On Tue, 18 Nov 2014, Matthew Fortune wrote: > > > I admit to being a bit more nervous about 4.9 but the test coverage > > > seems thorough enough. I guess I would have been less concerned if the > > > optimisation was still just tied to TARGET_MICROMIPS for the 4.9 > > branch. > > > > > > Catherine, what do you think? > > > > > This is okay for 4.9 IMO. > > OK
FWIW we've been using this change since Oct 2012 with no issues (as I noted it was meant to be included with the original microMIPS support submission, but was lost in transit) and also GAS has code to relax out-of-range branches to jumps in non-PIC standard MIPS code under the same condition this RTL insn uses, so even if a wrong branch slipped through here (which it doesn't), then GAS would fix it up. See gas/config/tc-mips.c (md_apply_fix) <BFD_RELOC_16_PCREL_S2> for the relaxation piece if interested. Maciej