[PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[Kamlesh Gurudasani]

Add support for MCRC64 engine to calculate 64-bit CRC in Full-CPU mode

MCRC64 engine calculates 64-bit cyclic redundancy checks (CRC)
according to the ISO 3309 standard.

The ISO 3309 64-bit CRC model parameters are as follows:
    Generator Polynomial: x^64 + x^4 + x^3 + x + 1
    Polynomial Value: 0x000000000000001B
    Initial value: 0x0000000000000000
    Reflected Input: False
    Reflected Output: False
    Xor Final: 0x0000000000000000

Tested with
CONFIG_CRYPTO_MANAGER_DISABLE_TESTS is not set
CONFIG_CRYPTO_MANAGER_EXTRA_TESTS=y

and tcrypt,
sudo modprobe tcrypt mode=329 sec=1

User space application implemented using algif_hash,
https://gist.github.com/ti-kamlesh/73abfcc1a33318bb3b199d36b6209e59

Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
---
Changes in v2:
- Add generic implementation of crc64-iso
- Fixes according to review comments
- Link to v1: https://lore.kernel.org/r/20230719-mcrc-upstream-v1-0-dc8798a24c47@ti.com

---
Kamlesh Gurudasani (6):
      lib: add ISO 3309 model crc64
      crypto: crc64 - add crc64-iso framework
      dt-bindings: crypto: Add Texas Instruments MCRC64
      crypto: ti - add driver for MCRC64 engine
      arm64: dts: ti: k3-am62: Add dt node, cbass_main ranges for MCRC64
      arm64: defconfig: enable TI MCRC64 module

 Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml |  47 ++++++++
 MAINTAINERS                                             |   7 ++
 arch/arm64/boot/dts/ti/k3-am62-main.dtsi                |   7 ++
 arch/arm64/boot/dts/ti/k3-am62.dtsi                     |   1 +
 arch/arm64/configs/defconfig                            |   2 +
 crypto/Kconfig                                          |  11 ++
 crypto/Makefile                                         |   1 +
 crypto/crc64_iso_generic.c                              | 119 ++++++++++++++++++
 crypto/tcrypt.c                                         |   5 +
 crypto/testmgr.c                                        |   7 ++
 crypto/testmgr.h                                        | 404 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 drivers/crypto/Kconfig                                  |   1 +
 drivers/crypto/Makefile                                 |   1 +
 drivers/crypto/ti/Kconfig                               |  10 ++
 drivers/crypto/ti/Makefile                              |   2 +
 drivers/crypto/ti/mcrc64.c                              | 442 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 include/linux/crc64.h                                   |   5 +
 lib/crc64-iso.c                                         | 126 +++++++++++++++++++
 lib/crc64.c                                             |  27 +++++
 lib/gen_crc64table.c                                    |   6 +
 20 files changed, 1231 insertions(+)
---
base-commit: 21ef7b1e17d039053edaeaf41142423810572741
change-id: 20230719-mcrc-upstream-7ae9a75cab37

Best regards,
-- 
Kamlesh Gurudasani <kamlesh@ti.com>

[PATCH v2 1/6] lib: add ISO 3309 model crc64

[Kamlesh Gurudasani]

Add the polynomial to the crc64 table generation, and provide a
generic library routine implementing the algorithm.

64-bit cyclic redundancy checks (CRC) according to the ISO 3309 standard.

The ISO 3309 64-bit CRC model parameters are as follows:
    Generator Polynomial: x^64 + x^4 + x^3 + x + 1
    Polynomial Value: 0x000000000000001B
    Initial value: 0x0000000000000000
    Reflected Input: False
    Reflected Output: False
    Xor Final: 0x0000000000000000

Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
---
 include/linux/crc64.h |  1 +
 lib/crc64.c           | 27 +++++++++++++++++++++++++++
 lib/gen_crc64table.c  |  6 ++++++
 3 files changed, 34 insertions(+)

diff --git a/include/linux/crc64.h b/include/linux/crc64.h
index e044c60d1e61..70202da51c2c 100644
--- a/include/linux/crc64.h
+++ b/include/linux/crc64.h
@@ -10,6 +10,7 @@
 #define CRC64_ROCKSOFT_STRING "crc64-rocksoft"
 
 u64 __pure crc64_be(u64 crc, const void *p, size_t len);
+u64 __pure crc64_iso_generic(u64 crc, const void *p, size_t len);
 u64 __pure crc64_rocksoft_generic(u64 crc, const void *p, size_t len);
 
 u64 crc64_rocksoft(const unsigned char *buffer, size_t len);
diff --git a/lib/crc64.c b/lib/crc64.c
index 61ae8dfb6a1c..e066c8b2a511 100644
--- a/lib/crc64.c
+++ b/lib/crc64.c
@@ -22,6 +22,11 @@
  * x^24 + x^23 + x^22 + x^21 + x^19 + x^17 + x^13 + x^12 + x^10 + x^9 +
  * x^7 + x^4 + x + 1
  *
+ * crc64iso[256] table is from the ISO specification polynomial
+ * defined as,
+ *
+ * x^64 + x^4 + x^3 + x + 1
+ *
  * crc64rocksoft[256] table is from the Rocksoft specification polynomial
  * defined as,
  *
@@ -63,6 +68,28 @@ u64 __pure crc64_be(u64 crc, const void *p, size_t len)
 }
 EXPORT_SYMBOL_GPL(crc64_be);
 
+/**
+ * crc64_iso_generic - Calculate bitwise ISO CRC64
+ * @crc: seed value for computation. 0 for a new CRC calculation, or the
+ *	 previous crc64 value if computing incrementally.
+ * @p: pointer to buffer over which CRC64 is run
+ * @len: length of buffer @p
+ */
+u64 __pure crc64_iso_generic(u64 crc, const void *p, size_t len)
+{
+	size_t i, t;
+
+	const unsigned char *_p = p;
+
+	for (i = 0; i < len; i++) {
+		t = ((crc >> 56) ^ (*_p++)) & 0xFF;
+		crc = crc64isotable[t] ^ (crc << 8);
+	}
+
+	return crc;
+}
+EXPORT_SYMBOL_GPL(crc64_iso_generic);
+
 /**
  * crc64_rocksoft_generic - Calculate bitwise Rocksoft CRC64
  * @crc: seed value for computation. 0 for a new CRC calculation, or the
diff --git a/lib/gen_crc64table.c b/lib/gen_crc64table.c
index 55e222acd0b8..ec2add75ec8b 100644
--- a/lib/gen_crc64table.c
+++ b/lib/gen_crc64table.c
@@ -17,9 +17,11 @@
 #include <stdio.h>
 
 #define CRC64_ECMA182_POLY 0x42F0E1EBA9EA3693ULL
+#define CRC64_ISO_POLY 0x000000000000001BULL
 #define CRC64_ROCKSOFT_POLY 0x9A6C9329AC4BC9B5ULL
 
 static uint64_t crc64_table[256] = {0};
+static uint64_t crc64_iso_table[256] = {0};
 static uint64_t crc64_rocksoft_table[256] = {0};
 
 static void generate_reflected_crc64_table(uint64_t table[256], uint64_t poly)
@@ -82,6 +84,9 @@ static void print_crc64_tables(void)
 	printf("static const u64 ____cacheline_aligned crc64table[256] = {\n");
 	output_table(crc64_table);
 
+	printf("\nstatic const u64 ____cacheline_aligned crc64isotable[256] = {\n");
+	output_table(crc64_iso_table);
+
 	printf("\nstatic const u64 ____cacheline_aligned crc64rocksofttable[256] = {\n");
 	output_table(crc64_rocksoft_table);
 }
@@ -89,6 +94,7 @@ static void print_crc64_tables(void)
 int main(int argc, char *argv[])
 {
 	generate_crc64_table(crc64_table, CRC64_ECMA182_POLY);
+	generate_crc64_table(crc64_iso_table, CRC64_ISO_POLY);
 	generate_reflected_crc64_table(crc64_rocksoft_table, CRC64_ROCKSOFT_POLY);
 	print_crc64_tables();
 	return 0;

-- 
2.34.1

[PATCH v2 2/6] crypto: crc64 - add crc64-iso framework

[Kamlesh Gurudasani]

Hardware specific features may be able to calculate a crc64-iso, so provide
a framework for drivers to register their implementation. If nothing is
registered, fallback to the generic table lookup implementation. The
implementation is modeled after the crc64-rocksoft equivalent

Add testmgr, tcrypt tests and vectors for 64-bit cyclic
redundancy checks (CRC) according to the ISO 3309 standard.

The ISO 3309 64-bit CRC model parameters are as follows:
    Generator Polynomial: x^64 + x^4 + x^3 + x + 1
    Polynomial Value: 0x000000000000001B
    Initial value: 0x0000000000000000
    Reflected Input: False
    Reflected Output: False
    Xor Final: 0x0000000000000000

Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
---
 crypto/Kconfig             |  11 +++
 crypto/Makefile            |   1 +
 crypto/crc64_iso_generic.c | 119 ++++++++++++++++++++++++++++
 crypto/tcrypt.c            |   5 ++
 crypto/testmgr.c           |   7 ++
 crypto/testmgr.h           | 404 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 include/linux/crc64.h      |   4 +
 lib/crc64-iso.c            | 126 ++++++++++++++++++++++++++++++
 8 files changed, 677 insertions(+)

diff --git a/crypto/Kconfig b/crypto/Kconfig
index 650b1b3620d8..6a6e69aa25da 100644
--- a/crypto/Kconfig
+++ b/crypto/Kconfig
@@ -1146,6 +1146,17 @@ config CRYPTO_CRCT10DIF
 
 	  CRC algorithm used by the SCSI Block Commands standard.
 
+config CRYPTO_CRC64_ISO
+	tristate "CRC64 based on ISO 3309 Model algorithm"
+	depends on CRC64
+	select CRYPTO_HASH
+	help
+	  CRC64 CRC algorithm based on the ISO 3309 Model CRC Algorithm
+
+	  Generator polynomial: x^64 + x^4 + x^3 + x + 1
+	  Polynomial value: 0x000000000000001b
+	  See https://en.wikipedia.org/wiki/Cyclic_redundancy_check
+
 config CRYPTO_CRC64_ROCKSOFT
 	tristate "CRC64 based on Rocksoft Model algorithm"
 	depends on CRC64
diff --git a/crypto/Makefile b/crypto/Makefile
index 953a7e105e58..c087553818f6 100644
--- a/crypto/Makefile
+++ b/crypto/Makefile
@@ -157,6 +157,7 @@ obj-$(CONFIG_CRYPTO_MICHAEL_MIC) += michael_mic.o
 obj-$(CONFIG_CRYPTO_CRC32C) += crc32c_generic.o
 obj-$(CONFIG_CRYPTO_CRC32) += crc32_generic.o
 obj-$(CONFIG_CRYPTO_CRCT10DIF) += crct10dif_common.o crct10dif_generic.o
+obj-$(CONFIG_CRYPTO_CRC64_ISO) += crc64_iso_generic.o
 obj-$(CONFIG_CRYPTO_CRC64_ROCKSOFT) += crc64_rocksoft_generic.o
 obj-$(CONFIG_CRYPTO_AUTHENC) += authenc.o authencesn.o
 obj-$(CONFIG_CRYPTO_LZO) += lzo.o lzo-rle.o
diff --git a/crypto/crc64_iso_generic.c b/crypto/crc64_iso_generic.c
new file mode 100644
index 000000000000..9fea7a3eb650
--- /dev/null
+++ b/crypto/crc64_iso_generic.c
@@ -0,0 +1,119 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#include <linux/crc64.h>
+#include <linux/module.h>
+#include <crypto/internal/hash.h>
+#include <asm/unaligned.h>
+
+static int chksum_cra_init(struct crypto_tfm *tfm)
+{
+	u64 *key = crypto_tfm_ctx(tfm);
+
+	*key = 0;
+	return 0;
+}
+
+static int chksum_init(struct shash_desc *desc)
+{
+	u64 *key = crypto_shash_ctx(desc->tfm);
+	u64 *crc = shash_desc_ctx(desc);
+
+	*crc = *key;
+	return 0;
+}
+
+static int chksum_update(struct shash_desc *desc, const u8 *data,
+			 unsigned int length)
+{
+	u64 *crc = shash_desc_ctx(desc);
+
+	*crc = crc64_iso_generic(*crc, data, length);
+	return 0;
+}
+
+static int chksum_final(struct shash_desc *desc, u8 *out)
+{
+	u64 *crc = shash_desc_ctx(desc);
+
+	put_unaligned_le64(*crc, out);
+	return 0;
+}
+
+static int __chksum_finup(u64 crc, const u8 *data, unsigned int len, u8 *out)
+{
+	crc = crc64_iso_generic(crc, data, len);
+
+	put_unaligned_le64(crc, out);
+	return 0;
+}
+
+static int chksum_finup(struct shash_desc *desc, const u8 *data,
+			unsigned int len, u8 *out)
+{
+	u64 *crc = shash_desc_ctx(desc);
+
+	return __chksum_finup(*crc, data, len, out);
+}
+
+static int chksum_digest(struct shash_desc *desc, const u8 *data,
+			 unsigned int length, u8 *out)
+{
+	u64 *key = crypto_shash_ctx(desc->tfm);
+
+	return __chksum_finup(*key, data, length, out);
+}
+
+/*
+ * Setting the seed allows arbitrary accumulators and flexible XOR policy
+ */
+static int chksum_setkey(struct crypto_shash *tfm, const u8 *key,
+			 unsigned int keylen)
+{
+	u64 *mctx = crypto_shash_ctx(tfm);
+
+	if (keylen != sizeof(u64))
+		return -EINVAL;
+
+	*mctx = get_unaligned_le64(key);
+	return 0;
+}
+
+static struct shash_alg alg = {
+	.digestsize	=	sizeof(u64),
+	.setkey		=	chksum_setkey,
+	.init		=	chksum_init,
+	.update		=	chksum_update,
+	.final		=	chksum_final,
+	.finup		=	chksum_finup,
+	.digest		=	chksum_digest,
+	.descsize	=	sizeof(u64),
+	.base		=	{
+		.cra_name		=	CRC64_ISO_STRING,
+		.cra_driver_name	=	"crc64-iso-generic",
+		.cra_priority		=	200,
+		.cra_flags		=       CRYPTO_ALG_OPTIONAL_KEY,
+		.cra_blocksize		=	1,
+		.cra_ctxsize		=       sizeof(u64),
+		.cra_module		=	THIS_MODULE,
+		.cra_init		=       chksum_cra_init,
+	}
+};
+
+static int __init crc64_iso_init(void)
+{
+	return crypto_register_shash(&alg);
+}
+
+static void __exit crc64_iso_exit(void)
+{
+	crypto_unregister_shash(&alg);
+}
+
+module_init(crc64_iso_init);
+module_exit(crc64_iso_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Kamlesh Gurudasani <kamlesh@ti.com>");
+MODULE_DESCRIPTION("ISO model CRC64 calculation");
+MODULE_ALIAS_CRYPTO("crc64-iso");
+MODULE_ALIAS_CRYPTO("crc64-iso-generic");
diff --git a/crypto/tcrypt.c b/crypto/tcrypt.c
index 202ca1a3105d..a41d52a879d6 100644
--- a/crypto/tcrypt.c
+++ b/crypto/tcrypt.c
@@ -2327,6 +2327,11 @@ static int do_test(const char *alg, u32 type, u32 mask, int m, u32 num_mb)
 				generic_hash_speed_template);
 		if (mode > 300 && mode < 400) break;
 		fallthrough;
+	case 329:
+		test_hash_speed("crc64-iso", sec, generic_hash_speed_template);
+		if (mode > 300 && mode < 400)
+			break;
+		fallthrough;
 	case 399:
 		break;
 
diff --git a/crypto/testmgr.c b/crypto/testmgr.c
index 216878c8bc3d..bd0414b99287 100644
--- a/crypto/testmgr.c
+++ b/crypto/testmgr.c
@@ -4679,6 +4679,13 @@ static const struct alg_test_desc alg_test_descs[] = {
 		.suite = {
 			.hash = __VECS(crc32c_tv_template)
 		}
+	}, {
+		.alg = "crc64-iso",
+		.test = alg_test_hash,
+		.fips_allowed = 1,
+		.suite = {
+			.hash = __VECS(crc64_iso_tv_template)
+		}
 	}, {
 		.alg = "crc64-rocksoft",
 		.test = alg_test_hash,
diff --git a/crypto/testmgr.h b/crypto/testmgr.h
index 5ca7a412508f..5225abd96818 100644
--- a/crypto/testmgr.h
+++ b/crypto/testmgr.h
@@ -5209,6 +5209,410 @@ static const struct hash_testvec rmd160_tv_template[] = {
 	}
 };
 
