nested EPT: fixing wrong handling for L2 guest's direct mmio access
authorYang Zhang <yang.z.zhang@Intel.com>
Fri, 17 Jan 2014 15:00:21 +0000 (16:00 +0100)
committerJan Beulich <jbeulich@suse.com>
Fri, 17 Jan 2014 15:00:21 +0000 (16:00 +0100)
commit0b988ba711171b39aed9851cfe90fded50f775c5
tree04f2f4f69ab4cbb174a3cc923f74aaed88d72806
parente9af61b969906976188609379183cb304935f448
nested EPT: fixing wrong handling for L2 guest's direct mmio access

L2 guest will access the physical device directly(nested VT-d). For such access,
Shadow EPT table should point to device's MMIO. But in current logic, L0 doesn't
distinguish the MMIO whether from qemu or physical device when building shadow EPT table.
This is wrong. This patch will setup the correct shadow EPT table for such MMIO ranges.

Signed-off-by: Yang Zhang <yang.z.zhang@Intel.com>
Acked-by: Tim Deegan <tim@xen.org>
xen/arch/x86/mm/hap/nested_hap.c
xen/include/asm-x86/hvm/nestedhvm.h