Re: [Xen-devel] [PATCH] x86/boot: re-arrange how/when we do disk I/O

[Paul Durrant <Paul.Durrant@xxxxxxxxxx>]

> -----Original Message-----
> From: Julien Grall [mailto:julien.grall@xxxxxxx]
> Sent: 26 June 2017 14:04
> To: Jan Beulich <JBeulich@xxxxxxxx>
> Cc: Andrew Cooper <Andrew.Cooper3@xxxxxxxxxx>; Paul Durrant
> <Paul.Durrant@xxxxxxxxxx>; xen-devel <xen-devel@xxxxxxxxxxxxxxxxxxxx>;
> Lars Kurth <lars.kurth@xxxxxxxxxx>
> Subject: Re: [PATCH] x86/boot: re-arrange how/when we do disk I/O
> 
> Hi,
> 
> On 12/06/17 17:59, Julien Grall wrote:
> > Hi Jan,
> >
> > On 12/06/17 16:27, Jan Beulich wrote:
> >>>>> On 12.06.17 at 17:11, <JBeulich@xxxxxxxx> wrote:
> >>> We place the trampoline no lower than at 256k, so we have ample space
> >>> to read the MBRs of BIOS disks into an aligned buffer right below the
> >>> trampoline (not doing so has been found to be a problem on a buggy
> BIOS
> >>> coming with a Skull Canyon NUC). To facilitate that move MBR reading
> >>> past EDD info retrieval.
> >>>
> >>> Also add a wrap check to the EDD info retrieval loop, to match that in
> >>> the MBR reading one.
> >>>
> >>> Reported-by: Paul Durrant <Paul.Durrant@xxxxxxxxxx>
> >>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
> >>> Tested-by: Paul Durrant <Paul.Durrant@xxxxxxxxxx>
> >>
> >> As to 4.9 inclusion, I think I should have mentioned right away
> >> that such a change of course isn't without risk: We're trying to
> >> work around a BIOS bug by relying on other information
> >> coming from the BIOS being correct / usable. Obviously there is
> >> the risk that this regresses with some other BIOS having a bug
> >> elsewhere.
> >>
> >> I'm therefore leaning towards suggesting to take this change
> >> only for master right now, and consider pulling it into the 4.9
> >> branch in a couple of weeks time, when the patch has seen at
> >> least a little more testing on a wider set of systems.
> >
> > I agree on your conclusion. I would suggest to add a release note if it
> > is not backported before the release.
> >
> > I will keep track of this.
> 
> I haven't seen any update on this. Can someone describe briefly the
> problem so we can add it in the section "known issues" of the release notes?
> 

Julien,

  The symptom is an apparent failure to progress beyond the grub stage when 
trying to boot Xen on a skull canyon. You see the messages about loading the 
modules, but then nothing more except maybe a reset of the vga mode (depending 
on what your xen command line looks like).
  The problem is a buggy BIOS which wedges in the int13 handler in some cases. 
Xen 4.9 happens to tickle this bug whereas Xen 4.8 got away with it.

  Cheers,

    Paul

> Cheers,
> 
> --
> Julien Grall

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel