Created attachment 27591 [details] lspci This is on eMachines E640G laptop. I compiled kernel from Git (at commit 36423a5ed5e4ea95ceedb68fad52965033e11639) and when I try to run with Radeon KMS enabled first kernel just stops booting, then after about a minute ro so I get some more output which looks like an oops. I'll attach the screenshots later. if I disable KMS the kernel boots fine.
Created attachment 27611 [details] config
Created attachment 27621 [details] first-boot-screen This is where booting stops for the first time
Created attachment 27631 [details] second-boot-screen This is where the boot stops for the second time after abou a minute.
OK, I've updated the sources to commit d4348c678977c7093438bbbf2067c49396ae941b and this time the kernel does boot (although the one minute long pause is still there). I've checked dmesg and it looks like it can't find the firmware: [ 1.328311] [drm] Loading CEDAR Microcode [ 1.927991] ACPI: Battery Slot [BAT1] (battery present) [ 61.920928] r600_cp: Failed to load firmware "radeon/CEDAR_pfp.bin" [ 61.921022] [drm:evergreen_startup] *ERROR* Failed to load firmware! But I checked and the fitmware is there: # ls -l /lib/firmware/radeon/ | grep CEDAR -rw-r--r-- 1 root root 5504 avg 26 19:30 CEDAR_me.bin -rw-r--r-- 2 root root 4480 avg 26 19:30 CEDAR_pfp.bin -rw-r--r-- 4 root root 3072 avg 26 19:30 CEDAR_rlc.bin
Created attachment 28051 [details] dmesg This is the complete dmesg after booting to the end
Make sure the firmware is included in your initrd if you are using one.
I'm not using initrd. Should I be using it? On another laptop I also don't use initrd and it does load R300 Microcode.
Is the radeon drm built in or built as a module? If it's built in, you'll have to add the firmware to your kernel image IIRC. If it's a module, make sure it's looking in the right place.
Aaah now I know what was going on. Just found out this now http://www.x.org/wiki/radeonBuildHowTo#TroubleshootingExtraFirmwareforR600.2BAC8-R700.2BAC8-Evergreen So I added these configs and now it loads firmware/microcode just fine. So I guess this bug should be closed now? Unless handling the case of not existing/found firmware should still be improved?
I closed as invalid since the problem was on my side and with firmware all works finr for quite some time now
Actually close