Re: [PATCH] USI->now, gettimeofday() and debugging

[David Fifield <david () bamsoftware com>]

On Wed, May 05, 2010 at 12:46:18PM -0500, Kris Katterjohn wrote:
> Hey everyone,
>
> The UltraScanInfo{} "now" member is used to keep the near-current time
> so gettimeofday() doesn't have to be called over and over while port
> scanning.  However I found a couple of places where "now" can vary
> depending on if debugging is enabled.  I've attached a patch and
> describe the changes below.
>
> In processData(), "now" can be updated with gettimeofday if debugging
> is enabled (but not otherwise) even just before using "now" to compare
> times for probe timeout, moving a probe to the bench, etc.  So it seems
> Nmap will behave differently here just based on debugging, which surely
> doesn't seem correct.  Here I update "now" at the beginning regardless
> of debugging.
>
> In doAnyOutstandingRetransmits(), gettimeofday() is called for "now"
> back to back if debugging is enabled, which also isn't good.  Here I
> just reorder a few lines and have it called once.
>
> The ends of both of these functions update "now" again only if
> debugging is enabled, so I decided to do it update it either way so
> following functions get the same time independent of debugging.

I don't think this causes any problems. It looks good.

David Fifield
_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/