Re: OpenSSL Security Advisory

2009-01-09 Thread Remo Inverardi
One way to exploit this flaw would be for a remote attacker who is in control of a malicious server or who can use a 'man in the middle' attack to present a malformed SSL/TLS signature from a certificate chain to a vulnerable client, bypassing validation. In my opinion, this statement is not

OpenSSL and LSB Build Tools

2008-04-23 Thread Remo Inverardi
Using the attached patch, OpenSSL can be built using lsbcc3 on Linux: $ ./Configure linux-generic32-lsb $ make CC=lsbcc3 Since this might be of general interest, could you please consider the attached patch for future versions of OpenSSL? Regards, Remo diff -ru openssl-0.9.8g/Configure

Unsigned certificate requests

2006-07-17 Thread Remo Inverardi
We are currently working on a project, where a special kind of RA has submit an unsigned CSR to a CA. An unsigned CSR is a request where the signature field is unassigned (null), because private key operations are inpractical and the RA has already verified that the requester is in possession

[openssl.org #1347] Field 'attributes' of X509_REQ_INFO

2006-06-11 Thread Remo Inverardi via RT
Steve wrote in thread 'Field attributes of X509_REQ_INFO': I see. Well that is probably best fixed within X509_REQ_add_extensions() itself by checking to see if the field is NULL and allocating it if necessary. Here's a patch. While looking at the code, I've also checked the other functions

Re: Field 'attributes' of X509_REQ_INFO

2006-06-08 Thread Remo Inverardi
Steve, I see. Well that is probably best fixed within X509_REQ_add_extensions() itself by checking to see if the field is NULL and allocating it if necessary. Here's a patch. While looking at the code, I've also checked the other functions in x509_req.c. They all seem to either allocate the

Re: Timing Attacks against OpenSSL

2003-03-18 Thread Remo Inverardi
While OpenSSL definitely does provide for blinding, several widely-deployed applications don't seem to enable this option. One reason is it doesn't appear possible to enable blinding from the SSL library itself. After reading the paper, and looking at the OpenSSL RSA blinding code, I feel

Increasing the struct_ref member of engine_st

2002-08-16 Thread Remo Inverardi
While some OpenSSL objects (DH, DSA, RSA) offer an *_up_ref function to increase the reference count, others have their internal structure publicly available (EVP_PKEY, X509, etc), so you can use CRYPTO_add() to increase the reference count. In the process of writing wrapper classes for

Re: COMPILE INSTALL PROBLEMS

2000-02-28 Thread Remo Inverardi
Tom, WE LIKE FREE SOFTWARE. WE LIKE IT WHEN IT WORKS. WE LIKE YOUR SOFTWARE WHEN IT WORKS. Hey, no reason for crying out. I downloaded the 0.9.4 release and everything works great. Just make sure you read the docs carefully and do exactly as you are told. This is, as you mentioned, free

Buffer overflows in OpenSSL 0.9.4 on Windows

2000-01-30 Thread Remo Inverardi
with BoundsChecker or something similar enabled? I've got the buffer overflows and some leeks. I am pretty sure that it's not me, even though it may be possible that I'm using some of the OpenSSL APIs the wrong way. Thanks for sharing any similar experiences you had. Remo Inverardi, ABACUS Research

Emulating standard socket read and write behaviour

2000-01-18 Thread Remo Inverardi
I was wondering if anybody ever managed to write functions which emulate the standard socket read and write behaviour. My web server uses the following logic to read from a socket: a) do a select and wait until data is available b) do a read (recv) c) if read returns 0 goto c1 if read

Re: Checking for memory leaks

2000-01-17 Thread Remo Inverardi
Before the program exits, call EVP_cleanup() and ERR_free_strings() to free the memory allocated in these steps. Mmh, how come I've never read about these functions before? Is there any place where I can get an idea of what every function in the library is good for? I'm sorry, but I didn't find

Checking for memory leaks

2000-01-14 Thread Remo Inverardi
When compiling with Visual C++, I usually put the following lines somewhere in my debug code: #ifdef _DEBUG _CrtSetDbgFlag(_CRTDBG_CHECK_ALWAYS_DF || _CRTDBG_DELAY_FREE_MEM_DF || _CRTDBG_LEAK_CHECK_DF); #endif and #ifdef _DEBUG

Compiler warning in bn_div.c, line 241

2000-01-13 Thread Remo Inverardi
I tried compiling my OpenSSL 0.9.4 distribution on Windows NT using Visual C++ 6.0 and debugging information enabled. My do_nt.bat file looks as follows: perl util\mkfiles.pl MINFO perl util\mk1mf.pl debug no-asm VC-NT ms\nt.mak perl util\mk1mf.pl debug dll no-asm VC-NT ms\ntdll.mak

Windows Sockets

2000-01-13 Thread Remo Inverardi
How do you guys set up the windows sockets when using OpenSSL? Do they have to be blocking or non-blocking? Are there any socket options which do not work with OpenSSL? Thanks for any information. If there is any documentation online which may provide me with the answers, please tell me where to

Re: Compiler warning in bn_div.c, line 241

2000-01-13 Thread Remo Inverardi
Andrew, there is a minor cast error in crypto\bn\bn_div.c. Replace line 241 which reads: Thank you, Remo __ OpenSSL Project http://www.openssl.org Development Mailing List