2 * Oct 15, 2000 Matt Domsch <Matt_Domsch@dell.com>
3 * Nicer crc32 functions/docs submitted by linux@horizon.com. Thanks!
4 * Code was from the public domain, copyright abandoned. Code was
5 * subsequently included in the kernel, thus was re-licensed under the
8 * Oct 12, 2000 Matt Domsch <Matt_Domsch@dell.com>
9 * Same crc32 function was used in 5 other places in the kernel.
10 * I made one version, and deleted the others.
11 * There are various incantations of crc32(). Some use a seed of 0 or ~0.
12 * Some xor at the end with ~0. The generic crc32() function takes
13 * seed as an argument, and doesn't xor at the end. Then individual
14 * users can do whatever they need.
15 * drivers/net/smc9194.c uses seed ~0, doesn't xor with ~0.
16 * fs/jffs2 uses seed 0, doesn't xor with ~0.
17 * fs/partitions/efi.c uses seed ~0, xor's with ~0.
19 * This source code is licensed under the GNU General Public License,
20 * Version 2. See the file COPYING for more details.
23 #include <linux/crc32.h>
24 #include <linux/kernel.h>
25 #include <linux/module.h>
26 #include <linux/compiler.h>
27 #include <linux/types.h>
28 #include <linux/init.h>
29 #include <linux/atomic.h>
30 #include "crc32defs.h"
32 # define tole(x) __constant_cpu_to_le32(x)
38 # define tobe(x) __constant_cpu_to_be32(x)
42 #include "crc32table.h"
44 MODULE_AUTHOR("Matt Domsch <Matt_Domsch@dell.com>");
45 MODULE_DESCRIPTION("Ethernet CRC32 calculations");
46 MODULE_LICENSE("GPL");
48 #if CRC_LE_BITS == 8 || CRC_BE_BITS == 8
51 crc32_body(u32 crc, unsigned char const *buf, size_t len, const u32 (*tab)[256])
53 # ifdef __LITTLE_ENDIAN
54 # define DO_CRC(x) crc = t0[(crc ^ (x)) & 255] ^ (crc >> 8)
55 # define DO_CRC4 crc = t3[(crc) & 255] ^ \
56 t2[(crc >> 8) & 255] ^ \
57 t1[(crc >> 16) & 255] ^ \
60 # define DO_CRC(x) crc = t0[((crc >> 24) ^ (x)) & 255] ^ (crc << 8)
61 # define DO_CRC4 crc = t0[(crc) & 255] ^ \
62 t1[(crc >> 8) & 255] ^ \
63 t2[(crc >> 16) & 255] ^ \
68 const u32 *t0=tab[0], *t1=tab[1], *t2=tab[2], *t3=tab[3];
71 if (unlikely((long)buf & 3 && len)) {
74 } while ((--len) && ((long)buf)&3);
77 /* load data 32 bits wide, xor data 32 bits wide. */
80 for (--b; len; --len) {
81 crc ^= *++b; /* use pre increment for speed */
85 /* And the last few bytes */
87 u8 *p = (u8 *)(b + 1) - 1;
89 DO_CRC(*++p); /* use pre increment for speed */
98 * crc32_le() - Calculate bitwise little-endian Ethernet AUTODIN II CRC32
99 * @crc: seed value for computation. ~0 for Ethernet, sometimes 0 for
100 * other uses, or the previous crc32 value if computing incrementally.
101 * @p: pointer to buffer over which CRC is run
102 * @len: length of buffer @p
104 u32 __pure crc32_le(u32 crc, unsigned char const *p, size_t len);
108 * In fact, the table-based code will work in this case, but it can be
109 * simplified by inlining the table in ?: form.
112 u32 __pure crc32_le(u32 crc, unsigned char const *p, size_t len)
117 for (i = 0; i < 8; i++)
118 crc = (crc >> 1) ^ ((crc & 1) ? CRCPOLY_LE : 0);
122 #else /* Table-based approach */
124 u32 __pure crc32_le(u32 crc, unsigned char const *p, size_t len)
126 # if CRC_LE_BITS == 8
127 const u32 (*tab)[] = crc32table_le;
129 crc = __cpu_to_le32(crc);
130 crc = crc32_body(crc, p, len, tab);
131 return __le32_to_cpu(crc);
132 # elif CRC_LE_BITS == 4
135 crc = (crc >> 4) ^ crc32table_le[crc & 15];
136 crc = (crc >> 4) ^ crc32table_le[crc & 15];
139 # elif CRC_LE_BITS == 2
142 crc = (crc >> 2) ^ crc32table_le[crc & 3];
143 crc = (crc >> 2) ^ crc32table_le[crc & 3];
144 crc = (crc >> 2) ^ crc32table_le[crc & 3];
145 crc = (crc >> 2) ^ crc32table_le[crc & 3];
153 * crc32_be() - Calculate bitwise big-endian Ethernet AUTODIN II CRC32
154 * @crc: seed value for computation. ~0 for Ethernet, sometimes 0 for
155 * other uses, or the previous crc32 value if computing incrementally.
156 * @p: pointer to buffer over which CRC is run
157 * @len: length of buffer @p
159 u32 __pure crc32_be(u32 crc, unsigned char const *p, size_t len);
163 * In fact, the table-based code will work in this case, but it can be
164 * simplified by inlining the table in ?: form.
167 u32 __pure crc32_be(u32 crc, unsigned char const *p, size_t len)
172 for (i = 0; i < 8; i++)
174 (crc << 1) ^ ((crc & 0x80000000) ? CRCPOLY_BE :
180 #else /* Table-based approach */
181 u32 __pure crc32_be(u32 crc, unsigned char const *p, size_t len)
183 # if CRC_BE_BITS == 8
184 const u32 (*tab)[] = crc32table_be;
186 crc = __cpu_to_be32(crc);
187 crc = crc32_body(crc, p, len, tab);
188 return __be32_to_cpu(crc);
189 # elif CRC_BE_BITS == 4
192 crc = (crc << 4) ^ crc32table_be[crc >> 28];
193 crc = (crc << 4) ^ crc32table_be[crc >> 28];
196 # elif CRC_BE_BITS == 2
199 crc = (crc << 2) ^ crc32table_be[crc >> 30];
200 crc = (crc << 2) ^ crc32table_be[crc >> 30];
201 crc = (crc << 2) ^ crc32table_be[crc >> 30];
202 crc = (crc << 2) ^ crc32table_be[crc >> 30];
209 EXPORT_SYMBOL(crc32_le);
210 EXPORT_SYMBOL(crc32_be);
213 * A brief CRC tutorial.
215 * A CRC is a long-division remainder. You add the CRC to the message,
216 * and the whole thing (message+CRC) is a multiple of the given
217 * CRC polynomial. To check the CRC, you can either check that the
218 * CRC matches the recomputed value, *or* you can check that the
219 * remainder computed on the message+CRC is 0. This latter approach
220 * is used by a lot of hardware implementations, and is why so many
221 * protocols put the end-of-frame flag after the CRC.
223 * It's actually the same long division you learned in school, except that
224 * - We're working in binary, so the digits are only 0 and 1, and
225 * - When dividing polynomials, there are no carries. Rather than add and
226 * subtract, we just xor. Thus, we tend to get a bit sloppy about
227 * the difference between adding and subtracting.
229 * A 32-bit CRC polynomial is actually 33 bits long. But since it's
230 * 33 bits long, bit 32 is always going to be set, so usually the CRC
231 * is written in hex with the most significant bit omitted. (If you're
232 * familiar with the IEEE 754 floating-point format, it's the same idea.)
234 * Note that a CRC is computed over a string of *bits*, so you have
235 * to decide on the endianness of the bits within each byte. To get
236 * the best error-detecting properties, this should correspond to the
237 * order they're actually sent. For example, standard RS-232 serial is
238 * little-endian; the most significant bit (sometimes used for parity)
239 * is sent last. And when appending a CRC word to a message, you should
240 * do it in the right order, matching the endianness.
