Everhart, Glenn From: JONESD@er6.eng.ohio-state.edu Sent: Wednesday, June 10, 1998 12:12 AM To: Info-VAX@Mvb.Saic.Com Subject: Re: SSH server for VMS ? In message <6lkelc$ekn$1@riksun.riken.go.jp>, ichihara@rikaxp.riken.go.jp (Takashi Ichihara) writes: > Thank you very much for the quick response and useful information. >I got it and now seems to be working in our AXP environment >(OpenVMS Alpha V6.2-1H3, UCX 4.0 (EC0-3), DEC-C V5.6-003). And your SSLeay version? > Following is just for your information. In my environment above, I >encountered an warning in linking ( undefined symbol: PT_INITIALIZE ) >as follows > The SSH server is a multi-threaded application, by default SSLeay is not thread-safe. Rather than gear SSLeay towards any particular threading package, Eric Young provides a means for multi-threaded applications to supply their own locking callouts when they need to make SSLeay threads aware. The pt_initialize() function loads DECthreads-based functions into these locking callouts. The pt_initialize() function is part bss_mst.c, which is a module I wrote as part of SSL support for the OSU HTTP server. I thought bss_mst.c had been rolled into the standard SSLeay distribution, but I may be wrong. (pt_initialize should logically be in a different file that the rest of bss_mst.c, but I wanted to minimize the number of new files to present Mr. Young with). Bss_mst.c, along with the rest of the SSL shim for the OSU server is available at: http://kcgl1.eng.ohio-state.edu/www/preview/tserver_ssl_v3a.zip Having said all that, I built a version of the SSH server with the lock debug trace enabled and it seems the only locking is in SSLeay's random number generator. While it is only using the locks in 1 module, this is enough to warrant including pt_initialize() call in the main program. David L. Jones | Phone: (614) 292-6929 Ohio State University | Internet: 140 W. 19th St. Rm. 231a | jonesd@kcgl1.eng.ohio-state.edu Columbus, OH 43210 | vman+@osu.edu Disclaimer: Dogs can't tell it's not bacon.