diff options
Diffstat (limited to 'openssl/bugs')
| -rw-r--r-- | openssl/bugs/MS | 7 | ||||
| -rw-r--r-- | openssl/bugs/SSLv3 | 49 | ||||
| -rw-r--r-- | openssl/bugs/alpha.c | 91 | ||||
| -rw-r--r-- | openssl/bugs/dggccbug.c | 45 | ||||
| -rw-r--r-- | openssl/bugs/sgiccbug.c | 57 | ||||
| -rw-r--r-- | openssl/bugs/sslref.dif | 26 | ||||
| -rw-r--r-- | openssl/bugs/stream.c | 131 | ||||
| -rw-r--r-- | openssl/bugs/ultrixcc.c | 45 |
8 files changed, 451 insertions, 0 deletions
diff --git a/openssl/bugs/MS b/openssl/bugs/MS new file mode 100644 index 000000000..a1dcfb90d --- /dev/null +++ b/openssl/bugs/MS @@ -0,0 +1,7 @@ +If you use the function that does an fopen inside the DLL, it's malloc +will be used and when the function is then written inside, more +hassles +.... + + +think about it. diff --git a/openssl/bugs/SSLv3 b/openssl/bugs/SSLv3 new file mode 100644 index 000000000..a75a1652d --- /dev/null +++ b/openssl/bugs/SSLv3 @@ -0,0 +1,49 @@ +So far... + +ssl3.netscape.com:443 does not support client side dynamic +session-renegotiation. + +ssl3.netscape.com:444 (asks for client cert) sends out all the CA RDN +in an invalid format (the outer sequence is removed). + +Netscape-Commerce/1.12, when talking SSLv2, accepts a 32 byte +challenge but then appears to only use 16 bytes when generating the +encryption keys. Using 16 bytes is ok but it should be ok to use 32. +According to the SSLv3 spec, one should use 32 bytes for the challenge +when opperating in SSLv2/v3 compatablity mode, but as mentioned above, +this breaks this server so 16 bytes is the way to go. + +www.microsoft.com - when talking SSLv2, if session-id reuse is +performed, the session-id passed back in the server-finished message +is different from the one decided upon. + +ssl3.netscape.com:443, first a connection is established with RC4-MD5. +If it is then resumed, we end up using DES-CBC3-SHA. It should be +RC4-MD5 according to 7.6.1.3, 'cipher_suite'. +Netscape-Enterprise/2.01 (https://merchant.netscape.com) has this bug. +It only really shows up when connecting via SSLv2/v3 then reconnecting +via SSLv3. The cipher list changes.... +NEW INFORMATION. Try connecting with a cipher list of just +DES-CBC-SHA:RC4-MD5. For some weird reason, each new connection uses +RC4-MD5, but a re-connect tries to use DES-CBC-SHA. So netscape, when +doing a re-connect, always takes the first cipher in the cipher list. + +If we accept a netscape connection, demand a client cert, have a +non-self-signed CA which does not have it's CA in netscape, and the +browser has a cert, it will crash/hang. Works for 3.x and 4.xbeta + +Netscape browsers do not really notice the server sending a +close notify message. I was sending one, and then some invalid data. +netscape complained of an invalid mac. (a fork()ed child doing a +SSL_shutdown() and still sharing the socket with its parent). + +Netscape, when using export ciphers, will accept a 1024 bit temporary +RSA key. It is supposed to only accept 512. + +If Netscape connects to a server which requests a client certificate +it will frequently hang after the user has selected one and never +complete the connection. Hitting "Stop" and reload fixes this and +all subsequent connections work fine. This appears to be because +Netscape wont read any new records in when it is awaiting a server +done message at this point. The fix is to send the certificate request +and server done messages in one record. diff --git a/openssl/bugs/alpha.c b/openssl/bugs/alpha.c new file mode 100644 index 000000000..701d6a7c7 --- /dev/null +++ b/openssl/bugs/alpha.c @@ -0,0 +1,91 @@ +/* bugs/alpha.c */ +/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com) + * All rights reserved. + * + * This package is an SSL implementation written + * by Eric Young (eay@cryptsoft.com). + * The implementation was written so as to conform with Netscapes SSL. + * + * This library is free for commercial and non-commercial use as long as + * the following conditions are aheared to. The following conditions + * apply to all code found in this distribution, be it the RC4, RSA, + * lhash, DES, etc., code; not just the SSL code. The SSL documentation + * included with this distribution is covered by the same copyright terms + * except that the holder is Tim Hudson (tjh@cryptsoft.com). + * + * Copyright remains Eric Young's, and as such any Copyright notices in + * the code are not to be removed. + * If this package is used in a product, Eric Young should be given attribution + * as the author of the parts of the library used. + * This can be in the form of a textual message at program startup or + * in documentation (online or textual) provided with the package. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * "This product includes cryptographic software written by + * Eric Young (eay@cryptsoft.com)" + * The word 'cryptographic' can be left out if the rouines from the library + * being used are not cryptographic related :-). + * 4. If you include any Windows specific code (or a derivative thereof) from + * the apps directory (application code) you must include an acknowledgement: + * "This product includes software written by Tim Hudson (tjh@cryptsoft.com)" + * + * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + * + * The licence and distribution terms for any publically available version or + * derivative of this code cannot be changed. i.e. this code cannot simply be + * copied and put under another distribution licence + * [including the GNU Public Licence.] + */ + +/* while not exactly a bug (ASN1 C leaves this undefined) it is + * something to watch out for. This was fine on linux/NT/Solaris but not + * Alpha */ + +/* it is basically an example of + * func(*(a++),*(a++)) + * which parameter is evaluated first? It is not defined in ASN1 C. + */ + +#include <stdio.h> + +#define TYPE unsigned int + +void func(a,b) +TYPE *a; +TYPE b; + { + printf("%ld -1 == %ld\n",a[0],b); + } + +main() + { + TYPE data[5]={1L,2L,3L,4L,5L}; + TYPE *p; + int i; + + p=data; + + for (i=0; i<4; i++) + { + func(p,*(p++)); + } + } diff --git a/openssl/bugs/dggccbug.c b/openssl/bugs/dggccbug.c new file mode 100644 index 000000000..30e07a60e --- /dev/null +++ b/openssl/bugs/dggccbug.c @@ -0,0 +1,45 @@ +/* NOCW */ +/* dggccbug.c */ +/* bug found by Eric Young (eay@cryptsoft.com) - May 1995 */ + +#include <stdio.h> + +/* There is a bug in + * gcc version 2.5.8 (88open OCS/BCS, DG-2.5.8.3, Oct 14 1994) + * as shipped with DGUX 5.4R3.10 that can be bypassed by defining + * DG_GCC_BUG in my code. + * The bug manifests itself by the vaule of a pointer that is + * used only by reference, not having it's value change when it is used + * to check for exiting the loop. Probably caused by there being 2 + * copies of the valiable, one in a register and one being an address + * that is passed. */ + +/* compare the out put from + * gcc dggccbug.c; ./a.out + * and + * gcc -O dggccbug.c; ./a.out + * compile with -DFIXBUG to remove the bug when optimising. + */ + +void inc(a) +int *a; + { + (*a)++; + } + +main() + { + int p=0; +#ifdef FIXBUG + int dummy; +#endif + + while (p<3) + { + fprintf(stderr,"%08X\n",p); + inc(&p); +#ifdef FIXBUG + dummy+=p; +#endif + } + } diff --git a/openssl/bugs/sgiccbug.c b/openssl/bugs/sgiccbug.c new file mode 100644 index 000000000..178239d49 --- /dev/null +++ b/openssl/bugs/sgiccbug.c @@ -0,0 +1,57 @@ +/* NOCW */ +/* sgibug.c */ +/* bug found by Eric Young (eay@mincom.oz.au) May 95 */ + +#include <stdio.h> + +/* This compiler bug it present on IRIX 5.3, 5.1 and 4.0.5 (these are + * the only versions of IRIX I have access to. + * defining FIXBUG removes the bug. + * (bug is still present in IRIX 6.3 according to + * Gage <agage@forgetmenot.Mines.EDU> + */ + +/* Compare the output from + * cc sgiccbug.c; ./a.out + * and + * cc -O sgiccbug.c; ./a.out + */ + +static unsigned long a[4]={0x01234567,0x89ABCDEF,0xFEDCBA98,0x76543210}; +static unsigned long b[4]={0x89ABCDEF,0xFEDCBA98,0x76543210,0x01234567}; +static unsigned long c[4]={0x77777778,0x8ACF1357,0x88888888,0x7530ECA9}; + +main() + { + unsigned long r[4]; + sub(r,a,b); + fprintf(stderr,"input a= %08X %08X %08X %08X\n",a[3],a[2],a[1],a[0]); + fprintf(stderr,"input b= %08X %08X %08X %08X\n",b[3],b[2],b[1],b[0]); + fprintf(stderr,"output = %08X %08X %08X %08X\n",r[3],r[2],r[1],r[0]); + fprintf(stderr,"correct= %08X %08X %08X %08X\n",c[3],c[2],c[1],c[0]); + } + +int sub(r,a,b) +unsigned long *r,*a,*b; + { + register unsigned long t1,t2,*ap,*bp,*rp; + int i,carry; +#ifdef FIXBUG + unsigned long dummy; +#endif + + ap=a; + bp=b; + rp=r; + carry=0; + for (i=0; i<4; i++) + { + t1= *(ap++); + t2= *(bp++); + t1=(t1-t2); +#ifdef FIXBUG + dummy=t1; +#endif + *(rp++)=t1&0xffffffff; + } + } diff --git a/openssl/bugs/sslref.dif b/openssl/bugs/sslref.dif new file mode 100644 index 000000000..0aa92bfe6 --- /dev/null +++ b/openssl/bugs/sslref.dif @@ -0,0 +1,26 @@ +The February 9th, 1995 version of the SSL document differs from +https://www.netscape.com in the following ways. +===== +The key material for generating a SSL_CK_DES_64_CBC_WITH_MD5 key is +KEY-MATERIAL-0 = MD5[MASTER-KEY,"0",CHALLENGE,CONNECTION-ID] +not +KEY-MATERIAL-0 = MD5[MASTER-KEY,CHALLENGE,CONNECTION-ID] +as specified in the documentation. +===== +From the section 2.6 Server Only Protocol Messages + +If the SESSION-ID-HIT flag is non-zero then the CERTIFICATE-TYPE, +CERTIFICATE-LENGTH and CIPHER-SPECS-LENGTH fields will be zero. + +This is not true for https://www.netscape.com. The CERTIFICATE-TYPE +is returned as 1. +===== +I have not tested the following but it is reported by holtzman@mit.edu. + +SSLref clients wait to recieve a server-verify before they send a +client-finished. Besides this not being evident from the examples in +2.2.1, it makes more sense to always send all packets you can before +reading. SSLeay was waiting in the server to recieve a client-finish +before sending the server-verify :-). I have changed SSLeay to send a +server-verify before trying to read the client-finished. + diff --git a/openssl/bugs/stream.c b/openssl/bugs/stream.c new file mode 100644 index 000000000..c3b5e867d --- /dev/null +++ b/openssl/bugs/stream.c @@ -0,0 +1,131 @@ +/* bugs/stream.c */ +/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com) + * All rights reserved. + * + * This package is an SSL implementation written + * by Eric Young (eay@cryptsoft.com). + * The implementation was written so as to conform with Netscapes SSL. + * + * This library is free for commercial and non-commercial use as long as + * the following conditions are aheared to. The following conditions + * apply to all code found in this distribution, be it the RC4, RSA, + * lhash, DES, etc., code; not just the SSL code. The SSL documentation + * included with this distribution is covered by the same copyright terms + * except that the holder is Tim Hudson (tjh@cryptsoft.com). + * + * Copyright remains Eric Young's, and as such any Copyright notices in + * the code are not to be removed. + * If this package is used in a product, Eric Young should be given attribution + * as the author of the parts of the library used. + * This can be in the form of a textual message at program startup or + * in documentation (online or textual) provided with the package. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. All advertising materials mentioning features or use of this software + * must display the following acknowledgement: + * "This product includes cryptographic software written by + * Eric Young (eay@cryptsoft.com)" + * The word 'cryptographic' can be left out if the rouines from the library + * being used are not cryptographic related :-). + * 4. If you include any Windows specific code (or a derivative thereof) from + * the apps directory (application code) you must include an acknowledgement: + * "This product includes software written by Tim Hudson (tjh@cryptsoft.com)" + * + * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + * + * The licence and distribution terms for any publically available version or + * derivative of this code cannot be changed. i.e. this code cannot simply be + * copied and put under another distribution licence + * [including the GNU Public Licence.] + */ + +#include <stdio.h> +#include <openssl/rc4.h> +#ifdef OPENSSL_NO_DES +#include <des.h> +#else +#include <openssl/des.h> +#endif + +/* show how stream ciphers are not very good. The mac has no affect + * on RC4 while it does for cfb DES + */ + +main() + { + fprintf(stderr,"rc4\n"); + rc4(); + fprintf(stderr,"cfb des\n"); + des(); + } + +int des() + { + des_key_schedule ks; + des_cblock iv,key; + int num; + static char *keystr="01234567"; + static char *in1="0123456789ABCEDFdata 12345"; + static char *in2="9876543210abcdefdata 12345"; + unsigned char out[100]; + int i; + + des_set_key((des_cblock *)keystr,ks); + + num=0; + memset(iv,0,8); + des_cfb64_encrypt(in1,out,26,ks,(des_cblock *)iv,&num,1); + for (i=0; i<26; i++) + fprintf(stderr,"%02X ",out[i]); + fprintf(stderr,"\n"); + + num=0; + memset(iv,0,8); + des_cfb64_encrypt(in2,out,26,ks,(des_cblock *)iv,&num,1); + for (i=0; i<26; i++) + fprintf(stderr,"%02X ",out[i]); + fprintf(stderr,"\n"); + } + +int rc4() + { + static char *keystr="0123456789abcdef"; + RC4_KEY key; + unsigned char in[100],out[100]; + int i; + + RC4_set_key(&key,16,keystr); + in[0]='\0'; + strcpy(in,"0123456789ABCEDFdata 12345"); + RC4(key,26,in,out); + + for (i=0; i<26; i++) + fprintf(stderr,"%02X ",out[i]); + fprintf(stderr,"\n"); + + RC4_set_key(&key,16,keystr); + in[0]='\0'; + strcpy(in,"9876543210abcdefdata 12345"); + RC4(key,26,in,out); + + for (i=0; i<26; i++) + fprintf(stderr,"%02X ",out[i]); + fprintf(stderr,"\n"); + } diff --git a/openssl/bugs/ultrixcc.c b/openssl/bugs/ultrixcc.c new file mode 100644 index 000000000..7ba75b140 --- /dev/null +++ b/openssl/bugs/ultrixcc.c @@ -0,0 +1,45 @@ +#include <stdio.h> + +/* This is a cc optimiser bug for ultrix 4.3, mips CPU. + * What happens is that the compiler, due to the (a)&7, + * does + * i=a&7; + * i--; + * i*=4; + * Then uses i as the offset into a jump table. + * The problem is that a value of 0 generates an offset of + * 0xfffffffc. + */ + +main() + { + f(5); + f(0); + } + +int f(a) +int a; + { + switch(a&7) + { + case 7: + printf("7\n"); + case 6: + printf("6\n"); + case 5: + printf("5\n"); + case 4: + printf("4\n"); + case 3: + printf("3\n"); + case 2: + printf("2\n"); + case 1: + printf("1\n"); +#ifdef FIX_BUG + case 0: + ; +#endif + } + } + |