About Kernel Documentation Linux Kernel Contact Linux Resources Linux Blog

Documentation / block / biodoc.txt

Custom Search

Based on kernel version 4.2. Page generated on 2015-09-09 12:08 EST.

1		Notes on the Generic Block Layer Rewrite in Linux 2.5
2		=====================================================
4	Notes Written on Jan 15, 2002:
5		Jens Axboe <jens.axboe@oracle.com>
6		Suparna Bhattacharya <suparna@in.ibm.com>
8	Last Updated May 2, 2002
9	September 2003: Updated I/O Scheduler portions
10		Nick Piggin <npiggin@kernel.dk>
12	Introduction:
14	These are some notes describing some aspects of the 2.5 block layer in the
15	context of the bio rewrite. The idea is to bring out some of the key
16	changes and a glimpse of the rationale behind those changes.
18	Please mail corrections & suggestions to suparna@in.ibm.com.
20	Credits:
21	---------
23	2.5 bio rewrite:
24		Jens Axboe <jens.axboe@oracle.com>
26	Many aspects of the generic block layer redesign were driven by and evolved
27	over discussions, prior patches and the collective experience of several
28	people. See sections 8 and 9 for a list of some related references.
30	The following people helped with review comments and inputs for this
31	document:
32		Christoph Hellwig <hch@infradead.org>
33		Arjan van de Ven <arjanv@redhat.com>
34		Randy Dunlap <rdunlap@xenotime.net>
35		Andre Hedrick <andre@linux-ide.org>
37	The following people helped with fixes/contributions to the bio patches
38	while it was still work-in-progress:
39		David S. Miller <davem@redhat.com>
42	Description of Contents:
43	------------------------
45	1. Scope for tuning of logic to various needs
46	  1.1 Tuning based on device or low level driver capabilities
47		- Per-queue parameters
48		- Highmem I/O support
49		- I/O scheduler modularization
50	  1.2 Tuning based on high level requirements/capabilities
51		1.2.1 Request Priority/Latency
52	  1.3 Direct access/bypass to lower layers for diagnostics and special
53	      device operations
54		1.3.1 Pre-built commands
55	2. New flexible and generic but minimalist i/o structure or descriptor
56	   (instead of using buffer heads at the i/o layer)
57	  2.1 Requirements/Goals addressed
58	  2.2 The bio struct in detail (multi-page io unit)
59	  2.3 Changes in the request structure
60	3. Using bios
61	  3.1 Setup/teardown (allocation, splitting)
62	  3.2 Generic bio helper routines
63	    3.2.1 Traversing segments and completion units in a request
64	    3.2.2 Setting up DMA scatterlists
65	    3.2.3 I/O completion
66	    3.2.4 Implications for drivers that do not interpret bios (don't handle
67	 	  multiple segments)
68	    3.2.5 Request command tagging
69	  3.3 I/O submission
70	4. The I/O scheduler
71	5. Scalability related changes
72	  5.1 Granular locking: Removal of io_request_lock
73	  5.2 Prepare for transition to 64 bit sector_t
74	6. Other Changes/Implications
75	  6.1 Partition re-mapping handled by the generic block layer
76	7. A few tips on migration of older drivers
77	8. A list of prior/related/impacted patches/ideas
78	9. Other References/Discussion Threads
80	---------------------------------------------------------------------------
82	Bio Notes
83	--------
85	Let us discuss the changes in the context of how some overall goals for the
86	block layer are addressed.
88	1. Scope for tuning the generic logic to satisfy various requirements
90	The block layer design supports adaptable abstractions to handle common
91	processing with the ability to tune the logic to an appropriate extent
92	depending on the nature of the device and the requirements of the caller.
93	One of the objectives of the rewrite was to increase the degree of tunability
94	and to enable higher level code to utilize underlying device/driver
95	capabilities to the maximum extent for better i/o performance. This is
96	important especially in the light of ever improving hardware capabilities
97	and application/middleware software designed to take advantage of these
98	capabilities.
100	1.1 Tuning based on low level device / driver capabilities
102	Sophisticated devices with large built-in caches, intelligent i/o scheduling
103	optimizations, high memory DMA support, etc may find some of the
104	generic processing an overhead, while for less capable devices the
105	generic functionality is essential for performance or correctness reasons.
106	Knowledge of some of the capabilities or parameters of the device should be
107	used at the generic block layer to take the right decisions on
108	behalf of the driver.
110	How is this achieved ?
112	Tuning at a per-queue level:
114	i. Per-queue limits/values exported to the generic layer by the driver
116	Various parameters that the generic i/o scheduler logic uses are set at
117	a per-queue level (e.g maximum request size, maximum number of segments in
118	a scatter-gather list, hardsect size)
120	Some parameters that were earlier available as global arrays indexed by
121	major/minor are now directly associated with the queue. Some of these may
122	move into the block device structure in the future. Some characteristics
123	have been incorporated into a queue flags field rather than separate fields
124	in themselves.  There are blk_queue_xxx functions to set the parameters,
125	rather than update the fields directly
127	Some new queue property settings:
129		blk_queue_bounce_limit(q, u64 dma_address)
130			Enable I/O to highmem pages, dma_address being the
131			limit. No highmem default.
133		blk_queue_max_sectors(q, max_sectors)
134			Sets two variables that limit the size of the request.
136			- The request queue's max_sectors, which is a soft size in
137			units of 512 byte sectors, and could be dynamically varied
138			by the core kernel.
140			- The request queue's max_hw_sectors, which is a hard limit
141			and reflects the maximum size request a driver can handle
142			in units of 512 byte sectors.
144			The default for both max_sectors and max_hw_sectors is
145			255. The upper limit of max_sectors is 1024.
147		blk_queue_max_phys_segments(q, max_segments)
148			Maximum physical segments you can handle in a request. 128
149			default (driver limit). (See 3.2.2)
151		blk_queue_max_hw_segments(q, max_segments)
152			Maximum dma segments the hardware can handle in a request. 128
153			default (host adapter limit, after dma remapping).
154			(See 3.2.2)
156		blk_queue_max_segment_size(q, max_seg_size)
157			Maximum size of a clustered segment, 64kB default.
159		blk_queue_hardsect_size(q, hardsect_size)
160			Lowest possible sector size that the hardware can operate
161			on, 512 bytes default.
