"Hacker" Tomu FPGA (Fomu) v0.0 support #34
Conversation
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
icecube doesn't care about init values, but yosys does and you can't satisfy them with HW RAM module. So here we remove all the init values and we make sure the reads are not dependent on the reset line Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
…atches If you don't assign all 'reg's in a process, this effectively describes a latch, and the HW doesn't have any HW latches which leads yosys to create a logic loop, which is definitely not good in FPGA ! Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
icecube doesn't care about init values, but yosys does and you can't satisfy them with HW RAM module. So here we remove all the init values and we make sure the reads are not dependent on the reset line Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
…atches If you don't assign all 'reg's in a process, this effectively describes a latch, and the HW doesn't have any HW latches which leads yosys to create a logic loop, which is definitely not good in FPGA ! Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
|
Hey @osresearch |
|
@osresearch - I've merged a bunch of the dependent pull requests. I think it would be a good idea to merge #40 and this pull request in the next couple of days. Could you update me on what you think the state is? |
mithro
changed the title
|
@osresearch - I would love to merge this support. Are you going to have any time to clean it up or should I try and find time to do it? |
|
This needs to be re-tested with a more recent checkout, although I don't have my programming jig handy. I can test with the existing bootloader and write into the user application space, which will hopefully be sufficient to validate. |
|
I've merged this with master and it seems to work fine on my one Fomu to test with. The 48 MHz clock "passes" timing at 38 MHz, which is unchanged from before. |
Successfully flashed the Fomu v0.0 "Hacker" board (ice40up5k-uwg30) with this build (using
fomu-flash) and was able to upload a user program withtinyprog -p.arachne-pnrdoesn't produce a working bitsream for me andnextpnr-ice40seems very unstable as to whether or not things will work. The/dev/ttyACM0enumeration and programming works with the seed 12345678, even thoughicetimereports that it is only 19.69 MHz.