[llvm] r245588 - Fix a debug location handling bug in GVN.

David Blaikie via llvm-commits llvm-commits at lists.llvm.org
Thu Aug 20 16:15:18 PDT 2015


Got it a bit shorter, but not necessarily more obvious... I assume the
inlining is necessary, so probably good to use
__attribute__((always_inline) on that nested function. not sure what else.

int a, b;
void f1(int *p1) {
  if (b)
    a = 1;
  if (a && *p1)
    f4();
}
int f2(int);
void f3(void) {
  a = f2(1);
  f1(&a);
}





On Thu, Aug 20, 2015 at 2:10 PM, David Blaikie <dblaikie at gmail.com> wrote:

>
>
> On Thu, Aug 20, 2015 at 2:01 PM, Adrian Prantl <aprantl at apple.com> wrote:
>
>>
>> > On Aug 20, 2015, at 12:27 PM, David Blaikie <dblaikie at gmail.com> wrote:
>> >
>> >
>> >
>> > On Thu, Aug 20, 2015 at 11:40 AM, Adrian Prantl <aprantl at apple.com>
>> wrote:
>> >
>> >> On Aug 20, 2015, at 11:34 AM, David Blaikie <dblaikie at gmail.com>
>> wrote:
>> >>
>> >>
>> >>
>> >> On Thu, Aug 20, 2015 at 11:23 AM, Adrian Prantl via llvm-commits <
>> llvm-commits at lists.llvm.org> wrote:
>> >> Author: adrian
>> >> Date: Thu Aug 20 13:23:56 2015
>> >> New Revision: 245588
>> >>
>> >> URL: http://llvm.org/viewvc/llvm-project?rev=245588&view=rev
>> >> Log:
>> >> Fix a debug location handling bug in GVN.
>> >> Caught by the famous "DebugLoc describes the currect SubProgram"
>> assertion.
>> >>
>> >> Huzzah! \o/ ;) (sorry, I know they're a pain to track down, but it's a
>> great way to find debug info quality bugs that would produce really shitty
>> debug info if it weren't for this assertion)
>> >
>> > Oh yes.
>> >
>> >>
>> >> When GVN is removing a nonlocal load it updates the debug location of
>> the
>> >> SSA value it replaced the load with with the one of the load.
>> >>
>> >> I guess I'm not quite following here (knowing next to nothing about
>> GVN) - could you explain this in more detail? What transformation is GVN
>> performing on a call?
>> >
>> > GVN is eliminating the load (and store) to @f and replaces all uses of
>> the load with the value that is stored in @f, which is %call. It then used
>> to transfer the DebugLoc of the load to the value it’s replacing it with
>> (in this case %call). This triggers the assertion because %call happens to
>> be, well, a call.
>> >
>> >>
>> >> In the
>> >> testcase this actually overwrites a valid debug location with an empty
>> one.
>> >>
>> >> In reality GVN has to make an arbitrary choice between two equally
>> valid
>> >> debug locations. This patch changes to behavior to only update the
>> >> location if the value doesn't already have a debug location.
>> >>
>> >> Added:
>> >>     llvm/trunk/test/DebugInfo/gvn.ll
>> >> Modified:
>> >>     llvm/trunk/lib/Transforms/Scalar/GVN.cpp
>> >>
>> >> Modified: llvm/trunk/lib/Transforms/Scalar/GVN.cpp
>> >> URL:
>> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/GVN.cpp?rev=245588&r1=245587&r2=245588&view=diff
>> >>
>> ==============================================================================
>> >> --- llvm/trunk/lib/Transforms/Scalar/GVN.cpp (original)
>> >> +++ llvm/trunk/lib/Transforms/Scalar/GVN.cpp Thu Aug 20 13:23:56 2015
>> >> @@ -1744,7 +1744,8 @@ bool GVN::processNonLocalLoad(LoadInst *
>> >>      if (isa<PHINode>(V))
>> >>        V->takeName(LI);
>> >>      if (Instruction *I = dyn_cast<Instruction>(V))
>> >> -      I->setDebugLoc(LI->getDebugLoc());
>> >> +      if (LI->getDebugLoc())
>> >> +        I->setDebugLoc(LI->getDebugLoc());
>> >>      if (V->getType()->getScalarType()->isPointerTy())
>> >>        MD->invalidateCachedPointerInfo(V);
>> >>      markInstructionForDeletion(LI);
>> >>
>> >> Added: llvm/trunk/test/DebugInfo/gvn.ll
>> >> URL:
>> http://llvm.org/viewvc/llvm-project/llvm/trunk/test/DebugInfo/gvn.ll?rev=245588&view=auto
>> >>
>> ==============================================================================
>> >> --- llvm/trunk/test/DebugInfo/gvn.ll (added)
>> >> +++ llvm/trunk/test/DebugInfo/gvn.ll Thu Aug 20 13:23:56 2015
>> >> @@ -0,0 +1,135 @@
>> >> +; RUN: opt < %s -O2 -gvn -S | FileCheck %s
>> >> +;
>> >> +; Produced at -O2 from:
>> >>
>> >> Do you have the original reproduction -cc1 line, by chance? I'd be
>> curious to see if this test case can be simplified a bit further to make
>> the interesting part(s) more obvious.
