DebugInfo: Use distinct inlinedAt MDLocations to avoid separate inlined calls being coalesced

When two calls from the same MDLocation are inlined they currently get
treated as one inlined function call (creating difficulty debugging,
duplicate variables, etc).

Clang worked around this by including column information on inline calls
which doesn't address LTO inlining or calls to the same function from
the same line and column (such as through a macro). It also didn't
address ctor and member function calls.

By making the inlinedAt locations distinct, every call site has an
explicitly distinct location that cannot be coalesced with any other
call.

This can produce linearly (2x in the worst case where every call is
inlined and the call instruction has a non-call instruction at the same
location) more debug locations. Any increase beyond that are in cases
where the Clang workaround was insufficient and the new scheme is
creating necessary distinct nodes that were being erroneously coalesced
previously.

After this change to LLVM the incomplete workarounds in Clang. That
should reduce the number of debug locations (in a build without column
info, the default on Darwin, not the default on Linux) by not creating
pseudo-distinct locations for every call to an inline function.

(oh, and I made the inlined-at chain rebuilding iterative instead of
recursive because I was having trouble wrapping my head around it the
way it was - open to discussion on the right design for that function
(including going back to a recursive solution))

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@226736 91177308-0d34-0410-b5e6-96231b3b80d8

lib/Transforms/Utils/InlineFunction.cpp

@@ -823,20 +823,42 @@ static bool hasLifetimeMarkers(AllocaInst *AI) {
   return false;
 }
 
-/// updateInlinedAtInfo - Helper function used by fixupLineNumbers to
-/// recursively update InlinedAtEntry of a DebugLoc.
-static DebugLoc updateInlinedAtInfo(const DebugLoc &DL, 
-                                    const DebugLoc &InlinedAtDL,
-                                    LLVMContext &Ctx) {
-  if (MDNode *IA = DL.getInlinedAt(Ctx)) {
-    DebugLoc NewInlinedAtDL 
-      = updateInlinedAtInfo(DebugLoc::getFromDILocation(IA), InlinedAtDL, Ctx);
-    return DebugLoc::get(DL.getLine(), DL.getCol(), DL.getScope(Ctx),
-                         NewInlinedAtDL.getAsMDNode(Ctx));
+/// Rebuild the entire inlined-at chain for this instruction so that the top of
+/// the chain now is inlined-at the new call site.
+static DebugLoc
+updateInlinedAtInfo(DebugLoc DL, MDLocation *InlinedAtNode,
+                    LLVMContext &Ctx,
+                    DenseMap<const MDLocation *, MDLocation *> &IANodes) {
+  SmallVector<MDLocation*, 3> InlinedAtLocations;
+  MDLocation *Last = InlinedAtNode;
+  DebugLoc CurInlinedAt = DL;
+
+  // Gather all the inlined-at nodes
+  while (MDLocation *IA =
+             cast_or_null<MDLocation>(CurInlinedAt.getInlinedAt(Ctx))) {
+    // Skip any we've already built nodes for
+    if (MDLocation *Found = IANodes[IA]) {
+      Last = Found;
+      break;
+    }
+
+    InlinedAtLocations.push_back(IA);
+    CurInlinedAt = DebugLoc::getFromDILocation(IA);
   }
 
-  return DebugLoc::get(DL.getLine(), DL.getCol(), DL.getScope(Ctx),
-                       InlinedAtDL.getAsMDNode(Ctx));
+  // Starting from the top, rebuild the nodes to point to the new inlined-at
+  // location (then rebuilding the rest of the chain behind it) and update the
+  // map of already-constructed inlined-at nodes.
+  for (auto I = InlinedAtLocations.rbegin(), E = InlinedAtLocations.rend();
+       I != E; ++I) {
+    const MDLocation *MD = *I;
+    Last = IANodes[MD] = MDLocation::getDistinct(
+        Ctx, MD->getLine(), MD->getColumn(), MD->getScope(), Last);
+  }
+
+  // And finally create the normal location for this instruction, referring to
+  // the new inlined-at chain.
+  return DebugLoc::get(DL.getLine(), DL.getCol(), DL.getScope(Ctx), Last);
 }
 
 /// fixupLineNumbers - Update inlined instructions' line numbers to 
@@ -847,6 +869,20 @@ static void fixupLineNumbers(Function *Fn, Function::iterator FI,
   if (TheCallDL.isUnknown())
     return;
 
