Bug 141202 - machine freezes unless noapic is given
Summary: machine freezes unless noapic is given
Status: RESOLVED INVALID
Alias: None
Product: SUSE Linux 10.1
Classification: openSUSE
Component: Kernel (show other bugs)
Version: Alpha 4
Hardware: x86-64 Other
: P5 - None : Normal (vote)
Target Milestone: ---
Assignee: Andreas Kleen
QA Contact: E-mail List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-01-02 23:13 UTC by Rihards Olups
Modified: 2006-01-18 07:12 UTC (History)
1 user (show)

See Also:
Found By: Other
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---

Attachments
hwinfo output with and without noapic (35.14 KB, application/x-bzip2)
2006-01-04 15:57 UTC, Rihards Olups 		Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Rihards Olups 2006-01-02 23:13:50 UTC 
Asrock S939 SIS756 motherboard
CPU: AMD Athlon(tm) 64 Processor 3200+ stepping 02
2.6.15-rc5-git3-2-default

in default configuration, machine freezes shortly after the boot. this is usually triggered by intensive disk access.

logfiles did not contain anything very suspicios except some messages about apic errors.

seems that adding noapic to kernel commandline helped - i've been using this machine for several hours without a glitch.
Comment 1 Olaf Kirch 2006-01-04 09:19:55 UTC 
Could you please supply the output of hwinfo with and without noapic?
Comment 2 Rihards Olups 2006-01-04 15:57:04 UTC 
Created attachment 61954 [details]
hwinfo output with and without noapic
Comment 3 Olaf Kirch 2006-01-09 15:18:59 UTC 
Hm, the boot message log shows

<7>APIC error on CPU0: 04(40)		(exactly once)
<7>APIC error on CPU0: 40(40)		(many many times)

04 is Send accept error
40 is Received illegal vector

Andi, any clue what might cause this?
Comment 4 Thomas Renninger 2006-01-09 18:52:27 UTC 
Could this have to do that the whole ACPI subsystem could not be loaded and some IO/apic configuration might be missing(#141375)?:

 ACPI-0311: *** Error: ns_search_and_enter: Bad character in ACPI Name: 43035350
    ACPI-0292: *** Error: Looking up [0x43035350] (NON-ASCII)
 in namespace, AE_BAD_CHARACTER
    ACPI-0115: *** Error: acpi_load_tables: Could not load namespace: AE_BAD_CHARACTER
    ACPI-0123: *** Error: acpi_load_tables: Could not load tables: AE_BAD_CHARACTER
ACPI: Unable to load the System Description Tables
Comment 5 Andreas Kleen 2006-01-09 18:57:27 UTC 
Yes, many modern systems don't work in APIC mode without ACPI since
the mptables are untested.

Did ACPI work with earlier kernels on the system?
Comment 6 Rihards Olups 2006-01-10 08:00:30 UTC 
suse 10.0 refused to install (did not find any ide or sata blockdevices), so 10.1a4 is the first operating system on this machine. temporalily upgrading to rc6-git<something> did not help.
Comment 7 Andreas Kleen 2006-01-17 19:37:06 UTC 
Should probably figure out why the ACPI stack can't parse your ACPI tables. 
Can you please attach acpidmp output?
Comment 8 Rihards Olups 2006-01-18 06:55:52 UTC 
it was already found out at bug 141375 - seems that two bytes are swapped in bios. in any case, acpidmp is also attached there.

i informed asrock the next day, but have received no answer so far. maybe somebody else with more influential email address can bug them ;)
Comment 9 Andreas Kleen 2006-01-18 07:12:44 UTC 
Ok BIOS bug -> INVALID.