Add default cfg_t and debug_module_config_t constructor to libriscv

[jerryz123]

Adding additional config options to cfg_t has, up to now, required propagating the flag to all code which instantiates cfg_t, including projects which use Spike as a library.

Now, the default constructor for cfg_t will construct the default spike parameters.

This should be merged after #1522 to avoid merge conflicts.

[aswaterman]

I didn't look at the code, but the rationale checks out. Probably worth pinging this thread after #1522 is merged.

[scottj97]

I'm not sure I understand the removal of cfg_arg_t. (There's only one place left that uses it after this change.)

@rswarbrick introduced this in #946 to keep track of which config vars had been overridden from their initial values (e.g. via command-line). Do we not need that capability anymore?

[jerryz123]

@rswarbrick introduced this in #946 to keep track of which config vars had been overridden from their initial values (e.g. via command-line). Do we not need that capability anymore?

That capability was never used once the cfg_t was passed into sim_t. I believe some other code in spike main uses the cfg_arg_t ... the explicit_hartids stuff.

My understanding is that the overridden API in cfg_t was added to eventually support a use case where the user provides cmdline args, and a DTS/DTB, and spike would implement some code to check that the DTS/DTB is consistent with the command line args (#938). However, none of the consistency checking has been implemented. I'm not actually sure this "consistency check" approach makes sense... do we need to support the use case of providing both DTB and cmdline args which affect the same system configurations?

[jerryz123]

@aswaterman @scottj97 This is ready now.

[aswaterman]

Still OK with me; let's wait for @scottj97

[scottj97]

I don’t care

…

On Sun, Dec 10, 2023 at 7:36 PM Andrew Waterman ***@***.***> wrote: Still OK with me; let's wait for @scottj97 <https://github.com/scottj97> — Reply to this email directly, view it on GitHub <#1526 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAVCTTUPEPFMCHTJYPNBFLYIZPTVAVCNFSM6AAAAABALZIKR2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNBZGE4TENJVGA> . You are receiving this because you were mentioned.Message ID: ***@***.***>