Re: [PATCH 3/4] xen: more debugging in the e820 parsing

[Stefano Stabellini <stefano.stabellini@eu.citrix.com>]

On Wed, 13 Apr 2011, Konrad Rzeszutek Wilk wrote:
> > I have a machine here in which the e820 printed by Xen and the e820
> > in dom0 *before* any modifications differs:
> > 
> > (XEN) Xen-e820 RAM map:
> > (XEN)  0000000000000000 - 000000000009fc00 (usable)
> > (XEN)  000000000009fc00 - 00000000000a0000 (reserved)
> > (XEN)  00000000000e0000 - 0000000000100000 (reserved)
> > (XEN)  0000000000100000 - 00000000beb96000 (usable)
> > (XEN)  00000000beb96000 - 00000000bed97000 (ACPI NVS)
> > (XEN)  00000000bed97000 - 00000000bf651000 (usable)
> > (XEN)  00000000bf651000 - 00000000bf6e9000 (ACPI NVS)
> > (XEN)  00000000bf6e9000 - 00000000bf6ec000 (usable)
> > (XEN)  00000000bf6ec000 - 00000000bf6ff000 (ACPI data)
> > (XEN)  00000000bf6ff000 - 00000000bf700000 (usable)
> > 
> > [    0.000000] e820_region: type=1 start=0000000000000000 end=000000000009fc00
> > [    0.000000] e820_region: type=2 start=000000000009fc00 end=00000000000a0000
> > [    0.000000] e820_region: type=2 start=00000000000e0000 end=0000000000100000
> > [    0.000000] e820_region: type=1 start=0000000000100000 end=00000000beb96000
> > [    0.000000] e820_region: type=4 start=00000000beb96000 end=00000000bed97000
> > [    0.000000] e820_region: type=1 start=00000000bed97000 end=00000000bf651000
> > [    0.000000] e820_region: type=4 start=00000000bf651000 end=00000000bf6e9000
> > [    0.000000] e820_region: type=1 start=00000000bf6e9000 end=00000000bf6ec000
> > [    0.000000] e820_region: type=3 start=00000000bf6ec000 end=00000000bf6ff000
> > [    0.000000] e820_region: type=1 start=00000000bf6ff000 end=00000000bf700000
> 
> 
> > [    0.000000] e820_region: type=2 start=00000000fec00000 end=00000000fec01000
> > [    0.000000] e820_region: type=2 start=00000000fee00000 end=00000000fee01000
> 
> Ah, so the IOAPIC regions don't show up in the E820. Do they show up in the
> E820 printed by the Linux kernel? If I use the Xen E820 output and what the
> guest prints for its E820 I seem to get even this "hidden" area.


Yes, that's right. However it might be confusing if you are trying to
debug the transformations made by xen/setup.c to the e820...
In any case I don't feel strongly about this, it is just that I found
myself adding these printk's more than once so I thought that it might
be good to have them in any case for debugging. On the other hand I add
printk's all the time so surely it won't kill me to add these ones too.