You are complete right. I disabled TLSEXT, because of the code size.
Up to now, the result is very clear now.
Lee
On Fri, Jan 10, 2014 at 12:30 PM, Dave Thompson <dthomp...@prinpay.com>wrote:
> (Due to Outlook upgrade, I can no longer reply inline to richtext. Sorry.)
>
>
>
> Is it possible your failing client openssl was built with –no-tlsext ?
>
> Both ServerHello replies do contain the RI extension (because the
>
> cipherlist contained SCSV_ERI); if the code to parse extensions
>
> isn’t executed, then the packet length will mismatch.
>
>
>
> Maybe if TLSEXT is disabled client shouldn’t send SCSV_ERI,
>
> since it is promising capability we can’t deliver on. OTOH,
>
> more and more TLS usage depends on extensions, so
>
> disabling TLSEXT is increasingly a bad idea.
>
>
>
>
>
> *From:* owner-openssl-us...@openssl.org [mailto:
> owner-openssl-us...@openssl.org] *On Behalf Of *yueyue papa
> *Sent:* Thursday, January 09, 2014 22:21
> *To:* openssl-users@openssl.org
> *Subject:* *** Spam *** about ssl3_get_server_hello
>
>
>
> The 0.9.8h openSSL version is working correct in my system. Recently, I
> updated to 1.0.1e openSSL, I meet the SSL connection failure.
>
>
>
> The problem is in here
>
>
>
> case SSL3_ST_CR_SRVR_HELLO_A:
>
> case SSL3_ST_CR_SRVR_HELLO_B:
>
> ret=ssl3_get_server_hello(s);
>
> if (ret <= 0)
>
>
>
> ###### My embedded environment connect log#########
>
> ssl3_connect state[5000]
>
> ssl3_connect state[1110]
>
> ssl sock_write(), want in1 = 66
>
> sock_write=: length = 66 [0x42]
>
> 0x20a66fcb 16 03 00 00 3D 01 00 00 - 39 03 00 52 CF 0B D8 E8 |
> ....=...9..R....
>
> 0x20a66fdb C2 06 F6 5B AD 42 CB 65 - 29 1D 76 3C 82 EF 36 F3 |
> ...[.B.e).v<..6.
>
> 0x20a66feb 81 89 CE AD 68 24 DF 98 - 7C A8 64 00 00 12 00 35 |
> ....h$..|.d....5
>
> 0x20a66ffb 00 0A 00 2F 00 05 00 04 - 00 09 00 08 00 03 00 FF |
> .../............
>
> 0x20a6700b 01 00 - | ..
>
> ssl3_connect state[1120]
>
> ssl sock_read(), want out1 = 5
>
> sock_read=: length = 5 [0x5]
>
> 0x20a62e7b 16 03 00 00 51 - |
> ....Q
>
> ssl sock_read(), want out1 = 81
>
> sock_read=: length = 81 [0x51]
>
> 0x20a62e80 02 00 00 4D 03 00 52 CF - 60 34 AE D2 CF C7 5D A7 |
> ...M..R.`4....].
>
> 0x20a62e90 15 71 39 36 08 8B D3 DB - C8 E8 15 23 94 AE EB E6 |
> .q96.......#....
>
> 0x20a62ea0 3E CE 3B 55 63 60 20 2F - 3D 93 96 71 B2 FE 3A 10 |
> >.;Uc` /=..q..:.
>
> 0x20a62eb0 53 31 84 59 06 B2 5F 09 - 62 A0 89 91 A3 20 E2 9A |
> S1.Y.._.b.... ..
>
> 0x20a62ec0 5F 32 24 75 4A 28 5E 00 - 05 00 00 05 FF 01 00 01 |
> _2$uJ(^.........
>
> 0x20a62ed0 00 - | .
>
> #############
>
>
>
> ######My PC connect log #########
>
> CONNECTED(00000180)
>
> write to 0x1fee3f0 [0x1e72fab] (152 bytes => 152 (0x98))
>
> 0000 - 16 03 00 00 93 01 00 00-8f 03 00 52 cf 5a 9e ee ...........R.Z..
>
> 0010 - f2 c6 eb ac bd 82 4d cf-34 17 88 5e e7 5d e4 f3 ......M.4..^.]..
>
> 0020 - ba b8 d0 04 12 88 bf 50-e4 4d 57 00 00 68 c0 14 .......P.MW..h..
>
> 0030 - c0 0a c0 22 c0 21 00 39-00 38 00 88 00 87 c0 0f ...".!.9.8......
>
> 0040 - c0 05 00 35 00 84 c0 12-c0 08 c0 1c c0 1b 00 16 ...5............
>
> 0050 - 00 13 c0 0d c0 03 00 0a-c0 13 c0 09 c0 1f c0 1e ................
>
> 0060 - 00 33 00 32 00 9a 00 99-00 45 00 44 c0 0e c0 04 .3.2.....E.D....
>
> 0070 - 00 2f 00 96 00 41 00 07-c0 11 c0 07 c0 0c c0 02 ./...A..........
>
> 0080 - 00 05 00 04 00 15 00 12-00 09 00 14 00 11 00 08 ................
>
> 0090 - 00 06 00 03 00 ff 01 .......
>
> 0098 - <SPACES/NULS>
>
> read from 0x1fee3f0 [0x1e6ea5b] (5 bytes => 5 (0x5))
>
> 0000 - 16 03 00 00 51 ....Q
>
> read from 0x1fee3f0 [0x1e6ea60] (81 bytes => 81 (0x51))
>
> 0000 - 02 00 00 4d 03 00 52 cf-5a 99 be 19 c0 7e 80 55 ...M..R.Z....~.U
>
> 0010 - d6 4c 2d af 05 41 f8 19-79 24 a3 66 ba 82 9b 05 .L-..A..y$.f....
>
> 0020 - cd 31 eb 4e 1d 14 20 6b-db c4 0f 3d 82 82 06 52 .1.N.. k...=...R
>
> 0030 - 52 66 54 b9 91 d8 2b ec-c4 27 c4 11 75 bb ec f6 RfT...+..'..u...
>
> 0040 - 64 9f ca bc e3 c9 12 c0-11 00 00 05 ff 01 00 01 d...............
>
> 0051 - <SPACES/NULS>
>
> read from 0x1fee3f0 [0x1e6ea5b] (5 bytes => 5 (0x5))
>
> 0000 - 16 03 00 0c c2 .....
>
> read from 0x1fee3f0 [0x1e6ea60] (3266 bytes => 3266 (0xCC2))
>
>
>
> #########
>
>
>
> It is clear that received data are the same, only the send data are
> different. It is caused my different cipher supported.
>
>
>
> In ssl3_get_server_hello process,
>
>
>
> #ifndef OPENSSL_NO_TLSEXT
>
> /* TLS extensions*/
>
> if (s->version >= SSL3_VERSION)
>
> {
>
> if (!ssl_parse_serverhello_tlsext(s,&p,d,n, &al))
>
> {
>
> /* 'al' set by
> ssl_parse_serverhello_tlsext */
>
>
> SSLerr(SSL_F_SSL3_GET_SERVER_HELLO,SSL_R_PARSE_TLSEXT);
>
> goto f_err;
>
> }
>
> if (ssl_check_serverhello_tlsext(s) <= 0)
>
> {
>
>
> SSLerr(SSL_F_SSL3_GET_SERVER_HELLO,SSL_R_SERVERHELLO_TLSEXT);
>
> goto err;
>
> }
>
> }
>
> #endif
>
>
>
> if (p != (d+n)) <<=== The error is caused here !!!!
>
> {
>
> /* wrong packet length */
>
> al=SSL_AD_DECODE_ERROR;
>
>
> SSLerr(SSL_F_SSL3_GET_SERVER_HELLO,SSL_R_BAD_PACKET_LENGTH);
>
> goto f_err;
>
> }
>
>
>
> ########
>
>
>
> The ssl3_get_server_hello () is working correct in the PC platform, only
> meets problem in the embedded environment. The different is the supported
> ciphers in different environment.
>
>
>
> If i remove the check: if (p != (d+n)), the SSL connection does not have
> any problem. I do not think the problem is in the condition check.
>
>
>
> Any suggestion is welcome.
>
>
>
> Lee
>
>
>
>
>
>
>
