About Kernel Documentation Linux Kernel Contact Linux Resources Linux Blog

Documentation / this_cpu_ops.txt




Custom Search

Based on kernel version 4.13.3. Page generated on 2017-09-23 13:56 EST.

1	===================
2	this_cpu operations
3	===================
4	
5	:Author: Christoph Lameter, August 4th, 2014
6	:Author: Pranith Kumar, Aug 2nd, 2014
7	
8	this_cpu operations are a way of optimizing access to per cpu
9	variables associated with the *currently* executing processor. This is
10	done through the use of segment registers (or a dedicated register where
11	the cpu permanently stored the beginning of the per cpu	area for a
12	specific processor).
13	
14	this_cpu operations add a per cpu variable offset to the processor
15	specific per cpu base and encode that operation in the instruction
16	operating on the per cpu variable.
17	
18	This means that there are no atomicity issues between the calculation of
19	the offset and the operation on the data. Therefore it is not
20	necessary to disable preemption or interrupts to ensure that the
21	processor is not changed between the calculation of the address and
22	the operation on the data.
23	
24	Read-modify-write operations are of particular interest. Frequently
25	processors have special lower latency instructions that can operate
26	without the typical synchronization overhead, but still provide some
27	sort of relaxed atomicity guarantees. The x86, for example, can execute
28	RMW (Read Modify Write) instructions like inc/dec/cmpxchg without the
29	lock prefix and the associated latency penalty.
30	
31	Access to the variable without the lock prefix is not synchronized but
32	synchronization is not necessary since we are dealing with per cpu
33	data specific to the currently executing processor. Only the current
34	processor should be accessing that variable and therefore there are no
35	concurrency issues with other processors in the system.
36	
37	Please note that accesses by remote processors to a per cpu area are
38	exceptional situations and may impact performance and/or correctness
39	(remote write operations) of local RMW operations via this_cpu_*.
40	
41	The main use of the this_cpu operations has been to optimize counter
42	operations.
43	
44	The following this_cpu() operations with implied preemption protection
45	are defined. These operations can be used without worrying about
46	preemption and interrupts::
47	
48		this_cpu_read(pcp)
49		this_cpu_write(pcp, val)
50		this_cpu_add(pcp, val)
51		this_cpu_and(pcp, val)
52		this_cpu_or(pcp, val)
53		this_cpu_add_return(pcp, val)
54		this_cpu_xchg(pcp, nval)
55		this_cpu_cmpxchg(pcp, oval, nval)
56		this_cpu_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)
57		this_cpu_sub(pcp, val)
58		this_cpu_inc(pcp)
59		this_cpu_dec(pcp)
60		this_cpu_sub_return(pcp, val)
61		this_cpu_inc_return(pcp)
62		this_cpu_dec_return(pcp)
63	
64	
65	Inner working of this_cpu operations
66	------------------------------------
67	
68	On x86 the fs: or the gs: segment registers contain the base of the
69	per cpu area. It is then possible to simply use the segment override
70	to relocate a per cpu relative address to the proper per cpu area for
71	the processor. So the relocation to the per cpu base is encoded in the
72	instruction via a segment register prefix.
73	
74	For example::
75	
76		DEFINE_PER_CPU(int, x);
77		int z;
78	
79		z = this_cpu_read(x);
80	
81	results in a single instruction::
82	
83		mov ax, gs:[x]
84	
85	instead of a sequence of calculation of the address and then a fetch
86	from that address which occurs with the per cpu operations. Before
87	this_cpu_ops such sequence also required preempt disable/enable to
88	prevent the kernel from moving the thread to a different processor
89	while the calculation is performed.
