[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH 00/11] target/arm: Fix neon reg offsets
|
From: |
Peter Maydell |
|
Subject: |
Re: [PATCH 00/11] target/arm: Fix neon reg offsets |
|
Date: |
Wed, 28 Oct 2020 16:48:58 +0000 |
On Wed, 28 Oct 2020 at 03:27, Richard Henderson
<richard.henderson@linaro.org> wrote:
>
> Much of the existing usage of neon_reg_offset is broken for
> big-endian hosts, as it computes the offset of the first
> 32-bit unit, not the offset of the entire vector register.
>
> Fix this by separating out the different usages. Make the
> whole thing look a bit more like the aarch64 code.
I haven't reviewed this yet but it fixes a lot of the
problems I saw in my risu run on an s390x box, and I
don't see any regressions on x86-64. However these still
fail on s390x compared to an x86-64 host:
insn_VPADD_float_f16.risu.bin FAIL
insn_VPMAX_float_f16.risu.bin FAIL
insn_VPMIN_float_f16.risu.bin FAIL
insn_VSDOT_s.risu.bin FAIL
insn_VUDOT_s.risu.bin FAIL
thanks
-- PMM
- [PATCH 04/11] target/arm: Use neon_element_offset in vfp_reg_offset, (continued)
- [PATCH 04/11] target/arm: Use neon_element_offset in vfp_reg_offset, Richard Henderson, 2020/10/27
- [PATCH 06/11] target/arm: Expand read/write_neon_element32 to all MemOp, Richard Henderson, 2020/10/27
- [PATCH 05/11] target/arm: Add read/write_neon_element32, Richard Henderson, 2020/10/27
- [PATCH 07/11] target/arm: Rename neon_load_reg32 to vfp_load_reg32, Richard Henderson, 2020/10/27
- [PATCH 08/11] target/arm: Add read/write_neon_element64, Richard Henderson, 2020/10/27
- [PATCH 09/11] target/arm: Rename neon_load_reg64 to vfp_load_reg64, Richard Henderson, 2020/10/27
- [PATCH 11/11] target/arm: Improve do_prewiden_3d, Richard Henderson, 2020/10/27
- [PATCH 10/11] target/arm: Simplify do_long_3d and do_2scalar_long, Richard Henderson, 2020/10/27
- Re: [PATCH 00/11] target/arm: Fix neon reg offsets,
Peter Maydell <=
- Re: [PATCH 00/11] target/arm: Fix neon reg offsets, Peter Maydell, 2020/10/28