+static const struct hash_testvec crc64_iso_tv_template[] = {
+	{
+		.psize = 0,
+		.digest = "\x00\x00\x00\x00\x00\x00\x00\x00",
+	},
+	{
+		.plaintext = "\x01\x02\x03\x04\x05\x06\x07\x08"
+			     "\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
+			     "\x11\x12\x13\x14\x15\x16\x17\x18"
+			     "\x19\x1a\x1b\x1c\x1d\x1e\x1f\x20"
+			     "\x21\x22\x23\x24\x25\x26\x27\x28",
+		.psize = 40,
+		.digest = "\xaf\x45\xba\x7d\xf2\xda\xa0\xaa",
+	},
+	{
+		.plaintext = "\x29\x2a\x2b\x2c\x2d\x2e\x2f\x30"
+			     "\x31\x32\x33\x34\x35\x36\x37\x38"
+			     "\x39\x3a\x3b\x3c\x3d\x3e\x3f\x40"
+			     "\x41\x42\x43\x44\x45\x46\x47\x48"
+			     "\x49\x4a\x4b\x4c\x4d\x4e\x4f\x50",
+		.psize = 40,
+		.digest = "\x81\x55\x2e\x76\xf8\xd0\xaa\xa0",
+	},
+	{
+		.plaintext = "\x51\x52\x53\x54\x55\x56\x57\x58"
+			     "\x59\x5a\x5b\x5c\x5d\x5e\x5f\x60"
+			     "\x61\x62\x63\x64\x65\x66\x67\x68"
+			     "\x69\x6a\x6b\x6c\x6d\x6e\x6f\x70"
+			     "\x71\x72\x73\x74\x75\x76\x77\x78",
+		.psize = 40,
+		.digest = "\xc6\xb6\x26\x82\x0d\x25\x5f\x55",
+	},
+	{
+		.plaintext = "\x79\x7a\x7b\x7c\x7d\x7e\x7f\x80"
+			     "\x81\x82\x83\x84\x85\x86\x87\x88"
+			     "\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90"
+			     "\x91\x92\x93\x94\x95\x96\x97\x98"
+			     "\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0",
+		.psize = 40,
+		.digest = "\x20\x8a\xe6\x59\xdf\xf7\x8d\x87",
+	},
+	{
+		.plaintext = "\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8"
+			     "\xa9\xaa\xab\xac\xad\xae\xaf\xb0"
+			     "\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8"
+			     "\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0"
+			     "\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8",
+		.psize = 40,
+		.digest = "\x19\x9e\xba\xff\x70\x58\x22\x28",
+
+	},
+	{
+		.plaintext = "\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0"
+			     "\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8"
+			     "\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0"
+			     "\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8"
+			     "\xe9\xea\xeb\xec\xed\xee\xef\xf0",
+		.psize = 40,
+		.digest = "\xa3\xdc\x11\x98\x16\x3e\x44\x4e",
+	},
+	{
+		.plaintext = "\x29\x2a\x2b\x2c\x2d\x2e\x2f\x30"
+			     "\x31\x32\x33\x34\x35\x36\x37\x38"
+			     "\x39\x3a\x3b\x3c\x3d\x3e\x3f\x40"
+			     "\x41\x42\x43\x44\x45\x46\x47\x48"
+			     "\x49\x4a\x4b\x4c\x4d\x4e\x4f\x50",
+		.psize = 40,
+		.digest = "\x81\x55\x2e\x76\xf8\xd0\xaa\xa0",
+	},
+	{
+		.plaintext = "\x51\x52\x53\x54\x55\x56\x57\x58"
+			     "\x59\x5a\x5b\x5c\x5d\x5e\x5f\x60"
+			     "\x61\x62\x63\x64\x65\x66\x67\x68"
+			     "\x69\x6a\x6b\x6c\x6d\x6e\x6f\x70"
+			     "\x71\x72\x73\x74\x75\x76\x77\x78",
+		.psize = 40,
+		.digest = "\xc6\xb6\x26\x82\x0d\x25\x5f\x55",
+	},
+	{
+		.plaintext = "\x79\x7a\x7b\x7c\x7d\x7e\x7f\x80"
+			     "\x81\x82\x83\x84\x85\x86\x87\x88"
+			     "\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90"
+			     "\x91\x92\x93\x94\x95\x96\x97\x98"
+			     "\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0",
+		.psize = 40,
+		.digest = "\x20\x8a\xe6\x59\xdf\xf7\x8d\x87",
+	},
+	{
+		.plaintext = "\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8"
+			     "\xa9\xaa\xab\xac\xad\xae\xaf\xb0"
+			     "\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8"
+			     "\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0"
+			     "\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8",
+		.psize = 40,
+		.digest = "\x19\x9e\xba\xff\x70\x58\x22\x28",
+	},
+	{
+		.plaintext = "\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0"
+			     "\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8"
+			     "\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0"
+			     "\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8"
+			     "\xe9\xea\xeb\xec\xed\xee\xef\xf0",
+		.psize = 40,
+		.digest = "\xa3\xdc\x11\x98\x16\x3e\x44\x4e",
+	},
+	{
+		.key = "\xff\xff\xff\xff\xff\xff\xff\xff",
+		.ksize = 8,
+		.plaintext = "\x01\x02\x03\x04\x05\x06\x07\x08"
+			     "\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
+			     "\x11\x12\x13\x14\x15\x16\x17\x18"
+			     "\x19\x1a\x1b\x1c\x1d\x1e\x1f\x20"
+			     "\x21\x22\x23\x24\x25\x26\x27\x28"
+			     "\x29\x2a\x2b\x2c\x2d\x2e\x2f\x30"
+			     "\x31\x32\x33\x34\x35\x36\x37\x38"
+			     "\x39\x3a\x3b\x3c\x3d\x3e\x3f\x40"
+			     "\x41\x42\x43\x44\x45\x46\x47\x48"
+			     "\x49\x4a\x4b\x4c\x4d\x4e\x4f\x50"
+			     "\x51\x52\x53\x54\x55\x56\x57\x58"
+			     "\x59\x5a\x5b\x5c\x5d\x5e\x5f\x60"
+			     "\x61\x62\x63\x64\x65\x66\x67\x68"
+			     "\x69\x6a\x6b\x6c\x6d\x6e\x6f\x70"
+			     "\x71\x72\x73\x74\x75\x76\x77\x78"
+			     "\x79\x7a\x7b\x7c\x7d\x7e\x7f\x80"
+			     "\x81\x82\x83\x84\x85\x86\x87\x88"
+			     "\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90"
+			     "\x91\x92\x93\x94\x95\x96\x97\x98"
+			     "\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0"
+			     "\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8"
+			     "\xa9\xaa\xab\xac\xad\xae\xaf\xb0"
+			     "\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8"
+			     "\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0"
+			     "\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8"
+			     "\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0"
+			     "\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8"
+			     "\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0"
+			     "\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8"
+			     "\xe9\xea\xeb\xec\xed\xee\xef\xf0",
+		.psize = 240,
+		.digest = "\x8b\xa6\xd7\x91\xb4\x74\x96\x84",
+	}, {
+		.key = "\xff\xff\xff\xff\xff\xff\xff\xff",
+		.ksize = 8,
+		.plaintext = "\x6e\x05\x79\x10\xa7\x1b\xb2\x49"
+			     "\xe0\x54\xeb\x82\x19\x8d\x24\xbb"
+			     "\x2f\xc6\x5d\xf4\x68\xff\x96\x0a"
+			     "\xa1\x38\xcf\x43\xda\x71\x08\x7c"
+			     "\x13\xaa\x1e\xb5\x4c\xe3\x57\xee"
+			     "\x85\x1c\x90\x27\xbe\x32\xc9\x60"
+			     "\xf7\x6b\x02\x99\x0d\xa4\x3b\xd2"
+			     "\x46\xdd\x74\x0b\x7f\x16\xad\x21"
+			     "\xb8\x4f\xe6\x5a\xf1\x88\x1f\x93"
+			     "\x2a\xc1\x35\xcc\x63\xfa\x6e\x05"
+			     "\x9c\x10\xa7\x3e\xd5\x49\xe0\x77"
+			     "\x0e\x82\x19\xb0\x24\xbb\x52\xe9"
+			     "\x5d\xf4\x8b\x22\x96\x2d\xc4\x38"
+			     "\xcf\x66\xfd\x71\x08\x9f\x13\xaa"
+			     "\x41\xd8\x4c\xe3\x7a\x11\x85\x1c"
+			     "\xb3\x27\xbe\x55\xec\x60\xf7\x8e"
+			     "\x02\x99\x30\xc7\x3b\xd2\x69\x00"
+			     "\x74\x0b\xa2\x16\xad\x44\xdb\x4f"
+			     "\xe6\x7d\x14\x88\x1f\xb6\x2a\xc1"
+			     "\x58\xef\x63\xfa\x91\x05\x9c\x33"
+			     "\xca\x3e\xd5\x6c\x03\x77\x0e\xa5"
+			     "\x19\xb0\x47\xde\x52\xe9\x80\x17"
+			     "\x8b\x22\xb9\x2d\xc4\x5b\xf2\x66"
+			     "\xfd\x94\x08\x9f\x36\xcd\x41\xd8"
+			     "\x6f\x06\x7a\x11\xa8\x1c\xb3\x4a"
+			     "\xe1\x55\xec\x83\x1a\x8e\x25\xbc"
+			     "\x30\xc7\x5e\xf5\x69\x00\x97\x0b"
+			     "\xa2\x39\xd0\x44\xdb\x72\x09\x7d"
+			     "\x14\xab\x1f\xb6\x4d\xe4\x58\xef"
+			     "\x86\x1d\x91\x28\xbf\x33\xca\x61"
+			     "\xf8\x6c\x03\x9a\x0e\xa5\x3c\xd3"
+			     "\x47\xde\x75\x0c\x80\x17\xae\x22"
+			     "\xb9\x50\xe7\x5b\xf2\x89\x20\x94"
+			     "\x2b\xc2\x36\xcd\x64\xfb\x6f\x06"
+			     "\x9d\x11\xa8\x3f\xd6\x4a\xe1\x78"
+			     "\x0f\x83\x1a\xb1\x25\xbc\x53\xea"
+			     "\x5e\xf5\x8c\x00\x97\x2e\xc5\x39"
+			     "\xd0\x67\xfe\x72\x09\xa0\x14\xab"
+			     "\x42\xd9\x4d\xe4\x7b\x12\x86\x1d"
+			     "\xb4\x28\xbf\x56\xed\x61\xf8\x8f"
+			     "\x03\x9a\x31\xc8\x3c\xd3\x6a\x01"
+			     "\x75\x0c\xa3\x17\xae\x45\xdc\x50"
+			     "\xe7\x7e\x15\x89\x20\xb7\x2b\xc2"
+			     "\x59\xf0\x64\xfb\x92\x06\x9d\x34"
+			     "\xcb\x3f\xd6\x6d\x04\x78\x0f\xa6"
+			     "\x1a\xb1\x48\xdf\x53\xea\x81\x18"
+			     "\x8c\x23\xba\x2e\xc5\x5c\xf3\x67"
+			     "\xfe\x95\x09\xa0\x37\xce\x42\xd9"
+			     "\x70\x07\x7b\x12\xa9\x1d\xb4\x4b"
+			     "\xe2\x56\xed\x84\x1b\x8f\x26\xbd"
+			     "\x31\xc8\x5f\xf6\x6a\x01\x98\x0c"
+			     "\xa3\x3a\xd1\x45\xdc\x73\x0a\x7e"
+			     "\x15\xac\x20\xb7\x4e\xe5\x59\xf0"
+			     "\x87\x1e\x92\x29\xc0\x34\xcb\x62"
+			     "\xf9\x6d\x04\x9b\x0f\xa6\x3d\xd4"
+			     "\x48\xdf\x76\x0d\x81\x18\xaf\x23"
+			     "\xba\x51\xe8\x5c\xf3\x8a\x21\x95"
+			     "\x2c\xc3\x37\xce\x65\xfc\x70\x07"
+			     "\x9e\x12\xa9\x40\xd7\x4b\xe2\x79"
+			     "\x10\x84\x1b\xb2\x26\xbd\x54\xeb"
+			     "\x5f\xf6\x8d\x01\x98\x2f\xc6\x3a"
+			     "\xd1\x68\xff\x73\x0a\xa1\x15\xac"
+			     "\x43\xda\x4e\xe5\x7c\x13\x87\x1e"
+			     "\xb5\x29\xc0\x57\xee\x62\xf9\x90"
+			     "\x04\x9b\x32\xc9\x3d\xd4\x6b\x02"
