Patchwork [1/3] crc32c: Implement CRC32c with slicing-by-8 algorithm

login
register
mail settings
Submitter Darrick J. Wong
Date Sept. 1, 2011, 12:33 a.m.
Message ID <20110901003324.32645.51490.stgit@elm3c44.beaverton.ibm.com>
Download mbox | patch
Permalink /patch/112747/
State Superseded
Headers show

Comments

Darrick J. Wong - Sept. 1, 2011, 12:33 a.m.
The existing CRC32c implementation uses Sarwate's algorithm to calculate the
code one byte at a time.  Using slicing-by-8, we can process buffers 8 bytes at
a time, for a substantial increase in performance.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
---
 crypto/Makefile      |   11 +
 crypto/crc32c.c      |  635 ++++++++++++++++++++++++++++++++++++++++++--------
 crypto/crc32c_defs.h |   34 +++
 3 files changed, 576 insertions(+), 104 deletions(-)
 create mode 100644 crypto/crc32c_defs.h



--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/crypto/Makefile b/crypto/Makefile
index ce5a813..00811ef 100644
--- a/crypto/Makefile
+++ b/crypto/Makefile
@@ -94,3 +94,14 @@  obj-$(CONFIG_CRYPTO_USER_API_SKCIPHER) += algif_skcipher.o
 #
 obj-$(CONFIG_XOR_BLOCKS) += xor.o
 obj-$(CONFIG_ASYNC_CORE) += async_tx/
+
+hostprogs-y	:= gen_crc32ctable
+clean-files	:= crc32ctable.h
+
+$(obj)/crc32c.o: $(obj)/crc32c_table.h
+
+quiet_cmd_crc32c = GEN     $@
+      cmd_crc32c = $< > $@
+
+$(obj)/crc32c_table.h: $(obj)/gen_crc32ctable
+	$(call cmd,crc32c)
diff --git a/crypto/crc32c.c b/crypto/crc32c.c
index 3f9ad28..d18f6a1 100644
--- a/crypto/crc32c.c
+++ b/crypto/crc32c.c
@@ -33,6 +33,35 @@ 
  * Software Foundation; either version 2 of the License, or (at your option)
  * any later version.
  *
+ * The current crc32c implementation is adapted from Bob Pearson's slice-by-8
+ * crc32 kernel patch from mid-2011.
+ *
+ * August 26, 2011 Darrick J. Wong <djwong at us.ibm.com>
+ * Reuse Bob Pearson's slice-by-8 implementation for e2fsprogs.
+ *
+ * July 20, 2011 Bob Pearson <rpearson at systemfabricworks.com>
+ * added slice by 8 algorithm to the existing conventional and
+ * slice by 4 algorithms.
+ *
+ * Oct 15, 2000 Matt Domsch <Matt_Domsch@dell.com>
+ * Nicer crc32 functions/docs submitted by linux@horizon.com.  Thanks!
+ * Code was from the public domain, copyright abandoned.  Code was
+ * subsequently included in the kernel, thus was re-licensed under the
+ * GNU GPL v2.
+ *
+ * Oct 12, 2000 Matt Domsch <Matt_Domsch@dell.com>
+ * Same crc32 function was used in 5 other places in the kernel.
+ * I made one version, and deleted the others.
+ * There are various incantations of crc32().  Some use a seed of 0 or ~0.
+ * Some xor at the end with ~0.  The generic crc32() function takes
+ * seed as an argument, and doesn't xor at the end.  Then individual
+ * users can do whatever they need.
+ *   drivers/net/smc9194.c uses seed ~0, doesn't xor with ~0.
+ *   fs/jffs2 uses seed 0, doesn't xor with ~0.
+ *   fs/partitions/efi.c uses seed ~0, xor's with ~0.
+ *
+ * This source code is licensed under the GNU General Public License,
+ * Version 2.  See the file COPYING for more details.
  */
 
 #include <crypto/internal/hash.h>
@@ -40,6 +69,7 @@ 
 #include <linux/module.h>
 #include <linux/string.h>
 #include <linux/kernel.h>
+#include "crc32c_defs.h"
 
 #define CHKSUM_BLOCK_SIZE	1
 #define CHKSUM_DIGEST_SIZE	4
@@ -52,92 +82,398 @@  struct chksum_desc_ctx {
 	u32 crc;
 };
 
-/*
- * This is the CRC-32C table
- * Generated with:
- * width = 32 bits
- * poly = 0x1EDC6F41
- * reflect input bytes = true
- * reflect output bytes = true
- */
+#if CRC_LE_BITS > 8
+# define tole(x) (__force u32) __constant_cpu_to_le32(x)
+#else
+# define tole(x) (x)
+#endif
 
-static const u32 crc32c_table[256] = {
-	0x00000000L, 0xF26B8303L, 0xE13B70F7L, 0x1350F3F4L,
-	0xC79A971FL, 0x35F1141CL, 0x26A1E7E8L, 0xD4CA64EBL,
-	0x8AD958CFL, 0x78B2DBCCL, 0x6BE22838L, 0x9989AB3BL,
-	0x4D43CFD0L, 0xBF284CD3L, 0xAC78BF27L, 0x5E133C24L,
-	0x105EC76FL, 0xE235446CL, 0xF165B798L, 0x030E349BL,
-	0xD7C45070L, 0x25AFD373L, 0x36FF2087L, 0xC494A384L,
-	0x9A879FA0L, 0x68EC1CA3L, 0x7BBCEF57L, 0x89D76C54L,
-	0x5D1D08BFL, 0xAF768BBCL, 0xBC267848L, 0x4E4DFB4BL,
-	0x20BD8EDEL, 0xD2D60DDDL, 0xC186FE29L, 0x33ED7D2AL,
-	0xE72719C1L, 0x154C9AC2L, 0x061C6936L, 0xF477EA35L,
-	0xAA64D611L, 0x580F5512L, 0x4B5FA6E6L, 0xB93425E5L,
-	0x6DFE410EL, 0x9F95C20DL, 0x8CC531F9L, 0x7EAEB2FAL,
-	0x30E349B1L, 0xC288CAB2L, 0xD1D83946L, 0x23B3BA45L,
-	0xF779DEAEL, 0x05125DADL, 0x1642AE59L, 0xE4292D5AL,
-	0xBA3A117EL, 0x4851927DL, 0x5B016189L, 0xA96AE28AL,
-	0x7DA08661L, 0x8FCB0562L, 0x9C9BF696L, 0x6EF07595L,
-	0x417B1DBCL, 0xB3109EBFL, 0xA0406D4BL, 0x522BEE48L,
-	0x86E18AA3L, 0x748A09A0L, 0x67DAFA54L, 0x95B17957L,
-	0xCBA24573L, 0x39C9C670L, 0x2A993584L, 0xD8F2B687L,
-	0x0C38D26CL, 0xFE53516FL, 0xED03A29BL, 0x1F682198L,
-	0x5125DAD3L, 0xA34E59D0L, 0xB01EAA24L, 0x42752927L,
-	0x96BF4DCCL, 0x64D4CECFL, 0x77843D3BL, 0x85EFBE38L,
-	0xDBFC821CL, 0x2997011FL, 0x3AC7F2EBL, 0xC8AC71E8L,
-	0x1C661503L, 0xEE0D9600L, 0xFD5D65F4L, 0x0F36E6F7L,
-	0x61C69362L, 0x93AD1061L, 0x80FDE395L, 0x72966096L,
-	0xA65C047DL, 0x5437877EL, 0x4767748AL, 0xB50CF789L,
-	0xEB1FCBADL, 0x197448AEL, 0x0A24BB5AL, 0xF84F3859L,
-	0x2C855CB2L, 0xDEEEDFB1L, 0xCDBE2C45L, 0x3FD5AF46L,
-	0x7198540DL, 0x83F3D70EL, 0x90A324FAL, 0x62C8A7F9L,
-	0xB602C312L, 0x44694011L, 0x5739B3E5L, 0xA55230E6L,
-	0xFB410CC2L, 0x092A8FC1L, 0x1A7A7C35L, 0xE811FF36L,
-	0x3CDB9BDDL, 0xCEB018DEL, 0xDDE0EB2AL, 0x2F8B6829L,
-	0x82F63B78L, 0x709DB87BL, 0x63CD4B8FL, 0x91A6C88CL,
-	0x456CAC67L, 0xB7072F64L, 0xA457DC90L, 0x563C5F93L,
-	0x082F63B7L, 0xFA44E0B4L, 0xE9141340L, 0x1B7F9043L,
-	0xCFB5F4A8L, 0x3DDE77ABL, 0x2E8E845FL, 0xDCE5075CL,
-	0x92A8FC17L, 0x60C37F14L, 0x73938CE0L, 0x81F80FE3L,
-	0x55326B08L, 0xA759E80BL, 0xB4091BFFL, 0x466298FCL,
-	0x1871A4D8L, 0xEA1A27DBL, 0xF94AD42FL, 0x0B21572CL,
-	0xDFEB33C7L, 0x2D80B0C4L, 0x3ED04330L, 0xCCBBC033L,
-	0xA24BB5A6L, 0x502036A5L, 0x4370C551L, 0xB11B4652L,
-	0x65D122B9L, 0x97BAA1BAL, 0x84EA524EL, 0x7681D14DL,
-	0x2892ED69L, 0xDAF96E6AL, 0xC9A99D9EL, 0x3BC21E9DL,
-	0xEF087A76L, 0x1D63F975L, 0x0E330A81L, 0xFC588982L,
-	0xB21572C9L, 0x407EF1CAL, 0x532E023EL, 0xA145813DL,
-	0x758FE5D6L, 0x87E466D5L, 0x94B49521L, 0x66DF1622L,
-	0x38CC2A06L, 0xCAA7A905L, 0xD9F75AF1L, 0x2B9CD9F2L,
-	0xFF56BD19L, 0x0D3D3E1AL, 0x1E6DCDEEL, 0xEC064EEDL,
-	0xC38D26C4L, 0x31E6A5C7L, 0x22B65633L, 0xD0DDD530L,
-	0x0417B1DBL, 0xF67C32D8L, 0xE52CC12CL, 0x1747422FL,
-	0x49547E0BL, 0xBB3FFD08L, 0xA86F0EFCL, 0x5A048DFFL,
-	0x8ECEE914L, 0x7CA56A17L, 0x6FF599E3L, 0x9D9E1AE0L,
-	0xD3D3E1ABL, 0x21B862A8L, 0x32E8915CL, 0xC083125FL,
-	0x144976B4L, 0xE622F5B7L, 0xF5720643L, 0x07198540L,
-	0x590AB964L, 0xAB613A67L, 0xB831C993L, 0x4A5A4A90L,
-	0x9E902E7BL, 0x6CFBAD78L, 0x7FAB5E8CL, 0x8DC0DD8FL,
-	0xE330A81AL, 0x115B2B19L, 0x020BD8EDL, 0xF0605BEEL,
-	0x24AA3F05L, 0xD6C1BC06L, 0xC5914FF2L, 0x37FACCF1L,
-	0x69E9F0D5L, 0x9B8273D6L, 0x88D28022L, 0x7AB90321L,
-	0xAE7367CAL, 0x5C18E4C9L, 0x4F48173DL, 0xBD23943EL,
-	0xF36E6F75L, 0x0105EC76L, 0x12551F82L, 0xE03E9C81L,
-	0x34F4F86AL, 0xC69F7B69L, 0xD5CF889DL, 0x27A40B9EL,
-	0x79B737BAL, 0x8BDCB4B9L, 0x988C474DL, 0x6AE7C44EL,
-	0xBE2DA0A5L, 0x4C4623A6L, 0x5F16D052L, 0xAD7D5351L
-};
+#if CRC_BE_BITS > 8
+# define tobe(x) (__force u32) __constant_cpu_to_be32(x)
+#else
+# define tobe(x) (x)
+#endif
 
