Re: [ANNOUNCE] OpenSSL 0.9.6f released
On Fri, Aug 09, 2002 at 12:35:40AM +0200, Richard Levitte - VMS Whacker wrote: > -BEGIN PGP SIGNED MESSAGE- > > > OpenSSL version 0.9.6f released > === Why is the util/cygwin.sh file not in the distribution anymore? I don't find a word of that mentioned in the CHANGES file. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.6f released
On Fri, Aug 09, 2002 at 10:01:09PM +0200, Corinna Vinschen wrote: > On Fri, Aug 09, 2002 at 12:35:40AM +0200, Richard Levitte - VMS Whacker wrote: > > -BEGIN PGP SIGNED MESSAGE- > > > > > > OpenSSL version 0.9.6f released > > === > > Why is the util/cygwin.sh file not in the distribution anymore? > I don't find a word of that mentioned in the CHANGES file. I've just seen that's fixed with 0.9.6g. Thanks, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: IMPORTANT: Please try these specific snapshots
On Thu, Nov 14, 2002 at 12:02:28AM +0100, Richard Levitte - VMS Whacker wrote: > openssl-0.9.6-stable-SNAP-200211xx.tar.gz non-engine version > [...] > openssl-0.9.7-stable-SNAP-200211xx.tar.gz Hi, a few problems. The Configure script still uses the old deprecated -m486 instead of the -march=i486 option: 0.9.6-stable-SNAP-20021112: == --- Configure.orig 2002-11-14 12:14:43.0 +0100 +++ Configure 2002-11-14 12:15:00.0 +0100 @@ -477,7 +477,7 @@ my %table=( # Cygwin "Cygwin-pre1.3", "gcc:-DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O2 -m486 -Wall::(unknown)::BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}::win32", -"Cygwin", "gcc:-DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O2 -m486 -WallBN_LLONG ${x86_gcc_des} ${x86_gcc_opts}::win32:cygwin-shared:::.dll", +"Cygwin", "gcc:-DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O2 -march=i486 +-WallBN_LLONG ${x86_gcc_des} ${x86_gcc_opts}::win32:cygwin-shared:::.dll", # Ultrix from Bernhard Simon <[EMAIL PROTECTED]> "ultrix-cc","cc:-std1 -O -Olimit 1000 -DL_ENDIAN::(unknown)::", == 0.9.7-stable-SNAP-20021112: == --- Configure.orig 2002-11-14 12:20:39.0 +0100 +++ Configure 2002-11-14 12:20:52.0 +0100 @@ -520,7 +520,7 @@ my %table=( # Cygwin "Cygwin-pre1.3", "gcc:-DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O3 -m486 -Wall::(unknown):CYGWIN32::BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}::win32", -"Cygwin", "gcc:-DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O3 -m486 -Wall:::CYGWIN32::BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_out_asm}:win32:cygwin-shared:::.dll", +"Cygwin", "gcc:-DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O3 -march=i486 +-Wall:::CYGWIN32::BN_LLONG ${x86_gcc_des} +${x86_gcc_opts}:${x86_out_asm}:win32:cygwin-shared:::.dll", # DJGPP "DJGPP", "gcc:-I/dev/env/DJDIR/watt32/inc -DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O2 -Wall:::MSDOS:-L/dev/env/DJDIR/watt32/lib -lwatt:BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}::", == The Cygwin build script in the util subdir suffers from a missing `make depend': Both openssl versions: == --- util/cygwin.sh.orig 2002-11-14 12:16:03.0 +0100 +++ util/cygwin.sh 2002-11-14 12:07:51.0 +0100 @@ -96,6 +96,8 @@ fi get_openssl_version +make depend || exit 1 + make || exit 1 base_install == Other than that, the util/cygwin.sh file still has wrong permissions which prevents building the official Cygwin version without having to chmod +x the script first. Even though `make depend' has been called, the following happens in the 0.9.6-stable-SNAP-20021112 build: making all in test... make[1]: Entering directory `/src/openssl-0.9.6-stable-SNAP-20021112/test' gcc -I../include -DTHREADS -DDSO_WIN32 -DNO_IDEA -DNO_RC5 -DNO_MDC2 -DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O2 -march=i486 -Wall -c -o bntest.o bntest.c gcc -o bntest -I../include -DTHREADS -DDSO_WIN32 -DNO_IDEA -DNO_RC5 -DNO_MDC2 -DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O2 -march=i486 -Wall bntest.o -L.. -lcrypto make[1]: *** No rule to make target `ideatest.o', needed by `ideatest'. Stop. make[1]: Leaving directory `/src/openssl-0.9.6-stable-SNAP-20021112/test' make: *** [sub_all] Error 1 This doesn't happen in 0.9.7-stable-SNAP-20021112. Sorry, I have no patch for this one. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:vinschen@;redhat.com __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: IMPORTANT: Please try these specific snapshots
On Thu, Nov 14, 2002 at 01:32:24PM +0100, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Thu, 14 Nov 2002 12:53:03 >+0100, Corinna Vinschen <[EMAIL PROTECTED]> said: > > vinschen> The Configure script still uses the old deprecated -m486 instead of the > vinschen> -march=i486 option: > > Patch applied. Sorry to say that but you didn't apply them correctly. It's not -march=486 but -march=i486. Please note the additional 'i'. Still testing, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:vinschen@;redhat.com __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: IMPORTANT: Please try these specific snapshots
On Fri, Nov 15, 2002 at 11:28:04AM +0100, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Fri, 15 Nov 2002 11:06:21 >+0100, Corinna Vinschen <[EMAIL PROTECTED]> said: > > vinschen> Sorry to say that but you didn't apply them correctly. It's not > vinschen> > vinschen> -march=486 > vinschen> > vinschen> but > vinschen> > vinschen> -march=i486. > > OK, fixed and committed. > > Sorry about that. No worries. I've just tested building both, 0.9.6h and 0.9.7 snapshots as of today. No problems with 0.9.6h left. I checked also running ssh/sshd with the new DLLs and it seems running fine. With 0.9.7 on the other hand I had build problems, which are probably also given in 0.9.6, just not visible in my build environment. Let me explain from the 0.9.7 point of view. First, there's a build problem which only affects Windows systems due to the way DLLs are searched for. In contrast to Unixy systems, shared libs are search in $PATH. Therefore I propose the following patch to test/Makefile: --- test/Makefile.ssl.orig 2002-11-15 12:20:59.0 +0100 +++ test/Makefile.ssl 2002-11-15 12:15:52.0 +0100 @@ -118,8 +118,8 @@ tests: exe apps $(TESTS) apps: @(cd ..; $(MAKE) DIRS=apps all) -SET_SO_PATHS=LIBPATH="`cd ..; pwd`"; LD_LIBRARY_PATH=$$LIBPATH; DYLD_LIBRARY_PATH=$$LIBPATH; SHLIB_PATH=$$LIBPATH; \ - export LD_LIBRARY_PATH DYLD_LIBRARY_PATH SHLIB_PATH LIBPATH +SET_SO_PATHS=LIBPATH="`cd ..; pwd`"; LD_LIBRARY_PATH=$$LIBPATH; +DYLD_LIBRARY_PATH=$$LIBPATH; SHLIB_PATH=$$LIBPATH; PATH="$$LIBPATH:$$PATH" \ + export LD_LIBRARY_PATH DYLD_LIBRARY_PATH SHLIB_PATH LIBPATH PATH alltests: \ test_des test_idea test_sha test_md4 test_md5 test_hmac \ But there's still another problem left then. Three applications die with access violations, openssl, rc4test and bltest. What I did then was, to build everything again. But before, I removed my /usr/include/openssl directory which is (obviously) the directory present from a 0.9.6 installation. Now everything worked fine. That means, the problem here is that gcc (JFYI, Cygwin is now on version 3.2) searches the system include paths before searching the paths given with -I on the command line, iff the includes are bracketed using <...> instead of "...". I can see two possible solutions to circumvent the problem: Either prepend -I. -I- to all other -I options like gcc -I. -I- -I.. -I../include according to the gcc documentation: `-I-' Split the include path. Any directories specified with `-I' options before `-I-' are searched only for headers requested with `#include "FILE"'; they are not searched for `#include '. If additional directories are specified with `-I' options after the `-I-', those directories are searched for all `#include' directives. In addition, `-I-' inhibits the use of the directory of the current file directory as the first search directory for `#include "FILE"'. or change all #include to #include "openssl/..." I don't see another solution, currently. HTH, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:vinschen@;redhat.com __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: IMPORTANT: Please try these specific snapshots
Hi Richard, On Fri, Nov 15, 2002 at 02:12:11PM +0100, Richard Levitte - VMS Whacker wrote: > vinschen> -SET_SO_PATHS=LIBPATH="`cd ..; pwd`"; LD_LIBRARY_PATH=$$LIBPATH; >DYLD_LIBRARY_PATH=$$LIBPATH; SHLIB_PATH=$$LIBPATH; \ > vinschen> - export LD_LIBRARY_PATH DYLD_LIBRARY_PATH SHLIB_PATH LIBPATH > vinschen> +SET_SO_PATHS=LIBPATH="`cd ..; pwd`"; LD_LIBRARY_PATH=$$LIBPATH; >DYLD_LIBRARY_PATH=$$LIBPATH; SHLIB_PATH=$$LIBPATH; PATH="$$LIBPATH:$$PATH" \ > vinschen> + export LD_LIBRARY_PATH DYLD_LIBRARY_PATH SHLIB_PATH LIBPATH >PATH > > You're missing a ; just before export... You're right. I didn't miss it due to the way variable assignments work in the shell. > However, since you $PATH might not be properly set (because of the > missing ;), it may be that running the tests actually use 0.9.6 DLLs. No, the missing ; doesn't hurt: A=b export A is equivalent to A=b; export A > Had it been an include file error, you would probably not have been > able to build... I'm still pretty confident that it's an include file problem. Note that removing the /usr/include/openssl directory was the *only* difference between two full make runs. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:vinschen@;redhat.com __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: IMPORTANT: Please try these specific snapshots
On Fri, Nov 15, 2002 at 02:49:17PM +0100, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Fri, 15 Nov 2002 13:23:57 >+0100, Corinna Vinschen <[EMAIL PROTECTED]> said: > > vinschen> First, there's a build problem which only affects Windows systems due > vinschen> to the way DLLs are searched for. In contrast to Unixy systems, shared > vinschen> libs are search in $PATH. Therefore I propose the following patch to > vinschen> test/Makefile: > > Can you check if the attached patch helps? I'm (very positively) surprised. This patch also heals the problem with the crashing testsuite applications. I had even restored the /usr/include/openssl directory. So my analyzis was actually incorrect. I'm not disappointed about that fact, though :-) However, since the introduction of a new version collides with the DLLs of an already installed 0.9.6 version, there's a need to change the names of the produced DLLs (but not the .dll.a files) from now on. So, from 0.9.7 on, the DLLs should be named like 'cygssl-0.9.7.dll' and 'cygcrypto-0.9.7.dll' while the static and link libs still should be named libssl.a, libssl.dll.a, libcrypto.a and libcrypto.dll.a. May I propose the attached patch to 0.9.7? There will probably be another change needed for util/cygwin.sh but I will have to have another look into that issue first. Thanks, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:vinschen@;redhat.com --- Makefile.ssl.orig 2002-11-15 16:17:21.0 +0100 +++ Makefile.ssl2002-11-15 16:17:17.0 +0100 @@ -244,7 +244,7 @@ clean-shared: fi; \ ( set -x; rm -f lib$$i$(SHLIB_EXT) ); \ if [ "$(PLATFORM)" = "Cygwin" ]; then \ - ( set -x; rm -f cyg$$i$(SHLIB_EXT) lib$$i$(SHLIB_EXT).a ); \ + ( set -x; rm -f cyg$$i-$(SHLIB_VERSION_NUMBER)$(SHLIB_EXT) +lib$$i$(SHLIB_EXT).a ); \ fi; \ done @@ -294,7 +294,7 @@ do_darwin-shared: do_cygwin-shared: libs='-L. ${SHLIBDEPS}'; for i in ${SHLIBDIRS}; do \ - ( set -x; ${CC} -shared -o cyg$$i.dll \ + ( set -x; ${CC} -shared -o cyg$$i-${SHLIB_VERSION_NUMBER}.dll \ -Wl,-Bsymbolic \ -Wl,--whole-archive lib$$i.a \ -Wl,--out-implib,lib$$i.dll.a \ @@ -745,7 +745,7 @@ install: all install_docs chmod 555 $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new; \ mv $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i; \ else \ - c=`echo $$i | sed 's/^lib/cyg/'`; \ + c=`echo $$i | sed +'s/^lib\(.*\)/cyg\1-$(SHLIB_VERSION_NUMBER)/'`; \ cp $$c $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new; \ chmod 755 $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new; \ mv $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c; \
Re: IMPORTANT: Please try these specific snapshots
On Fri, Nov 15, 2002 at 04:35:53PM +0100, Richard Levitte - VMS Whacker wrote: > Are symlinks possible to do with Cygwin? If they are, I propose we do > it the same way as under Unix (where '.so.0.9.7' ends the file name. > In the Cygwin case, it would end with '.0.9.7.dll' instead...). That > way, it won't be a special case for Cygwin, and the library version > matching will hopefully work well enough. Also, then one can have > symlinks called cygcrypto.dll pointing at the correct library du jour. That's exactly how it can't work. The DLL search algorithm is inside of Windows and it doesn't work using symlinks (resp. shortcuts under Windows) unfortunately. Corinna P.S.: It's ok to send mail to the mailing list only since I'm subscribed. Thanks. -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:vinschen@;redhat.com __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: IMPORTANT: Please try these specific snapshots
On Fri, Nov 15, 2002 at 05:00:04PM +0100, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Fri, 15 Nov 2002 16:46:40 >+0100, Corinna Vinschen <[EMAIL PROTECTED]> said: > > vinschen> That's exactly how it can't work. The DLL search algorithm is inside > vinschen> of Windows and it doesn't work using symlinks (resp. shortcuts under > vinschen> Windows) unfortunately. > > OK, another question: does Windows DLLs have any version information > inside that's used for comparison, or is it just informative? AFAIK DLLs are only picked up by name when searching for resolving symbols. Actually, we tried to find a workaround in Cygwin already long ago but Windows is not that cooperative here. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:vinschen@;redhat.com __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: IMPORTANT: Please try these specific snapshots
On Fri, Nov 15, 2002 at 05:45:34PM +0100, Richard Levitte - VMS Whacker wrote: > OK, I'm applying your change of the DLL names. I've transfered that > naming to 0.9.8-dev (which has a different mechanism for building > shared libraries), do you have the possibility to check that the next > 0.9.8-dev snapshot works as well (the name is > "openssl-SNAL-2002mmdd.tar.gz")? No need to hurry, though... That's good. I will not manage to do that the next days. Any chance you could remind me end of next week? Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:vinschen@;redhat.com __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.6h released
On Fri, Dec 06, 2002 at 01:32:50AM +0100, Richard Levitte - VMS Whacker wrote: > -BEGIN PGP SIGNED MESSAGE- > > > OpenSSL version 0.9.6h released > === Hi, I just found that the openssl-0.9.6h.tar.gz archive has the following version definition in include/openssl/opensslv.h: #define OPENSSL_VERSION_NUMBER 0x00906080L This disallows usage of the new openssl version with e. g. an ssh linked against 0.9.6g. Shouldn't that be #define OPENSSL_VERSION_NUMBER 0x0090608fL ? Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Make depend fails for gcc (0.9.6h)
On Sun, Dec 15, 2002 at 01:31:52AM -0800, Doug Kaufman wrote: > I just compiled openssl-0.9.6h (from CVS) with Cygwin. I configured > with no-idea no-rc5 no-mdc2 no-shared. Make depend fails because > Cygwin has no program named "makedepend", which is called by > util/domd. I borrowed the domd from 0.9.7, which has a provision for > gcc being the makedepend program, and the rest of the compilation went > fine, including "make test" and "make install". I am not sure I see > a simple fix, since in 0.9.7 the Configure script determines if the > makedepend program should be gcc or not. Do we need to put this in the > Configure script and modify Makefile.org and util/domd accordingly? > > I presume that this applies to any gcc compilation where there is no > separate "makedepend" program. >Doug `makedepend' is typically part of the XFree86 package, in Cygwin as well as on Linux systems. Just install the Cygwin XFree86 package. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Make depend fails for gcc (0.9.6h)
On Sun, Dec 15, 2002 at 12:58:18PM +0100, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Sun, 15 Dec 2002 12:02:34 >+0100, Corinna Vinschen <[EMAIL PROTECTED]> said: > > vinschen> `makedepend' is typically part of the XFree86 package, in Cygwin as > vinschen> well as on Linux systems. Just install the Cygwin XFree86 package. > > I must admit it would be a little silly to install XFree86 just to get > hold of makedepend. > > (and it's beyond me why makedepend still hasn't become a separate > package, but that's not your fault) You can't rely on it's existance and that's not a Cygwin problem. It's just that most today's Linux/Unix installations have X installed by default so you'll never see that problem. Just curious: What about Apple's OS X? It has no X so it probably has no makedepend, too. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.7 released
On Tue, Dec 31, 2002 at 01:34:07AM +0100, Richard Levitte - VMS Whacker wrote: > -BEGIN PGP SIGNED MESSAGE- > > > OpenSSL version 0.9.7 released > === Hi, sorry for the late response but I didn't manage to test earlier :-( I have two patches attached which are necessary to build 0.9.7 for the Cygwin release. The Makefile.org patch is necessary to build the Cygwin DLLs at all. It fixes a wrong sed expression. The patch to util/cygwin.sh is required by the naming scheme used in the Cygwin net release. However, currently I'm unfortunately unable to release a Cygwin net version of 0.9.7 due to a linker problem, which results in dropped symbols in the link stub library. The most prominent dropped symbol is RC4. Building OpenSSH with this libs results in ssh and sshd crashing immediately :-((( I'm still in the process of trying to find out why that happens. Btw., does anybody have an idea where a symbol ".RC4_end" could come from? Corinna --- openssl-0.9.7.orig/Makefile.org 2002-12-30 00:03:16.0 +0100 +++ openssl-0.9.7/Makefile.org 2003-01-07 19:25:50.0 +0100 @@ -749,7 +749,7 @@ install: all install_docs chmod 555 $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new; \ mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i; \ else \ - c=`echo $$i | sed 's/^lib\(.*\)/cyg\1-$(SHLIB_VERSION_NUMBER)/'`; \ + c=`echo $$i | sed +'s/^lib\(.*\).dll/cyg\1-$(SHLIB_VERSION_NUMBER).dll/'`; \ cp $$c $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new; \ chmod 755 $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new; \ mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c; \ --- openssl-0.9.7.orig/util/cygwin.sh 2002-11-14 13:39:51.0 +0100 +++ openssl-0.9.7/util/cygwin.sh2003-01-07 19:25:15.0 +0100 @@ -114,13 +114,13 @@ strip usr/bin/*.exe usr/bin/*.dll # Runtime package find etc usr/bin usr/doc usr/ssl/certs usr/ssl/man/man[157] usr/ssl/misc \ usr/ssl/openssl.cnf usr/ssl/private -empty -o \! -type d | -tar cjfT openssl-${VERSION}-${SUBVERSION}.tar.bz2 - +tar cjfT openssl2-${VERSION}-${SUBVERSION}.tar.bz2 - # Development package find usr/include usr/lib usr/ssl/man/man3 -empty -o \! -type d | -tar cjfT openssl-devel-${VERSION}-${SUBVERSION}.tar.bz2 - +tar cjfT openssl2-devel-${VERSION}-${SUBVERSION}.tar.bz2 - -ls -l openssl-${VERSION}-${SUBVERSION}.tar.bz2 -ls -l openssl-devel-${VERSION}-${SUBVERSION}.tar.bz2 +ls -l openssl2-${VERSION}-${SUBVERSION}.tar.bz2 +ls -l openssl2-devel-${VERSION}-${SUBVERSION}.tar.bz2 cleanup -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.7 released
On Thu, Jan 09, 2003 at 05:10:55PM +0100, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Thu, 9 Jan 2003 17:05:12 >+0100, Corinna Vinschen <[EMAIL PROTECTED]> said: > > vinschen> Btw., does anybody have an idea where a symbol ".RC4_end" could come from? > > crypto/rc4/asm/rx86unix.cpp in my case (on Linux)... > > I've applied your patches, and will commit them promptly. Please > check the next snapshot. Thanks, I just found out that the symbol comes from rx86unix.cpp in my case, too. What I don't understand is the following. crypto/rc4/Makefile.ssl contains the following: RC4_ENC=rc4_enc.o # or use #RC4_ENC=asm/rx86-elf.o #RC4_ENC=asm/rx86-out.o #RC4_ENC=asm/rx86-sol.o #RC4_ENC=asm/rx86bdsi.o Even though it's supposed to build rc4_enc.o to get the RC4 function(s), it builds asm/rx86-out.o instead. However, I found out, that the linker problem disappears if I build the libraries using rc4_enc.o instead of asm/rx86-out.o. Frankly, I have NO idea why! Until I found a solution for that linker problem (which is a linker bug, apparently) I'd like to build the Cygwin version using rc4_enc.o. How can I do that most cleanly? Thanks in advance, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.7 released
On Thu, Jan 09, 2003 at 05:34:39PM +0100, Andy Polyakov wrote: > > Btw., does anybody have an idea where a symbol ".RC4_end" could come from? > > I should have an idea... To start which I didn't expect to see .RC4_end > symbol be visible at all. Assembler should have hidden it as names > starting with . are supposed to be internal. I can double-check it later > today. As for dropped RC4 symbol. I'm not really following... Did 'make > test' pass? No. The problem is that when creating the link stub, the _RC4 symbol is treated as a data or bss symbol, not a text symbol. Linking against that link stub creates an illegal reference into the DLL ==> rc4test crashes. The same happens with _DES_ede3_cbc_encrypt and _DES_ncbc_encrypt, resulting in destest crashing. This all doesn't happen when linking against the static crypto lib, btw., it's justy a problem of the dynamic linking process. > > c=`echo $$i | sed 's/^lib\(.*\).dll/cyg\1-$(SHLIB_VERSION_NUMBER).dll/'`; > ^ this of course works, but it probably > should be \., not just dot:-) Oops, you're right. Thanks, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.7 released
On Thu, Jan 09, 2003 at 05:53:57PM +0100, Andy Polyakov wrote: > As for .*_end symbols. Apparently there're more... Any particular reason > why are you complaining just about .RC4_end? No. Just the one I found first due to the ssh crash. > > Until I found a solution for that linker problem (which is a linker bug, > > apparently) I'd like to build the Cygwin version using rc4_enc.o. How > > can I do that most cleanly? > > By fixing rx86-out.o:-) A. That doesn't help w/o details. I don't see anything wrong with that file. I have no idea why our linker treats RC4 from that file (and other functions from assembler files, apparently) as data. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.7 released
On Thu, Jan 09, 2003 at 06:25:57PM +0100, Andy Polyakov wrote: > I meant "be patient:-)" As for details, you should have provided more of > them:-) At least that you've configured with shared flag and that 'make > test' fails. In other words something we can *easily* reproduce. Cheers. Uhm... I thought that was clear from the context. I was talking about building DLLs already in my first posting and my fix to Makefile.org was related to that. And I don't think it's openssl's fault. It's very likely a linker problem when generating the DLL and link stub from the static lib (which is working fine even with the assembler code). I just wasn't able to find out why the linker fails so far. I'm going to ask some linker experts... Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.7 released
On Thu, Jan 09, 2003 at 06:32:05PM +0100, Andy Polyakov wrote: > > > > > Until I found a solution for that linker problem (which is a linker bug, > > > > > apparently) I'd like to build the Cygwin version using rc4_enc.o. How > > > > > can I do that most cleanly? > > > > > > > > By fixing rx86-out.o:-) A. > > Did it ever work? Assembler support in cygwin-shared build that is? BN I can't tell. I never saw that problem before and in 0.9.6h all tests pass. Sorry, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.7 released
On Thu, Jan 09, 2003 at 10:35:16PM +0100, Corinna Vinschen wrote: > The "2" is just sort of a counter which is needed to allow more than > [...] Hi, I have chosen another packaging method: Version 0.9.6h will consist only of the needed DLLs to link applications build against it. The package will be called openssl096-0.9.6h-1.tar.bz2 The new 0.9.7 version will become the openssl package again, as the 0.9.6 version before. So the new packages are still just called "openssl": openssl-0.9.7-1.tar.bz2 openssl-devel-0.9.7-1.tar.bz2 openssl-0.9.7-1-src.tar.bz2 That means, the patch to util/cygwin.sh can be reverted to create the normal openssh packages again instead of openssh2. Is that ok with you? Would you revert the cygwin.sh patch then? Thanks in advance, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.7 released
On Thu, Jan 09, 2003 at 08:58:29PM +0100, Andy Polyakov wrote: > I found the documentation for .def ... .endef, but it doesn't say a > thing about the values of .scl and .type. So I just copied from an .s > file generated by compiler. Verify the values with your people. This is from the gas info pages: --- SNIP --- `.scl CLASS' Set the storage-class value for a symbol. This directive may only be used inside a `.def'/`.endef' pair. Storage class may flag whether a symbol is static or external, or it may record further symbolic debugging information. The `.scl' directive is primarily associated with COFF output; when configured to generate `b.out' output format, `as' accepts this directive but ignores it. `.type INT' (COFF version) == This directive, permitted only within `.def'/`.endef' pairs, records the integer INT as the type attribute of a symbol table entry. `.type' is associated only with COFF format output; when `as' is configured for `b.out' output, it accepts this directive but ignores it. --- SNAP --- The value needed for .scl is C_EXT = 2 and the value for .type is (DT_FCN << N_BTSHFT) = 32 so that should be ok. > Oh! BTW, I probably have to object against calling the package openssl2, > openssl097 is probably more appropriate. A. The "2" is just sort of a counter which is needed to allow more than one package installed at the same time. E.g. we have two concurrent versions of libreadline called libreadline4 and libreadline5. This is *not* related to the official version number of the packages, which in case of the libreadline packages are 4.1 and 4.3. This official version number is part of the full package name, e.g. in case of libreadline currently: libreadline4-4.1-2.tar.bz2 libreadline5-4.3-2.tar.bz2 so in case of openssl this will become openssl-0.9.6h-2.tar.bz2 openssl2-0.9.7-1.tar.bz2 The next version, 0.9.8, would become openssl3-0.9.8-1.tar.bz2 The names of the real DLLs don't contain this counter: cygcrypto.dll == 0.9.6x, no version number for backward compatibility cygcrypto.dll-0.9.7 cygcrypto.dll-0.9.8 > *** crypto/perlasm/x86asm.pl.orig Thu Jul 18 14:38:03 2002 > --- crypto/perlasm/x86asm.plThu Jan 9 20:41:52 2003 > *** > *** 91,97 > #undef SIZE > #undef TYPE > #define SIZE(a,b) > ! #define TYPE(a,b) > #endif /* __CYGWIN || __DJGPP */ > #endif > > --- 91,97 > #undef SIZE > #undef TYPE > #define SIZE(a,b) > ! #define TYPE(a,b) .def a; .scl 2; .type 32; .endef > #endif /* __CYGWIN || __DJGPP */ > #endif Builds fine, all tests pass, ssh works again with this patch. Thanks very much! Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.7 released
On Fri, Jan 10, 2003 at 07:27:35PM +0100, Corinna Vinschen wrote: > > [...] > Hi, I have chosen another packaging method: > > Version 0.9.6h will consist only of the needed DLLs to link applications > build against it. > > The package will be called > > openssl096-0.9.6h-1.tar.bz2 > > The new 0.9.7 version will become the openssl package again, as the > 0.9.6 version before. So the new packages are still just called > "openssl": > > openssl-0.9.7-1.tar.bz2 > openssl-devel-0.9.7-1.tar.bz2 > openssl-0.9.7-1-src.tar.bz2 > > That means, the patch to util/cygwin.sh can be reverted to create > the normal openssh packages again instead of openssh2. > > Is that ok with you? Would you revert the cygwin.sh patch then? Hi, in the meantime I released Cygwin's openssl using the above packaging scheme. Would you mind to revert the cygwin.sh changes? The following patch should do it: Index: util/cygwin.sh === RCS file: /home/cvs/cvsroot/src/openssl/util/cygwin.sh,v retrieving revision 1.3 diff -u -p -r1.3 cygwin.sh --- util/cygwin.sh 2003/01/09 16:15:46 1.3 +++ util/cygwin.sh 2003/01/13 14:35:19 @@ -114,13 +114,13 @@ strip usr/bin/*.exe usr/bin/*.dll # Runtime package find etc usr/bin usr/doc usr/ssl/certs usr/ssl/man/man[157] usr/ssl/misc \ usr/ssl/openssl.cnf usr/ssl/private -empty -o \! -type d | -tar cjfT openssl2-${VERSION}-${SUBVERSION}.tar.bz2 - +tar cjfT openssl-${VERSION}-${SUBVERSION}.tar.bz2 - # Development package find usr/include usr/lib usr/ssl/man/man3 -empty -o \! -type d | -tar cjfT openssl2-devel-${VERSION}-${SUBVERSION}.tar.bz2 - +tar cjfT openssl-devel-${VERSION}-${SUBVERSION}.tar.bz2 - -ls -l openssl2-${VERSION}-${SUBVERSION}.tar.bz2 -ls -l openssl2-devel-${VERSION}-${SUBVERSION}.tar.bz2 +ls -l openssl-${VERSION}-${SUBVERSION}.tar.bz2 +ls -l openssl-devel-${VERSION}-${SUBVERSION}.tar.bz2 cleanup Thanks, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: IMPORTANT: please test snapshot openssl-0.9.7-stable-SNAP-20030214.tar.gz
On Fri, Feb 14, 2003 at 06:08:34PM +0100, Bodo Moeller wrote: > Please test snapshot openssl-0.9.7-stable-SNAP-20030214.tar.gz > (or later), which will be available today around 8 p.m. GMT at > ftp://ftp.openssl.org/snapshot;type=d >. openssl-0.9.7-stable-SNAP-20030216 on current Cygwin 1.3.20. Building and installing shared and static libs using the utils/cygwin.sh script (but with idea, rc5 and mdc2) works fine. All tests pass. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ADVISORY] Timing Attack on OpenSSL
Hi, is it recommended to apply the below patch to 0.9.6i as well? We're still releasing both versions, 0.9.6i and 0.9.7a in the Cygwin net distro. Corinna On Mon, Mar 17, 2003 at 08:47:01AM +, Ben Laurie wrote: > I expect a release to follow shortly. > > -- > http://www.apache-ssl.org/ben.html http://www.thebunker.net/ > > "There is no limit to what a man can do or how far he can go if he > doesn't mind who gets the credit." - Robert Woodruff > OpenSSL v0.9.7a and 0.9.6i vulnerability > > [...] > We strongly advise upgrading OpenSSL in all cases, as a precaution. > Index: crypto/rsa/rsa_eay.c > === > RCS file: /e/openssl/cvs/openssl/crypto/rsa/rsa_eay.c,v > retrieving revision 1.28.2.3 > diff -u -r1.28.2.3 rsa_eay.c > --- crypto/rsa/rsa_eay.c 30 Jan 2003 17:37:46 - 1.28.2.3 > +++ crypto/rsa/rsa_eay.c 16 Mar 2003 10:34:13 - > @@ -195,6 +195,25 @@ > return(r); > } > > +static int rsa_eay_blinding(RSA *rsa, BN_CTX *ctx) > + { > + int ret = 1; > + CRYPTO_w_lock(CRYPTO_LOCK_RSA); > + /* Check again inside the lock - the macro's check is racey */ > + if(rsa->blinding == NULL) > + ret = RSA_blinding_on(rsa, ctx); > + CRYPTO_w_unlock(CRYPTO_LOCK_RSA); > + return ret; > + } > + > +#define BLINDING_HELPER(rsa, ctx, err_instr) \ > + do { \ > + if(((rsa)->flags & RSA_FLAG_BLINDING) && \ > + ((rsa)->blinding == NULL) && \ > + !rsa_eay_blinding(rsa, ctx)) \ > + err_instr \ > + } while(0) > + > /* signing */ > static int RSA_eay_private_encrypt(int flen, const unsigned char *from, >unsigned char *to, RSA *rsa, int padding) > @@ -239,8 +258,8 @@ > goto err; > } > > - if ((rsa->flags & RSA_FLAG_BLINDING) && (rsa->blinding == NULL)) > - RSA_blinding_on(rsa,ctx); > + BLINDING_HELPER(rsa, ctx, goto err;); > + > if (rsa->flags & RSA_FLAG_BLINDING) > if (!BN_BLINDING_convert(&f,rsa->blinding,ctx)) goto err; > > @@ -318,8 +337,8 @@ > goto err; > } > > - if ((rsa->flags & RSA_FLAG_BLINDING) && (rsa->blinding == NULL)) > - RSA_blinding_on(rsa,ctx); > + BLINDING_HELPER(rsa, ctx, goto err;); > + > if (rsa->flags & RSA_FLAG_BLINDING) > if (!BN_BLINDING_convert(&f,rsa->blinding,ctx)) goto err; > > Index: crypto/rsa/rsa_lib.c > === > RCS file: /e/openssl/cvs/openssl/crypto/rsa/rsa_lib.c,v > retrieving revision 1.30.2.2 > diff -u -r1.30.2.2 rsa_lib.c > --- crypto/rsa/rsa_lib.c 30 Jan 2003 17:37:46 - 1.30.2.2 > +++ crypto/rsa/rsa_lib.c 16 Mar 2003 10:34:13 - > @@ -72,7 +72,13 @@ > > RSA *RSA_new(void) > { > - return(RSA_new_method(NULL)); > + RSA *r=RSA_new_method(NULL); > + > +#ifndef OPENSSL_NO_FORCE_RSA_BLINDING > + r->flags|=RSA_FLAG_BLINDING; > +#endif > + > + return r; > } > > void RSA_set_default_method(const RSA_METHOD *meth) -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
[BUG?] 0.9.7c Makefile sets weird permissions
Hi, I just tried to build 0.9.7c on Cygwin. I got an error message when I created the Cygwin package. The reason was that the package creation script failed to remove the usr/lib/pkgconfig directory after creating the package files. What happened is that the top level Makefile creates a directory usr/lib/pkgconfig and then after copying the openssl.pc file into it it sets the permissions of this directory to 644 (line 837). This results in a error message in the util/cygwin.sh script. Shouldn't line 837 set the permissions of the copied openssl.pc file instead of the permissions of the directory? chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/lib/pkgconfig/openssl.pc instead of chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/lib/pkgconfig Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Win32 compiles under cygwin
On May 13 00:03, Steven Reddie wrote: > Yes, I was surprised too. I don't know if it's because Cygwin's ar has been > made compatible with Microsoft's lib format, or if the two formats have > always been the compatible. They always have been compatible, being PE/COFF format. An upcoming change in gcc is to produce DWARF-2 debugging sections also in PE/COFF under Cygwin. LINK will very likely not be able to deal with them, but I suspect they won't be generated with -mno-cygwin option. Corinna -- Corinna Vinschen Cygwin Co-Project Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Win32 compiles under cygwin
On May 10 17:53, Mark Jaffe wrote: > I have one other issue I need resolution on: when I run the make file > under cygwin, the resulting libraries are exactly what I get on unix: > libssl.a and libcrypto.a. What I want to know is how do I get > ssleay32.dll and libeay32.dll? These are required to link m2crypto on > Win32. Just to add a data point, what you get when compiling under Cygwin are 6 files: libcrypto.a libcrypto.dll.a cygcrypto-0.9.7.dll libssl.a libssl.dll.a cygssl-0.9.7.dll The *.a files are static libs, *.dll.a are the import libs used to link against the DLLs. Both DLLs expect a cygwin1.dll to be loadable at runtime. Corinna -- Corinna Vinschen Cygwin Co-Project Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
[PATCH] util/cygwin.sh: Install docs under /usr/share/doc
Hi, would some kind person from the core developer team mind to apply the following patch to the util/cygwin.sh file? A while ago, Cygwin moved from installing documentation under /usr/doc to /usr/share/doc. The below patch accomodates that. This patch should go into 0.9.7 and into mainline. Thanks in advance, Corinna Index: util/cygwin.sh === RCS file: /home/cvs/cvsroot/src/openssl/util/cygwin.sh,v retrieving revision 1.4 diff -p -u -r1.4 cygwin.sh --- util/cygwin.sh 13 Jan 2003 14:47:52 - 1.4 +++ util/cygwin.sh 28 Sep 2004 09:55:16 - @@ -39,7 +39,7 @@ function base_install() function doc_install() { - DOC_DIR=${INSTALL_PREFIX}/usr/doc/openssl + DOC_DIR=${INSTALL_PREFIX}/usr/share/doc/openssl mkdir -p ${DOC_DIR} cp CHANGES CHANGES.SSLeay INSTALL LICENSE NEWS README ${DOC_DIR} @@ -49,7 +49,7 @@ function doc_install() function create_cygwin_readme() { - README_DIR=${INSTALL_PREFIX}/usr/doc/Cygwin + README_DIR=${INSTALL_PREFIX}/usr/share/doc/Cygwin README_FILE=${README_DIR}/openssl-${VERSION}.README mkdir -p ${README_DIR} @@ -112,7 +112,7 @@ cd ${INSTALL_PREFIX} strip usr/bin/*.exe usr/bin/*.dll # Runtime package -find etc usr/bin usr/doc usr/ssl/certs usr/ssl/man/man[157] usr/ssl/misc \ +find etc usr/bin usr/share/doc usr/ssl/certs usr/ssl/man/man[157] usr/ssl/misc \ usr/ssl/openssl.cnf usr/ssl/private -empty -o \! -type d | tar cjfT openssl-${VERSION}-${SUBVERSION}.tar.bz2 - # Development package -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [PATCH] util/cygwin.sh: Install docs under /usr/share/doc
On Sep 28 14:03, Goetz Babin-Ebell wrote: > Hello Richard, > > Richard Levitte - VMS Whacker wrote: > >In message <[EMAIL PROTECTED]> on Tue, 28 Sep 2004 > >11:55:56 +0200, Corinna Vinschen <[EMAIL PROTECTED]> said: > > > >vinschen> would some kind person from the core developer team mind to > >vinschen> apply the following patch to the util/cygwin.sh file? A > >vinschen> while ago, Cygwin moved from installing documentation under > >vinschen> /usr/doc to /usr/share/doc. The below patch accomodates > >vinschen> that. This patch should go into 0.9.7 and into mainline. > > > >Applied in all active branches. > > This will break openssl with older version of cygwin. No, this will break nothing. When installing a newer version of openssl the docs will just be installed in /usr/share/doc instead of /usr/doc, that's all. There's no moving of sensitive directories like /usr/ssl involved. When installing the official way (using Cygwin's setup.exe), the same files will just get installed in the right place. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: OpenSSL 0.9.7f released
On Mar 22 22:38, Dr. Stephen Henson wrote: > OpenSSL version 0.9.7f released Unfortunately I can't build this version for Cygwin: gcc -shared -Wl,--image-base,0xFE0 -o cygcrypto-0.9.7.dll \ -Wl,-Bsymbolic -Wl,--whole-archive libcrypto.a \ -Wl,--out-implib,libcrypto.dll.a -Wl,--no-whole-archive -L. libcrypto.a(rand_unix.o)(.text+0x0):rand_unix.c: \ multiple definition of `_RAND_poll' libcrypto.a(rand_win.o)(.text+0x0):rand_win.c: first defined here The reason is that somebody added Cygwin between 0.9.7e and 0.9.7f to the list of sytems which should build crypto/rand/rand_win.c. That's not correct. Cygwin is a Unix systems on top of Windows and it provides its own /dev/random and /dev/urandom for a long time already, so it should be using crypto/rand/rand_unix.c. /dev/random and /dev/urandom on Cygwin are using the underlying Windows Crypto provider to generate random numbers. So, please apply the following patch. I did that in the version of OpenSSL 0.9.7f released as part of the Cygwin net distribution. Thanks, Corinna --- crypto/rand/rand_win.c.ORIG 2005-03-23 11:09:10.493307500 +0100 +++ crypto/rand/rand_win.c 2005-03-23 11:09:16.353643500 +0100 @@ -113,7 +113,7 @@ #include #include "rand_lcl.h" -#if defined(OPENSSL_SYS_WINDOWS) || defined(OPENSSL_SYS_WIN32) || defined(OPENSSL_SYS_WIN32_CYGWIN) +#if defined(OPENSSL_SYS_WINDOWS) || defined(OPENSSL_SYS_WIN32) #include #ifndef _WIN32_WINNT # define _WIN32_WINNT 0x0400 -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: OpenSSL 0.9.7f released
On Mar 23 04:55, vijay basav wrote: > hi > i have built openssl-0.9.7e version for cygwin for mips2 processor . Never heaerd of that. Since when is Cygwin running on Mips? We only support it on x86. Corinna __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 1 released
On May 19 22:12, Richard Levitte - VMS Whacker wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > > OpenSSL version 0.9.8 Beta 1 > > > OpenSSL - The Open Source toolkit for SSL/TLS > http://www.openssl.org/ > > OpenSSL is currently in a release cycle. The first beta is now released. > The beta release is available for download via HTTP and FTP from the > following master locations (the various FTP mirrors you can find under > http://www.openssl.org/source/mirror.html): > > o http://www.openssl.org/source/ > o ftp://ftp.openssl.org/source/ > > The file names of the beta are: > > o openssl-0.9.8-beta1.tar.gz Builds and tests fine on Cygwin, but `make install' is seriously broken for Cygwin. First, the crypto and ssl DLLs are now created without version in the DLL name again (cygssl.dll instead of cygssl-0.9.8.dll) as it was back in 0.9.6. This had been fixed in 0.9.7. Pulling this into the 0.9.8 Makefiles is trivial. Second, the engine installation only supports .so shared libs installations and building the engine DLLs misses the version information in the DLL names, too. The patch to accomodate Cygwin is somewhat more work, but should be basically copy/paste from the top level Makefiles. In both cases I'll send a patch to fix that. It's just rather late and especially the engine Makefile change needs some testing, so give me a couple of days. Thanks, Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 1 released
On May 20 01:43, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Fri, 20 May 2005 00:15:40 +0200, Corinna > Vinschen <[EMAIL PROTECTED]> said: > > vinschen> In both cases I'll send a patch to fix that. It's just > vinschen> rather late and especially the engine Makefile change needs > vinschen> some testing, so give me a couple of days. > > If you get it to us before Monday, it will be there in the next > planned beta (see http://www.openssl.org/news/state.html). Cool. I see what I can do. We're expecting a rainy weekend anyway :-) Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 1 released
On May 20 11:21, Corinna Vinschen wrote: > On May 20 01:43, Richard Levitte - VMS Whacker wrote: > > In message <[EMAIL PROTECTED]> on Fri, 20 May 2005 00:15:40 +0200, Corinna > > Vinschen <[EMAIL PROTECTED]> said: > > > > vinschen> In both cases I'll send a patch to fix that. It's just > > vinschen> rather late and especially the engine Makefile change needs > > vinschen> some testing, so give me a couple of days. > > > > If you get it to us before Monday, it will be there in the next > > planned beta (see http://www.openssl.org/news/state.html). > > Cool. I see what I can do. We're expecting a rainy weekend anyway :-) Sorry for a possibly dumb question, but I don't quite understand the engines concept. Is there a good reason that the engine shared libs are created w/o version information? The problem is that Makefile.shared is called from engines/Makefile without LIBVERSION being set to a sensible value, as it's done in the top level Makefile: LIBVERSION=${SHLIB_MAJOR}.${SHLIB_MINOR} Even worse, the engines Makefile doesn't even have the version information set anywhere, so it's not clear to me how I can generate the engines with version info. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 1 released
