Bug 5177
Summary: | intermittent system hang - over temperature? - ASUS P5AD2 | ||
---|---|---|---|
Product: | ACPI | Reporter: | Alex Unigovsky (unik) |
Component: | Power-Thermal | Assignee: | Len Brown (lenb) |
Status: | REJECTED UNREPRODUCIBLE | ||
Severity: | normal | CC: | acpi-bugzilla, protasnb, rodd, xoen |
Priority: | P2 | ||
Hardware: | i386 | ||
OS: | Linux | ||
Kernel Version: | 2.6.13 | Subsystem: | |
Regression: | --- | Bisected commit-id: | |
Attachments: |
Output of lspci -vvx on 2.6.13 with ACPI2.0/APIC turned off
Full dmesg of system booting 2.6.13-mm1 (with vestfb-tng and sk98lin patches) and ACPI2.0/APIC turned off in BIOS |
Description
Alex Unigovsky
2005-09-03 03:55:32 UTC
Created attachment 5875 [details]
Output of lspci -vvx on 2.6.13 with ACPI2.0/APIC turned off
If i need to supply additional info, please let me know. I'll do my best :)
Created attachment 5910 [details]
Full dmesg of system booting 2.6.13-mm1 (with vestfb-tng and sk98lin patches) and ACPI2.0/APIC turned off in BIOS
Maybe this can help... But I see that this bug is just ignored... :(
if you have ACPI and the IOAPIC enabled in the bios, can you boot with "acpi=noirq" or "pci=noacpi" and get working devices and HT at the same time? Well, yes, acpi=noirq helped. Now I see 2 CPUs and am able to boot and use SiL controller. Thanks a lot. Didn't try pci=noacpi, should I? The only thing that bugs me is this "ACPI Interpreter Disabled" thing in dmesg, but I think that I'm supposed to see it :) How else can I help? ACPI and APIC were enabled in BIOS for the "acpi=noirq" boot, yes? Please attach the complete dmesg and /proc/interrupts from this boot. Also, please attach the output from acpidump, available in pmmtools here: http://ftp.kernel.org/pub/linux/kernel/people/lenb/acpi/utils Are you able to cappture the "debug" console output for the failure case? If the dmesg above shows no clues we'll need it. After booting into 2.6.14-rc1-mm1 I had no more error/warning messages when booting with ACPI and APIC enabled, and I didn't have to issue acpi=noirq. Hyper-threading worked fine, and all ATA/SATA/whatever controllers were initialized successfully (the only strange thing I noted is the longer-than-usual init time for IT8212, but it never gave any error). That is good news. The bad news is that the system is very unstable, and freezes just about 30 minutes to several hours after booting. I have no clue how to collect diagnostics/debugging info, because I cannot predict the freeze - X just stops responding, keyboard goes down (completely - i.e. numlock doesn't switch). Can you tell me a way that I can diagnose a problem in this situation? Or maybe some suggestions... Kernel 2.6.13-rc1-mm1 compiled with GCC 4.0.1. Dmesg/interrupts/acpidump/etc will follow shortly. The only other patch applied was sk98lin from syskonnect.com, because sky2 (even with enabled workarounds) still didn't work for me. PS: Sorry if these problems are unrelated, I'm quite new to kernel bugzilla, and didn't know where to post. And sorry for my bad English :) On 2.6.14-rc2 (vanilla + reiser4 from -mm) ACPI works fine (no errors, everything detects/inits fine, HT works). I will post further info after I get a more powerful CPU cooler, because I suspect in-work freezes are caused by CPU temp (according to sensors) reaching 70 celsius and staying there). Sorry for the delay. Forgot to say, the temperature issue only exists on 2.6.13-14. On 2.6.12, CPU temp always stays at 54-56 celsius. It looks like the problem was fixed. Any objections to closing the bug? Thanks. If you are still seeing a problem, please re-open this bug report, and... test if it still happens with CONFIG_HWMON=n and ACPI enabled, or with "acpi=off" If it fails with ACPI enabled, but works with "acpi=off", please attach the complete dmesg from the ACPI boot and the output from acpidump. |