[PATCH 0/5] x86/mm: out-of-memory errors in p2m ops

[Tim Deegan <tim@xen.org>]

These patches begin to address error handling when the shadow/hap
pool can't allocate any more p2m pages.

Patch 1 is a bit of preliminary tidying.
Patch 2 prints a warning on error so it can be diagnosed.
Patches 3-5 add checks to direct callers of the p2m allocation functions.

With these patches applied, pool allocation errors will be
propagated as far as the set_p2m_entry() interface.
Unfortunately there are many callers of that that don't check the
return code -- PoD and mem_paging are the major cuplrits, but
there are other cases where the error is discarded further up the
stack (e.g. some callers of set_mmio_p2m_entry()).  Chasing those
up is work for another day.

Cheers,

Tim.