Date: Fri, 21 Sep 2007 16:55:24 -0500 From: Craig Boston <cb@severious.net> To: current@freebsd.org Subject: Re: nvidia driver on recent current? Message-ID: <20070921215514.GA1114@nowhere> In-Reply-To: <20070921202523.GB4044@nowhere> References: <20070921202523.GB4044@nowhere>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Sep 21, 2007 at 03:25:24PM -0500, Craig Boston wrote: > Some of the addresses look very suspicious, so it seems the stack is > likely corrupted. Now that I think about it, the stack isn't corrupt. Those are just the arguments for the ?? functions that kgdb can't work out the prototype for. A little extra info, this is the newest nvidia driver -- 100.14.11. It happens on a kernel from last week, as well as an 8/31 build that I tried rolling back to. Before that it was stable for months, but as it's somewhat intermittent it may take a while to locate the exact change. Running multiple GL clients at once seems to be a good way, and it almost always happens just as one terminates. This is an SMP system. I habitually add -DKSE -DSMP to the nvidia driver CFLAGS between the 'make patch' and 'make build' steps in the port install, but taking those out doesn't seem to have any effect. Craig
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070921215514.GA1114>