[Xen-ia64-devel] RE: [PATCH] Add prototype for __get_free_pages

["Tian, Kevin" <kevin.tian@xxxxxxxxx>]

Actually all the patches I sent out yesterday including this one is for
non-VTI xenolinux, when I debug xen0+xenU. ;-) Currently there's a bit
problem to boot VTI dom0, and Anthony has a fix who will push out soon.

>-----Original Message-----
>From: Magenheimer, Dan (HP Labs Fort Collins)
>[mailto:dan.magenheimer@xxxxxx]
>Sent: Friday, September 02, 2005 11:18 PM
>To: Tian, Kevin
>Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
>Subject: RE: [PATCH] Add prototype for __get_free_pages
>
>Good catch!  I wonder why this didn't affect non-VTI?  It seems
>like it should have.
>
>> -----Original Message-----
>> From: Tian, Kevin [mailto:kevin.tian@xxxxxxxxx]
>> Sent: Friday, September 02, 2005 2:42 AM
>> To: Magenheimer, Dan (HP Labs Fort Collins)
>> Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
>> Subject: [PATCH] Add prototype for __get_free_pages
>>
>> Due to upgrading to 2.6.13, missing prototypes for several new memory
>> allocate interfaces actually destroy the xenheap. The machine just
MCA
>> once xenU is created and executed in metaphysical mode.
>>
>> Missing prototype makes allocated address from __get_free_pages
>> truncated to 32bit:
>>      d->arch.mm->pgd = pgd_alloc(d->arch.mm);
>>
>> In this case, 0xe0000000040f8000 is truncated to 0x40f8000 as the pgd
>> base address. Later when Xen tries to access this table for querying
>> guest physical -> machine mapping, page fault handler doesn't
consider
>> it as legal hypervisor address and something goes weird. Finally MCA
>> happens.
>>
>> Signed-off-by Kevin Tian <Kevin.tian@xxxxxxxxx>
>>
>> Thanks,
>> Kevin
>>
>>

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel