This is the mail archive of the gcc-patches@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]

Re: [PATCH] Fix PR37870, ICE in extract_bit_field_1


Ian Lance Taylor <iant@google.com> writes:

> We are not supposed to take a subreg of a floating point mode.  We
> currently permit word_mode subregs of a floating point value, which is
> probably why your patch works.  But we want to get away from that.
> See validate_subeg.  Conceptually this is because machines with
> separate floating point registers generally don't have access to the
> individual words in the register, so we wind up forcing the value into
> memory anyhow.
>
> The right choice here may be to somehow push the value into memory.

Or, if there is a larger integer mode, copy the value into a new
register with that mode.

Ian


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