Comment 4 for bug 524064

I see the exact same problem with ATI hardware. So, I think this is reported against the wrong package, and I think the duplicate that is marked is incorrect. I am not doing anything more advanced than I have done for years. Bumping the client limit will just mean we hit a higher limit after a little longer. I'm pretty convinced it's a client issue, and there's a leaky client, but I'm having a lot of trouble pinning the client responsible down.