Mailing-List: contact cygwin-developers-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-developers-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin-developers AT sources DOT redhat DOT com Message-ID: <397C2BE5.DA513A25@cygnus.com> Date: Mon, 24 Jul 2000 13:43:33 +0200 From: Corinna Vinschen X-Mailer: Mozilla 4.73 [en] (X11; I; Linux 2.2.14-SMP i686) X-Accept-Language: de, en MIME-Version: 1.0 To: cygdev , cygpatch Subject: [PATCH]: New implementation for /dev/[u]random Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi all, I have checked in a patch which changes the implementation of /dev/random and /dev/urandom as follows: - If initialization of the system crypto provider or retrieving the entropy source fails, reading from /dev/random fails as well. - If that happens when using /dev/urandom, a pseudo random number generator (the same as used in DJGPP) is used as a fallback entropy source. - In either case it's now possible to thrill the entropy source (system _and_ pseudo) by writing to the device. The difference in handling /dev/random in contrast to /dev/urandom is reasoned by it's purpose as a source for "very high quality randomness" as it's described in the Linux man page random(4) while /dev/urandom is explicitly allowed to use a substitute. For clearness: `sshd' uses /dev/urandom so it should work now under any circumstances. Corinna -- Corinna Vinschen Cygwin Developer Cygnus Solutions, a Red Hat company