Bump oss-cad-suite toolchain

[nathanaelhuffman]

We need some fixes in ghdl for more modern vhdl files such as ghdl/ghdl#2682

I've installed the new toolchain, we'll see how ci goes. We can walk this back as the old toolchain is still present on the build machine.

[nathanaelhuffman]

Yay! Don't you love being on the hairy edge of fitting on some of these ignition designs?

Post tool-chain bump, we get a no-fit on the following designs:
ignition_target_psc_rev_a_reset_button
ignition_target_sidecar_rev_a_reset_button

Since neither of these boards is supported and we're not using reset_button builds regardless, thoughts about just dropping these builds altogether?

The other alternatives are the continue to hold the bsv-designs back at this currently supported oss-cad-suite and pray we don't need to bump again, or attempt to make these go on a diet somehow?

[494/525] /opt/oss-cad-suite-20250211/bin/nextpnr-ice40 -q --lp1k --package qn84 --freq 50 -l env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/psc/ignition_target_psc_rev_a_reset_button.config.log --pcf ../hdl/ip/bsv/ignition/ignition_target.pcf --json env/be270e8514fe38fe936f25783690bd7a63efc533/hdl/projects/psc/ignition_target_psc_rev_a_reset_button_top.json --asc env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/psc/ignition_target_psc_rev_a_reset_button.config
FAILED: env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/psc/ignition_target_psc_rev_a_reset_button.config env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/psc/ignition_target_psc_rev_a_reset_button.config.log 
/opt/oss-cad-suite-20250211/bin/nextpnr-ice40 -q --lp1k --package qn84 --freq 50 -l env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/psc/ignition_target_psc_rev_a_reset_button.config.log --pcf ../hdl/ip/bsv/ignition/ignition_target.pcf --json env/be270e8514fe38fe936f25783690bd7a63efc533/hdl/projects/psc/ignition_target_psc_rev_a_reset_button_top.json --asc env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/psc/ignition_target_psc_rev_a_reset_button.config
ERROR: Unable to find legal placement for all cells, design is probably at utilisation limit.
0 warnings, 1 error
[495/525] /opt/oss-cad-suite-20250211/bin/nextpnr-ice40 -q --lp1k --package qn84 --freq 50 -l env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/sidecar/ignition_target/ignition_target_sidecar_rev_a_reset_button.config.log --pcf ../hdl/ip/bsv/ignition/ignition_target.pcf --json env/c9579aa6a23c50e82d40c688f8b2849c0a997c7e/hdl/projects/sidecar/ignition_target/ignition_target_rev_a_reset_button_top.json --asc env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/sidecar/ignition_target/ignition_target_sidecar_rev_a_reset_button.config
FAILED: env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/sidecar/ignition_target/ignition_target_sidecar_rev_a_reset_button.config env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/sidecar/ignition_target/ignition_target_sidecar_rev_a_reset_button.config.log 
/opt/oss-cad-suite-20250211/bin/nextpnr-ice40 -q --lp1k --package qn84 --freq 50 -l env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/sidecar/ignition_target/ignition_target_sidecar_rev_a_reset_button.config.log --pcf ../hdl/ip/bsv/ignition/ignition_target.pcf --json env/c9579aa6a23c50e82d40c688f8b2849c0a997c7e/hdl/projects/sidecar/ignition_target/ignition_target_rev_a_reset_button_top.json --asc env/2a4205778a478477ecd30a6f1010902e609bfefc/hdl/projects/sidecar/ignition_target/ignition_target_sidecar_rev_a_reset_button.config
ERROR: Unable to find legal placement for all cells, design is probably at utilisation limit.
0 warnings, 1 error

[Aaron-Hartwig]

I am good with removing those targets! I ran into this last time I bumped the toolchain and that is what pushed me to create the no button target! At this point it is becoming a maintenance burden to keep those other ones around and they're adding little value outside the bench top context in the lab.

[arjenroodselaar]

I drove by for unrelated reasons but happened to see this. If you need to bring these back for some reason you could try running the synth step with -noabc9, which uses the older synthesis code path which may fit again. And a more recent version of the Bluespec compiler may result in slightly smaller logic to begin with, but that's more moving parts. Cheers!