[Libre-soc-dev] microwatt-libre-soc interoperable verilator snapshots / debugging

lkcl luke.leighton at gmail.com
Sun Jan 9 19:27:58 GMT 2022



On January 9, 2022 6:49:51 PM UTC, lkcl <luke.leighton at gmail.com> wrote:

>not if they are 100% compatible because it is [compatible]
>register/memory state being saved not [internal,
>implementation-specific] architectural state.

and, there is actually not that much difference.

if there was, it would not be possible for a kernel like linux to perform a context-switch in an architecturally-independent way.

one other possibility is to add the facility to execute one single instruction via the DMI interface.

this would be incredibly laborious, constructing a sequence of instructions that performed the exact same job of a context-switch.

and, unfortunately, because it is relying on the very same instruction execution engine and architecture in HDL that is to be tested, could end up triggering the very same bug that is desired to be tested!

"let's test for a bug in addi oh wait we have to use addi via the DMI interface to upload the registers" isn't going to fly :)

l.



More information about the Libre-soc-dev mailing list