Rich,
Thanks for the updated patch.
Please do not start new threads for a continuation of an existing
thread. This makes it difficult to track in the archives.
On Tue, 2015-10-20 at 23:41 -0400, Rich Felker wrote:
> Attached is a hopefully near-ready-for-commit version of the SH/FDPIC
> patch. I believe I've addressed all comments by Oleg and Kaz on the
> previous versions of the patch. I'm still working on drafting the
> Changelog entry (there's a lot to go in it, and I might very well be
> going into more detail than is needed).
Other than the missing ChangeLog: How did you test the patch?
> One thing I've considered doing, since TARGET_FDPIC implies flag_pic
> now, is removing all parts of the patch that just replace checks for
> flag_pic with (flag_pic || TARGET_FDPIC). Would doing this be
> desirable? It shrinks the patch a bit but of course more strongly
> codes the assumption that TARGET_FDPIC implies flag_pic.
If FDPIC only ever will make sense in combination with flag_pic != 0,
then I guess this could be done. If you do that, please add a comment
above this hunk:
> + if (TARGET_FDPIC && !flag_pic)
> + flag_pic = 2;
Some other nits:
> rtx lab = function_symbol (func_addr_rtx, "__movmemSI12_i4",
> SFUNC_STATIC).lab;
Break overlong lines to fit into 80 columns. E.g.
rtx lab = function_symbol (func_addr_rtx, "__movmemSI12_i4",
SFUNC_STATIC).lab;
> if (TARGET_FDPIC
> && (TARGET_SHMEDIA || TARGET_SHCOMPACT || !TARGET_SH2))
> sorry ("non-SH2 FDPIC");
Drop SH5 stuff.
> if (TARGET_FDPIC)
> {
> emit_move_insn (gen_rtx_REG (Pmode, PIC_REG),
> sh_get_fdpic_reg_initial_val ());
> }
>
Remove braces around single statements.
> return (GET_CODE (x) != CONST_DOUBLE
> || mode == DFmode || mode == SFmode
> || mode == DImode || GET_MODE (x) == VOIDmode);
Remove unnecessary parens around return statements.
When applying the patch I'm getting:
patching file gcc/config/sh/sh-protos.h
(Stripping trailing CRs from patch; use --binary to disable.)
Maybe something with your editor settings?
Cheers,
Oleg