+			     "\x76\x0d\xa4\x18\xaf\x46\xdd\x51"
+			     "\xe8\x7f\x16\x8a\x21\xb8\x2c\xc3"
+			     "\x5a\xf1\x65\xfc\x93\x07\x9e\x35"
+			     "\xcc\x40\xd7\x6e\x05\x79\x10\xa7"
+			     "\x1b\xb2\x49\xe0\x54\xeb\x82\x19"
+			     "\x8d\x24\xbb\x2f\xc6\x5d\xf4\x68"
+			     "\xff\x96\x0a\xa1\x38\xcf\x43\xda"
+			     "\x71\x08\x7c\x13\xaa\x1e\xb5\x4c"
+			     "\xe3\x57\xee\x85\x1c\x90\x27\xbe"
+			     "\x32\xc9\x60\xf7\x6b\x02\x99\x0d"
+			     "\xa4\x3b\xd2\x46\xdd\x74\x0b\x7f"
+			     "\x16\xad\x21\xb8\x4f\xe6\x5a\xf1"
+			     "\x88\x1f\x93\x2a\xc1\x35\xcc\x63"
+			     "\xfa\x6e\x05\x9c\x10\xa7\x3e\xd5"
+			     "\x49\xe0\x77\x0e\x82\x19\xb0\x24"
+			     "\xbb\x52\xe9\x5d\xf4\x8b\x22\x96"
+			     "\x2d\xc4\x38\xcf\x66\xfd\x71\x08"
+			     "\x9f\x13\xaa\x41\xd8\x4c\xe3\x7a"
+			     "\x11\x85\x1c\xb3\x27\xbe\x55\xec"
+			     "\x60\xf7\x8e\x02\x99\x30\xc7\x3b"
+			     "\xd2\x69\x00\x74\x0b\xa2\x16\xad"
+			     "\x44\xdb\x4f\xe6\x7d\x14\x88\x1f"
+			     "\xb6\x2a\xc1\x58\xef\x63\xfa\x91"
+			     "\x05\x9c\x33\xca\x3e\xd5\x6c\x03"
+			     "\x77\x0e\xa5\x19\xb0\x47\xde\x52"
+			     "\xe9\x80\x17\x8b\x22\xb9\x2d\xc4"
+			     "\x5b\xf2\x66\xfd\x94\x08\x9f\x36"
+			     "\xcd\x41\xd8\x6f\x06\x7a\x11\xa8"
+			     "\x1c\xb3\x4a\xe1\x55\xec\x83\x1a"
+			     "\x8e\x25\xbc\x30\xc7\x5e\xf5\x69"
+			     "\x00\x97\x0b\xa2\x39\xd0\x44\xdb"
+			     "\x72\x09\x7d\x14\xab\x1f\xb6\x4d"
+			     "\xe4\x58\xef\x86\x1d\x91\x28\xbf"
+			     "\x33\xca\x61\xf8\x6c\x03\x9a\x0e"
+			     "\xa5\x3c\xd3\x47\xde\x75\x0c\x80"
+			     "\x17\xae\x22\xb9\x50\xe7\x5b\xf2"
+			     "\x89\x20\x94\x2b\xc2\x36\xcd\x64"
+			     "\xfb\x6f\x06\x9d\x11\xa8\x3f\xd6"
+			     "\x4a\xe1\x78\x0f\x83\x1a\xb1\x25"
+			     "\xbc\x53\xea\x5e\xf5\x8c\x00\x97"
+			     "\x2e\xc5\x39\xd0\x67\xfe\x72\x09"
+			     "\xa0\x14\xab\x42\xd9\x4d\xe4\x7b"
+			     "\x12\x86\x1d\xb4\x28\xbf\x56\xed"
+			     "\x61\xf8\x8f\x03\x9a\x31\xc8\x3c"
+			     "\xd3\x6a\x01\x75\x0c\xa3\x17\xae"
+			     "\x45\xdc\x50\xe7\x7e\x15\x89\x20"
+			     "\xb7\x2b\xc2\x59\xf0\x64\xfb\x92"
+			     "\x06\x9d\x34\xcb\x3f\xd6\x6d\x04"
+			     "\x78\x0f\xa6\x1a\xb1\x48\xdf\x53"
+			     "\xea\x81\x18\x8c\x23\xba\x2e\xc5"
+			     "\x5c\xf3\x67\xfe\x95\x09\xa0\x37"
+			     "\xce\x42\xd9\x70\x07\x7b\x12\xa9"
+			     "\x1d\xb4\x4b\xe2\x56\xed\x84\x1b"
+			     "\x8f\x26\xbd\x31\xc8\x5f\xf6\x6a"
+			     "\x01\x98\x0c\xa3\x3a\xd1\x45\xdc"
+			     "\x73\x0a\x7e\x15\xac\x20\xb7\x4e"
+			     "\xe5\x59\xf0\x87\x1e\x92\x29\xc0"
+			     "\x34\xcb\x62\xf9\x6d\x04\x9b\x0f"
+			     "\xa6\x3d\xd4\x48\xdf\x76\x0d\x81"
+			     "\x18\xaf\x23\xba\x51\xe8\x5c\xf3"
+			     "\x8a\x21\x95\x2c\xc3\x37\xce\x65"
+			     "\xfc\x70\x07\x9e\x12\xa9\x40\xd7"
+			     "\x4b\xe2\x79\x10\x84\x1b\xb2\x26"
+			     "\xbd\x54\xeb\x5f\xf6\x8d\x01\x98"
+			     "\x2f\xc6\x3a\xd1\x68\xff\x73\x0a"
+			     "\xa1\x15\xac\x43\xda\x4e\xe5\x7c"
+			     "\x13\x87\x1e\xb5\x29\xc0\x57\xee"
+			     "\x62\xf9\x90\x04\x9b\x32\xc9\x3d"
+			     "\xd4\x6b\x02\x76\x0d\xa4\x18\xaf"
+			     "\x46\xdd\x51\xe8\x7f\x16\x8a\x21"
+			     "\xb8\x2c\xc3\x5a\xf1\x65\xfc\x93"
+			     "\x07\x9e\x35\xcc\x40\xd7\x6e\x05"
+			     "\x79\x10\xa7\x1b\xb2\x49\xe0\x54"
+			     "\xeb\x82\x19\x8d\x24\xbb\x2f\xc6"
+			     "\x5d\xf4\x68\xff\x96\x0a\xa1\x38"
+			     "\xcf\x43\xda\x71\x08\x7c\x13\xaa"
+			     "\x1e\xb5\x4c\xe3\x57\xee\x85\x1c"
+			     "\x90\x27\xbe\x32\xc9\x60\xf7\x6b"
+			     "\x02\x99\x0d\xa4\x3b\xd2\x46\xdd"
+			     "\x74\x0b\x7f\x16\xad\x21\xb8\x4f"
+			     "\xe6\x5a\xf1\x88\x1f\x93\x2a\xc1"
+			     "\x35\xcc\x63\xfa\x6e\x05\x9c\x10"
+			     "\xa7\x3e\xd5\x49\xe0\x77\x0e\x82"
+			     "\x19\xb0\x24\xbb\x52\xe9\x5d\xf4"
+			     "\x8b\x22\x96\x2d\xc4\x38\xcf\x66"
+			     "\xfd\x71\x08\x9f\x13\xaa\x41\xd8"
+			     "\x4c\xe3\x7a\x11\x85\x1c\xb3\x27"
+			     "\xbe\x55\xec\x60\xf7\x8e\x02\x99"
+			     "\x30\xc7\x3b\xd2\x69\x00\x74\x0b"
+			     "\xa2\x16\xad\x44\xdb\x4f\xe6\x7d"
+			     "\x14\x88\x1f\xb6\x2a\xc1\x58\xef"
+			     "\x63\xfa\x91\x05\x9c\x33\xca\x3e"
+			     "\xd5\x6c\x03\x77\x0e\xa5\x19\xb0"
+			     "\x47\xde\x52\xe9\x80\x17\x8b\x22"
+			     "\xb9\x2d\xc4\x5b\xf2\x66\xfd\x94"
+			     "\x08\x9f\x36\xcd\x41\xd8\x6f\x06"
+			     "\x7a\x11\xa8\x1c\xb3\x4a\xe1\x55"
+			     "\xec\x83\x1a\x8e\x25\xbc\x30\xc7"
+			     "\x5e\xf5\x69\x00\x97\x0b\xa2\x39"
+			     "\xd0\x44\xdb\x72\x09\x7d\x14\xab"
+			     "\x1f\xb6\x4d\xe4\x58\xef\x86\x1d"
+			     "\x91\x28\xbf\x33\xca\x61\xf8\x6c"
+			     "\x03\x9a\x0e\xa5\x3c\xd3\x47\xde"
+			     "\x75\x0c\x80\x17\xae\x22\xb9\x50"
+			     "\xe7\x5b\xf2\x89\x20\x94\x2b\xc2"
+			     "\x36\xcd\x64\xfb\x6f\x06\x9d\x11"
+			     "\xa8\x3f\xd6\x4a\xe1\x78\x0f\x83"
+			     "\x1a\xb1\x25\xbc\x53\xea\x5e\xf5"
+			     "\x8c\x00\x97\x2e\xc5\x39\xd0\x67"
+			     "\xfe\x72\x09\xa0\x14\xab\x42\xd9"
+			     "\x4d\xe4\x7b\x12\x86\x1d\xb4\x28"
+			     "\xbf\x56\xed\x61\xf8\x8f\x03\x9a"
+			     "\x31\xc8\x3c\xd3\x6a\x01\x75\x0c"
+			     "\xa3\x17\xae\x45\xdc\x50\xe7\x7e"
+			     "\x15\x89\x20\xb7\x2b\xc2\x59\xf0"
+			     "\x64\xfb\x92\x06\x9d\x34\xcb\x3f"
+			     "\xd6\x6d\x04\x78\x0f\xa6\x1a\xb1"
+			     "\x48\xdf\x53\xea\x81\x18\x8c\x23"
+			     "\xba\x2e\xc5\x5c\xf3\x67\xfe\x95"
+			     "\x09\xa0\x37\xce\x42\xd9\x70\x07"
+			     "\x7b\x12\xa9\x1d\xb4\x4b\xe2\x56"
+			     "\xed\x84\x1b\x8f\x26\xbd\x31\xc8"
+			     "\x5f\xf6\x6a\x01\x98\x0c\xa3\x3a"
+			     "\xd1\x45\xdc\x73\x0a\x7e\x15\xac"
+			     "\x20\xb7\x4e\xe5\x59\xf0\x87\x1e"
+			     "\x92\x29\xc0\x34\xcb\x62\xf9\x6d"
+			     "\x04\x9b\x0f\xa6\x3d\xd4\x48\xdf"
+			     "\x76\x0d\x81\x18\xaf\x23\xba\x51"
+			     "\xe8\x5c\xf3\x8a\x21\x95\x2c\xc3"
+			     "\x37\xce\x65\xfc\x70\x07\x9e\x12"
+			     "\xa9\x40\xd7\x4b\xe2\x79\x10\x84"
+			     "\x1b\xb2\x26\xbd\x54\xeb\x5f\xf6"
+			     "\x8d\x01\x98\x2f\xc6\x3a\xd1\x68"
+			     "\xff\x73\x0a\xa1\x15\xac\x43\xda"
+			     "\x4e\xe5\x7c\x13\x87\x1e\xb5\x29"
+			     "\xc0\x57\xee\x62\xf9\x90\x04\x9b"
+			     "\x32\xc9\x3d\xd4\x6b\x02\x76\x0d"
+			     "\xa4\x18\xaf\x46\xdd\x51\xe8\x7f"
+			     "\x16\x8a\x21\xb8\x2c\xc3\x5a\xf1"
+			     "\x65\xfc\x93\x07\x9e\x35\xcc\x40"
+			     "\xd7\x6e\x05\x79\x10\xa7\x1b\xb2"
+			     "\x49\xe0\x54\xeb\x82\x19\x8d\x24"
+			     "\xbb\x2f\xc6\x5d\xf4\x68\xff\x96"
+			     "\x0a\xa1\x38\xcf\x43\xda\x71\x08"
+			     "\x7c\x13\xaa\x1e\xb5\x4c\xe3\x57"
+			     "\xee\x85\x1c\x90\x27\xbe\x32\xc9"
+			     "\x60\xf7\x6b\x02\x99\x0d\xa4\x3b"
+			     "\xd2\x46\xdd\x74\x0b\x7f\x16\xad"
+			     "\x21\xb8\x4f\xe6\x5a\xf1\x88\x1f"
+			     "\x93\x2a\xc1\x35\xcc\x63\xfa\x6e"
+			     "\x05\x9c\x10\xa7\x3e\xd5\x49\xe0"
+			     "\x77\x0e\x82\x19\xb0\x24\xbb\x52"
+			     "\xe9\x5d\xf4\x8b\x22\x96\x2d\xc4"
+			     "\x38\xcf\x66\xfd\x71\x08\x9f\x13"
+			     "\xaa\x41\xd8\x4c\xe3\x7a\x11\x85"
+			     "\x1c\xb3\x27\xbe\x55\xec\x60\xf7"
+			     "\x8e\x02\x99\x30\xc7\x3b\xd2\x69"
+			     "\x00\x74\x0b\xa2\x16\xad\x44\xdb"
+			     "\x4f\xe6\x7d\x14\x88\x1f\xb6\x2a"
+			     "\xc1\x58\xef\x63\xfa\x91\x05\x9c"
+			     "\x33\xca\x3e\xd5\x6c\x03\x77\x0e"
+			     "\xa5\x19\xb0\x47\xde\x52\xe9\x80"
+			     "\x17\x8b\x22\xb9\x2d\xc4\x5b\xf2"
+			     "\x66\xfd\x94\x08\x9f\x36\xcd\x41"
+			     "\xd8\x6f\x06\x7a\x11\xa8\x1c\xb3"
+			     "\x4a\xe1\x55\xec\x83\x1a\x8e\x25"
+			     "\xbc\x30\xc7\x5e\xf5\x69\x00\x97"
+			     "\x0b\xa2\x39\xd0\x44\xdb\x72\x09"
+			     "\x7d\x14\xab\x1f\xb6\x4d\xe4\x58"
+			     "\xef\x86\x1d\x91\x28\xbf\x33\xca"
+			     "\x61\xf8\x6c\x03\x9a\x0e\xa5\x3c"
+			     "\xd3\x47\xde\x75\x0c\x80\x17\xae"
+			     "\x22\xb9\x50\xe7\x5b\xf2\x89\x20"
+			     "\x94\x2b\xc2\x36\xcd\x64\xfb\x6f"
+			     "\x06\x9d\x11\xa8\x3f\xd6\x4a\xe1"
+			     "\x78\x0f\x83\x1a\xb1\x25\xbc\x53"
+			     "\xea\x5e\xf5\x8c\x00\x97\x2e\xc5"
+			     "\x39\xd0\x67\xfe\x72\x09\xa0\x14"
+			     "\xab\x42\xd9\x4d\xe4\x7b\x12\x86"
+			     "\x1d\xb4\x28\xbf\x56\xed\x61\xf8"
+			     "\x8f\x03\x9a\x31\xc8\x3c\xd3\x6a"
+			     "\x01\x75\x0c\xa3\x17\xae\x45\xdc"
+			     "\x50\xe7\x7e\x15\x89\x20\xb7\x2b"
+			     "\xc2\x59\xf0\x64\xfb\x92\x06\x9d"
+			     "\x34\xcb\x3f\xd6\x6d\x04\x78\x0f"
+			     "\xa6\x1a\xb1\x48\xdf\x53\xea\x81"
+			     "\x18\x8c\x23\xba\x2e\xc5\x5c\xf3"
+			     "\x67\xfe\x95\x09\xa0\x37\xce\x42"
+			     "\xd9\x70\x07\x7b\x12\xa9\x1d\xb4"
+			     "\x4b\xe2\x56\xed\x84\x1b\x8f\x26"
+			     "\xbd\x31\xc8\x5f\xf6\x6a\x01\x98",
+		.psize = 2048,
+		.digest = "\x4b\x82\xa5\x0e\x72\x01\x0b\xc6",
+	}
+};
+
 static const u8 zeroes[4096] = { [0 ... 4095] = 0 };
 static const u8 ones[4096] = { [0 ... 4095] = 0xff };
 
diff --git a/include/linux/crc64.h b/include/linux/crc64.h
index 70202da51c2c..10b792080374 100644
--- a/include/linux/crc64.h
+++ b/include/linux/crc64.h
@@ -8,11 +8,15 @@
 #include <linux/types.h>
 
 #define CRC64_ROCKSOFT_STRING "crc64-rocksoft"
+#define CRC64_ISO_STRING "crc64-iso"
 
 u64 __pure crc64_be(u64 crc, const void *p, size_t len);
 u64 __pure crc64_iso_generic(u64 crc, const void *p, size_t len);
 u64 __pure crc64_rocksoft_generic(u64 crc, const void *p, size_t len);
 
+u64 crc64_iso(const unsigned char *buffer, size_t len);
+u64 crc64_iso_update(u64 crc, const unsigned char *buffer, size_t len);
+
 u64 crc64_rocksoft(const unsigned char *buffer, size_t len);
 u64 crc64_rocksoft_update(u64 crc, const unsigned char *buffer, size_t len);
 
diff --git a/lib/crc64-iso.c b/lib/crc64-iso.c
new file mode 100644
index 000000000000..d6e803124fa0
--- /dev/null
+++ b/lib/crc64-iso.c
@@ -0,0 +1,126 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/crc64.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <crypto/hash.h>
+#include <crypto/algapi.h>
+#include <linux/static_key.h>
+#include <linux/notifier.h>
+
+static struct crypto_shash __rcu *crc64_iso_tfm;
+static DEFINE_STATIC_KEY_TRUE(crc64_iso_fallback);
+static DEFINE_MUTEX(crc64_iso_mutex);
+static struct work_struct crc64_iso_rehash_work;
+
+static int crc64_iso_notify(struct notifier_block *self, unsigned long val, void *data)
+{
+	struct crypto_alg *alg = data;
+
+	if (val != CRYPTO_MSG_ALG_LOADED ||
+	    strcmp(alg->cra_name, CRC64_ISO_STRING))
+		return NOTIFY_DONE;
+
+	schedule_work(&crc64_iso_rehash_work);
+	return NOTIFY_OK;
+}
+
+static void crc64_iso_rehash(struct work_struct *work)
+{
+	struct crypto_shash *new, *old;
+
+	mutex_lock(&crc64_iso_mutex);
+	old = rcu_dereference_protected(crc64_iso_tfm,
+					lockdep_is_held(&crc64_iso_mutex));
+	new = crypto_alloc_shash(CRC64_ISO_STRING, 0, 0);
+	if (IS_ERR(new)) {
+		mutex_unlock(&crc64_iso_mutex);
+		return;
+	}
+	rcu_assign_pointer(crc64_iso_tfm, new);
+	mutex_unlock(&crc64_iso_mutex);
+
+	if (old) {
+		synchronize_rcu();
+		crypto_free_shash(old);
+	} else {
+		static_branch_disable(&crc64_iso_fallback);
+	}
+}
+
+static struct notifier_block crc64_iso_nb = {
+	.notifier_call = crc64_iso_notify,
+};
+
+u64 crc64_iso_update(u64 crc, const unsigned char *buffer, size_t len)
+{
+	struct {
+		struct shash_desc shash;
+		u64 crc;
+	} desc;
+	int err;
+
+	if (static_branch_unlikely(&crc64_iso_fallback))
+		return crc64_iso_generic(crc, buffer, len);
+
+	rcu_read_lock();
+	desc.shash.tfm = rcu_dereference(crc64_iso_tfm);
+	desc.crc = crc;
+	err = crypto_shash_update(&desc.shash, buffer, len);
+	rcu_read_unlock();
+
+	WARN_ON_ONCE(err);
+
+	return desc.crc;
+}
+EXPORT_SYMBOL_GPL(crc64_iso_update);
+
+u64 crc64_iso(const unsigned char *buffer, size_t len)
+{
+	return crc64_iso_update(0, buffer, len);
+}
+EXPORT_SYMBOL_GPL(crc64_iso);
+
+static int __init crc64_iso_mod_init(void)
+{
+	INIT_WORK(&crc64_iso_rehash_work, crc64_iso_rehash);
+	crypto_register_notifier(&crc64_iso_nb);
+	crc64_iso_rehash(&crc64_iso_rehash_work);
+	return 0;
+}
+
+static void __exit crc64_iso_mod_fini(void)
+{
+	crypto_unregister_notifier(&crc64_iso_nb);
+	cancel_work_sync(&crc64_iso_rehash_work);
+	crypto_free_shash(rcu_dereference_protected(crc64_iso_tfm, 1));
+}
+
+module_init(crc64_iso_mod_init);
+module_exit(crc64_iso_mod_fini);
+
+static int crc64_iso_transform_show(char *buffer, const struct kernel_param *kp)
+{
+	struct crypto_shash *tfm;
+	int len;
+
+	if (static_branch_unlikely(&crc64_iso_fallback))
+		return sprintf(buffer, "fallback\n");
+
+	rcu_read_lock();
+	tfm = rcu_dereference(crc64_iso_tfm);
+	len = snprintf(buffer, PAGE_SIZE, "%s\n",
+		       crypto_shash_driver_name(tfm));
+	rcu_read_unlock();
+
+	return len;
+}
+
+module_param_call(transform, NULL, crc64_iso_transform_show, NULL, 0444);
+
+MODULE_AUTHOR("Kamlesh Gurudasani <kamlesh@ti.com>");
+MODULE_DESCRIPTION("ISO model CRC64 calculation (library API)");
+MODULE_LICENSE("GPL");
+MODULE_SOFTDEP("pre: crc64");

-- 
2.34.1

[PATCH v2 3/6] dt-bindings: crypto: Add Texas Instruments MCRC64

[Kamlesh Gurudasani]

Add binding for Texas Instruments MCRC64

MCRC64 engine calculates 64-bit cyclic redundancy checks (CRC)
according to the ISO 3309 standard.

The ISO 3309 64-bit CRC model parameters are as follows:
    Generator Polynomial: x^64 + x^4 + x^3 + x + 1
    Polynomial Value: 0x000000000000001B
    Initial value: 0x0000000000000000
    Reflected Input: False
    Reflected Output: False
    Xor Final: 0x0000000000000000

Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
---
 Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml | 47 +++++++++++++++++++++++++++++++++++++++++++++++
 MAINTAINERS                                             |  5 +++++
 2 files changed, 52 insertions(+)

diff --git a/Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml b/Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml
new file mode 100644
index 000000000000..38bc7efebd68
--- /dev/null
+++ b/Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml
@@ -0,0 +1,47 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/crypto/ti,mcrc64.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Texas Instruments MCRC64
+
+description: The MCRC64 engine calculates 64-bit cyclic redundancy checks
+  (CRC) according to the ISO 3309 standard.
+
+maintainers:
+  - Kamlesh Gurudasani <kamlesh@ti.com>
+
+properties:
+  compatible:
+    const: ti,am62-mcrc64
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    maxItems: 1
+
+  power-domains:
+    maxItems: 1
+
+required:
+  - compatible
+  - reg
+  - clocks
+  - power-domains
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/soc/ti,sci_pm_domain.h>
+
+    crc@30300000 {
+      compatible = "ti,am62-mcrc64";
+      reg = <0x30300000 0x1000>;
+      clocks = <&k3_clks 116 0>;
+      power-domains = <&k3_pds 116 TI_SCI_PD_EXCLUSIVE>;
+    };
+
+...
diff --git a/MAINTAINERS b/MAINTAINERS
index 02a3192195af..66b51f43d196 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -21481,6 +21481,11 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/iio/adc/ti,lmp92064.yaml
 F:	drivers/iio/adc/ti-lmp92064.c
 
+TI MEMORY CYCLIC REDUNDANCY CHECK (MCRC64) DRIVER
+M:	Kamlesh Gurudasani <kamlesh@ti.com>
+S:	Maintained
+F:	Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml
+
 TI PCM3060 ASoC CODEC DRIVER
 M:	Kirill Marinushkin <kmarinushkin@birdec.com>
 L:	alsa-devel@alsa-project.org (moderated for non-subscribers)

-- 
2.34.1

[PATCH v2 4/6] crypto: ti - add driver for MCRC64 engine

[Kamlesh Gurudasani]

Add support for MCRC64 engine to calculate 64-bit CRC in Full-CPU mode.

In Full-CPU mode, the CPU does the data patterns transfer and signature
verification all by itself, only CRC calculation is being done by MCRC64
engine.

MCRC64 engine calculates 64-bit cyclic redundancy checks (CRC)
according to the ISO 3309 standard.

The ISO 3309 64-bit CRC model parameters are as follows:
    Generator Polynomial: x^64 + x^4 + x^3 + x + 1
    Polynomial Value: 0x000000000000001B
    Initial value: 0x0000000000000000
    Reflected Input: False
    Reflected Output: False
    Xor Final: 0x0000000000000000

Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
---
 MAINTAINERS                |   2 +
 drivers/crypto/Kconfig     |   1 +
 drivers/crypto/Makefile    |   1 +
 drivers/crypto/ti/Kconfig  |  10 +++
 drivers/crypto/ti/Makefile |   2 +
 drivers/crypto/ti/mcrc64.c | 442 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 6 files changed, 458 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index 66b51f43d196..0ceca7dbede9 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -21483,8 +21483,10 @@ F:	drivers/iio/adc/ti-lmp92064.c
 
 TI MEMORY CYCLIC REDUNDANCY CHECK (MCRC64) DRIVER
 M:	Kamlesh Gurudasani <kamlesh@ti.com>
+L:	linux-crypto@vger.kernel.org
 S:	Maintained
 F:	Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml
+F:	drivers/crypto/ti/mcrc64.c
 
 TI PCM3060 ASoC CODEC DRIVER
 M:	Kirill Marinushkin <kmarinushkin@birdec.com>
diff --git a/drivers/crypto/Kconfig b/drivers/crypto/Kconfig
index c761952f0dc6..2101f92ead66 100644
--- a/drivers/crypto/Kconfig
+++ b/drivers/crypto/Kconfig
@@ -796,5 +796,6 @@ config CRYPTO_DEV_SA2UL
 
 source "drivers/crypto/aspeed/Kconfig"
 source "drivers/crypto/starfive/Kconfig"
+source "drivers/crypto/ti/Kconfig"
 