-/*
- * Steps through buffer one byte at at time, calculates reflected
- * crc using table.
- */
+#include "crc32c_table.h"
+
+#if CRC_LE_BITS == 32
+/* slice by 4 algorithm */
+static u32 crc32c_le_body(u32 crc, u8 const *buf, size_t len)
+{
+	const u8 *p8;
+	const u32 *p32;
+	size_t init_bytes;
+	size_t words;
+	size_t end_bytes;
+	size_t i;
+	u32 q;
+	u8 i0, i1, i2, i3;
+
+	crc = (__force u32) __cpu_to_le32(crc);
+
+	/* unroll loop into 'init_bytes' odd bytes followed by
+	 * 'words' aligned 4 byte words followed by
+	 * 'end_bytes' odd bytes at the end */
+	p8 = buf;
+	p32 = (u32 *)PTR_ALIGN(p8, 4);
+	init_bytes = min((uintptr_t)p32 - (uintptr_t)p8, len);
+	words = (len - init_bytes) >> 2;
+	end_bytes = (len - init_bytes) & 3;
+
+	for (i = 0; i < init_bytes; i++) {
+#ifdef __LITTLE_ENDIAN
+		i0 = *p8++ ^ crc;
+		crc = t0_le[i0] ^ (crc >> 8);
+#else
+		i0 = *p8++ ^ (crc >> 24);
+		crc = t0_le[i0] ^ (crc << 8);
+#endif
+	}
+
+	/* using pre-increment below slightly faster */
+	p32--;
 
