[Libre-soc-isa] [Bug 968] document shift-and-add instruction

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=968

--- Comment #14 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to Andrey Miroshnikov from comment #11)
> (In reply to Luke Kenneth Casson Leighton from comment #10)
> > 64-bit on both operands, it makes no difference whatsoever.
> Ah ok, although the operands are unsigned, the sign extension allows to do
> address offsetting.
> 
> > 
> > compare:
> > 
> >    ADD( RA, SIGNEXTEND( LOWERHALF(RB) )
> > 
> > with:
> > 
> >    ADD( RA, ZEROEXTEND( LOWERHALF(RB) )
> > 
> > then set RB=0x0000_0000_8000_0000 and compare the two.
> 
> Thanks, that info was missing from bitmanip.mdwn.

yes. you don't put massive explanations into pseudocode.
it goes into notes (here) or a discussion page.

> Where are these statements located?

they are not. i wrote them for you, just now.

> > because it is more compact when doing address offsets.  takes up less
> > space.  RA base 64 bit, offsets into data structure need only be 32bit.
> 
> Does that mean shadduw pseudocode is missing a EXTS() statement?

no it does not.

> There is no explicit sign-extension in the pseudocode.

correct.  if the desired behaviour was to perform a signed 32bit add
then not only would EXTS be used, the instruction would be called "shaddsw"
(or sigh shaddws)

> Adjusted based on your explanation:
> https://git.libre-soc.org/?p=libreriscv.git;a=commitdiff;
> h=95a74ba3566333ac498bb1110781e563f86b4c06

read again.  i said *no* on changing the pseudocode markdown for
ISACaller.  why did you then change the pseudocode markdown used
by ISACaller?

please revert

@@ -122,17 +124,21 @@ shadd r4, r1, r2, 3
 Pseudocode:

     shift <- sm + 1                                    # Shift is between 1-4
-    n <- (RB)[XLEN/2:XLEN-1]               # Limit RB to upper word (32-bits)
+    n <- (RB)[32:63]                           # Only use lower 32-bits of RB
     sum[0:63] <- (n << shift) + (RA)    # Shift n, add RA
     RT <- sum                                      # Result stored in RT

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