RE: DT with memory bank of size 0 (WAS: Re: AW: AW: Colibri imx8qxp: Missing kernel boot module)

[Daniel Wagner2 <Daniel.Wagner2@xxxxxxxxxxxxxxxxxx>]

Hi Julien,

> >>> Do you have Linux running on baremetal on this board? If so would
> >>> you mind to dump the DT from the userspace (via /proc/device-tree)
> >>> this time?
> >
> > I do have linux running on baremetal on the plattform.
> > You were right, after the boot, the memory node contains the second
> > bank with size=0.
> >
> > memory@80000000 {
> >     device_type = "memory";
> >     reg = < 0x00 0x80200000 0x00 0x7fe00000 0x08 0x80000000 0x00 0x00
> >;
> > };
> 
> Thank you for the testing. I am glad we finally found the reason!
> 
> >>>
> >>> In any case, we may want to relax the check in Xen. Any opinions?
> >>
> >> Yeah, ignoring a bank with size=0 is fine. I checked the epapr and it
> > doesn't
> >> specify that size=0 is invalid, so I think it is actually better to
> >> ignore
> > it and
> >> continue even from a spec perspective.
> 
> I have sent a patch to ignore empty bank ([1]). Would you mind to test it?

Yes, Xen is now able to boot Dom0, even if the chosen node comes after 
the memory node in the dtb.

> The patch is based on unstable, but it should be easy to apply it on older
release.

Your patch was applied to our sources, which are based on 
Branch "imx_5.4.24_2.1.0" of imx-xen [2], 
with no modification needed.

Thank you for your support and help!

Best regards,
Daniel

> [1] https://lore.kernel.org/xen-devel/20200918171116.29005-1-
> julien@xxxxxxx/

[2] git://source.codeaurora.org/external/imx/imx-xen.git

Attachment: smime.p7s
Description: S/MIME cryptographic signature