On May 20 14:45, Corinna Vinschen wrote: > On May 20 11:21, Corinna Vinschen wrote: > > On May 20 01:43, Richard Levitte - VMS Whacker wrote: > > > In message <[EMAIL PROTECTED]> on Fri, 20 May 2005 00:15:40 +0200, > > > Corinna Vinschen <[EMAIL PROTECTED]> said: > > > > > > vinschen> In both cases I'll send a patch to fix that. It's just > > > vinschen> rather late and especially the engine Makefile change needs > > > vinschen> some testing, so give me a couple of days. > > > > > > If you get it to us before Monday, it will be there in the next > > > planned beta (see http://www.openssl.org/news/state.html). > > > > Cool. I see what I can do. We're expecting a rainy weekend anyway :-) Ok, below is a patch which should solve the problems on Cygwin. It tweaks Makefile.org, Makefile.shared, engines/Makefile and util/cygwin.sh. A short description of the changes: Makefile.org: - The libclean target now also removes DLLs in toplevel and engines, as well as *.a files in engines which are at least created when building engines on Cygwin. - The install_sw target installs DLLs on Cygwin with version number again, as introduced with OpenSSL 0.9.7. This allows cooperation of multiple versions of OpenSSL on the same machine, as on other platforms. Makefile.shared: - The link_o.cygwin target contains a small tweak which allows to build DLLs with and without version number, depending of LIBVERSION being empty or not. Since LIBVERSION is not set in the engines Makefile, engines are built without version numbers. - The link_a.cygwin target creates the cygcrypto DLL with another base address. This is necessary to avoid collision with the Cygwin DLL itself. The address 0x6120 used right now collides at least with the Cygwin heap in the upcoming 1.5.17 release. Since the Cygwin DLL is the only DLL in a Cygwin system which must not be rebased, I chose a new, save base address (0x6300) for the cygcrypto DLL instead. - The link_a.cygwin target also takes version numbering in the DLL name into account. engines/Makefile: - The install target installs the engines not as cygFOO.dll, but as libFOO.so instead. This might come as a surprise, but since engines are run-time loaded using dlopen and the dlfcn_name_converter is used also for Cygwin, that's the way to go :-) What I'm not sure about is, if the engines shouldn't better be installed into /usr/bin. Otherwise they are only loadable using full paths, or when $PATH also contains the engine installation path, according to Windows DLL loading rules. I'd be grateful about a hint here... util/cygwin.sh: - Accomodates the engines installation now. - Installs also the certificates from the certs directory. This was never correct before! Shame on me. Thanks for considering, Corinna --- openssl-0.9.8-beta1.ORIG/Makefile.org 2005-05-19 03:48:45.0 +0200 +++ openssl-0.9.8-beta1/Makefile.org2005-05-20 16:19:21.128963500 +0200 @@ -276,7 +276,7 @@ Makefile: Makefile.org Configure config @false libclean: - rm -f *.map *.so *.so.* engines/*.so *.a */lib */*/lib + rm -f *.map *.so *.so.* *.dll engines/*.so engines/*.dll *.a engines/*.a */lib */*/lib clean: libclean rm -f shlib/*.o *.o core a.out fluff rehash.time testlog make.log cctest cctest.c @@ -476,13 +476,13 @@ install_sw: chmod 555 $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new; \ mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i; \ else \ - c=`echo $$i | sed 's/^lib/cyg/'`; \ + c=`echo $$i | sed 's/^lib\(.*\)\.dll\.a/cyg\1-$(SHLIB_VERSION_NUMBER).dll/'`; \ cp $$c $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new; \ chmod 755 $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new; \ mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c; \ - cp $$i.a $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a.new; \ - chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a.new; \ - mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a.new $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a; \ + cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new; \ + chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new; \ + mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new $(INSTALL_PREFIX)$(INS
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 1 released
On May 21 11:44, Richard Levitte - VMS Whacker wrote: > Corinna Vinschen said: > vinschen> Even worse, the engines Makefile doesn't even have the > vinschen> version information set anywhere, so it's not clear to me > vinschen> how I can generate the engines with version info. > > You simply don't :-). Thanks, that's what I figured. So my patch I've sent yesterday should be basically correct. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 1 released
On May 21 18:42, Richard Levitte - VMS Whacker wrote: > Corinna Vinschen said: > > vinschen> Ok, below is a patch which should solve the problems on > vinschen> Cygwin. It tweaks Makefile.org, Makefile.shared, > vinschen> engines/Makefile and util/cygwin.sh. > > Applied. Thanks! What about my question: >> What I'm not sure about is, if the engines shouldn't better be installed >> into /usr/bin. Otherwise they are only loadable using full paths, or when >> $PATH also contains the engine installation path, according to Windows >> DLL loading rules. I'd be grateful about a hint here... Is that something to worry about at all? Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 1 released
On May 21 19:29, Richard Levitte - VMS Whacker wrote: > The following from CHANGES should explain it: > > *) Change the ENGINE framework to automatically load engines > dynamically from specific directories [...] > time with the environment variable OPENSSL_ENGINES. > [Geoff Thorpe and Richard Levitte] It does. I should start reading documentation... Thanks, Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 1 released
Hi Andy, On May 21 22:36, Andy Polyakov wrote: > Speaking of documentation:-) Is cygwin memory layout documented > somewhere? What I'm trying to ask is if you can provide a pointer. If > it's documented in commentary section in source code, then could you be > so kind to give some approximate directions where to look? A. It's not exactly documented, no. If you download the sources, there's a file how-cygheap-works.txt, which explains the basics about cygwin's own heap which is inherited to child applications on fork and exec. Other than that, there are just a couple of shared memory areas for various puposes (process list, tape "driver", etc) and the Cygwin DLL itself. The DLL is loaded to 0x6100, the cygheap is always expected to be right after the end of the DLL's loading area. The shared memory regions have been accidentally loaded to 0x0400 since 1.5.13 up to current 1.5.16, but will reside beginning at 0x6000 with the upcoming 1.5.17. If you need more details, you will have to look into the sources, I assume. If you have questions, feel free to send PM, I think more of that is pretty OT for this list. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Problems with engines in 0.9.8-beta1
Hi, now that I had first contact with engines, I thought it might be better to give them some testing. It turns out that loading the engines works fine, but there are three engines which don't work, and it seems to me that this isn't a Cygwin specific problem. These three engines don't load: lib4758_cca.so libgmp.so libncipher.so One of them is not a mystery. libgmp.so is just empty, because OPENSSL_USE_GMP wasn't set when building. So that can be ignored. However, I found that lib4758_cca.so and libncipher.so don't load, because the engine id differs from the engine name. The engine id of lib4758_cca.so is "4758cca" instead of "4758_cca", the id of libncipher.so is "chil" instead of "ncipher". The result is this. If I try, for instance, to open the engine "ncipher" with openssl: OpenSSL> engine ncipher 1608:error:260B606D:engine routines:DYNAMIC_LOAD:init failed:eng_dyn.c:521: 1608:error:2606A074:engine routines:ENGINE_by_id:no such engine:eng_list.c:411:id=ncipher This is because the loaded bind function bind_fn() rejects being initialized, because the incoming id "ncipher" is not equal the engine's own id engine_hwcrhk_id == "chil". OTOH, when trying to load an engine "chil", this happens: OpenSSL> engine chil 3668:error:25066067:DSO support routines:DLFCN_LOAD:could not load the shared library:dso_dlfcn.c:162:filename(/usr/lib/engines/libchil.so): dlopen, Win32 error 126 3668:error:25070067:DSO support routines:DSO_load:could not load the shared library:dso_lib.c:244: 3668:error:260B6084:engine routines:DYNAMIC_LOAD:dso not found:eng_dyn.c:450: 3668:error:2606A074:engine routines:ENGINE_by_id:no such engine:eng_list.c:411:id=chil It tries to open an engine called libchil.so, which isn't present. The problem is equivalent for lib4758_cca.so. So it seems, either the engine ids are just wrong and should be changed, or the names of the created shared libs is wrong and should be changed. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 2 released
On May 24 06:21, Richard Levitte - VMS Whacker wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > > OpenSSL version 0.9.8 Beta 2 > Builds and tests well on Cygwin. The naming problem of the 4758_cca and ncipher engines I reported on 2005-05-22 is still there, though. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: Problems with engines in 0.9.8-beta1
On May 24 23:16, Geoff Thorpe wrote: > On May 22, 2005 08:17 am, Corinna Vinschen wrote: > > now that I had first contact with engines, I thought it might be > > better to give them some testing. > > Yes, thanks for doing so :-) Sure. I just can't test if they *really* work since I'm obviously missing the stuff they're trying to connect to. > > However, I found that lib4758_cca.so and libncipher.so don't load, > > because the engine id differs from the engine name. > > > > The engine id of lib4758_cca.so is "4758cca" instead of "4758_cca", > > the id of libncipher.so is "chil" instead of "ncipher". > > Yes, this is unfortunate. I've just committed a fix to the 0.9.8-stable > and HEAD branches that will tolerate both names when binding as a dynamic > engine. It'd still be preferable to change the names of the generated > shared libraries too so that the default name with static and dynamic use > is the same (ie. using 'chil' for a built-in engine and 'ncipher' for an > external engine make much sense). Right now the existing change will just > allow you to dynamically bind using 'ncipher'. Please try out the next > nightly snapshot if you're able. Works. But I agree, it be more correct to be able to load a "chil" and a "4758cca" engine. > Richard, any idea of how safe it would be to change the names of the two > shared librariesy at this stage of the 0.9.8 betas? I'm reluctant to > charge ahead for fear of breaking the strange builds (win32, VMS, > cygwin, ...) WHAT? Cygwin a strange build? How dare you... ;-) Thanks, Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: Problems with engines in 0.9.8-beta1
On May 29 15:35, Geoff Thorpe wrote: > Corinna, Richard, > > I've commited changes to HEAD and the 0.9.8-stable branches, please let me > know if you see any problems. I tried CVS HEAD and this looks good to me, as far as I can test it. Thanks, Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 5 released
On Jun 13 06:19, Richard Levitte - VMS Whacker wrote: > OpenSSL version 0.9.8 Beta 5 > Builds OOTB and tests run fine on Cygwin. Thanks, Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 5 released
On Jun 13 18:59, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Mon, 13 Jun 2005 10:57:10 +0200, Corinna > Vinschen <[EMAIL PROTECTED]> said: > > vinschen> On Jun 13 06:19, Richard Levitte - VMS Whacker wrote: > vinschen> > OpenSSL version 0.9.8 Beta 5 > vinschen> > > vinschen> > vinschen> Builds OOTB and tests run fine on Cygwin. > > Which version(s)? The latest release 1.5.17 and current CVS. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 6 released
On Jun 21 08:21, Richard Levitte - VMS Whacker wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > > OpenSSL version 0.9.8 Beta 6 (FINAL!) Build and tests fine on Cygwin. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 beta 6 released
On Jun 22 10:25, Corinna Vinschen wrote: > On Jun 21 08:21, Richard Levitte - VMS Whacker wrote: > > -BEGIN PGP SIGNED MESSAGE- > > Hash: SHA1 > > > > > > OpenSSL version 0.9.8 Beta 6 (FINAL!) > > Build and tests fine on Cygwin. Just one installation quirk. The engines are installed non-stripped right now. I forgot to tweak the util/cygwin.sh script to strip them before creating the install archives. Any problem to apply the below last minute patch to util/cygwin.sh? Corinna --- util/cygwin.sh.ORIG 2005-06-22 13:56:02.659532200 +0200 +++ util/cygwin.sh 2005-06-22 13:54:51.822851000 +0200 @@ -119,7 +119,9 @@ create_cygwin_readme create_profile_files cd ${INSTALL_PREFIX} -strip usr/bin/*.exe usr/bin/*.dll +chmod u+w usr/lib/engines/*.so +strip usr/bin/*.exe usr/bin/*.dll usr/lib/engines/*.so +chmod u-w usr/lib/engines/*.so # Runtime package find etc usr/bin usr/lib/engines usr/share/doc usr/ssl/certs \ -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1135] 0.9.8-beta7-dev and DJGPP
On Jun 24 14:29, [EMAIL PROTECTED] via RT wrote: > > The OpenSSL 0.9.8-stable snapshot from 24 June 2005 configures, > builds, tests, and installs without problem on DJGPP. Same for Cygwin. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCE] OpenSSL 0.9.8 released
On Jul 5 21:40, Richard Levitte - VMS Whacker wrote: > OpenSSL version 0.9.8 released > == I'm pleased to announce that OpenSSL 0.9.8 is now packaged in the Cygwin net distribution. See http://cygwin.com/ml/cygwin-announce/2005-07/msg3.html Thanks to the OpenSSL team, Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[PATCH] 0.9.8: Install bug on case-preserving filesystems
Hi, the first day of 0.9.8 in Cygwin resulted in the following bug report on the Cygwin list: > Have just done a fresh install of cygwin to compile a Linux app that uses > DES. > > According to "man des", I should still be able to include , > but that > file doesn't seem to exist in this release. The reporter was right and I tried to figure out what has gone wrong so that the des (and, FWIW, the bf) header files were missing. The cause of these files missing is a "cleanup" in the Makefiles of crypto/des and crypto/bf. Both directories have a file called "INSTALL" in them. The Makefile rule is: install: @do something in a subshell The problem only happens on filesystems which are not case-sensitive but only case-preserving, like filesystems on Windows or MacOS. What happens is that make(1) tries to find if the rule is up to date. While doing this, make calls stat("install"), which returns with success on case-preserving filesystems! As a result, the install rule is "up to date" and the make action will never be called. This did *not* happen in 0.9.7, which had the "install" rule created so that the above install bug couldn't happen: install: installs installs: @do something in a subshell This circumvents the above problem on case-preserving filesystems. I would like to ask for reverting this change in the affected directories crypto/des and crypto/bf. The patch is below. I'm going to release a new Cygwin version of OpenSSL 0.9.8 today which has this fix applied and which installs the des and bf header files correctly. Thanks for considering, Corinna --- crypto/des/Makefile.save2005-07-05 22:33:16.046875000 +0200 +++ crypto/des/Makefile 2005-07-07 10:24:53.31250 +0200 @@ -88,7 +88,9 @@ links: @$(PERL) $(TOP)/util/mklink.pl ../../test $(TEST) @$(PERL) $(TOP)/util/mklink.pl ../../apps $(APPS) -install: +install: installs + +installs: @[ -n "$(INSTALLTOP)" ] # should be set by top Makefile... @headerlist="$(EXHEADER)"; for i in $$headerlist ; \ do \ --- crypto/bf/Makefile.save 2005-07-05 22:33:19.53125 +0200 +++ crypto/bf/Makefile 2005-07-07 10:24:40.78125 +0200 @@ -62,7 +62,9 @@ links: @$(PERL) $(TOP)/util/mklink.pl ../../test $(TEST) @$(PERL) $(TOP)/util/mklink.pl ../../apps $(APPS) -install: +install: installs + +installs: @[ -n "$(INSTALLTOP)" ] # should be set by top Makefile... @headerlist="$(EXHEADER)"; for i in $$headerlist ; \ do \ -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [PATCH] 0.9.8: Install bug on case-preserving filesystems
