Blame view
Documentation/this_cpu_ops.txt
11.2 KB
79ab3b0d2 this_cpu_ops.txt:... |
1 |
=================== |
a1b2a555d percpu: add docum... |
2 |
this_cpu operations |
79ab3b0d2 this_cpu_ops.txt:... |
3 4 5 6 |
=================== :Author: Christoph Lameter, August 4th, 2014 :Author: Pranith Kumar, Aug 2nd, 2014 |
a1b2a555d percpu: add docum... |
7 8 |
this_cpu operations are a way of optimizing access to per cpu |
ac490f4dc Documentation: th... |
9 10 11 12 |
variables associated with the *currently* executing processor. This is done through the use of segment registers (or a dedicated register where the cpu permanently stored the beginning of the per cpu area for a specific processor). |
a1b2a555d percpu: add docum... |
13 |
|
ac490f4dc Documentation: th... |
14 15 |
this_cpu operations add a per cpu variable offset to the processor specific per cpu base and encode that operation in the instruction |
a1b2a555d percpu: add docum... |
16 |
operating on the per cpu variable. |
ac490f4dc Documentation: th... |
17 |
This means that there are no atomicity issues between the calculation of |
a1b2a555d percpu: add docum... |
18 |
the offset and the operation on the data. Therefore it is not |
ac490f4dc Documentation: th... |
19 |
necessary to disable preemption or interrupts to ensure that the |
a1b2a555d percpu: add docum... |
20 21 22 23 24 |
processor is not changed between the calculation of the address and the operation on the data. Read-modify-write operations are of particular interest. Frequently processors have special lower latency instructions that can operate |
ac490f4dc Documentation: th... |
25 26 27 |
without the typical synchronization overhead, but still provide some sort of relaxed atomicity guarantees. The x86, for example, can execute RMW (Read Modify Write) instructions like inc/dec/cmpxchg without the |
a1b2a555d percpu: add docum... |
28 29 30 31 32 33 34 |
lock prefix and the associated latency penalty. Access to the variable without the lock prefix is not synchronized but synchronization is not necessary since we are dealing with per cpu data specific to the currently executing processor. Only the current processor should be accessing that variable and therefore there are no concurrency issues with other processors in the system. |
ac490f4dc Documentation: th... |
35 36 37 38 39 40 41 42 43 |
Please note that accesses by remote processors to a per cpu area are exceptional situations and may impact performance and/or correctness (remote write operations) of local RMW operations via this_cpu_*. The main use of the this_cpu operations has been to optimize counter operations. The following this_cpu() operations with implied preemption protection are defined. These operations can be used without worrying about |
79ab3b0d2 this_cpu_ops.txt:... |
44 |
preemption and interrupts:: |
ac490f4dc Documentation: th... |
45 |
|
ac490f4dc Documentation: th... |
46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 |
this_cpu_read(pcp) this_cpu_write(pcp, val) this_cpu_add(pcp, val) this_cpu_and(pcp, val) this_cpu_or(pcp, val) this_cpu_add_return(pcp, val) this_cpu_xchg(pcp, nval) this_cpu_cmpxchg(pcp, oval, nval) this_cpu_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2) this_cpu_sub(pcp, val) this_cpu_inc(pcp) this_cpu_dec(pcp) this_cpu_sub_return(pcp, val) this_cpu_inc_return(pcp) this_cpu_dec_return(pcp) Inner working of this_cpu operations ------------------------------------ |
a1b2a555d percpu: add docum... |
65 66 67 68 69 |
On x86 the fs: or the gs: segment registers contain the base of the per cpu area. It is then possible to simply use the segment override to relocate a per cpu relative address to the proper per cpu area for the processor. So the relocation to the per cpu base is encoded in the instruction via a segment register prefix. |
79ab3b0d2 this_cpu_ops.txt:... |
70 |
For example:: |
a1b2a555d percpu: add docum... |
71 72 73 74 75 |
