Re: [PATCH 2/4] TCG support for AVX

[Paolo Bonzini]

On 4/18/22 21:45, Paul Brook wrote:
> > Massively too large for a single patch, I'm afraid. This needs
> > to be split, probably into at least twenty patches, which each
> > are a reviewable chunk of code that does one coherent thing.
> Hmm, I'mm see what I can do.
>
> Unfortunately the table driven decoding means that going from two to
> three operands tends to be a bit all or nothing just to get the thing
> to compile.

Hi Paul, welcome back and thanks for this huge work.  It should be
possible at least to split the patch as follows (at least that's
what _I_ would do in order to review it):

* mechanical changes to translate.c

-    [0x10] = { SSE_SPECIAL, SSE_SPECIAL, SSE_SPECIAL, SSE_SPECIAL }, /* 
movups, movupd, movss, movsd */
+    [0x10] = SSE_SPECIAL, /* movups, movupd, movss, movsd */
+    [0x11] = SSE_SPECIAL, /* movups, movupd, movss, movsd */
+    [0x12] = SSE_SPECIAL, /* movlps, movlpd, movsldup, movddup */
+    [0x13] = SSE_SPECIAL, /* movlps, movlpd */

* mechanical introduction of XMM_OFFSET()

* non-AVX/VEX changes (e.g. SSE_OPF_3DNOW)

* decoding fixes for SSE instructions (CHECK_NO_VEX)

* fix zeroing of high bits

* 3-operand decoding and helpers

* AVX/AVX2 support (existing instructions)

* AVX/AVX2 support (new instructions)

I can do some of the work too since I was planning to do this
anyway (but have hardly started yet).

Paolo