> I don't think he was quite that definitive. "One reading of the manual > suggests ...". However, Richard's interpretation is the same as I've had > for eons. You can return the original class, a narrower class or NO_REGS.
Yes, I also read the manual and discussed the patch with Richard offline, but was a bit lazy in the description ;) > I've tested the attached patched in Thumb >>> >>> mode, on ARMv5, A9 and A9hf and on cross A15 without regression. >>> >>> Yvan >>> >>> 2013-11-20 Yvan Roux <yvan.r...@linaro.org> >>> >>> PR target/58785 >>> * config/arm/arm.c (arm_preferred_reload_class): Only return >>> LO_REGS >>> when rclass is GENERAL_REGS. > > This is fine for the trunk as it implements Richard's recommendation and > brings the code into compliance with the semantics of P_R_C. I'll also add the testcase I put in bugzilla and that was in Vlad's first patch, if it's ok for you. Yvan
