v4: https://lists.gnu.org/archive/html/qemu-devel/2023-05/msg01406.html
v4 -> v5: Use pre-decrement on BE to have nicer values in memory (Alex).
          Fix the comment saying that only LE is supported (Alex).

v3: https://lists.gnu.org/archive/html/qemu-devel/2023-04/msg04756.html
v3 -> v4: Trivial rebase on top of Alex's series.

v2: https://lists.gnu.org/archive/html/qemu-devel/2023-04/msg04154.html
v2 -> v3: The idea with sharing the QEMU headers with the tests seems
          to be controversial. Just rework the test to work without
          the explicit byte swaps.

v1: https://lists.gnu.org/archive/html/qemu-devel/2023-04/msg03746.html
v1 -> v2: Use cpu_to_le16() and friends (Thomas).

Hi,

I noticed that Alex added "undefine MULTIARCH_TESTS" to
tests/tcg/s390x/Makefile.softmmu-target in the plugin patch, and
thought that it may better to just enable them, which this series
does.

Patch 1 fixes an endianness issue in the memory test.

Patch 2 enables the multiarch system test. The main difficulty is
outputting characters via SCLP, which is sidestepped by reusing the
pc-bios/s390-ccw implementation.

Best regards,
Ilya

Ilya Leoshkevich (2):
  tests/tcg/multiarch: Make the system memory test work on big-endian
  tests/tcg/s390x: Enable the multiarch system tests

 tests/tcg/multiarch/system/memory.c     | 67 ++++++++++++++++---------
 tests/tcg/s390x/Makefile.softmmu-target | 40 +++++++++------
 tests/tcg/s390x/console.c               | 12 +++++
 tests/tcg/s390x/head64.S                | 31 ++++++++++++
 4 files changed, 112 insertions(+), 38 deletions(-)
 create mode 100644 tests/tcg/s390x/console.c
 create mode 100644 tests/tcg/s390x/head64.S

-- 
2.40.1