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:
111 	 * compaction threads can race against page migration functions  in isolate_movable_page()
115 	 * being (wrongly) re-isolated while it is under migration,  in isolate_movable_page()
163  * from where they were once taken off for compaction/migration.
203  * Restore a potential migration pte to a working pte entry
227 		/* PMD-mapped THP migration entry */  in remove_migration_pte()
241 		 * Recheck VMA as permissions can change since migration started  in remove_migration_pte()
291  * Get rid of all migration entries and replace them by
308  * Something used the pte of a page under migration. We need to
309  * get to the page and wait until migration is finished.
331 	 * Once page cache replacement of page migration started, page_count  in __migration_entry_wait()
688  *                    Migration functions
741 			 * async migration. Release the taken locks  in buffer_migrate_lock_buffers()
837  * Migration function for pages with buffers. This function can only be used
886 	 * migration. Writeout may mean we loose the lock and the  in writeout()
888 	 * At this point we know that the migration attempt cannot  in writeout()
903  * Default handling if a filesystem does not provide a migration function.
909 		/* Only writeback pages in full synchronous migration */  in fallback_migrate_page()
963 			 * for page migration.  in move_to_new_page()
973 		 * isolation step. In that case, we shouldn't try migration.  in move_to_new_page()
1052 		 * Only in the case of a full synchronous migration is it  in __unmap_and_move()
1074 	 * of migration. File cache pages are no problem because of page_lock()  in __unmap_and_move()
1075 	 * File Caches may use write_page() or lock_page() in migration, then,  in __unmap_and_move()
1122 		/* Establish migration ptes */  in __unmap_and_move()
1146 	 * If migration is successful, decrease refcount of the newpage  in __unmap_and_move()
1221 	 * If migration is successful, releases reference grabbed during  in unmap_and_move()
1257  * Counterpart of unmap_and_move_page() for hugepage migration.
1260  * because there is no race between I/O and migration for hugepage.
1268  * hugepage migration fails without data corruption.
1270  * There is also no race when direct I/O is issued on the page under migration,
1271  * because then pte is replaced with migration swap entry and direct I/O code
1272  * will wait in the page fault for migration to complete.
1287 	 * This check is necessary because some callers of hugepage migration  in unmap_and_move_huge_page()
1290 	 * kicking migration.  in unmap_and_move_huge_page()
1383 	 * If migration was not successful and there's a freeing callback, use  in unmap_and_move_huge_page()
1397  *		   supplied as the target for the page migration
1401  *			as the target of the page migration.
1402  * @put_new_page:	The function used to free target pages if migration
1405  * @mode:		The migration mode that specifies the constraints for
1406  *			page migration, if any.
1407  * @reason:		The reason for page migration.
1446 			 * during migration.  in migrate_pages()
1464 				 * THP migration might be unsupported or the  in migrate_pages()
1509 				 * removed from migration page list and not  in migrate_pages()
1564 		 * clear __GFP_RECLAIM to make the migration callback  in alloc_migration_target()
1771 			/* The page is successfully queued for migration */  in do_pages_move()
1984  * Returns true if this is a safe migration target node for misplaced NUMA
2038 	 * migrate_misplaced_transhuge_page() skips page migration's usual  in numamigrate_isolate_page()
2040 	 * has been isolated: a GUP pin, or any other pin, prevents migration.  in numamigrate_isolate_page()
2056 	 * disappearing underneath us during migration.  in numamigrate_isolate_page()
2154 	/* Prepare a page as a migration target */  in migrate_misplaced_transhuge_page()
2419 		 * any kind of migration. Side effect is that it "freezes" the  in migrate_vma_collect_pmd()
2432 		 * set up a special migration page table entry now.  in migrate_vma_collect_pmd()
2440 			/* Setup special migration page table entry */  in migrate_vma_collect_pmd()
2490  * @migrate: migrate struct containing all migration information
2522  * migrate_page_move_mapping(), except that here we allow migration of a
2546 		 * GUP will fail for those. Yet if there is a pending migration  in migrate_vma_check_page()
2547 		 * a thread might try to wait on the pte migration entry and  in migrate_vma_check_page()
2549 		 * differentiate a regular pin from migration wait. Hence to  in migrate_vma_check_page()
2551 		 * infinite loop (one stoping migration because the other is  in migrate_vma_check_page()
2552 		 * waiting on pte migration entry). We always return true here.  in migrate_vma_check_page()
2572  * @migrate: migrate struct containing all migration information
2598 			 * a deadlock between 2 concurrent migration where each  in migrate_vma_prepare()
2679  * migrate_vma_unmap() - replace page mapping with special migration pte entry
2680  * @migrate: migrate struct containing all migration information
2682  * Replace page mapping (CPU page table pte) with a special migration pte entry
2738  * @args: contains the vma, start, and pfns arrays for the migration
2759  * with MIGRATE_PFN_MIGRATE flag in src array unless this is a migration from
2769  * properly set the destination entry like for regular migration.  Note that
2771  * migration was successful for those entries after calling migrate_vma_pages()
2772  * just like for regular migration.
2975  * @migrate: migrate struct containing all migration information
2978  * struct page. This effectively finishes the migration from source page to the
3058  * @migrate: migrate struct containing all migration information
3060  * This replaces the special migration pte entry with either a mapping to the
3061  * new page if migration was successful for that page, or to the original page