On Tue, 19 Jan 2021 at 17:46, Richard Henderson <richard.hender...@linaro.org> wrote: > > On 1/19/21 4:38 AM, Peter Maydell wrote: > >> - case 's': > >> - /* qemu_st8_i32 data constraint */ > >> - ct->regs = 0xf; > >> -#ifdef CONFIG_SOFTMMU > >> - tcg_regset_reset_reg(ct->regs, TCG_REG_L0); > >> - tcg_regset_reset_reg(ct->regs, TCG_REG_L1); > >> -#endif > >> - break; > > > > But in the old code the 's' constraint is 0xf for both > > x86-64 and i386. > > That's perhaps laziness, or simply the lack of names in the old code. It > logically should be BYTEL, because that's where byte stores go from. > > In the end it doesn't matter, because this constraint is *only* used by i386. > The opcode INDEX_op_qemu_st8_i32 is not used by x86_64 at all. See > TCG_TARGET_HAS_qemu_st8_i32 in tcg-target.h.
OK. Can we keep actual changes in separate commits from just-refactoring changes, please? thanks -- PMM
