I just had a test run with the new 2004.2 universal live-CD. boot: gentoo acpi=on (had some more options, but they should not matter) I used verbose boot messages. There was an unresolved symbol error while probing the acpi-module thermal.o. Screen scrolled too fast to read the details (symbol name). I think, this module is not impportant for normal tasks with the live-CD, but someone might want to look into this. Reproducible: Always Steps to Reproduce: 1. Boot computer with live-CD 2004.2 in CD-ROM drive. 2. >>> boot: gentoo acpi=on 3. watch verbose boot messages to spot missing symbol message for thermal.o Actual Results: thermal.o did not load Expected Results: thermal.o should have been loaded. Test hardware: pentium4 (2GHz) / 1 GB RAM ASUS P4T-E (i850) with latest acpi-bios boot-CDROM is connected to the onboard ide controller.
You do realize that a desktop board won't have the proper hardware for the thermal module, right? More than likely, the failure is because it was unable to find the specified hardware. Without more information, there isn't anything we can do about it. Does it show up in a dmesg? If you can get more information on it, REOPEN this bug.
You are right. My "expected results" were not very precise. I do realize that the board might not have an acpi-button/ac/fan/thermal and the modules will not be loaded in this case, but the resulting message should be "... thermal.o: No such device" instead of "unresolved symbol acpi_processor_set_thermal_limit_Re8a3605f". This message can not be found in dmesg. fan.o was probed prior to thermal.o resulting in "no such device" (which is correct). processor.o was probed after thermal.o resulting a message about the failiure of thermal.o When trying to modprobe thermal.o later everything works as expected ("no such device"). The error seems to be limited to the boot process. Because of the missing hardware I cannot test, if the missing symbol error occurs when the acpi-thermal/processor device is present or not. If the error is limited to absence of the acpi devices, it is cosmetic and this bug can be closed again. Otherwise it should be fixed (maybe by checking module deps) in order to enable acpi modules for notebooks. Someone with acpi-hardware (e.g. a notebook) should verify this and update this bug. To reproduce error try boot: gentoo-nofb acpi=on (I also used dokeymap). You can find the error message at some point after "hardware initialization" (textmod progress-bar).
Hrrrmn... what really is bothering me is why thermal is even trying to load before processor. The thermal module requires processor. I'll look and see what I can do, it's probably an issue with autoconfig. I'll look into it. Also, thank you for getting me more information very quickly. If only more bug posters were so quick to help out... =]
Did this also happen on the smp kernel?
REOPEN if you have more information on this