On Wed, Jan 06, 2021 at 11:05:05 +0100, Tim Wiederhake wrote:
> Signed-off-by: Tim Wiederhake <twiederh@xxxxxxxxxx>
> ---
> ...64-cpuid-Atom-P5362-processor-disabled.xml | 8 +
> ..._64-cpuid-Atom-P5362-processor-enabled.xml | 10 +
> .../x86_64-cpuid-Atom-P5362-processor.json | 2415 +++++++++++++++++
> .../x86_64-cpuid-Atom-P5362-processor.xml | 61 +
We should remove the "-processor" part of the file names as it's
redundant. I'll send a patch for cpu-gather.py script to do that
automatically next time we parse data from such CPU.
> 4 files changed, 2494 insertions(+)
> create mode 100644 tests/cputestdata/x86_64-cpuid-Atom-P5362-processor-disabled.xml
> create mode 100644 tests/cputestdata/x86_64-cpuid-Atom-P5362-processor-enabled.xml
> create mode 100644 tests/cputestdata/x86_64-cpuid-Atom-P5362-processor.json
> create mode 100644 tests/cputestdata/x86_64-cpuid-Atom-P5362-processor.xml
The new test files are not used anywhere. The cputest.c file should be
updated to start testing this CPU:
DO_TEST_CPUID(VIR_ARCH_X86_64, "Atom-P5362", JSON_MODELS_REQUIRED);
and once done, you can run
VIR_TEST_REGENERATE_OUTPUT=1 tests/cputest
to generate the test results. And of course they should be manually
checked whether they make sense. This will show the CPU detection
completely fails on this CPU as it is detected as either IvyBridge or
Westmere three different CPU models depending on where the CPU model is
supposed to be used (capabilities vs. domain capabilities).
Jirka
