[openssl.org #1287] Bug in crypto/rc2/rc2speed.c
I've found a bug in crypto/rc2/rc2speed.c. It is revealed when compiling the 0.9.7 code (although this code is unchanged for 0.9.8) on MacOSX against the 10.2.8 sdk, and the result is an undefined HZ. At line crypto/rc2/rc2speed.c:105 are the preprocessor directives to define HZ if it is not already defined: #ifndef HZ #ifndef CLK_TCK #define HZ 100.0 #endif #else /* CLK_TCK */ #define HZ ((double)CLK_TCK) #endif The #endif and #else are out of order, causing HZ to be undefined if CLK_TCK is defined (as it is on the MacOS10.2.8 sdk). It could also cause a duplicate definition of HZ had already been defined. This isn't noticed on the newer versions of MacOS as CLK_TCK is not defined. It should be: #ifndef HZ #ifndef CLK_TCK #define HZ 100.0 #else /* CLK_TCK */ #define HZ ((double)CLK_TCK) #endif #endif Regards, David Litwin BigFix __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
the Sun covenant language
Hi, The Sun covenant language is deleted for the following files before, openssl/apps/ecparam.c openssl/apps/speed.c openssl/crypto/bn/bn.h openssl/crypto/bn/bntest.c openssl/crypto/ec/ec.h openssl/crypto/ec/ec2_mult.c openssl/crypto/ec/ec2_smpl.c openssl/crypto/ec/ec_curve.c openssl/crypto/ec/ec_cvt.c openssl/crypto/ec/ec_lcl.h openssl/crypto/ec/ectest.c openssl/crypto/ecdh/ecdh.h openssl/crypto/ecdh/ecdhtest.c openssl/crypto/ecdh/ech_key.c openssl/crypto/ecdh/ech_lib.c openssl/crypto/ecdh/ech_ossl.c openssl/crypto/ecdsa/ecdsatest.c openssl/crypto/engine/tb_ecdh.c openssl/ssl/s3_clnt.c openssl/ssl/s3_lib.c openssl/ssl/s3_srvr.c openssl/ssl/tls1.h (CVS COMMIT MAIL) http://cvs.openssl.org/chngview?cn=8610 the covenant language is not deleted about openssl/crypto/bn/bn_gf2m.c. Will there be the circumstances that cannot delete the covenant language? In addition, where will a pertinence point of ECC Code described in the covenant language be? Will it be the whole openssl/crypto/bn/bn_gf2m.c? --- * In addition, Sun covenants to all licensees who provide a reciprocal * covenant with respect to their own patents if any, not to sue under * current and future patent claims necessarily infringed by the making, * using, practicing, selling, offering for sale and/or otherwise * disposing of the ECC Code as delivered hereunder (or portions thereof), * provided that such covenant shall not apply: * 1) for code that a licensee deletes from the ECC Code; * 2) separates from the ECC Code; or * 3) for infringements caused by: * i) the modification of the ECC Code or * ii) the combination of the ECC Code with other software or * devices where such combination causes the infringement. --- Thanks. Takurou Saitou __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[openssl.org #1281] Solaris9: 'make test' failure in evp_test
Confirming identical results on Solaris 10 x86 patchlevel 118844_28, built with Sun C 5.8 2005/10/13. Any word or suggestions to help debug? gcc compilation does not exhibit this issue. [EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[openssl.org #1281] Solaris9: 'make test' failure in evp_test
via RT schrieb: Confirming identical results on Solaris 10 x86 patchlevel 118844_28, built with Sun C 5.8 2005/10/13. Any word or suggestions to help debug? gcc compilation does not exhibit this issue. Some info follows: - env output _=/bin/env MANPATH=/usr/man:/usr/local/man:/usr/openwin/man:/opt/csw/man SSH_TTY=/dev/pts/8 PATH=/bin:/usr/bin:/usr/sbin:/opt/csw/bin:/opt/csw/sbin:/usr/ccs/bin:/opt/SUNWspro/bin EDITOR=vi LOGNAME=sunger USER=sunger SHELL=/bin/ksh HOME=/export/home/sunger FCEDIT=vi SSH_CLIENT=172.16.10.75 1267 22 TERM=xterm PWD=/export/home/sunger/software/build/openssl-0.9.8a TZ=Europe/Berlin ENV=/export/home/sunger/.kshrc - ./config output attached as file - warning from make output attached as file - make output without warnings attached as file - make test output and now my speculations :-) 1) compiler options - during make -xtarget=ultra -xarch=v8plus is used - on http://developers.sun.com/prodtech/cc/articles/options.html SUN descripted -xtarget=ultra3 -xarch=v8plusa 2) may be it helps to use some GNU tools during make test I saw some messages about bc - may be there are some problems during make with some other SUN tools 3) compiler optimisation May be the SUN compiler will use optimisation a little bit to intensiv. May be it helps to recompile the package with lower optimisation. Unfortunatly my expirience with c-programming is to short to well interpret this error. But my favorit is No. 2 and No. 3 . Hopefully that helps. best regards, Steffen Unger -- Steffen Unger [EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[PATCH] compile without DH
the attached patch fixes the compile if no-dh is specified for Configure
or
if no-dh but not no-ecdh is specified
regards,
Peter
--
Peter Meerwald
Kaigasse 3 / 8
A-5020 Salzburg / AUSTRIA
+43-664-218 (mobile)diff -ur openssl-0.9.8a-orig/ssl/ssl_cert.c openssl-0.9.8a-work/ssl/ssl_cert.c
--- openssl-0.9.8a-orig/ssl/ssl_cert.c 2005-08-08 21:39:29.0 +0200
+++ openssl-0.9.8a-work/ssl/ssl_cert.c 2006-02-24 18:28:59.140625000 +0100
@@ -296,7 +296,7 @@
return(ret);
-#ifndef OPENSSL_NO_DH /* avoid 'unreferenced label' warning if OPENSSL_NO_DH
is defined */
+#if !defined(OPENSSL_NO_DH) || !defined(OPENSSL_NO_ECDH)
err:
#endif
#ifndef OPENSSL_NO_RSA
diff -ur openssl-0.9.8a-orig/ssl/ssltest.c openssl-0.9.8a-work/ssl/ssltest.c
--- openssl-0.9.8a-orig/ssl/ssltest.c 2005-10-02 14:26:06.0 +0200
+++ openssl-0.9.8a-work/ssl/ssltest.c 2006-02-24 18:19:10.640625000 +0100
@@ -2234,6 +2234,7 @@
dh-length = 160;
return(dh);
}
+#endif
static int do_test_cipherlist(void)
{
@@ -2287,4 +2288,3 @@
return 1;
}
-#endif
Re: [PATCH] compile without DH
Peter Meerwald wrote: the attached patch fixes the compile if no-dh is specified for Configure or if no-dh but not no-ecdh is specified ok, patch applied (0.9.8 and head). Please test a recent snapshot. Thanks, Nils __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
