1 ============== 2 Page fragments 3 ============== 4 5 A page fragment is an arbitrary-length arbitrary-offset area of memory 6 which resides within a 0 or higher order compound page. Multiple 7 fragments within that page are individually refcounted, in the page's 8 reference counter. 9 10 The page_frag functions, page_frag_alloc and page_frag_free, provide a 11 simple allocation framework for page fragments. This is used by the 12 network stack and network device drivers to provide a backing region of 13 memory for use as either an sk_buff->head, or to be used in the "frags" 14 portion of skb_shared_info. 15 16 In order to make use of the page fragment APIs a backing page fragment 17 cache is needed. This provides a central point for the fragment allocation 18 and tracks allows multiple calls to make use of a cached page. The 19 advantage to doing this is that multiple calls to get_page can be avoided 20 which can be expensive at allocation time. However due to the nature of 21 this caching it is required that any calls to the cache be protected by 22 either a per-cpu limitation, or a per-cpu limitation and forcing interrupts 23 to be disabled when executing the fragment allocation. 24 25 The network stack uses two separate caches per CPU to handle fragment 26 allocation. The netdev_alloc_cache is used by callers making use of the 27 netdev_alloc_frag and __netdev_alloc_skb calls. The napi_alloc_cache is 28 used by callers of the __napi_alloc_frag and napi_alloc_skb calls. The 29 main difference between these two calls is the context in which they may be 30 called. The "netdev" prefixed functions are usable in any context as these 31 functions will disable interrupts, while the "napi" prefixed functions are 32 only usable within the softirq context. 33 34 Many network device drivers use a similar methodology for allocating page 35 fragments, but the page fragments are cached at the ring or descriptor 36 level. In order to enable these cases it is necessary to provide a generic 37 way of tearing down a page cache. For this reason __page_frag_cache_drain 38 was implemented. It allows for freeing multiple references from a single 39 page via a single call. The advantage to doing this is that it allows for 40 cleaning up the multiple references that were added to a page in order to 41 avoid calling get_page per allocation. 42 43 Alexander Duyck, Nov 29, 2016.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.