audio: midi: do not cast into unsigned long in UMP_* macros

[titouanc]

In commit 5e83d22, all bit masking
operations in audio/midi.h were modified to use BIT_MASK(n) instead of
literal hexadecimal constants. However, masking a number with a BIT_MASK
promotes it to an unsigned long (while UMPs are packets of uint32_t aka
unsigned on most platforms). This results in annoying warnings from LOG_*
calls, such as:

warning: format '%lX' expects argument of type 'long unsigned int',
but argument 4 has type 'uint32_t' {aka 'unsigned int'}

There is no point in making these long integers, so let's revert the
bit masks as they were before, with literal hexadecimal values.

This should remove some CI annoyance:

• Fixes samples: usb: sample.usb_device_next.midi: build failure on NXP series platforms #97242
• Fixes audio: midi: implicit cast to unsigned long result from UMP_xxx() macros #97298
• This is an alternative to subsys: usb: device_next: class: fix format specifier warning in debug log #97251
• This is an alternative to audio: midi: revert macros from using BIT_MASK #97297

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[etienne-lms]

Thanks for the fix. LGTM.

That said, I wonder if BIT(), BIT_MASK() and GENMASK() should be changed to use U instead of UL qualifiers? @nashif, @kartben, would that make sense?

[titouanc]

I wonder if BIT(), BIT_MASK() and GENMASK() should be changed to use U instead of UL qualifiers

I'd expect this to introduce some breakages (similar to the one being fixed here) all over the tree. If engaging in such a refactor; it could make sense have a long variant like BITL(), BIT_MASKL() and GENMASKL() ?

[dleach02]

Can we please expedite this one as it is blocking some of our PRs in CI
@jfischer-no