Skip to contentSkip to conversation

3.11 boot failure caused by commit crypto

BUG Zhao Hongjiang 4 participants 9 messages Jul 24, 2013 – Aug 1, 2013

Subsystems (6)
x86 Library code Crypto API Arm STM32 +2 more subsystems
lore
Share & tools
lore

Zhao Hongjiang

author
#0 lore raw

When i boot my system with the latest 3.11-rc2 kernel, something occur as follow:

Waiting for device /dev/disk/by-id/... to appear: ............ Could not find //dev/disk/by-id/... Want me to fall back to /dev/disk/by-id/..... ?

I found out the patch commit id: 2d31e518a42828df7877bca23a958627d60408bc (crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework) cause this problem by bisect.

I revert this patch and this problem go away, the boot become successful.

maintainer
#1 lore raw
Quoted 9 lines · Zhao Hongjiang

When i boot my system with the latest 3.11-rc2 kernel, something occur as follow:

Waiting for device /dev/disk/by-id/... to appear: ............ Could not find //dev/disk/by-id/... Want me to fall back to /dev/disk/by-id/..... ?

I found out the patch commit id: 2d31e518a42828df7877bca23a958627d60408bc (crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework) cause this problem by bisect.

I revert this patch and this problem go away, the boot become successful.

OK, I will revert this until we have a proper solution in place that does not break boot on systems like yours.

commit e70308ec0e4bff344fcfdf160de40e1150552c5f

Author: Herbert Xu <[email protected]>
Date:   Wed Jul 24 17:04:16 2013 +1000

Revert "crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework"

This reverts commits
67822649d7305caf3dd50ed46c27b99c94eff996
39761214eefc6b070f29402aa1165f24d789b3f7
0b95a7f85718adcbba36407ef88bba0a7379ed03
31d939625a9a20b1badd2d4e6bf6fd39fa523405
2d31e518a42828df7877bca23a958627d60408bc

Unfortunately this change broke boot on some systems that used an initrd which does not include the newly created crct10dif modules. As these modules are required by sd_mod under certain configurations this is a serious problem.

Signed-off-by Herbert Xu <[email protected]>
arch/x86/crypto/Makefile +0 -2
index 7d6ba9d..6c63c35 100644
@@ -27,7 +27,6 @@ obj-$(CONFIG_CRYPTO_SHA1_SSSE3) += sha1-ssse3.o
 obj-$(CONFIG_CRYPTO_CRC32_PCLMUL) += crc32-pclmul.o
 obj-$(CONFIG_CRYPTO_SHA256_SSSE3) += sha256-ssse3.o
 obj-$(CONFIG_CRYPTO_SHA512_SSSE3) += sha512-ssse3.o
-obj-$(CONFIG_CRYPTO_CRCT10DIF_PCLMUL) += crct10dif-pclmul.o
 
 # These modules require assembler to support AVX.
 ifeq ($(avx_supported),yes)

@@ -82,4 +81,3 @@ crc32c-intel-$(CONFIG_64BIT) += crc32c-pcl-intel-asm_64.o
 crc32-pclmul-y := crc32-pclmul_asm.o crc32-pclmul_glue.o
 sha256-ssse3-y := sha256-ssse3-asm.o sha256-avx-asm.o sha256-avx2-asm.o sha256_ssse3_glue.o
 sha512-ssse3-y := sha512-ssse3-asm.o sha512-avx-asm.o sha512-avx2-asm.o sha512_ssse3_glue.o
-crct10dif-pclmul-y := crct10dif-pcl-asm_64.o crct10dif-pclmul_glue.o
arch/x86/crypto/crct10dif-pcl-asm_64.S deleted +0 -643

643 lines not on this page · raw message

arch/x86/crypto/crct10dif-pclmul_glue.c deleted +0 -151

151 lines not on this page · raw message

crypto/Kconfig +0 -19
index 904ffe8..2754f2b 100644
@@ -376,25 +376,6 @@ config CRYPTO_CRC32_PCLMUL
 	  which will enable any routine to use the CRC-32-IEEE 802.3 checksum
 	  and gain better performance as compared with the table implementation.
 