+  auto &Ctx = Fn->getContext();
+  auto *InlinedAtNode = cast<MDLocation>(TheCallDL.getAsMDNode(Ctx));
+
+  // Create a unique call site, not to be confused with any other call from the
+  // same location.
+  InlinedAtNode = MDLocation::getDistinct(
+      Ctx, InlinedAtNode->getLine(), InlinedAtNode->getColumn(),
+      InlinedAtNode->getScope(), InlinedAtNode->getInlinedAt());
+
+  // Cache the inlined-at nodes as they're built so they are reused, without
+  // this every instruction's inlined-at chain would become distinct from each
+  // other.
+  DenseMap<const MDLocation *, MDLocation *> IANodes;
+
   for (; FI != Fn->end(); ++FI) {
     for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
          BI != BE; ++BI) {
@@ -864,7 +900,7 @@ static void fixupLineNumbers(Function *Fn, Function::iterator FI,
 
         BI->setDebugLoc(TheCallDL);
       } else {
-        BI->setDebugLoc(updateInlinedAtInfo(DL, TheCallDL, BI->getContext()));
+        BI->setDebugLoc(updateInlinedAtInfo(DL, InlinedAtNode, BI->getContext(), IANodes));
         if (DbgValueInst *DVI = dyn_cast<DbgValueInst>(BI)) {
           LLVMContext &Ctx = BI->getContext();
           MDNode *InlinedAt = BI->getDebugLoc().getInlinedAt(Ctx);

test/DebugInfo/inline-debug-info-multiret.ll

@@ -10,7 +10,7 @@
 ; CHECK: br label %invoke.cont, !dbg ![[MD]]
 ; The branch instruction has the source location of line 9 and its inlined location
 ; has the source location of line 14.
-; CHECK: ![[INL:[0-9]+]] = !MDLocation(line: 14, scope: {{.*}})
+; CHECK: ![[INL:[0-9]+]] = distinct !MDLocation(line: 14, scope: {{.*}})
 ; CHECK: ![[MD]] = !MDLocation(line: 9, scope: {{.*}}, inlinedAt: ![[INL]])
 
 ; ModuleID = 'test.cpp'

test/DebugInfo/inline-debug-info.ll

@@ -28,11 +28,11 @@
 ; CHECK: _Z4testi.exit:
 ; Make sure the branch instruction created during inlining has a debug location,
 ; so the range of the inlined function is correct.
-; CHECK: br label %invoke.cont, !dbg ![[MD:[0-9]+]]
+; CHECK: br label %invoke.cont, !dbg [[MD:![0-9]+]]
 ; The branch instruction has the source location of line 9 and its inlined location
 ; has the source location of line 14.
-; CHECK: ![[INL:[0-9]+]] = !MDLocation(line: 14, scope: {{.*}})
-; CHECK: ![[MD]] = !MDLocation(line: 9, scope: {{.*}}, inlinedAt: ![[INL]])
+; CHECK: [[INL:![0-9]*]] = distinct !MDLocation(line: 14, scope: {{.*}})
+; CHECK: [[MD]] = !MDLocation(line: 9, scope: {{.*}}, inlinedAt: [[INL]])
 
 ; ModuleID = 'test.cpp'
 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-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"

test/DebugInfo/inline-no-debug-info.ll

@@ -21,10 +21,11 @@
 
 ; Debug location of the code in caller() and of the inlined code that did not
 ; have any debug location before.
-; CHECK-DAG: [[A]] = !MDLocation(line: 4, scope: !{{[01-9]+}})
+; CHECK-DAG: [[A]] = !MDLocation(line: 4, scope: !{{[0-9]+}})
 
 ; Debug location of the inlined code.
-; CHECK-DAG: [[B]] = !MDLocation(line: 2, scope: !{{[01-9]+}}, inlinedAt: [[A]])
+; CHECK-DAG: [[B]] = !MDLocation(line: 2, scope: !{{[0-9]+}}, inlinedAt: [[A_INL:![0-9]*]])
+; CHECK-DAG: [[A_INL]] = distinct !MDLocation(line: 4, scope: !{{[0-9]+}})
 
 
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"

