[Libre-soc-bugs] [Bug 432] New: mulld pseudocode overflow calculation incorrect

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=432

            Bug ID: 432
           Summary: mulld pseudocode overflow calculation incorrect
           Product: Libre-SOC's first SoC
           Version: unspecified
          Hardware: All
                OS: All
            Status: CONFIRMED
          Severity: normal
          Priority: ---
         Component: Specification
          Assignee: lkcl at lkcl.net
          Reporter: programmerjake at gmail.com
                CC: libre-soc-bugs at lists.libre-soc.org
   NLnet milestone: ---

I was reading some of the pseudocode, and discovered that the overflow
calculation is incorrect.

it currently is:
prod[0:127] <- MULS((RA), (RB))
RT <- prod[64:127]
overflow <- ((prod[0:63] != 0x0000_0000_0000_0000) &
             (prod[0:63] != 0xffff_ffff_ffff_ffff))

it should be:
prod[0:127] <- MULS((RA), (RB))
RT <- prod[64:127]
overflow <- ((prod[0:64] != 0x0_0000_0000_0000_0000) &
             (prod[0:64] != 0x1_ffff_ffff_ffff_ffff))

since otherwise a product such as 1 << 63 is marked as not overflowing when it
is clearly out of range since it's greater than the max i64. likewise for -1 <<
64, which is smaller than the min i64.

reporting as a bug rather than just altering the pseudocode so we can check any
existing code that is built off the incorrect pseudocode

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