-static u32 crc32c(u32 crc, const u8 *data, unsigned int length)
+	for (i = 0; i < words; i++) {
+#ifdef __LITTLE_ENDIAN
+		q = *++p32 ^ crc;
+		i3 = q;
+		i2 = q >> 8;
+		i1 = q >> 16;
+		i0 = q >> 24;
+		crc = t3_le[i3] ^ t2_le[i2] ^ t1_le[i1] ^ t0_le[i0];
+#else
+		q = *++p32 ^ crc;
+		i3 = q >> 24;
+		i2 = q >> 16;
+		i1 = q >> 8;
+		i0 = q;
+		crc = t3_le[i3] ^ t2_le[i2] ^ t1_le[i1] ^ t0_le[i0];
+#endif
+	}
+
+	p8 = (u8 *)(++p32);
+
+	for (i = 0; i < end_bytes; i++) {
+#ifdef __LITTLE_ENDIAN
+		i0 = *p8++ ^ crc;
+		crc = t0_le[i0] ^ (crc >> 8);
+#else
+		i0 = *p8++ ^ (crc >> 24);
+		crc = t0_le[i0] ^ (crc << 8);
+#endif
+	}
+
+	return __le32_to_cpu((__force __le32)crc);
+}
+#endif
+
+#if CRC_BE_BITS == 32
+static u32 crc32c_be_body(u32 crc, u8 const *buf, size_t len)
 {
-	while (length--)
-		crc = crc32c_table[(crc ^ *data++) & 0xFFL] ^ (crc >> 8);
+	const u8 *p8;
+	const u32 *p32;
+	size_t init_bytes;
+	size_t words;
+	size_t end_bytes;
+	size_t i;
+	u32 q;
+	u8 i0, i1, i2, i3;
+
+	crc = (__force u32) __cpu_to_be32(crc);
+
+	p8 = buf;
+	p32 = (u32 *)PTR_ALIGN(p8, 4);
+	init_bytes = min((uintptr_t)p32 - (uintptr_t)p8, len);
+	words = (len - init_bytes) >> 2;
+	end_bytes = (len - init_bytes) & 3;
+
+	for (i = 0; i < init_bytes; i++) {
+#ifdef __LITTLE_ENDIAN
+		i0 = *p8++ ^ crc;
+		crc = t0_be[i0] ^ (crc >> 8);
+#else
+		i0 = *p8++ ^ (crc >> 24);
+		crc = t0_be[i0] ^ (crc << 8);
+#endif
+	}
+
+	p32--;
+
+	for (i = 0; i < words; i++) {
+#ifdef __LITTLE_ENDIAN
+		q = *++p32 ^ crc;
+		i3 = q;
+		i2 = q >> 8;
+		i1 = q >> 16;
+		i0 = q >> 24;
+		crc = t3_be[i3] ^ t2_be[i2] ^ t1_be[i1] ^ t0_be[i0];
+#else
+		q = *++p32 ^ crc;
+		i3 = q >> 24;
+		i2 = q >> 16;
+		i1 = q >> 8;
+		i0 = q;
+		crc = t3_be[i3] ^ t2_be[i2] ^ t1_be[i1] ^ t0_be[i0];
+#endif
+	}
+
+	p8 = (u8 *)(++p32);
+
+	for (i = 0; i < end_bytes; i++) {
+#ifdef __LITTLE_ENDIAN
+		i0 = *p8++ ^ crc;
+		crc = t0_be[i0] ^ (crc >> 8);
+#else
+		i0 = *p8++ ^ (crc >> 24);
+		crc = t0_be[i0] ^ (crc << 8);
+#endif
+	}
+
+	return __be32_to_cpu((__force __be32)crc);
+}
+#endif
+
+#if CRC_LE_BITS == 64
+/* slice by 8 algorithm */
+static u32 crc32c_le_body(u32 crc, u8 const *buf, size_t len)
+{
+	const u8 *p8;
+	const u32 *p32;
+	size_t init_bytes;
+	size_t words;
+	size_t end_bytes;
+	size_t i;
+	u32 q;
+	u8 i0, i1, i2, i3;
+
+	crc = (__force u32) __cpu_to_le32(crc);
+
+	p8 = buf;
+	p32 = (u32 *)PTR_ALIGN(p8, 8);
+	init_bytes = min((uintptr_t)p32 - (uintptr_t)p8, len);
+	words = (len - init_bytes) >> 3;
+	end_bytes = (len - init_bytes) & 7;
+
+	for (i = 0; i < init_bytes; i++) {
+#ifdef __LITTLE_ENDIAN
+		i0 = *p8++ ^ crc;
+		crc = t0_le[i0] ^ (crc >> 8);
+#else
+		i0 = *p8++ ^ (crc >> 24);
+		crc = t0_le[i0] ^ (crc << 8);
+#endif
+	}
+
+	p32--;
+
+	for (i = 0; i < words; i++) {
+#ifdef __LITTLE_ENDIAN
+		q = *++p32 ^ crc;
+		i3 = q;
+		i2 = q >> 8;
+		i1 = q >> 16;
+		i0 = q >> 24;
+		crc = t7_le[i3] ^ t6_le[i2] ^ t5_le[i1] ^ t4_le[i0];
+
+		q = *++p32;
+		i3 = q;
+		i2 = q >> 8;
+		i1 = q >> 16;
+		i0 = q >> 24;
+		crc ^= t3_le[i3] ^ t2_le[i2] ^ t1_le[i1] ^ t0_le[i0];
+#else
+		q = *++p32 ^ crc;
+		i3 = q >> 24;
+		i2 = q >> 16;
+		i1 = q >> 8;
+		i0 = q;
+		crc = t7_le[i3] ^ t6_le[i2] ^ t5_le[i1] ^ t4_le[i0];
+
+		q = *++p32;
+		i3 = q >> 24;
+		i2 = q >> 16;
+		i1 = q >> 8;
+		i0 = q;
+		crc ^= t3_le[i3] ^ t2_le[i2] ^ t1_le[i1] ^ t0_le[i0];
+#endif
+	}
+
+	p8 = (u8 *)(++p32);
+
+	for (i = 0; i < end_bytes; i++) {
+#ifdef __LITTLE_ENDIAN
+		i0 = *p8++ ^ crc;
+		crc = t0_le[i0] ^ (crc >> 8);
+#else
+		i0 = *p8++ ^ (crc >> 24);
+		crc = t0_le[i0] ^ (crc << 8);
+#endif
+	}
+
+	return __le32_to_cpu(crc);
+}
+#endif
+
+#if CRC_BE_BITS == 64
+static u32 crc32c_be_body(u32 crc, u8 const *buf, size_t len)
+{
+	const u8 *p8;
+	const u32 *p32;
+	size_t init_bytes;
+	size_t words;
+	size_t end_bytes;
+	size_t i;
+	u32 q;
+	u8 i0, i1, i2, i3;
+
+	crc = (__force u32) __cpu_to_be32(crc);
+
+	p8 = buf;
+	p32 = (u32 *)PTR_ALIGN(p8, 8);
+	init_bytes = min((uintptr_t)p32 - (uintptr_t)p8, len);
+	words = (len - init_bytes) >> 3;
+	end_bytes = (len - init_bytes) & 7;
+
+	for (i = 0; i < init_bytes; i++) {
+#ifdef __LITTLE_ENDIAN
+		i0 = *p8++ ^ crc;
+		crc = t0_be[i0] ^ (crc >> 8);
+#else
+		i0 = *p8++ ^ (crc >> 24);
+		crc = t0_be[i0] ^ (crc << 8);
+#endif
+	}
+
+	p32--;
+
+	for (i = 0; i < words; i++) {
+#ifdef __LITTLE_ENDIAN
+		q = *++p32 ^ crc;
+		i3 = q;
+		i2 = q >> 8;
+		i1 = q >> 16;
+		i0 = q >> 24;
+		crc = t7_be[i3] ^ t6_be[i2] ^ t5_be[i1] ^ t4_be[i0];
+
+		q = *++p32;
+		i3 = q;
+		i2 = q >> 8;
+		i1 = q >> 16;
+		i0 = q >> 24;
+		crc ^= t3_be[i3] ^ t2_be[i2] ^ t1_be[i1] ^ t0_be[i0];
+#else
+		q = *++p32 ^ crc;
+		i3 = q >> 24;
+		i2 = q >> 16;
+		i1 = q >> 8;
+		i0 = q;
+		crc = t7_be[i3] ^ t6_be[i2] ^ t5_be[i1] ^ t4_be[i0];
+
+		q = *++p32;
+		i3 = q >> 24;
+		i2 = q >> 16;
+		i1 = q >> 8;
+		i0 = q;
+		crc ^= t3_be[i3] ^ t2_be[i2] ^ t1_be[i1] ^ t0_be[i0];
+#endif
+	}
+
+	p8 = (u8 *)(++p32);
+
+	for (i = 0; i < end_bytes; i++) {
+#ifdef __LITTLE_ENDIAN
+		i0 = *p8++ ^ crc;
+		crc = t0_be[i0] ^ (crc >> 8);
+#else
+		i0 = *p8++ ^ (crc >> 24);
+		crc = t0_be[i0] ^ (crc << 8);
+#endif
+	}
+
+	return __be32_to_cpu(crc);
+}
+#endif
+
+/**
+ * crc32c_le() - Calculate bitwise little-endian CRC32c.
+ * @crc: seed value for computation.  ~0 for ext4, sometimes 0 for
+ *	other uses, or the previous crc32c value if computing incrementally.
+ * @p: pointer to buffer over which CRC is run
+ * @len: length of buffer @p
+ */
+static u32 crc32c_le(u32 crc, unsigned char const *p, size_t len)
+{
+#if CRC_LE_BITS == 1
+	int i;
+	while (len--) {
+		crc ^= *p++;
+		for (i = 0; i < 8; i++)
+			crc = (crc >> 1) ^ ((crc & 1) ? CRCPOLY_LE : 0);
+	}
+# elif CRC_LE_BITS == 2
+	while (len--) {
+		crc ^= *p++;
+		crc = (crc >> 2) ^ t0_le[crc & 0x03];
+		crc = (crc >> 2) ^ t0_le[crc & 0x03];
+		crc = (crc >> 2) ^ t0_le[crc & 0x03];
+		crc = (crc >> 2) ^ t0_le[crc & 0x03];
+	}
+# elif CRC_LE_BITS == 4
+	while (len--) {
+		crc ^= *p++;
+		crc = (crc >> 4) ^ t0_le[crc & 0x0f];
+		crc = (crc >> 4) ^ t0_le[crc & 0x0f];
+	}
+# elif CRC_LE_BITS == 8
+	while (len--) {
+		crc ^= *p++;
+		crc = (crc >> 8) ^ t0_le[crc & 0xff];
+	}
+# else
+	crc = crc32c_le_body(crc, p, len);
+# endif
+	return crc;
+}
 
