Re: [PATCH/RFA] SH TLS support

[kaz Kojima <kkojima at rr dot iij4u dot or dot jp>]

Joern Rennecke <amylaar at onetel dot net dot uk> wrote:
> Right, since we have no SH TLS implementation right now, an implementation
> with some inefficiencies is a vast improvement from having none at all.
> But we should document the fact that a better solution is possible.
> We often use comments with ??? where we do something that is inefficient
> (or in some cases even wrong)-: in the md file to mark places that
> could be improved on later.
> 
> You patch is OK if you put such comments in there.
> 
> Something like
> ;; ??? this should be a define_insn_and_split
> ;; See the thread [PATCH/RFA] SH TLS support on gcc-patches
> ;; <http://gcc.gnu.org/ml/gcc-patches/2003-02/msg01898.html>
> ;; for details.

It's reasonable. I'd like to check in the patch with a such comment.
Thanks a lot for reviewing and suggestions.

> Or you could summarize this discussion and put it once into comments,
> and refer to these comments in the three places where these long fixed
> sequences are, but that might a bit much work for little gain -
> I hope this does get eventually mproved before the gcc-patches web archives
> move or get abandomed.

Sure. Losing these archives would be a nightmare.

Regards,
	kaz