242 * Just like with ordinary division, the remainder is always smaller than
243 * the divisor (the CRC polynomial) you're dividing by. Each step of the
244 * division, you take one more digit (bit) of the dividend and append it
245 * to the current remainder. Then you figure out the appropriate multiple
246 * of the divisor to subtract to being the remainder back into range.
247 * In binary, it's easy - it has to be either 0 or 1, and to make the
248 * XOR cancel, it's just a copy of bit 32 of the remainder.
250 * When computing a CRC, we don't care about the quotient, so we can
251 * throw the quotient bit away, but subtract the appropriate multiple of
252 * the polynomial from the remainder and we're back to where we started,
253 * ready to process the next bit.
255 * A big-endian CRC written this way would be coded like:
256 * for (i = 0; i < input_bits; i++) {
257 * multiple = remainder & 0x80000000 ? CRCPOLY : 0;
258 * remainder = (remainder << 1 | next_input_bit()) ^ multiple;
260 * Notice how, to get at bit 32 of the shifted remainder, we look
261 * at bit 31 of the remainder *before* shifting it.
263 * But also notice how the next_input_bit() bits we're shifting into
264 * the remainder don't actually affect any decision-making until
265 * 32 bits later. Thus, the first 32 cycles of this are pretty boring.
266 * Also, to add the CRC to a message, we need a 32-bit-long hole for it at
267 * the end, so we have to add 32 extra cycles shifting in zeros at the
268 * end of every message,
270 * So the standard trick is to rearrage merging in the next_input_bit()
271 * until the moment it's needed. Then the first 32 cycles can be precomputed,
272 * and merging in the final 32 zero bits to make room for the CRC can be
274 * This changes the code to:
275 * for (i = 0; i < input_bits; i++) {
276 * remainder ^= next_input_bit() << 31;
277 * multiple = (remainder & 0x80000000) ? CRCPOLY : 0;
278 * remainder = (remainder << 1) ^ multiple;
280 * With this optimization, the little-endian code is simpler:
281 * for (i = 0; i < input_bits; i++) {
282 * remainder ^= next_input_bit();
283 * multiple = (remainder & 1) ? CRCPOLY : 0;
284 * remainder = (remainder >> 1) ^ multiple;
287 * Note that the other details of endianness have been hidden in CRCPOLY
288 * (which must be bit-reversed) and next_input_bit().
290 * However, as long as next_input_bit is returning the bits in a sensible
291 * order, we can actually do the merging 8 or more bits at a time rather
292 * than one bit at a time:
293 * for (i = 0; i < input_bytes; i++) {
294 * remainder ^= next_input_byte() << 24;
295 * for (j = 0; j < 8; j++) {
296 * multiple = (remainder & 0x80000000) ? CRCPOLY : 0;
297 * remainder = (remainder << 1) ^ multiple;
300 * Or in little-endian:
301 * for (i = 0; i < input_bytes; i++) {
302 * remainder ^= next_input_byte();
303 * for (j = 0; j < 8; j++) {
304 * multiple = (remainder & 1) ? CRCPOLY : 0;
305 * remainder = (remainder << 1) ^ multiple;
308 * If the input is a multiple of 32 bits, you can even XOR in a 32-bit
309 * word at a time and increase the inner loop count to 32.
311 * You can also mix and match the two loop styles, for example doing the
312 * bulk of a message byte-at-a-time and adding bit-at-a-time processing
313 * for any fractional bytes at the end.
315 * The only remaining optimization is to the byte-at-a-time table method.
316 * Here, rather than just shifting one bit of the remainder to decide
317 * in the correct multiple to subtract, we can shift a byte at a time.
318 * This produces a 40-bit (rather than a 33-bit) intermediate remainder,
319 * but again the multiple of the polynomial to subtract depends only on
320 * the high bits, the high 8 bits in this case.
322 * The multiple we need in that case is the low 32 bits of a 40-bit
323 * value whose high 8 bits are given, and which is a multiple of the
324 * generator polynomial. This is simply the CRC-32 of the given
327 * Two more details: normally, appending zero bits to a message which
328 * is already a multiple of a polynomial produces a larger multiple of that
329 * polynomial. To enable a CRC to detect this condition, it's common to
330 * invert the CRC before appending it. This makes the remainder of the
331 * message+crc come out not as zero, but some fixed non-zero value.
