[Libre-soc-dev] daily kan-ban update 04jan2022
luke.leighton at gmail.com
Tue Jan 4 15:51:56 GMT 2022
today i tried running linux-5.7 under microwatt-verilator: after an hour it went into a loop writing to the same location.
debugging that is going to be... difficult, precisely because it takes an entire hour to get to that point.
also i found that simply commenting out one line of code (pick one) from the mmu.bin executable causes mmu_test_4 to succeed rather than fail.
given that running the exact same program using the same libresoc core passes under an alternative verilator simulation i suspect a problem with microwatt-verilator rather than the libresoc core itself.
i may instead try some different approaches to see if more info can be had before diving in to what could turn into a week's worth of coding.
More information about the Libre-soc-dev