[Libre-soc-bugs] [Bug 550] binutils support needed for svp64

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=550

--- Comment #89 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to Dmitry Selyutin from comment #88)
> I'm also curious about these entries in minor_31.csv:
> 
> 0b1101111010,SHIFT_ROT,OP_EXTSWSLI,NONE,CONST_SH,RS,RA,NONE,CR0,0,0,ZERO,0,
> NONE,0,0,0,0,0,0,RC,0,0,extswsli,XS,
> 0b1101111011,SHIFT_ROT,OP_EXTSWSLI,NONE,CONST_SH,RS,RA,NONE,CR0,0,0,ZERO,0,
> NONE,0,0,0,0,0,0,RC,0,0,extswsli,XS,

https://github.com/antonblanchard/microwatt/blob/6ff3b2499ceaf911252f634d44469fefb58952ae/decode1.vhdl#L255

let's have a look at the spec (v3.0C p108 book I 3.3.14.2.1

       445 sh Rc
    21     30 31

MSB0 bit 30 is for the "sh" field.  a "convenient" way to ensure that
only 9 bits of what is normally considered to be the XO field both go
to EXTSWSLI would be... to have two entries.

> Almost the same except for opcode. I assume the latter is `extswsli.`?
> 
> UPD: same for isel.

likewise for isel except there are 32 identical entries.

pick one. doesn't matter which because they're identical.  the relationship
is by instruction name not any other fields.

the link from svp64_struct to binutils is by name.

all and any other relationships are for other purposes not in any way
related to the job needed to be done here.

-- 
You are receiving this mail because:
You are on the CC list for the bug.