90	
91	Consider the following this_cpu operation::
92	
93		this_cpu_inc(x)
94	
95	The above results in the following single instruction (no lock prefix!)::
96	
97		inc gs:[x]
98	
99	instead of the following operations required if there is no segment
100	register::
101	
102		int *y;
103		int cpu;
104	
105		cpu = get_cpu();
106		y = per_cpu_ptr(&x, cpu);
107		(*y)++;
108		put_cpu();
109	
110	Note that these operations can only be used on per cpu data that is
111	reserved for a specific processor. Without disabling preemption in the
112	surrounding code this_cpu_inc() will only guarantee that one of the
113	per cpu counters is correctly incremented. However, there is no
114	guarantee that the OS will not move the process directly before or
115	after the this_cpu instruction is executed. In general this means that
116	the value of the individual counters for each processor are
117	meaningless. The sum of all the per cpu counters is the only value
118	that is of interest.
119	
120	Per cpu variables are used for performance reasons. Bouncing cache
121	lines can be avoided if multiple processors concurrently go through
122	the same code paths.  Since each processor has its own per cpu
123	variables no concurrent cache line updates take place. The price that
124	has to be paid for this optimization is the need to add up the per cpu
125	counters when the value of a counter is needed.
126	
127	
128	Special operations
129	------------------
130	
131	::
132	
133		y = this_cpu_ptr(&x)
134	
135	Takes the offset of a per cpu variable (&x !) and returns the address
136	of the per cpu variable that belongs to the currently executing
137	processor.  this_cpu_ptr avoids multiple steps that the common
138	get_cpu/put_cpu sequence requires. No processor number is
139	available. Instead, the offset of the local per cpu area is simply
140	added to the per cpu offset.
141	
142	Note that this operation is usually used in a code segment when
143	preemption has been disabled. The pointer is then used to
144	access local per cpu data in a critical section. When preemption
145	is re-enabled this pointer is usually no longer useful since it may
146	no longer point to per cpu data of the current processor.
147	
148	
149	Per cpu variables and offsets
150	-----------------------------
151	
152	Per cpu variables have *offsets* to the beginning of the per cpu
153	area. They do not have addresses although they look like that in the
154	code. Offsets cannot be directly dereferenced. The offset must be
155	added to a base pointer of a per cpu area of a processor in order to
156	form a valid address.
157	
158	Therefore the use of x or &x outside of the context of per cpu
159	operations is invalid and will generally be treated like a NULL
160	pointer dereference.
161	
162	::
163	
164		DEFINE_PER_CPU(int, x);
165	
166	In the context of per cpu operations the above implies that x is a per
167	cpu variable. Most this_cpu operations take a cpu variable.
168	
169	::
170	
171		int __percpu *p = &x;
172	
173	&x and hence p is the *offset* of a per cpu variable. this_cpu_ptr()
174	takes the offset of a per cpu variable which makes this look a bit
175	strange.
176	
177	
178	Operations on a field of a per cpu structure
179	--------------------------------------------
180	
181	Let's say we have a percpu structure::
182	
183		struct s {
184			int n,m;
185		};
186	
187		DEFINE_PER_CPU(struct s, p);
188	
189	
190	Operations on these fields are straightforward::
191	
192		this_cpu_inc(p.m)
193	
194		z = this_cpu_cmpxchg(p.m, 0, 1);
195	
196	
197	If we have an offset to struct s::
198	
199		struct s __percpu *ps = &p;
200	
201		this_cpu_dec(ps->m);
202	
203		z = this_cpu_inc_return(ps->n);
204	
205	
206	The calculation of the pointer may require the use of this_cpu_ptr()
207	if we do not make use of this_cpu ops later to manipulate fields::
208	
209		struct s *pp;
210	
211		pp = this_cpu_ptr(&p);
212	
213		pp->m--;
214	
215		z = pp->n++;
216	
217	
218	Variants of this_cpu ops
219	------------------------
220	
221	this_cpu ops are interrupt safe. Some architectures do not support
222	these per cpu local operations. In that case the operation must be
223	replaced by code that disables interrupts, then does the operations
224	that are guaranteed to be atomic and then re-enable interrupts. Doing
225	so is expensive. If there are other reasons why the scheduler cannot
226	change the processor we are executing on then there is no reason to
227	disable interrupts. For that purpose the following __this_cpu operations
228	are provided.