163	New queue flags:
165		QUEUE_FLAG_CLUSTER (see 3.2.2)
166		QUEUE_FLAG_QUEUED (see 3.2.4)
169	ii. High-mem i/o capabilities are now considered the default
171	The generic bounce buffer logic, present in 2.4, where the block layer would
172	by default copyin/out i/o requests on high-memory buffers to low-memory buffers
173	assuming that the driver wouldn't be able to handle it directly, has been
174	changed in 2.5. The bounce logic is now applied only for memory ranges
175	for which the device cannot handle i/o. A driver can specify this by
176	setting the queue bounce limit for the request queue for the device
177	(blk_queue_bounce_limit()). This avoids the inefficiencies of the copyin/out
178	where a device is capable of handling high memory i/o.
180	In order to enable high-memory i/o where the device is capable of supporting
181	it, the pci dma mapping routines and associated data structures have now been
182	modified to accomplish a direct page -> bus translation, without requiring
183	a virtual address mapping (unlike the earlier scheme of virtual address
184	-> bus translation). So this works uniformly for high-memory pages (which
185	do not have a corresponding kernel virtual address space mapping) and
186	low-memory pages.
188	Note: Please refer to Documentation/DMA-API-HOWTO.txt for a discussion
189	on PCI high mem DMA aspects and mapping of scatter gather lists, and support
190	for 64 bit PCI.
192	Special handling is required only for cases where i/o needs to happen on
193	pages at physical memory addresses beyond what the device can support. In these
194	cases, a bounce bio representing a buffer from the supported memory range
195	is used for performing the i/o with copyin/copyout as needed depending on
196	the type of the operation.  For example, in case of a read operation, the
197	data read has to be copied to the original buffer on i/o completion, so a
198	callback routine is set up to do this, while for write, the data is copied
199	from the original buffer to the bounce buffer prior to issuing the
200	operation. Since an original buffer may be in a high memory area that's not
201	mapped in kernel virtual addr, a kmap operation may be required for
202	performing the copy, and special care may be needed in the completion path
203	as it may not be in irq context. Special care is also required (by way of
204	GFP flags) when allocating bounce buffers, to avoid certain highmem
205	deadlock possibilities.
207	It is also possible that a bounce buffer may be allocated from high-memory
208	area that's not mapped in kernel virtual addr, but within the range that the
209	device can use directly; so the bounce page may need to be kmapped during
210	copy operations. [Note: This does not hold in the current implementation,
211	though]
213	There are some situations when pages from high memory may need to
214	be kmapped, even if bounce buffers are not necessary. For example a device
215	may need to abort DMA operations and revert to PIO for the transfer, in
216	which case a virtual mapping of the page is required. For SCSI it is also
217	done in some scenarios where the low level driver cannot be trusted to
218	handle a single sg entry correctly. The driver is expected to perform the
219	kmaps as needed on such occasions using the __bio_kmap_atomic and bio_kmap_irq
220	routines as appropriate. A driver could also use the blk_queue_bounce()
221	routine on its own to bounce highmem i/o to low memory for specific requests
222	if so desired.
224	iii. The i/o scheduler algorithm itself can be replaced/set as appropriate
226	As in 2.4, it is possible to plugin a brand new i/o scheduler for a particular
227	queue or pick from (copy) existing generic schedulers and replace/override
228	certain portions of it. The 2.5 rewrite provides improved modularization
229	of the i/o scheduler. There are more pluggable callbacks, e.g for init,
230	add request, extract request, which makes it possible to abstract specific
231	i/o scheduling algorithm aspects and details outside of the generic loop.
232	It also makes it possible to completely hide the implementation details of
233	the i/o scheduler from block drivers.
235	I/O scheduler wrappers are to be used instead of accessing the queue directly.
236	See section 4. The I/O scheduler for details.
238	1.2 Tuning Based on High level code capabilities
240	i. Application capabilities for raw i/o
242	This comes from some of the high-performance database/middleware
243	requirements where an application prefers to make its own i/o scheduling
244	decisions based on an understanding of the access patterns and i/o
245	characteristics
247	ii. High performance filesystems or other higher level kernel code's
248	capabilities
250	Kernel components like filesystems could also take their own i/o scheduling
251	decisions for optimizing performance. Journalling filesystems may need
252	some control over i/o ordering.
254	What kind of support exists at the generic block layer for this ?
256	The flags and rw fields in the bio structure can be used for some tuning
257	from above e.g indicating that an i/o is just a readahead request, or priority
258	settings (currently unused). As far as user applications are concerned they
259	would need an additional mechanism either via open flags or ioctls, or some
260	other upper level mechanism to communicate such settings to block.
262	1.2.1 Request Priority/Latency
264	Todo/Under discussion:
265	Arjan's proposed request priority scheme allows higher levels some broad
266	  control (high/med/low) over the priority  of an i/o request vs other pending
267	  requests in the queue. For example it allows reads for bringing in an
268	  executable page on demand to be given a higher priority over pending write
269	  requests which haven't aged too much on the queue. Potentially this priority
270	  could even be exposed to applications in some manner, providing higher level
271	  tunability. Time based aging avoids starvation of lower priority
272	  requests. Some bits in the bi_rw flags field in the bio structure are
273	  intended to be used for this priority information.
276	1.3 Direct Access to Low level Device/Driver Capabilities (Bypass mode)
277	    (e.g Diagnostics, Systems Management)
279	There are situations where high-level code needs to have direct access to
280	the low level device capabilities or requires the ability to issue commands
281	to the device bypassing some of the intermediate i/o layers.
282	These could, for example, be special control commands issued through ioctl
283	interfaces, or could be raw read/write commands that stress the drive's
284	capabilities for certain kinds of fitness tests. Having direct interfaces at
285	multiple levels without having to pass through upper layers makes
286	it possible to perform bottom up validation of the i/o path, layer by
287	layer, starting from the media.
289	The normal i/o submission interfaces, e.g submit_bio, could be bypassed
290	for specially crafted requests which such ioctl or diagnostics
291	interfaces would typically use, and the elevator add_request routine
292	can instead be used to directly insert such requests in the queue or preferably
293	the blk_do_rq routine can be used to place the request on the queue and
294	wait for completion. Alternatively, sometimes the caller might just
295	invoke a lower level driver specific interface with the request as a
296	parameter.
298	If the request is a means for passing on special information associated with
299	the command, then such information is associated with the request->special
300	field (rather than misuse the request->buffer field which is meant for the
301	request data buffer's virtual mapping).
