initialize variables patch for bn_nist.c

ese variables before using, I would like to apply the patch. OK ? Here is original topic on GitHub. https://github.com/libressl-portable/openbsd/pull/19 Best Regards, kinichiro inoguchi diff --git src/lib/libssl/src/crypto/bn/bn_nist.c src/lib/libssl/src/crypto/bn/bn_nist.c index 4d3a612..a87309

Parsing tlsext strictly in libressl

Hi, I verified that regression test src/regress/lib/libssl/unit/tls_ext_alpn.c fails on these cases; - proto_invalid_len5, 7, 8 - proto_invalid_missing1 - 5 - proto_invalid_missing8, 9 To correct these failures, ssl_parse_clienthello_tlsext() and ssl_parse_serverhello_tlsext() in

Re: Add libtls functionality for OCSP, and OCSP stapling support

to verify by OCSP stapling will be always single ...) Best regards, Kinichiro Inoguchi

add error check to ocsp_test.c

Hi, I would like to add error check for CAfile loading since some OS doesn't have /etc/ssl/cert.pem. Best regards, Kinichiro Inoguchi diff --git src/regress/lib/libcrypto/ocsp/ocsp_test.c src/regress/lib/libcrypto/ocsp/ocsp_test.c index 8867536..31594fa 100644 --- src/regress/lib/libcrypto/ocsp

openssl s_time error with -time option

m git repo of libressl-portable. Best regards, Kinichiro Inoguchi diff --git src/usr.bin/openssl/apps.c src/usr.bin/openssl/apps.c index d1d0d14..525166e 100644 --- src/usr.bin/openssl/apps.c +++ src/usr.bin/openssl/apps.c @@ -2218,7 +2218,8 @@ options_parse(int argc, char **argv, struct option *o

Re: Fix boundary issue in chacha code

Sorry for my misunderstanding, and thanks for teaching me. I had read C99 standard document. http://www.open-std.org/jtc1/sc22/wg14/www/docs/n1124.pdf On p.130 6.7.8 Initialization "32 EXAMPLE 8 The declaration" says, - char s[] = "abc", t[3] = "abc"; defines ‘‘plain’’ char array

LibreSSL selects weak digest for (EC)DH

Hi, I would like to fix this SNI issue. reported by @davidben https://github.com/libressl-portable/openbsd/issues/69 #3560: OpenSSL selects weak digest for (EC)DH https://rt.openssl.org/Ticket/Display.html?id=3560 original OpenSSL commit is here.

Re: LibreSSL selects weak digest for (EC)DH

Thanks, Brent. I appreciate if you commit this. Kinichiro

Re: Explicitly cast the return variable in tls_load_file()

I could have an answer that this compilation error was a bug of compiler, and that bug will be tracked. https://software.intel.com/en-us/forums/intel-c-compiler/topic/698109 I saw the type of buf was changed in cvs, then I can avoid this compilation problem. Thanks. Kinichiro

Re: Explicitly cast the return variable in tls_load_file()

Thanks, that is apparently better than I suggested and reasonable. And I confirmed it can also avoid the issue. I appreciate if this is applied. And, yes I believe compilation error is bug of compiler, not source code. I posted this compilation error to Intel C++ compiler forum on Sunday.

Explicitly cast the return variable in tls_load_file()

I would like to cast the return variable explicitly in tls_load_file(). This fix also avoiding Intel C++ compiler "assertion failed" described here. https://github.com/libressl-portable/portable/issues/209#issuecomment-249587024 ok ? Index: tls_util.c

LibreSSL should not allow too many consecutive warning alerts

Hi, This patch is for CVE-2016-8610. See http://seclists.org/oss-sec/2016/q4/224 . - Don't allow too many consecutive warning alerts up to MAX_WARN_ALERT_COUNT OpenSSL seems not to fix this issue on branch 1.0.1. Then I refer to these 2 commits on branch 1.0.2. - Don't allow too many