On Jul 7 15:51, Richard Levitte wrote: > Doug Kaufman writes: > > >> install: > >>@do something in a subshell > >> > >>The problem only happens on filesystems which are not case-sensitive but > >>only case-preserving, like filesystems on Windows or MacOS. > > > >I can confirm that the des.h and blowfish.h headers are not installed in > >either the DJGPP build or in the MINGW build (via > >Configure/-mno-cygwin). > > I'll see what I can do with that, tomorrow evening. My patch should be everything needed. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [PATCH] 0.9.8: Install bug on case-preserving filesystems
On Jul 8 12:14, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Thu, 7 Jul 2005 11:01:11 +0200, Corinna > Vinschen <[EMAIL PROTECTED]> said: > > vinschen> The problem only happens on filesystems which are not > vinschen> case-sensitive but only case-preserving, like filesystems on > vinschen> Windows or MacOS. > vinschen> > vinschen> What happens is that make(1) tries to find if the rule is up > vinschen> to date. While doing this, make calls stat("install"), > vinschen> which returns with success on case-preserving filesystems! > > I've applied a fix which looks a little different from yours: > > Index: crypto/des/Makefile > === > RCS file: /e/openssl/cvs/openssl/crypto/des/Makefile,v > retrieving revision 1.5 > diff -u -r1.5 Makefile > --- crypto/des/Makefile 16 May 2005 16:55:22 - 1.5 > +++ crypto/des/Makefile 8 Jul 2005 10:10:55 - > @@ -88,7 +88,10 @@ > @$(PERL) $(TOP)/util/mklink.pl ../../test $(TEST) > @$(PERL) $(TOP)/util/mklink.pl ../../apps $(APPS) > > -install: > +# We need to use force because 'install' matches 'INSTALL' on case > +# insensitive systems > +FRC.install: > +install: FRC.install > @[ -n "$(INSTALLTOP)" ] # should be set by top Makefile... > @headerlist="$(EXHEADER)"; for i in $$headerlist ; \ > do \ > > > (did the same thing in the crypto/bf directory) Looks good. Thanks, Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: v0.9.8 and MingW inside MSYS
On Jul 8 22:08, Matthias Mohr wrote: > Hi everybody, > > I tried to compile OpenSSL v0.9.8 inside Windows with > MingW environment inside MingW's MSYS environment > (this is a small unix like environment like cygwin, but much > smaller and native to windows). Just for the records. MSYS *is* Cygwin, a branch from a very old version of Cygwin. It's no way more or less "native" than Cygwin itself. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [PATCH] libcrypto without executable stack
On Nov 9 01:19, Mike Frysinger wrote: > On Wed, Nov 09, 2005 at 12:00:19AM +0100, Dirk Mueller wrote: > > the appended patch makes libcrypto.so compile without executable stack > > requirements. it should be portable accross all versions of binutils (and > > doesn't affect any non-linux platform anyway). > > it will break non-ELF builds though (but maybe the script isnt used for > non-ELF targets so thats OK?) It's also used for Cygwin and the patch breaks the Cygwin build. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [PATCH] libcrypto without executable stack
On Nov 9 13:57, Dirk Mueller wrote: > On Wednesday 09 November 2005 10:45, Corinna Vinschen wrote: > > It's also used for Cygwin and the patch breaks the Cygwin build. > > I don't have a cygwin toolchain around, but can you tell me the error message > so that I can work on fixing it? x86cpuid-cof.s: Assembler messages: x86cpuid-cof.s:0: Warning: end of file not at end of a line; newline inserted x86cpuid-cof.s:165: Error: junk at end of line, first unrecognized character is `-' > does the attached patch work? Yes, it works. The resulting asm file does not contain the new .section pseudo op. There are two problems with this section in PE/COFF, first, the section name must not contain a dash (for whatever reason), and second, COFF sections have only up to two parameters as described in the gas info pages: .section NAME[, "FLAGS"] .section NAME[, SUBSEGMENT] Btw., the first asm message indicates that a \n is missing. You should add this at the end of the section string to avoid the warning. HTH, Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: Alternate RAND_poll for XP/2003 Server/Vista
On Dec 2 00:45, Steven Reddie wrote: > That's an interesting blog article. The 2nd comment is by the author and > lists the entropy sources. I recall there was discussion on this list quite > some time ago where it was stated that OpenSSL wouldn't use only the > CryptoAPI random number generator since Microsoft hadn't provided details of > how the entropy was gathered. Perhaps the information in that post provides > enough detail to warrant dropping all of the heap walking guff that has been > known to trip up OpenSSL on occasion. > > Adria, RAND_poll already calls CryptGenRandom doesn't it? You could > probably just comment out all of the other code to get the same result. I'm wondering about this anyway. While the exact code of CryptGenRandom isn't open source, MSDN has a quite extensive description how the random numbers are generated by CryptGenRandom, see the Remarks section in http://msdn.microsoft.com/library/default.asp?url=/library/en-us/seccrypto/security/cryptgenrandom.asp which also talks about the entropy sources used. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: Alternate RAND_poll for XP/2003 Server/Vista
On Dec 1 12:43, Rick Jones wrote: > Corinna Vinschen wrote: > >I'm wondering about this anyway. While the exact code of CryptGenRandom > >isn't open source, MSDN has a quite extensive description how the random > >numbers are generated by CryptGenRandom, see the Remarks section in > >http://msdn.microsoft.com/library/default.asp?url=/library/en-us/seccrypto/security/cryptgenrandom.asp > >which also talks about the entropy sources used. > > Being otherwise blissfully ignorant of things Windows, does this snippet > from the URL above: > > "With Microsoft CSPs, CryptGenRandom uses the same random number generator > used by other security components." > > imply that CryptGenRandom might be satisfied by code other than that from > Microsoft described in the URL above? You omited the next sentence: "This allows numerous processes to contribute to a system-wide seed." I understand this as "every process using one of the Microsoft CSPs will internally access the same random number generator." As if, say, every process uses /dev/random on Linux. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: Alternate RAND_poll for XP/2003 Server/Vista
On Dec 2 13:38, Steven Reddie wrote: > Two sentences after that one are these: > > "To form the seed for the random number generator, a calling > application supplies bits > it might have-for instance, mouse or keyboard timing input-that are > then added to both > the stored seed and various system data and user data such as the > process ID and thread ID, > the system clock, the system time, the system counter, memory > status, free disk clusters, > the hashed user environment block. This result is SHA-1 hashed, and > the output is used to seed > an RC4 stream, which is then used as the random stream and used to > update the stored seed." > > So although all processes access the same PRNG it seems that there is > provision to "mix it up" a little. > > I imagine that an installed Cryptographic Provider could provide its own > PRNG and Windows could be configured to use that instead. If you install another CSP, you can use this CSP and this CSP in turn can use another source for its cryptographic magic, including the sources for its random number generator. The above description is only valid for the predefined CSPs as delivered by default by the OS[1]. If you decide to use CryptGenRandom, you also have to call CryptAcquireContext[2]. This function gets the CSP as an argument. If you use NULL here, as the openssl library does, the CSP is the one set as the user default CSP. The user default CSP can be set by a call to CryptSetProvider[3], and is then used as the default provider for this user. Every process started after this call gets the new CSP. This means, that the openssl library can not be sure to use the same CSP all the time, or even a trustworthy one (for a given value of trust). But, if you call CryptAcquireContext with one of the Microsoft predefined CSPs, you can rely on the fact that the same random number generator is used all the time and the description above describes how the random numbers are generated then. Whatever your trust level in relation to Microsoft is, you know at least which CSP is used. FWIW, the Cygwin implementation of /dev/random and /dev/urandom uses the "Microsoft Base Cryptographic Provider 1.0", MS_DEF_PROV, so at least it's using the same cryptographic source all the time. Shouldn't a fixed CSP be used for the native Windows random number generator in crypt/rand/rand_win.c, too? Corinna [1] http://msdn.microsoft.com/library/default.asp?url=/library/en-us/seccrypto/security/cryptographic_provider_names.asp [2] http://msdn.microsoft.com/library/default.asp?url=/library/en-us/seccrypto/security/cryptacquirecontext.asp [3] http://msdn.microsoft.com/library/default.asp?url=/library/en-us/seccrypto/security/cryptsetprovider.asp > > -Original Message- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > On Behalf Of Corinna Vinschen > Sent: Friday, 2 December 2005 7:59 AM > To: openssl-dev@openssl.org > Subject: Re: Alternate RAND_poll for XP/2003 Server/Vista > > On Dec 1 12:43, Rick Jones wrote: > > Corinna Vinschen wrote: > > >I'm wondering about this anyway. While the exact code of > > >CryptGenRandom isn't open source, MSDN has a quite extensive > > >description how the random numbers are generated by CryptGenRandom, > > >see the Remarks section in > > >http://msdn.microsoft.com/library/default.asp?url=/library/en-us/secc > > >rypto/security/cryptgenrandom.asp > > >which also talks about the entropy sources used. > > > > Being otherwise blissfully ignorant of things Windows, does this > > snippet from the URL above: > > > > "With Microsoft CSPs, CryptGenRandom uses the same random number > > generator used by other security components." > > > > imply that CryptGenRandom might be satisfied by code other than that > > from Microsoft described in the URL above? > > You omited the next sentence: > > "This allows numerous processes to contribute to a system-wide seed." > > I understand this as "every process using one of the Microsoft CSPs will > internally access the same random number generator." As if, say, every > process uses /dev/random on Linux. > > > Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat, Inc. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1365] Re: IPv6 support in openssl's BIO
On Jul 21 11:23, Andy Polyakov via RT wrote: > > I sent a patch to provide a solution for platforms that have addrinfo. > > Basically, if you do not have addrinfo, you cannot expect IPv6 to > > work. > > My comment was not about having IPv6 working without addrinfo, but about > not breaking platforms which don't have one. > > > I was looking for a feedback about this particular solution for > > platforms that support IPv6. > > And feedback is "no, it's not the way it should be done, because we > can't have different branches for platforms that have addrinfo and those > which don't." > > > Supporting legacy platforms was not my primary goal _at this time_ > > Well, it is ours at all times and you have to respect that:-) Well, the > goal is not really support legacy platform at all costs, but rather not > discontinue if compatibility can be achieved by maintaining certain > discipline. > [...] > OK, to be absolutely sincere. It's not about what Microsoft [or RedHat] > supports for the moment, but about exercising above mentioned > discipline. It's not about looking for excuses, but exploring maximum > possible extent of portability. Is it possible to write code which > adapts itself to run-time environment? Regardless whether run-time is > supported by vendor or not? How complicated is it? Forget that you're > RedHat employee, become programmer instead:-) But this is getting > off-topic. DSO_global_lookup was not the point I wanted to emphasize! > DSO_global_lookup is mentioned only because you'll have to cope with it > in HEAD branch. The main point is that changes of this character should > be *conditional* for backward compatibility and you have to play by this > rules. A. Just as a side note to Jan. Cygwin is a Red Hat product as well and the current release doesn't support IPv6 and getaddrinfo for now. I have a developer branch which adds IPv6 support to Cygwin, but this work will hit the streets only in a few weeks or months. Of course I wouldn't like to see OpenSSL suddenly breaking on current version of Cygwin just because IPv6 (resp. getaddrinfo) is used unconditionally ;-) Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: OPENSSL_gmtime on platforms that don't have a safe gmtime function
On Aug 11 12:29, Joe Gluck wrote: > The OPENSSL_gmtime in o_time.c (that gets called from other places like > ASN1_UTCTIME_cmp_time_t in a_utctm.c) does not use the safe version of > gmtime in lots of platforms including: > OPENSSL_SYS_WIN32 > OPENSSL_SYS_OS2 > __CYGWIN32__ > OPENSSL_SYS_MACOSX > OPENSSL_SYS_SUNOS > This could cause problems in multi-thread environments, why not use mutexes > to wrap the gmtime and memcpy, like other places in the library? Please drop Cygwin from this list since it has gmtime_r for ages. I'm sorry that I missed that. See patch below. Corinna --- crypto/o_time.c.orig2006-08-13 10:23:23.815738500 +0200 +++ crypto/o_time.c 2006-08-13 10:23:34.211600500 +0200 @@ -73,7 +73,7 @@ struct tm *OPENSSL_gmtime(const time_t * { struct tm *ts = NULL; -#if defined(OPENSSL_THREADS) && !defined(OPENSSL_SYS_WIN32) && !defined(OPENSSL_SYS_OS2) && !defined(__CYGWIN32__) && (!defined(OPENSSL_SYS_VMS) || defined(gmtime_r)) && !defined(OPENSSL_SYS_MACOSX) && !defined(OPENSSL_SYS_SUNOS) +#if defined(OPENSSL_THREADS) && !defined(OPENSSL_SYS_WIN32) && !defined(OPENSSL_SYS_OS2) && (!defined(OPENSSL_SYS_VMS) || defined(gmtime_r)) && !defined(OPENSSL_SYS_MACOSX) && !defined(OPENSSL_SYS_SUNOS) /* should return &data, but doesn't on some systems, so we don't even look at the return value */ gmtime_r(timer,result); -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: Cross compile OpenSSL in Linux using MinGW32
