Steam Winsock regression

[Scott Ritchie]

On Mon, 2005-12-19 at 14:59 -0800, James Liggett wrote:
> On Mon, 2005-12-19 at 10:00 +0100, Andreas Mohr wrote:
> > I don't think Scott meant it that way ;-)
> > Who cares about Wine, here's a change that makes half the machines in the
> > network lose connection randomly, that's much bigger things to worry about
> > potentially than simply non-working Wine socket functionality...
> Of course that's the case here, maybe I didn't say it like I should have
> and it came off not quite right. Sorry, it was late at night and I was
> tired ;-) In fact, I was thinking the same thing. Hell, I almost
> uninstalled wine because of it thinking I should stay away from
> something that unstable and potentially dangerous. But maybe it's not
> all Wine's fault. Last week my brother and I got together with a friend
> of his to play a couple late-night matches of CS Source. I was playing
> on Windows XP since Source games don't work so well on Wine yet. For
> some reason I can't explain, my network's router kept failing
> spontaneously while we were playing, in a similar manner in wine with
> the patch. It's possible that Marcus's winsock patch might have exposed
> a bug in Source/Steam network code, but I can't prove that yet. 
> > 
> > Thus one should probably attempt to investigate a bit more.
> Absolutely. What can I do to help out here?
> 
> James

Actually, I meant it might be a security issue for 2 reasons.  The first
is that taking out other machines is obviously a problem.  The second,
however, was that you reported it took out the internet of the machine
Wine was running on until a restart.  A userlevel application shouldn't
be able to do that, which would mean there might be an internet-killing
exploit in the form of a winelib app.

Thanks,
Scott Ritchie