-config CRYPTO_CRCT10DIF
-	tristate "CRCT10DIF algorithm"
-	select CRYPTO_HASH
-	help
-	  CRC T10 Data Integrity Field computation is being cast as
-	  a crypto transform.  This allows for faster crc t10 diff
-	  transforms to be used if they are available.
-
-config CRYPTO_CRCT10DIF_PCLMUL
-	tristate "CRCT10DIF PCLMULQDQ hardware acceleration"
-	depends on X86 && 64BIT && CRC_T10DIF
-	select CRYPTO_HASH
-	help
-	  For x86_64 processors with SSE4.2 and PCLMULQDQ supported,
-	  CRC T10 DIF PCLMULQDQ computation can be hardware
-	  accelerated PCLMULQDQ instruction. This option will create
-	  'crct10dif-plcmul' module, which is faster when computing the
-	  crct10dif checksum as compared with the generic table implementation.
-
 config CRYPTO_GHASH
 	tristate "GHASH digest algorithm"
 	select CRYPTO_GF128MUL
crypto/Makefile +0 -1
index 62af87d..a8e9b0f 100644
@@ -83,7 +83,6 @@ obj-$(CONFIG_CRYPTO_ZLIB) += zlib.o
 obj-$(CONFIG_CRYPTO_MICHAEL_MIC) += michael_mic.o
 obj-$(CONFIG_CRYPTO_CRC32C) += crc32c.o
 obj-$(CONFIG_CRYPTO_CRC32) += crc32.o
-obj-$(CONFIG_CRYPTO_CRCT10DIF) += crct10dif.o
 obj-$(CONFIG_CRYPTO_AUTHENC) += authenc.o authencesn.o
 obj-$(CONFIG_CRYPTO_LZO) += lzo.o
 obj-$(CONFIG_CRYPTO_842) += 842.o
crypto/crct10dif.c deleted +0 -178

178 lines not on this page · raw message

crypto/tcrypt.c +0 -8
index 25a5934..66d254c 100644
@@ -1174,10 +1174,6 @@ static int do_test(int m)
 		ret += tcrypt_test("ghash");
 		break;
 
-	case 47:
-		ret += tcrypt_test("crct10dif");
-		break;
-
 	case 100:
 		ret += tcrypt_test("hmac(md5)");
 		break;

@@ -1502,10 +1498,6 @@ static int do_test(int m)
 		test_hash_speed("crc32c", sec, generic_hash_speed_template);
 		if (mode > 300 && mode < 400) break;
 
-	case 320:
-		test_hash_speed("crct10dif", sec, generic_hash_speed_template);
-		if (mode > 300 && mode < 400) break;
-
 	case 399:
 		break;
crypto/testmgr.c +0 -10
index 2f00607..ecddf92 100644
@@ -2046,16 +2046,6 @@ static const struct alg_test_desc alg_test_descs[] = {
 			}
 		}
 	}, {
-		.alg = "crct10dif",
-		.test = alg_test_hash,
-		.fips_allowed = 1,
-		.suite = {
-			.hash = {
-				.vecs = crct10dif_tv_template,
-				.count = CRCT10DIF_TEST_VECTORS
-			}
-		}
-	}, {
 		.alg = "cryptd(__driver-cbc-aes-aesni)",
 		.test = alg_test_null,
 		.fips_allowed = 1,
crypto/testmgr.h +0 -33
index 7d44aa3..1e701bc 100644
@@ -450,39 +450,6 @@ static struct hash_testvec rmd320_tv_template[] = {
 	}
 };
 
