Don't shallow clone submodules (revert #1064)

[jerryz123]

Related PRs / Issues:

Reverts #1064. Shallow clones cause myriad problems,

Type of change:

• Bug fix
• New feature
• Other enhancement

Impact:

• RTL change
• Software change (RISC-V software)
• Build system change
• Other

Contributor Checklist:

• Did you set main as the base branch?
• Did you delete any extraneous prints/debugging code?
• (If applicable) Did you add documentation for the feature?
• (If applicable) Did you add a test demonstrating the PR?

• (If applicable) Did you mark the PR as "Please Backport"?

Release Notes

[jerryz123]

It appears firesim points to some backport

[abejgonzalez]

LGTM. I think we should add the option in the init-submods step to shallow clone things (can you add an issue for this) but this solves the majority of our problems.

[abejgonzalez]

@jerryz123 is this ready to merge?

[jerryz123]

Not yet. I need to check why this still seems to have some git errors in CI

[jerryz123]

I don't really know why this is failing. Was not able to reproduce locally.

[DorianXGH]

Did you manage to get the same commit hashes for the submodules ? Idk, it seems like the CI pipeline fails to find the right commits.

[abejgonzalez]

Have you fully reverted both #1125 and #1064. Seems like you should just move the submodule init back outside the search_submodule loop to fully revert. IDK if that solves the problem but worth a shot.

[jerryz123]

Those changes only affect the check-commit script? The error in CI shouldn't be related to that, (I'm looking at the prepare-chipyard jobs).

[abejgonzalez]

Hmm. On further look, I'm also quite stumped. I've gone ahead and added some debug prints but this is still very weird.

[abejgonzalez]

Pending fix, remove my checkmark

[abejgonzalez]

This should be fixed now. I think this was some weird issue with the FireSim tests / CI that was fixed with the extra cd.

[abejgonzalez]

The tests are failing because FireSim needs to get the libdwarf fix.

[jerryz123]

I don't see how the firesim script would have affected the init-submodules for the chipyard-only jobs, but it looks like they work now?

[jerryz123]

Do you want to just merge this and fix libdwarf in another PR?

[abejgonzalez]

Yea it was really weird, it looked like the firesim- jobs in the workflow would run first, mess up the repository, fail, then go on to the prepare- steps. The cd helps in general but is probably not the main fix (was probably a combo of this fix + me clearing the old repos from the CI machine itself).

I think this PR is ready to merge. I think @sagark is planning on minor releasing soon which would bump FireSim in CY for the libdwarf fix.

[jerryz123]

Ok. I'll merge it now then. Thanks for looking at it.