333 * The same problem applies to zero bits prepended to the message, and
334 * a similar solution is used. Instead of starting with a remainder of
335 * 0, an initial remainder of all ones is used. As long as you start
336 * the same way on decoding, it doesn't make a difference.
344 #if 0 /*Not used at present */
346 buf_dump(char const *prefix, unsigned char const *buf, size_t len)
348 fputs(prefix, stdout);
350 printf(" %02x", *buf++);
356 static void bytereverse(unsigned char *buf, size_t len)
359 unsigned char x = bitrev8(*buf);
364 static void random_garbage(unsigned char *buf, size_t len)
367 *buf++ = (unsigned char) random();
370 #if 0 /* Not used at present */
371 static void store_le(u32 x, unsigned char *buf)
373 buf[0] = (unsigned char) x;
374 buf[1] = (unsigned char) (x >> 8);
375 buf[2] = (unsigned char) (x >> 16);
376 buf[3] = (unsigned char) (x >> 24);
380 static void store_be(u32 x, unsigned char *buf)
382 buf[0] = (unsigned char) (x >> 24);
383 buf[1] = (unsigned char) (x >> 16);
384 buf[2] = (unsigned char) (x >> 8);
385 buf[3] = (unsigned char) x;
389 * This checks that CRC(buf + CRC(buf)) = 0, and that
390 * CRC commutes with bit-reversal. This has the side effect
391 * of bytewise bit-reversing the input buffer, and returns
392 * the CRC of the reversed buffer.
394 static u32 test_step(u32 init, unsigned char *buf, size_t len)
399 crc1 = crc32_be(init, buf, len);
400 store_be(crc1, buf + len);
401 crc2 = crc32_be(init, buf, len + 4);
403 printf("\nCRC cancellation fail: 0x%08x should be 0\n",
406 for (i = 0; i <= len + 4; i++) {
407 crc2 = crc32_be(init, buf, i);
408 crc2 = crc32_be(crc2, buf + i, len + 4 - i);
410 printf("\nCRC split fail: 0x%08x\n", crc2);
413 /* Now swap it around for the other test */
415 bytereverse(buf, len + 4);
416 init = bitrev32(init);
417 crc2 = bitrev32(crc1);
418 if (crc1 != bitrev32(crc2))
419 printf("\nBit reversal fail: 0x%08x -> 0x%08x -> 0x%08x\n",
420 crc1, crc2, bitrev32(crc2));
421 crc1 = crc32_le(init, buf, len);
423 printf("\nCRC endianness fail: 0x%08x != 0x%08x\n", crc1,
425 crc2 = crc32_le(init, buf, len + 4);
427 printf("\nCRC cancellation fail: 0x%08x should be 0\n",
430 for (i = 0; i <= len + 4; i++) {
431 crc2 = crc32_le(init, buf, i);
432 crc2 = crc32_le(crc2, buf + i, len + 4 - i);
434 printf("\nCRC split fail: 0x%08x\n", crc2);
446 unsigned char buf1[SIZE + 4];
447 unsigned char buf2[SIZE + 4];
448 unsigned char buf3[SIZE + 4];
450 u32 crc1, crc2, crc3;
452 for (i = 0; i <= SIZE; i++) {
453 printf("\rTesting length %d...", i);
455 random_garbage(buf1, i);
456 random_garbage(buf2, i);
457 for (j = 0; j < i; j++)
458 buf3[j] = buf1[j] ^ buf2[j];
460 crc1 = test_step(INIT1, buf1, i);
461 crc2 = test_step(INIT2, buf2, i);
462 /* Now check that CRC(buf1 ^ buf2) = CRC(buf1) ^ CRC(buf2) */
463 crc3 = test_step(INIT1 ^ INIT2, buf3, i);
464 if (crc3 != (crc1 ^ crc2))
465 printf("CRC XOR fail: 0x%08x != 0x%08x ^ 0x%08x\n",
468 printf("\nAll test complete. No failures expected.\n");
472 #endif /* UNITTEST */