Hi Again,

I've been looking into the below further and if it is
a leak then I think the neatest solution isn't to
create a separate free function for the compression
methods stack, but to free it in EVP_cleanup() in the
exact same manner that the ciphers and digests that
are dynamically created in SSL_library_init() are
freed there also.

> Hi List,
> 
> I think that the dynamic allocation of the global
> stack of compression methods (ssl_comp_methods)
> during
> SSL_library_init() which doesn't have a
> corresponding
> free causes a memory leak.
> 
> I can see that around where this is allocated in
> load_builtin_compressions(), memory checking is
> turned
> off and on, which leads me to believe that the
> creator
> of this code considered this to be a once of
> allocation and not a leak.
> 
> This may be true for the scenario where the ssl
> library is loaded and initialised once for the life
> of
> a process. However, in a scenario where the ssl
> library is repeatedly loaded, initialised, and
> unloaded within a single process, wouldn't this
> allocation would occur every time or am I missing
> something?
> 
> If this is a memory leak, then I'd be willing to
> submit a patch with some kind of
> SSL_free_comp_methods() function.
> 
> Thanks,
> 

Send instant messages to your online friends http://au.messenger.yahoo.com 
______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       [email protected]
Automated List Manager                           [EMAIL PROTECTED]

Reply via email to