qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [RFC PATCH v2 05/44] target/loongarch: Implement vadd/vsub


From: Richard Henderson
Subject: Re: [RFC PATCH v2 05/44] target/loongarch: Implement vadd/vsub
Date: Wed, 29 Mar 2023 08:22:23 -0700
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Thunderbird/102.8.0

On 3/29/23 02:59, gaosong wrote:

在 2023/3/29 上午3:59, Richard Henderson 写道:
On 3/27/23 20:05, Song Gao wrote:
+    func(mop, vd_ofs, vj_ofs, vk_ofs, 16, 16);

Oh, reading about ASXD and 256-bit vectors makes me wonder if it would be better to plan ahead and have a function, or DisasContext member, for the length of the vector.

like arm:

/* Return the byte size of the "whole" vector register, VL / 8.  */
static inline int vec_full_reg_size(DisasContext *s)
{
     return s->vl;
}

What I'm confused about is what is the difference between s->vl and s->vec_len ?

The first is for aarch64 SVE.
The second is for armv5 VFP (which was removed from armv8).


r~



reply via email to

[Prev in Thread] Current Thread [Next in Thread]