Get info about packages/pins from the IDE files. #57
Conversation
Signed-off-by: YRabbit <rabbit@yrabbit.cyou>
|
Well, we'll see how the CI goes. Should break on examples because nextpnr is old, but it will create bases. |
apycula/tiled_fuzzer.py
Outdated
| "device": "GW1N-4-PBGA256-6", | ||
| "partnumber": "GW1N-LV4PG256C6/I5", | ||
| }, | ||
| "GW1N-1": { | ||
| "package": "LQ144", | ||
| "package": "LQFP144", | ||
| "header": 1, # stupid note in excel |
apycula/chipdb.py
Outdated
| return (pkgs, res, res_bank_pins) | ||
|
|
||
| # returns ({partnumber: (package, device)}, {pins}, {bank_pins}) | ||
| def xls_pinout(device): |
There was a problem hiding this comment.
maybe this shouldn't be called xls when it's not excel based ;)
|
Exactly as predicted! So what's up with the nextpnr side? I mean, isn't the chipdb format the same? |
apycula/gowin_bba.py
Outdated
| @@ -147,6 +147,17 @@ def write_timing(b, timing): | |||
| b.u32(len(timing)) | |||
| b.ref(blk) | |||
|
|
|||
| def write_partnumber_packages(b, db): | |||
| import ipdb; ipdb.set_trace() | |||
There was a problem hiding this comment.
uh breakpoints should go I think haha
There was a problem hiding this comment.
Exactly as predicted! So what's up with the nextpnr side? I mean, isn't the chipdb format the same?
The structure has changed: a table with packages has been added.
You can put the next patch on nexpnr
Signed-off-by: YRabbit <rabbit@yrabbit.cyou>
Signed-off-by: YRabbit <rabbit@yrabbit.cyou>
Signed-off-by: YRabbit <rabbit@yrabbit.cyou>
|
The speed is determined from the partnumber, the match is stored in the database so the input regular expression is simplified and allows more models:) Of course this is only friendly with the last commit for nextpnr: yrabbit/nextpnr@74b4f69 |
|
Looks good to me! |
|
Wait actually if this is an incompatible bba change it should bump the DB version |
|
Line 186 in b1dedc0 |
Signed-off-by: YRabbit <rabbit@yrabbit.cyou>
Signed-off-by: YRabbit <rabbit@yrabbit.cyou>
Signed-off-by: YRabbit rabbit@yrabbit.cyou