+/**
+ * crc32c_be() - Calculate bitwise big-endian CRC32c.
+ * @crc: seed value for computation.  ~0 for ext4, sometimes 0 for
+ *	other uses, or the previous crc32c value if computing incrementally.
+ * @p: pointer to buffer over which CRC is run
+ * @len: length of buffer @p
+ */
+static u32 crc32c_be(u32 crc, unsigned char const *p, size_t len)
+{
+#if CRC_BE_BITS == 1
+	int i;
+	while (len--) {
+		crc ^= *p++ << 24;
+		for (i = 0; i < 8; i++)
+			crc = (crc << 1) ^
+			      ((crc & 0x80000000) ? CRCPOLY_BE : 0);
+	}
+# elif CRC_BE_BITS == 2
+	while (len--) {
+		crc ^= *p++ << 24;
+		crc = (crc << 2) ^ t0_be[crc >> 30];
+		crc = (crc << 2) ^ t0_be[crc >> 30];
+		crc = (crc << 2) ^ t0_be[crc >> 30];
+		crc = (crc << 2) ^ t0_be[crc >> 30];
+	}
+# elif CRC_BE_BITS == 4
+	while (len--) {
+		crc ^= *p++ << 24;
+		crc = (crc << 4) ^ t0_be[crc >> 28];
+		crc = (crc << 4) ^ t0_be[crc >> 28];
+	}
+# elif CRC_BE_BITS == 8
+	while (len--) {
+		crc ^= *p++ << 24;
+		crc = (crc << 8) ^ t0_be[crc >> 24];
+	}
+# else
+	crc = crc32c_be_body(crc, p, len);
+# endif
 	return crc;
 }
 