229	
230	These operations have no guarantee against concurrent interrupts or
231	preemption. If a per cpu variable is not used in an interrupt context
232	and the scheduler cannot preempt, then they are safe. If any interrupts
233	still occur while an operation is in progress and if the interrupt too
234	modifies the variable, then RMW actions can not be guaranteed to be
235	safe::
236	
237		__this_cpu_read(pcp)
238		__this_cpu_write(pcp, val)
239		__this_cpu_add(pcp, val)
240		__this_cpu_and(pcp, val)
241		__this_cpu_or(pcp, val)
242		__this_cpu_add_return(pcp, val)
243		__this_cpu_xchg(pcp, nval)
244		__this_cpu_cmpxchg(pcp, oval, nval)
245		__this_cpu_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)
246		__this_cpu_sub(pcp, val)
247		__this_cpu_inc(pcp)
248		__this_cpu_dec(pcp)
249		__this_cpu_sub_return(pcp, val)
250		__this_cpu_inc_return(pcp)
251		__this_cpu_dec_return(pcp)
252	
253	
254	Will increment x and will not fall-back to code that disables
255	interrupts on platforms that cannot accomplish atomicity through
256	address relocation and a Read-Modify-Write operation in the same
257	instruction.
258	
259	
260	&this_cpu_ptr(pp)->n vs this_cpu_ptr(&pp->n)
261	--------------------------------------------
262	
263	The first operation takes the offset and forms an address and then
264	adds the offset of the n field. This may result in two add
265	instructions emitted by the compiler.
266	
267	The second one first adds the two offsets and then does the
268	relocation.  IMHO the second form looks cleaner and has an easier time
269	with (). The second form also is consistent with the way
270	this_cpu_read() and friends are used.
271	
272	
273	Remote access to per cpu data
274	------------------------------
275	
276	Per cpu data structures are designed to be used by one cpu exclusively.
277	If you use the variables as intended, this_cpu_ops() are guaranteed to
278	be "atomic" as no other CPU has access to these data structures.
279	
280	There are special cases where you might need to access per cpu data
281	structures remotely. It is usually safe to do a remote read access
282	and that is frequently done to summarize counters. Remote write access
283	something which could be problematic because this_cpu ops do not
284	have lock semantics. A remote write may interfere with a this_cpu
285	RMW operation.
286	
287	Remote write accesses to percpu data structures are highly discouraged
288	unless absolutely necessary. Please consider using an IPI to wake up
289	the remote CPU and perform the update to its per cpu area.
290	
291	To access per-cpu data structure remotely, typically the per_cpu_ptr()
292	function is used::
293	
294	
295		DEFINE_PER_CPU(struct data, datap);
296	
297		struct data *p = per_cpu_ptr(&datap, cpu);
298	
299	This makes it explicit that we are getting ready to access a percpu
300	area remotely.
301	
302	You can also do the following to convert the datap offset to an address::
303	
304		struct data *p = this_cpu_ptr(&datap);
305	
306	but, passing of pointers calculated via this_cpu_ptr to other cpus is
307	unusual and should be avoided.
308	
309	Remote access are typically only for reading the status of another cpus
310	per cpu data. Write accesses can cause unique problems due to the
311	relaxed synchronization requirements for this_cpu operations.
312	
313	One example that illustrates some concerns with write operations is
314	the following scenario that occurs because two per cpu variables
315	share a cache-line but the relaxed synchronization is applied to
316	only one process updating the cache-line.
317	
318	Consider the following example::
319	
320	
321		struct test {
322			atomic_t a;
323			int b;
324		};
325	
326		DEFINE_PER_CPU(struct test, onecacheline);
327	
328	There is some concern about what would happen if the field 'a' is updated
329	remotely from one processor and the local processor would use this_cpu ops
330	to update field b. Care should be taken that such simultaneous accesses to
331	data within the same cache line are avoided. Also costly synchronization
332	may be necessary. IPIs are generally recommended in such scenarios instead
333	of a remote write to the per cpu area of another processor.
334	
335	Even in cases where the remote writes are rare, please bear in
336	mind that a remote write will evict the cache line from the processor
337	that most likely will access it. If the processor wakes up and finds a
338	missing local cache line of a per cpu area, its performance and hence
339	the wake up times will be affected.
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.