| Age | Commit message (Collapse) | Author |
|---|
|
- Refactor the flow typing implementation in the type-checker. This
should fix an issue involving riscv_platform. Specifically it should
now work better when an if statement contains multiple conditions
combined with and/or, only some of which imply constraints at the
type level. This change also simplifies the implementation of flow
typing, and removes some obscure features that were hardly used -
specifically, flow typing could modify types, but this was fairly
obscure and doesn't seem to affect any of our specifications. More
testing is needed to ensure that this change hasn't inadvertantly
broken anything, but it does pass all our tests and continue to
typecheck arm, riscv and cheri.
- Also adds a option for generating faster undefined functions for
enum and variant types. Previously I tried to optimise away such
functions in the C backend, because they could be slow and cause
considerable uneccessary allocation, however this was error prone
and it turns out a much simpler solution is to simply make the
functions themselves much faster, at the cost of hard-coding certain
decisions about what undefined means for these types at compile tile
(which is fine for fast emulation). This almost doubles the
performance of the generated C code.
- Add a wrapper for right shift to avoid UB when shifting by 64 or
more places.
|
|
|
|
c_backend. still need to merge changes to prelude.sail.
|
|
solves a problem where the Hashtbl module was encountering a stack overflow booting CheriBSD. There seems to be little or no performance impact.
|
|
|
|
|
|
|
|
|
|
|
|
arichardson-patch-1
|
|
Looks like Jenkins is still on OCaml 4.02.3. We should probably upgrade to 4.05 at some point.
|
|
Also fixes to C backend for compiling MIPS spec to C
- Fix an issue with const correctness in internal_vector_update
functions generated by C backend
- Add builtins for MIPS to sail.h
- Fix an issue where reg_deref didn't work when called on pointers to
large bitvectors, i.e. vectors containing references to large
bitfields as in the MIPS TLB code
- Various bug fixes and changes for running U-boot on ARM model,
including for interpreter and OCaml compilation.
- Fix memory leak issues and incorrect shadowing for foreach loops
- Update C header file. Fixes memory leak in memory read/write
builtins.
- Add aux constructor to ANF representation to hold environment
information.
- Fix undefined behavior caused by optimisation left shifting uint64_t
vectors 64 or more times. Unfortunately there's more issues because
the same happens for X >> 64 right shifts. It would make sense for
this to be zero, because that would guarantee the property that ((X
>> n) >> m) == (X >> (n + m)) but we probably need to do (X >> (n -
1) >> 1) in the optimisation to ensure that we don't cause
UB. Shifting by 63 and then by 1 is well-defined, but shifting by 64
in one go isn't according to the C standard. This issue with
right-shifts only occurs for zero-length vectors, so it's not a huge
deal, but it's still annoying.
- Add versions of print_bits and print_int that print to
stderr. Follows OCaml convention of print/prerr. Should make things
more explicit. Different backends had different ideas about where
print should output to, not every backend needs to have this
(e.g. theorem prover backends don't need to print) but having both
stderr and stdout seperate and clear is useful for executable models
(UART needs to be stdout, debug messages should be stderr).
|
|
|
|
|
|
|
|
Dump format is the same as for the cap GPRs with DEBUG CAP HWREG as the prefix
|
|
|
|
Some tweaks to installation instructions for latest OCaml versions (4.05 and 4.06)
|
|
(Similar to the proper translation for function definitions)
|
|
|
|
Otherwise it has occasional problems working out the return type
|
|
Useful for partial test cases (e.g., some of the typechecking tests)
Also a bonus warning for such functions in normal use
|
|
|
|
vector_access is a bit hacky at the moment - it expects a constraint to
be shown between the index and the list size, but we don't track list
sizes in general
|
|
(<5% on a simple test) but dramatically reduces memory usage compared to having a hash table entry per byte!
|
|
dtb to be mapped.
|
|
declaration"
This reverts commit 895f868cd537277ba61dfc427fee0e288af7e226.
These are actually treated as Ints (although you could pretend they
weren't and it mostly worked).
|
|
Stops (e.g.) an Int being used as a Type, including when no kind was
declared. The following commit will remove the test for the latter
case.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
platform ocaml files into _sbuild directory generated by sail and then running ocamlbuild there.
|
|
|
|
platform build.
|
|
|
|
|
|
|
|
|
|
|
|
build until we add suitable platform definitions/stubs.
The platform bits are not yet hooked into the model, but only into the build, so are untested.
|
|
default (off by default).
|
|
|
|
|
|
Worked around problem with the model where it tries to use bound variables
in patterns
|
|
zero not only when rt is the zero register (zr).
|