@@ -146,7 +482,7 @@  static u32 crc32c(u32 crc, const u8 *data, unsigned int length)
  * crc using table.
  */
 
-static int chksum_init(struct shash_desc *desc)
+static int crc32c_init(struct shash_desc *desc)
 {
 	struct chksum_ctx *mctx = crypto_shash_ctx(desc->tfm);
 	struct chksum_desc_ctx *ctx = shash_desc_ctx(desc);
@@ -156,12 +492,21 @@  static int chksum_init(struct shash_desc *desc)
 	return 0;
 }
 
+static int crc32c_cra_init(struct crypto_tfm *tfm)
+{
+	struct chksum_ctx *mctx = crypto_tfm_ctx(tfm);
+
+	mctx->key = ~0;
+	return 0;
+}
+
+/* Little Endian version of algorithm */
 /*
  * Setting the seed allows arbitrary accumulators and flexible XOR policy
  * If your algorithm starts with ~0, then XOR with ~0 before you set
  * the seed.
  */
-static int chksum_setkey(struct crypto_shash *tfm, const u8 *key,
+static int crc32c_le_setkey(struct crypto_shash *tfm, const u8 *key,
 			 unsigned int keylen)
 {
 	struct chksum_ctx *mctx = crypto_shash_ctx(tfm);
@@ -174,16 +519,16 @@  static int chksum_setkey(struct crypto_shash *tfm, const u8 *key,
 	return 0;
 }
 
