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: [RFC] Don't completely scalarize a record if it contains bit-field (PR tree-optimization/45144)




On Jul 30, 2010, at 11:54 AM, kenner@vlsi1.ultra.nyu.edu (Richard Kenner) wrote:

What we IMHO need is a pass late in the gimple pipeline which will
optimize adjacent bitfield operations and lower to BIT_FIELD_REF ops
or something similar, because bitfield ops are too hard to be handled
efficiently after the expansion.

That's an interesting idea. Does anyone else have an idea as to the best plan here?

Note that fold-const.c has done this in some cases for a very long time.

Interesting because we have this extact pass here at cavium. I can post the patch set that adds it to 4.3.3 if anyone wants to look at them.



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