Re: chachatest.c patch

Hi, I have a question for this patch. Do you refer to the document below ? https://tools.ietf.org/html/draft-strombergson-chacha-test-vectors-01 Best regards, Kinichiro Inoguchi On Sat, Mar 18, 2017 at 07:56:19AM -0700, Steven Roberts wrote: > TC3 contained the data for TC4. > TC4 con

Re: [libcrypto] Don't build empty ecp_nistp* objects

Absolutely. ok inoguchi@ On Mon, Jul 17, 2017 at 06:26:30AM -0500, Brent Cook wrote: > OPENSSL_NO_EC_NISTP_64_GCC_128 has been defined in opensslfeatures.h for a > long time, which effectively means that ecp_nistp* are all empty files. So, > there is no reason to build them in the first place.

Re: Reported problem: postfix, libressl 2.6.x, DHE-RSA-AES256-SHA

Hi, In my OpenBSD 6.2, LibreSSL server and OpenSSL client can communicate successfully via cipher DHE-RSA-AES256-SHA. -- # Server LibreSSL 2.6.3 $ openssl s_server -cert server.pem

Re: Reported problem: postfix, libressl 2.6.x, DHE-RSA-AES256-SHA

Same kind of issue had been raised for Postfix and LibreSSL 2.2.2 2 years ago. http://postfix.1071664.n5.nabble.com/SSL-accept-errors-after-recent-upgrade-to-LibreSSL-2-2-2-td78774.html In this case, SSLv2 caused some wrong behavior.

Re: LibreSSL patch for AIX

Hi, I would like to see that patch. Since I thought autotools could adjust the differences between OSs, I didn't know that a special patch needed for AIX. Best regards, Kinichiro Inoguchi

Re: openssl(1) s_socket.c: don't leak `name' (CID #154702)

I feel that "error case free" should be done in do_accept() rather than caller. After strdup(), there are 2 "return (0)". How about adding "free(*host)" before these 2 "return (0)" ? I worried that error return occurs before strdup() in do_accept(). On Sun, Aug 19, 2018 at 10:40:55AM +0200, Theo

Re: CID #183499: don't leak db in RSA_padding_check_PKCS1_OAEP()

It looks good to me. OK inoguchi@ On Sun, Aug 19, 2018 at 08:44:24AM +0200, Theo Buehler wrote: > Coverity complains about the case where EVP_Digest() fails, but there > are a couple more. > > Index: rsa/rsa_oaep.c > === > RCS

Re: openssl(1) s_socket.c: don't leak `name' (CID #154702)

OK @inoguchi 2018/08/19 22:44 "Theo Buehler" : > On Sun, Aug 19, 2018 at 09:53:32PM +0900, Kinichiro Inoguchi wrote: > > I feel that "error case free" should be done in do_accept() rather than > caller. > > After strdup(), there are 2 "return (0)&

Re: Replace getprogname() to argv[0] in bnaddsub

ectory hierarchy into a context that bounces stderr to > the wrong place. > > Instead, the Windows compat should have a have a getprogname() stub > which does the right thing. > > Kinichiro Inoguchi wrote: > > > To run regress bnaddsub on Windows, I would like to supe

Replace getprogname() to argv[0] in bnaddsub

To run regress bnaddsub on Windows, I would like to supersede getprogname with argv[0] since it is not on Windows. OK ? Index: regress/lib/libcrypto/bn/addsub/bnaddsub.c === RCS file:

Re: Replace getprogname() to argv[0] in bnaddsub

__progname is not portable, and I once requested to eliminate it. For porable reason, I prefer to using argv[0]. Anyway, thanks for your comment. Kinichiro Inoguchi On Tue, Jul 17, 2018 at 12:01:59PM -0300, Gleydson Soares wrote: > On Tue, Jul 17, 2018 at 10:09:37PM +0900, Kinichiro Inogu

