Re: [v6 PATCH 21/21] selftests/x86: Add tests for User-Mode Instruction Prevention

From: Ricardo Neri
Date: Fri Mar 10 2017 - 18:38:41 EST


On Wed, 2017-03-08 at 07:56 -0800, Andy Lutomirski wrote:
> On Tue, Mar 7, 2017 at 4:32 PM, Ricardo Neri
> <ricardo.neri-calderon@xxxxxxxxxxxxxxx> wrote:
> > Certain user space programs that run on virtual-8086 mode may utilize
> > instructions protected by the User-Mode Instruction Prevention (UMIP)
> > security feature present in new Intel processors: SGDT, SIDT and SMSW. In
> > such a case, a general protection fault is issued if UMIP is enabled. When
> > such a fault happens, the kernel catches it and emulates the results of
> > these instructions with dummy values. The purpose of this new
> > test is to verify whether the impacted instructions can be executed without
> > causing such #GP. If no #GP exceptions occur, we expect to exit virtual-
> > 8086 mode from INT 0x80.
> >
> > The instructions protected by UMIP are executed in representative use
> > cases:
> > a) the memory address of the result is given in the form of a displacement
> > from the base of the data segment
> > b) the memory address of the result is given in a general purpose register
> > c) the result is stored directly in a general purpose register.
> >
> > Unfortunately, it is not possible to check the results against a set of
> > expected values because no emulation will occur in systems that do not have
> > the UMIP feature. Instead, results are printed for verification.
>
> You could pre-initialize the result buffer to a bunch of non-matching
> values (1, 2, 3, ...) and then check that all the invocations of the
> same instruction gave the same value.

Yes, I can do this. Alternatively, I can check in the test program if
the CPU has UMIP and only run the tests in that case.

>
> If you do this, maybe make it a follow-up patch -- see other email.

Great! Thank you!

Thanks and BR,
Ricardo