DEFINE_PER_CPU(int, x); int z; z = this_cpu_read(x); |
79ab3b0d2 this_cpu_ops.txt:... |
76 |
results in a single instruction:: |
a1b2a555d percpu: add docum... |
77 78 79 80 |
mov ax, gs:[x] instead of a sequence of calculation of the address and then a fetch |
ac490f4dc Documentation: th... |
81 |
from that address which occurs with the per cpu operations. Before |
a1b2a555d percpu: add docum... |
82 83 84 |
this_cpu_ops such sequence also required preempt disable/enable to prevent the kernel from moving the thread to a different processor while the calculation is performed. |
79ab3b0d2 this_cpu_ops.txt:... |
85 |
Consider the following this_cpu operation:: |
a1b2a555d percpu: add docum... |
86 87 |
this_cpu_inc(x) |
79ab3b0d2 this_cpu_ops.txt:... |
88 |
The above results in the following single instruction (no lock prefix!):: |
a1b2a555d percpu: add docum... |
89 90 91 92 |
inc gs:[x] instead of the following operations required if there is no segment |
79ab3b0d2 this_cpu_ops.txt:... |
93 |
register:: |
a1b2a555d percpu: add docum... |
94 95 96 97 98 99 100 101 |
int *y; int cpu; cpu = get_cpu(); y = per_cpu_ptr(&x, cpu); (*y)++; put_cpu(); |
ac490f4dc Documentation: th... |
102 |
Note that these operations can only be used on per cpu data that is |
a1b2a555d percpu: add docum... |
103 104 |
reserved for a specific processor. Without disabling preemption in the surrounding code this_cpu_inc() will only guarantee that one of the |
ac490f4dc Documentation: th... |
105 |
per cpu counters is correctly incremented. However, there is no |
a1b2a555d percpu: add docum... |
106 107 108 109 110 111 112 113 114 |
guarantee that the OS will not move the process directly before or after the this_cpu instruction is executed. In general this means that the value of the individual counters for each processor are meaningless. The sum of all the per cpu counters is the only value that is of interest. Per cpu variables are used for performance reasons. Bouncing cache lines can be avoided if multiple processors concurrently go through the same code paths. Since each processor has its own per cpu |
ac490f4dc Documentation: th... |
115 |
variables no concurrent cache line updates take place. The price that |
a1b2a555d percpu: add docum... |
116 |
has to be paid for this optimization is the need to add up the per cpu |
ac490f4dc Documentation: th... |
117 |
counters when the value of a counter is needed. |
a1b2a555d percpu: add docum... |
118 |
|
79ab3b0d2 this_cpu_ops.txt:... |
119 120 121 122 |
Special operations ------------------ :: |
a1b2a555d percpu: add docum... |
123 124 125 126 127 128 129 |
y = this_cpu_ptr(&x) Takes the offset of a per cpu variable (&x !) and returns the address of the per cpu variable that belongs to the currently executing processor. this_cpu_ptr avoids multiple steps that the common get_cpu/put_cpu sequence requires. No processor number is |
ac490f4dc Documentation: th... |
130 131 |
available. Instead, the offset of the local per cpu area is simply added to the per cpu offset. |
a1b2a555d percpu: add docum... |
132 |
|
ac490f4dc Documentation: th... |
133 134 135 136 137 |
Note that this operation is usually used in a code segment when preemption has been disabled. The pointer is then used to access local per cpu data in a critical section. When preemption is re-enabled this pointer is usually no longer useful since it may no longer point to per cpu data of the current processor. |
a1b2a555d percpu: add docum... |
138 139 140 141 |
Per cpu variables and offsets ----------------------------- |
ac490f4dc Documentation: th... |
142 |
Per cpu variables have *offsets* to the beginning of the per cpu |
a1b2a555d percpu: add docum... |
143 144 |
area. They do not have addresses although they look like that in the code. Offsets cannot be directly dereferenced. The offset must be |
ac490f4dc Documentation: th... |
145 |
added to a base pointer of a per cpu area of a processor in order to |
