(no commit message)

diff --git a/openpower/sv/bitmanip/appendix.mdwn b/openpower/sv/bitmanip/appendix.mdwn
index 4b8e20f72658be18a7b6e8ffd743ba7cdd7f8640..bb784ffdb5607a5244e0ae09a6375099e0756db0 100644 (file)
--- a/openpower/sv/bitmanip/appendix.mdwn
+++ b/openpower/sv/bitmanip/appendix.mdwn
@@ -132,7 +132,8 @@ have a second hidden (implicit) destination register, RS.
 RS is calculated as RT+VL, where all scalar operations
 assume VL=1.  With `sv.msubx` *creating* a pair of Vector
 results, `sv.weirdaddx` correspondingly has to pick the
-pair up in order to carry on the algorithm.
+pairs up, containing the split lo-hi 128-bit products,
+in order to carry on the algorithm.
 
 **msubx RT, RA, RB, RC** (RS=RT+VL for SVP64, RS=RT+1 for scalar)
 
@@ -141,9 +142,9 @@ pair up in order to carry on the algorithm.
     RT <- sub[64:127]
     RS <- sub[0:63]
 
-**weirdaddx RT, RA, RB** (RS=RT+VL for SVP64, RS=RT+1 for scalar)
+**weirdaddx RT, RA, RB** (RS=RB+VL for SVP64, RS=RB+1 for scalar)
 
-    cat[0:127] = (RB) || (RS)
+    cat[0:127] = (RS) || (RB)
     sum[0:127] = cat + EXTZ(RA) + [1]*128
     rhi[0:63] = sum[0:63]
     if (RA) <= 1 then rhi = rhi + ([0]*63 || 1)
@@ -152,14 +153,15 @@ pair up in order to carry on the algorithm.
 
 These two combine as, simply:
 
-    # RS=RT+VL, assume VL=8, therefore RS starts at r8.v
+    # assume VL=8, therefore RS starts at r8.v
     # q       : r16
     # dividend: r20.v
     # divisor : r28.v
     # carry   : r40
     li r17, 0
     sv.msubx r0.v, r16, r20.v, r28.v
-    sv.weirdaddx r0.v, r17, r8.v
+    # here, RS=RB+VL, therefore again RS starts at r8.v
+    sv.weirdaddx r0.v, r17, r0.v
 
 As a result, a big-integer subtract and multiply may be carried out
 in only 3 instructions, one of which is setting a scalar integer to