Re: Openssl in multithreaded environment segfaults

2004-02-20 Thread Frédéric Giudicelli
Sorry, I meant the SSL_CTX initialization code. -- FrÃdÃric Giudicelli http://www.newpki.org Alberto Alonso wrote: Changing the code to use one SSL_CTX per thread is a large task because of the complexity of the whole system. The initialization code is probably easier to look at: I use: extern

Re: Openssl in multithreaded environment segfaults

2004-02-20 Thread Alberto Alonso
Changing the code to use one SSL_CTX per thread is a large task because of the complexity of the whole system. The initialization code is probably easier to look at: I use: extern pthread_mutex_t *ggsyscom_ssllock_cs; extern long *ggsyscom_ssllock_count; which are defined on the main.c file as

Re: Openssl in multithreaded environment segfaults

2004-02-20 Thread Frédéric Giudicelli
I assume Stephen mentionned EVP_CIPHER_CTX because he wasn't sure which "ctx" you were talking about. I can garantee you that shared SSL_CTX work fine, as long as the mutexes are initialized the proper way. Did you try to use one SSL_CTX per thread to see if it worked ? If you have a problem wit

Re: Openssl in multithreaded environment segfaults

2004-02-20 Thread Alberto Alonso
Actually I already initialize the locking callbacks. What got me confused is the EVP_CIPHER_CTX stuff as I have never seen that. Thanks, Alberto On Fri, 2004-02-20 at 17:33, FrÃdÃric Giudicelli wrote: > If you share the return of the SSL_CTX_new call, then you need to > initialize the mutexes

Re: Openssl in multithreaded environment segfaults

2004-02-20 Thread Frédéric Giudicelli
If you share the return of the SSL_CTX_new call, then you need to initialize the mutexes in libcrypto, see CRYPTO_set_locking_callback for this. The second option is to have one SSL_CTX per thread, as mentionned by Stephen. Regards, -- Frédéric Giudicelli http://www.newpki.org Alberto Alonso wr

Re: Openssl in multithreaded environment segfaults

2004-02-20 Thread Alberto Alonso
On Fri, 2004-02-20 at 16:34, Dr. Stephen Henson wrote: > Looks like a race condition of some sort. > > Well first thing I'd suggest is using the latest 0.9.7 snapshot and seeing if > you still have the problem. I will try that. > > Also compile OpenSSL with debugging symbols so the precise loca

Re: Openssl in multithreaded environment segfaults

2004-02-20 Thread Dr. Stephen Henson
On Fri, Feb 20, 2004, Alberto Alonso wrote: > I'm trying to use openssl-0.9.7a-23 (the RPM from Fedora 1) > in a multithreaded environment. > > Basically I have a server with about 5 worker threads and > a client with a varying number of threads that connect to > the server. > > When using arou

Openssl in multithreaded environment segfaults

2004-02-20 Thread Alberto Alonso
I'm trying to use openssl-0.9.7a-23 (the RPM from Fedora 1) in a multithreaded environment. Basically I have a server with about 5 worker threads and a client with a varying number of threads that connect to the server. When using around 20 threads at the client I can't see any problems. But w