Re:OpenBSD errata
by trengrj on Friday December 04, 2015 @02:49PM
(#10678633)
LibreSSL seemed to be pulling ahead with two of the OpenSSL vulnerabilities not applicable.
It irks me that OpenSSL would keep a vulnerability from the LibreSSL team since August. I hope people using OpenSSL will begin to switch over now. LibreSSL has shown that it is no going away, and that it approaches security in a far better way.
I just updated and recompiled everything a few days ago. Bad timing. Kinda sucks that manual patching is a pain a lot of the time, but oh well, dependencies will do that I guess.
Re:OpenBSD errata
by jjuhl on Friday December 04, 2015 @03:34PM
(#10678956)
Seems like the OpenBSD guys are a little bitter. And rightfully so. OpenSSL has a lot of improvement work to do - both with their code base and their collaboration efforts.
Re:OpenBSD errata
by chx on Friday December 04, 2015 @03:32PM
(#10678938)
> We did not merge this because it gave miod@ a bad feeling.
Not even sure what the right reaction to that is. On one hand, good for you for skipping an insecure patch but on the other hand, only a bad feeling? You didn't realize it's a security hole?
Re:OpenBSD errata
by zeeboo on Friday December 04, 2015 @03:40PM
(#10678980)
Neither did the people writing it, or merging it. Security isn't an obvious property of code or there would be no security bugs. You have to rely on heuristics, and in this case the heuristics were right.
Re:OpenBSD errata
by jjuhl on Friday December 04, 2015 @03:51PM
(#10679051)
When reviewing code you sometimes miss things. It's not an exact science.
You also sometimes get a bad feeling without being able to say exactly what's wrong - it just feels/looks bad. That's when you should seek additional input from your peers.
Rejecting code based on 'smells wrong' can be sane. The submitter should then explain better why the code is sane.
Re:OpenBSD errata
by chops on Friday December 04, 2015 @03:42PM
(#10678996)
Surely sometimes when you glance over a patch that even without actually working through the code, you get a vibe that something doesn't seem quite right. I know I've gotten patches like this before, and my code isn't nearly as mission-critical and complicated as the SSL libraries used by basically everyone on the planet.
It irks me that OpenSSL would keep a vulnerability from the LibreSSL team since August. I hope people using OpenSSL will begin to switch over now. LibreSSL has shown that it is no going away, and that it approaches security in a far better way.
Not even sure what the right reaction to that is. On one hand, good for you for skipping an insecure patch but on the other hand, only a bad feeling? You didn't realize it's a security hole?
You also sometimes get a bad feeling without being able to say exactly what's wrong - it just feels/looks bad. That's when you should seek additional input from your peers.
Rejecting code based on 'smells wrong' can be sane. The submitter should then explain better why the code is sane.
CVE-2015-13XX should be CVE-2015-31XX.