5.3-RELEASE TODO

[Adam K Kirchhoff]

Scott Long wrote:

>On Fri, 24 Sep 2004, Adam K Kirchhoff wrote:
>  
>
>>David Syphers wrote:
>>
>>    
>>
>>>On Friday 24 September 2004 01:43 am, Oliver Brandmueller wrote:
>>>
>>>
>>>      
>>>
>>>>On Fri, Sep 24, 2004 at 01:41:22AM -0600, Scott Long wrote:
>>>>
>>>>
>>>>        
>>>>
>>>>>This is a list of open issues that need to be resolved for FreeBSD 5.3.
>>>>>If you have any updates for this list, please e-mail re at FreeBSD.org.
>>>>>
>>>>>Show stopper defects for 5.3-RELEASE
>>>>>
>>>>>
>>>>>          
>>>>>
>>>>There seems to be an issue with SMP and ACPI which came into the tree
>>>>somewhere after BETA3, see the corresponding threads in current@
>>>>
>>>>SMP problems with -CURRENT? (and BETA5)
>>>>
>>>>
>>>>        
>>>>
>>>This thread mentions APIC several times (one person saying the problems went
>>>away when a kernel "without device apic" was made, and someone else saying,
>>>"clearly there was a bug introduced in the ACPI or APIC code").
>>>
>>>
>>>      
>>>
>>That would be me.  Here's a summary of what I encountered (all this can
>>be found in the "SMP problems with -CURRENT (and BETA5)":
>>
>>With HTT enabled in the BIOS:
>>
>>1) With ACPI and APIC enabled: Locking up after detecting my hard drives
>>and launching the second processor.
>>
>>2) With ACPI disabled and APIC enabled:  Crashing with a "Fatal trap 9".
>>
>>3) With both ACPI and APIC disabled: Success.
>>
>>With HTT disabled in the BIOS:
>>
>>1) With ACPI and APIC enabled:  Locking up after detecting my hard drives.
>>
>>And, I'd agree with your assesment that, no matter the cause, this is a
>>pretty significant show-stopping bug :-)
>>
>>Adam
>>
>>
>>    
>>
>
>Agreed.  John and Nate, can you two please look into this and report back?
>Also, Adam, would it be possible for you to step through the RELENG_5
>commits and find the point where the problem occurs?
>
>  
>

I can tell you right now that it's definitely between BETA4 and 
BETA5...  I probably won't have much time to through the various commits 
for a little over a week, though.   Once I'm back, though, I'll gladly 
give it a shot if no one beats me to it.

Adam