linux-master/mm/hugetlb.c

548 hugetlb_resv_map_add(struct resv_map *map, struct list_head *rg, long from,
555 		nrg = get_file_region_entry_from_cache(map, from, to);
556 		record_hugetlb_cgroup_uncharge_info(cg, h, map, nrg);
558 		coalesce_file_region(map, nrg);
691  * map.  Regions will be taken from the cache to fill in this range.
699  * Return the number of new huge pages added to the map.  This number is greater
754  * Examine the existing reserve map and determine how many
758  * map to add the specified range [f, t).  region_chg does
760  * map.  A number of new file_region structures is added to the cache as a
769  * reservation map for the range [f, t).  This number is greater or equal to
819  * Delete the specified range [f, t) from the reserve map.  If the
824  * Returns the number of huge pages deleted from the reserve map.
845 		 * may be a "placeholder" entry in the map which is of the form
928  * the reserve map region for a page.  The huge page itself was free'ed
931  * these counts, the reserve map entry which could not be deleted will
956  * Count and return the number of huge pages in the reserve map
1027  * bits of the reservation map pointer, which are always clear due to
1045  * manner to a shared mapping.  A shared mapping has a region map associated
1046  * with the underlying file, this region map represents the backing file
1048  * after the page is instantiated.  A private mapping has a region map
1050  * reference it, this region map represents those offsets which have consumed
1121 	/* Clear out any active regions before we release the map. */
1163 static void set_vma_resv_map(struct vm_area_struct *vma, struct resv_map *map)
1168 	set_vma_private_data(vma, (unsigned long)map);
1204 	 * - For MAP_PRIVATE mappings, this is the reserve map which does
1271 		 * be a region map for all pages.  The only situation where
1272 		 * there is no region map is if a hole was punched via
1293 		 * call to vma_needs_reserves().  The reserve map for
2718  *    to the associated reservation map.
2775  * to add the page to the reservation map.  If the page allocation fails,
2781  * is not the case is if a reserve map was changed between calls.  It
2791  * map was created during huge page allocation and must be removed.  It is to
2820 		 * 1 page, and that adding to resv map a 1 page entry can only
2867 	 * Subtle - The reserve map for private mappings has the
2869 	 * entry is in the reserve map, it means a reservation exists.
2870 	 * If an entry exists in the reserve map, it means the
2873 	 * value returned from reserve map manipulation routines above.
2921  *    not set.  However, alloc_hugetlb_folio always updates the reserve map.
2925  * to adjust the reservation map.  This case deals primarily with private
2926  * mappings.  Adjust the reserve map here to be consistent with global
2928  * reserve map indicates there is a reservation present.
2930  * In case 2, simply undo reserve map modifications done by alloc_hugetlb_folio.
2940 			 * Rare out of memory condition in reserve map
2958 			 * This indicates there is an entry in the reserve map
2982 			 * reserve map.
2984 			 * For shared mappings, no entry in the map indicates
3155 	 * Examine the region/reserve map to determine if the process
3169 	 * reserves as indicated by the region/reserve map. Check
3181 		 * map, there could be reservations associated with the
3243 		 * The page was added to the reservation map between
5179 	 * This new VMA should share its siblings reservation map if present.
5180 	 * The VMA will only ever have a valid reservation map pointer where
5182 	 * has a reference to the reservation map it cannot disappear until
7187 	 * When the last VMA disappears, the region map says how much
7192 	 * consumed reservations are stored in the map. Hence, nothing
7204 			 * map between region_chg and region_add.  This