 endif # CRYPTO_HW
diff --git a/drivers/crypto/Makefile b/drivers/crypto/Makefile
index d859d6a5f3a4..f1a151b73ff1 100644
--- a/drivers/crypto/Makefile
+++ b/drivers/crypto/Makefile
@@ -41,6 +41,7 @@ obj-$(CONFIG_CRYPTO_DEV_SAHARA) += sahara.o
 obj-$(CONFIG_CRYPTO_DEV_SL3516) += gemini/
 obj-y += stm32/
 obj-$(CONFIG_CRYPTO_DEV_TALITOS) += talitos.o
+obj-$(CONFIG_ARCH_K3) += ti/
 obj-$(CONFIG_CRYPTO_DEV_VIRTIO) += virtio/
 obj-$(CONFIG_CRYPTO_DEV_VMX) += vmx/
 obj-$(CONFIG_CRYPTO_DEV_BCM_SPU) += bcm/
diff --git a/drivers/crypto/ti/Kconfig b/drivers/crypto/ti/Kconfig
new file mode 100644
index 000000000000..548d4cb905bf
--- /dev/null
+++ b/drivers/crypto/ti/Kconfig
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: GPL-2.0-only
+config CRYPTO_DEV_TI_MCRC64
+	tristate "Support for TI MCRC64 engine"
+	depends on ARCH_K3
+	select CRYPTO_HASH
+	help
+	  This enables support for the MCRC64 engine
+	  which can be found on TI K3 SOC.
+	  MCRC64 engine calculates 64-bit cyclic redundancy checks (CRC)
+	  according to the ISO 3309 standard.
diff --git a/drivers/crypto/ti/Makefile b/drivers/crypto/ti/Makefile
new file mode 100644
index 000000000000..94ffc2576137
--- /dev/null
+++ b/drivers/crypto/ti/Makefile
@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0-only
+obj-$(CONFIG_CRYPTO_DEV_TI_MCRC64) += mcrc64.o
diff --git a/drivers/crypto/ti/mcrc64.c b/drivers/crypto/ti/mcrc64.c
new file mode 100644
index 000000000000..9a7634bf1c21
--- /dev/null
+++ b/drivers/crypto/ti/mcrc64.c
@@ -0,0 +1,442 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) Texas Instruments 2023 - http://www.ti.com
+ * Author: Kamlesh Gurudasani <kamlesh@ti.com>
+ */
+
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/mod_devicetable.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+
+#include <crypto/internal/hash.h>
+
+#include <asm/unaligned.h>
+
+#define DRIVER_NAME		"mcrc64"
+#define CHKSUM_DIGEST_SIZE	8
+#define CHKSUM_BLOCK_SIZE	1
+
+/* Registers */
+#define CRC_CTRL0 0x0000 /* CRC Global Control Register 0 */
+#define CH_PSA_SWRE(ch) BIT(((ch) - 1) << 3) /* PSA Software Reset  */
+
+#define CRC_CTRL1 0x0008 /* CRC Global Control Register 1 */
+#define PWDN BIT(0) /* Power Down  */
+
+#define CRC_CTRL2 0x0010 /* CRC Global Control Register 2 */
+#define CH_MODE(ch, m) ((m) << (((ch) - 1) << 3))
+
+#define PSA_SIGREGL(ch) ((0x6 + (4 * ((ch) - 1))) << 4) /* Signature register */
+
+#define MCRC64_ALG_MASK 0x8000000000000000
+#define MCRC64_CRC64_POLY 0x000000000000001b
+
+#define MCRC64_AUTOSUSPEND_DELAY 50
+
+enum mcrc64_mode {
+	MCRC64_MODE_DATA_CAPTURE = 0,
+	MCRC64_MODE_AUTO,
+	MCRC64_MODE_SEMI_CPU,
+	MCRC64_MODE_FULL_CPU,
+	MCRC64_MODE_INVALID,
+};
+
+enum mcrc64_channel {
+	MCRC64_CHANNEL_1 = 1,
+	MCRC64_CHANNEL_2,
+	MCRC64_CHANNEL_3,
+	MCRC64_CHANNEL_4,
+	MCRC64_CHANNEL_INVALID,
+};
+
+struct mcrc64_data {
+	struct list_head list;
+	struct device *dev;
+	void __iomem *regs;
+};
+
+struct mcrc64_list {
+	struct list_head dev_list;
+	spinlock_t lock; /* protect dev_list */
+};
+
+static struct mcrc64_list mcrc64_dev_list = {
+	.dev_list = LIST_HEAD_INIT(mcrc64_dev_list.dev_list),
+	.lock = __SPIN_LOCK_UNLOCKED(mcrc64_dev_list.lock),
+};
+
+struct mcrc64_tfm_ctx {
+	struct mcrc64_data *dev_data;
+	u64 key;
+};
+
+struct mcrc64_desc_ctx {
+	u64 signature;
+};
+
+static struct mcrc64_data *mcrc64_get_dev(struct mcrc64_tfm_ctx *tctx)
+{
+	struct mcrc64_data *dev_data;
+
+	if (tctx->dev_data)
+		return tctx->dev_data;
+
+	spin_lock_bh(&mcrc64_dev_list.lock);
+	dev_data = list_first_entry(&mcrc64_dev_list.dev_list, struct mcrc64_data, list);
+	if (dev_data)
+		list_move_tail(&dev_data->list, &mcrc64_dev_list.dev_list);
+	spin_unlock_bh(&mcrc64_dev_list.lock);
+
+	return dev_data;
+}
+
+static int mcrc64_set_mode(void __iomem *regs, u32 channel, u32 mode)
+{
+	u32 mode_set_val;
+	u32 crc_ctrl2_reg = 0;
+
+	if (mode < 0 || mode >= MCRC64_MODE_INVALID)
+		return -EINVAL;
+
+	if (channel <= 0 || channel >= MCRC64_CHANNEL_INVALID)
+		return -EINVAL;
+
+	mode_set_val = crc_ctrl2_reg | CH_MODE(channel, mode);
+
+	/* Write CRC_CTRL2, set mode */
+	writel_relaxed(mode_set_val, regs + CRC_CTRL2);
+
+	return 0;
+}
+
+static int mcrc64_reset_signature(void __iomem *regs, u32 channel)
+{
+	u32 crc_ctrl0_reg, reset_val, reset_undo_val;
+
+	if (channel <= 0 || channel >= MCRC64_CHANNEL_INVALID)
+		return -EINVAL;
+
+	/* reset PSA */
+	crc_ctrl0_reg = readl_relaxed(regs + CRC_CTRL0);
+
+	reset_val = crc_ctrl0_reg | CH_PSA_SWRE(channel);
+	reset_undo_val = crc_ctrl0_reg & ~CH_PSA_SWRE(channel);
+
+	/* Write CRC_CTRL0 register, reset PSA register */
+	writel_relaxed(reset_val, regs + CRC_CTRL0);
+	writel_relaxed(reset_undo_val, regs + CRC_CTRL0);
+
+	return 0;
+}
+
+/* This helper implements crc64 calculation using CPU */
+static u64 mcrc64_calculate_sw_crc(u64 crc, u8 byte)
+{
+	u64 bit = 0;
+	u8 j;
+
+	for (j = 0; j < 8; j++) {
+		bit = crc & MCRC64_ALG_MASK;
+		crc <<= 1;
+		if (byte & (0x80 >> j))
+			bit ^= MCRC64_ALG_MASK;
+		if (bit)
+			crc ^= MCRC64_CRC64_POLY;
+	}
+
+	return crc;
+}
+
+static int mcrc64_calculate_crc(void __iomem *regs, u32 channel,
+				const u8 *d8, size_t length, u64 *crc64)
+{
+	void __iomem *psa_reg;
+	u64 signature = 0;
+
+	if (channel <= 0 || channel >= MCRC64_CHANNEL_INVALID)
+		return -EINVAL;
+
+	psa_reg = regs + PSA_SIGREGL(channel);
+
+	for (; length >= sizeof(u64); d8 += sizeof(u64), length -= sizeof(u64))
+		writeq_relaxed(cpu_to_be64p((u64 *)d8), psa_reg);
+
+	signature = readq_relaxed(psa_reg);
+
+	if (length) {
+		while (length--)
+			signature = mcrc64_calculate_sw_crc(signature, *d8++);
+
+		/* set capture mode */
+		int ret = mcrc64_set_mode(regs, MCRC64_CHANNEL_1,
+					MCRC64_MODE_DATA_CAPTURE);
+		if (ret)
+			return ret;
+
+		ret = mcrc64_reset_signature(regs, MCRC64_CHANNEL_1);
+		if (ret)
+			return ret;
+
+		writeq_relaxed(signature, psa_reg);
+
+		ret = mcrc64_set_mode(regs, MCRC64_CHANNEL_1,
+				      MCRC64_MODE_FULL_CPU);
+		if (ret)
+			return ret;
+	}
+
+	*crc64 = signature;
+
+	return 0;
+}
+
+static int mcrc64_cra_init(struct crypto_tfm *tfm)
+{
+	struct mcrc64_tfm_ctx *tctx = crypto_tfm_ctx(tfm);
+	struct mcrc64_data *dev_data;
+
+	dev_data = mcrc64_get_dev(tctx);
+	if (!dev_data)
+		return -ENODEV;
+
+	pm_runtime_get_sync(dev_data->dev);
+
+	tctx->key = 0;
+
+	return 0;
+}
+
+static void mcrc64_cra_exit(struct crypto_tfm *tfm)
+{
+	struct mcrc64_tfm_ctx *tctx = crypto_tfm_ctx(tfm);
+	struct mcrc64_data *dev_data;
+
+	dev_data = mcrc64_get_dev(tctx);
+
+	pm_runtime_mark_last_busy(dev_data->dev);
+	pm_runtime_put_autosuspend(dev_data->dev);
+}
+
+static int mcrc64_setkey(struct crypto_shash *tfm, const u8 *key,
+			 unsigned int keylen)
+{
+	struct mcrc64_tfm_ctx *tctx = crypto_shash_ctx(tfm);
+
+	if (keylen != sizeof(u64))
+		return -EINVAL;
+
+	tctx->key = get_unaligned_le64(key);
+
+	return 0;
+}
+
+static int mcrc64_init(struct shash_desc *desc)
+{
+	struct mcrc64_desc_ctx *ctx = shash_desc_ctx(desc);
+	struct mcrc64_tfm_ctx *tctx = crypto_shash_ctx(desc->tfm);
+	struct mcrc64_data *dev_data;
+	void __iomem *psa_reg;
+
+	dev_data = mcrc64_get_dev(tctx);
+	if (!dev_data)
+		return -ENODEV;
+
+	pm_runtime_get_sync(dev_data->dev);
+
+	/* set capture mode */
+	int ret = mcrc64_set_mode(dev_data->regs, MCRC64_CHANNEL_1,
+				  MCRC64_MODE_DATA_CAPTURE);
+	if (ret)
+		return ret;
+
+	/* reset PSA */
+	psa_reg = dev_data->regs + PSA_SIGREGL(MCRC64_CHANNEL_1);
+	ret =  mcrc64_reset_signature(dev_data->regs, MCRC64_CHANNEL_1);
+	if (ret)
+		return ret;
+
+	/* write key  */
+	writeq_relaxed(tctx->key, psa_reg);
+
+	/* set full cpu mode */
+	ret = mcrc64_set_mode(dev_data->regs, MCRC64_CHANNEL_1,
+			      MCRC64_MODE_FULL_CPU);
+	if (ret)
+		return ret;
+
+	ctx->signature = readq_relaxed(psa_reg);
+
+	return 0;
+}
+
+static int mcrc64_update(struct shash_desc *desc, const u8 *d8,
+			 unsigned int length)
+{
+	struct mcrc64_desc_ctx *ctx = shash_desc_ctx(desc);
+	struct mcrc64_tfm_ctx *tctx = crypto_shash_ctx(desc->tfm);
+	struct mcrc64_data *dev_data;
+
+	dev_data = mcrc64_get_dev(tctx);
+	if (!dev_data)
+		return -ENODEV;
+
+	return mcrc64_calculate_crc(dev_data->regs, MCRC64_CHANNEL_1,
+				    d8, length, &ctx->signature);
+}
+
+static int mcrc64_final(struct shash_desc *desc, u8 *out)
+{
+	struct mcrc64_desc_ctx *ctx = shash_desc_ctx(desc);
+
+	/* Send computed CRC */
+	put_unaligned_le64(ctx->signature, out);
+	return 0;
+}
+
+static int mcrc64_finup(struct shash_desc *desc, const u8 *data,
+			unsigned int length, u8 *out)
+{
+	return mcrc64_update(desc, data, length) ?:
+		mcrc64_final(desc, out);
+}
+
+static int mcrc64_digest(struct shash_desc *desc, const u8 *data,
+			 unsigned int length, u8 *out)
+{
+	return mcrc64_init(desc) ?: mcrc64_finup(desc, data, length, out);
+}
+
+static unsigned int refcnt;
+static DEFINE_MUTEX(refcnt_lock);
+static struct shash_alg algs[] = {
+	/* CRC-64 */
+	{
+		.setkey		= mcrc64_setkey,
+		.init		= mcrc64_init,
+		.update		= mcrc64_update,
+		.final		= mcrc64_final,
+		.finup		= mcrc64_finup,
+		.digest		= mcrc64_digest,
+		.descsize	= sizeof(struct mcrc64_desc_ctx),
+		.digestsize	= CHKSUM_DIGEST_SIZE,
+		.base		= {
+			.cra_name		= "crc64-iso",
+			.cra_driver_name	= "mcrc64",
+			.cra_priority		= 300,
+			.cra_flags		= CRYPTO_ALG_OPTIONAL_KEY,
+			.cra_blocksize		= CHKSUM_BLOCK_SIZE,
+			.cra_alignmask		= 7,
+			.cra_ctxsize		= sizeof(struct mcrc64_tfm_ctx),
+			.cra_module		= THIS_MODULE,
+			.cra_init		= mcrc64_cra_init,
+			.cra_exit		= mcrc64_cra_exit,
+		}
+	}
+};
+
+static int mcrc64_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct mcrc64_data *dev_data;
+	int ret;
+
+	dev_data = devm_kzalloc(dev, sizeof(*dev_data), GFP_KERNEL);
+	if (!dev_data)
+		return -ENOMEM;
+
+	dev_data->dev = dev;
+	dev_data->regs = devm_platform_ioremap_resource(pdev, 0);
+
+	platform_set_drvdata(pdev, dev_data);
+
+	spin_lock(&mcrc64_dev_list.lock);
+	list_add(&dev_data->list, &mcrc64_dev_list.dev_list);
+	spin_unlock(&mcrc64_dev_list.lock);
+
+	mutex_lock(&refcnt_lock);
+	if (!refcnt) {
+		ret = crypto_register_shashes(algs, ARRAY_SIZE(algs));
+		if (ret) {
+			mutex_unlock(&refcnt_lock);
+			dev_err(dev, "Failed to register\n");
+			return ret;
+		}
+	}
+	refcnt++;
+	mutex_unlock(&refcnt_lock);
+
+	pm_runtime_set_autosuspend_delay(dev, MCRC64_AUTOSUSPEND_DELAY);
+	pm_runtime_use_autosuspend(dev);
+
+	pm_runtime_get_noresume(dev);
+	pm_runtime_set_active(dev);
+	pm_runtime_enable(dev);
+
+	pm_runtime_put_sync(dev);
+
+	return 0;
+}
+
+static int mcrc64_remove(struct platform_device *pdev)
+{
+	struct mcrc64_data *dev_data = platform_get_drvdata(pdev);
+	int ret;
+
+	ret = pm_runtime_resume_and_get(dev_data->dev);
+	if (ret < 0)
+		return ret;
+
+	spin_lock(&mcrc64_dev_list.lock);
+	list_del(&dev_data->list);
+	spin_unlock(&mcrc64_dev_list.lock);
+
+	mutex_lock(&refcnt_lock);
+	if (!--refcnt)
+		crypto_unregister_shashes(algs, ARRAY_SIZE(algs));
+	mutex_unlock(&refcnt_lock);
+
+	pm_runtime_disable(dev_data->dev);
+	pm_runtime_put_noidle(dev_data->dev);
+
+	return 0;
+}
+
+static int __maybe_unused mcrc64_suspend(struct device *dev)
+{
+	return	pm_runtime_force_suspend(dev);
+}
+
+static int __maybe_unused mcrc64_resume(struct device *dev)
+{
+	return pm_runtime_force_resume(dev);
+}
+
+static const struct dev_pm_ops mcrc64_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(mcrc64_suspend,
+				mcrc64_resume)
+};
+
+static const struct of_device_id of_match[] = {
+	{ .compatible = "ti,am62-mcrc64", },
+	{},
+};
+MODULE_DEVICE_TABLE(of, of_match);
+
+static struct platform_driver mcrc64_driver = {
+	.probe	= mcrc64_probe,
+	.remove = mcrc64_remove,
+	.driver = {
+		.name		= DRIVER_NAME,
+		.pm		= &mcrc64_pm_ops,
+		.of_match_table = of_match,
+	},
+};
+
+module_platform_driver(mcrc64_driver);
+
+MODULE_AUTHOR("Kamlesh Gurudasani <kamlesh@ti.com>");
+MODULE_DESCRIPTION("Texas Instruments MCRC64 driver");
+MODULE_LICENSE("GPL");

-- 
2.34.1

[PATCH v2 5/6] arm64: dts: ti: k3-am62: Add dt node, cbass_main ranges for MCRC64

[Kamlesh Gurudasani]

Add the address space for MCRC64 to the ranges property of the
cbass_main node and add dt node for MCRC64 engine

Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
---
 arch/arm64/boot/dts/ti/k3-am62-main.dtsi | 7 +++++++
 arch/arm64/boot/dts/ti/k3-am62.dtsi      | 1 +
 2 files changed, 8 insertions(+)

diff --git a/arch/arm64/boot/dts/ti/k3-am62-main.dtsi b/arch/arm64/boot/dts/ti/k3-am62-main.dtsi
index 5f19ef46d44c..84a7e7d43d02 100644
--- a/arch/arm64/boot/dts/ti/k3-am62-main.dtsi
+++ b/arch/arm64/boot/dts/ti/k3-am62-main.dtsi
@@ -201,6 +201,13 @@ crypto: crypto@40900000 {
 		dma-names = "tx", "rx1", "rx2";
 	};
 
+	crc: crc@30300000 {
+		compatible = "ti,am62-mcrc64";
+		reg = <0x00 0x30300000 0x00 0x1000>;
+		clocks = <&k3_clks 116 0>;
+		power-domains = <&k3_pds 116 TI_SCI_PD_EXCLUSIVE>;
+	};
+
 	secure_proxy_sa3: mailbox@43600000 {
 		compatible = "ti,am654-secure-proxy";
 		#mbox-cells = <1>;
diff --git a/arch/arm64/boot/dts/ti/k3-am62.dtsi b/arch/arm64/boot/dts/ti/k3-am62.dtsi
index 5e72c445f37a..7acd8009aa28 100644
--- a/arch/arm64/boot/dts/ti/k3-am62.dtsi
+++ b/arch/arm64/boot/dts/ti/k3-am62.dtsi
@@ -65,6 +65,7 @@ cbass_main: bus@f0000 {
 			 <0x00 0x30040000 0x00 0x30040000 0x00 0x00080000>, /* PRUSS-M */
 			 <0x00 0x30101000 0x00 0x30101000 0x00 0x00010100>, /* CSI window */
 			 <0x00 0x30200000 0x00 0x30200000 0x00 0x00010000>, /* DSS */
+			 <0x00 0x30300000 0x00 0x30300000 0x00 0x00001000>, /* MCRC64 */
 			 <0x00 0x31000000 0x00 0x31000000 0x00 0x00050000>, /* USB0 DWC3 Core window */
 			 <0x00 0x31100000 0x00 0x31100000 0x00 0x00050000>, /* USB1 DWC3 Core window */
 			 <0x00 0x40900000 0x00 0x40900000 0x00 0x00030000>, /* SA3UL */

-- 
2.34.1

[PATCH v2 6/6] arm64: defconfig: enable TI MCRC64 module

[Kamlesh Gurudasani]

K3 devices include MCRC64 engine for crc64 calculation.
Enable module to be built for K3 devices.

Also enable algif_hash module, which is needed to access MCRC64 module
from userspace.

Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
---
 arch/arm64/configs/defconfig | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index bf13d5c46578..4d555a125315 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -1535,6 +1535,7 @@ CONFIG_CRYPTO_TEST=m
 CONFIG_CRYPTO_ECHAINIV=y
 CONFIG_CRYPTO_MICHAEL_MIC=m
 CONFIG_CRYPTO_ANSI_CPRNG=y
+CONFIG_CRYPTO_USER_API_HASH=m
 CONFIG_CRYPTO_USER_API_RNG=m
 CONFIG_CRYPTO_CHACHA20_NEON=m
 CONFIG_CRYPTO_GHASH_ARM64_CE=y
@@ -1558,6 +1559,7 @@ CONFIG_CRYPTO_DEV_HISI_ZIP=m
 CONFIG_CRYPTO_DEV_HISI_HPRE=m
 CONFIG_CRYPTO_DEV_HISI_TRNG=m
 CONFIG_CRYPTO_DEV_SA2UL=m
+CONFIG_CRYPTO_DEV_TI_MCRC64=m
 CONFIG_DMA_RESTRICTED_POOL=y
 CONFIG_CMA_SIZE_MBYTES=32
 CONFIG_PRINTK_TIME=y

-- 
2.34.1

Re: [PATCH v2 5/6] arm64: dts: ti: k3-am62: Add dt node, cbass_main ranges for MCRC64

[Nishanth Menon]

On 00:58-20230811, Kamlesh Gurudasani wrote:
> Add the address space for MCRC64 to the ranges property of the
> cbass_main node and add dt node for MCRC64 engine
> 
> Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
> ---

Please mark device tree changes in the same series (if you are posting) as
"DONOTMERGE" to help driver maintainers understand that the dts come in
via SoC tree (most maintainers do not need this, but it keeps things
sane and adjust your expectation that this will hit the next window from
where the driver gets in).

-- 
Regards,
Nishanth Menon
Key (0xDDB5849D1736249D) / Fingerprint: F8A2 8693 54EB 8232 17A3  1A34 DDB5 849D 1736 249D

Re: [PATCH v2 6/6] arm64: defconfig: enable TI MCRC64 module

[Nishanth Menon]

On 00:58-20230811, Kamlesh Gurudasani wrote:
> K3 devices include MCRC64 engine for crc64 calculation.
> Enable module to be built for K3 devices.
> 
> Also enable algif_hash module, which is needed to access MCRC64 module
> from userspace.
> 
> Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
> ---

There are few things to improve in this series, but we can discuss this
as part of defconfig merge

See thread: https://lore.kernel.org/linux-arm-kernel/ae2ad056-96de-41b7-8df4-1d9c0f5c469b@app.fastmail.com/
for additional info.

K3 devices is too broad, you want to specify specific boards that will
benefit out of this.

