1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 11:13:28 +01:00
llvm-mirror/test/CodeGen/X86/fp128-extract.ll
Fangrui Song e24be6085b [X86] Emit @PLT for x86-64 and keep unadorned symbols for x86-32
This essentially reverts the x86-64 side effect of r327198.

For x86-32, @PLT (R_386_PLT32) is not suitable in -fno-pic mode so the
code forces MO_NO_FLAG (like a forced dso_local) (https://bugs.llvm.org//show_bug.cgi?id=36674#c6).

For x86-64, both `call/jmp foo` and `call/jmp foo@PLT` emit R_X86_64_PLT32
(https://sourceware.org/bugzilla/show_bug.cgi?id=22791) so there is no
difference using @PLT. Using @PLT is actually favorable because this drops
a difference with -fpie/-fpic code and makes it possible to avoid a canonical
PLT entry when taking the address of an undefined function symbol.
2020-12-05 13:17:47 -08:00

29 lines
1.3 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -O2 -mtriple=x86_64-linux-android -mattr=+mmx -enable-legalize-types-checking | FileCheck %s
; RUN: llc < %s -O2 -mtriple=x86_64-linux-gnu -mattr=+mmx -enable-legalize-types-checking | FileCheck %s
; Test the softened result of extractelement op code.
define fp128 @TestExtract(<2 x double> %x) nounwind {
; CHECK-LABEL: TestExtract:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: subq $40, %rsp
; CHECK-NEXT: movaps %xmm0, (%rsp) # 16-byte Spill
; CHECK-NEXT: callq __extenddftf2
; CHECK-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; CHECK-NEXT: movaps (%rsp), %xmm0 # 16-byte Reload
; CHECK-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1]
; CHECK-NEXT: callq __extenddftf2
; CHECK-NEXT: movaps %xmm0, %xmm1
; CHECK-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; CHECK-NEXT: addq $40, %rsp
; CHECK-NEXT: jmp __multf3@PLT # TAILCALL
entry:
; Simplified instruction pattern from the output of llvm before r289042,
; for a boost function ...::insert<...>::traverse<...>().
%a = fpext <2 x double> %x to <2 x fp128>
%0 = extractelement <2 x fp128> %a, i32 0
%1 = extractelement <2 x fp128> %a, i32 1
%2 = fmul fp128 %0, %1
ret fp128 %2
}