[LLD][ELF][ARM] Do not insert interworking thunks for non STT_FUNC sy…

…mbols

ELF for the ARM architecture requires linkers to provide
interworking for symbols that are of type STT_FUNC. Interworking for
other symbols must be encoded directly in the object file. LLD was always
providing interworking, regardless of the symbol type, this breaks some
programs that have branches from Thumb state targeting STT_NOTYPE symbols
that have bit 0 clear, but they are in fact internal labels in a Thumb
function. LLD treats these symbols as ARM and inserts a transition to Arm.

This fixes the problem for in range branches, R_ARM_JUMP24,
R_ARM_THM_JUMP24 and R_ARM_THM_JUMP19. This is expected to be the vast
majority of problem cases as branching to an internal label close to the
function.

There is at least one follow up patch required.
- R_ARM_CALL and R_ARM_THM_CALL may do interworking via BL/BLX
  substitution.

In theory range-extension thunks can be altered to not change state when
the symbol type is not STT_FUNC. I will need to check with ld.bfd to see if
this is the case in practice.

Fixes (part of) ClangBuiltLinux/linux#773

Differential Revision: https://reviews.llvm.org/D73474

(cherry picked from commit 4f38ab2)

lld/ELF/Arch/ARM.cpp

@@ -275,8 +275,8 @@ bool ARM::needsThunk(RelExpr expr, RelType type, const InputFile *file,
   case R_ARM_PLT32:
   case R_ARM_JUMP24:
     // Source is ARM, all PLT entries are ARM so no interworking required.
-    // Otherwise we need to interwork if Symbol has bit 0 set (Thumb).
-    if (expr == R_PC && ((s.getVA() & 1) == 1))
+    // Otherwise we need to interwork if STT_FUNC Symbol has bit 0 set (Thumb).
+    if (s.isFunc() && expr == R_PC && (s.getVA() & 1))
       return true;
     LLVM_FALLTHROUGH;
   case R_ARM_CALL: {
@@ -286,8 +286,8 @@ bool ARM::needsThunk(RelExpr expr, RelType type, const InputFile *file,
   case R_ARM_THM_JUMP19:
   case R_ARM_THM_JUMP24:
     // Source is Thumb, all PLT entries are ARM so interworking is required.
-    // Otherwise we need to interwork if Symbol has bit 0 clear (ARM).
-    if (expr == R_PLT_PC || ((s.getVA() & 1) == 0))
+    // Otherwise we need to interwork if STT_FUNC Symbol has bit 0 clear (ARM).
+    if (expr == R_PLT_PC || (s.isFunc() && (s.getVA() & 1) == 0))
       return true;
     LLVM_FALLTHROUGH;
   case R_ARM_THM_CALL: {

lld/test/ELF/arm-thumb-interwork-ifunc.s

@@ -0,0 +1,52 @@
+// REQUIRES: arm
+// RUN: llvm-mc --triple=armv7a-linux-gnueabihf -arm-add-build-attributes -filetype=obj -o %t.o %s
+// RUN: ld.lld %t.o -o %t
+// RUN: llvm-objdump -triple armv7a-none-linux-gnueabi -d --no-show-raw-insn %t
+
+/// Non-preemptible ifuncs are called via a PLT entry which is always Arm
+/// state, expect the ARM callers to go direct to the PLT entry, Thumb
+/// branches are indirected via state change thunks, the bl is changed to blx.
+
+ .syntax unified
+ .text
+ .balign 0x1000
+ .type foo STT_GNU_IFUNC
+ .globl foo
+foo:
+ bx lr
+
+ .section .text.1, "ax", %progbits
+ .arm
+ .global _start
+_start:
+ b foo
+ bl foo
+
+ .section .text.2, "ax", %progbits
+ .thumb
+ .global thumb_caller
+thumb_caller:
+ b foo
+ b.w foo
+ bl foo
+
+// CHECK: 00012004 _start:
+// CHECK-NEXT: b       #36
+// CHECK-NEXT: bl      #32
+
+// CHECK: 0001200c thumb_caller:
+// CHECK-NEXT: b.w     #8
+// CHECK-NEXT: b.w     #4
+// CHECK-NEXT: blx     #24
+
+// CHECK: 00012018 __Thumbv7ABSLongThunk_foo:
+// CHECK-NEXT: movw    r12, #8240
+// CHECK-NEXT: movt    r12, #1
+// CHECK-NEXT: bx      r12
+
+// CHECK: Disassembly of section .iplt:
+
+// CHECK: 00012030 $a:
+// CHECK-NEXT: add     r12, pc, #0, #12
+// CHECK-NEXT: add     r12, r12, #4096
+// CHECK-NEXT: ldr     pc, [r12, #8]!

