On Sun, 11 Dec 2011 00:13:08 -0600 Bruce Dubbs <[email protected]> wrote:
> Mostly for DJ, > > I finally finished going through Xorg. I always get a sense of > satisfaction when bringing it up for the the first time on a new system. > It's the same feeling when I start LFS for the first time after a major > update. > > DRI and glxgears/glxinfo worked right off the bat. glxgears was only > about 550 FPS or so (my main system is 9900 FPS with a proprietary > nvidia driver), but that's mostly a HW issue. If you use the new glxgears from the mesa-demos it syncs with the refresh rate of the monitor so it always gives me 60 fps > I reviewed your instructions and I think they are extremely well done. > I did make a couple of formatting changes and touchups here and there, > but I liked what you did. > > Some of the dependencies needed to be updated and I did that. I think > we are pretty much up to speed there. > > After the builds, we have 'Xorg-7.6-2 Configuration', but we don't ever > tell the user how to test xorg. I guessed at 'startx' and that brought > up twm. We need to say that somewhere. xinit works too. > > The default startx brings up three xterms and an xclock. We also need > to tell the user that the leftmost xterm is a login xterm and that > exiting from that xterm will terminate the X session. > > BTW, I built in /opt/xorg. I like to do that so I can build another > version of xorg without changing the current programs that would be the > case if installing in /usr. > > On my system, the 3rd xterm (xterm -geometry 80x20+494-0) is completely > covered by the other two. We probably should mention that possibility. > > Looking at the Xorg.0.log I had a couple of entries that I don't > completely understand. > > 1. (WW) Open ACPI failed (/var/run/acpid.socket) > > I've seen that before. I suppose we might want to consider adding apcid > to the book. > > 2. (II) NV: driver for NVIDIA chipsets: many > (WW) Falling back to old probe method for vesa > (WW) Falling back to old probe method for fbdev > > I'm not sure what these warnings mean. I'm sure they're harmless, but I > wonder why they're there. I think it's because you don't have kernel mode setting enabled in you kernel config. > Later it says: > > (II) UnloadModule: "vesa" > (II) Unloading vesa > (II) UnloadModule: "fbdev" > (II) Unloading fbdev > (II) UnloadModule: "fbdevhw" > (II) Unloading fbdevhw > > Perhaps I just need to set /etc/X11/xorg.conf.d/videocard-0.conf > > 3. (EE) open /dev/fb0: No such file or directory > > Is this a kernel configuration issue? I have CONFIG_FB=y, but not > CONFIG_FB_DDC or CONFIG_FB_BOOT_VESA_SUPPORT. I do think it's harmless, > but I'm just trying to run down the details. > > 4. It seems to ID the monitor and video card correctly: > (II) NV: driver for NVIDIA chipsets: ... > (--) NV: Found NVIDIA Quadro NVS 55/280 PCI at 01@00:00:0 > (II) NV(0): Monitor name: DELL IN1910N > > I'm not sure how to get NOUVEAU working. I think libdrm > --enable-nouveau-experimental-api needs to be set, but I'm not sure what > else. I think Mesa needs --with-gallium-drivers=nouveau, you need to install the driver xf86-video-nouveau which is only available through git, you need to recompile your kernel to enable the kernel part of the nouveau stack and you need to put driver=nouveau in your xorg.conf. It's a bit of hastle to get it all set up but once you do it makes recompiling the kernel a lot easier because the code is in the kernel. The Nvidia closed source blob makes recompiling the kernel such a pain! I think the nouveau people have done an excellent job. It's all open source. I stopped buying Nvidia hardware because they won't pay for a developer to work on nouveau. Andy -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
