Created attachment 5102 [details] gfx hwinfo/lsmod/Xorg.0.log without modprobe radeon Initial summary: X uses fbdev instead of radeon for rv200 unless modprobe radeon before starting X Also happens with kernel 3.12.8. openSUSE 13.1's 3.11.10 and Rawhide's 3.15.rc0 do not require manually modprobing the radeon kernel driver for X to use the radeon X driver for rv200. This manual intervention for Cauldron's X to use the optimal driver should not be required.
Created attachment 5103 [details] Xorg.0.log after modprobe radeon Note lspci -v and lsmod output have been prepended.
CC: (none) => anssi.hannula, thierry.vignaud, tmb
Can you attach output of lspicdrake -v
Created attachment 5115 [details] lspicdrake -v
Enable debugging of the display driver helper: # sed -i "s,^# DEBUG,DEBUG," /sbin/display_driver_helper then reboot, and when X loads up with the wrong driver, attach /dev/ddh_debug.
display_driver_helper was nowhere to be found on host m7ncd's Cauldron / filesystem. drakx-kbd-mouse-x11 was not installed. X uses radeon fully automatically since installing it. Rawhide with 3.15.rc0 and openSUSE with 3.11.10 load radeon long before graphical.target is reached, and neither have display_driver_helper. Why does Cauldron need this? If it's truly needed, why didn't whatever depends on it have it pulled in automatically when it was installed? cf. https://lists.fedoraproject.org/pipermail/devel/2014-April/197399.html
display_driver_helper is used to avoid loading radeon driver when a proprietary driver or a generic vesa driver is configured. I guess we should a) modify the udev rule to always load the driver so that if drakx stuff is not installed (and make sure the proprietary drivers depend on the drakx stuff if they don't already), or b) move display_driver_helper to some package in basesystem. To test (a), edit /lib/udev/rules.d/80-drivers.rules from: SUBSYSTEM=="pci", ATTR{class}=="0x03*", DRIVER!="?*", TEST=="/initrd", RUN+="/sbin/display_driver_helper --load $env{MODALIAS}", GOTO="drivers_end" to: SUBSYSTEM=="pci", ATTR{class}=="0x03*", DRIVER!="?*", TEST=="/sbin/display_driver_helper", TEST=="/initrd", RUN+="/sbin/display_driver_helper --load $env{MODALIAS}", GOTO="drivers_end" I'm not immediately sure whether (a) or (b) is the right way to go, though... Thomas, any opinion?
My 80-drivers.rules contains nothing I can spot that is like what you want to change from: # do not edit this file, it will be overwritten on update ACTION=="remove", GOTO="drivers_end" ENV{MODALIAS}=="?*", RUN{builtin}="kmod load $env{MODALIAS}" SUBSYSTEM=="tifm", ENV{TIFM_CARD_TYPE}=="SD", RUN{builtin}="kmod load tifm_sd" SUBSYSTEM=="tifm", ENV{TIFM_CARD_TYPE}=="MS", RUN{builtin}="kmod load tifm_ms" SUBSYSTEM=="memstick", RUN{builtin}="kmod load ms_block mspro_block" SUBSYSTEM=="i2o", RUN{builtin}="kmod load i2o_block" SUBSYSTEM=="module", KERNEL=="parport_pc", RUN{builtin}="kmod load ppdev" KERNEL=="mtd*ro", ENV{MTD_FTL}=="smartmedia", RUN{builtin}="kmod load sm_ftl" LABEL="drivers_end"
Are you sure you didn't accidentally check on a non-Mageia system? There should be Mageia-specific stuff below the ACTION=="remove" line. I've verified that systemd-208-14.mga5.x86_64 does contain the lines.
(In reply to Anssi Hannula from comment #8) > Are you sure you didn't accidentally check on a non-Mageia system? That is what I had done. :-p I made the comment 6 edit, renamed display_driver_helper display_driver_helpeR, rebooted, and the radeon kernel driver automatically loaded @[25.480256] (on the 3.13.9-desktop-1.mga5 i686 system).
still a problem with 3.14.3 kernel and x11-driver-video-ati-7.3.0-1
Looks to be fixed on host m7ncd as of 3.15.6 kernel and 7.4.0 driver.
Status: NEW => RESOLVEDResolution: (none) => FIXED