qemu/tests/decode, branch master

tests/: spelling fixes

2023-09-08T10:08:52Z

with some rewording in tests/qemu-iotests/298 tests/qtest/fuzz/generic_fuzz.c tests/unit/test-throttle.c as suggested by Eric. Signed-off-by: Michael Tokarev Reviewed-by: Eric Blake

decodetree: Add --output-null for meson testing

2023-06-01T02:56:42Z

Using "-o /dev/null" fails on Windows. Rather that working around this in meson, add a separate command-line option so that we can use python's os.devnull. Reported-by: Thomas Huth Fixes: 656666dc7d1b ("tests/decode: Convert tests to meson") Signed-off-by: Richard Henderson Reviewed-by: Thomas Huth Message-Id: <20230531232510.66985-1-richard.henderson@linaro.org>

tests/decode: Add tests for various named-field cases

2023-05-30T17:55:39Z

Add some tests for various cases of named-field use, both ones that should work and ones that should be diagnosed as errors. Signed-off-by: Peter Maydell Reviewed-by: Richard Henderson Message-Id: <20230523120447.728365-7-peter.maydell@linaro.org>

tests/decode: Convert tests to meson

2023-05-30T17:55:39Z

Signed-off-by: Richard Henderson

decodetree: Extend argument set syntax to allow types

2021-05-01T18:45:35Z

Rather than force all structure members to be 'int', allow the type of the member to be specified. Reviewed-by: Luis Pires Signed-off-by: Richard Henderson

decodetree: Improve identifier matching

2020-09-14T13:23:19Z

Only argument set members have to be C identifiers, everything else gets prefixed during conversion to C. Some places just checked the leading character, and some places matched a leading character plus a C identifier. Convert everything to match full identifiers, including the [&%@&] prefix, and drop the full C identifier requirement. Reported-by: Peter Maydell Signed-off-by: Richard Henderson Reviewed-by: Philippe Mathieu-Daudé Message-id: 20200903192334.1603773-1-richard.henderson@linaro.org Signed-off-by: Peter Maydell

decodetree: Drop check for less than 2 patterns in a group

2020-06-09T16:19:48Z

While it makes little sense for the end product to have a group containing only a single pattern, avoiding this case within an incremental patch set is troublesome. Because this is expected to be a transient condition, do not bother "optimizing" this case, e.g. by folding away the group. Reviewed-by: Peter Maydell Signed-off-by: Richard Henderson

tests/decode: Test non-overlapping groups

2020-06-08T18:09:48Z

Reviewed-by: Peter Maydell Signed-off-by: Richard Henderson

decodetree: Allow group covering the entire insn space

2020-06-08T17:36:47Z

This is an edge case for sure, but the logic that disallowed this case was faulty. Further, a few fixes scattered about can allow this to work. Reviewed-by: Peter Maydell Signed-off-by: Richard Henderson

decodetree: Allow !function with no input bits

2019-08-19T15:03:41Z

Call this form a "parameter", returning a value extracted from the DisasContext. Reviewed-by: Philippe Mathieu-Daude Signed-off-by: Richard Henderson