303	For passing request data, the caller must build up a bio descriptor
304	representing the concerned memory buffer if the underlying driver interprets
305	bio segments or uses the block layer end*request* functions for i/o
306	completion. Alternatively one could directly use the request->buffer field to
307	specify the virtual address of the buffer, if the driver expects buffer
308	addresses passed in this way and ignores bio entries for the request type
309	involved. In the latter case, the driver would modify and manage the
310	request->buffer, request->sector and request->nr_sectors or
311	request->current_nr_sectors fields itself rather than using the block layer
312	end_request or end_that_request_first completion interfaces.
313	(See 2.3 or Documentation/block/request.txt for a brief explanation of
314	the request structure fields)
316	[TBD: end_that_request_last should be usable even in this case;
317	Perhaps an end_that_direct_request_first routine could be implemented to make
318	handling direct requests easier for such drivers; Also for drivers that
319	expect bios, a helper function could be provided for setting up a bio
320	corresponding to a data buffer]
322	<JENS: I dont understand the above, why is end_that_request_first() not
323	usable? Or _last for that matter. I must be missing something>
324	<SUP: What I meant here was that if the request doesn't have a bio, then
325	 end_that_request_first doesn't modify nr_sectors or current_nr_sectors,
326	 and hence can't be used for advancing request state settings on the
327	 completion of partial transfers. The driver has to modify these fields 
328	 directly by hand.
329	 This is because end_that_request_first only iterates over the bio list,
330	 and always returns 0 if there are none associated with the request.
331	 _last works OK in this case, and is not a problem, as I mentioned earlier
332	>
334	1.3.1 Pre-built Commands
336	A request can be created with a pre-built custom command  to be sent directly
337	to the device. The cmd block in the request structure has room for filling
338	in the command bytes. (i.e rq->cmd is now 16 bytes in size, and meant for
339	command pre-building, and the type of the request is now indicated
340	through rq->flags instead of via rq->cmd)
342	The request structure flags can be set up to indicate the type of request
343	in such cases (REQ_PC: direct packet command passed to driver, REQ_BLOCK_PC:
344	packet command issued via blk_do_rq, REQ_SPECIAL: special request).
346	It can help to pre-build device commands for requests in advance.
347	Drivers can now specify a request prepare function (q->prep_rq_fn) that the
348	block layer would invoke to pre-build device commands for a given request,
349	or perform other preparatory processing for the request. This is routine is
350	called by elv_next_request(), i.e. typically just before servicing a request.
351	(The prepare function would not be called for requests that have REQ_DONTPREP
352	enabled)
354	Aside:
355	  Pre-building could possibly even be done early, i.e before placing the
356	  request on the queue, rather than construct the command on the fly in the
357	  driver while servicing the request queue when it may affect latencies in
358	  interrupt context or responsiveness in general. One way to add early
359	  pre-building would be to do it whenever we fail to merge on a request.
360	  Now REQ_NOMERGE is set in the request flags to skip this one in the future,
361	  which means that it will not change before we feed it to the device. So
362	  the pre-builder hook can be invoked there.
365	2. Flexible and generic but minimalist i/o structure/descriptor.
367	2.1 Reason for a new structure and requirements addressed
369	Prior to 2.5, buffer heads were used as the unit of i/o at the generic block
370	layer, and the low level request structure was associated with a chain of
371	buffer heads for a contiguous i/o request. This led to certain inefficiencies
372	when it came to large i/o requests and readv/writev style operations, as it
373	forced such requests to be broken up into small chunks before being passed
374	on to the generic block layer, only to be merged by the i/o scheduler
375	when the underlying device was capable of handling the i/o in one shot.
376	Also, using the buffer head as an i/o structure for i/os that didn't originate
377	from the buffer cache unnecessarily added to the weight of the descriptors
378	which were generated for each such chunk.
380	The following were some of the goals and expectations considered in the
381	redesign of the block i/o data structure in 2.5.
383	i.  Should be appropriate as a descriptor for both raw and buffered i/o  -
384	    avoid cache related fields which are irrelevant in the direct/page i/o path,
385	    or filesystem block size alignment restrictions which may not be relevant
386	    for raw i/o.
387	ii. Ability to represent high-memory buffers (which do not have a virtual
388	    address mapping in kernel address space).
389	iii.Ability to represent large i/os w/o unnecessarily breaking them up (i.e
390	    greater than PAGE_SIZE chunks in one shot)
391	iv. At the same time, ability to retain independent identity of i/os from
392	    different sources or i/o units requiring individual completion (e.g. for
393	    latency reasons)
394	v.  Ability to represent an i/o involving multiple physical memory segments
395	    (including non-page aligned page fragments, as specified via readv/writev)
396	    without unnecessarily breaking it up, if the underlying device is capable of
397	    handling it.
398	vi. Preferably should be based on a memory descriptor structure that can be
399	    passed around different types of subsystems or layers, maybe even
400	    networking, without duplication or extra copies of data/descriptor fields
401	    themselves in the process
402	vii.Ability to handle the possibility of splits/merges as the structure passes
403	    through layered drivers (lvm, md, evms), with minimal overhead.
405	The solution was to define a new structure (bio)  for the block layer,
406	instead of using the buffer head structure (bh) directly, the idea being
407	avoidance of some associated baggage and limitations. The bio structure
408	is uniformly used for all i/o at the block layer ; it forms a part of the
409	bh structure for buffered i/o, and in the case of raw/direct i/o kiobufs are
410	mapped to bio structures.
412	2.2 The bio struct
414	The bio structure uses a vector representation pointing to an array of tuples
415	of <page, offset, len> to describe the i/o buffer, and has various other
416	fields describing i/o parameters and state that needs to be maintained for
417	performing the i/o.
419	Notice that this representation means that a bio has no virtual address
420	mapping at all (unlike buffer heads).
422	struct bio_vec {
423	       struct page     *bv_page;
424	       unsigned short  bv_len;
425	       unsigned short  bv_offset;
426	};
428	/*
429	 * main unit of I/O for the block layer and lower layers (ie drivers)
430	 */
431	struct bio {
432	       struct bio          *bi_next;    /* request queue link */
433	       struct block_device *bi_bdev;	/* target device */
434	       unsigned long       bi_flags;    /* status, command, etc */
435	       unsigned long       bi_rw;       /* low bits: r/w, high: priority */
437	       unsigned int	bi_vcnt;     /* how may bio_vec's */
438	       struct bvec_iter	bi_iter;	/* current index into bio_vec array */
440	       unsigned int	bi_size;     /* total size in bytes */
441	       unsigned short 	bi_phys_segments; /* segments after physaddr coalesce*/
442	       unsigned short	bi_hw_segments; /* segments after DMA remapping */
443	       unsigned int	bi_max;	     /* max bio_vecs we can hold
444	                                        used as index into pool */
445	       struct bio_vec   *bi_io_vec;  /* the actual vec list */
446	       bio_end_io_t	*bi_end_io;  /* bi_end_io (bio) */
447	       atomic_t		bi_cnt;	     /* pin count: free when it hits zero */
448	       void             *bi_private;
449	};
451	With this multipage bio design:
453	- Large i/os can be sent down in one go using a bio_vec list consisting
454	  of an array of <page, offset, len> fragments (similar to the way fragments
455	  are represented in the zero-copy network code)
456	- Splitting of an i/o request across multiple devices (as in the case of
457	  lvm or raid) is achieved by cloning the bio (where the clone points to
458	  the same bi_io_vec array, but with the index and size accordingly modified)
459	- A linked list of bios is used as before for unrelated merges (*) - this
460	  avoids reallocs and makes independent completions easier to handle.
