On 03/27/2018 05:25 PM, Christian Borntraeger wrote: > > > On 03/27/2018 03:29 PM, David Hildenbrand wrote: >> On 27.03.2018 13:55, Janosch Frank wrote: >>> The IBM team wants to join in on testing s390x KVM thoroughly, so I >>> prepared a lot of memory related tests and some IBM z14 related tests. >>> >>> v3: >>> * Merged sthyi inline assembly and improved report messages >>> * Cosmetical fixes >>> * Other inline assembly fixes >>> >>> v2: >>> * Improved inline assembly for gs and vector tests >>> * Improved test skipping messages for tests needing specific >>> failities >>> * General improvements from review >>> >>> Janosch Frank (8): >>> s390x: Add sthyi tests >>> s390x: Add privileged operation exceptions handling >>> s390x: Add storage keys tests >>> s390x: Add diag10 tests >>> s390x: Add pfmf tests >>> s390x: Add cmm tests >>> s390x: Add vector tests >>> s390x: Add simple guarded storage tests >>> >>> lib/s390x/asm/arch_def.h | 28 ++++++++ >>> lib/s390x/asm/mem.h | 52 +++++++++++++++ >>> lib/s390x/interrupt.c | 7 ++ >>> s390x/Makefile | 7 ++ >>> s390x/cmm.c | 69 +++++++++++++++++++ >>> s390x/diag10.c | 81 ++++++++++++++++++++++ >>> s390x/gs.c | 170 +++++++++++++++++++++++++++++++++++++++++++++++ >>> s390x/pfmf.c | 144 +++++++++++++++++++++++++++++++++++++++ >>> s390x/skey.c | 94 ++++++++++++++++++++++++++ >>> s390x/sthyi.c | 169 ++++++++++++++++++++++++++++++++++++++++++++++ >>> s390x/sthyi.h | 131 ++++++++++++++++++++++++++++++++++++ >>> s390x/unittests.cfg | 21 ++++++ >>> s390x/vector.c | 134 +++++++++++++++++++++++++++++++++++++ >>> 13 files changed, 1107 insertions(+) >>> create mode 100644 lib/s390x/asm/mem.h >>> create mode 100644 s390x/cmm.c >>> create mode 100644 s390x/diag10.c >>> create mode 100644 s390x/gs.c >>> create mode 100644 s390x/pfmf.c >>> create mode 100644 s390x/skey.c >>> create mode 100644 s390x/sthyi.c >>> create mode 100644 s390x/sthyi.h >>> create mode 100644 s390x/vector.c >>> >> >> In general, looks good to me. But I somewhat dislike creating so many >> new test. >> >> Can we at least merge pfmf+skey and maybe move sthyi+diag10+(cmm) to >> intercept.c? > > FWIW, I personally prefer to have simple testcases that only do one thing as it simplifies > the process to reproducing the bug and the fixing. > I should really proof read my email before pressing the send button.... Another reason to keep things separate is the following: by just looking at the size of cmm.c (69 lines) it becomes obvious that a future patch can and should extend this testcase. If you now cramp everything together you can no longer detect simple testcases that want to be extended.
