intel_adsp_ace15_mtpm.conf: temporarily disable CONFIG_MODULES

[marc-hb]

Temporarily disable CONFIG_MODULES to get MTL working again.

Avoids crash #9308

[marc-hb]

MTL https://sof-ci.01.org/sofpr/PR9313/build6572/devicetest/index.html is not finished yet but I looked at it on the internal site and everything is 100% green so far.

This one-line CONFIG_ change moves us from "NOTHING works on MTL" back to "everything works on MTL". This is NOT an actual fix #9308 but this restores project velocity while not even reverting a single line of actual code.

This does not even lose much LLEXT coverage because LNL is still being tested with LLEXT!
https://sof-ci.01.org/sofpr/PR9313/build6570/build/firmware_community/lnl.log

I don't think there can be any least intrusive way to get MTL CI back up and running and things moving again.

EDIT: kudos to @fredoh9 for leading me to this minimal solution: #9308 (comment)

The 2 pause failures in LNL https://sof-ci.01.org/sofpr/PR9313/build6571/devicetest/index.html are known (thesofproject/linux#5109 and #9123) and unrelated. This PR makes absolutely zero LNL change.

No cavs failure in https://sof-ci.01.org/sofpr/PR9313/build6573/devicetest/index.html, just a model missing.

[andyross]

Seems reasonable. Weird that the resume crash is MTL-specific though

[fredoh9]

Wow, good finding for minimum change. 👍

[marc-hb]

Weird that the resume crash is MTL-specific though

Yes: this seems like a very complex issue. All the relevant PRs passed in isolation, see details in #9308. I think this shows that bug is complex and unlikely to be root-caused and fixed in a short time.

[kv2019i]

Good to go. I don't think we have a choice, we need to get CI back functional and this is a minimal change. Some tests still running, but so far lal looks good. Please @lyakh and @tmleman follow up to the PR to reenable CONFIG_MODULES when fix is found.