summaryrefslogtreecommitdiff
path: root/src/main/scala/Chisel/Core.scala
AgeCommit message (Collapse)Author
2015-10-19Core.scala API documentationducky
2015-10-16Remove old Literal object; fold the parseLit into UIntAndrew Waterman
2015-10-08Handle IllegalArgumentExcetions that mean a cloneType is neededPalmer Dabbelt
When trying to cloneType() on classes that take a builtin (like Int), we get an IllegalArgumentException instead of a InvocationTargetException. This change prints a nice error message instead of a stack trace.
2015-10-01Flip Mem/SeqMem argument order; deprecate old orderingAndrew Waterman
2015-09-25Improve coverage of default cloneType method; add commentsAndrew Waterman
2015-09-24Add new-style Vec constructor; deprecate old versionducky
2015-09-18Improve IR class hierarchyAndrew Waterman
- Rename Alias to Node to match FIRRTL notion - Remove poorly-named Immediate and replace root of hierarchy with Arg
2015-09-18Correct some scaladoc commentsAndrew Waterman
2015-09-18Use FIRRTL idiom for SeqMem read-enablesAndrew Waterman
Emit read-enables as mux(ren, addr, poison).
2015-09-01Add scaladoc to UInt/SInt companion objectsAndrew Waterman
2015-09-01Disallow external use of Vec/UInt/SInt constructorsAndrew Waterman
Use the companion objects instead.
2015-08-31Fix val io = new Bundle{...}.flipAndrew Waterman
Now, we emit all I/Os inside a bundle named io.
2015-08-28Use FIRRTL smem for SeqMemAndrew Waterman
Read enables and read-write ports aren't working yet.
2015-08-28Add poison nodeAndrew Waterman
2015-08-27Redefine masked Mem writes for Mem[Vec]Andrew Waterman
2015-08-27Fix bug where flipping top-level I/O had no effectAndrew Waterman
The fix is to propagate the flip to the fields in the bundle.
2015-08-27Vec.apply is for types; Vec.fill is for rvaluesAndrew Waterman
2015-08-27Add chisel2 scaladoc for 'when'.Jim Lawson
2015-08-26Remove Mem from Data hierarchyAndrew Waterman
Just like Reg, state elements are not Data.
2015-08-26Simplify Module internal data structuresAndrew Waterman
2015-08-26Simplify I/O zero-initializationAndrew Waterman
2015-08-26import relevant scaladoc from chisel(2).Jim Lawson
2015-08-20Prevent some defs from being marked as Bundle fieldsAndrew Waterman
2015-08-20Remove Port/Kind IR nodes, which merely wrap DataAndrew Waterman
2015-08-13Make error reporting reentrantAndrew Waterman
2015-08-13Make temporary names locally unique, rather than globally soAndrew Waterman
2015-08-13Add back missing () on toBits declarationAndrew Waterman
2015-08-13Tighten permissions on some classes & membersAndrew Waterman
2015-08-13fun with ##Henry Cook
2015-08-13Clean up UInt/SInt/Bool companion objectsAndrew Waterman
2015-08-13Don't fold constants in the frontendAndrew Waterman
We need to make a similar change for extract, pending a FIRRTL bug fix.
2015-08-13Check validity of bit extract rangesAndrew Waterman
2015-08-13Avoid importing for single useAndrew Waterman
2015-08-13FP stuff doesn't belong in DataAndrew Waterman
2015-08-13Cleanup DynamicContextHenry Cook
2015-08-13re-privatize class Namespace, fix use of Module/Bundle child namespacesHenry Cook
2015-08-13clean up Id and Builder.globalRefMapHenry Cook
2015-08-13refactor NamespaceHenry Cook
2015-08-13Streamline files, breaking up Core.scala and resorting some smaller onesHenry Cook
2015-08-12Emitter no longer mutates the refMapAndrew Waterman
2015-08-12params and paramsScope objectsHenry Cook
2015-08-12Marshal the global mutable state into one objectAndrew Waterman
Hopefully, the Chisel core is now thread-safe.
2015-08-11Miscellaneous cleanupsAndrew Waterman
2015-08-11Emit newline at EOFAndrew Waterman
2015-08-11Remove useless call to getWidthAndrew Waterman
2015-08-10Fix Mux type safetyAndrew Waterman
The implementation is a total kludge, but at least it's not broken.
2015-08-10Make Bits.toBool safeAndrew Waterman
It now fails if the width is unknown or is not equal to 1. We could consider relaxing this later, defining it as this.orR.
2015-08-10Force toUInt = asUInt, toSInt = asSIntAndrew Waterman
2015-08-10Don't use cloneType for primopsAndrew Waterman
Doing so results in incorrect code for Bools, because the widths of some Bool primops' results are greater than 1. The alternative would be to make Bool not extend UInt.
2015-08-10Remove redundant castAndrew Waterman