[llvm] [X86] Respect code models more when determining if a global reference can fit in 32 bits (PR #75386)

[Reid Kleckner via llvm-commits]

================
@@ -1072,7 +1127,7 @@ define dso_local float @load_constant_pool(float %x) #0 {
 ;
 ; MEDIUM-STATIC-LABEL: load_constant_pool:
 ; MEDIUM-STATIC:       # %bb.0:
-; MEDIUM-STATIC-NEXT:    movabsq ${{\.?LCPI[0-9]+_[0-9]+}}, %rax
+; MEDIUM-STATIC-NEXT:    movl ${{\.?LCPI[0-9]+_[0-9]+}}, %eax
----------------
rnk wrote:

These constant pool changes look correct.

https://github.com/llvm/llvm-project/pull/75386