This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug target/78804] [RX] -m64bit-doubles does not work


https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78804

Oleg Endo <olegendo at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ian at airs dot com

--- Comment #14 from Oleg Endo <olegendo at gcc dot gnu.org> ---
(In reply to joseph@codesourcery.com from comment #13)
> fp-bit != soft-fp.

Ugh, sorry.  I didn't realize that they are two distinct pieces.  I was really
just referring to fp-bit.  I have not tried soft-fp on RX.  But if it uses
bit-fields in the same way as fp-bit does at the moment, it will have issues on
RX as well.

Adding Ian to CC, who is listed as fp-bit maintainer.

> soft-fp always uses bit-fields (with the order 
> depending on the endianness).  It's possible that in some cases you need 
> to ensure the declared types of the bit-fields are such that no padding 
> gets inserted and so you have the right layout.

Since we're at it ... how to ensure bitfield layout and enclosing struct size? 
I think fp-bit already tries to do that with __attribute__((packed)) and yet it
fails...

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]