On Oct 20 13:33, Victor B. Wagner wrote: > On 2006.10.20 at 10:56:35 +0200, Andy Polyakov wrote: > > >It is not perfect to, because it assumes that if one uses mingw32 > > >target, there is always some Unix emulation environment (i.e. cygwin, > > >msys or real Unix in case of cross-builds). > > > > As implied earlier I'd actually prefer this, i.e. mingw build to > > *require* Unix emulation environment. Is it really unreasonable? In > > I think it is reasonable. Unless it would break some non-gcc build > (i.e Visual Studio, Borland or some netware one). > > > other words I'd simply scrap "$IsMK1MF=1 if ($target eq "mingw" && $^O > > ne "cygwin" && !is_msys());" line for good. A. > > Now, some further info. > > Next problem I've encountered building current CVS state of 0.9.9 > was error in e_os.h > "ws2tcpip.h is not compatible with winsock.h". It was fixed by removal > of #include from mentioned file. > > I'm not an expert on Win32 tcpip history and cannot tell whether it is > problem of my mingw32 runtime headers or something also. ws2tcpip.h is incompatible with winsock.h since winsock.h is only meant for supporting old Winsock 1.1 applications. A "modern" Winsock 2 application should include winsock2.h and ws2tcpip.h. > Next problem was "dereferencing pointer to incomplete type" in > line 728 of b_sock.c. It seems that symbol AF_INET6 is somehow declared > (may be cross-compiler picks some native header), but appropriate > structures are not defined. I've temporary solved this problem by > replacing The IPv6 stuff is defined in ws2tcpip.h... And here's another problem. The functions getaddrinfo, getnameinfo and freeaddrinfo are only exported by ws2_32.dll beginning with Windows XP. There's an earlier implementation for Windows 2000 which is exported by a DLL called wship6.dll. There's no v6 for 9x or NT4. Consequentially, on native Windows (not Cygwin) the functions should not be linked against, but instead corresponding function pointers should be loaded at runtime from either ws2_32.dll or wship6.dll using LoadLibrary/GetProcAddress. HTH, Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: Cross compile OpenSSL in Linux using MinGW32
On Oct 20 14:28, Victor B. Wagner wrote: > On 2006.10.20 at 11:49:39 +0200, Corinna Vinschen wrote: > > ws2tcpip.h is incompatible with winsock.h since winsock.h is only meant > > for supporting old Winsock 1.1 applications. A "modern" Winsock 2 > > application should include winsock2.h and ws2tcpip.h. > > So, it is line 455 in e_os.h which is offending, not line 278? Line 455 looks wrong to me. winsock2.h is already included in line 277 so I don't see how another include of winsock.h in line 455 could be right. > > And here's another problem. The functions getaddrinfo, getnameinfo and > > freeaddrinfo are only exported by ws2_32.dll beginning with Windows XP. > > There's an earlier implementation for Windows 2000 which is exported by > > a DLL called wship6.dll. There's no v6 for 9x or NT4. Consequentially, > > on native Windows (not Cygwin) the functions should not be linked > > against, but instead corresponding function pointers should be loaded at > > runtime from either ws2_32.dll or wship6.dll using > > LoadLibrary/GetProcAddress. > > So, use IPV6 on native windows requires considerable changes anyway? I wouldn't say it's considerable. Just a tweak to the loading of getaddrinfo/freeaddrinfo in crypto/bio/b_sock.c, AFAICS. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: Cross compile OpenSSL in Linux using MinGW32
On Oct 20 15:21, Victor B. Wagner wrote: > On 2006.10.20 at 13:01:01 +0200, Corinna Vinschen wrote: > > > > So, use IPV6 on native windows requires considerable changes anyway? > > > > I wouldn't say it's considerable. Just a tweak to the loading of > > getaddrinfo/freeaddrinfo in crypto/bio/b_sock.c, AFAICS. > > Implementing of dynamic loading by hand is tricky thing anyway. Huh? > One have to declare function pointers and provide code which would fill > them with correct value. Which is already done in crypto/bio/b_sock.c. Did you look into the code? > And this code should be clever enough to find > appropriate DLL (provided that most Windows systems out there have > both). LoadLibrary? That's nothing new and already used in crypto/dso/dso_win32.c. ws2_32.dll exists on all systems starting with Windows 95 OSR2, but that doesn't mean the IPv6 functions are available. On Windows 2000, IPv6 is only available if wship6.dll is installed. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: Cross compile OpenSSL in Linux using MinGW32
On Oct 20 13:51, Andy Polyakov wrote: > Keep in mind that mingw defines _WIN32_WINNT=0x333, i.e. the intention > was to target all NT versions [note that 0x333 actually covers even for > Windows 9x, which has at least all 0x333 stubs, so that application can > actually start]. As for winsock versioning. Upon latest modifications to > b_sock.c I considered linking with wsock32 to be sufficient/appropriate > for following reason. Systems equipped with ws2_32.dll do have wsock32 > too, and this wsock32.dll is actually linked with ws2_32.dll. Meaning > that [legacy] application linked with wsock32 alone will actually bring > even ws2_32.dll into address space. Now note that b_sock.c makes > *global* lookups for getaddrinfo, meaning that application linked with > wsock32 alone will actually find getaddrinfo even if it resides in > ws2_32! So that the fact that latest headers [those defining struct This is a very thin ice approach. When you use wsock32, it's using Winsock 1.1. There are incompatibilities between Winsock 1.1 and Winsock 2, which are solved by using different header files. Including winsock.h and winsock2.h concurrently is wrong. It's also wrong to include winsock.h when linking against ws2_32.dll and it's wrong to include winsock2.h when linking against wsock32.dll. For instance, several socket options have different values. As an example, IP_TOS is defined as the value 3 under Winsock 2, but it was defined as the value 8 under Winsock 1.1. > addrinfo] are included, but elder library is linked with is actually > intentional. Yes, it requires certain programming discipline, but it's > [considered] doable. As for IPv6. If w2k supports it only through > additional library, I'd say "is it really a problem not to have IPv6 on > pre-XP?" A. Seriously, I'd consider Winsock 1.1 as the one which should be left behind, rather than Windows 2000 users. As I said in another mail, Winsock 2 is by default available since Windows 95 OSR2 and NT4. Even for the original non-OSR2 release of Windows 95 is a Microsoft package with a Winsock2 implementation available. On the other hand, Windows 2000 is still officially supported by Microsoft, in contrast to Windows 95 and, FWIW, wsock32.dll. It's really not that hard. Always use ws2_32.dll instead of wsock32.dll, never include winsock.h and, last but not least, if loading getaddrinfo/ freeaddrinfo from ws2_32.dll fails, try again by loading it from wship6.dll. If that fails, IPv6 is not available. However, I'm not sure if the DSO_global_lookup approach also covers wship6.dll automatically on W2K. Somebody would have to try it. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: Cross compile OpenSSL in Linux using MinGW32
On Oct 20 15:05, Andy Polyakov wrote: > >Seriously, I'd consider Winsock 1.1 as the one which should be left > >behind, rather than Windows 2000 users. > > Windows 2000 users are not left behind, IPv6 on 2000 would be. That's what I meant :) > Keep in mind 0x333, it's 3.51. If we switch to ws2_32, I'd insist on > bumping _WIN_WIN32 to 0x400. Shall we do that? I personally have no > objections to that. Me neither, but then again, I'm not a user nor am I maintainer of a native Windows OpenSSL... > DSO_global_lookup looks only through currently loaded dlls, and never > attempts to load any new. That's why I said I don't know if it would already work. It's possible that wship6.dll is automatically loaded into the processes VM when ws2_32.dll is loaded in which case DSO_global_lookup would just do the "right thing"(tm). > On bright side one can simply throw in > LoadLibrary("wship6.dll") literally anywhere, e.g. next to WSAStartup. A. Yup. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[PATCH] apps/ocsp.c: Fix non-POSIX #include
Hi, building apps/ocsp.c fails on Cygwin like this: gcc [...] -c -o ocsp.o ocsp.c ocsp.c: In function `query_responder': ocsp.c:1262: error: storage size of 'tv' isn't known ocsp.c:1290: warning: implicit declaration of function `select' ocsp.c:1262: warning: unused variable `tv' make[2]: *** [ocsp.o] Error 1 The reason is that time.h is included instead of sys/time.h. The latter is required to get the definition of struct timeval per POSIX (1), and it's not required to include sys/time.h from time.h (2). The implicit declaration of select is a result of missing the sys/select.h include (3). The below patch fixes that. Index: apps/ocsp.c === RCS file: /home/cvs/cvsroot/src/openssl/apps/ocsp.c,v retrieving revision 1.39 diff -p -u -r1.39 ocsp.c --- apps/ocsp.c 25 Dec 2006 10:54:14 - 1.39 +++ apps/ocsp.c 29 Mar 2007 09:25:37 - @@ -63,7 +63,8 @@ declared properly on Compaq platforms (at least with DEC C). */ -#include +#include +#include #include "apps.h" #include #include Thanks, Corinna (1) http://www.opengroup.org/onlinepubs/009695399/basedefs/sys/time.h.html (2) http://www.opengroup.org/onlinepubs/009695399/basedefs/time.h.html (3) http://www.opengroup.org/onlinepubs/009695399/functions/select.html -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[PATCH] Makefile.shared: Link all shared libs with --enable-auto-image-base
Hi, the below patches are supposed to help with an annoying DLL problem on Cygwin. Since Windows doesn't provide a native fork implementation, Cygwin emulates fork by creating a new process and overwrites the forkee's memory with data from the forker. This method only works if the OS doesn't randomly decide where to put DLLs into the memory of the new process. However, DLLs are linked with a default image base of 0x1000, thus requiring the runtime loader to relocate the DLLs to some other memory base, which sometimes results in different locations for DLLs in forker and forkee. This in turn breaks fork. One way to ease the pain is to rebase the DLLs manually, using a tool called rebase. If all DLLs in a Cygwin installation are based on different memory locations the problem is much less frequently observed, next to never. The disadvantage of this approach is that it requires the manual intervention. What we're trying to do now in Cygwin is this: Binutils' ld allows to create pseudo-random image base addresses by giving the command line option --enable-auto-image-base. It chooses the image base for the DLL by creating a hash value from the DLL's name. This lowers the chance of DLL collisions a lot, and manual rebasing is most of the time not necessary anymore. The below two patches change Makefile.shared so that the shared libs on Cygwin are built with --enable-auto-image-base. The first patch is against current CVS, the second patch is against the 0.9.8 release. I hope both are ok for inclusion. Thanks in advance, Corinna === Current CVS === Index: Makefile.shared === RCS file: /home/cvs/cvsroot/src/openssl/Makefile.shared,v retrieving revision 1.63 diff -p -u -r1.63 Makefile.shared --- Makefile.shared 25 Mar 2007 15:20:35 - 1.63 +++ Makefile.shared 29 Mar 2007 09:34:02 - @@ -255,7 +255,7 @@ link_o.cygwin: SHLIB_SOVER=${LIBVERSION:+"-$(LIBVERSION)"}; \ ALLSYMSFLAGS='-Wl,--whole-archive'; \ NOALLSYMSFLAGS='-Wl,--no-whole-archive'; \ - SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,-s,-Bsymbolic"; \ + SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,--enable-auto-image-base -Wl,-s,-Bsymbolic"; \ $(LINK_SO_O) link_a.cygwin: @ $(CALC_VERSIONS); \ @@ -273,8 +273,7 @@ link_a.cygwin: extras="$$extras rc.o"; \ ALLSYMSFLAGS='-Wl,--whole-archive'; \ NOALLSYMSFLAGS='-Wl,--no-whole-archive'; \ - base=; [ $(LIBNAME) = "crypto" ] && base=-Wl,--image-base,0x6300; \ - SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared $$base -Wl,-s,-Bsymbolic -Wl,--out-implib,lib$(LIBNAME).dll.a $$extras"; \ + SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,--enable-auto-image-base -Wl,-s,-Bsymbolic -Wl,--out-implib,lib$(LIBNAME).dll.a $$extras"; \ [ -f apps/$$dll_name ] && rm apps/$$dll_name; \ [ -f test/$$dll_name ] && rm test/$$dll_name; \ $(LINK_SO_A) || exit 1; \ === 0.9.8 === --- Makefile.shared 2007-01-18 17:05:21.540291400 +0100 +++ Makefile.shared 2007-03-29 11:38:48.552858700 +0200 @@ -242,7 +242,7 @@ link_o.cygwin: SHLIB_SOVER=${LIBVERSION:+"-$(LIBVERSION)"}; \ ALLSYMSFLAGS='-Wl,--whole-archive'; \ NOALLSYMSFLAGS='-Wl,--no-whole-archive'; \ - SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,-Bsymbolic -Wl,--out-implib,lib$(LIBNAME).dll.a"; \ + SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,--enable-auto-image-base -Wl,-Bsymbolic -Wl,--out-implib,lib$(LIBNAME).dll.a"; \ $(LINK_SO_O) link_a.cygwin: @ $(CALC_VERSIONS); \ @@ -253,8 +253,7 @@ link_a.cygwin: SHLIB_SOVER=-$(LIBVERSION); \ ALLSYMSFLAGS='-Wl,--whole-archive'; \ NOALLSYMSFLAGS='-Wl,--no-whole-archive'; \ - base=; [ $(LIBNAME) = "crypto" ] && base=-Wl,--image-base,0x6300; \ - SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared $$base -Wl,-Bsymbolic -Wl,--out-implib,lib$(LIBNAME).dll.a"; \ + SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,--enable-auto-image-base -Wl,-Bsymbolic -Wl,--out-implib,lib$(LIBNAME).dll.a"; \ [ -f apps/$$SHLIB$$SHLIB_SUFFIX ] && rm apps/$$SHLIB$$SHLIB_SUFFIX; \ [ -f test/$$SHLIB$$SHLIB_SUFFIX ] && rm test/$$SHLIB$$SHLIB_SUFFIX; \ $(LINK_SO_A) || exit 1; \ === end === -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1517] [PATCH] Makefile.shared: Link all shared libs with --enable-auto-image-base