461	- Code that traverses the req list can find all the segments of a bio
462	  by using rq_for_each_segment.  This handles the fact that a request
463	  has multiple bios, each of which can have multiple segments.
464	- Drivers which can't process a large bio in one shot can use the bi_iter
465	  field to keep track of the next bio_vec entry to process.
466	  (e.g a 1MB bio_vec needs to be handled in max 128kB chunks for IDE)
467	  [TBD: Should preferably also have a bi_voffset and bi_vlen to avoid modifying
468	   bi_offset an len fields]
470	(*) unrelated merges -- a request ends up containing two or more bios that
471	    didn't originate from the same place.
473	bi_end_io() i/o callback gets called on i/o completion of the entire bio.
475	At a lower level, drivers build a scatter gather list from the merged bios.
476	The scatter gather list is in the form of an array of <page, offset, len>
477	entries with their corresponding dma address mappings filled in at the
478	appropriate time. As an optimization, contiguous physical pages can be
479	covered by a single entry where <page> refers to the first page and <len>
480	covers the range of pages (up to 16 contiguous pages could be covered this
481	way). There is a helper routine (blk_rq_map_sg) which drivers can use to build
482	the sg list.
484	Note: Right now the only user of bios with more than one page is ll_rw_kio,
485	which in turn means that only raw I/O uses it (direct i/o may not work
486	right now). The intent however is to enable clustering of pages etc to
487	become possible. The pagebuf abstraction layer from SGI also uses multi-page
488	bios, but that is currently not included in the stock development kernels.
489	The same is true of Andrew Morton's work-in-progress multipage bio writeout 
490	and readahead patches.
492	2.3 Changes in the Request Structure
494	The request structure is the structure that gets passed down to low level
495	drivers. The block layer make_request function builds up a request structure,
496	places it on the queue and invokes the drivers request_fn. The driver makes
497	use of block layer helper routine elv_next_request to pull the next request
498	off the queue. Control or diagnostic functions might bypass block and directly
499	invoke underlying driver entry points passing in a specially constructed
500	request structure.
502	Only some relevant fields (mainly those which changed or may be referred
503	to in some of the discussion here) are listed below, not necessarily in
504	the order in which they occur in the structure (see include/linux/blkdev.h)
505	Refer to Documentation/block/request.txt for details about all the request
506	structure fields and a quick reference about the layers which are
507	supposed to use or modify those fields.
509	struct request {
510		struct list_head queuelist;  /* Not meant to be directly accessed by
511						the driver.
512						Used by q->elv_next_request_fn
513						rq->queue is gone
514						*/
515		.
516		.
517		unsigned char cmd[16]; /* prebuilt command data block */
518		unsigned long flags;   /* also includes earlier rq->cmd settings */
519		.
520		.
521		sector_t sector; /* this field is now of type sector_t instead of int
522				    preparation for 64 bit sectors */
523		.
524		.
526		/* Number of scatter-gather DMA addr+len pairs after
527		 * physical address coalescing is performed.
528		 */
529		unsigned short nr_phys_segments;
531		/* Number of scatter-gather addr+len pairs after
532		 * physical and DMA remapping hardware coalescing is performed.
533		 * This is the number of scatter-gather entries the driver
534		 * will actually have to deal with after DMA mapping is done.
535		 */
536		unsigned short nr_hw_segments;
538		/* Various sector counts */
539		unsigned long nr_sectors;  /* no. of sectors left: driver modifiable */
540		unsigned long hard_nr_sectors;  /* block internal copy of above */
541		unsigned int current_nr_sectors; /* no. of sectors left in the
542						   current segment:driver modifiable */
543		unsigned long hard_cur_sectors; /* block internal copy of the above */
544		.
545		.
546		int tag;	/* command tag associated with request */
547		void *special;  /* same as before */
548		char *buffer;   /* valid only for low memory buffers up to
549				 current_nr_sectors */
550		.
551		.
552		struct bio *bio, *biotail;  /* bio list instead of bh */
553		struct request_list *rl;
554	}
556	See the rq_flag_bits definitions for an explanation of the various flags
557	available. Some bits are used by the block layer or i/o scheduler.
559	The behaviour of the various sector counts are almost the same as before,
560	except that since we have multi-segment bios, current_nr_sectors refers
561	to the numbers of sectors in the current segment being processed which could
562	be one of the many segments in the current bio (i.e i/o completion unit).
563	The nr_sectors value refers to the total number of sectors in the whole
564	request that remain to be transferred (no change). The purpose of the
565	hard_xxx values is for block to remember these counts every time it hands
566	over the request to the driver. These values are updated by block on
567	end_that_request_first, i.e. every time the driver completes a part of the
568	transfer and invokes block end*request helpers to mark this. The
569	driver should not modify these values. The block layer sets up the
570	nr_sectors and current_nr_sectors fields (based on the corresponding
571	hard_xxx values and the number of bytes transferred) and updates it on
572	every transfer that invokes end_that_request_first. It does the same for the
573	buffer, bio, bio->bi_iter fields too.
575	The buffer field is just a virtual address mapping of the current segment
576	of the i/o buffer in cases where the buffer resides in low-memory. For high
577	memory i/o, this field is not valid and must not be used by drivers.
579	Code that sets up its own request structures and passes them down to
580	a driver needs to be careful about interoperation with the block layer helper
581	functions which the driver uses. (Section 1.3)
583	3. Using bios
585	3.1 Setup/Teardown
587	There are routines for managing the allocation, and reference counting, and
588	freeing of bios (bio_alloc, bio_get, bio_put).
