FedoraThis forum is for the discussion of the Fedora Project.
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
Ok, so I have a problem. I've gone to the Fedoraforum.org guide, and I also followed the RPMFusion tips section. Neither of them worked well for me.
When I added the rdblacklist=nouveau to grub.conf it still loaded the nouveau kernel (gave me the graphical bootup for Fedora 12. If I hit the up arrow during this load (and before I put in my LUKs passphrase for my encrypted /home) I could no longer enter the passphrase. It wouldn't accept it and I couldn't mount /home.
I also did the dracut thing with the boot image, which removed the nouveau driver from the init image. But, I was back to not being able to put in my LUKS passphrase and couldn't mount /home.
I haven't tried using the Nvidia drivers from the Nvidia website, but after spending 2 hours trying all this crap and then trying to restore my system (which I luckily was able to do) I'm not too inclined to keep messing with it.
Anyone else had the same problem with an encrypted partition trying to mount it at boot time? I'm guessing that when I disabled nouveau it reverted to the vesa driver. Seems like something is broken there with LUKS.
My solution was just to run mkinitfs without the option to load any video driver with the initial RAM file system (which disables the graphical boot screen, but I like to watch the boot messages.) Then there's no driver loaded when the nVidia driver gets loaded by the akmod process.
And, FYI, I have to use the nVidia driver since nouveau has a "bug" that crashes my on-board nVidia driver (a MPC67) after five minutes to five hours of use. (When the driver crashes, so does the X-server, and all unsaved work. At least the screen doesn't turn blue, eh?) I reported the bug during Rawhide testing, but, as far as I know, it's not yet fixed.
I've tried every FC12 recipe including this one and none work. I had a few problems with FC11 where I used kmod-nvidia but had to add "vmalloc=256M" to the kernel line in grub.conf. lspci reports
02:00.0 VGA compatible controller: nVidia Corporation GeForce 8200 (rev a2)
03:00.0 VGA compatible controller: nVidia Corporation GeForce 8400 GS (rev a1)
I've tried kmod-nvidia and the nVidia download (the nvidia.ko build goes without a hitch. At startx, the machine hangs and needs a power down reboot. The only way I can do anything is from run-level 3. I am running the latest 32 bit PAE kernel. Having troubles with 64 bit Python. If I use rdblacklist=nouveau the boot hangs after "Starting udev:" and then requires a power down, wait, power up reboot. Same thing happens if I build a nouveau free initrd, but then I have spend 3 more hours rebuilding my FC12 evaluation system. Am using an MCP78 south bridge with a fan hot melt glued to it (121 F).
This gives me 2 DVI ports for my pair of 24" monitors (1920 x 1200). Thank the Gnome for workspaces or I would have had to buy 6 more monitors :-). Note: VGA is too fuzzy. I do software development. I need 4 or 5 reference manuals open, a debugger, a profiler, a test bed, and all of my other development tools open. It would be nice to have everything visible at the same time, but I make do. Nouveau is not ready for prime time.
Last edited by bvn; 05-21-2010 at 02:07 AM.