Re: [PATCH 2/8] xen/mmu: Add the notion of identity (1-1) mapping.

From: Ian Campbell
Date: Tue Jan 04 2011 - 11:59:46 EST

On Tue, 2011-01-04 at 16:53 +0000, Ian Campbell wrote:
> If I do "__set_phys_to_machine(X, X)" where X happens to currently
> correspond to p2m_mid_missing won't that cause all pfn entries in the
> range covered by p2m_top[topidx] (i.e. quite a few, like 512^2 pages
> or
> something) to switch from missing to identity? Similarly for
> p2m_top[topidx][mididx]?
> Perhaps ranges of identity bits are often well aligned with the
> boundaries in the p2m 3-level tree but wouldn't that just be
> coincidence?

I wonder if it would make sense to have a debugfs file which exports the
p2m, for the purposes of eye-balling it for this sort of issue?
comparing to the e820 etc...

Maybe the whole thing in raw form would be overkill but spitting out a
list of ranges of identity, invalid, normal pages, plus annotation
regarding whether those come from a p2m_mid_identity style page or a
leaf node, might be occasionally useful.

Ian Campbell
Current Noise: Place of Skulls - The Watchers

Call toll free number before digging.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at