Fixup test to accept different Opcode values

[dyung]

Summary

The test is searching for a specific opcode when it doesn't really need to. This causes a problem when the opcode changes for whatever reason, so update the test with a change similar to r274774 to make the test more robust.

[kristof.beyls]

LGTM - I've got the same change as part of https://reviews.llvm.org/D28079 - that part already got approved there.
Feel free to commit.

[dyung]

In D28487#641115, @kristof.beyls wrote:

LGTM - I've got the same change as part of https://reviews.llvm.org/D28079 - that part already got approved there.
Feel free to commit.

Thanks! I've committed the change in r291605.

llvm/trunk/test/CodeGen/X86/implicit-null-checks.mir

 # CHECK-LABEL: name:            imp_null_check_with_bitwise_op_4
 alignment:       4
 tracksRegLiveness: true
 liveins:
   - { reg: '%rdi' }
   - { reg: '%rsi' }
 # CHECK:  bb.0.entry:
 # CHECK:  %rbx = MOV64rr %rdx
-# CHECK-NEXT:  %rdi = FAULTING_LOAD_OP %bb.3.is_null, 260, killed %rbx, killed %rdi, 1, _, 0, _, implicit-def dead %eflags :: (load 4 from %ir.x)
+# CHECK-NEXT:  %rdi = FAULTING_LOAD_OP %bb.3.is_null, {{[0-9]+}}, killed %rbx, killed %rdi, 1, _, 0, _, implicit-def dead %eflags :: (load 4 from %ir.x)
 
 body:             |
   bb.0.entry:
     successors: %bb.3.is_null, %bb.1.not_null
     liveins: %rsi, %rdi, %rdx
 
     TEST64rr %rdi, %rdi, implicit-def %eflags
     JE_1 %bb.3.is_null, implicit %eflags