test/Transforms/Inline/debug-info-duplicate-calls.ll

@@ -0,0 +1,121 @@
+; RUN: opt < %s -always-inline -S | FileCheck %s
+
+; Original input generated from clang -emit-llvm -S -c -mllvm -disable-llvm-optzns
+;
+; #define CALLS1 f2(); f2();
+; #define CALLS2 f4(); f4();
+; void f1();
+; inline __attribute__((always_inline)) void f2() {
+;   f1();
+; }
+; inline __attribute__((always_inline)) void f3() {
+;   CALLS1
+; }
+; inline __attribute__((always_inline)) void f4() {
+;   f3();
+; }
+; void f() {
+;   CALLS2
+; }
+
+; There should be unique locations for all 4 of these instructions, correctly
+; describing the inlining that has occurred, even in the face of duplicate call
+; site locations.
+
+; The nomenclature used for the tags here is <function name>[cs<number>] where
+; 'cs' is an abbreviation for 'call site' and the number indicates which call
+; site from within the named function this is. (so, given the above inlining, we
+; should have 4 calls to 'f1', two from the first call to f4 and two from the
+; second call to f4)
+
+; CHECK: call void @_Z2f1v(), !dbg [[fcs1_f4_f3cs1_f2:![0-9]+]]
+; CHECK: call void @_Z2f1v(), !dbg [[fcs1_f4_f3cs2_f2:![0-9]+]]
+; CHECK: call void @_Z2f1v(), !dbg [[fcs2_f4_f3cs1_f2:![0-9]+]]
+; CHECK: call void @_Z2f1v(), !dbg [[fcs2_f4_f3cs2_f2:![0-9]+]]
+
+; CHECK-DAG: [[F:![0-9]+]] = {{.*}} ; [ DW_TAG_subprogram ] {{.*}} [f]
+; CHECK-DAG: [[F2:![0-9]+]] = {{.*}} ; [ DW_TAG_subprogram ] {{.*}} [f2]
+; CHECK-DAG: [[F3:![0-9]+]] = {{.*}} ; [ DW_TAG_subprogram ] {{.*}} [f3]
+; CHECK-DAG: [[F4:![0-9]+]] = {{.*}} ; [ DW_TAG_subprogram ] {{.*}} [f4]
+
+; CHECK: [[fcs1_f4_f3cs1_f2]] = {{.*}}, scope: [[F2]], inlinedAt: [[fcs1_f4_f3cs1:![0-9]+]])
+; CHECK: [[fcs1_f4_f3cs1]] = {{.*}}, scope: [[F3]], inlinedAt: [[fcs1_f4:![0-9]+]])
+; CHECK: [[fcs1_f4]] = {{.*}}, scope: [[F4]], inlinedAt: [[fcs1:![0-9]+]])
+; CHECK: [[fcs1]] = {{.*}}, scope: [[F]])
+; CHECK: [[fcs1_f4_f3cs2_f2]] = {{.*}}, scope: [[F2]], inlinedAt: [[fcs1_f4_f3cs2:![0-9]+]])
+; CHECK: [[fcs1_f4_f3cs2]] = {{.*}}, scope: [[F3]], inlinedAt: [[fcs1_f4]])
+
+; CHECK: [[fcs2_f4_f3cs1_f2]] = {{.*}}, scope: [[F2]], inlinedAt: [[fcs2_f4_f3cs1:![0-9]+]])
+; CHECK: [[fcs2_f4_f3cs1]] = {{.*}}, scope: [[F3]], inlinedAt: [[fcs2_f4:![0-9]+]])
+; CHECK: [[fcs2_f4]] = {{.*}}, scope: [[F4]], inlinedAt: [[fcs2:![0-9]+]])
+; CHECK: [[fcs2]] = {{.*}}, scope: [[F]])
+; CHECK: [[fcs2_f4_f3cs2_f2]] = {{.*}}, scope: [[F2]], inlinedAt: [[fcs2_f4_f3cs2:![0-9]+]])
+; CHECK: [[fcs2_f4_f3cs2]] = {{.*}}, scope: [[F3]], inlinedAt: [[fcs2_f4]])
+
+$_Z2f4v = comdat any
+
+$_Z2f3v = comdat any
+
+$_Z2f2v = comdat any
+
+; Function Attrs: uwtable
+define void @_Z1fv() #0 {
+entry:
+  call void @_Z2f4v(), !dbg !13
+  call void @_Z2f4v(), !dbg !13
+  ret void, !dbg !14
+}
+
+; Function Attrs: alwaysinline inlinehint uwtable
+define linkonce_odr void @_Z2f4v() #1 comdat {
+entry:
+  call void @_Z2f3v(), !dbg !15
+  ret void, !dbg !16
+}
+
+; Function Attrs: alwaysinline inlinehint uwtable
+define linkonce_odr void @_Z2f3v() #1 comdat {
+entry:
+  call void @_Z2f2v(), !dbg !17
+  call void @_Z2f2v(), !dbg !17
+  ret void, !dbg !18
+}
+
+; Function Attrs: alwaysinline inlinehint uwtable
+define linkonce_odr void @_Z2f2v() #1 comdat {
+entry:
+  call void @_Z2f1v(), !dbg !19
+  ret void, !dbg !20
+}
+
+declare void @_Z2f1v() #2
+
+attributes #0 = { uwtable "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #1 = { alwaysinline inlinehint uwtable "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+attributes #2 = { "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+
+!llvm.dbg.cu = !{!0}
+!llvm.module.flags = !{!10, !11}
+!llvm.ident = !{!12}
+
+!0 = !{!"0x11\004\00clang version 3.7.0 (trunk 226474) (llvm/trunk 226478)\000\00\000\00\002", !1, !2, !2, !3, !2, !2} ; [ DW_TAG_compile_unit ] [/tmp/dbginfo/debug-info-duplicate-calls.cpp] [DW_LANG_C_plus_plus]
+!1 = !{!"debug-info-duplicate-calls.cpp", !"/tmp/dbginfo"}
+!2 = !{}
+!3 = !{!4, !7, !8, !9}
+!4 = !{!"0x2e\00f\00f\00\0013\000\001\000\000\00256\000\0013", !1, !5, !6, null, void ()* @_Z1fv, null, null, !2} ; [ DW_TAG_subprogram ] [line 13] [def] [f]
+!5 = !{!"0x29", !1}                               ; [ DW_TAG_file_type ] [/tmp/dbginfo/debug-info-duplicate-calls.cpp]
+!6 = !{!"0x15\00\000\000\000\000\000\000", null, null, null, !2, null, null, null} ; [ DW_TAG_subroutine_type ] [line 0, size 0, align 0, offset 0] [from ]
+!7 = !{!"0x2e\00f4\00f4\00\0010\000\001\000\000\00256\000\0010", !1, !5, !6, null, void ()* @_Z2f4v, null, null, !2} ; [ DW_TAG_subprogram ] [line 10] [def] [f4]
+!8 = !{!"0x2e\00f3\00f3\00\007\000\001\000\000\00256\000\007", !1, !5, !6, null, void ()* @_Z2f3v, null, null, !2} ; [ DW_TAG_subprogram ] [line 7] [def] [f3]
+!9 = !{!"0x2e\00f2\00f2\00\004\000\001\000\000\00256\000\004", !1, !5, !6, null, void ()* @_Z2f2v, null, null, !2} ; [ DW_TAG_subprogram ] [line 4] [def] [f2]
+!10 = !{i32 2, !"Dwarf Version", i32 4}
+!11 = !{i32 2, !"Debug Info Version", i32 2}
+!12 = !{!"clang version 3.7.0 (trunk 226474) (llvm/trunk 226478)"}
+!13 = !MDLocation(line: 14, column: 3, scope: !4)
+!14 = !MDLocation(line: 15, column: 1, scope: !4)
+!15 = !MDLocation(line: 11, column: 3, scope: !7)
+!16 = !MDLocation(line: 12, column: 1, scope: !7)
+!17 = !MDLocation(line: 8, column: 3, scope: !8)
+!18 = !MDLocation(line: 9, column: 1, scope: !8)
+!19 = !MDLocation(line: 5, column: 3, scope: !9)
+!20 = !MDLocation(line: 6, column: 1, scope: !9)

