https://bugzilla.novell.com/show_bug.cgi?id=352648 User jeffm@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=352648#c1 Jeff Mahoney changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jeffm@novell.com AssignedTo|kernel-maintainers@forge.provo.novell.com |astarikovskiy@novell.com --- Comment #1 from Jeff Mahoney 2008-01-09 10:57:38 MST --- There doesn't necessarily have to be any ACPI-related modules. ACPI is built into the kernel. Only things like acer_acpi and cpufreq end up being modules. Technically we don't officially support external modules, but if it's working elsewhere we might want to take a look anyway. Please attach the output of "hwinfo" -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=352648 User cathectic@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=352648#c3 Carlos Corbacho changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |cathectic@gmail.com --- Comment #3 from Carlos Corbacho 2008-01-26 18:00:04 MST --- Jeff, I'm the maintainer of acer_acpi. Basically, acer_acpi is nowadays a two part driver - a generic ACPI-WMI driver (wmi-acer), and an Acer specific driver (acer_acpi) on top. The problems with SuSE have only started happening after I made this split (rather than trying to hardcode in paths to ACPI methods, which is known to break on newer laptops). Inexplicably, this has broken only on OpenSuSE 10.3 (I've only received three reports to date though, so perhaps it's not terribly popular with your users, or it doesn't break for everyone - I have no idea). The WMI driver is where the failure is currently occurring, and the code in there is a backport of the stuff I'm currently submitting upstream. So whilst I appreciate you don't "officially" support out-of-tree modules, this would be a worthwhile exception to make, since it will be upstream in the near future. I've also had two other reports where acer_acpi inexplicably fails on known supported laptops (Aspire 5920 and Extensa 5220), and the only common culprit is OpenSuSE 10.3 - again, in all cases, it's been reported that on other distributions, acer_acpi works on the same hardware just fine. So, as for culprits, I'm open to suggestions - does SuSE ship any other drivers that try to claim PNP0C14? Or are there any compiler bugs, etc that could cause this issue that are currently present in OpenSuSE? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=352648 User lolkaantimat@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=352648#c5 --- Comment #5 from Evgeniy Ivanov 2008-01-29 03:59:26 MST --- (In reply to comment #4 from Jeff Mahoney)

Can you try using an openSUSE 11 Alpha kernel? The kernel itself should be quite stable as it's not that far off from 2.6.24. You can get it from: http://ftp.suse.com/pub/projects/kernel/kotd/HEAD/

No problem. Which one? This one: kernel-default-2.6.24-20080126120552.i586.rpm? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=352648 User lolkaantimat@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=352648#c7 --- Comment #7 from Evgeniy Ivanov 2008-01-30 05:57:09 MST --- I wasn't able to build acer_acpi with this kernel. kernel-source-2.6.24-20080129095057 kernel-default-2.6.24-20080129095057 I did mrproper in the /usr/src/linux and then copied config-2.6.24* and did make olconfig && make prepare. After I copied /linux-2.6.24-20080129095057-obj/i386/default/Module.symvers to /ust/src/linux. Also I changed KERNELSRC in acer_acpi's Makefile to /usr/src/linux. =================== KERNELSRC=/usr/src/linux KERNELVERSION=2.6.24-20080129095057-default .. obj-m += wmi-acer.o acer_acpi.o CC=gcc EXTRA_CFLAGS+=-c -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing -fno-common -pipe INCLUDE=-I$(KERNELSRC)/include .. wmi-acer.ko: $(SOURCE) $(MAKE) -C $(KERNELSRC) SUBDIRS=$(PWD) modules .. The last line gives strange error: make -C /usr/src/linux SUBDIRS=/workspace/build/acpi/acer_acpi-0.11_rc3 modules make[1]: Entering directory `/usr/src/linux-2.6.24-20080129095057' CC [M] /workspace/build/acpi/acer_acpi-0.11_rc3/wmi-acer.o /bin/sh: scripts/genksyms/genksyms: No such file or directory make[2]: *** [/workspace/build/acpi/acer_acpi-0.11_rc3/wmi-acer.o] Error 1 make[1]: *** [_module_/workspace/build/acpi/acer_acpi-0.11_rc3] Error 2 make[1]: Leaving directory `/usr/src/linux-2.6.24-20080129095057' make: *** [wmi-acer.ko] Ошибка 2 What should I do with the kernelsource? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=352648 User lolkaantimat@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=352648#c9 Evgeniy Ivanov changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|lolkaantimat@gmail.com | --- Comment #9 from Evgeniy Ivanov 2008-01-31 03:54:36 MST --- Rafał, thanks. I will try 0.9.1 (2.6.24 loads slower than stable). I've tried acer_acpi 0.11 on the 2.6.24. It works. I don't know why 0.10 has such problems with 2.6.22, but it doesn't make any sense because it works with newer kernel. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=352648 User cathectic@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=352648#c11 --- Comment #11 from Carlos Corbacho 2008-02-04 10:15:47 MST --- (In reply to comment #10 from Alexey Starikovskiy)

Problem is resolved with using appropriate external module version.

I would like to note here that I disagree with this assessment and 'resolution'. The 'correct' solution, on SuSE at least, appears to be to use 2.6.24 or newer kernel releases, as there is something in your 2.6.22 & 2.6.23 kernels that breaks acer_acpi. Downgrading to an old, unsupported and deprecated acer_acpi version is simply avoiding a problem with SuSE; it is not a solution, and certainly not one I would support. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=352648 User astarikovskiy@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=352648#c13 --- Comment #13 from Alexey Starikovskiy 2008-02-04 10:47:23 MST --- There is recomendation to use 2.6.22 with 0.9.1 version of acer_acpi. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=352648 User cathectic@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=352648#c15 --- Comment #15 from Carlos Corbacho 2008-02-04 12:52:34 MST --- Is there an easy way to detect this then, and for acer_acpi switch to the pre-2.6.23 behaviour on 2.6.22 for SuSE only kernels? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=352648 User cathectic@gmail.com added comment https://bugzilla.novell.com/show_bug.cgi?id=352648#c17 --- Comment #17 from Carlos Corbacho 2008-02-05 11:00:45 MST --- Well, that's what I'm after. I already have a test for the 2.6.23 vanilla ACPI changes in wmi-acer.c which is: #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,23) So it wouldn't be much of a stretch to change that to: #if (LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,23)) || (defined(CONFIG_SUSE_KERNEL) && LINUX_VERSION_CODE == KERNEL_VERSION(2,6,22)) Would that work? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.