I suggest to keep this as "DONOTMERGE" to indicate this should'nt go via
subsystem maintainer tree (most maintainers are  aware of it, but
explicitly calling it out helps keep things sane)


>  arch/arm64/configs/defconfig | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
> index bf13d5c46578..4d555a125315 100644
> --- a/arch/arm64/configs/defconfig
> +++ b/arch/arm64/configs/defconfig
> @@ -1535,6 +1535,7 @@ CONFIG_CRYPTO_TEST=m
>  CONFIG_CRYPTO_ECHAINIV=y
>  CONFIG_CRYPTO_MICHAEL_MIC=m
>  CONFIG_CRYPTO_ANSI_CPRNG=y
> +CONFIG_CRYPTO_USER_API_HASH=m
>  CONFIG_CRYPTO_USER_API_RNG=m
>  CONFIG_CRYPTO_CHACHA20_NEON=m
>  CONFIG_CRYPTO_GHASH_ARM64_CE=y
> @@ -1558,6 +1559,7 @@ CONFIG_CRYPTO_DEV_HISI_ZIP=m
>  CONFIG_CRYPTO_DEV_HISI_HPRE=m
>  CONFIG_CRYPTO_DEV_HISI_TRNG=m
>  CONFIG_CRYPTO_DEV_SA2UL=m
> +CONFIG_CRYPTO_DEV_TI_MCRC64=m
>  CONFIG_DMA_RESTRICTED_POOL=y
>  CONFIG_CMA_SIZE_MBYTES=32
>  CONFIG_PRINTK_TIME=y
> 
> -- 
> 2.34.1
> 

-- 
Regards,
Nishanth Menon
Key (0xDDB5849D1736249D) / Fingerprint: F8A2 8693 54EB 8232 17A3  1A34 DDB5 849D 1736 249D

Re: [PATCH v2 2/6] crypto: crc64 - add crc64-iso framework

[Eric Biggers]

On Fri, Aug 11, 2023 at 12:58:49AM +0530, Kamlesh Gurudasani wrote:
> diff --git a/lib/crc64-iso.c b/lib/crc64-iso.c
> new file mode 100644
> index 000000000000..d6e803124fa0
[...]
> +u64 crc64_iso_update(u64 crc, const unsigned char *buffer, size_t len)
> +{
> +	struct {
> +		struct shash_desc shash;
> +		u64 crc;
> +	} desc;
> +	int err;
> +
> +	if (static_branch_unlikely(&crc64_iso_fallback))
> +		return crc64_iso_generic(crc, buffer, len);
> +
> +	rcu_read_lock();
> +	desc.shash.tfm = rcu_dereference(crc64_iso_tfm);
> +	desc.crc = crc;
> +	err = crypto_shash_update(&desc.shash, buffer, len);
> +	rcu_read_unlock();
> +
> +	WARN_ON_ONCE(err);
> +
> +	return desc.crc;
> +}
> +EXPORT_SYMBOL_GPL(crc64_iso_update);
> +
> +u64 crc64_iso(const unsigned char *buffer, size_t len)
> +{
> +	return crc64_iso_update(0, buffer, len);
> +}
> +EXPORT_SYMBOL_GPL(crc64_iso);

These functions are never called.

Why are you trying to add unused code to the kernel?

- Eric

Re: [EXTERNAL] Re: [PATCH v2 2/6] crypto: crc64 - add crc64-iso framework

[Kamlesh Gurudasani]

Eric Biggers <ebiggers@kernel.org> writes:

> On Fri, Aug 11, 2023 at 12:58:49AM +0530, Kamlesh Gurudasani wrote:
>> diff --git a/lib/crc64-iso.c b/lib/crc64-iso.c
>> new file mode 100644
>> index 000000000000..d6e803124fa0
> [...]
>> +u64 crc64_iso_update(u64 crc, const unsigned char *buffer, size_t len)
>> +{
>> +	struct {
>> +		struct shash_desc shash;
>> +		u64 crc;
>> +	} desc;
>> +	int err;
>> +
>> +	if (static_branch_unlikely(&crc64_iso_fallback))
>> +		return crc64_iso_generic(crc, buffer, len);
>> +
>> +	rcu_read_lock();
>> +	desc.shash.tfm = rcu_dereference(crc64_iso_tfm);
>> +	desc.crc = crc;
>> +	err = crypto_shash_update(&desc.shash, buffer, len);
>> +	rcu_read_unlock();
>> +
>> +	WARN_ON_ONCE(err);
>> +
>> +	return desc.crc;
>> +}
>> +EXPORT_SYMBOL_GPL(crc64_iso_update);
>> +
>> +u64 crc64_iso(const unsigned char *buffer, size_t len)
>> +{
>> +	return crc64_iso_update(0, buffer, len);
>> +}
>> +EXPORT_SYMBOL_GPL(crc64_iso);
>
> These functions are never called.
>
> Why are you trying to add unused code to the kernel?
>
> - Eric
Thanks for the review, Eric.

Will remove this in next revision.

Regards,
Kamlesh

Re: [PATCH v2 3/6] dt-bindings: crypto: Add Texas Instruments MCRC64

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 3009 bytes --]

On Fri, Aug 11, 2023 at 12:58:50AM +0530, Kamlesh Gurudasani wrote:
> Add binding for Texas Instruments MCRC64
> 
> MCRC64 engine calculates 64-bit cyclic redundancy checks (CRC)
> according to the ISO 3309 standard.
> 
> The ISO 3309 64-bit CRC model parameters are as follows:
>     Generator Polynomial: x^64 + x^4 + x^3 + x + 1
>     Polynomial Value: 0x000000000000001B
>     Initial value: 0x0000000000000000
>     Reflected Input: False
>     Reflected Output: False
>     Xor Final: 0x0000000000000000
> 
> Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
> ---
>  Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml | 47 +++++++++++++++++++++++++++++++++++++++++++++++
>  MAINTAINERS                                             |  5 +++++
>  2 files changed, 52 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml b/Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml
> new file mode 100644
> index 000000000000..38bc7efebd68
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml
> @@ -0,0 +1,47 @@
> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/crypto/ti,mcrc64.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: Texas Instruments MCRC64
> +
> +description: The MCRC64 engine calculates 64-bit cyclic redundancy checks

A newline after "description" please.

> +  (CRC) according to the ISO 3309 standard.
> +
> +maintainers:
> +  - Kamlesh Gurudasani <kamlesh@ti.com>
> +
> +properties:
> +  compatible:
> +    const: ti,am62-mcrc64

Is the am62 an SoC or a family of SoCs? I googled a wee bit for am62 &
there seems to be an am625 and an am623?

Otherwise, this looks good to me.

> +
> +  reg:
> +    maxItems: 1
> +
> +  clocks:
> +    maxItems: 1
> +
> +  power-domains:
> +    maxItems: 1
> +
> +required:
> +  - compatible
> +  - reg
> +  - clocks
> +  - power-domains
> +
> +additionalProperties: false
> +
> +examples:
> +  - |
> +    #include <dt-bindings/soc/ti,sci_pm_domain.h>
> +
> +    crc@30300000 {
> +      compatible = "ti,am62-mcrc64";
> +      reg = <0x30300000 0x1000>;
> +      clocks = <&k3_clks 116 0>;
> +      power-domains = <&k3_pds 116 TI_SCI_PD_EXCLUSIVE>;
> +    };
> +
> +...
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 02a3192195af..66b51f43d196 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -21481,6 +21481,11 @@ S:	Maintained
>  F:	Documentation/devicetree/bindings/iio/adc/ti,lmp92064.yaml
>  F:	drivers/iio/adc/ti-lmp92064.c
>  
> +TI MEMORY CYCLIC REDUNDANCY CHECK (MCRC64) DRIVER
> +M:	Kamlesh Gurudasani <kamlesh@ti.com>
> +S:	Maintained
> +F:	Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml
> +
>  TI PCM3060 ASoC CODEC DRIVER
>  M:	Kirill Marinushkin <kmarinushkin@birdec.com>
>  L:	alsa-devel@alsa-project.org (moderated for non-subscribers)
> 
> -- 
> 2.34.1
> 

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH v2 3/6] dt-bindings: crypto: Add Texas Instruments MCRC64

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 3532 bytes --]

On Fri, Aug 11, 2023 at 04:34:33PM +0100, Conor Dooley wrote:
> On Fri, Aug 11, 2023 at 12:58:50AM +0530, Kamlesh Gurudasani wrote:
> > Add binding for Texas Instruments MCRC64
> > 
> > MCRC64 engine calculates 64-bit cyclic redundancy checks (CRC)
> > according to the ISO 3309 standard.
> > 
> > The ISO 3309 64-bit CRC model parameters are as follows:
> >     Generator Polynomial: x^64 + x^4 + x^3 + x + 1
> >     Polynomial Value: 0x000000000000001B
> >     Initial value: 0x0000000000000000
> >     Reflected Input: False
> >     Reflected Output: False
> >     Xor Final: 0x0000000000000000
> > 
> > Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
> > ---
> >  Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml | 47 +++++++++++++++++++++++++++++++++++++++++++++++
> >  MAINTAINERS                                             |  5 +++++
> >  2 files changed, 52 insertions(+)
> > 
> > diff --git a/Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml b/Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml
> > new file mode 100644
> > index 000000000000..38bc7efebd68
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml
> > @@ -0,0 +1,47 @@
> > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> > +%YAML 1.2
> > +---
> > +$id: http://devicetree.org/schemas/crypto/ti,mcrc64.yaml#
> > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > +
> > +title: Texas Instruments MCRC64
> > +
> > +description: The MCRC64 engine calculates 64-bit cyclic redundancy checks
> 
> A newline after "description" please.
> 
> > +  (CRC) according to the ISO 3309 standard.
> > +
> > +maintainers:
> > +  - Kamlesh Gurudasani <kamlesh@ti.com>
> > +
> > +properties:
> > +  compatible:
> > +    const: ti,am62-mcrc64
> 
> Is the am62 an SoC or a family of SoCs? I googled a wee bit for am62 &
> there seems to be an am625 and an am623?

Or is it an am62p5, in which case the compatible should contain
ti,am62p5 I suppose. Sorry for my confusion here, its not really clear
me too since I've been seeing many different-but-similar product names
the last few days.

Thanks,
Conor.

> 
> Otherwise, this looks good to me.
> 
> > +
> > +  reg:
> > +    maxItems: 1
> > +
> > +  clocks:
> > +    maxItems: 1
> > +
> > +  power-domains:
> > +    maxItems: 1
> > +
> > +required:
> > +  - compatible
> > +  - reg
> > +  - clocks
> > +  - power-domains
> > +
> > +additionalProperties: false
> > +
> > +examples:
> > +  - |
> > +    #include <dt-bindings/soc/ti,sci_pm_domain.h>
> > +
> > +    crc@30300000 {
> > +      compatible = "ti,am62-mcrc64";
> > +      reg = <0x30300000 0x1000>;
> > +      clocks = <&k3_clks 116 0>;
> > +      power-domains = <&k3_pds 116 TI_SCI_PD_EXCLUSIVE>;
> > +    };
> > +
> > +...
> > diff --git a/MAINTAINERS b/MAINTAINERS
> > index 02a3192195af..66b51f43d196 100644
> > --- a/MAINTAINERS
> > +++ b/MAINTAINERS
> > @@ -21481,6 +21481,11 @@ S:	Maintained
> >  F:	Documentation/devicetree/bindings/iio/adc/ti,lmp92064.yaml
> >  F:	drivers/iio/adc/ti-lmp92064.c
> >  
> > +TI MEMORY CYCLIC REDUNDANCY CHECK (MCRC64) DRIVER
> > +M:	Kamlesh Gurudasani <kamlesh@ti.com>
> > +S:	Maintained
> > +F:	Documentation/devicetree/bindings/crypto/ti,mcrc64.yaml
> > +
> >  TI PCM3060 ASoC CODEC DRIVER
> >  M:	Kirill Marinushkin <kmarinushkin@birdec.com>
> >  L:	alsa-devel@alsa-project.org (moderated for non-subscribers)
> > 
> > -- 
> > 2.34.1
> > 



[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH v2 2/6] crypto: crc64 - add crc64-iso framework

[Eric Biggers]

On Fri, Aug 11, 2023 at 12:58:49AM +0530, Kamlesh Gurudasani wrote:
> diff --git a/include/linux/crc64.h b/include/linux/crc64.h
> index 70202da51c2c..10b792080374 100644
> --- a/include/linux/crc64.h
> +++ b/include/linux/crc64.h
> @@ -8,11 +8,15 @@
>  #include <linux/types.h>
>  
>  #define CRC64_ROCKSOFT_STRING "crc64-rocksoft"
> +#define CRC64_ISO_STRING "crc64-iso"
>  
>  u64 __pure crc64_be(u64 crc, const void *p, size_t len);
>  u64 __pure crc64_iso_generic(u64 crc, const void *p, size_t len);
>  u64 __pure crc64_rocksoft_generic(u64 crc, const void *p, size_t len);
>  
> +u64 crc64_iso(const unsigned char *buffer, size_t len);
> +u64 crc64_iso_update(u64 crc, const unsigned char *buffer, size_t len);
> +
>  u64 crc64_rocksoft(const unsigned char *buffer, size_t len);
>  u64 crc64_rocksoft_update(u64 crc, const unsigned char *buffer, size_t len);

Is "crc64-iso" clear enough, or should it be "crc64-iso3309"?  There are
thousands of ISO standards.  Different CRC variants are specified by different
ISO standards.  Is this particular variant indeed commonly referred to as simply
the "ISO" CRC-64?  Even if it's currently the case that all other CRCs in ISO
standards are different widths than 64 bits (which may be unlikely?), I'm not
sure we should count on no CRC-64 variant ever being standardized by ISO.

- Eric

Re: [PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[Eric Biggers]

On Fri, Aug 11, 2023 at 12:58:47AM +0530, Kamlesh Gurudasani wrote:
> Add support for MCRC64 engine to calculate 64-bit CRC in Full-CPU mode
> 
> MCRC64 engine calculates 64-bit cyclic redundancy checks (CRC)
> according to the ISO 3309 standard.
> 
> The ISO 3309 64-bit CRC model parameters are as follows:
>     Generator Polynomial: x^64 + x^4 + x^3 + x + 1
>     Polynomial Value: 0x000000000000001B
>     Initial value: 0x0000000000000000
>     Reflected Input: False
>     Reflected Output: False
>     Xor Final: 0x0000000000000000
> 
> Tested with
> CONFIG_CRYPTO_MANAGER_DISABLE_TESTS is not set
> CONFIG_CRYPTO_MANAGER_EXTRA_TESTS=y
> 
> and tcrypt,
> sudo modprobe tcrypt mode=329 sec=1
> 
> User space application implemented using algif_hash,
> https://gist.github.com/ti-kamlesh/73abfcc1a33318bb3b199d36b6209e59
> 
> Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>

I do not see any in-kernel user of this CRC variant being introduced, which
leaves algif_hash as the only use case.

Can you elaborate on the benefit this brings to your application?  Yes, it
allows you to use your hardware CRC engine.  But, that comes with all the
overhead from the syscalls, algif_hash, and the driver.  How does performance
compare to a properly optimized software CRC implementation on your platform,
i.e. an implementation using carryless multiplication instructions (e.g. ARMv8
CE) if available on your platform, otherwise an implementation using the
slice-by-8 or slice-by-16 method?

- Eric

Re: [EXTERNAL] Re: [PATCH v2 2/6] crypto: crc64 - add crc64-iso framework

[Kamlesh Gurudasani]

Eric Biggers <ebiggers@kernel.org> writes:

> On Fri, Aug 11, 2023 at 12:58:49AM +0530, Kamlesh Gurudasani wrote:
>> diff --git a/include/linux/crc64.h b/include/linux/crc64.h
>> index 70202da51c2c..10b792080374 100644
>> --- a/include/linux/crc64.h
>> +++ b/include/linux/crc64.h
>> @@ -8,11 +8,15 @@
>>  #include <linux/types.h>
>>  
>>  #define CRC64_ROCKSOFT_STRING "crc64-rocksoft"
>> +#define CRC64_ISO_STRING "crc64-iso"
>>  
>>  u64 __pure crc64_be(u64 crc, const void *p, size_t len);
>>  u64 __pure crc64_iso_generic(u64 crc, const void *p, size_t len);
>>  u64 __pure crc64_rocksoft_generic(u64 crc, const void *p, size_t len);
>>  
>> +u64 crc64_iso(const unsigned char *buffer, size_t len);
>> +u64 crc64_iso_update(u64 crc, const unsigned char *buffer, size_t len);
>> +
>>  u64 crc64_rocksoft(const unsigned char *buffer, size_t len);
>>  u64 crc64_rocksoft_update(u64 crc, const unsigned char *buffer, size_t len);
>
> Is "crc64-iso" clear enough, or should it be "crc64-iso3309"?  There are
> thousands of ISO standards.  Different CRC variants are specified by different
> ISO standards.  Is this particular variant indeed commonly referred to as simply
> the "ISO" CRC-64?  Even if it's currently the case that all other CRCs in ISO
> standards are different widths than 64 bits (which may be unlikely?), I'm not
> sure we should count on no CRC-64 variant ever being standardized by ISO.
>
> - Eric
https://en.wikipedia.org/wiki/Cyclic_redundancy_check

Last entry CRC-64-ISO in the table.
It is mentioned as crc64-iso and that's the
only 64-bit CRC standardized by ISO. But I do agree that crc64-iso3309 would
be more specific, will change it to crc64-iso3309 in next
revision. Thanks.