a1b2a555d percpu: add docum... |
146 147 148 149 150 |
form a valid address. Therefore the use of x or &x outside of the context of per cpu operations is invalid and will generally be treated like a NULL pointer dereference. |
79ab3b0d2 this_cpu_ops.txt:... |
151 |
:: |
ac490f4dc Documentation: th... |
152 |
DEFINE_PER_CPU(int, x); |
a1b2a555d percpu: add docum... |
153 |
|
ac490f4dc Documentation: th... |
154 155 |
In the context of per cpu operations the above implies that x is a per cpu variable. Most this_cpu operations take a cpu variable. |
a1b2a555d percpu: add docum... |
156 |
|
79ab3b0d2 this_cpu_ops.txt:... |
157 |
:: |
ac490f4dc Documentation: th... |
158 |
int __percpu *p = &x; |
a1b2a555d percpu: add docum... |
159 |
|
ac490f4dc Documentation: th... |
160 161 162 |
&x and hence p is the *offset* of a per cpu variable. this_cpu_ptr() takes the offset of a per cpu variable which makes this look a bit strange. |
a1b2a555d percpu: add docum... |
163 164 165 166 |
Operations on a field of a per cpu structure -------------------------------------------- |
79ab3b0d2 this_cpu_ops.txt:... |
167 |
Let's say we have a percpu structure:: |
a1b2a555d percpu: add docum... |
168 169 170 171 172 173 |
struct s { int n,m; }; DEFINE_PER_CPU(struct s, p); |
79ab3b0d2 this_cpu_ops.txt:... |
174 |
Operations on these fields are straightforward:: |
a1b2a555d percpu: add docum... |
175 176 177 178 |
this_cpu_inc(p.m) z = this_cpu_cmpxchg(p.m, 0, 1); |
79ab3b0d2 this_cpu_ops.txt:... |
179 |
If we have an offset to struct s:: |
a1b2a555d percpu: add docum... |
180 181 |
struct s __percpu *ps = &p; |
ac490f4dc Documentation: th... |
182 |
this_cpu_dec(ps->m); |
a1b2a555d percpu: add docum... |
183 184 185 186 187 |
z = this_cpu_inc_return(ps->n); The calculation of the pointer may require the use of this_cpu_ptr() |
79ab3b0d2 this_cpu_ops.txt:... |
188 |
if we do not make use of this_cpu ops later to manipulate fields:: |
a1b2a555d percpu: add docum... |
189 190 191 192 193 194 195 196 197 198 199 |
struct s *pp; pp = this_cpu_ptr(&p); pp->m--; z = pp->n++; Variants of this_cpu ops |
79ab3b0d2 this_cpu_ops.txt:... |
200 |
------------------------ |
a1b2a555d percpu: add docum... |
201 |
|
ac490f4dc Documentation: th... |
202 |
this_cpu ops are interrupt safe. Some architectures do not support |
a1b2a555d percpu: add docum... |
203 204 |
these per cpu local operations. In that case the operation must be replaced by code that disables interrupts, then does the operations |
ac490f4dc Documentation: th... |
205 |
that are guaranteed to be atomic and then re-enable interrupts. Doing |
a1b2a555d percpu: add docum... |
206 207 |
so is expensive. If there are other reasons why the scheduler cannot change the processor we are executing on then there is no reason to |
ac490f4dc Documentation: th... |
208 209 210 211 212 213 214 215 |
disable interrupts. For that purpose the following __this_cpu operations are provided. These operations have no guarantee against concurrent interrupts or preemption. If a per cpu variable is not used in an interrupt context and the scheduler cannot preempt, then they are safe. If any interrupts still occur while an operation is in progress and if the interrupt too modifies the variable, then RMW actions can not be guaranteed to be |
79ab3b0d2 this_cpu_ops.txt:... |
216 |
safe:: |
ac490f4dc Documentation: th... |
217 |
|
ac490f4dc Documentation: th... |
218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 |
__this_cpu_read(pcp) __this_cpu_write(pcp, val) __this_cpu_add(pcp, val) __this_cpu_and(pcp, val) __this_cpu_or(pcp, val) __this_cpu_add_return(pcp, val) __this_cpu_xchg(pcp, nval) __this_cpu_cmpxchg(pcp, oval, nval) __this_cpu_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2) __this_cpu_sub(pcp, val) __this_cpu_inc(pcp) __this_cpu_dec(pcp) __this_cpu_sub_return(pcp, val) __this_cpu_inc_return(pcp) __this_cpu_dec_return(pcp) Will increment x and will not fall-back to code that disables |
a1b2a555d percpu: add docum... |
