; REQUIRES: x86 ; RUN: llc -mtriple=i686-pc-windows-msvc -filetype=obj -o %T/lto-lazy-reference-quadruple.obj %S/Inputs/lto-lazy-reference-quadruple.ll ; RUN: llvm-as -o %T/lto-lazy-reference-dummy.bc %S/Inputs/lto-lazy-reference-dummy.ll ; RUN: rm -f %t.lib ; RUN: llvm-ar cru %t.lib %T/lto-lazy-reference-quadruple.obj %T/lto-lazy-reference-dummy.bc ; RUN: llvm-as -o %t.obj %s ; RUN: lld-link /out:%t.exe /entry:main /subsystem:console %t.obj %t.lib target datalayout = "e-m:x-p:32:32-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:32-n8:16:32-a:0:32-S32" target triple = "i686-pc-windows-msvc18.0.0" ; Define fltused, since we don't link against the MS C runtime but are ; using floats. @_fltused = dllexport global i32 0, align 4 define double @main(double %x) { entry: ; When compiled, this defines the __real@40800000 symbol, which already has a ; lazy definition in the lib file from lto-lazy-reference-quadruple.obj. This ; test makes sure we *don't* try to take the definition from the lazy ; reference, because that can bring in new references to bitcode files after ; LTO, such as lto-lazy-reference-dummy.bc in this case. %mul = fmul double %x, 4.0 ret double %mul }