| Age | Commit message (Collapse) | Author |
|---|
|
This change allows the RISC-V spec to compile to C, but more testing
is needed to ensure it works correctly.
|
|
Make the C l-expression type in Sail more generic and expressive, and
refactor the generation of conversions into a seperate
codegen_conversion function, which can handle more complex cases than
the previous more ad-hoc method.
|
|
Test that basic bi-directional mappings compile correctly
Test that a minimal file importing the prelude compiles correctly
|
|
Ensure that this works even when the union types are dependent in the wrong order, before topologically sorting definitions. We do this by calling fix_variant_ctyps on all cdefs by passing a list of
prior cdefs to specialize_variants.
|
|
|
|
Test the builtin functions by compiling them to C, OCaml, and OCaml
via Lem. Split up some of the longer builtin test programs to avoid
stack overflows when compiling to OCaml, as 3000+ line long blocks can
cause issues with some re-writing steps.
Also test constant-folding with builtins (this should reduce the
asserts in these files to assert true), and also test constant folding
with the C compilation.
Fix a bug whereby vectors with heap-allocated elements were not
initialized correctly.
Fix a bug caused by compiling and optimising empty vector literals.
Fix an OCaml test case that broke due to the ref type being used. Now
uses references to registers.
Fix a bug where Sail would output big integers that lem can't
parse. Checks if integer is between Int32.min_int and Int32.max_int
and if not, use integerOfString to represent the integer. Really this
should be fixed in Lem.
Make the python test runner script the default for testing builtins
and running the C compilation tests in test/run_tests.sh
Add a ocaml_build_dir option that sets a custom build directory for
OCaml. This is needed for running OCaml tests in parallel so the
builds don't clobber one another.
|
|
Add new python test runner script, which allows tests to be run in
parallel before collecting the results. This makes the tests run a lot
faster, especially for the builtins and C compilation tests. Also
handles reporting errors mushc more nicely than the previous way of
doing it in shell script.
|
|
|
|
|
|
We now generate anonymous types in the correct order, but post specialisation more dependencies can occur between named types, so an additional
sorting step is needed to ensure that these happen in the correct order. In theory we could end up with circular dependencies here that don't exist
at the Sail source level, but this shouldn't occur often (or ever) in practice. I think this is fixable but it would require some code generator
changes.
|
|
lifted types
Add a test case for nested variant constructors
|
|
|
|
Especially for return expressions, which fixes a test case
|
|
|
|
|
|
- Fix ambiguities in parser.mly
- Ensure that no new identifiers are bound in or-patterns and
not-patterns, by adding a no_bindings switch to the
environment. These patterns shouldn't generate any bogus flow typing
constraints because we just pass through the original environment
without adding any possible constraints (although this does mean we
don't get any flow typing from negated numeric literals right now,
which is a TODO).
- Reformat some code to match surrounding code.
- Add a typechecking test case for not patterns
- Add a typechecking test case for or patterns
At least at the front end everything should work now, but we need to
do a little bit more to rewrite these patterns away for lem etc.
|
|
|
|
|
|
Fixes monomorphisation on files using mappings.
Also extended constant propagation to handle pattern matches on
bitvector expressions (because an earlier rewrite replaces the literals).
Also moved L_undef rewriting because monomorphisation can handle them
but not the replacement functions.
|
|
|
|
Uses previous stage to deal with (e.g.) guards.
New option -dcoq_warn_nonex tells you where all of the extra default
cases were added.
|
|
|
|
|
|
|
|
Also use zero-initialised memory. Apparently some tests access unitialised
memory, and the default behaviour of the Lem shallow embedding is to fail in
this case.
|
|
|
|
Use nondeterministic choice by default instead of a deterministic bitstream
generator in the state, which is slightly awkward to reason about, because
every use of undefined_boolS changes the state. The previous behaviour can be
implemented as Sail code, if desired.
Also add a default implementation of internal_pick that nondeterministically
chooses an element from a list.
|
|
|
|
|
|
Add a test case in test/c/eq_struct.sail. Ensure that the macro EQUAL(type) will always give a valid equality function for any
builtin type in sail.h.
|
|
Non bitvector literals for decreasing vectors were not being reversed
correctly, so the list of capability registers was effectively in
reverse order.
Added a test case to test/c/ based on this aspect of CHERI
|
|
We should test before the first iteration in case 'to' starts out as
less than 'from'.
|
|
vector_update_subrange wasn't setting its return length correctly
|
|
|
|
|
|
Necessary to prevent redundant clauses that Coq will reject
(There's still a problem if you use a variable rather than a wildcard,
see the test)
|
|
|
|
|
|
|
|
Previously the ANF->IR translation cared too much about how things
were allocated in C, so it had to constantly check whether things
needed to be allocated on the stack or heap, and generate different
cequences of IR instructions depending on either. This change removes
the ialloc IR instruction, and changes iinit and idecl so that the
code generator now generates different C for the same IR instructions
based on the variable types involved.
The next change in this vein would be to merge icopy and iconvert at
the IR level so that conversions between uint64_t and large-bitvectors
are inserted by the code generator. This would be good because it
would make the ANF->IR translation more robust to changes in the types
of variables caused by flow-typing, and optimization passes could
convert large bitvectors to uint64_t as local changes.
|
|
(still need to sort out some string stuff, though)
|
|
Fix a bug involving indentifers on the left hand side of assignment
statements not being shadowed correctly within foreach loops.
Make the different between different types of integer division
explicit in at least the C compilation for now. fdiv_int is division
rounding towards -infinity (floor). while tdiv_int is truncating
towards zero. Same for fmod_int and tmod_int.
|
|
|
|
|
|
Plus
- Complete solver support for inequalities
- Reduce exponentials in solver
|
|
It can be proved almost entirely by symbolic execution (in <15s) _if_
the right definitions are in the compset. It took a lot of interactive
stumbling about to discover that LUPDATE was missing from the standard
list compset.
|
|
The proof now gets through simulation of the first instruction of the test.
|
|
|
|
|
|
|