lld/test/ELF/arm-thumb-interwork-notfunc.s

@@ -0,0 +1,73 @@
+// REQUIRES: arm
+// RUN: llvm-mc --triple=armv7a-linux-gnueabihf -arm-add-build-attributes -filetype=obj -o %t.o %s
+// RUN: ld.lld %t.o -o %t
+// RUN: llvm-objdump --no-show-raw-insn -d %t | FileCheck %s
+
+.syntax unified
+.section .arm_target, "ax", %progbits
+.balign 0x1000
+.arm
+arm_func_with_notype:
+.type arm_func_with_explicit_notype, %notype
+arm_func_with_explicit_notype:
+ bx lr
+
+.section .thumb_target, "ax", %progbits
+.balign 4
+.thumb
+thumb_func_with_notype:
+.type thumb_func_with_explicit_notype, %notype
+thumb_func_with_explicit_notype:
+ bx lr
+
+/// All the symbols that are targets of the branch relocations do not have
+/// type STT_FUNC. LLD should not insert interworking thunks as non STT_FUNC
+/// symbols have no state information, the ABI assumes the user has manually
+/// done the interworking.
+.section .arm_caller, "ax", %progbits
+.balign 4
+.arm
+.global _start
+_start:
+ b .arm_target
+ b arm_func_with_notype
+ b arm_func_with_explicit_notype
+ b .thumb_target
+ b thumb_func_with_notype
+ b thumb_func_with_explicit_notype
+
+ .section .thumb_caller, "ax", %progbits
+ .thumb
+ .balign 4
+ b.w .arm_target
+ b.w arm_func_with_notype
+ b.w arm_func_with_explicit_notype
+ b.w .thumb_target
+ b.w thumb_func_with_notype
+ b.w thumb_func_with_explicit_notype
+ beq.w .arm_target
+ beq.w arm_func_with_notype
+ beq.w arm_func_with_explicit_notype
+ beq.w .thumb_target
+ beq.w thumb_func_with_notype
+ beq.w thumb_func_with_explicit_notype
+
+// CHECK: 00012008 _start:
+// CHECK-NEXT: 12008: b       #-16
+// CHECK-NEXT: 1200c: b       #-20
+// CHECK-NEXT: 12010: b       #-24
+// CHECK-NEXT: 12014: b       #-24
+// CHECK-NEXT: 12018: b       #-28
+// CHECK-NEXT: 1201c: b       #-32
+// CHECK:      12020: b.w     #-36
+// CHECK-NEXT: 12024: b.w     #-40
+// CHECK-NEXT: 12028: b.w     #-44
+// CHECK-NEXT: 1202c: b.w     #-44
+// CHECK-NEXT: 12030: b.w     #-48
+// CHECK-NEXT: 12034: b.w     #-52
+// CHECK-NEXT: 12038: beq.w   #-60
+// CHECK-NEXT: 1203c: beq.w   #-64
+// CHECK-NEXT: 12040: beq.w   #-68
+// CHECK-NEXT: 12044: beq.w   #-68
+// CHECK-NEXT: 12048: beq.w   #-72
+// CHECK-NEXT: 1204c: beq.w   #-76

lld/test/ELF/arm-thunk-edgecase.s

@@ -18,6 +18,7 @@
 .arm
 .section .text_armfunc, "ax", %progbits
 .globl armfunc
+.type armfunc, %function
 armfunc:
 	b	thumbfunc