sv2v converts SystemVerilog (IEEE 1800-2017) to Verilog (IEEE 1364-2005), with an emphasis on supporting synthesizable language constructs.
The primary goal of this project is to create a completely free and open-source tool for converting SystemVerilog to Verilog. While methods for performing this conversion already exist, they generally either rely on commercial tools, or are limited in scope.
This project was originally developed to target Yosys, and so allows for disabling the conversion of (passing through) those SystemVerilog features that Yosys supports.
The idea for this project was shared with me while I was an undergraduate at Carnegie Mellon University as part of a joint Computer Science and Electrical and Computer Engineering research project on open hardware under Professors Ken Mai and Dave Eckhardt. I have greatly enjoyed collaborating with the team at CMU since January 2019, even after my graduation the following May.
All of sv2v's dependencies are free and open-source.
- Build Dependencies
- Haskell Stack - Haskell build system
- Haskell dependencies are managed in
sv2v.cabal
- Test Dependencies
- Icarus Verilog - for Verilog simulation
- shUnit2 - test framework
- Python 3.x - for evaluating certain test cases
Binaries for Ubuntu, macOS, and Windows are available on the releases page. If your system is not covered, or you would like to build the latest commit, simple instructions for building from source are below.
You must have Stack installed to build sv2v. Then you can:
git clone https://github.com/zachjs/sv2v.git
cd sv2v
make
This creates the executable at ./bin/sv2v
. Stack takes care of installing
exact (compatible) versions of the compiler and sv2v's build dependencies.
You can install the binary to your local bin path (typically ~/.local/bin
) by
running stack install
, or copy over the executable manually.
sv2v takes in a list of files and prints the converted Verilog to stdout
by
default. Users should typically pass all of their SystemVerilog source files to
sv2v at once so it can properly resolve packages, interfaces, type parameters,
etc., across files. Using --write=adjacent
will create a converted .v
for
every .sv
input file rather than printing to stdout
. --write
/-w
can also
be used to specify a path to a .v
output file. Undefined modules and
interfaces can be automatically loaded from library directories using
--libdir
/-y
.
Users may specify include
search paths, define macros during preprocessing,
and exclude some of the conversions. Specifying -
as an input file will read
from stdin
.
Below is the current usage printout.
sv2v [OPTIONS] [FILES]
Preprocessing:
-I --incdir=DIR Add a directory to the include search path
-y --libdir=DIR Add a directory to the library search path used
when looking for undefined modules and interfaces
-D --define=NAME[=VALUE] Define a macro for preprocessing
--siloed Lex input files separately, so macros from
earlier files are not defined in later files
--skip-preprocessor Disable preprocessing of macros, comments, etc.
Conversion:
--pass-through Dump input without converting
-E --exclude=CONV Exclude a particular conversion (Always, Assert,
Interface, Logic, SeverityTask, or UnbasedUnsized)
-v --verbose Retain certain conversion artifacts
-w --write=MODE/FILE/DIR How to write output; default is 'stdout'; use
'adjacent' to create a .v file next to each input;
use a path ending in .v to write to a file; use a
path to an existing directory to create a .v within
for each converted module
--top=NAME Remove uninstantiated modules except the given
top module; can be used multiple times
Other:
--oversized-numbers Disable standard-imposed 32-bit limit on unsized
number literals (e.g., 'h1_ffff_ffff, 4294967296)
--dump-prefix=PATH Create intermediate output files with the given
path prefix; used for internal debugging
--bugpoint=SUBSTR Reduce the input by pruning modules, wires, etc.,
that aren't needed to produce the given output or
error substring when converted
--help Display this help message
--version Print version information
--numeric-version Print just the version number
sv2v supports most synthesizable SystemVerilog features. Current notable
exceptions include defparam
on interface instances, certain synthesizable
usages of parameterized classes, and the bind
keyword. Assertions are also
supported, but are simply dropped during conversion.
If you find a bug or have a feature request, please create an issue. Preference will be given to issues that include examples or test cases.
This project contains a preprocessor, lexer, and parser, and an abstract syntax tree representation for a subset of the SystemVerilog specification. The parser is not very strict. The AST allows for the representation of syntactically (and semantically) invalid Verilog. The goal is to be more general in the representation to enable more standardized and straightforward conversion procedures. This could be extended into an independent and more fully-featured front end if there is significant interest.
Once the test dependencies are installed, tests can be run with make test
.
GitHub Actions is used to automatically test commits. Please review the test
documentation for guidance on adding, debugging, and interpreting tests.
There is also a SystemVerilog compliance suite that tests open-source tools' SystemVerilog support. Although not every test in the suite is applicable, it has been a valuable asset in finding edge cases.
This project was originally forked from Tom Hawkin's Verilog parser. While the front end has changed substantially to support the larger SystemVerilog standard, his project was a great starting point.
Reid Long was invaluable in developing this tool, providing significant tests and advice, and isolating many bugs.
Edric Kusuma helped me with the ins and outs of SystemVerilog, with which I had no prior experience, and has also helped with test cases.
Since sv2v's public release, many people have taken the time to file detailed bug reports and feature requests. I greatly appreciate their help in furthering the project.