590	This makes use of Ingo Molnar's mempool implementation, which enables
591	subsystems like bio to maintain their own reserve memory pools for guaranteed
592	deadlock-free allocations during extreme VM load. For example, the VM
593	subsystem makes use of the block layer to writeout dirty pages in order to be
594	able to free up memory space, a case which needs careful handling. The
595	allocation logic draws from the preallocated emergency reserve in situations
596	where it cannot allocate through normal means. If the pool is empty and it
597	can wait, then it would trigger action that would help free up memory or
598	replenish the pool (without deadlocking) and wait for availability in the pool.
599	If it is in IRQ context, and hence not in a position to do this, allocation
600	could fail if the pool is empty. In general mempool always first tries to
601	perform allocation without having to wait, even if it means digging into the
602	pool as long it is not less that 50% full.
604	On a free, memory is released to the pool or directly freed depending on
605	the current availability in the pool. The mempool interface lets the
606	subsystem specify the routines to be used for normal alloc and free. In the
607	case of bio, these routines make use of the standard slab allocator.
609	The caller of bio_alloc is expected to taken certain steps to avoid
610	deadlocks, e.g. avoid trying to allocate more memory from the pool while
611	already holding memory obtained from the pool.
612	[TBD: This is a potential issue, though a rare possibility
613	 in the bounce bio allocation that happens in the current code, since
614	 it ends up allocating a second bio from the same pool while
615	 holding the original bio ]
617	Memory allocated from the pool should be released back within a limited
618	amount of time (in the case of bio, that would be after the i/o is completed).
619	This ensures that if part of the pool has been used up, some work (in this
620	case i/o) must already be in progress and memory would be available when it
621	is over. If allocating from multiple pools in the same code path, the order
622	or hierarchy of allocation needs to be consistent, just the way one deals
623	with multiple locks.
625	The bio_alloc routine also needs to allocate the bio_vec_list (bvec_alloc())
626	for a non-clone bio. There are the 6 pools setup for different size biovecs,
627	so bio_alloc(gfp_mask, nr_iovecs) will allocate a vec_list of the
628	given size from these slabs.
630	The bio_get() routine may be used to hold an extra reference on a bio prior
631	to i/o submission, if the bio fields are likely to be accessed after the
632	i/o is issued (since the bio may otherwise get freed in case i/o completion
633	happens in the meantime).
635	The bio_clone() routine may be used to duplicate a bio, where the clone
636	shares the bio_vec_list with the original bio (i.e. both point to the
637	same bio_vec_list). This would typically be used for splitting i/o requests
638	in lvm or md.
640	3.2 Generic bio helper Routines
642	3.2.1 Traversing segments and completion units in a request
644	The macro rq_for_each_segment() should be used for traversing the bios
645	in the request list (drivers should avoid directly trying to do it
646	themselves). Using these helpers should also make it easier to cope
647	with block changes in the future.
649		struct req_iterator iter;
650		rq_for_each_segment(bio_vec, rq, iter)
651			/* bio_vec is now current segment */
653	I/O completion callbacks are per-bio rather than per-segment, so drivers
654	that traverse bio chains on completion need to keep that in mind. Drivers
655	which don't make a distinction between segments and completion units would
656	need to be reorganized to support multi-segment bios.
658	3.2.2 Setting up DMA scatterlists
660	The blk_rq_map_sg() helper routine would be used for setting up scatter
661	gather lists from a request, so a driver need not do it on its own.
663		nr_segments = blk_rq_map_sg(q, rq, scatterlist);
665	The helper routine provides a level of abstraction which makes it easier
666	to modify the internals of request to scatterlist conversion down the line
667	without breaking drivers. The blk_rq_map_sg routine takes care of several
668	things like collapsing physically contiguous segments (if QUEUE_FLAG_CLUSTER
669	is set) and correct segment accounting to avoid exceeding the limits which
670	the i/o hardware can handle, based on various queue properties.
672	- Prevents a clustered segment from crossing a 4GB mem boundary
673	- Avoids building segments that would exceed the number of physical
674	  memory segments that the driver can handle (phys_segments) and the
675	  number that the underlying hardware can handle at once, accounting for
676	  DMA remapping (hw_segments)  (i.e. IOMMU aware limits).
678	Routines which the low level driver can use to set up the segment limits:
680	blk_queue_max_hw_segments() : Sets an upper limit of the maximum number of
681	hw data segments in a request (i.e. the maximum number of address/length
682	pairs the host adapter can actually hand to the device at once)
684	blk_queue_max_phys_segments() : Sets an upper limit on the maximum number
685	of physical data segments in a request (i.e. the largest sized scatter list
686	a driver could handle)
688	3.2.3 I/O completion
690	The existing generic block layer helper routines end_request,
691	end_that_request_first and end_that_request_last can be used for i/o
692	completion (and setting things up so the rest of the i/o or the next
693	request can be kicked of) as before. With the introduction of multi-page
694	bio support, end_that_request_first requires an additional argument indicating
695	the number of sectors completed.
697	3.2.4 Implications for drivers that do not interpret bios (don't handle
698	 multiple segments)
700	Drivers that do not interpret bios e.g those which do not handle multiple
701	segments and do not support i/o into high memory addresses (require bounce
702	buffers) and expect only virtually mapped buffers, can access the rq->buffer
703	field. As before the driver should use current_nr_sectors to determine the
704	size of remaining data in the current segment (that is the maximum it can
705	transfer in one go unless it interprets segments), and rely on the block layer
706	end_request, or end_that_request_first/last to take care of all accounting
707	and transparent mapping of the next bio segment when a segment boundary
708	is crossed on completion of a transfer. (The end*request* functions should
709	be used if only if the request has come down from block/bio path, not for
710	direct access requests which only specify rq->buffer without a valid rq->bio)
712	3.2.5 Generic request command tagging
714 Tag helpers
716	Block now offers some simple generic functionality to help support command
717	queueing (typically known as tagged command queueing), ie manage more than
718	one outstanding command on a queue at any given time.
720		blk_queue_init_tags(struct request_queue *q, int depth)
722		Initialize internal command tagging structures for a maximum
723		depth of 'depth'.
725		blk_queue_free_tags((struct request_queue *q)
727		Teardown tag info associated with the queue. This will be done
728		automatically by block if blk_queue_cleanup() is called on a queue
729		that is using tagging.
731	The above are initialization and exit management, the main helpers during
732	normal operations are:
734		blk_queue_start_tag(struct request_queue *q, struct request *rq)
736		Start tagged operation for this request. A free tag number between
737		0 and 'depth' is assigned to the request (rq->tag holds this number),
738		and 'rq' is added to the internal tag management. If the maximum depth
739		for this queue is already achieved (or if the tag wasn't started for
740		some other reason), 1 is returned. Otherwise 0 is returned.
742		blk_queue_end_tag(struct request_queue *q, struct request *rq)
744		End tagged operation on this request. 'rq' is removed from the internal
745		book keeping structures.
747	To minimize struct request and queue overhead, the tag helpers utilize some
748	of the same request members that are used for normal request queue management.
749	This means that a request cannot both be an active tag and be on the queue
750	list at the same time. blk_queue_start_tag() will remove the request, but
751	the driver must remember to call blk_queue_end_tag() before signalling
752	completion of the request to the block layer. This means ending tag
753	operations before calling end_that_request_last()! For an example of a user
754	of these helpers, see the IDE tagged command queueing support.
756	Certain hardware conditions may dictate a need to invalidate the block tag
757	queue. For instance, on IDE any tagged request error needs to clear both
758	the hardware and software block queue and enable the driver to sanely restart
759	all the outstanding requests. There's a third helper to do that:
761		blk_queue_invalidate_tags(struct request_queue *q)
763		Clear the internal block tag queue and re-add all the pending requests
764		to the request queue. The driver will receive them again on the
765		next request_fn run, just like it did the first time it encountered
766		them.
768 Tag info
770	Some block functions exist to query current tag status or to go from a
771	tag number to the associated request. These are, in no particular order:
773		blk_queue_tagged(q)
775		Returns 1 if the queue 'q' is using tagging, 0 if not.
777		blk_queue_tag_request(q, tag)
779		Returns a pointer to the request associated with tag 'tag'.
781		blk_queue_tag_depth(q)
783		Return current queue depth.
785		blk_queue_tag_queue(q)
787		Returns 1 if the queue can accept a new queued command, 0 if we are
788		at the maximum depth already.
790		blk_queue_rq_tagged(rq)
792		Returns 1 if the request 'rq' is tagged.
794 Internal structure
796	Internally, block manages tags in the blk_queue_tag structure:
798		struct blk_queue_tag {
799			struct request **tag_index;	/* array or pointers to rq */
800			unsigned long *tag_map;		/* bitmap of free tags */
801			struct list_head busy_list;	/* fifo list of busy tags */
802			int busy;			/* queue depth */
803			int max_depth;			/* max queue depth */
804		};
806	Most of the above is simple and straight forward, however busy_list may need
807	a bit of explaining. Normally we don't care too much about request ordering,
808	but in the event of any barrier requests in the tag queue we need to ensure
809	that requests are restarted in the order they were queue. This may happen
810	if the driver needs to use blk_queue_invalidate_tags().
812	3.3 I/O Submission
814	The routine submit_bio() is used to submit a single io. Higher level i/o
815	routines make use of this:
817	(a) Buffered i/o:
818	The routine submit_bh() invokes submit_bio() on a bio corresponding to the
819	bh, allocating the bio if required. ll_rw_block() uses submit_bh() as before.
821	(b) Kiobuf i/o (for raw/direct i/o):
822	The ll_rw_kio() routine breaks up the kiobuf into page sized chunks and
823	maps the array to one or more multi-page bios, issuing submit_bio() to
824	perform the i/o on each of these.
826	The embedded bh array in the kiobuf structure has been removed and no
827	preallocation of bios is done for kiobufs. [The intent is to remove the
828	blocks array as well, but it's currently in there to kludge around direct i/o.]
829	Thus kiobuf allocation has switched back to using kmalloc rather than vmalloc.
831	Todo/Observation:
833	 A single kiobuf structure is assumed to correspond to a contiguous range
834	 of data, so brw_kiovec() invokes ll_rw_kio for each kiobuf in a kiovec.
835	 So right now it wouldn't work for direct i/o on non-contiguous blocks.
836	 This is to be resolved.  The eventual direction is to replace kiobuf
837	 by kvec's.
839	 Badari Pulavarty has a patch to implement direct i/o correctly using
840	 bio and kvec.
843	(c) Page i/o:
844	Todo/Under discussion:
846	 Andrew Morton's multi-page bio patches attempt to issue multi-page
847	 writeouts (and reads) from the page cache, by directly building up
848	 large bios for submission completely bypassing the usage of buffer
849	 heads. This work is still in progress.
851	 Christoph Hellwig had some code that uses bios for page-io (rather than
852	 bh). This isn't included in bio as yet. Christoph was also working on a
853	 design for representing virtual/real extents as an entity and modifying
854	 some of the address space ops interfaces to utilize this abstraction rather
855	 than buffer_heads. (This is somewhat along the lines of the SGI XFS pagebuf
856	 abstraction, but intended to be as lightweight as possible).
858	(d) Direct access i/o:
859	Direct access requests that do not contain bios would be submitted differently
860	as discussed earlier in section 1.3.
862	Aside:
864	  Kvec i/o:
866	  Ben LaHaise's aio code uses a slightly different structure instead
867	  of kiobufs, called a kvec_cb. This contains an array of <page, offset, len>
868	  tuples (very much like the networking code), together with a callback function
869	  and data pointer. This is embedded into a brw_cb structure when passed
870	  to brw_kvec_async().
872	  Now it should be possible to directly map these kvecs to a bio. Just as while
873	  cloning, in this case rather than PRE_BUILT bio_vecs, we set the bi_io_vec
874	  array pointer to point to the veclet array in kvecs.
876	  TBD: In order for this to work, some changes are needed in the way multi-page
877	  bios are handled today. The values of the tuples in such a vector passed in
878	  from higher level code should not be modified by the block layer in the course
879	  of its request processing, since that would make it hard for the higher layer
880	  to continue to use the vector descriptor (kvec) after i/o completes. Instead,
881	  all such transient state should either be maintained in the request structure,
882	  and passed on in some way to the endio completion routine.
885	4. The I/O scheduler
886	I/O scheduler, a.k.a. elevator, is implemented in two layers.  Generic dispatch
887	queue and specific I/O schedulers.  Unless stated otherwise, elevator is used
888	to refer to both parts and I/O scheduler to specific I/O schedulers.
890	Block layer implements generic dispatch queue in block/*.c.
891	The generic dispatch queue is responsible for requeueing, handling non-fs
892	requests and all other subtleties.
894	Specific I/O schedulers are responsible for ordering normal filesystem
895	requests.  They can also choose to delay certain requests to improve
896	throughput or whatever purpose.  As the plural form indicates, there are
897	multiple I/O schedulers.  They can be built as modules but at least one should
898	be built inside the kernel.  Each queue can choose different one and can also
899	change to another one dynamically.