Regards,
Kamlesh

Re: [EXTERNAL] Re: [PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[Kamlesh Gurudasani]

Eric Biggers <ebiggers@kernel.org> writes:

> On Fri, Aug 11, 2023 at 12:58:47AM +0530, Kamlesh Gurudasani wrote:
>> Add support for MCRC64 engine to calculate 64-bit CRC in Full-CPU mode
>> 
>> MCRC64 engine calculates 64-bit cyclic redundancy checks (CRC)
>> according to the ISO 3309 standard.
>> 
>> The ISO 3309 64-bit CRC model parameters are as follows:
>>     Generator Polynomial: x^64 + x^4 + x^3 + x + 1
>>     Polynomial Value: 0x000000000000001B
>>     Initial value: 0x0000000000000000
>>     Reflected Input: False
>>     Reflected Output: False
>>     Xor Final: 0x0000000000000000
>> 
>> Tested with
>> CONFIG_CRYPTO_MANAGER_DISABLE_TESTS is not set
>> CONFIG_CRYPTO_MANAGER_EXTRA_TESTS=y
>> 
>> and tcrypt,
>> sudo modprobe tcrypt mode=329 sec=1
>> 
>> User space application implemented using algif_hash,
>> https://gist.github.com/ti-kamlesh/73abfcc1a33318bb3b199d36b6209e59
>> 
>> Signed-off-by: Kamlesh Gurudasani <kamlesh@ti.com>
>
> I do not see any in-kernel user of this CRC variant being introduced, which
> leaves algif_hash as the only use case.
>
> Can you elaborate on the benefit this brings to your application?  Yes, it
> allows you to use your hardware CRC engine.  But, that comes with all the
> overhead from the syscalls, algif_hash, and the driver.  How does performance
> compare to a properly optimized software CRC implementation on your platform,
> i.e. an implementation using carryless multiplication instructions (e.g. ARMv8
> CE) if available on your platform, otherwise an implementation using the
> slice-by-8 or slice-by-16 method?
>
> - Eric
Hi Eric,

We are more interested in offload than performance, with splice system
call and DMA mode in driver(will be implemented after this series gets
merged), good amount of cpu cycles will be saved.

There is one more mode(auto mode) in mcrc64 which helps to verify crc64
values against pre calculated crc64, saving the efforts of comparing in
userspace.

Current generic implementation of crc64-iso(part of this series)
gives 173 Mb/s of speed as opposed to mcrc64 which gives speed of 812
Mb/s when tested with tcrypt.

Regard,
Kamlesh

Re: [EXTERNAL] Re: [PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[Eric Biggers]

On Fri, Aug 18, 2023 at 02:36:34PM +0530, Kamlesh Gurudasani wrote:
> Hi Eric,
> 
> We are more interested in offload than performance, with splice system
> call and DMA mode in driver(will be implemented after this series gets
> merged), good amount of cpu cycles will be saved.

So it's for power usage, then?  Or freeing up CPU for other tasks?

> There is one more mode(auto mode) in mcrc64 which helps to verify crc64
> values against pre calculated crc64, saving the efforts of comparing in
> userspace.

Is there any path forward to actually support this?

> 
> Current generic implementation of crc64-iso(part of this series)
> gives 173 Mb/s of speed as opposed to mcrc64 which gives speed of 812
> Mb/s when tested with tcrypt.

This doesn't answer my question, which to reiterate was:

    How does performance compare to a properly optimized software CRC
    implementation on your platform, i.e. an implementation using carryless
    multiplication instructions (e.g. ARMv8 CE) if available on your platform,
    otherwise an implementation using the slice-by-8 or slice-by-16 method?

The implementation you tested was slice-by-1.  Compared to that, it's common for
slice-by-8 to speed up CRCs by about 4 times and for folding with carryless
multiplication to speed up CRCs by 10-30 times, sometimes limited only by memory
bandwidth.  I don't know what specific results you would get on your specific
CPU and for this specific CRC, and you could certainly see something different
if you e.g. have some low-end embedded CPU.  But those are the typical results
I've seen for other CRCs on different CPUs.  So, a software implementation may
be more attractive than you realize.  It could very well be the case that a
PMULL based CRC implementation actually ends up with less CPU load than your
"hardware offload", when taking into syscall, algif_hash, and driver overhead...

- Eric

Re: [EXTERNAL] Re: [EXTERNAL] Re: [PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[Kamlesh Gurudasani]

Eric Biggers <ebiggers@kernel.org> writes:

> On Fri, Aug 18, 2023 at 02:36:34PM +0530, Kamlesh Gurudasani wrote:
>> Hi Eric,
>> 
>> We are more interested in offload than performance, with splice system
>> call and DMA mode in driver(will be implemented after this series gets
>> merged), good amount of cpu cycles will be saved.
>
> So it's for power usage, then?  Or freeing up CPU for other tasks?
It is for freeing CPU for other tasks

>
>> There is one more mode(auto mode) in mcrc64 which helps to verify crc64
>> values against pre calculated crc64, saving the efforts of comparing in
>> userspace.
>
> Is there any path forward to actually support this?
>
>> 
>> Current generic implementation of crc64-iso(part of this series)
>> gives 173 Mb/s of speed as opposed to mcrc64 which gives speed of 812
>> Mb/s when tested with tcrypt.
>
> This doesn't answer my question, which to reiterate was:
>
>     How does performance compare to a properly optimized software CRC
>     implementation on your platform, i.e. an implementation using carryless
>     multiplication instructions (e.g. ARMv8 CE) if available on your platform,
>     otherwise an implementation using the slice-by-8 or slice-by-16 method?
>
> The implementation you tested was slice-by-1.  Compared to that, it's common for
> slice-by-8 to speed up CRCs by about 4 times and for folding with carryless
> multiplication to speed up CRCs by 10-30 times, sometimes limited only by memory
> bandwidth.  I don't know what specific results you would get on your specific
> CPU and for this specific CRC, and you could certainly see something different
> if you e.g. have some low-end embedded CPU.  But those are the typical results
> I've seen for other CRCs on different CPUs.  So, a software implementation may
> be more attractive than you realize.  It could very well be the case that a
> PMULL based CRC implementation actually ends up with less CPU load than your
> "hardware offload", when taking into syscall, algif_hash, and driver overhead...
>
> - Eric
Hi Eric, thanks for your detailed and valuable inputs.

As per your suggestion, we did some profiling. 

Use case is to calculate crc32/crc64 for file input from user space.

Instead of directly implementing PMULL based CRC64, we made first comparison between 
Case 1.
CRC32 (splice() + kernel space SW driver) 
https://gist.github.com/ti-kamlesh/5be75dbde292e122135ddf795fad9f21

Case 2.
CRC32(mmap() + userspace armv8 crc32 instruction implementation)
(tried read() as well to get contents of file, but that lost to mmap() so not mentioning number here)
https://gist.github.com/ti-kamlesh/002df094dd522422c6cb62069e15c40d

Case 3.
CRC64 (splice() + MCRC64 HW)
https://gist.github.com/ti-kamlesh/98b1fc36c9a7c3defcc2dced4136b8a0


Overall, overhead of userspace + af_alg + driver in (Case 1) and
( Case 3) is ~0.025s, which is constant for any file size.
This is calculated using real time to calculate crc  -
driver time (time spend inside init() + update() +final()) = overhead ~0.025s    



+-------------------+-----------------------------+-----------------------+------------------------+------------------------+
|                   |                             |                       |                        |                        |
| File size         | 120mb(ideal size for us)    | 20mb                  | 15mb                   | 5mb                    |
+===================+=============================+=======================+========================+========================+
|                   |                             |                       |                        |                        |
| CRC32 (Case 1)    | Driver time 0.155s          | Driver time 0.0325s   | Driver time 0.019s     | Driver time 0.0062s    |
|                   |    real time 0.18s          |    real time 0.06s    |    real time 0.04s     |    real time 0.03s     |
|                   |    overhead 0.025s          |    overhead 0.025s    |    overhead 0.021s     |    overhead ~0.023s    |
+-------------------+-----------------------------+-----------------------+------------------------+------------------------+
|                   |                             |                       |                        |                        |
| CRC32 (Case 2)    | Real time 0.30s             | Real time 0.05s       | Real time 0.04s        | Real time 0.02s        |
+-------------------+-----------------------------+-----------------------+------------------------+------------------------+
|                   |                             |                       |                        |                        |
| CRC64 (Case 3)    | Driver time   0.385s        | Driver time 0.0665s   | Driver time 0.0515s    | Driver time 0.019s     |
|                   |    real time 0.41s          |    real time 0.09s    |    real time 0.08s     |    real time 0.04s     |
|                   |    overhead 0.025s          |    overhead 0.025s    |    overhead ~0.025s    |    overhead ~0.021s    |
+-------------------+-----------------------------+-----------------------+------------------------+------------------------+

Here, if we consider similar numbers for crc64 PMULL implementation as
crc32 (case 2) , we save good number of cpu cycles using mcrc64
in case of files bigger than 5-10mb as most of the time is being spent in HW offload.

Regards,
Kamlesh

Re: [EXTERNAL] Re: [EXTERNAL] Re: [PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[Kamlesh Gurudasani]

Eric Biggers <ebiggers@kernel.org> writes:

> On Fri, Aug 18, 2023 at 02:36:34PM +0530, Kamlesh Gurudasani wrote:
>> Hi Eric,
>> 
>> We are more interested in offload than performance, with splice system
>> call and DMA mode in driver(will be implemented after this series gets
>> merged), good amount of cpu cycles will be saved.
>
> So it's for power usage, then?  Or freeing up CPU for other tasks?
>

It's for freeing up CPU for other tasks

>> There is one more mode(auto mode) in mcrc64 which helps to verify crc64
>> values against pre calculated crc64, saving the efforts of comparing in
>> userspace.
>
> Is there any path forward to actually support this?
>
>> 
>> Current generic implementation of crc64-iso(part of this series)
>> gives 173 Mb/s of speed as opposed to mcrc64 which gives speed of 812
>> Mb/s when tested with tcrypt.
>
> This doesn't answer my question, which to reiterate was:
>
>     How does performance compare to a properly optimized software CRC
>     implementation on your platform, i.e. an implementation using carryless
>     multiplication instructions (e.g. ARMv8 CE) if available on your platform,
>     otherwise an implementation using the slice-by-8 or slice-by-16 method?
>
> The implementation you tested was slice-by-1.  Compared to that, it's common for
> slice-by-8 to speed up CRCs by about 4 times and for folding with carryless
> multiplication to speed up CRCs by 10-30 times, sometimes limited only by memory
> bandwidth.  I don't know what specific results you would get on your specific
> CPU and for this specific CRC, and you could certainly see something different
> if you e.g. have some low-end embedded CPU.  But those are the typical results
> I've seen for other CRCs on different CPUs.  So, a software implementation may
> be more attractive than you realize.  It could very well be the case that a
> PMULL based CRC implementation actually ends up with less CPU load than your
> "hardware offload", when taking into syscall, algif_hash, and driver overhead...
>
> - Eric

Hi Eric, thanks for your detailed and valuable inputs.

As per your suggestion, we did some profiling. 

Use case is to calculate crc32/crc64 for file input from user space.

Instead of directly implementing PMULL based CRC64, we made first comparison between 
Case 1.
CRC32 (splice() + kernel space SW driver) 
https://gist.github.com/ti-kamlesh/5be75dbde292e122135ddf795fad9f21

Case 2.
CRC32(mmap() + userspace armv8 crc32 instruction implementation)
(tried read() as well to get contents of file, but that lost to mmap() so not mentioning number here)
https://gist.github.com/ti-kamlesh/002df094dd522422c6cb62069e15c40d

Case 3.
CRC64 (splice() + MCRC64 HW)
https://gist.github.com/ti-kamlesh/98b1fc36c9a7c3defcc2dced4136b8a0


Overall, overhead of userspace + af_alg + driver in (Case 1) and ( Case 3) is ~0.025s, which is constant for any file size.
This is calculated using real time to calculate crc  - driver time (time spend inside init() + update() +final()) = overhead ~0.025s    

Here, if we consider similar numbers for crc64 PMULL implementation as crc32 (case 2) , we save good number of cpu cycles using mcrc64
in case of files bigger than 5-10mb as most of the time is being spent in HW offload.

+-------------------+-----------------------------+-----------------------+------------------------+------------------------+
|                   |                             |                       |                        |                        |
| File size         | 120mb(ideal size for us)    | 20mb                  | 15mb                   | 5mb                    |
+===================+=============================+=======================+========================+========================+
|                   |                             |                       |                        |                        |
| CRC32 (Case 1)    | Driver time 0.155s          | Driver time 0.0325s   | Driver time 0.019s     | Driver time 0.0062s    |
|                   |    real time 0.18s          |    real time 0.06s    |    real time 0.04s     |    real time 0.03s     |
|                   |    overhead 0.025s          |    overhead 0.025s    |    overhead 0.021s     |    overhead ~0.023s    |
+-------------------+-----------------------------+-----------------------+------------------------+------------------------+
|                   |                             |                       |                        |                        |
| CRC32 (Case 2)    | Real time 0.30s             | Real time 0.05s       | Real time 0.04s        | Real time 0.02s        |
+-------------------+-----------------------------+-----------------------+------------------------+------------------------+
|                   |                             |                       |                        |                        |
| CRC64 (Case 3)    | Driver time   0.385s        | Driver time 0.0665s   | Driver time 0.0515s    | Driver time 0.019s     |
|                   |    real time 0.41s          |    real time 0.09s    |    real time 0.08s     |    real time 0.04s     |
|                   |    overhead 0.025s          |    overhead 0.025s    |    overhead ~0.025s    |    overhead ~0.021s    |
+-------------------+-----------------------------+-----------------------+------------------------+------------------------+

Re: [EXTERNAL] Re: [EXTERNAL] Re: [PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[Kamlesh Gurudasani]

Eric Biggers <ebiggers@kernel.org> writes:

> On Fri, Aug 18, 2023 at 02:36:34PM +0530, Kamlesh Gurudasani wrote:
>> Hi Eric,
>> 
>> We are more interested in offload than performance, with splice system
>> call and DMA mode in driver(will be implemented after this series gets
>> merged), good amount of cpu cycles will be saved.
>
> So it's for power usage, then?  Or freeing up CPU for other tasks?
>
It's for freeing CPU fpr other tasks
>> There is one more mode(auto mode) in mcrc64 which helps to verify crc64
>> values against pre calculated crc64, saving the efforts of comparing in
>> userspace.
>
> Is there any path forward to actually support this?
>
>> 
>> Current generic implementation of crc64-iso(part of this series)
>> gives 173 Mb/s of speed as opposed to mcrc64 which gives speed of 812
>> Mb/s when tested with tcrypt.
>
> This doesn't answer my question, which to reiterate was:
>
>     How does performance compare to a properly optimized software CRC
>     implementation on your platform, i.e. an implementation using carryless
>     multiplication instructions (e.g. ARMv8 CE) if available on your platform,
>     otherwise an implementation using the slice-by-8 or slice-by-16 method?
>
> The implementation you tested was slice-by-1.  Compared to that, it's common for
> slice-by-8 to speed up CRCs by about 4 times and for folding with carryless
> multiplication to speed up CRCs by 10-30 times, sometimes limited only by memory
> bandwidth.  I don't know what specific results you would get on your specific
> CPU and for this specific CRC, and you could certainly see something different
> if you e.g. have some low-end embedded CPU.  But those are the typical results
> I've seen for other CRCs on different CPUs.  So, a software implementation may
> be more attractive than you realize.  It could very well be the case that a
> PMULL based CRC implementation actually ends up with less CPU load than your
> "hardware offload", when taking into syscall, algif_hash, and driver overhead...
>
> - Eric

Hi Eric, thanks for your detailed and valuable inputs.