>> >
>> > Testcase-golfing! :-)
>> >
>> > It’s very hard to reduce it any more from source, because it needs this
>> specific chain of simplify-cfg and gvn:
>> >   clang -cc1 -triple arm64-apple-ios -emit-llvm -gdwarf-2 -O2 file.c -o
>> -
>> >
>> > Hmm, can't seem to reproduce it with this (& I don't mind just the full
>> command line that crashes clang - rather than having to take it from clang
>> to opt, etc, if you have that) - any ideas? (I'm trying to reproduce at
>> r245521, FWIW)
>>
>> This is the setup I used for delta/creduce:
>>
>> $ cat test.sh
>> set -e
>> clang-3.8 -cc1 -triple arm64-apple-ios -emit-llvm -main-file-name
>> action.c -gdwarf-2 -O2 -x c action.i -o action.ll &>/dev/null
>> FileCheck test.sh <action.ll
>> echo SUCCESS
>> # CHECK: tail call i32 @pid_for_task({{.*\)[^,]+$}}
>>
>
> Ah, sorry, I see - you were checking for the IR, rather than the crash.
> Fair (I was expecting a crash - but that would've require more (unnecessary
> for the issue at hand) code). I'll work with that :)
>
>
>>
>> $ cat action.i
>> struct context {
>>   int cur_pid
>> };
>> int a, b, c, f, d;
>> int pid_for_task(int);
>> sample(struct context *p1)
>> {
>>   if (c)
>>     b = a;
>>   if (a && p1->cur_pid)
>>     sample_internal();
>> }
>> callback() {
>>   f = pid_for_task(d);
>>   sample(&f);
>> }
>>
>> I noticed that when running the test through opt running through opt -gvn
>> alone did not reproduce the error, but running through opt -O2 -gvn did,
>> although none of the passes running before gvn (or after) changed the IR,
>> so it is possible that there is nondeterminism involved.
>> I also tried to add an assertion into GVN and running bugpoint on the IR,
>> but it failed to reduce the testcase any further.
>>
>> -- adrian
>> >
>> >
>> > -- adrian
>> >
>> >> +; struct context {
>> >> +;   int cur_pid
>> >> +; };
>> >> +; int a, b, c, f, d;
>> >> +; int pid_for_task(int);
>> >> +; sample(struct context *p1)
>> >> +; {
>> >> +;   if (c)
>> >> +;     b = a;
>> >> +;   if (a && p1->cur_pid)
>> >> +;     sample_internal();
>> >> +; }
>> >> +; callback() {
>> >> +;   f = pid_for_task(d);
>> >> +;   sample(&f);
>> >> +; }
>> >> +
>> >> +target datalayout = "e-m:o-i64:64-i128:128-n32:64-S128"
>> >> +target triple = "arm64-apple-ios"
>> >> +
>> >> +%struct.context = type { i32 }
>> >> +
>> >> + at c = common global i32 0, align 4
>> >> + at a = common global i32 0, align 4
>> >> + at b = common global i32 0, align 4
>> >> + at d = common global i32 0, align 4
>> >> + at f = common global i32 0, align 4
>> >> +
>> >> +; Function Attrs: nounwind
>> >> +declare i32 @sample_internal(...)
