[llvm] [ARM]Adjust cost of muls in SMLAL patterns (PR #122713)

[Nashe Mncube via llvm-commits]

================
@@ -0,0 +1,80 @@
+; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple thumbv8.1-m.main -mattr=+dsp  < %s | FileCheck %s
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple thumbv8.1-m.main < %s | FileCheck %s --check-prefix=CHECK-NO-DSP
+define i64 @test(i16 %a, i16 %b) {
+; CHECK-LABEL: 'test'
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %as = sext i16 %a to i32
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bs = sext i16 %b to i32
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: %m = mul i32 %as, %bs
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %ms = sext i32 %m to i64
+; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %ms
+;
+; CHECK-NO-DSP-LABEL: 'test'
+; CHECK-NO-DSP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %as = sext i16 %a to i32
+; CHECK-NO-DSP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %bs = sext i16 %b to i32
+; CHECK-NO-DSP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %m = mul i32 %as, %bs
+; CHECK-NO-DSP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %ms = sext i32 %m to i64
+; CHECK-NO-DSP-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i64 %ms
+;
+    %as = sext i16 %a to i32
+    %bs = sext i16 %b to i32
----------------
nasherm wrote:

Done

https://github.com/llvm/llvm-project/pull/122713