[llvm] [StackSafetyAnalysis] Don't call getTruncateOrZeroExtend for pointers of different sizes (PR #79804)

[Fangrui Song via llvm-commits]

================
@@ -0,0 +1,19 @@
+; RUN: opt %s -disable-output -S -passes="print<stack-safety-local>" 2>&1 | FileCheck %s
+
+; Datalayout from AMDGPU, p5 is 32 bits and p is 64.
+; We used to call SCEV getTruncateOrZeroExtend on %x.ascast/%x which caused an assertion failure.
+
+; CHECK:      @a dso_preemptable
+; CHECK-NEXT:   args uses:
+; CHECK-NEXT:     x[]: full-set
+; CHECK-NEXT:   allocas uses:
+
+target datalayout = "e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-p7:160:256:256:32-p8:128:128-p9:192:256:256:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1-ni:7:8:9"
+
----------------
MaskRay wrote:

Set `target triple` and specify `; REQUIRES: amdgpu-registered-target`

https://github.com/llvm/llvm-project/pull/79804