[llvm] Adding support in llvm-exegesis for Aarch64 for handling FPR64/128, PPR16 and ZPR128 reg class. (PR #127564)

[Sjoerd Meijer via llvm-commits]

================
@@ -2,22 +2,30 @@
 
 # ppr register class initialization testcase 
 # ideally we should use PTRUE_{B/H?S/D} instead of FADDV_VPZ_D for isolated testcase; but exegesis does not support PTRUE_{B/H?S/D} yet;
-# RUN: llvm-exegesis -mcpu=neoverse-v2 -mode=latency -opcode-name=FADDV_VPZ_D 2>&1 | FileCheck %s --check-prefix=PPR
+# RUN: llvm-exegesis -mcpu=neoverse-v2 -mode=latency --dump-object-to-disk=FADDV_VPZ_D.o --opcode-name=FADDV_VPZ_D 2>&1 | FileCheck %s --check-prefix=PPR
+# RUN: llvm-objdump -d FADDV_VPZ_D.o | FileCheck %s --check-prefix=PPR_ASM
 # PPR-NOT: setRegTo is not implemented, results will be unreliable
 # PPR: assembled_snippet: {{.*}}C0035FD6
+# PPR_ASM: {{0|4}}:	{{.*}} ptrue p{{[0-9]|1[0-5]}}
 
 # zpr register class initialization testcase 
 # ideally we should use DUPM_ZI instead of FADDV_VPZ_S for isolated testcase; but exegesis does not support DUPM_ZI yet;
-# RUN: llvm-exegesis -mcpu=neoverse-v2 -mode=latency -opcode-name=FADDV_VPZ_S 2>&1 | FileCheck %s --check-prefix=ZPR
+# RUN: llvm-exegesis -mcpu=neoverse-v2 -mode=latency --dump-object-to-disk=FADDV_VPZ_S.o --opcode-name=FADDV_VPZ_S 2>&1 | FileCheck %s --check-prefix=ZPR
+# RUN: llvm-objdump -d FADDV_VPZ_S.o | FileCheck %s --check-prefix=ZPR_ASM
 # ZPR-NOT: setRegTo is not implemented, results will be unreliable
 # ZPR: assembled_snippet: {{.*}}C0035FD6
+# ZPR_ASM: {{4|8}}: {{.*}} dupm z{{[0-9]|[1-2][0-9]|3[0-1]}}.s, #0x1
 
 # fpr64 register class initialization testcase
-# RUN: llvm-exegesis -mcpu=neoverse-v2 -mode=latency -opcode-name=ADDVv4i16v 2>&1 | FileCheck %s --check-prefix=FPR64
+# RUN: llvm-exegesis -mcpu=neoverse-v2 -mode=latency --dump-object-to-disk=ADDVv4i16v.o --opcode-name=ADDVv4i16v 2>&1 | FileCheck %s --check-prefix=FPR64
+# RUN: llvm-objdump -d ADDVv4i16v.o | FileCheck %s --check-prefix=FPR64-ASM
 # FPR64-NOT: setRegTo is not implemented, results will be unreliable
 # FPR64: assembled_snippet: {{.*}}C0035FD6
+# FPR64-ASM: {{0|4}}:	{{.*}} fmov d{{[0-9]|[1-2][0-9]|3[0-1]}}, #2.0{{.*}}
----------------
sjoerdmeijer wrote:

You don't need the `{{.*}}` at the end.

https://github.com/llvm/llvm-project/pull/127564