Kernel gets an OOPS during load of apm. apm: BIOS version 1.2 Flags 0x03 (Driver version 1.16) Unable to handle kernel paging request at virtual address 805259d8 printing eip: 0000a286 *pde = 00000000 Oops: 0000 CPU: 0 EIP: 0010:[<0000a286>] Not tainted EFLAGS: 00010046 eax: 0000530e ebx: 00000000 ecx: 00000102 edx: 00000000 esi: c059bdae edi: 00000292 ebp: 00000000 esp: c10f5f88 ds: 0000 es: 0000 ss: 0018 Process kapmd (pid: 3, stackpage=c10f5000) Stack: 00000000 00000292 00000018 00000018 00000000 00000000 00000000 c0125214 c059bdae c10f424c 0008e000 ffd8ec1e 0000530e 00000000 00000102 c10f5fc8 0000b3c0 00000f11 0000b435 c0125214 00000000 00000018 00000018 c106bf6c Call Trace: [<00000292>] [<0008e000>] [<0000530e>] [<0000b3c0>] [<00000f11>] [<0000b435>] [<000008a6>] [<0000b3c0>] Code: 2e ff 1d d8 59 12 c0 0f 92 c3 5d 5f 07 1f 89 c5 8b 44 24 2c I am also attaching my .config for you to look over.
Created attachment 12175 [details] kernel configuration
Had that, too, in -r6. Turned off kernel exec protection in GRsec, and the problem went away.
I have a similar problem since I changed my kernel on my dual PIII (Abit VP6 motherboard) from vanilla-2.4.21 to gentoo-sources-2.4.20-r6 (this change is due to netfilter modules I need for my firewall... but it's another story... :-P). Sometimes when I load apm module (I'm on a smp kernel, so I load apm with option 'power_off=1') I've got several error messages like those reported by Chris Gianelloni. My problem, as I can see from my error messages, it's related with line 1669 of apm.c.
Why not use ACPI instead of APM? APM is deprecated and not fully supported by pretty much any motherboard.
Lots of older machines rely on APM and either do not have ACPI support at all, or have buggy ACPI support in their bios. So APM is still a real solution for a lot of users.
This particular machine only has APM support. ACPI is the same as no power management at all, since it does nothing.
Same here: APM works without side effects, ACPI does not.
Any more news on this? Could you possibly try a 2.6 kernel and see if the errors persist? Have you tried a kernel newer than 2.4.20?
Personally, I don't have the machine running anymore. I'm using all 2.6 here now, anyway. I don't know if the crash would have occurred on a newer kernel, since I never got a newer one on that machine. I would say we just RESOLVE-CANTFIX this one and be done with it, since this kernel has been out of portage for quite a long time now. This bug is almost a year old... *grin*
Done and Done.
Hmmm, has anyone checked what happens when grsec with kernel exec protection is used together with apm in a 2.6 kernel? I'm currently at gentoo-dev-sources-2.6.4-r1, and have no problems with apm, but that's without grsec...
