RE: [Xen-devel] X86_64 "assert" when booting 64-bit image.

["Petersson, Mats" <Mats.Petersson@xxxxxxx>]

> -----Original Message-----
> From: Keir Fraser [mailto:Keir.Fraser@xxxxxxxxxxxx] 
> Sent: 09 February 2006 16:31
> To: Petersson, Mats
> Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
> Subject: Re: [Xen-devel] X86_64 "assert" when booting 64-bit image.
> 
> 
> On 9 Feb 2006, at 16:16, Petersson, Mats wrote:
> 
> >> That debug code is totally ancient. Ian may know whether 
> it has any 
> >> relevance any more.
> >
> > Thanks Keir.
> >
> > I #if 0'd out the test and it flies through the rest of the stuff 
> > until the point where it got without debug. But I don't think the 
> > checking code was added purely because it seemed fun to add in the 
> > first place, so I'm still a bit concerned that it may actually be 
> > pointing at something that causes a problem... Is it really 
> safe to remove it?
> 
> Mats,
> 
> I think that the correct thing to do is to remove that whole 
> middle portion of __shadow_status(). That is, the entire 
> outermost 'if' 
> statement. (That is, the 'if ( VALID_MFN()....' all the way 
> to 'return 0; }').
> 
> Can you please try that out and see how it works for you?

I've done that [I did it first using a #if 0, but I've now hit the
"delete" key for it...] - Patch attached. 

I don't know if I need to add this for removing existing lines of code:
Signed off by: Mats Petersson  mats.petersson@xxxxxxx

--
Mats
> 
> I actually think there is another problem here. 
> PGT_fl1_shadow shadow pages are looked up by the first guest 
> pfn in that superpage extent, but that first guest pfn may 
> itself be a pagetable page, and no pfn can currently have 
> more than one 'shadow status'. That needs more investigation though...
> 
>   -- Keir
> 
> 
>

Attachment: shadow.h.patch
Description: shadow.h.patch

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