Created attachment 48082 [details] dmesg after booting without noapic an dnolapic Boots very slowly without noapic and nolapic, presumably from some sort of bug with ACPI/Bios/Firmware... I have tried this with both mainline and the Archlinux default kernel, both of which have the same issues, though the Arch kernel boots a little more quickly without noapic/noalpic. Lots of these bugs: [Firmware Bug]: ACPI: BIOS _OSI(Linux) query ignored I fear this is affecting performance, as without apic and lapic dynticks don't work correctly, to my knowledge. Attached is a dmesg from booting without noapic and nolapic, a dmesg with noapic nolapic, as well as the output of lspci and lsusb. Kernel: Linux mattop 2.6.37 #2 ZEN SMP PREEMPT Wed Feb 16 23:17:41 CST 2011 x86_64 Intel(R) Core(TM) i7 CPU Q 740 @ 1.73GHz GenuineIntel GNU/Linux Specs: Processor: Intel Core i7-740QM quad-core processor Chipset: Intel HM55 Graphics: NVIDIA GeForce GTX 460M 1.5GB GDDR5 VRAM WLAN: 802.11 b/g/n (@ 2.4GHz), Bluetooth v2.1 LAN: 10/100/1000 Mbps Ethernet Memory: 8GB DDR3
Created attachment 48092 [details] dmesg with noapic and nolapic
Created attachment 48102 [details] output of lspci
Created attachment 48112 [details] output of lsusb
After some more digging it seems this is also related to EFI/UEFI. I am unable to load uefivars, as modprobe reports the device doesn't exist on this system, leading me to believe whatever firmware it uses hasn't had support yet been built into the kernel...it'd be nice to have that so I can use grub2 with EFI for faster booting.
> [Firmware Bug]: ACPI: BIOS _OSI(Linux) query ignored please attach the output from acpidump so we can verify that you can ignore those. re: speed issue please try booting with "intel_idle.max_cstate=0"
Created attachment 67462 [details] acpidump The speed issue is gone as of ... .38 or so. There are still firmware issues and I've even forced OSI as Linux, which makes the bugs say they're using Linux instead of the default Windows emulation thing... Either way, I recently fixed my DSDT and tried it, and that ended up actually making my plugged in status and some other ACPI functions break. Attached is the acpidump. I'll add the original and fixed versions of the dsdt files below too.
Created attachment 67472 [details] Asus G73JW original dsdt
Created attachment 67482 [details] Asus G73JW "fixed" dsdt
FYI, the repeated OSI(Linux) warning has been reduced to one, as of this patch in Linux 3.1 merge window: commit 8997621bb2daaf19a4e9d82f118224159d8054e2 Author: Len Brown <len.brown@intel.com> Date: Tue Aug 2 00:45:48 2011 -0400 ACPI print OSI(Linux) warning only once This message gets repeated on some machines: https://bugzilla.kernel.org/show_bug.cgi?id=29292 Signed-off-by: Len Brown <len.brown@intel.com>
So you at least took the annoying factor back down to 1 from something like 50...that's pretty nice. I guess I still need to figure out the best way to fix my DSDT? It compiles without errors or warnings with the fixes I found, it's just as I said it tends to break some things...
I merged in your acpi fixes with my local repo (a clone of git.zen-kernel.org/zen-stable/ ) and it merged in well, but I still get repeated firmware errors...it's a 2.6 kernel, but I'd have expected it to work there too. I'll try 3.1 when I get a chance, with your patch.
It's great that the kernel bugzilla is back. Can you please verify if the problem still exists in the latest upstream kernel?
bug closed as there is no response from the bug reporter. please feel free to reopen it if the problem still exists in the latest upstream kernel.