* `misc/__init__.py`:
* checks if there's a `yosys-abc` in the same directory - if yes, sets the variable `sys._pyosys_abc`
* checks if there's a `share` in the same directory - if yes, sets the variable `sys._pyosys_share_dirname`
* `yosys.cc::init_share_dirname`: check for `sys._pyosys_share_dirname`, use it at the highest priority if Python is enabled
* `yosys.cc::init_abc_executable_name`: check for `sys._pyosys_abc`, use it at at the highest priority if Python is enabled
* `Makefile`: add new target, `share`, to only create the extra targets
* `setup.py`: compile libyosys.so, yosys-abc and share, and copy them all as part of the pyosys build
* `test/arch/ecp5/add_sub.py`: ported `add_sub.ys` to Python to act as a test for the share directory and abc with Python wheels, used in CI
* Created `setup.py`: Python package manifest to build `pyosys` wheels with a custom extension to build and include `libyosys.so` using Make
* `.gitignore`: Added byproducts of the Python wheel build process
* `Makefile`: Added `-undefined dynamic_lookup` to `libyosys.so` so missing symbols can be resolved by importing into a Python interpreter
* `kernel/yosys.cc`: Gated `PyImport_AppendInittab` with `!Py_IsInitialized`; as of Python 3.12, the interpreter is already initialized and `PyImport_AppendInittab` would cause an exception to be raised
* Created `wheels.yml`: CI workflow for building wheels for CPython on:
* Linux (glibc, musl) and Darwin
* x86-64 and arm64
* Less brittle method of adding script dirname to sys.path
* Check if scriptfp successfully opens before using it
* Move `log_error` to after `PyErr_Print()` is called
This matches the behavior of running a Python interpreter, where the
first element of sys.path is the dirname of the script being run.
This allows importing of files and modules in the same directory without
messing with PYTHONPATH or similar.
This adds a Python equivalent to the `-c` option, where scripts importing `libyosys` can be imported and used.
Most of the work for this was already done to enable Python passes a couple years back, so this is a relatively small changeset.
It adds `DriveBit`, `DriveChunk` and `DriveSpec` types which are similar
to `SigBit`, `SigChunk` and `SigSpec` but can also directly represent
cell ports, undriven bits and multiple drivers. For indexing an RTLIL
module and for querying signal drivers it comes with a `DriverMap` type
which is somewhat similar to a `SigMap` but is guaranteed to produce
signal drivers as returned representatives.
A `DriverMap` can also optionally preserve connections via intermediate
wires (e.g. querying the driver of a cell input port will return a
connected intermediate wire, querying the driver of that wire will
return the cell output port that's driving the wire).
This is already supported by `SigSpec` and since both `SigChunk` and
`SigSpec` implement `extract` which is the multi-bit variant of this,
there is no good reason for `SigChunk` to not support
`SigBit operator[](int offset)`.
When implementing custom log_... functions or custom overloads for the
core log functions like log_signal it is necessary to return `char *`
that are valid long enough.
The log_... functions implemented in log.cc use either `log_id_cache` or
`string_buf` which both are cleared on log_pop.
This commit adds a public `log_str` function which stores its argument
in the `log_id_cache` and returns the stored copy, such that custom
log functions outside of log.cc can also create strings that remain
valid until the next `log_pop`.
This adds a generic non-recursive implementation of Tarjan's linear time
SCC algorithm that produces components in topological order. It can be
instantiated to work directly on any graph representation for which the
enumerate_nodes and enumerate_successors interface can be implemented.
Only the `B` input (the shift amount) can be marked as signed on a
`$shiftx` cell. Adapt the helper accordingly and prevent it from
creating invalid RTLIL when called with `is_signed` set. Previously
it would mark both `A` and `B` as signed.
Previously `extract` on a `SigSpec` would always unpack it. Since a
significant amount of `SigSpec`s have one or few chunks, it's worth
having a dedicated implementation.
This is especially true, since the RTLIL frontend calls into this for
every `wire [lhs:rhs]` slice, making this `extract` take up 40% when
profiling `read_rtlil` with one of the largest coarse grained RTLIL
designs I had on hand.
With this change the `read_rtlil` profile looks like I would expect it
to look like, but I noticed that a lot of the other core RTLIL methods
also are a bit too eager with unpacking or implementing
`SigChunk`/`Const` overloads that just convert to a single chunk
`SigSpec` and forward to the implementation for that, when a direct
implementation would avoid temporary std::vector allocations. While not
relevant for `read_rtlil`, to me it looks like there might be a few easy
overall performance gains to be had by addressing this more generally.
This adjusts the way the headers kernel/{yosys,rtlil,register,log}.h
include each other to avoid the need of including headers outside of
include guards as well as avoiding the inclusion of rtlil.h in the
middle of yosys.h with rtlil.h depending on the prefix of yosys.h, and
the suffix of yosys.h depending on rtlil.h.
To do this I moved some of the declaration in yosys.h into a new header
yosys_common.h. I'm not sure if that is strictly necessary.
Including any of these files still results in the declarations of all
these headers being included, so this shouldn't be a breaking change for
any passes or external plugins.
My main motivation for this is that ccls's (clang based language server)
include guard handling gets confused by the previous way the includes
were done. It often ends up treating the include guard as a generic
disabled preprocessor conditional, breaking navigation and highlighting
for the core RTLIL data structures.
Additionally I think avoiding cyclic includes in the middle of header
files that depend on includes being outside of include guards will also
be less confusing for developers reading the code, not only for tools
like ccls.
Before this commit, the combination of `_` and `0` format characters
would produce a result like `000000001010_1010`.
After this commit, it would be `0000_0000_1010_1010`.
This has a slight quirk where a format like `{:020_b}` results in
the output `0_0000_0000_1010_1010`, which is one character longer than
requested. Python has the same behavior, and it's not clear what would
be strictly speaking correct, so Python behavior is implemented.
Emil J. Tywoniak
Philippe Sauter
Mohamed Gaber
Martin Povišer
Miodrag Milanović
Mike Inouye
Roland Coeurjoly
Claire Xenia Wolf
N. Engelhardt
Emily Schmidt
Jannis Harder
Krystine Sherwin
Emil J
Alexander von Gluck
George Rennie
Catherine