[Dario Lombardo <dario.lombardo.ml@xxxxxxxxx>]

I can't answer for all the "tainted" checks on coverity, but the ones I'm patching seem reasonable.

Quick example

num = tvb_get_ntohl(...)

for (i = 0; i < num; i++) { }

Nothing is going to crash in wireshark, but if a tainted packet contains 0xFFFFFFFF, why should we loop with this value, until we overcome the boudaries? The value is clearly wrong. My idea of patch is to check this value against a max, and stop the current call if the check fails. 

Currently there are 160 tainted checks open in coverity. How can we be sure that all of them are unuseful, hence we disable the check? What about look at them, and maybe mark some them as false positive? 

On Mon, Jul 11, 2016 at 12:46 PM, Jaap Keuter <jaap.keuter@xxxxxxxxx> wrote:

Hi List,

Since (not so) recently the Coverity code analysis has added a checker for so called tainted data. This data is considered coming from an external source (eg. the network) hence suspicious until validated. Using these tainted values is considered a risk. In general this is true, Wireshark on the other hand is intended and designed to handle suspicious / (very) possibly wrong network data (that’s what we’re using it for, amongst other things). So even though data is tainted, many cases the use of the TVB, etc. protects us from the problems envisioned by the checker writers.

So what to so with these Coverity issues. Before we start to implement all kinds of arbitrary checks (duplicating effort already handled by the tvb code) and limits (mostly arbitrary) we should consider is this checker is really valuable in this context.

Kind regards,
Jaap

___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
             mailto:wireshark-dev-request@xxxxxxxxxxxxx?subject=unsubscribe