Re: [Xen-devel] Current xen-unstable xen crashing in boot

[Keir Fraser <keir.fraser@xxxxxxxxxxxxx>]

Yes, the dependencies on xen.lds are broken, as only xen.lds.S is indicated
as a dependency. Completely clean build required.

I should remember to fix this, and similar for asm-offsets.s.

 -- Keir

On 15/08/2009 13:03, "Wei, Gang" <gang.wei@xxxxxxxxx> wrote:

> I used to meet similar problem before I do a clean build. The
> xen/arch/x86/xen.lds could not be regenerated without a clean build, which
> would cause __per_cpu_end still equal __per_cpu_start + (32 << 13).
> 
> Please confirm it is not caused by this factor.
> 
> Jimmy
> 
> Jeremy Fitzhardinge wrote:
>> On 08/14/09 13:48, Jeremy Fitzhardinge wrote:
>>>> (XEN) Early fatal page fault at e008:ffff828c8015415c
>> (cr2=ffff83007fc75008, ec=0000)
>>>> (XEN) Stack dump: 000000000007fc76 000000000000000c
>> ffff83007fc77ff0 ffff83007fc76000 ffff828c8026fe58
>> 0000000000000262 0000000000000020 0180000000000000
>> 000000000000010c ffff828c802a6380 ffff83007fc75008
>> 0000000000000001 ffff830000000000 ffff828c80219b0c
>> ffff83007fc76000 0000000000000000 ffff828c8015415c
>> 000000000000e008 0000000000010082 ffff828c8026fdd8
>> 0000000000000000 ffff828c80154158 ffff828c8026fde8
>> 0000000000000262 00000000000002e2 0000000000000086
>> 0000000000000262 0000000000000000 000000008026fe18
>> ffff83007fc75008 ffff828c00000027 ffff83007fc76000
>> ffff828c8021dd60 0000000000000002 000000007fc42000
>> ffff83000008bfc0 ffff83000008bf40 0000000000002000
>> ffff828c8026fe68 ffff828c80154f85 ffff828c8026fe78
>> ffff828c80154fa5 ffff828c8026ff18 ffff828c802343ef
>> 0000000000000000 0000000000000000 0000000000a1e000
>> 0000000000aca3e0 ffff83000008bf40 0000000000000000
>> 00000000ffffffff ffff83000000000c 0000000800000000
>> 000000010000006e 0000000000000003 00000000000002f8
>> 0000000000000000 0000000000000000 0000000000000000
>> 0000000000000000 0000000000000000 0000000000000000
>> 0000000000067e8c ffff828c801000b5 0000000000000000
>> 0000000000000000 0000000000000000 0000000000000000
>> 0000000000000000 0000000000000000 0000000000000000
>> 0000000000000000 0000000000000000 0000000000000000
>> 0000000000000000 0000000000000000 0000000000000000
>> 0000000000000000 0000000000000000 0000000000000000
>> 0000000000000000 0000000000000000 0000000000000000
>> 0000000000000000 0000000000000000 0000000000000000
>> 0000000000000000 0000000000000000 0000000000000000
>> 0000000000000000 00000000fffff000
>>>> 
>>>> This appears to be:
>>>> (gdb) x/i 0xffff828c8015415c
>>>> 0xffff828c8015415c <map_pages_to_xen+112>: mov    (%rax),%rax (gdb)
>>>> x/i 0xffff828c80154f85 0xffff828c80154f85
>>>> <__memguard_change_range+237>: leaveq
>>>> 
>>>> but I can't find a good enclosing frame from there (is
>> there a tool to pretty-print these stack traces?).
>>>> 
>>>> 
>>> 
>>> I bisected it down to:
>>> 
>>> The first bad revision is:
>>> changeset:   20038:1197585e32b7
>>> user:        Keir Fraser <keir.fraser@xxxxxxxxxx>
>>> date:        Fri Aug 07 17:29:50 2009 +0100
>>> summary:     x86: Increase default max CPUs to 64.
>>> 
>> 
>> Reverting this change on current tip works fine for me.
>> 
>>    J



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel