On microbenchmarks, stackguard does add some noticeable overhead to function calls, but in real life with daemons like ssh or apache, there is no measurable slowdown.
You can find more information by looking at his slides, which are in the middle of the picture library, or you can also download and view the PDF version that was sent to me by Crispin.
Here is also a link to a demo he made of
a program crashing with and without crashguard (1280x960)
![]() |
![]() |
![]() |
99/05/24 (22:13): Version 1.0
99/05/30 (11:12): Version 1.1. Added pdf of slides sent to me by Crispin Cowan