[SimplifyCFG] Switch not fully optimized

[nikic]

https://alive2.llvm.org/ce/z/vAWi88

define i8 @src(i8 noundef %arg) {
start:
  switch i8 %arg, label %unreachable [
    i8 0, label %case012
    i8 1, label %case1
    i8 2, label %case2
    i8 3, label %end
  ]

unreachable:
  unreachable

case1:
  br label %case012

case2:
  br label %case012

case012:
  %phi1 = phi i8 [ 3, %case2 ], [ 2, %case1 ], [ 1, %start ]
  br label %end

end:
  %phi2 = phi i8 [ %phi1, %case012 ], [ 4, %start ]
  ret i8 %phi2
}

We should combine this into a single phi, resulting in something like this:

define i8 @tgt(i8 noundef %arg) {
start:
  switch i8 %arg, label %unreachable [
    i8 0, label %end
    i8 1, label %case1
    i8 2, label %case2
    i8 3, label %case3
  ]

unreachable:
  unreachable

case1:
  br label %end

case2:
  br label %end

case3:
  br label %end

end:
  %phi = phi i8 [ 4, %case3 ], [ 3, %case2 ], [ 2, %case1 ], [ 1, %start ]
  ret i8 %phi
}

At which point it is subject to lookup table optimization and converted into "add 1".

[XChy]

A possible Solution here.

[XChy]

I agree with your idea of merging phis. However, here I found that lookup optimization doesn't fold the latter IR for the reason that "i8 is not type legal for lookup table" on my x86_64 machine. Actually, it's unfoldable for i32, i64, too. But I think such "LinearMapKind" lookup table should be machine-independent, which may be a bug? What's your opinion here.

[XChy]

#67275 resolves it.