On Mar 29 17:24, Corinna Vinschen via RT wrote: > Hi, > > the below patches are supposed to help with an annoying DLL problem on > Cygwin. [...] > > Index: Makefile.shared > === > RCS file: /home/cvs/cvsroot/src/openssl/Makefile.shared,v > retrieving revision 1.63 > diff -p -u -r1.63 Makefile.shared > --- Makefile.shared 25 Mar 2007 15:20:35 - 1.63 > +++ Makefile.shared 29 Mar 2007 09:34:02 - > @@ -255,7 +255,7 @@ link_o.cygwin: > SHLIB_SOVER=${LIBVERSION:+"-$(LIBVERSION)"}; \ > ALLSYMSFLAGS='-Wl,--whole-archive'; \ > NOALLSYMSFLAGS='-Wl,--no-whole-archive'; \ > - SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,-s,-Bsymbolic"; \ > + SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared > -Wl,--enable-auto-image-base -Wl,-s,-Bsymbolic"; \ > $(LINK_SO_O) > link_a.cygwin: > @ $(CALC_VERSIONS); \ > @@ -273,8 +273,7 @@ link_a.cygwin: > extras="$$extras rc.o"; \ > ALLSYMSFLAGS='-Wl,--whole-archive'; \ > NOALLSYMSFLAGS='-Wl,--no-whole-archive'; \ > - base=; [ $(LIBNAME) = "crypto" ] && base=-Wl,--image-base,0x6300; \ > - SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared $$base > -Wl,-s,-Bsymbolic -Wl,--out-implib,lib$(LIBNAME).dll.a $$extras"; \ > + SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared > -Wl,--enable-auto-image-base -Wl,-s,-Bsymbolic > -Wl,--out-implib,lib$(LIBNAME).dll.a $$extras"; \ > [ -f apps/$$dll_name ] && rm apps/$$dll_name; \ > [ -f test/$$dll_name ] && rm test/$$dll_name; \ > $(LINK_SO_A) || exit 1; \ Ping? Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1517] [PATCH] Makefile.shared: Link all shared libs with --enable-auto-image-base
Hi, On Apr 10 10:51, Corinna Vinschen wrote: > On Mar 29 17:24, Corinna Vinschen via RT wrote: > > Hi, > > > > the below patches are supposed to help with an annoying DLL problem on > > Cygwin. [...] > > > > Index: Makefile.shared > > === > > RCS file: /home/cvs/cvsroot/src/openssl/Makefile.shared,v > > retrieving revision 1.63 > > diff -p -u -r1.63 Makefile.shared > > --- Makefile.shared 25 Mar 2007 15:20:35 - 1.63 > > +++ Makefile.shared 29 Mar 2007 09:34:02 - > > @@ -255,7 +255,7 @@ link_o.cygwin: > > SHLIB_SOVER=${LIBVERSION:+"-$(LIBVERSION)"}; \ > > ALLSYMSFLAGS='-Wl,--whole-archive'; \ > > NOALLSYMSFLAGS='-Wl,--no-whole-archive'; \ > > - SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,-s,-Bsymbolic"; \ > > + SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared > > -Wl,--enable-auto-image-base -Wl,-s,-Bsymbolic"; \ > > $(LINK_SO_O) > > link_a.cygwin: > > @ $(CALC_VERSIONS); \ > > @@ -273,8 +273,7 @@ link_a.cygwin: > > extras="$$extras rc.o"; \ > > ALLSYMSFLAGS='-Wl,--whole-archive'; \ > > NOALLSYMSFLAGS='-Wl,--no-whole-archive'; \ > > - base=; [ $(LIBNAME) = "crypto" ] && base=-Wl,--image-base,0x6300; \ > > - SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared $$base > > -Wl,-s,-Bsymbolic -Wl,--out-implib,lib$(LIBNAME).dll.a $$extras"; \ > > + SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared > > -Wl,--enable-auto-image-base -Wl,-s,-Bsymbolic > > -Wl,--out-implib,lib$(LIBNAME).dll.a $$extras"; \ > > [ -f apps/$$dll_name ] && rm apps/$$dll_name; \ > > [ -f test/$$dll_name ] && rm test/$$dll_name; \ > > $(LINK_SO_A) || exit 1; \ > > Ping? it's about 6 weeks since I sent this patch. Is there some problem with it so it can't be applied? Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1517] [PATCH] Makefile.shared: Link all shared libs with --enable-auto-image-base
On May 19 22:02, Andy Polyakov via RT wrote: > please double-check http://cvs.openssl.org/chngview?cn=16237 Look good, thank you. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1517] [PATCH] Makefile.shared: Link all shared libs with --enable-auto-image-base
On May 19 21:54, Andy Polyakov via RT wrote: > >> Ping? > > > > it's about 6 weeks since I sent this patch. Is there some problem > > with it so it can't be applied? > > rules in question are used even in mingw build and why change this build > parameters? especially if we don't even know if suggested flag is > invalid in mingw ld context. i apologize it took time. a. No worries. The --enable-auto-image-base flag should work for mingw the same way as for Cygwin. The -mno-cygwin flag is not propagated to ld, and it's all about Windows DLLs anyway. However, the problem this flag is supposed to mitigate is a Cygwin-only problem. fork(2) just doesn't exist on MingW/native Windows. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1552] mingw patch for openssl-0.9.8e
On Jun 26 10:46, Roumen Petrov via RT wrote: > If a remember well borland compiler don't export variables. It seems to > me that gcc (mingw) don't export too. So that should use > OPENSSL_XXX_GLOBAL for both variables? gcc and FSF ld for i386 PE targets allow to export data symbols. A DATA keyword is supported in .def files for that purpose. > "Configure" set EXPORT_VAR_AS_FN for some win32 targets(msc, borlang, > mingw, but cigwin). Should "Configure" set EXPORT_VAR_AS_FN always if > build is for shared win32 platform ? Not for Cygwin, please. Corinna -- Corinna Vinschen Cygwin Project Co-Leader Red Hat __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[PATCH]: Configure OpenSSL for Cygwin with threading and shared library support
chmod 644 +$(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i; \ + else \ + c=`echo $$i | sed 's/^lib/cyg/'`; \ + cp $$c $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c; +\ + chmod 755 +$(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c; \ + cp $$i.a +$(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a; \ + chmod 644 +$(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a; \ + fi ); \ fi \ done; \ ( here="`pwd`"; \ Index: config === RCS file: /home/cvs/cvsroot/src/openssl096c/config,v retrieving revision 1.1.1.1 diff -u -p -r1.1.1.1 config --- config 2002/01/11 16:52:01 1.1.1.1 +++ config 2002/01/11 21:47:09 @@ -560,8 +560,7 @@ EOF # *-aix) OUT="aix-$CC" ;; # *-dgux) OUT="dgux" ;; mips-sony-newsos4) OUT="newsos4-gcc" ;; - *-*-cygwin32) OUT="CygWin32" - options="$options no-threads no-asm" ;; + *-*-cygwin32) OUT="CygWin32" ;; *) OUT=`echo $GUESSOS | awk -F- '{print $3}'`;; esac -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
[PATCH]: Configure OpenSSL for Cygwin with threading and shared library support
chmod 644 +$(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i; \ + else \ + c=`echo $$i | sed 's/^lib/cyg/'`; \ + cp $$c $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c; +\ + chmod 755 +$(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c; \ + cp $$i.a +$(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a; \ + chmod 644 +$(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a; \ + fi ); \ fi \ done; \ ( here="`pwd`"; \ Index: config === RCS file: /home/cvs/cvsroot/src/openssl096c/config,v retrieving revision 1.1.1.1 diff -u -p -r1.1.1.1 config --- config 2002/01/11 16:52:01 1.1.1.1 +++ config 2002/01/11 21:47:09 @@ -560,8 +560,7 @@ EOF # *-aix) OUT="aix-$CC" ;; # *-dgux) OUT="dgux" ;; mips-sony-newsos4) OUT="newsos4-gcc" ;; - *-*-cygwin32) OUT="CygWin32" - options="$options no-threads no-asm" ;; + *-*-cygwin32) OUT="CygWin32" ;; *) OUT=`echo $GUESSOS | awk -F- '{print $3}'`;; esac -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [PATCH]: Configure OpenSSL for Cygwin with threading and shared library support
On Sun, Jan 13, 2002 at 10:54:13AM -0800, Doug Kaufman wrote: > On Fri, 11 Jan 2002, Corinna Vinschen wrote: > > > I'm the OpenSSL maintainer for the Cygwin net distribution. > > When you release the new binary package, will your build exclude the > patented algorithms, so that it can be used without purchasing a > license in countries which recognize software patents such as the US? >Doug "purchasing a license"? I fear I don't understand. We have informed the BXA about offering OpenSSL on our site so what are you asking? I'm not a US citizen, perhaps that's the reason... Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [PATCH]: Configure OpenSSL for Cygwin with threading and shared library support
On Sun, Jan 13, 2002 at 09:00:37PM +0100, Richard Levitte - VMS Whacker wrote: > From: Corinna Vinschen <[EMAIL PROTECTED]> > > vinschen> - Makefile.org contains some changes to allow creating DLLs > vinschen> under Cygwin. The main change is adding a new target > vinschen> `do_cygwin-shared' which creates the cygcrypto.dll and cygssl.dll > vinschen> files as well as the corresponding link libraries libcrypto.dll.a > vinschen> and libssl.dll.a. > > Why the "cyg" prefix? All Cygwin shared libs have a `cyg' instead of a `lib' prefix. The reason is that in contrast to other operating systems you have multiple concurrent implementations of the same stuff, native, U/Win, Cygwin, Interix, etc. For that reason there was a community consensus to use another prefix to lower the chance to collide with one of the other implementations. > Also, since you're based in the US, you need to resend that patch with > a cc: to [EMAIL PROTECTED] Oh, I'm not based in the US. I'm in Germany and I'm not doing that as a Red Hat employee. The contribution of packages to the Cygwin net distribution is pure voluntary. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [PATCH]: Configure OpenSSL for Cygwin with threading and shared library support
On Sun, Jan 13, 2002 at 06:09:44PM -0800, Doug Kaufman wrote: > On Sun, 13 Jan 2002, Corinna Vinschen wrote: > > > On Sun, Jan 13, 2002 at 10:54:13AM -0800, Doug Kaufman wrote: > > > When you release the new binary package, will your build exclude the > > > patented algorithms, so that it can be used without purchasing a > > > license in countries which recognize software patents such as the US? > > >Doug > > > > "purchasing a license"? I fear I don't understand. We have informed > > the BXA about offering OpenSSL on our site so what are you asking? > > I'm not a US citizen, perhaps that's the reason... > > At least two of the algorithms used in openssl are subject to patents > [...] Thanks for the heads up. I got some additional feedback about these algorithms so I'm going to release 0.9.6c for Cygwin without IDEA, RC5 and MDC2. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [PATCH]: Configure OpenSSL for Cygwin with threading and shared library support
On Sun, Jan 13, 2002 at 09:19:58PM +0100, Corinna Vinschen wrote: > On Sun, Jan 13, 2002 at 09:00:37PM +0100, Richard Levitte - VMS Whacker wrote: > > From: Corinna Vinschen <[EMAIL PROTECTED]> > > > > vinschen> - Makefile.org contains some changes to allow creating DLLs > > vinschen> under Cygwin. The main change is adding a new target > > vinschen> `do_cygwin-shared' which creates the cygcrypto.dll and cygssl.dll > > vinschen> files as well as the corresponding link libraries libcrypto.dll.a > > vinschen> and libssl.dll.a. > > > > Why the "cyg" prefix? > > All Cygwin shared libs have a `cyg' instead of a `lib' prefix. > The reason is that in contrast to other operating systems you > have multiple concurrent implementations of the same stuff, > native, U/Win, Cygwin, Interix, etc. For that reason there > was a community consensus to use another prefix to lower the > chance to collide with one of the other implementations. > > > Also, since you're based in the US, you need to resend that patch with > > a cc: to [EMAIL PROTECTED] > > Oh, I'm not based in the US. I'm in Germany and I'm not doing > that as a Red Hat employee. The contribution of packages to > the Cygwin net distribution is pure voluntary. > > Corinna Hello, I don't want to be pushy but I'd like to know if I have to do something to get approval for that patch. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [PATCH]: Configure OpenSSL for Cygwin with threading and shared library support
On Tue, Jan 22, 2002 at 04:48:58PM +0100, Richard Levitte - VMS Whacker wrote: > From: Corinna Vinschen <[EMAIL PROTECTED]> > > vinschen> I don't want to be pushy but I'd like to know if I have to do > vinschen> something to get approval for that patch. > > I'm pretty sure I put it in my queue of patches to look through. > It'll probably get in as soon as I (or someone else) have done a > sanity check, unless it already has and I haven't noticed. Thanks for the feedback, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [PATCH]: Configure OpenSSL for Cygwin with threading and shared library support
On Tue, Jan 22, 2002 at 05:35:38PM +0100, Corinna Vinschen wrote: > On Tue, Jan 22, 2002 at 04:48:58PM +0100, Richard Levitte - VMS Whacker wrote: > > From: Corinna Vinschen <[EMAIL PROTECTED]> > > > > vinschen> I don't want to be pushy but I'd like to know if I have to do > > vinschen> something to get approval for that patch. > > > > I'm pretty sure I put it in my queue of patches to look through. > > It'll probably get in as soon as I (or someone else) have done a > > sanity check, unless it already has and I haven't noticed. > > Thanks for the feedback, > Corinna Hi, is this patch still in the loop? Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [PATCH]: Configure OpenSSL for Cygwin with threading and shared library support
On Wed, Feb 13, 2002 at 11:47:23AM +0100, Richard Levitte - VMS Whacker wrote: > One question remains before I go ahead: is it possible to have config > check the version of Cygwin and decide for another configuration > target if it's too old to provide thread safety. Since people with > older Cygwin installation may take the OpenSSL source and build it > there, it's of utmost importance that we get that part right, or we'll > just get a deluge of needless support requests from worried people. The output of uname -r (which is stored in $RELEASE) could be used for that. The following check could help: case $RELEASE in [bB]*|1.0|1.[12].*) options="$options no-threads no-asm no-shared" ;; esac This switches back to the old behaviour in case of the old B15-B20 versions and all versions from 1.0 up to 1.2.x while the new state is used for 1.3.x upwards. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [PATCH]: Configure OpenSSL for Cygwin with threading and shared library support
On Wed, Feb 13, 2002 at 02:54:46PM +0100, Richard Levitte - VMS Whacker wrote: > From: Corinna Vinschen <[EMAIL PROTECTED]> > > vinschen> case $RELEASE in > vinschen> [bB]*|1.0|1.[12].*) > vinschen> options="$options no-threads no-asm no-shared" ;; > vinschen> esac > vinschen> > vinschen> This switches back to the old behaviour in case of the old > vinschen> B15-B20 versions and all versions from 1.0 up to 1.2.x while > vinschen> the new state is used for 1.3.x upwards. > > Thanks for the info. > > OK, your patch is applied and slightly modified, and will be commited > in a moment. Please download the next snapshot (both of 0.9.6x and > 0.9.7-dev) and try them out. The next snapshot will be built around > 0100 CET tomorrow morning. Thanks! I'll test it. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Rename "CygWin32" to "Cygwin" acceptable?
