This patch implements local-dynamic TLS model support for the 64-bit

PowerPC target.  This is the last of the four models, so we now have 
full TLS support.

This is mostly a straightforward extension of the general dynamic model.
I had to use an additional Chain operand to tie ADDIS_DTPREL_HA to the
register copy following ADDI_TLSLD_L; otherwise everything above the
ADDIS_DTPREL_HA appeared dead and was removed.

As before, there are new test cases to test the assembly generation, and
the relocations output during integrated assembly.  The expected code
gen sequence can be read in test/CodeGen/PowerPC/tls-ld.ll.

There are a couple of things I think can be done more efficiently in the
overall TLS code, so there will likely be a clean-up patch forthcoming;
but for now I want to be sure the functionality is in place.

Bill

llvm-svn: 170003

2 files changed, 73 insertions, 0 deletions

diff --git a/llvm/test/CodeGen/PowerPC/tls-ld-obj.ll b/llvm/test/CodeGen/PowerPC/tls-ld-obj.ll
new file mode 100644
index 00000000000..4b9f340fdda
--- /dev/null
+++ b/llvm/test/CodeGen/PowerPC/tls-ld-obj.ll
@@ -0,0 +1,49 @@
+; RUN: llc -mcpu=pwr7 -O0 -filetype=obj -relocation-model=pic %s -o - | \
+; RUN: elf-dump --dump-section-data | FileCheck %s
+
+; Test correct relocation generation for thread-local storage using
+; the local dynamic model.
+
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+
+@a = hidden thread_local global i32 0, align 4
+
+define signext i32 @main() nounwind {
+entry:
+  %retval = alloca i32, align 4
+  store i32 0, i32* %retval
+  %0 = load i32* @a, align 4
+  ret i32 %0
+}
+
+; Verify generation of R_PPC64_GOT_TLSGD16_HA, R_PPC64_GOT_TLSGD16_LO,
+; and R_PPC64_TLSGD for accessing external variable a, and R_PPC64_REL24
+; for the call to __tls_get_addr.
+;
+; CHECK:       '.rela.text'
+; CHECK:       Relocation 0
+; CHECK-NEXT:  'r_offset'
+; CHECK-NEXT:  'r_sym', 0x[[SYM1:[0-9a-f]+]]
+; CHECK-NEXT:  'r_type', 0x00000056
+; CHECK:       Relocation 1
+; CHECK-NEXT:  'r_offset'
+; CHECK-NEXT:  'r_sym', 0x[[SYM1]]
+; CHECK-NEXT:  'r_type', 0x00000054
+; CHECK:       Relocation 2
+; CHECK-NEXT:  'r_offset'
+; CHECK-NEXT:  'r_sym', 0x[[SYM1]]
+; CHECK-NEXT:  'r_type', 0x0000006c
+; CHECK:       Relocation 3
+; CHECK-NEXT:  'r_offset'
+; CHECK-NEXT:  'r_sym', 0x{{[0-9a-f]+}}
+; CHECK-NEXT:  'r_type', 0x0000000a
+; CHECK:       Relocation 4
+; CHECK-NEXT:  'r_offset'
+; CHECK-NEXT:  'r_sym', 0x[[SYM1]]
+; CHECK-NEXT:  'r_type', 0x0000004d
+; CHECK:       Relocation 5
+; CHECK-NEXT:  'r_offset'
+; CHECK-NEXT:  'r_sym', 0x[[SYM1]]
+; CHECK-NEXT:  'r_type', 0x0000004b
+
diff --git a/llvm/test/CodeGen/PowerPC/tls-ld.ll b/llvm/test/CodeGen/PowerPC/tls-ld.ll
new file mode 100644
index 00000000000..1ebc6129e2a
--- /dev/null
+++ b/llvm/test/CodeGen/PowerPC/tls-ld.ll
@@ -0,0 +1,24 @@
+; RUN: llc -mcpu=pwr7 -O0 -relocation-model=pic < %s | FileCheck %s
+
+; Test correct assembly code generation for thread-local storage using
+; the local dynamic model.
+
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+
+@a = hidden thread_local global i32 0, align 4
+
+define signext i32 @main() nounwind {
+entry:
+  %retval = alloca i32, align 4
+  store i32 0, i32* %retval
+  %0 = load i32* @a, align 4
+  ret i32 %0
+}
+
+; CHECK:      addis [[REG:[0-9]+]], 2, a@got@tlsld@ha
+; CHECK-NEXT: addi 3, [[REG]], a@got@tlsld@l
+; CHECK-NEXT: bl __tls_get_addr(a@tlsld)
+; CHECK-NEXT: nop
+; CHECK-NEXT: addis [[REG2:[0-9]+]], 3, a@dtprel@ha
+; CHECK-NEXT: addi {{[0-9]+}}, [[REG2]], a@dtprel@l