[mips] BSEL's and BINS[RL] operands are reversed compared to the vselect node used in the pattern.

Summary:
Correct the match patterns and the lowerings that made the CodeGen tests pass despite the mistakes.

The original testcase that discovered the problem was SingleSource/UnitTests/SignlessType/factor.c in test-suite.
During review, we also found that some of the existing CodeGen tests were incorrect and fixed them:
* bitwise.ll: In bsel_v16i8 the IfSet/IfClear were reversed because bsel and bmnz have different operand orders and the test didn't correctly account for this. bmnz goes 'IfClear, IfSet, CondMask', while bsel goes 'CondMask, IfClear, IfSet'.
* vec.ll: In the cases where a bsel is emitted as a bmnz (they are the same operation with a different input tied to the result) the operands were in the wrong order.
* compare.ll and compare_float.ll: The bsel operand order was correct for a greater-than comparison, but a greater-than comparison instruction doesn't exist. Lowering this operation inverts the condition so the IfSet/IfClear need to be swapped to match.

The differences between BSEL, BMNZ, and BMZ and how they map to/from vselect are rather confusing. I've therefore added a note to MSA.txt to explain this in a single place in addition to the comments that explain each case.

Reviewers: matheusalmeida, jacksprat

Reviewed By: matheusalmeida

Differential Revision: http://llvm-reviews.chandlerc.com/D3028

llvm-svn: 203657

1 files changed, 8 insertions, 4 deletions

diff --git a/llvm/test/CodeGen/Mips/msa/compare_float.ll b/llvm/test/CodeGen/Mips/msa/compare_float.ll
index f5e8d9d9d6c..e93221b9361 100644
--- a/llvm/test/CodeGen/Mips/msa/compare_float.ll
+++ b/llvm/test/CodeGen/Mips/msa/compare_float.ll
@@ -525,7 +525,8 @@ define void @bsel_v4f32(<4 x float>* %d, <4 x float>* %a, <4 x float>* %b,
   %4 = fcmp ogt <4 x float> %1, %2
   ; CHECK-DAG: fclt.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
   %5 = select <4 x i1> %4, <4 x float> %1, <4 x float> %3
-  ; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
+  ; Note that IfSet and IfClr are swapped since the condition is inverted
+  ; CHECK-DAG: bsel.v [[R4]], [[R3]], [[R1]]
   store <4 x float> %5, <4 x float>* %d
   ; CHECK-DAG: st.w [[R4]], 0($4)
 
@@ -546,7 +547,8 @@ define void @bsel_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b,
   %4 = fcmp ogt <2 x double> %1, %2
   ; CHECK-DAG: fclt.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
   %5 = select <2 x i1> %4, <2 x double> %1, <2 x double> %3
-  ; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
+  ; Note that IfSet and IfClr are swapped since the condition is inverted
+  ; CHECK-DAG: bsel.v [[R4]], [[R3]], [[R1]]
   store <2 x double> %5, <2 x double>* %d
   ; CHECK-DAG: st.d [[R4]], 0($4)
 
@@ -565,7 +567,8 @@ define void @bseli_v4f32(<4 x float>* %d, <4 x float>* %a, <4 x float>* %b,
   %3 = fcmp ogt <4 x float> %1, %2
   ; CHECK-DAG: fclt.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
   %4 = select <4 x i1> %3, <4 x float> %1, <4 x float> zeroinitializer
-  ; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3:\$w[0-9]+]]
+  ; Note that IfSet and IfClr are swapped since the condition is inverted
+  ; CHECK-DAG: bsel.v [[R4]], [[R3:\$w[0-9]+]], [[R1]]
   store <4 x float> %4, <4 x float>* %d
   ; CHECK-DAG: st.w [[R4]], 0($4)
 
@@ -584,7 +587,8 @@ define void @bseli_v2f64(<2 x double>* %d, <2 x double>* %a, <2 x double>* %b,
   %3 = fcmp ogt <2 x double> %1, %2
   ; CHECK-DAG: fclt.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
   %4 = select <2 x i1> %3, <2 x double> %1, <2 x double> zeroinitializer
-  ; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3:\$w[0-9]+]]
+  ; Note that IfSet and IfClr are swapped since the condition is inverted
+  ; CHECK-DAG: bsel.v [[R4]], [[R3:\$w[0-9]+]], [[R1]]
   store <2 x double> %4, <2 x double>* %d
   ; CHECK-DAG: st.d [[R4]], 0($4)