Suppress shift warning for IAR compiler as well

[deepikabhavnani]

Description

Compiler gives invalid warning for shift operation, more details at https://github.com/ARMmbed/mbed-os/blob/master/drivers/MbedCRC.h#L23

Suppressing the warning for IAR compiler as well.

Pull request type

[X] Fix
[ ] Refactor
[ ] New target
[ ] Feature
[ ] Breaking change

Issue: #7251

[cmonr]

LGTM.

[cmonr]

+1 for the comment.

[cmonr]

/morph build

[cmonr]

@ARMmbed/mbed-os-maintainers So far, this is the second PR withing the last couple of days that I've seen resort to disabling warnings instead of updating code. It's fine, but it feels a bit risky imo.

Do we want to start doing something different in regards to these types of PRs, or just continue as is?

[mbed-ci]

Build : SUCCESS

Build number : 2393
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/7264/

Triggering tests

/morph test
/morph uvisor-test
/morph export-build
/morph mbed2-build

[mbed-ci]

Exporter Build : SUCCESS

Build number : 2029
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7264/

[kjbracey]

You've got to look at it case-by-case.

This one is absolutely a useful warning, so you definitely wouldn't want to disable it globally, but the particular code technique used here - achieving effective conditional compilation via if - triggers it. There's already a kill of the same warning present for another toolchain there, and the code is elegant and clear, and I wouldn't want to change it just to satisfy a machine that generates a spurious warning. (The observed negative shift can never actually be executed, so the total code is legal, but the warning is generated in the expression parser before it figures out the flowgraph).

My general feeling is that if the code gets less elegant or you start being coerced into poorer coding style because of a purely-advisory warning, the warning should go. Whether that's killed locally or globally depends on general utility of the warning versus how common false positives are.

Compilers can generate so many possible warnings, but their utility can range from "C required diagnostic, cos this isn't valid C" (so why not an error?) to "I've detected a real logic anomaly" (thanks!) to "this C code wouldn't be valid C++" (so?).

[mbed-ci]

Test : SUCCESS

Build number : 2180
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7264/2180