back.rtlil: always initialize the entire memory.
authorwhitequark <whitequark@whitequark.org>
Sat, 22 Dec 2018 05:27:42 +0000 (05:27 +0000)
committerwhitequark <whitequark@whitequark.org>
Sat, 22 Dec 2018 05:27:42 +0000 (05:27 +0000)
commit37b81309d3aacfc307f2dbd7be452608d7d9d26a
tree7f4fb2ddbf6c990c1a4f93932bbead895ac5c1be
parent99b778158d80a76f9487da6b1392818ceeaaf6e8
back.rtlil: always initialize the entire memory.

This avoids reading 'x from the memory in simulation. In general,
FPGA memories can only be initialized in block granularity, and
zero-initializing is cheap, so this is not a significant issue with
resource consumption.
nmigen/back/rtlil.py