I'm not sure the proposed patch actually addresses the original request (is a 4x reduction in the delay, from 2 seconds to .5 seconds, that much of an improvement?), but regardless, this is not just a usability issue, there's also a security issue here. Any changes to the behavior of PAM failure delays needs to be evaluated in the context of all PAM services on the system, not just for the sudo and desktop login cases.
I'm not sure the proposed patch actually addresses the original request (is a 4x reduction in the delay, from 2 seconds to .5 seconds, that much of an improvement?), but regardless, this is not just a usability issue, there's also a security issue here. Any changes to the behavior of PAM failure delays needs to be evaluated in the context of all PAM services on the system, not just for the sudo and desktop login cases.