[compiler-rt] [llvm] [msan] Add 32-bit platforms support (PR #109284)

[Vitaly Buka via llvm-commits]

================
@@ -371,17 +411,19 @@ declare <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float>) nounwind readnone
 define void @test_x86_sse_stmxcsr(ptr %a0) #0 {
 ; CHECK-LABEL: @test_x86_sse_stmxcsr(
 ; CHECK-NEXT:    [[TMP1:%.*]] = load i64, ptr @__msan_param_tls, align 8
+; CHECK-NEXT:    [[TMP5:%.*]] = load i64, ptr @__msan_va_arg_overflow_size_tls, align 8
+; CHECK-NEXT:    [[TMP6:%.*]] = add i64 0, [[TMP5]]
 ; CHECK-NEXT:    call void @llvm.donothing()
 ; CHECK-NEXT:    [[TMP2:%.*]] = ptrtoint ptr [[A0:%.*]] to i64
----------------
vitalybuka wrote:

there are a lot of conversions from ptr to i64, which is too much
but if msan works as is, let's keep this for follow up patches.

https://github.com/llvm/llvm-project/pull/109284