test/Transforms/Inline/debug-invoke.ll

@@ -6,7 +6,7 @@
 ; CHECK-NEXT: to label {{.*}} unwind label {{.*}}, !dbg [[INL_LOC:!.*]]
 ; CHECK: [[EMPTY:.*]] = !{}
 ; CHECK: [[INL_LOC]] = !MDLocation(line: 1, scope: [[EMPTY]], inlinedAt: [[INL_AT:.*]])
-; CHECK: [[INL_AT]] = !MDLocation(line: 2, scope: [[EMPTY]])
+; CHECK: [[INL_AT]] = distinct !MDLocation(line: 2, scope: [[EMPTY]])
 
 declare void @test()
 declare i32 @__gxx_personality_v0(...)

test/Transforms/Inline/inline_dbg_declare.ll

@@ -92,5 +92,6 @@ attributes #1 = { nounwind readnone }
 !22 = !MDLocation(line: 8, column: 14, scope: !9)
 !23 = !MDLocation(line: 9, column: 1, scope: !9)
 
-; CHECK: [[m23]] = !{!"0x101\00x\0016777217\000", !4, !5, !8, !22} ; [ DW_TAG_arg_variable ] [x] [line 1]
-; CHECK: [[m24]] = !MDLocation(line: 1, column: 17, scope: !4, inlinedAt: !22)
+; CHECK: [[m23]] = !{!"0x101\00x\0016777217\000", !4, !5, !8, [[CALL_SITE:![0-9]*]]} ; [ DW_TAG_arg_variable ] [x] [line 1]
+; CHECK: [[CALL_SITE]] = distinct !MDLocation(line: 8, column: 14, scope: !9)
+; CHECK: [[m24]] = !MDLocation(line: 1, column: 17, scope: !4, inlinedAt: [[CALL_SITE]])