901	A block layer call to the i/o scheduler follows the convention elv_xxx(). This
902	calls elevator_xxx_fn in the elevator switch (block/elevator.c). Oh, xxx
903	and xxx might not match exactly, but use your imagination. If an elevator
904	doesn't implement a function, the switch does nothing or some minimal house
905	keeping work.
907	4.1. I/O scheduler API
909	The functions an elevator may implement are: (* are mandatory)
910	elevator_merge_fn		called to query requests for merge with a bio
912	elevator_merge_req_fn		called when two requests get merged. the one
913					which gets merged into the other one will be
914					never seen by I/O scheduler again. IOW, after
915					being merged, the request is gone.
917	elevator_merged_fn		called when a request in the scheduler has been
918					involved in a merge. It is used in the deadline
919					scheduler for example, to reposition the request
920					if its sorting order has changed.
922	elevator_allow_merge_fn		called whenever the block layer determines
923					that a bio can be merged into an existing
924					request safely. The io scheduler may still
925					want to stop a merge at this point if it
926					results in some sort of conflict internally,
927					this hook allows it to do that. Note however
928					that two *requests* can still be merged at later
929					time. Currently the io scheduler has no way to
930					prevent that. It can only learn about the fact
931					from elevator_merge_req_fn callback.
933	elevator_dispatch_fn*		fills the dispatch queue with ready requests.
934					I/O schedulers are free to postpone requests by
935					not filling the dispatch queue unless @force
936					is non-zero.  Once dispatched, I/O schedulers
937					are not allowed to manipulate the requests -
938					they belong to generic dispatch queue.
940	elevator_add_req_fn*		called to add a new request into the scheduler
942	elevator_former_req_fn
943	elevator_latter_req_fn		These return the request before or after the
944					one specified in disk sort order. Used by the
945					block layer to find merge possibilities.
947	elevator_completed_req_fn	called when a request is completed.
949	elevator_may_queue_fn		returns true if the scheduler wants to allow the
950					current context to queue a new request even if
951					it is over the queue limit. This must be used
952					very carefully!!
954	elevator_set_req_fn
955	elevator_put_req_fn		Must be used to allocate and free any elevator
956					specific storage for a request.
958	elevator_activate_req_fn	Called when device driver first sees a request.
959					I/O schedulers can use this callback to
960					determine when actual execution of a request
961					starts.
962	elevator_deactivate_req_fn	Called when device driver decides to delay
963					a request by requeueing it.
965	elevator_init_fn*
966	elevator_exit_fn		Allocate and free any elevator specific storage
967					for a queue.
969	4.2 Request flows seen by I/O schedulers
970	All requests seen by I/O schedulers strictly follow one of the following three
971	flows.
973	 set_req_fn ->
975	 i.   add_req_fn -> (merged_fn ->)* -> dispatch_fn -> activate_req_fn ->
976	      (deactivate_req_fn -> activate_req_fn ->)* -> completed_req_fn
977	 ii.  add_req_fn -> (merged_fn ->)* -> merge_req_fn
978	 iii. [none]
980	 -> put_req_fn
982	4.3 I/O scheduler implementation
983	The generic i/o scheduler algorithm attempts to sort/merge/batch requests for
984	optimal disk scan and request servicing performance (based on generic
985	principles and device capabilities), optimized for:
986	i.   improved throughput
987	ii.  improved latency
988	iii. better utilization of h/w & CPU time
990	Characteristics:
992	i. Binary tree
993	AS and deadline i/o schedulers use red black binary trees for disk position
994	sorting and searching, and a fifo linked list for time-based searching. This
995	gives good scalability and good availability of information. Requests are
996	almost always dispatched in disk sort order, so a cache is kept of the next
997	request in sort order to prevent binary tree lookups.
999	This arrangement is not a generic block layer characteristic however, so
1000	elevators may implement queues as they please.
1002	ii. Merge hash
1003	AS and deadline use a hash table indexed by the last sector of a request. This
1004	enables merging code to quickly look up "back merge" candidates, even when
1005	multiple I/O streams are being performed at once on one disk.
1007	"Front merges", a new request being merged at the front of an existing request,
1008	are far less common than "back merges" due to the nature of most I/O patterns.
1009	Front merges are handled by the binary trees in AS and deadline schedulers.
1011	iii. Plugging the queue to batch requests in anticipation of opportunities for
1012	     merge/sort optimizations
1014	Plugging is an approach that the current i/o scheduling algorithm resorts to so
1015	that it collects up enough requests in the queue to be able to take
1016	advantage of the sorting/merging logic in the elevator. If the
1017	queue is empty when a request comes in, then it plugs the request queue
1018	(sort of like plugging the bath tub of a vessel to get fluid to build up)
1019	till it fills up with a few more requests, before starting to service
1020	the requests. This provides an opportunity to merge/sort the requests before
1021	passing them down to the device. There are various conditions when the queue is
1022	unplugged (to open up the flow again), either through a scheduled task or
1023	could be on demand. For example wait_on_buffer sets the unplugging going
1024	through sync_buffer() running blk_run_address_space(mapping). Or the caller
1025	can do it explicity through blk_unplug(bdev). So in the read case,
1026	the queue gets explicitly unplugged as part of waiting for completion on that
1027	buffer. For page driven IO, the address space ->sync_page() takes care of
1028	doing the blk_run_address_space().
1030	Aside:
1031	  This is kind of controversial territory, as it's not clear if plugging is
1032	  always the right thing to do. Devices typically have their own queues,
1033	  and allowing a big queue to build up in software, while letting the device be
1034	  idle for a while may not always make sense. The trick is to handle the fine
1035	  balance between when to plug and when to open up. Also now that we have
1036	  multi-page bios being queued in one shot, we may not need to wait to merge
1037	  a big request from the broken up pieces coming by.
1039	4.4 I/O contexts
1040	I/O contexts provide a dynamically allocated per process data area. They may
1041	be used in I/O schedulers, and in the block layer (could be used for IO statis,
1042	priorities for example). See *io_context in block/ll_rw_blk.c, and as-iosched.c
1043	for an example of usage in an i/o scheduler.