Hi, in OpenSSL, Cygwin is called "CygWin32" throughout. But the correct name and case usage of Cygwin is "Cygwin". No upper case Win, no 32. Question: Would renaming in the scripts be an acceptable patch? Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
[PATCH]: Latest `config' has syntax error
Hi, the patch to config has a syntax error. Patch below. Corinna Index: config === RCS file: /home/cvs/cvsroot/src/openssl/config,v retrieving revision 1.95 diff -u -p -r1.95 config --- config 2002/02/13 14:44:31 1.95 +++ config 2002/02/14 10:22:40 @@ -336,10 +336,11 @@ case "${SYSTEM}:${RELEASE}:${VERSION}:${ case "$RELEASE" in [bB]*|1.0|1.[12].*) echo "${MACHINE}-whatever-cygwin32_pre1.3"; exit 0 + ;; *) echo "${MACHINE}-whatever-cygwin32"; exit 0 - ;; - + ;; + esac esac # -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [PATCH]: Latest `config' has syntax error
On Thu, Feb 14, 2002 at 11:24:38AM +0100, Corinna Vinschen wrote: > Hi, > > the patch to config has a syntax error. Patch below. Oh, I've just seen that a patch has already been applied to openssl-0.9.6-stable-SNAP-20020213. But the problem seems still to be in CVS HEAD. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Build problem in CVS HEAD and openssl-0.9.6-stable-SNAP-20020213
Hi, I'm trying to build CVS HEAD as well as openssl-0.9.6-stable-SNAP-20020213 with the following options no-idea no-rc5 no-mdc2 as I'd do for a Cygwin release. Make complains in crypto/hmac: make[2]: *** No rule to make target `../../include/openssl/idea.h', needed by `hmac.o'. Stop. The dependencies in crypto/hmac/Makefile.ssl make hmac.o dependend on idea.h, rc5.h and mdc2.h but that seems pretty unnecessary since hmac.o is build even w/o these files. After tweaking Makefile in this directory, the build stops with the same problem in crypto/rsa so I assume it's sort of a generic problem? Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Rename "CygWin32" to "Cygwin" acceptable?
On Thu, Feb 14, 2002 at 11:37:30AM +0100, Richard Levitte - VMS Whacker wrote: > From: Corinna Vinschen <[EMAIL PROTECTED]> > > vinschen> in OpenSSL, Cygwin is called "CygWin32" throughout. But the > vinschen> correct name and case usage of Cygwin is "Cygwin". No upper > vinschen> case Win, no 32. > > Ah, OK, I'll fix that. Thanks! Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
[RFA] New script util/cygwin.sh
Hi, I've attached a file which I'd like to get into the ssl source tree. It's a script which automatically generates the binary runtime and development tar archives as they are provided by the Cygwin net distribution. It's definitely not a "must have" but I thought it would be helpful for others to reproduce a Cygwin release version... Thanks for considering, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] cygwin.sh Description: Bourne shell script
Re: [ANNOUNCE] OpenSSL 0.9.6d beta 1 released
On Fri, Apr 19, 2002 at 02:16:20PM -0700, Tim Rice wrote: > On Wed, 17 Apr 2002, Richard Levitte - VMS Whacker wrote: > > > The release of OpenSSL 0.9.6d is scheduled for Tuesday 2002-04-30. > > To make sure that it will work correctly, please test this version > > (especially on less common platforms), and report any problems to > > <[EMAIL PROTECTED]>. > > Solaris 7 solaris-sparcv8-gcc OK > Solaris 8 solaris-sparcv8-gcc OK > Caldera eDesktop 2.4 linux-elf OK > Caldera eServer 2.3.1 linux-elf OK > Caldera OpenLinux 3.1.1 linux-elf OK > RedHat 6.2linux-elf OK > SCO 3.2v4.2 sco3-gccOK > SCO 5.0.4 sco5-cc-pentium OK > SCO 5.0.4 sco5-gccOK > SCO 5.0.6 sco5-cc-pentium OK > SCO 5.0.6 sco5-gccOK > UnixWare 2.03 unixware-2.0-pentiumOK > UnixWare 2.1.3unixware-2.1-pentiumOK > UnixWare 7.1.1unixware-7-pentium OK > OpenUNIX 8.0.0OpenUNIX-8-pentium OK Cygwin i686-whatever-cygwinOK Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: OpenSSL 0.9.6d beta 1 testing results
On Fri, Apr 19, 2002 at 06:36:47PM -0400, Michael Kobar wrote: > --- Richard Levitte - VMS Whacker wrote: > > > > The release of OpenSSL 0.9.6d is scheduled for Tuesday 2002-04-30. > > To make sure that it will work correctly, please test this version > > (especially on less common platforms), and report any problems to > > <[EMAIL PROTECTED]>. > > All built with ./config and ./config no-idea no-rc5 > - > Solaris 2.6solaris-sparcv8-gcc Both OK > Solaris 7 solaris-x86-gcc Both OK > Solaris 7 solaris-sparcv9-gcc Both OK > RedHat 6.2 linux-elf Both OK > RedHat 7.1 linux-elf Both OK > Cygwin 1.3.10-1cygwinBoth FAILED > > The cygwin failures were sent to the list earlier today. I've build with ./config --prefix=/usr shared no-idea no-rc5 no-mdc2 as the Cygwin net distribution, and for testing purposes with ./config and both builds ran fine. Which Cygwin version are you using? Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Revised DOS patch for openssl-0.9.7
On Sat, Apr 20, 2002 at 03:09:18PM -0700, Doug Kaufman wrote: > Here is a revised patch for DOS (and other platforms that don't > support symbolic links). With this patch openssl-0.9.7 snapshot from > 19 April 2002 compiles without problems. Make test show no errors, > except inability to load 'random state' (related to no built-in > method of gathering entropy). I also built on Cygwin 1.3.2 with the > patch applied. Make and "make test" show no problems. The DJGPP build > was configured with "no-threads no-idea no-rc5". Cygwin build was > configured with "no-idea no-rc5". > Doug I don't understand what this patch has to do with Cygwin since Cygwin supports symbolc links since years now. Could you explain, please? Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Revised DOS patch for openssl-0.9.7
On Sun, Apr 21, 2002 at 12:53:06AM +0200, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Sun, 21 Apr 2002 00:42:24 >+0200, Corinna Vinschen <[EMAIL PROTECTED]> said: > > vinschen> I don't understand what this patch has to do with Cygwin since Cygwin > vinschen> supports symbolc links since years now. Could you explain, please? > > Note that Doug's changes shouldn't affect CygWin in that case. I > believe that he mentioned CygWin to mean "it didn't get broken by what > I did". I see. Thanks for the explanation, Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Revised DOS patch for openssl-0.9.7
On Sun, Apr 21, 2002 at 12:42:24AM +0200, Corinna Vinschen wrote: > On Sat, Apr 20, 2002 at 03:09:18PM -0700, Doug Kaufman wrote: > > Here is a revised patch for DOS (and other platforms that don't > > support symbolic links). With this patch openssl-0.9.7 snapshot from > > 19 April 2002 compiles without problems. Make test show no errors, > > except inability to load 'random state' (related to no built-in > > method of gathering entropy). I also built on Cygwin 1.3.2 with the > > patch applied. Make and "make test" show no problems. The DJGPP build > > was configured with "no-threads no-idea no-rc5". Cygwin build was > > configured with "no-idea no-rc5". > > Doug > > I don't understand what this patch has to do with Cygwin since Cygwin > supports symbolc links since years now. Could you explain, please? Oh and, btw., current Cygwin is 1.3.10, 1.3.11 coming soon. Why are you testing with a version which is almost a year old? Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Revised DOS patch for openssl-0.9.7
On Mon, Apr 22, 2002 at 08:21:42AM +0200, Richard Levitte - VMS Whacker wrote: > dkaufman> After running "make depend", the next error is due to lack > dkaufman> of "bn_sub_part_words". This was addressed in my patch for > dkaufman> both DJGPP and Cygwin. > > Ah, so that's what that is about. > > Tell me something, do you get those same errors when you configure > without excluding any algorithm (basically, just a plain "./config")? > If not, it might be worth looking into what gets different with the > exclusions... This is for the current version as in CVS. I did make depend && ./config && make on Cygwin and the above errors still strike: gcc -o openssl -DMONOLITH -I.. -I../include -DOPENSSL_SYSNAME_CYGWIN32 -DOPENSSL_THREADS -DDSO_WIN32 -DOPENSSL_NO_KRB5 -DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O3 -m486 -Wall openssl.o verify.o asn1pars.o req.o dgst.o dh.o dhparam.o enc.o passwd.o gendh.o errstr.o ca.o pkcs7.o crl2p7.o crl.o rsa.o rsautl.o dsa.o dsaparam.o ecdsa.o ecdsaparam.o x509.o genrsa.o gendsa.o s_server.o s_client.o speed.o s_time.o apps.o s_cb.o s_socket.o app_rand.o version.o sess_id.o ciphers.o nseq.o pkcs12.o pkcs8.o spkac.o smime.o rand.o engine.o ocsp.o -L.. -lssl -L.. -lcrypto ; \ fi ../libcrypto.a(bn_mul.o)(.text+0x44b):bn_mul.c: undefined reference to `bn_sub_part_words' [...] so it's pretty generic. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Revised DOS patch for openssl-0.9.7
On Mon, Apr 22, 2002 at 10:11:08AM +0200, Corinna Vinschen wrote: > gcc -o openssl -DMONOLITH -I.. -I../include -DOPENSSL_SYSNAME_CYGWIN32 >-DOPENSSL_THREADS -DDSO_WIN32 -DOPENSSL_NO_KRB5 -DTERMIOS -DL_ENDIAN >-fomit-frame-pointer -O3 -m486 -Wall openssl.o verify.o asn1pars.o req.o dgst.o dh.o >dhparam.o enc.o passwd.o gendh.o errstr.o ca.o pkcs7.o crl2p7.o crl.o rsa.o rsautl.o >dsa.o dsaparam.o ecdsa.o ecdsaparam.o x509.o genrsa.o gendsa.o s_server.o s_client.o >speed.o s_time.o apps.o s_cb.o s_socket.o app_rand.o version.o sess_id.o ciphers.o >nseq.o pkcs12.o pkcs8.o spkac.o smime.o rand.o engine.o ocsp.o -L.. -lssl -L.. >-lcrypto ; \ > fi > ../libcrypto.a(bn_mul.o)(.text+0x44b):bn_mul.c: undefined reference to >`bn_sub_part_words' > [...] > > so it's pretty generic. The problem is line 61 in crypto/bn/bn_mul.c: #if defined(OPENSSL_NO_ASM) || !(defined(__i386) || defined(__i386__))/* ... This is used to decide if `bn_sub_part_words' should be defined or not. Unfortunately this function doesn't exist in Cygwin but the __i386 and __i386__ defines are set by gcc's config file. Two questions: - Why is the above `#if' used for bn_sub_part_words() but not for bn_add_part_words()? - Should Cygwin always be build with "no-asm", perhaps? Or would it make more sense to redefine the above `#if'? Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Revised DOS patch for openssl-0.9.7
On Mon, Apr 22, 2002 at 07:51:27AM -0700, Doug Kaufman wrote: > On Mon, 22 Apr 2002, Corinna Vinschen wrote: > > > > ../libcrypto.a(bn_mul.o)(.text+0x44b):bn_mul.c: undefined reference to >`bn_sub_part_words' > > > [...] > > > > > > so it's pretty generic. > > > > The problem is line 61 in crypto/bn/bn_mul.c: > > > > #if defined(OPENSSL_NO_ASM) || !(defined(__i386) || defined(__i386__))/* ... > > I also built 0.9.7 under Cygwin, as "./Configure Cygwin", then "make" > and get the same error. See a fix in previously submitted patch. Oh, sorry, I've missed that part of the patch. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
util/cygwin.sh has wrong permissions
Hi, just a heads up, the util/cygwin.sh shell script has wrong permissions in the CVS repository as well as in the 0.9.6d release. The x bit isn't set, unfortunately. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #12] Re: Revised DOS patch for openssl-0.9.7
On Thu, May 30, 2002 at 08:00:43AM +0200, Richard Levitte via RT wrote: > > Hmm, out of curiosity, Ulf Moeller made a change to Configure > (revisions 1.328 and 1.314.2.12) that affects Cygwin. Could it be > that his change fixes this problem? I'm getting the following message in link stage now when building from current CVS: gcc -shared -o cygcrypto.dll -Wl,-Bsymbolic -Wl,--whole-archive libcrypto.a -W l,--out-implib,libcrypto.dll.a -Wl,--no-whole-archive -L. libcrypto.a(bn86-out.o)(.text+0x738):bn-586.s: multiple definition of `bn_sub_pa rt_words' libcrypto.a(bn_mul.o)(.text+0x0):bn_mul.c: first defined here Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #12] Re: Revised DOS patch for openssl-0.9.7
On Thu, May 30, 2002 at 10:30:10AM +0200, Richard Levitte - VMS Whacker wrote: > vinschen> I'm getting the following message in link stage now when building > vinschen> from current CVS: > vinschen> > vinschen> gcc -shared -o cygcrypto.dll -Wl,-Bsymbolic -Wl,--whole-archive >libcrypto.a -W > vinschen> l,--out-implib,libcrypto.dll.a -Wl,--no-whole-archive -L. > vinschen> libcrypto.a(bn86-out.o)(.text+0x738):bn-586.s: multiple definition of >`bn_sub_pa > vinschen> rt_words' > vinschen> libcrypto.a(bn_mul.o)(.text+0x0):bn_mul.c: first defined here > > Did you 'make clean'? $ make clean [...] $ config --prefix=/usr shared no-idea no-rc5 no-mdc2 [...] $ make [...] Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #12] Re: Revised DOS patch for openssl-0.9.7
On Thu, May 30, 2002 at 11:33:42AM +0200, Richard Levitte - VMS Whacker wrote: > In message <[EMAIL PROTECTED]> on Thu, 30 May 2002 11:01:53 >+0200, Corinna Vinschen <[EMAIL PROTECTED]> said: > > vinschen> On Thu, May 30, 2002 at 10:30:10AM +0200, Richard Levitte - VMS Whacker >wrote: > vinschen> > vinschen> I'm getting the following message in link stage now when >building > vinschen> > vinschen> from current CVS: > vinschen> > vinschen> > vinschen> > vinschen> gcc -shared -o cygcrypto.dll -Wl,-Bsymbolic >-Wl,--whole-archive libcrypto.a -W > vinschen> > vinschen> l,--out-implib,libcrypto.dll.a -Wl,--no-whole-archive -L. > vinschen> > vinschen> libcrypto.a(bn86-out.o)(.text+0x738):bn-586.s: multiple >definition of `bn_sub_pa > vinschen> > vinschen> rt_words' > vinschen> > vinschen> libcrypto.a(bn_mul.o)(.text+0x0):bn_mul.c: first defined here > vinschen> > > vinschen> > Did you 'make clean'? > vinschen> > vinschen> $ make clean > vinschen> [...] > vinschen> $ config --prefix=/usr shared no-idea no-rc5 no-mdc2 > vinschen> [...] > vinschen> $ make > vinschen> [...] > > And you're running with a fresh bn_mul.c, or does it implement > bn_sub_part_words() for Cygwin (i.e., do you have the proposed patch > applied)? > > In bn_mul.c, line 69 should read like this: > > #if defined(OPENSSL_NO_ASM) || !(defined(__i386) || defined(__i386__))/* Assembler >implementation exists only for x86 */ I'm sorry. I had the patch applied. Reverting to the above line fixed my problem. Now it builds fine on Cygwin. Corinna -- Corinna Vinschen Cygwin Developer Red Hat, Inc. mailto:[EMAIL PROTECTED] __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]