Test OpCooperativeMatrixLengthKHR (KhronosGroup#2135)

For now we are violating the specification by passing Matrix Id instead
of Matrix Type Id.

Place for improvement - to come up with the way to translate only the
type.

Author: vmaksimo

test/extensions/KHR/SPV_KHR_cooperative_matrix/cooperative_matrix.ll

@@ -22,14 +22,17 @@
 ; CHECK-SPIRV-DAG: TypeCooperativeMatrixKHR [[#MatTy2:]] [[#Int8Ty]] [[#Const0]] [[#Const12]] [[#Const48]] [[#Const3]]
 ; CHECK-SPIRV-DAG: TypeCooperativeMatrixKHR [[#MatTy3:]] [[#Int8Ty]] [[#Const2]] [[#Const48]] [[#Const12]] [[#Const3]]
 ; CHECK-SPIRV: CompositeConstruct [[#MatTy1]]
-; CHECK-SPIRV: CooperativeMatrixLoadKHR [[#MatTy2]]
+; CHECK-SPIRV: CooperativeMatrixLoadKHR [[#MatTy2]] [[#Load1:]]
+; TODO: Pass Matrix Type Id instead of Matrix Id to CooperativeMatrixLengthKHR.
+; CHECK-SPIRV: CooperativeMatrixLengthKHR [[#Int32Ty]] [[#]] [[#Load1]]
 ; CHECK-SPIRV: CooperativeMatrixLoadKHR [[#MatTy3]]
 ; CHECK-SPIRV: CooperativeMatrixMulAddKHR [[#MatTy1]]
 ; CHECK-SPIRV: CooperativeMatrixStoreKHR
 
 
 ; CHECK-LLVM: call spir_func target("spirv.CooperativeMatrixKHR", i32, 3, 12, 12, 3) @_Z26__spirv_CompositeConstructi(i32 0)
 ; CHECK-LLVM: call spir_func target("spirv.CooperativeMatrixKHR", i8, 0, 12, 48, 3) @_Z86__spirv_CooperativeMatrixLoadKHR_RPU3AS144__spirv_CooperativeMatrixKHR__char_0_12_48_3PU3AS4clii
+; CHECK-LLVM: call spir_func i32 @_Z34__spirv_CooperativeMatrixLengthKHRPU3AS144__spirv_CooperativeMatrixKHR__char_0_12_48_3(target("spirv.CooperativeMatrixKHR", i8, 0, 12, 48, 3)
 ; CHECK-LLVM: call spir_func target("spirv.CooperativeMatrixKHR", i8, 2, 48, 12, 3) @_Z86__spirv_CooperativeMatrixLoadKHR_RPU3AS144__spirv_CooperativeMatrixKHR__char_2_48_12_3PU3AS4cl
 ; CHECK-LLVM: call spir_func target("spirv.CooperativeMatrixKHR", i32, 3, 12, 12, 3) @_Z34__spirv_CooperativeMatrixMulAddKHRPU3AS144__spirv_CooperativeMatrixKHR__char_0_12_48_3PU3AS144__spirv_CooperativeMatrixKHR__char_2_48_12_3PU3AS144__spirv_CooperativeMatrixKHR__uint_3_12_12_3i(target("spirv.CooperativeMatrixKHR", i8, 0, 12, 48, 3) %{{.*}}, target("spirv.CooperativeMatrixKHR", i8, 2, 48, 12, 3) %{{.*}}, target("spirv.CooperativeMatrixKHR", i32, 3, 12, 12, 3)
 ; CHECK-LLVM: call spir_func void @_Z33__spirv_CooperativeMatrixStoreKHRPU3AS4iPU3AS144__spirv_CooperativeMatrixKHR__uint_3_12_12_3ili(ptr addrspace(4) %call.ascast.i.i, target("spirv.CooperativeMatrixKHR", i32, 3, 12, 12, 3)
@@ -105,6 +108,7 @@ for.body.i:                                       ; preds = %for.cond.i
   %add.ptr.i96.i = getelementptr inbounds i8, ptr addrspace(1) %add.ptr.i93.i, i64 %conv13.i
   %call.ascast.i66.i = addrspacecast ptr addrspace(1) %add.ptr.i96.i to ptr addrspace(4)
   %call1.i.i = tail call spir_func noundef target("spirv.CooperativeMatrixKHR", i8, 0, 12, 48, 3) @_Z32__spirv_CooperativeMatrixLoadKHR_1(ptr addrspace(4) noundef %call.ascast.i66.i, i64 noundef %_arg_K, i32 noundef 0, i32 noundef 1) #4
+  %len = tail call spir_func noundef i32 @_Z34__spirv_CooperativeMatrixLengthKHR(target("spirv.CooperativeMatrixKHR", i8, 0, 12, 48, 3) %call1.i.i)
   %div20.i = mul nsw i32 %k.0.i, 12
   %conv21.i = zext i32 %div20.i to i64
   %mul23.i = mul i64 %mul22.i, %conv21.i
@@ -136,6 +140,8 @@ _ZZZ15matrix_multiplyIiaLm24ELm96ELm24ELm96ELm24ELm24EEvR10big_matrixIT_XT5_EXT6
 ; Function Attrs: convergent
 declare dso_local spir_func noundef target("spirv.CooperativeMatrixKHR", i32, 3, 12, 12, 3) @_Z26__spirv_CompositeConstruct(i32 noundef) local_unnamed_addr #2
 
+declare dso_local spir_func noundef i32 @_Z34__spirv_CooperativeMatrixLengthKHR(target("spirv.CooperativeMatrixKHR", i8, 0, 12, 48, 3) noundef)
+
 ; Function Attrs: convergent
 declare dso_local spir_func noundef target("spirv.CooperativeMatrixKHR", i8, 0, 12, 48, 3) @_Z32__spirv_CooperativeMatrixLoadKHR_1(ptr addrspace(4) noundef, i64 noundef, i32 noundef, i32 noundef) local_unnamed_addr #2