As per your suggestion, we did some profiling. 

Use case is to calculate crc32/crc64 for file input from user space.

Instead of directly implementing PMULL based CRC64, we made first comparison between 
Case 1.
CRC32 (splice() + kernel space SW driver) 
https://gist.github.com/ti-kamlesh/5be75dbde292e122135ddf795fad9f21

Case 2.
CRC32(mmap() + userspace armv8 crc32 instruction implementation)
(tried read() as well to get contents of file, but that lost to mmap() so not mentioning number here)
https://gist.github.com/ti-kamlesh/002df094dd522422c6cb62069e15c40d

Case 3.
CRC64 (splice() + MCRC64 HW)
https://gist.github.com/ti-kamlesh/98b1fc36c9a7c3defcc2dced4136b8a0


Overall, overhead of userspace + af_alg + driver in (Case 1) and ( Case 3) is ~0.025s, which is constant for any file size.
This is calculated using real time to calculate crc  - driver time (time spend inside init() + update() +final()) = overhead ~0.025s    

Here, if we consider similar numbers for crc64 PMULL implementation as crc32 (case 2) , we save good number of cpu cycles using mcrc64
in case of files bigger than 5-10mb as most of the time is being spent
in HW offload.

╔═══════════════════╦═════════════════════════════╦═══════════════════════╦════════════════════════╦════════════════════════╗
â•‘                   â•‘                             â•‘                       â•‘                        â•‘                        â•‘
â•‘ File size         â•‘ 120mb(ideal size for us)    â•‘ 20mb                  â•‘ 15mb                   â•‘ 5mb                    â•‘
╠═══════════════════╬═════════════════════════════╬═══════════════════════╬════════════════════════╬════════════════════════╣
â•‘                   â•‘                             â•‘                       â•‘                        â•‘                        â•‘
â•‘ CRC32 (Case 1)    â•‘ Driver time 0.155s          â•‘ Driver time 0.0325s   â•‘ Driver time 0.019s     â•‘ Driver time 0.0062s    â•‘
â•‘                   â•‘    real time 0.18s          â•‘    real time 0.06s    â•‘    real time 0.04s     â•‘    real time 0.03s     â•‘
â•‘                   â•‘    overhead 0.025s          â•‘    overhead 0.025s    â•‘    overhead 0.021s     â•‘    overhead ~0.023s    â•‘
╠═══════════════════╬═════════════════════════════╬═══════════════════════╬════════════════════════╬════════════════════════╣
â•‘                   â•‘                             â•‘                       â•‘                        â•‘                        â•‘
â•‘ CRC32 (Case 2)    â•‘ Real time 0.30s             â•‘ Real time 0.05s       â•‘ Real time 0.04s        â•‘ Real time 0.02s        â•‘
╠═══════════════════╬═════════════════════════════╬═══════════════════════╬════════════════════════╬════════════════════════╣
â•‘                   â•‘                             â•‘                       â•‘                        â•‘                        â•‘
â•‘ CRC64 (Case 3)    â•‘ Driver time   0.385s        â•‘ Driver time 0.0665s   â•‘ Driver time 0.0515s    â•‘ Driver time 0.019s     â•‘
â•‘                   â•‘    real time 0.41s          â•‘    real time 0.09s    â•‘    real time 0.08s     â•‘    real time 0.04s     â•‘
â•‘                   â•‘    overhead 0.025s          â•‘    overhead 0.025s    â•‘    overhead ~0.025s    â•‘    overhead ~0.021s    â•‘
╚═══════════════════╩═════════════════════════════╩═══════════════════════╩════════════════════════╩════════════════════════╝

Re: [EXTERNAL] Re: [EXTERNAL] Re: [PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[Kamlesh Gurudasani]

Eric Biggers <ebiggers@kernel.org> writes:

Somehow couple of my earlier mails got blocked mailing list because of
table formatting, I guess. Resending. Accept my apologies for spamming. 

> On Fri, Aug 18, 2023 at 02:36:34PM +0530, Kamlesh Gurudasani wrote:
>> Hi Eric,
>> 
>> We are more interested in offload than performance, with splice system
>> call and DMA mode in driver(will be implemented after this series gets
>> merged), good amount of cpu cycles will be saved.
>
> So it's for power usage, then?  Or freeing up CPU for other tasks?
It's for freeing the CPU for other tasks

>
>> There is one more mode(auto mode) in mcrc64 which helps to verify crc64
>> values against pre calculated crc64, saving the efforts of comparing in
>> userspace.
>
> Is there any path forward to actually support this?
>
>> 
>> Current generic implementation of crc64-iso(part of this series)
>> gives 173 Mb/s of speed as opposed to mcrc64 which gives speed of 812
>> Mb/s when tested with tcrypt.
>
> This doesn't answer my question, which to reiterate was:
>
>     How does performance compare to a properly optimized software CRC
>     implementation on your platform, i.e. an implementation using carryless
>     multiplication instructions (e.g. ARMv8 CE) if available on your platform,
>     otherwise an implementation using the slice-by-8 or slice-by-16 method?
>
> The implementation you tested was slice-by-1.  Compared to that, it's common for
> slice-by-8 to speed up CRCs by about 4 times and for folding with carryless
> multiplication to speed up CRCs by 10-30 times, sometimes limited only by memory
> bandwidth.  I don't know what specific results you would get on your specific
> CPU and for this specific CRC, and you could certainly see something different
> if you e.g. have some low-end embedded CPU.  But those are the typical results
> I've seen for other CRCs on different CPUs.  So, a software implementation may
> be more attractive than you realize.  It could very well be the case that a
> PMULL based CRC implementation actually ends up with less CPU load than your
> "hardware offload", when taking into syscall, algif_hash, and driver overhead...
>
> - Eric

Hi Eric, thanks for your detailed and valuable inputs.

As per your suggestion, we did some profiling. 

Use case is to calculate crc32/crc64 for file input from user space.

Instead of directly implementing PMULL based CRC64, we made first comparison between Case 1.
CRC32 (splice() + kernel space SW driver)
https://gist.github.com/ti-kamlesh/5be75dbde292e122135ddf795fad9f21

Case 2.
CRC32(mmap() + userspace armv8 crc32 instruction implementation)
(tried read() as well to get contents of file,
but that lost to mmap() so not mentioning number here)
https://gist.github.com/ti-kamlesh/002df094dd522422c6cb62069e15c40d

Case 3.
CRC64 (splice() + MCRC64 HW)
https://gist.github.com/ti-kamlesh/98b1fc36c9a7c3defcc2dced4136b8a0


Overall, overhead of userspace + af_alg + driver in (Case 1) and ( Case
3) is ~0.025s, which is constant for any file size.
This is calculated using
real time to calculate crc  - driver time (time spend inside init() + update() +final()) = overhead ~0.025s    

Here, if we consider similar numbers for crc64 PMULL implementation as
crc32 (case 2) ,
we save good number of cpu cycles using mcrc64 in case of files bigger
than 5-10mb as most of the time is being spent in HW offload.



Comparison table:
https://gist.github.com/ti-kamlesh/8117b6f7120960a71541ab67c671602a

Re: [EXTERNAL] Re: [EXTERNAL] Re: [PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[Kamlesh Gurudasani]

Kamlesh Gurudasani <kamlesh@ti.com> writes:
...

> Hi Eric, thanks for your detailed and valuable inputs.
>
> As per your suggestion, we did some profiling. 
>
> Use case is to calculate crc32/crc64 for file input from user space.
>
> Instead of directly implementing PMULL based CRC64, we made first comparison between 
> Case 1.
> CRC32 (splice() + kernel space SW driver) 
> https://gist.github.com/ti-kamlesh/5be75dbde292e122135ddf795fad9f21
>
> Case 2.
> CRC32(mmap() + userspace armv8 crc32 instruction implementation)
> (tried read() as well to get contents of file, but that lost to mmap() so not mentioning number here)
> https://gist.github.com/ti-kamlesh/002df094dd522422c6cb62069e15c40d
>
> Case 3.
> CRC64 (splice() + MCRC64 HW)
> https://gist.github.com/ti-kamlesh/98b1fc36c9a7c3defcc2dced4136b8a0
>
>
> Overall, overhead of userspace + af_alg + driver in (Case 1) and
> ( Case 3) is ~0.025s, which is constant for any file size.
> This is calculated using real time to calculate crc  -
> driver time (time spend inside init() + update() +final()) = overhead ~0.025s    
>
>
>
> +-------------------+-----------------------------+-----------------------+------------------------+------------------------+
> |                   |                             |                       |                        |                        |
> | File size         | 120mb(ideal size for us)    | 20mb                  | 15mb                   | 5mb                    |
> +===================+=============================+=======================+========================+========================+
> |                   |                             |                       |                        |                        |
> | CRC32 (Case 1)    | Driver time 0.155s          | Driver time 0.0325s   | Driver time 0.019s     | Driver time 0.0062s    |
> |                   |    real time 0.18s          |    real time 0.06s    |    real time 0.04s     |    real time 0.03s     |
> |                   |    overhead 0.025s          |    overhead 0.025s    |    overhead 0.021s     |    overhead ~0.023s    |
> +-------------------+-----------------------------+-----------------------+------------------------+------------------------+
> |                   |                             |                       |                        |                        |
> | CRC32 (Case 2)    | Real time 0.30s             | Real time 0.05s       | Real time 0.04s        | Real time 0.02s        |
> +-------------------+-----------------------------+-----------------------+------------------------+------------------------+
> |                   |                             |                       |                        |                        |
> | CRC64 (Case 3)    | Driver time   0.385s        | Driver time 0.0665s   | Driver time 0.0515s    | Driver time 0.019s     |
> |                   |    real time 0.41s          |    real time 0.09s    |    real time 0.08s     |    real time 0.04s     |
> |                   |    overhead 0.025s          |    overhead 0.025s    |    overhead ~0.025s    |    overhead ~0.021s    |
> +-------------------+-----------------------------+-----------------------+------------------------+------------------------+
>
> Here, if we consider similar numbers for crc64 PMULL implementation as
> crc32 (case 2) , we save good number of cpu cycles using mcrc64
> in case of files bigger than 5-10mb as most of the time is being spent in HW offload.
>
> Regards,
> Kamlesh

Hi Eric,

Please let me know if above numbers make sense to you and I should send
next revision.

Regards,
Kamlesh

Re: [EXTERNAL] Re: [EXTERNAL] Re: [PATCH v2 0/6] Add support for Texas Instruments MCRC64 engine

[Kamlesh Gurudasani]

Kamlesh Gurudasani <kamlesh@ti.com> writes:

>>
>> Here, if we consider similar numbers for crc64 PMULL implementation as
>> crc32 (case 2) , we save good number of cpu cycles using mcrc64
>> in case of files bigger than 5-10mb as most of the time is being spent in HW offload.
>>
>> Regards,
>> Kamlesh
>
> Hi Eric,
>
> Please let me know if above numbers make sense to you and I should send
> next revision.

Hi Eric,

I understand that there is no in-kernel user for crc64-iso3309 and this
is new algorithm that we are trying to add in linux kernel.

As per your suggestion we did the calculations and it turns out to be we
are saving good number of cpu cycles with HW offload.

Also, there are some automotive customers who have a safety
requirement to offload any parameters that are in Linux to ensure FFI.

Let me know if you are willing to accept this driver, so that I can put
efforts to send next revision.

Regards,
Kamlesh

RE: Re: [PATCH v2 2/6] crypto: crc64 - add crc64-iso framework

[Elliott, Robert (Servers)]

> -----Original Message-----
> From: Kamlesh Gurudasani <kamlesh@ti.com>
> Sent: Friday, August 18, 2023 2:26 AM
> Subject: Re: [EXTERNAL] Re: [PATCH v2 2/6] crypto: crc64 - add crc64-iso
> framework
> 
> Eric Biggers <ebiggers@kernel.org> writes:
> 
> > Is "crc64-iso" clear enough, or should it be "crc64-iso3309"?  There are
> > thousands of ISO standards.  Different CRC variants are specified by
> different
> > ISO standards.  Is this particular variant indeed commonly referred to
> as simply
> > the "ISO" CRC-64?  Even if it's currently the case that all other CRCs
> in ISO
> > standards are different widths than 64 bits (which may be unlikely?),
> I'm not
> > sure we should count on no CRC-64 variant ever being standardized by
> ISO.
> >
> > - Eric
> https://en.wikipedia.org/wiki/Cyclic_redundancy_check
> 
> Last entry CRC-64-ISO in the table.
> It is mentioned as crc64-iso and that's the
> only 64-bit CRC standardized by ISO. 

ECMA-182 (DLT-1 tapes) was contributed to become ISO/IEC 13421 in 1993, so
that was another "64-bit CRC standardized by ISO." Plus, ISO could publish new
standards with new CRCs at any time.

> But I do agree that crc64-iso3309 would be more specific, will change it
> to crc64-iso3309 in next revision. Thanks.
> 
> Regards,
> Kamlesh

ISO-3309:1991 was withdrawn and revised by
ISO/IEC 3309:1993, which was withdrawn and revised by
ISO/IEC 13239:2002, which was confirmed in 2007 and is still current.

Apparently only the 1991 version defined a CRC-64; later versions dropped
that.

Is there really a demand for adding a 23 year old deprecated algorithm to
the kernel?

RE: Re: [PATCH v2 2/6] crypto: crc64 - add crc64-iso framework

[Kamlesh Gurudasani]

"Elliott, Robert (Servers)" <elliott@hpe.com> writes:

>> -----Original Message-----
>> From: Kamlesh Gurudasani <kamlesh@ti.com>
>> Sent: Friday, August 18, 2023 2:26 AM
>> Subject: Re: [EXTERNAL] Re: [PATCH v2 2/6] crypto: crc64 - add crc64-iso
>> framework
>> 
>> Eric Biggers <ebiggers@kernel.org> writes:
>> 
>> > Is "crc64-iso" clear enough, or should it be "crc64-iso3309"?  There are
>> > thousands of ISO standards.  Different CRC variants are specified by
>> different
>> > ISO standards.  Is this particular variant indeed commonly referred to
>> as simply
>> > the "ISO" CRC-64?  Even if it's currently the case that all other CRCs
>> in ISO
>> > standards are different widths than 64 bits (which may be unlikely?),
>> I'm not
>> > sure we should count on no CRC-64 variant ever being standardized by
>> ISO.
>> >
>> > - Eric
>> https://en.wikipedia.org/wiki/Cyclic_redundancy_check
>> 
>> Last entry CRC-64-ISO in the table.
>> It is mentioned as crc64-iso and that's the
>> only 64-bit CRC standardized by ISO. 
>
> ECMA-182 (DLT-1 tapes) was contributed to become ISO/IEC 13421 in 1993, so
> that was another "64-bit CRC standardized by ISO." Plus, ISO could publish new
> standards with new CRCs at any time.
>
>> But I do agree that crc64-iso3309 would be more specific, will change it
>> to crc64-iso3309 in next revision. Thanks.
>> 
>> Regards,
>> Kamlesh
>
> ISO-3309:1991 was withdrawn and revised by
> ISO/IEC 3309:1993, which was withdrawn and revised by
> ISO/IEC 13239:2002, which was confirmed in 2007 and is still current.
>
> Apparently only the 1991 version defined a CRC-64; later versions dropped
> that.
>
> Is there really a demand for adding a 23 year old deprecated algorithm to
> the kernel?
I understand your concern but a lot of TI's K3 based J7* and AM6* SOCs
have MCRC and MCRC64(Mostly on AM6* SOCs)
Where MCRC64 only supports above mentioned CRC64 algorithm and few
customers wants to use the hardware based CRC to ensure FFI, so we
actually need it.
If it is available in upstream and can be used easily, a lot of
customers would want to use it.

I'll look into the naming and will provide something specific to that
particular algo.

Kamlesh