Way back when, (see https://bugs.mageia.org/show_bug.cgi?id=1951 ), we had the issue of the free ATI driver requiring firmware even to operate in text mode. The workaround was to chroot the new partition, activate the nonfree distrib, install kernel-firmware-nonfree, and run mkinitrd -f xxx yyy to rebuild the initrd. Or, you could choose vesa during install, and do this directly. In current cauldron, a fresh install for a system with a Radeon HD47xx or earlier chooses the ati driver and gives a nice prompt about what to do in the new system. But the new system behaves differently now. It boots and shows text console messages, but almost immediately the screen goes black and a mouse cursor appears,as though X is starting. However, that's where it stays. The mouse cursor responds to the mouse in a jerky fashion, but nothing ever appears. This happens even if you set /etc/inittab to boot to level 3. If you go through the workaround and rebuild the initrd, you get an unbootable system. Dracut rebuilds the initrd with no errors, but booting stops almost immediately with dracut messages saying that the "resume is cancelled, device not found". Redoing the install as an update predictably installs no packages and rebuilds the initrd, but the booting situation doesn't change. So what should users with older radeon cards be doing now ?
Am I correct in guessing that this is a new problem with framebuffer on these cards and trying to start plymouth ?
Selecting Xorg:vesa as the driver during install produces a usable system.
However, if you install radeon-firmware, and use XFdrake to reset the X driver to the ati driver, systemd hangs (or rather, loops) at boot. The plymouth splash screen gets to 5 bubbles on the cauldron. If you ESC, you can see a screenful of systemd messages, with a couple of failures, one being atieventsd.service. If you ALT-F2, you get a blank screen (no getty), and if you ALT-F1 to get back, the plymouth splash is up on tty1 again. Hitting ESC gets you a flood of systemd messages, which end at the same point displayed originally, so I guess that systemd is failing, timing out, and retrying. If you use XFdrake to set the driver back to Xorg:vesa and reboot, you get the same hang/loop. The fresh install and the rest of this testing was done on today's cauldron.
lsinitrd shows that the radeon firmware is present, so it looks like the problem is atieventsd.
This is another reincarnation of bug#5262. It's really a shame that installing a third-party rpm can screw up systemd so badly. *** This bug has been marked as a duplicate of bug 5262 ***
Status: NEW => RESOLVEDResolution: (none) => DUPLICATE