-static int chksum_update(struct shash_desc *desc, const u8 *data,
+static int crc32c_le_update(struct shash_desc *desc, const u8 *data,
 			 unsigned int length)
 {
 	struct chksum_desc_ctx *ctx = shash_desc_ctx(desc);
 
-	ctx->crc = crc32c(ctx->crc, data, length);
+	ctx->crc = crc32c_le(ctx->crc, data, length);
 	return 0;
 }
 
-static int chksum_final(struct shash_desc *desc, u8 *out)
+static int crc32c_le_final(struct shash_desc *desc, u8 *out)
 {
 	struct chksum_desc_ctx *ctx = shash_desc_ctx(desc);
 
@@ -191,44 +536,96 @@  static int chksum_final(struct shash_desc *desc, u8 *out)
 	return 0;
 }
 
-static int __chksum_finup(u32 *crcp, const u8 *data, unsigned int len, u8 *out)
+static int __crc32c_le_finup(u32 *crcp, const u8 *data, unsigned int len,
+			     u8 *out)
 {
-	*(__le32 *)out = ~cpu_to_le32(crc32c(*crcp, data, len));
+	*(__le32 *)out = ~cpu_to_le32(crc32c_le(*crcp, data, len));
 	return 0;
 }
 
-static int chksum_finup(struct shash_desc *desc, const u8 *data,
+static int crc32c_le_finup(struct shash_desc *desc, const u8 *data,
 			unsigned int len, u8 *out)
 {
 	struct chksum_desc_ctx *ctx = shash_desc_ctx(desc);
 
-	return __chksum_finup(&ctx->crc, data, len, out);
+	return __crc32c_le_finup(&ctx->crc, data, len, out);
 }
 
-static int chksum_digest(struct shash_desc *desc, const u8 *data,
+static int crc32c_le_digest(struct shash_desc *desc, const u8 *data,
 			 unsigned int length, u8 *out)
 {
 	struct chksum_ctx *mctx = crypto_shash_ctx(desc->tfm);
 
-	return __chksum_finup(&mctx->key, data, length, out);
+	return __crc32c_le_finup(&mctx->key, data, length, out);
 }
 
-static int crc32c_cra_init(struct crypto_tfm *tfm)
+/* Big Endian version of algorithm */
+/*
+ * Setting the seed allows arbitrary accumulators and flexible XOR policy
+ * If your algorithm starts with ~0, then XOR with ~0 before you set
+ * the seed.
+ */
+static int crc32c_be_setkey(struct crypto_shash *tfm, const u8 *key,
+			 unsigned int keylen)
 {
-	struct chksum_ctx *mctx = crypto_tfm_ctx(tfm);
+	struct chksum_ctx *mctx = crypto_shash_ctx(tfm);
 
-	mctx->key = ~0;
+	if (keylen != sizeof(mctx->key)) {
+		crypto_shash_set_flags(tfm, CRYPTO_TFM_RES_BAD_KEY_LEN);
+		return -EINVAL;
+	}
+	mctx->key = be32_to_cpu(*(__be32 *)key);
 	return 0;
 }
 
-static struct shash_alg alg = {
+static int crc32c_be_update(struct shash_desc *desc, const u8 *data,
+			 unsigned int length)
+{
+	struct chksum_desc_ctx *ctx = shash_desc_ctx(desc);
+
+	ctx->crc = crc32c_be(ctx->crc, data, length);
+	return 0;
+}
+
+static int crc32c_be_final(struct shash_desc *desc, u8 *out)
+{
+	struct chksum_desc_ctx *ctx = shash_desc_ctx(desc);
+
+	*(__be32 *)out = ~cpu_to_be32p(&ctx->crc);
+	return 0;
+}
+
+static int __crc32c_be_finup(u32 *crcp, const u8 *data, unsigned int len,
+			     u8 *out)
+{
+	*(__be32 *)out = ~cpu_to_be32(crc32c_be(*crcp, data, len));
+	return 0;
+}
+
+static int crc32c_be_finup(struct shash_desc *desc, const u8 *data,
+			unsigned int len, u8 *out)
+{
+	struct chksum_desc_ctx *ctx = shash_desc_ctx(desc);
+
+	return __crc32c_be_finup(&ctx->crc, data, len, out);
+}
+
+static int crc32c_be_digest(struct shash_desc *desc, const u8 *data,
+			 unsigned int length, u8 *out)
+{
+	struct chksum_ctx *mctx = crypto_shash_ctx(desc->tfm);
+
+	return __crc32c_be_finup(&mctx->key, data, length, out);
+}
+
+static struct shash_alg alg_le = {
 	.digestsize		=	CHKSUM_DIGEST_SIZE,
-	.setkey			=	chksum_setkey,
-	.init		=	chksum_init,
-	.update		=	chksum_update,
-	.final		=	chksum_final,
-	.finup		=	chksum_finup,
-	.digest		=	chksum_digest,
+	.setkey			=	crc32c_le_setkey,
+	.init		=	crc32c_init,
+	.update		=	crc32c_le_update,
+	.final		=	crc32c_le_final,
+	.finup		=	crc32c_le_finup,
+	.digest		=	crc32c_le_digest,
 	.descsize		=	sizeof(struct chksum_desc_ctx),
 	.base			=	{
 		.cra_name		=	"crc32c",
@@ -242,14 +639,44 @@  static struct shash_alg alg = {
 	}
 };
 