1046	5. Scalability related changes
1048	5.1 Granular Locking: io_request_lock replaced by a per-queue lock
1050	The global io_request_lock has been removed as of 2.5, to avoid
1051	the scalability bottleneck it was causing, and has been replaced by more
1052	granular locking. The request queue structure has a pointer to the
1053	lock to be used for that queue. As a result, locking can now be
1054	per-queue, with a provision for sharing a lock across queues if
1055	necessary (e.g the scsi layer sets the queue lock pointers to the
1056	corresponding adapter lock, which results in a per host locking
1057	granularity). The locking semantics are the same, i.e. locking is
1058	still imposed by the block layer, grabbing the lock before
1059	request_fn execution which it means that lots of older drivers
1060	should still be SMP safe. Drivers are free to drop the queue
1061	lock themselves, if required. Drivers that explicitly used the
1062	io_request_lock for serialization need to be modified accordingly.
1063	Usually it's as easy as adding a global lock:
1065		static DEFINE_SPINLOCK(my_driver_lock);
1067	and passing the address to that lock to blk_init_queue().
1069	5.2 64 bit sector numbers (sector_t prepares for 64 bit support)
1071	The sector number used in the bio structure has been changed to sector_t,
1072	which could be defined as 64 bit in preparation for 64 bit sector support.
1074	6. Other Changes/Implications
1076	6.1 Partition re-mapping handled by the generic block layer
1078	In 2.5 some of the gendisk/partition related code has been reorganized.
1079	Now the generic block layer performs partition-remapping early and thus
1080	provides drivers with a sector number relative to whole device, rather than
1081	having to take partition number into account in order to arrive at the true
1082	sector number. The routine blk_partition_remap() is invoked by
1083	generic_make_request even before invoking the queue specific make_request_fn,
1084	so the i/o scheduler also gets to operate on whole disk sector numbers. This
1085	should typically not require changes to block drivers, it just never gets
1086	to invoke its own partition sector offset calculations since all bios
1087	sent are offset from the beginning of the device.
1090	7. A Few Tips on Migration of older drivers
1092	Old-style drivers that just use CURRENT and ignores clustered requests,
1093	may not need much change.  The generic layer will automatically handle
1094	clustered requests, multi-page bios, etc for the driver.
1096	For a low performance driver or hardware that is PIO driven or just doesn't
1097	support scatter-gather changes should be minimal too.
1099	The following are some points to keep in mind when converting old drivers
1100	to bio.
1102	Drivers should use elv_next_request to pick up requests and are no longer
1103	supposed to handle looping directly over the request list.
1104	(struct request->queue has been removed)
1106	Now end_that_request_first takes an additional number_of_sectors argument.
1107	It used to handle always just the first buffer_head in a request, now
1108	it will loop and handle as many sectors (on a bio-segment granularity)
1109	as specified.
1111	Now bh->b_end_io is replaced by bio->bi_end_io, but most of the time the
1112	right thing to use is bio_endio(bio, uptodate) instead.
1114	If the driver is dropping the io_request_lock from its request_fn strategy,
1115	then it just needs to replace that with q->queue_lock instead.
1117	As described in Sec 1.1, drivers can set max sector size, max segment size
1118	etc per queue now. Drivers that used to define their own merge functions i
1119	to handle things like this can now just use the blk_queue_* functions at
1120	blk_init_queue time.
1122	Drivers no longer have to map a {partition, sector offset} into the
1123	correct absolute location anymore, this is done by the block layer, so
1124	where a driver received a request ala this before:
1126		rq->rq_dev = mk_kdev(3, 5);	/* /dev/hda5 */
1127		rq->sector = 0;			/* first sector on hda5 */
1129	  it will now see
1131		rq->rq_dev = mk_kdev(3, 0);	/* /dev/hda */
1132		rq->sector = 123128;		/* offset from start of disk */
1134	As mentioned, there is no virtual mapping of a bio. For DMA, this is
1135	not a problem as the driver probably never will need a virtual mapping.
1136	Instead it needs a bus mapping (dma_map_page for a single segment or
1137	use dma_map_sg for scatter gather) to be able to ship it to the driver. For
1138	PIO drivers (or drivers that need to revert to PIO transfer once in a
1139	while (IDE for example)), where the CPU is doing the actual data
1140	transfer a virtual mapping is needed. If the driver supports highmem I/O,
1141	(Sec 1.1, (ii) ) it needs to use __bio_kmap_atomic and bio_kmap_irq to
1142	temporarily map a bio into the virtual address space.
1145	8. Prior/Related/Impacted patches
1147	8.1. Earlier kiobuf patches (sct/axboe/chait/hch/mkp)
1148	- orig kiobuf & raw i/o patches (now in 2.4 tree)
1149	- direct kiobuf based i/o to devices (no intermediate bh's)
1150	- page i/o using kiobuf
1151	- kiobuf splitting for lvm (mkp)
1152	- elevator support for kiobuf request merging (axboe)
1153	8.2. Zero-copy networking (Dave Miller)
1154	8.3. SGI XFS - pagebuf patches - use of kiobufs
1155	8.4. Multi-page pioent patch for bio (Christoph Hellwig)
1156	8.5. Direct i/o implementation (Andrea Arcangeli) since 2.4.10-pre11
1157	8.6. Async i/o implementation patch (Ben LaHaise)
1158	8.7. EVMS layering design (IBM EVMS team)
1159	8.8. Larger page cache size patch (Ben LaHaise) and
1160	     Large page size (Daniel Phillips)
1161	    => larger contiguous physical memory buffers
1162	8.9. VM reservations patch (Ben LaHaise)
1163	8.10. Write clustering patches ? (Marcelo/Quintela/Riel ?)
1164	8.11. Block device in page cache patch (Andrea Archangeli) - now in 2.4.10+
1165	8.12. Multiple block-size transfers for faster raw i/o (Shailabh Nagar,
1166	      Badari)
1167	8.13  Priority based i/o scheduler - prepatches (Arjan van de Ven)
1168	8.14  IDE Taskfile i/o patch (Andre Hedrick)
1169	8.15  Multi-page writeout and readahead patches (Andrew Morton)
1170	8.16  Direct i/o patches for 2.5 using kvec and bio (Badari Pulavarthy)
1172	9. Other References:
1174	9.1 The Splice I/O Model - Larry McVoy (and subsequent discussions on lkml,
1175	and Linus' comments - Jan 2001)
1176	9.2 Discussions about kiobuf and bh design on lkml between sct, linus, alan
1177	et al - Feb-March 2001 (many of the initial thoughts that led to bio were
1178	brought up in this discussion thread)
1179	9.3 Discussions on mempool on lkml - Dec 2001.
Hide Line Numbers
About Kernel Documentation Linux Kernel Contact Linux Resources Linux Blog

Information is copyright its respective author. All material is available from the Linux Kernel Source distributed under a GPL License. This page is provided as a free service by mjmwired.net.