236 237 238 |
interrupts on platforms that cannot accomplish atomicity through address relocation and a Read-Modify-Write operation in the same instruction. |
a1b2a555d percpu: add docum... |
239 240 241 242 |
&this_cpu_ptr(pp)->n vs this_cpu_ptr(&pp->n) -------------------------------------------- The first operation takes the offset and forms an address and then |
ac490f4dc Documentation: th... |
243 244 |
adds the offset of the n field. This may result in two add instructions emitted by the compiler. |
a1b2a555d percpu: add docum... |
245 246 247 248 249 |
The second one first adds the two offsets and then does the relocation. IMHO the second form looks cleaner and has an easier time with (). The second form also is consistent with the way this_cpu_read() and friends are used. |
ac490f4dc Documentation: th... |
250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 |
Remote access to per cpu data ------------------------------ Per cpu data structures are designed to be used by one cpu exclusively. If you use the variables as intended, this_cpu_ops() are guaranteed to be "atomic" as no other CPU has access to these data structures. There are special cases where you might need to access per cpu data structures remotely. It is usually safe to do a remote read access and that is frequently done to summarize counters. Remote write access something which could be problematic because this_cpu ops do not have lock semantics. A remote write may interfere with a this_cpu RMW operation. Remote write accesses to percpu data structures are highly discouraged unless absolutely necessary. Please consider using an IPI to wake up the remote CPU and perform the update to its per cpu area. To access per-cpu data structure remotely, typically the per_cpu_ptr() |
79ab3b0d2 this_cpu_ops.txt:... |
269 |
function is used:: |
ac490f4dc Documentation: th... |
270 271 272 273 274 275 276 277 |
DEFINE_PER_CPU(struct data, datap); struct data *p = per_cpu_ptr(&datap, cpu); This makes it explicit that we are getting ready to access a percpu area remotely. |
79ab3b0d2 this_cpu_ops.txt:... |
278 |
You can also do the following to convert the datap offset to an address:: |
ac490f4dc Documentation: th... |
279 280 281 282 283 284 285 286 287 288 289 290 291 292 |
struct data *p = this_cpu_ptr(&datap); but, passing of pointers calculated via this_cpu_ptr to other cpus is unusual and should be avoided. Remote access are typically only for reading the status of another cpus per cpu data. Write accesses can cause unique problems due to the relaxed synchronization requirements for this_cpu operations. One example that illustrates some concerns with write operations is the following scenario that occurs because two per cpu variables share a cache-line but the relaxed synchronization is applied to only one process updating the cache-line. |
79ab3b0d2 this_cpu_ops.txt:... |
293 |
Consider the following example:: |
ac490f4dc Documentation: th... |
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 |
struct test { atomic_t a; int b; }; DEFINE_PER_CPU(struct test, onecacheline); There is some concern about what would happen if the field 'a' is updated remotely from one processor and the local processor would use this_cpu ops to update field b. Care should be taken that such simultaneous accesses to data within the same cache line are avoided. Also costly synchronization may be necessary. IPIs are generally recommended in such scenarios instead of a remote write to the per cpu area of another processor. Even in cases where the remote writes are rare, please bear in mind that a remote write will evict the cache line from the processor that most likely will access it. If the processor wakes up and finds a missing local cache line of a per cpu area, its performance and hence the wake up times will be affected. |