Open seanlatias opened 4 years ago
We already know that the segfault happens in the Simplify
pass, but the cause is still unknown.
The data type of A
also doesn't seem to matter.
Just to confirm -- [0:7] means we are extracting the first 8 bits, right? Last Wed, I saw Xiangyi/Alga use [0:8] and [8:16]
No, we will be extracting the first 7 bits, which is the same as Python's convention. For Python, the range is upper-bound exclusive.
Can you take a look at the bit slicing support in other HW-oriented Python DSLs, like PyMTL and PyRTL? See if they are following this convention.
The following code leads to a segfault
The segfault only happens when the bit slices are the exact numbers as shown in the above code. Other numbers don't seem to create a segfault.