check-kmod-load-unload: use TOPDIR, don't rely on PATH

[marc-hb]

This was apparently the only test in run-all-tests.sh that required a PATH change.

[marc-hb]

https://sof-ci.01.org/softestpr/PR808/build924/devicetest/?model=APL_UP2_NOCODEC&testcase=check-suspend-resume-with-capture-5 is 5.6-rc1 regression thesofproject/linux#3283

The alsabat failure is also unrelated, this PR changes only one, different test.

[marc-hb]

Big regressions in https://sof-ci.01.org/softestpr/PR808/build927/devicetest/ unrelated to this, see thesofproject/linux#3292 for more details.

EDIT: fixed by revert thesofproject/linux#3294

[marc-hb]

SOFCI TEST

[greg-intel]

I'm not sure of the specific intent, but all the changes more or less make sense to me. I don't see any obvious issues, assuming the following isn't related:
2021-11-24 04:16:00 UTC Sub-Test: [REMOTE_ERROR] Caught error in kernel log 2021-11-24 04:16:00 UTC Sub-Test: [REMOTE_ERROR] Starting func_exit_handler(), exit status=1, FUNCNAME stack: 2021-11-24 04:16:00 UTC Sub-Test: [REMOTE_ERROR] die() @ /home/ubuntu/sof-test/test-case/../case-lib/lib.sh 2021-11-24 04:16:00 UTC Sub-Test: [REMOTE_ERROR] main() @ /home/ubuntu/sof-test/test-case/check-suspend-resume.sh:91 2021-11-24 04:16:00 UTC Sub-Test: [REMOTE_ERROR] main() @ /home/ubuntu/sof-test/test-case/check-suspend-resume.sh:91 2021-11-24 04:16:00 UTC Sub-Test: [REMOTE_INFO] Test Result: FAIL!

[marc-hb]

https://sof-ci.01.org/softestpr/PR808/build929/devicetest/?model=CML_RVP_SDW&testcase=check-suspend-resume-with-playback-5 is "preemptible" BUG thesofproject/linux#3283

sh-cml-sku0955-hda-02 crashed in the middle of the test run before running this test, so unrelated.

Everything else is green.

[greg-intel]

Last question, @marc-hb, Did the GitHub actions test these script changes?
[Update] I see now that the script is checked by sof-ci, and the other change is comments.

[greg-intel]

Changes look fairly straightforward.

[marc-hb]

Minor, zero functional change up for 10 days in review : merging.