/Documentation/virt/kvm/devices/ |
| D | vm.rst | 288 Allows userspace to stop migration mode, needed for PGSTE migration.
289 Setting this attribute when migration mode is not active will have no
298 Allows userspace to start migration mode, needed for PGSTE migration.
299 Setting this attribute when migration mode is already active will have
303 :Returns: -ENOMEM if there is not enough free memory to start migration mode;
310 Allows userspace to query the status of migration mode.
313 the data itself is either 0 if migration mode is disabled or 1
|
| /Linux-v5.15/include/uapi/linux/ |
| D | vfio.h | 413 * migration information. Field accesses from this structure are only supported
447 * the migration region should be used to resume the device.
478 * 3. Save the state during live migration
496 * migration region. The sequence to be followed for such transition is given
498 * 3. In live migration of user application, the state transitions from _RUNNING
502 * the device state data to application through the migration region.
505 * through the migration region.
508 * user must not be required to consume all migration data before the device
514 * received through the migration region to resume the device.
525 * The user application should read data_offset field from the migration
[all …]
|
| /Linux-v5.15/tools/perf/scripts/python/bin/ |
| D | sched-migration-report | 2 # description: sched migration overview
3 perf script $@ -s "$PERF_EXEC_PATH"/scripts/python/sched-migration.py
|
| /Linux-v5.15/Documentation/networking/ |
| D | net_failover.rst | 24 datapath. It also enables hypervisor controlled live migration of a VM with
76 Live Migration of a VM with SR-IOV VF & virtio-net in STANDBY mode
79 net_failover also enables hypervisor controlled live migration to be supported
83 Here is a sample script that shows the steps to initiate live migration on
|
| /Linux-v5.15/fs/ocfs2/dlm/ |
| D | dlmmaster.c | 822 /* if there is a migration in progress, let the migration in dlm_get_lock_resource()
824 * of the MIGRATION mle: either the migrate finished or in dlm_get_lock_resource()
836 mig ? "MIGRATION" : "BLOCK"); in dlm_get_lock_resource()
1591 mlog(0, "migration mle was found (%u->%u)\n", in dlm_master_request_handler()
1824 " from %u for migration\n", in dlm_assert_master_handler()
1829 " from %u for migration, ignoring\n", in dlm_assert_master_handler()
1931 mlog(0, "finishing off migration of lockres %.*s, " in dlm_assert_master_handler()
1977 * ref given by the master / migration request message. in dlm_assert_master_handler()
2119 * prevent migration from starting while we're still asserting in dlm_assert_master_worker()
2120 * our dominance. The reserved ast delays migration. in dlm_assert_master_worker()
[all …]
|
| /Linux-v5.15/Documentation/admin-guide/nfs/ |
| D | nfs-client.rst | 28 and transparent state migration, this identity string must not change
42 or loss of state during NFS reboot recovery or transparent state migration.
65 attribute. See `RFC3530 Section 6: Filesystem Migration and Replication`_ and
68 .. _RFC3530 Section 6\: Filesystem Migration and Replication: https://tools.ietf.org/html/rfc3530#s…
|
| /Linux-v5.15/drivers/gpu/drm/amd/amdkfd/ |
| D | kfd_svm.h | 71 * @migrate_mutex: to serialize range migration, validation and mapping update
93 * @granularity:migration granularity, log2 num pages
192 /* SVM API and HMM page migration work together, device memory type
193 * is initialized to not 0 when page migration register device memory.
|
| /Linux-v5.15/drivers/misc/sgi-gru/ |
| D | grutlbpurge.c | 106 * - intrablade process migration between cpus is not frequent but is
109 * blade because of intrablade migration
110 * - interblade migration is rare. Processes migrate their GRU context to
112 * - if interblade migration occurs, migration back to the original blade
|
| /Linux-v5.15/drivers/gpu/drm/i915/gem/selftests/ |
| D | i915_gem_migrate.c | 114 pr_err("Object failed migration to smem\n"); in lmem_pages_migrate_one()
132 pr_err("Object failed migration to lmem\n"); in lmem_pages_migrate_one()
215 /* Finally sync migration and check content. */ in igt_lmem_pages_migrate()
|
| /Linux-v5.15/Documentation/admin-guide/device-mapper/ |
| D | cache.rst | 31 Migration
35 Migration from slow device to fast device.
37 Migration from fast device to slow device.
118 Migration throttling
123 migration occurring at any one time. Currently we're not taking any
165 We can avoid copying data during migration if we know the block has
|
| /Linux-v5.15/arch/mips/include/asm/sn/sn0/ |
| D | hubmd.h | 41 #define MD_MIG_CANDIDATE 0x200040 /* Latest page migration candidate */
42 #define MD_MIG_CANDIDATE_CLR 0x200048 /* Clear page migration candidate */
384 /* Protection and migration field values */
395 #define MD_PPROT_MIGMD_SHFT 3 /* Migration mode */
408 #define MD_SPROT_MIGMD_SHFT 3 /* Migration mode */
414 /* Migration modes used in protection entries */
423 * Operations on page migration threshold register
455 * Operations on page migration count difference and absolute threshold
504 * Operations on page migration candidate register
|
| /Linux-v5.15/tools/testing/selftests/drivers/net/netdevsim/ |
| D | nexthop.sh | 420 log_test "Bucket migration after idle timer"
455 log_test "Bucket migration after idle timer (with delete)"
490 log_test "Bucket migration after $timer increase"
519 log_test "Bucket migration after $timer decrease"
556 log_test "Bucket migration after $timer increase"
597 log_test "Bucket migration after unbalanced timer"
634 log_test "Bucket migration after unbalanced timer (with delete)"
684 log_test "Bucket migration after unbalanced < idle timer"
|
| /Linux-v5.15/drivers/md/ |
| D | dm-cache-policy.h | 54 * if a migration has just been queued.
60 * Sometimes the core target can optimise a migration, eg, the
62 * In order to optimise it needs the migration immediately though
|