+static struct shash_alg alg_be = {
+	.digestsize		=	CHKSUM_DIGEST_SIZE,
+	.setkey			=	crc32c_be_setkey,
+	.init		=	crc32c_init,
+	.update		=	crc32c_be_update,
+	.final		=	crc32c_be_final,
+	.finup		=	crc32c_be_finup,
+	.digest		=	crc32c_be_digest,
+	.descsize		=	sizeof(struct chksum_desc_ctx),
+	.base			=	{
+		.cra_name		=	"crc32c-be",
+		.cra_driver_name	=	"crc32c-generic",
+		.cra_priority		=	100,
+		.cra_blocksize		=	CHKSUM_BLOCK_SIZE,
+		.cra_alignmask		=	3,
+		.cra_ctxsize		=	sizeof(struct chksum_ctx),
+		.cra_module		=	THIS_MODULE,
+		.cra_init		=	crc32c_cra_init,
+	}
+};
+
 static int __init crc32c_mod_init(void)
 {
-	return crypto_register_shash(&alg);
+	int ret;
+
+	ret = crypto_register_shash(&alg_le);
+	if (ret)
+		return ret;
+	ret = crypto_register_shash(&alg_be);
+	if (ret)
+		crypto_unregister_shash(&alg_le);
+	return ret;
 }
 
 static void __exit crc32c_mod_fini(void)
 {
-	crypto_unregister_shash(&alg);
+	crypto_unregister_shash(&alg_be);
+	crypto_unregister_shash(&alg_le);
 }
 
 module_init(crc32c_mod_init);
diff --git a/crypto/crc32c_defs.h b/crypto/crc32c_defs.h
new file mode 100644
index 0000000..977df8f
--- /dev/null
+++ b/crypto/crc32c_defs.h
@@ -0,0 +1,34 @@ 
+/*
+ * This is the CRC32c polynomial, as outlined by Castagnoli.
+ * x^32+x^28+x^27+x^26+x^25+x^23+x^22+x^20+x^19+x^18+x^14+x^13+x^11+x^10+x^9+
+ * x^8+x^6+x^0
+ */
+#define CRCPOLY_LE 0x82F63B78
+#define CRCPOLY_BE 0x1EDC6F41
+
+/* How many bits at a time to use.  Valid values are 1, 2, 4, 8, 32 and 64. */
+/* For less performance-sensitive, use 4 */
+#ifndef CRC_LE_BITS
+# define CRC_LE_BITS 64
+#endif
+#ifndef CRC_BE_BITS
+# define CRC_BE_BITS 64
+#endif
+
+/*
+ * Little-endian CRC computation.  Used with serial bit streams sent
+ * lsbit-first.  Be sure to use cpu_to_le32() to append the computed CRC.
+ */
+#if CRC_LE_BITS > 64 || CRC_LE_BITS < 1 || CRC_LE_BITS == 16 || \
+	CRC_LE_BITS & CRC_LE_BITS-1
+# error "CRC_LE_BITS must be one of {1, 2, 4, 8, 32, 64}"
+#endif
+
+/*
+ * Big-endian CRC computation.  Used with serial bit streams sent
+ * msbit-first.  Be sure to use cpu_to_be32() to append the computed CRC.
+ */
+#if CRC_BE_BITS > 64 || CRC_BE_BITS < 1 || CRC_BE_BITS == 16 || \
+	CRC_BE_BITS & CRC_BE_BITS-1
+# error "CRC_BE_BITS must be one of {1, 2, 4, 8, 32, 64}"
+#endif