This is the mail archive of the 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: ping: back end reinitialization hooks to support mixed mips16/nomips16 compilation in same file

On 7/8/07, Sandra Loosemore <> wrote:
Anyone had a chance to look over these patches yet?  Even some general feedback
or discussion about whether this is a good idea would be appreciated.

What's the advantage of controlling the target per function instead of per compilation unit? You can trivially split the compilation unit into a per function one...

Also this looks ugly ;)  And creating a backend object that can be
easily switched as you suggest will potentially cause a lot of indirection
overhead.  Maybe separate shared objects for libbackend (of course the
"right" parts of it) and explicitly calling the correct one would be a more
efficient (and easier to implement?) way?

That said - the only scenario I can see having more than one backend
available is if the compiler itself creates functions for the different
(sub-)architecture.  Like it would happen for auto-parallelization on
the Cell.


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