You are not logged in.
I've recently bought an old Fujitsu Lifebook S7110 featuring an Intel 915GM graphics chip.
The first attempt in trying to boot the archlinux-2010.05-core-CD resulted in a blank screen after the "udev hook". After I've read about the solution to pass the kernel the option "nomodeset", I finally succeeded in booting the CD. The installation was seamless. But my fresh ArchLinux installation inherited the same problem (that I needed the "nodemodeset" option). After installing X and the necessary intel-driver, X argued that it needed the kernel modesetting in order to start.
Following the instructions on http://wiki.archlinux.org/index.php/Intel, I've regenerated the initramfs.
After rebooting, the kernel could load the missing module i915 and could successfully switch to a nicer resolution. But sadly, the hook "udev" took more than 2 minutes to proceed.
In some cases, the system could not detect my root partiton any more and left me back in an emergency shell. In the other cases - after long waiting - the login prompt finally popped up, but the system was - according to "top" - occupied by "i915" and "udev" about 100%. Udev was so occupied that it did not even detect the USB-stick I hotplugged.
I've uploaded several logfiles at http://web403.webbox555.server-home.org/drake/lifebook/
Note the freshly plugged USB-stick detected by "lsusb". The file "udevd" was recorded by "udevd --debug". If you need more information about the problem, please let me know.
--
Unlike Arch, "grml" is very seamless: The live-distribution boots seamlessly, could automatically detect the graphics card and launch X (of course without kms and drm).
EDIT:
This problem is most likely a kernel bug. The most recent Ubuntu live CD wasn't able to boot properly without the 'nomodeset' kernel parameter. I tried was compiling the kernel 2.6.34 myself which was useless (same result). Finally, I fetched the kernel 2.6.35-rc6 from the official git-repositry, which was able to launch the system with kms. Now my dmesg gets flooded with messages in form '[drm:output_poll_execute] *ERROR* delayed enqueue failed -125'.
Last edited by aldekard (2010-07-28 20:07:28)
Offline