[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH v3 21/26] tcg-aarch64: Introduce tcg_out_insn_35
From: |
Claudio Fontana |
Subject: |
Re: [Qemu-devel] [PATCH v3 21/26] tcg-aarch64: Introduce tcg_out_insn_3507 |
Date: |
Wed, 9 Apr 2014 14:54:00 +0200 |
User-agent: |
Mozilla/5.0 (Windows NT 6.1; rv:24.0) Gecko/20100101 Thunderbird/24.0.1 |
On 03.04.2014 21:56, Richard Henderson wrote:
> Cleaning up the implementation of REV and REV16 at the same time.
>
> Reviewed-by: Claudio Fontana <address@hidden>
> Signed-off-by: Richard Henderson <address@hidden>
> ---
> tcg/aarch64/tcg-target.c | 22 ++++++++++++++--------
> 1 file changed, 14 insertions(+), 8 deletions(-)
During testing I found this patch causes a regression for big endian targets
(sparc).
Can you take a look?
I think it might be related to the extended form of the REV instruction needing
an additional 0x400. See below.
>
> diff --git a/tcg/aarch64/tcg-target.c b/tcg/aarch64/tcg-target.c
> index caaf8a2..de7490d 100644
> --- a/tcg/aarch64/tcg-target.c
> +++ b/tcg/aarch64/tcg-target.c
> @@ -327,6 +327,10 @@ typedef enum {
> I3506_CSEL = 0x1a800000,
> I3506_CSINC = 0x1a800400,
>
> + /* Data-processing (1 source) instructions. */
> + I3507_REV16 = 0x5ac00400,
> + I3507_REV = 0x5ac00800,
> +
> /* Data-processing (2 source) instructions. */
> I3508_LSLV = 0x1ac02000,
> I3508_LSRV = 0x1ac02400,
> @@ -545,6 +549,12 @@ static void tcg_out_insn_3506(TCGContext *s, AArch64Insn
> insn, TCGType ext,
> | tcg_cond_to_aarch64[c] << 12);
> }
>
> +static void tcg_out_insn_3507(TCGContext *s, AArch64Insn insn, TCGType ext,
> + TCGReg rd, TCGReg rn)
> +{
> + tcg_out32(s, insn | ext << 31 | rn << 5 | rd);
> +}
> +
> static void tcg_out_insn_3509(TCGContext *s, AArch64Insn insn, TCGType ext,
> TCGReg rd, TCGReg rn, TCGReg rm, TCGReg ra)
> {
> @@ -961,19 +971,15 @@ static void tcg_out_brcond(TCGContext *s, TCGMemOp ext,
> TCGCond c, TCGArg a,
> }
>
> static inline void tcg_out_rev(TCGContext *s, TCGType ext,
> - TCGReg rd, TCGReg rm)
> + TCGReg rd, TCGReg rn)
> {
> - /* using REV 0x5ac00800 */
> - unsigned int base = ext ? 0xdac00c00 : 0x5ac00800;
see the extended form 0xdac00c00 <-
> - tcg_out32(s, base | rm << 5 | rd);
> + tcg_out_insn(s, 3507, REV, ext, rd, rn);
> }
>
> static inline void tcg_out_rev16(TCGContext *s, TCGType ext,
> - TCGReg rd, TCGReg rm)
> + TCGReg rd, TCGReg rn)
> {
> - /* using REV16 0x5ac00400 */
> - unsigned int base = ext ? 0xdac00400 : 0x5ac00400;
> - tcg_out32(s, base | rm << 5 | rd);
> + tcg_out_insn(s, 3507, REV16, ext, rd, rn);
while this does not have it.
> }
>
> static inline void tcg_out_sxt(TCGContext *s, TCGType ext, TCGMemOp s_bits,
>
Ciao,
Claudio
- [Qemu-devel] [PATCH v3 13/26] tcg-aarch64: Implement tcg_register_jit, (continued)
- [Qemu-devel] [PATCH v3 13/26] tcg-aarch64: Implement tcg_register_jit, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 14/26] tcg-aarch64: Avoid add with zero in tlb load, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 15/26] tcg-aarch64: Use tcg_out_call for qemu_ld/st, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 16/26] tcg-aarch64: Use ADR to pass the return address to the ld/st helpers, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 17/26] tcg-aarch64: Use TCGMemOp in qemu_ld/st, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 19/26] tcg-aarch64: Implement TCG_TARGET_HAS_new_ldst, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 18/26] tcg-aarch64: Pass qemu_ld/st arguments directly, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 21/26] tcg-aarch64: Introduce tcg_out_insn_3507, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 20/26] tcg-aarch64: Support stores of zero, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 23/26] tcg-aarch64: Replace aarch64_ldst_op_data with TCGMemOp, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 22/26] tcg-aarch64: Merge aarch64_ldst_get_data/type into tcg_out_op, Richard Henderson, 2014/04/03
- [Qemu-devel] [PATCH v3 24/26] tcg-aarch64: Replace aarch64_ldst_op_data with AArch64LdstType, Richard Henderson, 2014/04/03