[llvm] [X86AsmParser] Check displacement overflow (PR #75747)

[Phoebe Wang via llvm-commits]

================
@@ -0,0 +1,36 @@
+# RUN: not llvm-mc -triple=x86_64 %s 2>&1 | FileCheck %s --check-prefixes=CHECK,64 --implicit-check-not=error: --implicit-check-not=warning:
+# RUN: llvm-mc -triple=i686 --defsym A16=1 %s 2>&1 | FileCheck %s --check-prefixes=CHECK,32 --implicit-check-not=error: --implicit-check-not=warning:
+
+.ifndef A16
+movq 0x80000000-1(%rip), %rax
+leaq -0x80000000(%rip), %rax
+
+# 64: [[#@LINE+1]]:17: error: displacement 2147483648 is not within [-2147483648, 2147483647]
+movq 0x80000000(%rip), %rax
+
+# 64: [[#@LINE+1]]:18: error: displacement -2147483649 is not within [-2147483648, 2147483647]
+leaq -0x80000001(%rip), %rax
+.endif
+
+movl 0xffffffff(%eax), %eax
+leal -0xffffffff(%eax), %eax
+
+# CHECK: [[#@LINE+1]]:19: warning: displacement 4294967296 shortened to signed 32-bit
+movl 0xffffffff+1(%eax), %eax
+
+# CHECK: [[#@LINE+1]]:20: warning: displacement -4294967296 shortened to signed 32-bit
+leal -0xffffffff-1(%eax), %eax
+
+{disp8} leal 0x100(%ebx), %eax
+{disp8} leal -0x100(%ebx), %eax
----------------
phoebewang wrote:

Should we emit `warning: displacement 256 is not within [-128, 127] for disp8 encoding`?

https://github.com/llvm/llvm-project/pull/75747