[Bug target/89251] [2018-q4m] Wrong datatype optimization on bitfield
david.koch at libertysurf dot fr
gcc-bugzilla@gcc.gnu.org
Fri Feb 8 12:58:00 GMT 2019
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89251
--- Comment #4 from Kochise <david.koch at libertysurf dot fr> ---
The pointer I access is volatile, not the uint32_t behind. Why would I ? The
'volatile' keyword has no meaning to change the data size being processed, only
removing the caching behavior of the compiler.
If the compiler can decide to change the specified size of non-volatile
datatypes, you really must consider reviewing some common practices. I wanted
this to be uint32_t, keep it that way, thanks.
Where in the standard is written that bitfield could be resized out from their
base datatype ? What's the point anyway ? And 'volatile' would "prevent that" ?
How unfriendly...
More information about the Gcc-bugs
mailing list