Date: Fri, 6 May 2011 18:24:46 +0200 From: Guido Falsi <mad@madpilot.net> To: Kostik Belousov <kostikbel@gmail.com> Cc: x11@freebsd.org Subject: Re: Xorg lockup in drmwtq state Message-ID: <20110506162446.GA84690@megatron.madpilot.net> In-Reply-To: <20110505163602.GY48734@deviant.kiev.zoral.com.ua> References: <20110505131736.GE29667@megatron.madpilot.net> <20110505163602.GY48734@deviant.kiev.zoral.com.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, May 05, 2011 at 07:36:02PM +0300, Kostik Belousov wrote: > 3. [Only if you have much time] Set hw.dri.0.debug to 1 and post > the kernel log somewhere. Hi again. I've tried for one hour to make it lockup while loggin dri debug info, but with no success. Maybe it's some timing race condition and debug output creates delays just enough to stop it from happening or make it much less frequent? Anyway, just when giving up, after I stopped the debug(by resetting the sysctl to zero) I switched to ttyv0, and when switching back to Xorg it crashed and dumped core. If it can serve anything I have the core dump. Unluckily I already deleted the debugging output (my fault, I'm sorry). Anyway I can try again this stress test next week. > > As a workaround, you might try to set LIBGL_ALWAYS_INDIRECT to 1 in > environment for Firefox, or to recompile graphics/cairo without > OpenGL support. The LIBGL_ALWAYS_INDIRECT workaround seems to work for now. But I have not tested this as thoroughly as I'd like to. I'll perform some more testing from monday(the PC showing the problem is my work one) -- Guido Falsi <mad@madpilot.net>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110506162446.GA84690>