Re: use BN_swap_ct() instead of BN_consttime_swap() in ec_GF2m_montgomery_point_multiply()

looks good to me. ok inoguchi@ On Sat, Jul 14, 2018 at 02:46:17PM +0200, Theo Buehler wrote: > The new BN_swap_ct() API is an improved version of the public > BN_consttime_swap() function: it allows for error checking doesn't > assert() and has fewer assumptions on the input. > > This

Re: BN_swap_ct() use size_t for byte count

I checked this diff libressl portable build. ok inoguchi@ On Sat, Jul 14, 2018 at 02:30:12PM +0200, Theo Buehler wrote: > As pointed out by jsing, using size_t for nwords would be more > appropriate for the new internal API BN_swap_ct(). Let's switch to it > and cast to an int internally after

Re: make X509_CRL_METHOD_free() NULL-safe

OK inoguchi@

Re: [PATCH v2] Provide static_ASN1_*(). From OpenSSL 1.1.0 API

Hi, This patch was applied, and thanks for your help. Sorry for late, since we couldn't have time to review. On Sat, Aug 17, 2019 at 07:11:28AM +0300, Stefan Strogin wrote: > v1->v2: > Use correct static_ASN1_ITEM_start macros instead of ASN1_ITEM_start in > static_ASN1_SEQUENCE_END_ref

Re: [PATCH] Extend OAEP support

I thought this patch could give an ability to handle OAEP label with openssl(1) pkeyutl command, and encryption works fine, but decryption fails. -- openssl genrsa -out rsakey.pem echo "abcd" | openssl pkeyutl -encrypt -inkey

Re: [PATCH] Provide static_ASN1_*(). From OpenSSL 1.1.0 API.

Hi, +#define static_ASN1_SEQUENCE_END_ref(stname, tname) \ + ;\ + ASN1_ITEM_start(tname) \ I think this should be "static_ASN1_ITEM_start" instead "ASN1_ITEM_start".

Re: openssl.1: Tag command names

> I like the idea! I agree. > To me it would be more logical to put .Tg above .Sh, but that is a minor > thing. I also think that it would better to place .Tg above .Sh . On Mon, Feb 17, 2020 at 11:20:34PM +0100, Remi Locherer wrote: > On Mon, Feb 17, 2020 at 05:19:27PM +0100, Klemens Nanni

Fix manual description in SSL_CTX_add_extra_chain_cert.3

I think both SSL_CTX_get_extra_chain_certs and SSL_CTX_get_extra_chain_certs_only should be described here. ok? Index: SSL_CTX_add_extra_chain_cert.3 === RCS file: /cvs/src/lib/libssl/man/SSL_CTX_add_extra_chain_cert.3,v retrieving

Re: Fix manual description in SSL_CTX_add_extra_chain_cert.3

and had not took consider about others. I should have thought about this from comprehensive perspective of view. I would like to abandon this patch for now. Regards, Kinichiro Inoguchi On Sun, Jan 12, 2020 at 10:46:26AM +0100, Ingo Schwarze wrote: > Hello Kinichiro-san, > > Kinichiro Inogu

Re: openssl(1) x509, change in serial number output between 6.5 and 6.6

On Thu, Apr 09, 2020 at 07:44:51PM +0100, Stuart Henderson wrote: > On 2020/04/09 20:13, Theo Buehler wrote: > > On Thu, Apr 09, 2020 at 05:56:55PM +0100, Stuart Henderson wrote: > > > Not new - this happened somewhere between 6.5 and 6.6 - but some > > > certificates are now showing up with bad

Re: [PATCH] gostr341001: support unwrapped private keys support

> There is no English specification for GOST PKCS8 files yet, > unfortunately. You can find similar pieces of code in OpenSSL's GOST > engine (https://github.com/gost-engine/engine/blob/master/gost_ameth.c#L347) > and in GnuTLS >

Re: [PATCH] ec: add support for several more GOST curves

Hi, I have a 3 questions, - parameter set values for Twisted Edwards - description in _ec_list_element_st - naming about object identifier details are described below. On Thu, Mar 26, 2020 at 09:25:57PM +0300, dbarysh...@gmail.com wrote: > From: Dmitry Baryshkov > > Add support for GOST