-#define CRCT10DIF_TEST_VECTORS	3
-static struct hash_testvec crct10dif_tv_template[] = {
-	{
-		.plaintext = "abc",
-		.psize  = 3,
-#ifdef __LITTLE_ENDIAN
-		.digest = "\x3b\x44",
-#else
-		.digest = "\x44\x3b",
-#endif
-	}, {
-		.plaintext = "1234567890123456789012345678901234567890"
-			     "123456789012345678901234567890123456789",
-		.psize	= 79,
-#ifdef __LITTLE_ENDIAN
-		.digest	= "\x70\x4b",
-#else
-		.digest	= "\x4b\x70",
-#endif
-	}, {
-		.plaintext =
-		"abcddddddddddddddddddddddddddddddddddddddddddddddddddddd",
-		.psize  = 56,
-#ifdef __LITTLE_ENDIAN
-		.digest = "\xe3\x9c",
-#else
-		.digest = "\x9c\xe3",
-#endif
-		.np     = 2,
-		.tap    = { 28, 28 }
-	}
-};
-
 /*
  * SHA1 test vectors  from from FIPS PUB 180-1
  * Long vector from CAVS 5.0
include/linux/crc-t10dif.h +0 -4
index b3cb71f..a9c96d8 100644
@@ -3,10 +3,6 @@
 
 #include <linux/types.h>
 
-#define CRC_T10DIF_DIGEST_SIZE 2
-#define CRC_T10DIF_BLOCK_SIZE 1
-
-__u16 crc_t10dif_generic(__u16 crc, const unsigned char *buffer, size_t len);
 __u16 crc_t10dif(unsigned char const *, size_t);
 
 #endif
lib/Kconfig +0 -2
index 339d5a4..fe01d41 100644
@@ -63,8 +63,6 @@ config CRC16
 
 config CRC_T10DIF
 	tristate "CRC calculation for the T10 Data Integrity Field"
-	select CRYPTO
-	select CRYPTO_CRCT10DIF
 	help
 	  This option is only needed if a module that's not in the
 	  kernel tree needs to calculate CRC checks for use with the
lib/crc-t10dif.c +43 -30

87 lines not on this page · raw message

Thanks,

Signature · 3 lines
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
#2 lore raw
Quoted 27 lines · Herbert Xu
Quoted 9 lines · Zhao Hongjiang

When i boot my system with the latest 3.11-rc2 kernel, something occur as follow:

Waiting for device /dev/disk/by-id/... to appear: ............ Could not find //dev/disk/by-id/... Want me to fall back to /dev/disk/by-id/..... ?

I found out the patch commit id: 2d31e518a42828df7877bca23a958627d60408bc (crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework) cause this problem by bisect.

I revert this patch and this problem go away, the boot become successful.

OK, I will revert this until we have a proper solution in place that does not break boot on systems like yours.

commit e70308ec0e4bff344fcfdf160de40e1150552c5f

Author: Herbert Xu <[email protected]>
Date:   Wed Jul 24 17:04:16 2013 +1000

Revert "crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework"

This reverts commits
67822649d7305caf3dd50ed46c27b99c94eff996
39761214eefc6b070f29402aa1165f24d789b3f7
0b95a7f85718adcbba36407ef88bba0a7379ed03
31d939625a9a20b1badd2d4e6bf6fd39fa523405
2d31e518a42828df7877bca23a958627d60408bc

Quoted 4 lines

Unfortunately this change broke boot on some systems that used an initrd which does not include the newly created crct10dif modules. As these modules are required by sd_mod under certain configurations this is a serious problem.

Herbert, what are your thoughts on a proper fix to initrd issue for crct10dif modules not getting included? Or can we
let the config option for the pclmul version to be either compiled out or compiled in for now for people who do want to use this?

Thanks.

Tim

maintainer
#3 lore raw
Quoted 5 lines · Tim Chen

Herbert, what are your thoughts on a proper fix to initrd issue for crct10dif modules not getting included? Or can we
let the config option for the pclmul version to be either compiled out or compiled in for now for people who do want to use this?

With the MODULES_SOFTDEP patch that Rusty recently merged, all we need to do is to use that in lib/crct10dif module. Once that patch hits mainline I'll take care of the rest.

