usb: cdc_acm: log related warning promoted to error

[cfriedt]

Describe the bug
It seems that a cdc_acm logging feature is triggering a warning that gets promoted to error in twister.

To Reproduce
Steps to reproduce the behavior:

1. twister -i -p degu_evk -s zephyr/tests/posix/common/portability.posix.common.tls.newlib
2. See error

Expected behavior
The test should build and pass.

Impact
What impact does this issue have on your progress (e.g., annoyance, showstopper)

Logs and console output

/__w/zephyr/zephyr/subsys/usb/device/class/cdc_acm.c
/__w/zephyr/zephyr/subsys/usb/device/class/cdc_acm.c:63:2: error: #warning "USB_CDC_ACM_LOG_LEVEL forced to LOG_LEVEL_NONE" [-Werror=cpp]
   63 | #warning "USB_CDC_ACM_LOG_LEVEL forced to LOG_LEVEL_NONE"
      |  ^~~~~~~
cc1: all warnings being treated as errors

Environment (please complete the following information):

• OS: (e.g. Linux, MacOS, Windows): any
• Toolchain (e.g Zephyr SDK, ...): Zephyr SDK 0.15.2
• Commit SHA or Version used: 1173db6

Additional context
It's likely that this has to do with a recent change in logging cc @nordic-krch

This is currently failing in #52977 even after a rebase.

[jfischer-no]

This is a kind of regression caused by #53032, degu_evk configuration has to be fixed.

[cfriedt]

This is a kind of regression caused by #53032, degu_evk configuration has to be fixed.

@jfischer-no - is there a fix for this issue that would prevent needing more future fixes in other areas?

[cfriedt]

This was not fixed by #53149. #52977 is still failing due to the following.

twister -i -p bl654_usb                 -s zephyr/tests/posix/common/portability.posix.common.tls
...
/Users/cfriedt/workspace/zephyrproject/zephyr/subsys/usb/device/class/cdc_acm.c:63:2: error: #warning "USB_CDC_ACM_LOG_LEVEL forced to LOG_LEVEL_NONE" [-Werror=cpp]
   63 | #warning "USB_CDC_ACM_LOG_LEVEL forced to LOG_LEVEL_NONE"
      |  ^~~~~~~

[cfriedt]

@stephanosio - thought I should pull you into this one. It just warmed up a bit. I also commented on #53032. Just concerned that we might have changes coming up to release that can trigger several regressions.

[desowin]

is there a fix for this issue that would prevent needing more future fixes in other areas?

A possible "fix" is to just remove the #warning. The downside is that the configured CDC ACM log level would be silently ignored.