diff options
| author | Maciej Dudek <mdudek@antmicro.com> | 2021-09-23 13:58:22 +0200 |
|---|---|---|
| committer | Maciej Dudek <mdudek@antmicro.com> | 2021-09-23 13:58:22 +0200 |
| commit | 8c97cbe3415b2292263752a0ea5c64ce73d39e51 (patch) | |
| tree | a52d4d000e645d147949a3b7ac03d76ba2a781be /json | |
| parent | b2e9ce46f1cf60eff5d6d0fe844fc056890dfe6a (diff) | |
| download | nextpnr-8c97cbe3415b2292263752a0ea5c64ce73d39e51.tar.gz nextpnr-8c97cbe3415b2292263752a0ea5c64ce73d39e51.tar.bz2 nextpnr-8c97cbe3415b2292263752a0ea5c64ce73d39e51.zip | |
Fix chain swap
Issue was due to dest_bels being not cleared between clusters unbindes, causing
newly bind bels to be unbinded and having their old bel value changed to new bel value.
Then when swap failed 2 cells were being bind to a single bel.
I tested leaving dest_bels in the function scope and moving it to the loop scope.
Code with dest_bels in the loop scope was faster than leaving it in the function scope,
and checking if the cell is in the processed cluster.
Signed-off-by: Maciej Dudek <mdudek@antmicro.com>
Diffstat (limited to 'json')
0 files changed, 0 insertions, 0 deletions