Cheers,

Signature · 3 lines
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
#4 lore raw
Quoted 10 lines · Herbert Xu
Quoted 5 lines · Tim Chen

Herbert, what are your thoughts on a proper fix to initrd issue for crct10dif modules not getting included? Or can we
let the config option for the pclmul version to be either compiled out or compiled in for now for people who do want to use this?

With the MODULES_SOFTDEP patch that Rusty recently merged, all we need to do is to use that in lib/crct10dif module. Once that patch hits mainline I'll take care of the rest.

Great. Thanks.

Tim

#5 lore raw
Quoted 8 lines · Tim Chen
Quoted 7 lines · Herbert Xu
Quoted 5 lines · Tim Chen

Herbert, what are your thoughts on a proper fix to initrd issue for crct10dif modules not getting included? Or can we
let the config option for the pclmul version to be either compiled out or compiled in for now for people who do want to use this?

Quoted 7 lines
Quoted 3 lines

With the MODULES_SOFTDEP patch that Rusty recently merged, all we need to do is to use that in lib/crct10dif module. Once that patch hits mainline I'll take care of the rest.

Great. Thanks.

Tim

Does this imply that you want me to push that to Linus now, and/or CC stable? Was planning for *next* merge window...

Cheers,
Rusty.

maintainer
#6 lore raw
Quoted 3 lines · Rusty Russell

Does this imply that you want me to push that to Linus now, and/or CC stable? Was planning for *next* merge window...

Next merge window is fine. However, it would be good to restore the new driver for that as well. So I see two ways to do this, either we push the crct10dif driver patches through your tree, on top of the MODULES_SOFTDEP patch, or I pull the MODULES_SOFTDEP patch into the crypto tree.

What would you like to do?

Cheers,

Signature · 3 lines
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
Acked-by Herbert Xu
#7 lore raw
Quoted 5 lines · Herbert Xu
Quoted 3 lines · Rusty Russell

Does this imply that you want me to push that to Linus now, and/or CC stable? Was planning for *next* merge window...

Quoted 7 lines

Next merge window is fine. However, it would be good to restore the new driver for that as well. So I see two ways to do this, either we push the crct10dif driver patches through your tree, on top of the MODULES_SOFTDEP patch, or I pull the MODULES_SOFTDEP patch into the crypto tree.

What would you like to do?

You're welcome to it: it's v. unlikely to conflict with anything else.

Here it is, fresh from my pending-rebases branch, with my Signed-off-by.

Cheers,
Rusty.

From: Andreas Robinson <[email protected]>
Subject: modules: add support for soft module dependencies

Additional and optional dependencies not found while building the kernel and modules, can now be declared explicitly.

Signed-off-by Andreas Robinson <[email protected]>
Acked-by Herbert Xu <[email protected]>
Signed-off-by Rusty Russell <[email protected]>
---
 include/linux/module.h | 5 +++++
 1 file changed, 5 insertions(+)
include/linux/module.h +5 -0
index 46f1ea0..504035f 100644
@@ -97,6 +97,11 @@ extern const struct gtype##_id __mod_##gtype##_table		\
 /* For userspace: you can also call me... */
 #define MODULE_ALIAS(_alias) MODULE_INFO(alias, _alias)
 
+/* Soft module dependencies. See man modprobe.d for details.
+ * Example: MODULE_SOFTDEP("pre: module-foo module-bar post: module-baz")
+ */
+#define MODULE_SOFTDEP(_softdep) MODULE_INFO(softdep, _softdep)
+
 /*
  * The following license idents are currently accepted as indicating free
  * software modules
maintainer
#8 lore raw
Quoted 4 lines · Rusty Russell

You're welcome to it: it's v. unlikely to conflict with anything else.

Here it is, fresh from my pending-rebases branch, with my Signed-off-by.

Great. I've added it to the cryptodev tree. Thanks Rusty!

Signature · 3 lines
Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt