Date: Sat, 30 Aug 2008 12:57:49 +0200 From: "Paul B. Mahol" <onemda@gmail.com> To: "Robert Noland" <rnoland@freebsd.org> Cc: freebsd-x11 <freebsd-x11@freebsd.org>, jimmiejaz@gmail.com Subject: Re: [CFT] drm updates Message-ID: <3a142e750808300357s49ebb105k589c749526fbd79c@mail.gmail.com> In-Reply-To: <3a142e750808291329l1836bc9bv8e94bdd96bd343aa@mail.gmail.com> References: <48B1A590.2040701@gmail.com> <3a142e750808270307k237c5660ha98e8bcdfbddc4dc@mail.gmail.com> <1219850374.61484.7.camel@squirrel.corp.cox.com> <3a142e750808281007r42bbe4d0u53be9c0074386c10@mail.gmail.com> <1219944411.61066.5.camel@squirrel.corp.cox.com> <3a142e750808281145m3d49839fv4665066d1791e4e6@mail.gmail.com> <1219953332.2212.4.camel@squirrel.corp.cox.com> <3a142e750808290354x2578a075r50ab0cd6edc21ed5@mail.gmail.com> <1220025309.37942.35.camel@squirrel.corp.cox.com> <3a142e750808291329l1836bc9bv8e94bdd96bd343aa@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
------=_Part_25509_22632719.1220093869564 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Content-Disposition: inline On 8/29/08, Paul B. Mahol <onemda@gmail.com> wrote: > On 8/29/08, Robert Noland <rnoland@freebsd.org> wrote: >> On Fri, 2008-08-29 at 12:54 +0200, Paul B. Mahol wrote: >>> On 8/28/08, Robert Noland <rnoland@freebsd.org> wrote: >>> > On Thu, 2008-08-28 at 20:45 +0200, Paul B. Mahol wrote: >>> >> On 8/28/08, Robert Noland <rnoland@freebsd.org> wrote: >>> >> > On Thu, 2008-08-28 at 19:07 +0200, Paul B. Mahol wrote: >>> >> >> On 8/27/08, Robert Noland <rnoland@freebsd.org> wrote: >>> >> >> > On Wed, 2008-08-27 at 12:07 +0200, Paul B. Mahol wrote: >>> >> >> >> On 8/26/08, Robert Noland <rnoland@freebsd.org> wrote: >>> >> >> >> > On Tue, 2008-08-26 at 17:51 +0200, Paul B. Mahol wrote: >>> >> >> >> >> Here is more kernel debug info (after updating libdrm and >>> >> >> >> >> testing >>> >> >> >> >> with >>> >> >> >> >> glxgears) >>> >> >> >> > >>> >> >> >> > I don't think that the two issues are related, but try this >>> >> >> >> > patch... >>> >> >> >> > It >>> >> >> >> > should at least, resolve the issue below. >>> >> >> >> > >>> >> >> >> > robert. >>> >> >> >> >>> >> >> >> Issue bellow is fixed, previous problem still exist. >>> >> >> >> >>> >> >> >> textdump of panic when unloading drm attached. >>> >> >> > >>> >> >> > Ok, The locking semantics are a nightmare... Give this a try, it >>> >> >> > should >>> >> >> > correct both the previous issue as well as this one. >>> >> >> > >>> >> >> > robert. >>> >> >> > >>> >> >> >>> >> >> Panic fixed, now only Xorg crashing when using glxgears (,...) >>> >> >> remains >>> >> >> ;) >>> >> > >>> >> > Ok, what hardware are you using again? Can you manually load drm >>> >> > modules and set hw.dri.0.debug=1 and send me the debug output? >>> >> > >>> >> > robert. >>> >> >>> >> Hardware: "Mobile 945GM/GU Express Integrated Graphics Controller" >>> >> Debug ouptut attached. >>> > >>> > Alright, try this... I was pointed at this from another thread, but >>> > your trace confirms why it doesn't work, at least with your driver. >>> > >>> > robert. >>> > >>> >>> Nothing improved with that patch, still crash with glxgears, and lock >>> with stellarium. >> >> ok, I don't understand why I'm not seeing any of this on my 965gm... If >> you want to send me another debug trace, I'll see if I can tell what is >> going on now... Also, please send me a copy of your xorg.conf. > > Well, actually it is: > > agp0: <Intel 82945GM (945GM GMCH) SVGA controller> on vgapci0 > agp0: detected 7932k stolen memory > agp0: aperture size is 256M > drm0: <Intel i945GM> on vgapci0 > >> >>> How to make sure that this is really drm problem and not i915, Xorg, >>> mesa-ports that need recompiling? >> >> If you are using current ports tree, then nothing should need rebuilding >> that I can think of. I have been using git libdrm and xf86-video-intel >> driver, but what is in ports now should be fine I think. The updated >> libdrm should only be required to get the new vblank code to allow it to >> disable vblank interrupts. >> >> There is a bug which prevents vblanks irqs from being disabled after a >> vt switch, but it shouldn't have a real impact on functionality. >> >> robert. >> >> > Ooops, there is still (with newest HEAD) panic when unloading agp module, textdump with backtrack and alltrace attached. ------=_Part_25509_22632719.1220093869564--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3a142e750808300357s49ebb105k589c749526fbd79c>