>> >> +
>> >> +; Function Attrs: nounwind
>> >> +define i32 @callback() #0 {
>> >> +entry:
>> >> +  %0 = load i32, i32* @d, align 4, !dbg !37
>> >> +
>> >> +  ; Verify that the call still has a debug location after GVN.
>> >> +  ; CHECK: %call = tail call i32 @pid_for_task(i32 %0) #{{[0-9]}},
>> !dbg
>> >> +  %call = tail call i32 @pid_for_task(i32 %0) #3, !dbg !37
>> >> +
>> >> +  store i32 %call, i32* @f, align 4, !dbg !37
>> >> +  tail call void @llvm.dbg.value(metadata %struct.context* bitcast
>> (i32* @f to %struct.context*), i64 0, metadata !25, metadata !26) #3, !dbg
>> !38
>> >> +  %1 = load i32, i32* @c, align 4, !dbg !40
>> >> +  %tobool.i = icmp eq i32 %1, 0, !dbg !40
>> >> +  %.pr.i = load i32, i32* @a, align 4, !dbg !41
>> >> +  br i1 %tobool.i, label %if.end.i, label %if.then.i, !dbg !42
>> >> +
>> >> +if.then.i:                                        ; preds = %entry
>> >> +  store i32 %.pr.i, i32* @b, align 4, !dbg !43
>> >> +  br label %if.end.i, !dbg !43
>> >> +
>> >> +if.end.i:                                         ; preds =
>> %if.then.i, %entry
>> >> +  %tobool1.i = icmp eq i32 %.pr.i, 0, !dbg !41
>> >> +
>> >> +  ; This instruction has no debug location -- in this
>> >> +  ; particular case it was removed by a bug in SimplifyCFG.
>> >> +  %2 = load i32, i32* @f, align 4
>> >> +
>> >> +  ; GVN is supposed to replace the load of @f with a direct reference
>> to %call.
>> >> +  ; CHECK: %tobool2.i = icmp eq i32 %call, 0, !dbg
>> >> +  %tobool2.i = icmp eq i32 %2, 0, !dbg !41
>> >> +
>> >> +  %or.cond = or i1 %tobool1.i, %tobool2.i, !dbg !41
>> >> +  br i1 %or.cond, label %sample.exit, label %if.then.3.i, !dbg !41
>> >> +
>> >> +if.then.3.i:                                      ; preds = %if.end.i
>> >> +  %call.i = tail call i32 bitcast (i32 (...)* @sample_internal to i32
>> ()*)() #3, !dbg !44
>> >> +  br label %sample.exit, !dbg !44
>> >> +
>> >> +sample.exit:                                      ; preds =
>> %if.end.i, %if.then.3.i
>> >> +  ret i32 undef, !dbg !45
>> >> +}
>> >> +
>> >> +declare i32 @pid_for_task(i32) #1
>> >> +
>> >> +; Function Attrs: nounwind readnone
>> >> +declare void @llvm.dbg.value(metadata, i64, metadata, metadata) #2
>> >> +
>> >> +attributes #0 = { nounwind }
>> >> +attributes #2 = { nounwind readnone }
>> >> +attributes #3 = { nounwind }
>> >> +
>> >> +!llvm.dbg.cu = !{!0}
>> >> +!llvm.module.flags = !{!22, !23}
>> >> +!llvm.ident = !{!24}
>> >> +
>> >> +!0 = distinct !DICompileUnit(language: DW_LANG_C99, file: !1,
>> producer: "clang version 3.8.0 (trunk 244473) (llvm/trunk 244644)",
>> isOptimized: false, runtimeVersion: 0, emissionKind: 1, enums: !2,
>> subprograms: !3, globals: !16)
>> >> +!1 = !DIFile(filename: "test.c", directory: "/")
>> >> +!2 = !{}
>> >> +!3 = !{!4, !13}
>> >> +!4 = !DISubprogram(name: "sample", scope: !5, file: !5, line: 6,
>> type: !6, isLocal: false, isDefinition: true, scopeLine: 7, flags:
>> DIFlagPrototyped, isOptimized: false, variables: !2)
>> >> +!5 = !DIFile(filename: "test.i", directory: "/")
>> >> +!6 = !DISubroutineType(types: !7)
>> >> +!7 = !{!8, !9}
>> >> +!8 = !DIBasicType(name: "int", size: 32, align: 32, encoding:
>> DW_ATE_signed)
>> >> +!9 = !DIDerivedType(tag: DW_TAG_pointer_type, baseType: !10, size:
>> 64, align: 64)
>> >> +!10 = !DICompositeType(tag: DW_TAG_structure_type, name: "context",
>> file: !5, line: 1, size: 32, align: 32, elements: !11)
>> >> +!11 = !{!12}
>> >> +!12 = !DIDerivedType(tag: DW_TAG_member, name: "cur_pid", scope: !10,
>> file: !5, line: 2, baseType: !8, size: 32, align: 32)
>> >> +!13 = !DISubprogram(name: "callback", scope: !5, file: !5, line: 13,
>> type: !14, isLocal: false, isDefinition: true, scopeLine: 13, isOptimized:
>> false, function: i32 ()* @callback, variables: !2)
>> >> +!14 = !DISubroutineType(types: !15)
>> >> +!15 = !{!8}
>> >> +!16 = !{!17, !18, !19, !20, !21}
>> >> +!17 = !DIGlobalVariable(name: "a", scope: !0, file: !5, line: 4,
>> type: !8, isLocal: false, isDefinition: true, variable: i32* @a)
>> >> +!18 = !DIGlobalVariable(name: "b", scope: !0, file: !5, line: 4,
>> type: !8, isLocal: false, isDefinition: true, variable: i32* @b)
>> >> +!19 = !DIGlobalVariable(name: "c", scope: !0, file: !5, line: 4,
>> type: !8, isLocal: false, isDefinition: true, variable: i32* @c)
>> >> +!20 = !DIGlobalVariable(name: "f", scope: !0, file: !5, line: 4,
>> type: !8, isLocal: false, isDefinition: true, variable: i32* @f)
>> >> +!21 = !DIGlobalVariable(name: "d", scope: !0, file: !5, line: 4,
>> type: !8, isLocal: false, isDefinition: true, variable: i32* @d)
>> >> +!22 = !{i32 2, !"Dwarf Version", i32 2}
>> >> +!23 = !{i32 2, !"Debug Info Version", i32 3}
>> >> +!24 = !{!"clang version 3.8.0 (trunk 244473) (llvm/trunk 244644)"}
>> >> +!25 = !DILocalVariable(name: "p1", arg: 1, scope: !4, file: !5, line:
>> 6, type: !9)
>> >> +!26 = !DIExpression()
>> >> +!27 = !DILocation(line: 6, scope: !4)
>> >> +!28 = !DILocation(line: 8, scope: !29)
>> >> +!29 = distinct !DILexicalBlock(scope: !4, file: !5, line: 8)
>> >> +!30 = !DILocation(line: 10, scope: !31)
>> >> +!31 = distinct !DILexicalBlock(scope: !4, file: !5, line: 10)
>> >> +!32 = !DILocation(line: 8, scope: !4)
>> >> +!33 = !DILocation(line: 9, scope: !29)
>> >> +!34 = !DILocation(line: 10, scope: !4)
>> >> +!35 = !DILocation(line: 11, scope: !31)
>> >> +!36 = !DILocation(line: 12, scope: !4)
>> >> +!37 = !DILocation(line: 14, scope: !13)
>> >> +!38 = !DILocation(line: 6, scope: !4, inlinedAt: !39)
>> >> +!39 = distinct !DILocation(line: 15, scope: !13)
>> >> +!40 = !DILocation(line: 8, scope: !29, inlinedAt: !39)
>> >> +!41 = !DILocation(line: 10, scope: !31, inlinedAt: !39)
>> >> +!42 = !DILocation(line: 8, scope: !4, inlinedAt: !39)
>> >> +!43 = !DILocation(line: 9, scope: !29, inlinedAt: !39)
>> >> +!44 = !DILocation(line: 11, scope: !31, inlinedAt: !39)
>> >> +!45 = !DILocation(line: 16, scope: !13)
>> >>
>> >>
>> >> _______________________________________________
>> >> llvm-commits mailing list
>> >> llvm-commits at lists.llvm.org
>> >> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150820/729c7f4d/attachment-0001.html>


More information about the llvm-commits mailing list