On 04/25/2017 02:50 PM, Andrew Cooper wrote:
On 24/04/17 18:54, Mohit Gambhir wrote:
Mohit Gambhir (2):
xtf/vpmu: Add Intel PMU MSR addresses
xtf/vpmu: MSR read/write tests for VPMU
arch/x86/include/arch/msr-index.h | 11 +
tests/vpmu/Makefile | 9 +
tests/vpmu/main.c | 504 ++++++++++++++++++++++++++++++++++++++
3 files changed, 524 insertions(+)
create mode 100644 tests/vpmu/Makefile
create mode 100644 tests/vpmu/main.c
Independently from the content of this series, how have you found the
XTF to use? Any issues or unexpected corner cases which can be improved?
I think overall it was fairly easy to use and was very useful for the
kind of testing we wanted to do for VPMU.
It helped me find 3 bugs - one hypervisor crash, one potential state
leak and one missing functionality.
I didnt find any bugs/corner cases in XTF but here are the few nice to
have features that you can consider adding -
1. As said in responses to one of the comments, I think it will be nice
to have some kind of
"expected failure" return type that indicates that it is a known bug.
Thats particularly
useful if the person writing the tests is not a Xen developer. You seem
to have indicated
that xtf_error() means that it is a known error but that wasn't
intuitive to me.
2. Its useful to print __LINE__ number from where an error/failure is
reported.
3. Is test_main() considered 1 test? If so it would be useful to define
a test at a finer granularity.
For instance in VPMU case, each MSR test should really be a separate
test. It would require
adding some semantics to names of the functions that define a test. That
way, its easier to
keep track of the health of the software I think.
4. Is there a way to build just one test? Something like make
test-hvm64-vpmu?
Thanks,
Mohit
~Andrew
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel
