[PATCH RFC V10 0/5] Basic guest memory introspection support

[PATCH RFC V10 0/5] Basic guest memory introspection support

[Razvan Cojocaru]

Hello,

We had to modify Xen in order to be able to detect rootkits in HVM
guests, in a way that allows an application that runs in dom0 (or a
similarly privileged domain) to control what the guest is allowed to do
once a threat is detected. This has been done over the mem_event mechanism.

To this end, we needed to:

1. Be able to execute the current instruction without allowing it to
write to memory. This is done based on new mem_event response
fields sent from the controlling application.

2. Have the guest as responsive as possible amid all the processing. So
we had to cache some information with each mem_event sent.

3. Not allow the hypervisor to disable sending information about
interesting MSR events.

4. Add an additional mem_event type, namely a VMCALL event - in order to
do its work, our application occasionally triggers VMCALLs in the guest
(not included in the current series, but included in the initial RFC
series).

5. Add an additional libxc function that allows triggering page faults
in the guest.

Changes since V9:
 - Addressed discussed issues (please see individual patches for
   details).

Patches:

xen: Handle resumed instruction based on previous mem_event reply
xen, libxc: Request page fault injection via libxc
xen, libxc: Force-enable relevant MSR events
xen: Optimize introspection access to guest state
xen: Emulate with no writes


Thanks,
Razvan Cojocaru

[PATCH RFC V10 0/5] Basic guest memory introspection support

[Razvan Cojocaru]

Hello,

As stated originally, we had to modify Xen in order to be able to detect
rootkits in HVM guests, in a way that allows an application that runs in
dom0 (or a similarly privileged domain) to control what the guest is
allowed to do once a threat is detected. This has been done over the
mem_event mechanism.

To this end, we needed to:

1. Be able to execute the current instruction without allowing it to
write to memory. This is done based on new mem_event response
fields sent from the controlling application.

2. Have the guest as responsive as possible amid all the processing. So
we had to cache some information with each mem_event sent.

3. Not allow the hypervisor to disable sending information about
interesting MSR events.

4. Add an additional mem_event type, namely a VMCALL event - in order to
do its work, our application occasionally triggers VMCALLs in the guest
(not included in the current series, but included in the initial RFC
series).

5. Add an additional libxc function that allows triggering page faults
in the guest.

Changes since V9:
 - The "xen, libxc: Request page fault injection via libxc" has
   been reworked to extend HVMOP_inject_trap instead of a new
   libxc hypercall.

Patches:

xen: Handle resumed instruction based on previous mem_event reply
xen, libxc: Request page fault injection via libxc
xen, libxc: Force-enable relevant MSR events
xen: Optimize introspection access to guest state
xen: Emulate with no writes


Thanks,
Razvan Cojocaru