llvm-mirror/test/CodeGen/X86/tls12.ll

; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu > %t
; RUN: grep {movb	%gs:i@NTPOFF, %al} %t
; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu > %t2
; RUN: grep {movb	%fs:i@TPOFF, %al} %t2

@i = thread_local global i8 15

define i8 @f() {
entry:
	%tmp1 = load i8* @i
	ret i8 %tmp1
}
Eliminate more uses of llvm-as and llvm-dis. llvm-svn: 81290 2009-09-09 01:54:48 +02:00			`; RUN: llc < %s -march=x86 -mtriple=i386-linux-gnu > %t`
Don't force promotion of return arguments on the callee. Some architectures (like x86) don't require it. This fixes bug 3779. llvm-svn: 67132 2009-03-18 00:43:59 +01:00			`; RUN: grep {movb %gs:i@NTPOFF, %al} %t`
Eliminate more uses of llvm-as and llvm-dis. llvm-svn: 81290 2009-09-09 01:54:48 +02:00			`; RUN: llc < %s -march=x86-64 -mtriple=x86_64-linux-gnu > %t2`
X86-64 TLS support for local exec and initial exec. llvm-svn: 68947 2009-04-13 15:02:49 +02:00			`; RUN: grep {movb %fs:i@TPOFF, %al} %t2`
optimize i8 and i16 tls values. llvm-svn: 66725 2009-03-11 23:40:04 +01:00
			`@i = thread_local global i8 15`

			`define i8 @f() {`
			`entry:`
			`%tmp1 = load i8* @i`
			`ret i8 %tmp1`
			`}`