Date: Mon, 09 Feb 2009 23:30:33 +0000 From: "Bruce M. Simpson" <bms@FreeBSD.org> To: Robert Noland <rnoland@FreeBSD.org> Cc: "S.N.Grigoriev" <serguey-grigoriev@yandex.ru>, Markus Hitter <mah@jump-ing.de>, freebsd-stable@freebsd.org Subject: Re: Unhappy Xorg upgrade Message-ID: <4990BC99.1070108@FreeBSD.org> In-Reply-To: <1234208586.1524.17.camel@ferret.2hip.net> References: <329181233306971@webmail57.yandex.ru> <985A59F2-20CC-4779-A000-018E52B5BFA9@jump-ing.de> <101781233319948@webmail36.yandex.ru> <4983A3AE.90804@FreeBSD.org> <498F901A.7000900@FreeBSD.org> <1234159237.23838.3.camel@ferret.2hip.net> <4990835A.3020303@FreeBSD.org> <1234208586.1524.17.camel@ferret.2hip.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Robert Noland wrote: > ... > Ok, that is odd... Once drm is loaded and X opens it, the ddx driver > should request that the irq handler be installed. At that point, dmesg > should show something resembling the following. > > vgapci0: child drm0 requested pci_enable_busmaster > info: [drm] AGP at 0xc0000000 256MB > info: [drm] Initialized i915 1.6.0 20080730 > drm0: [ITHREAD] > Yes, I normally see output similar to this when X is started. > ... > > Does the issue still occur if drm is disabled? > Just tried disabling DRM w/ 'Options "DRI" "off"' in Section "Device". DRM is indeed disabled -- no dmesg output and not loaded by X. However the problem is still there. BTW: This Radeon card does have MSI capabilities according to lspci, however they do not appear to be enabled either by FreeBSD or by X. I was about to point the finger at interrupt filters, however that blows that theory out of the water. FWIW the IBM T43 here has an i915GM, and USB is working just fine and dandy. The main data point which sticks out is the fact that the affected machine is amd64. Now that DRM has been disabled on my box, this would point the finger at the X userland. I don't see any obvious nasties in my Section "Device", although I do pass a BusID and BusType to prevent X from trying to use the second head with RandR (lots of pain with fubar DVI cables when I first purchased the monitor). I skimmed pci_user.c, thinking libpciaccess just thunks to it via /dev/pci, it appears there's no instrumentation there I can turn on to see what userland is actually frobbing. thanks, BMS
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4990BC99.1070108>