Christian Bruel <christian.br...@st.com> wrote: > The attached patch fixes an ice while building the linux kernel. Reduced > in the included testcase. > > The problem is that we are generating a movhi_reg_reg insn that accepts > only registers as operands. Spilling a pseudo on the stack results in an > invalid memory load/store constraints. > > The attached patch allows memory for reload. > Tested with the testsuite on sh4-linux and sh-superh-gcc. > No performance impact on a large number of benchmarks (EEMBC, CSIBe, > spec2006, ...) > > Oleg, since you moved out the r,r constraints from *mohi into > movhi_reg_reg, do you agree ? > > OK for trunk and 4.8 ? OK. Updates suggested by Oleg are also fine. When you incorporate them into the patch, please send it to the list for the record.
Regards, kaz