Re: [PATCH 2/2] gost: use ECerror to report EC errors

I had checked this by portable build and all regresses passed. I'm ok with this diff. On Thu, Mar 26, 2020 at 09:28:02PM +0300, dbarysh...@gmail.com wrote: > From: Dmitry Baryshkov > > GOST code uses GOSTerror(EC_R_foo) to report several errors. Use > ECerror(EC_R_foo) instead to make error

Re: [PATCH 1/2] gost: add missing error reporting

I had checked this by portable build and all regresses passed. I'm ok with this diff. On Thu, Mar 26, 2020 at 09:28:01PM +0300, dbarysh...@gmail.com wrote: > From: Dmitry Baryshkov > > Add few more error reports to help debugging. > > Sponsored by ROSA Linux. > > Signed-off-by: Dmitry

Re: [PATCH v2 1/2] ec: add support for several more GOST curves

Verified added curve parameters _EC_GOST_2012_256_TC26_A and _EC_GOST_2012_512_TC26_C are equivalent to the definition of https://tools.ietf.org/html/rfc7836#appendix-A.2 . Verified added curve parameter _EC_GOST_2012_512_Test is equivalent to

Re: [PATCH v2 2/2] gost: populate params tables with new curves

Hi, I have 2 questions. In GostR3410_512_params[], "A" and "TCA" have the same NID, "B" and "TCB" too. I thought these were redundant, but are there any reasons for this ? In GostR3410_512_params[], don't you need the record for NID_id_tc26_gost_3410_12_512_paramSetTest ? Best regards, On

Re: [PATCH] gostr341001: support unwrapped private keys support

Hi, Where can we see the specifcation for these 3 different format, wrapped in OCTET STRING, INTEGER and unwrapped but masked ? I tried to find but couldn't.

Re: [PATCH v3 2/2] gost: populate params tables with new curves

Confirmed that portable build and regresses succeeded. I'm ok with this patch. On Sun, Mar 29, 2020 at 02:48:05PM +0300, Dmitry Baryshkov wrote: > Allow users to specify new curves via strings. > > Sponsored by ROSA Linux > > Signed-off-by: Dmitry Baryshkov > --- >

Re: Update Windows getentropy implementation

On Mon, Nov 09, 2020 at 01:10:51PM -0600, Brent Cook wrote: > > This updates the getentropy implementation for Windows to use the newer > "Cryptography Next Generation APIs", replacing CryptGenRandom, which > already has been removed from applications built for the Windows Store. > > Tested with

Re: openssl(1): implement naccept

This builds fine and works good. ok inoguchi@ I have one comment additionally what jmc@ mentioned. On Sun, Aug 29, 2021 at 01:10:56PM +0100, Jason McIntyre wrote: > On Sun, Aug 29, 2021 at 02:00:44PM +0200, Theo Buehler wrote: > > Terminate the s_server after n clients connected to it. This is >

Compare pointer value with NULL in openssl(1) pkcs12

I would like to do some clean up for openssl(1) pkcs12. This diff changes pointer value checking to explicit comparison with NULL, and no functional changes here. This works fine for me on my local pc. ok? Index: pkcs12.c === RCS

Re: Compare pointer value with NULL in openssl(1) pkcs12

On Sat, Apr 23, 2022 at 08:31:50AM +0200, Theo Buehler wrote: > On Sat, Apr 23, 2022 at 01:45:12PM +0900, Kinichiro Inoguchi wrote: > > > I would like to do some clean up for openssl(1) pkcs12. > > This diff changes pointer value checking to explicit comparison with NULL, >

Check sk_push return value in openssl(1) pkcs12

This adds return value check for sk_X509_push. In error case, allocated memories are freed at 'export_end:'. After this diff, I would like to add more another return value check. ok? Index: pkcs12.c === RCS file: