[git] GCRYPT - branch, master, updated. libgcrypt-1.8.0-7-gcf1528e

by Jussi Kivilinna cvs at cvs.gnupg.org
Tue Aug 1 20:36:46 CEST 2017

This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "The GNU crypto library".

The branch, master has been updated
       via  cf1528e7f2761774d06ace0de48f39c96b52dc4f (commit)
       via  4a7aa30ae9f3ce798dd886c2f2d4164c43027748 (commit)
      from  b7cd44335d9cde43be6f693dca6399ed0762649c (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
commit cf1528e7f2761774d06ace0de48f39c96b52dc4f
Author: Jussi Kivilinna <jussi.kivilinna at iki.fi>
Date:   Sat Jul 29 14:34:23 2017 +0300

    Fix return value type for _gcry_md_extract
    * src/gcrypt-int.h (_gcry_md_extract): Use gpg_err_code_t instead of
    gpg_error_t for internal function return type.
    GnuPG-bug-id: 3314
    Signed-off-by: Jussi Kivilinna <jussi.kivilinna at iki.fi>

diff --git a/src/gcrypt-int.h b/src/gcrypt-int.h
index ddcafa5..ad719be 100644
--- a/src/gcrypt-int.h
+++ b/src/gcrypt-int.h
@@ -39,7 +39,7 @@ typedef struct mpi_ec_ctx_s *mpi_ec_t;
 /* Underscore prefixed internal versions of the public functions.
-   They return gpg_err_code and not gpg_error_t.  Some macros also
+   They return gpg_err_code_t and not gpg_error_t.  Some macros also
    need an underscore prefixed internal version.
    Note that the memory allocation functions and macros (xmalloc etc.)
@@ -120,8 +120,8 @@ gpg_err_code_t _gcry_md_ctl (gcry_md_hd_t hd, int cmd,
                           void *buffer, size_t buflen);
 void _gcry_md_write (gcry_md_hd_t hd, const void *buffer, size_t length);
 unsigned char *_gcry_md_read (gcry_md_hd_t hd, int algo);
-gpg_error_t _gcry_md_extract (gcry_md_hd_t hd, int algo, void *buffer,
-                              size_t length);
+gpg_err_code_t _gcry_md_extract (gcry_md_hd_t hd, int algo, void *buffer,
+                                 size_t length);
 void _gcry_md_hash_buffer (int algo, void *digest,
                            const void *buffer, size_t length);
 gpg_err_code_t _gcry_md_hash_buffers (int algo, unsigned int flags,

commit 4a7aa30ae9f3ce798dd886c2f2d4164c43027748
Author: Jussi Kivilinna <jussi.kivilinna at iki.fi>
Date:   Sat Jul 29 14:34:23 2017 +0300

    Fix building AArch32 CE implementations when target is ARMv6 arch
    * cipher/cipher-gcm-armv8-aarch32-ce.S: Select ARMv8 architecure.
    * cipher/rijndael-armv8-aarch32-ce.S: Ditto.
    * cipher/sha1-armv8-aarch32-ce.S: Ditto.
    * cipher/sha256-armv8-aarch32-ce.S: Ditto.
    * configure.ac (gcry_cv_gcc_inline_asm_aarch32_crypto): Ditto.
    Raspbian distribution defaults to ARMv6 architecture thus 'rbit'
    instruction is not available with default compiler flags. Patch
    adds explicit architecture selection for ARMv8 to enable 'rbit'
    usage with ARMv8/AArch32-CE assembly implementations of SHA,
    GHASH and AES.
    Reported-by: Chris Horry <zerbey at gmail.com>
    Signed-off-by: Jussi Kivilinna <jussi.kivilinna at iki.fi>

diff --git a/cipher/cipher-gcm-armv8-aarch32-ce.S b/cipher/cipher-gcm-armv8-aarch32-ce.S
index b61a787..1de66a1 100644
--- a/cipher/cipher-gcm-armv8-aarch32-ce.S
+++ b/cipher/cipher-gcm-armv8-aarch32-ce.S
@@ -24,6 +24,7 @@
 .syntax unified
+.arch armv8-a
 .fpu crypto-neon-fp-armv8
diff --git a/cipher/rijndael-armv8-aarch32-ce.S b/cipher/rijndael-armv8-aarch32-ce.S
index f375f67..5c8fa3c 100644
--- a/cipher/rijndael-armv8-aarch32-ce.S
+++ b/cipher/rijndael-armv8-aarch32-ce.S
@@ -24,6 +24,7 @@
 .syntax unified
+.arch armv8-a
 .fpu crypto-neon-fp-armv8
diff --git a/cipher/sha1-armv8-aarch32-ce.S b/cipher/sha1-armv8-aarch32-ce.S
index b0bc5ff..bf2b233 100644
--- a/cipher/sha1-armv8-aarch32-ce.S
+++ b/cipher/sha1-armv8-aarch32-ce.S
@@ -24,6 +24,7 @@
     defined(HAVE_GCC_INLINE_ASM_AARCH32_CRYPTO) && defined(USE_SHA1)
 .syntax unified
+.arch armv8-a
 .fpu crypto-neon-fp-armv8
diff --git a/cipher/sha256-armv8-aarch32-ce.S b/cipher/sha256-armv8-aarch32-ce.S
index 2041a23..2b17ab1 100644
--- a/cipher/sha256-armv8-aarch32-ce.S
+++ b/cipher/sha256-armv8-aarch32-ce.S
@@ -24,6 +24,7 @@
     defined(HAVE_GCC_INLINE_ASM_AARCH32_CRYPTO) && defined(USE_SHA256)
 .syntax unified
+.arch armv8-a
 .fpu crypto-neon-fp-armv8
diff --git a/configure.ac b/configure.ac
index 27faa7f..66e7cd6 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1619,6 +1619,7 @@ AC_CACHE_CHECK([whether GCC inline assembler supports AArch32 Crypto Extension i
                 ".syntax unified\n\t"
+                ".arch armv8-a\n\t"
                 ".fpu crypto-neon-fp-armv8\n\t"


Summary of changes:
 cipher/cipher-gcm-armv8-aarch32-ce.S | 1 +
 cipher/rijndael-armv8-aarch32-ce.S   | 1 +
 cipher/sha1-armv8-aarch32-ce.S       | 1 +
 cipher/sha256-armv8-aarch32-ce.S     | 1 +
 configure.ac                         | 1 +
 src/gcrypt-int.h                     | 6 +++---
 6 files changed, 8 insertions(+), 3 deletions(-)

The GNU crypto library

Gnupg-commits mailing list
Gnupg-commits at gnupg.org

More information about the Gcrypt-devel mailing list