https://issues.apache.org/bugzilla/show_bug.cgi?id=54357

--- Comment #31 from Alex Bligh <[email protected]> ---
(In reply to Kaspar Brand from comment #30)
> Created attachment 32053 [details]
> Patch v7 - store stapling certinfo in a global hash, based on work by Alex
> Bligh
> 
> Hmm, ok, so after another closer look, I think I found another issue with
> the existing code, actually: IINM, with the current certinfo_free code, we
> actually leak the OCSP_CERTID stored cinf->cid (which gets allocated by
> OCSP_cert_to_id).
> 
> I came to this conclusion when I was restructuring ssl_stapling_init_cert()
> a bit more, and am attaching my current version. I did some limited
> "real-world" testing, but more testing and further reviews are welcome and
> appreciated, of course.

Ah yes. Your v7 is still leaking it on server restart.

Whilst we could put in some form of pool handler, that is rather tiresome and I
worry about lifetime issues. I had been trying to keep SSL objects out of the
cert_info structure.

Do we need something like:
   apr_pool_cleanup_register(p, cid, OCSP_CERTID_free, apr_pool_cleanup_null);

in there?

-- 
You are receiving this mail because:
You are the assignee for the bug.

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to