[llvm] r205529 - Fix for PR 19261:

Eric Christopher echristo at gmail.com
Fri Apr 11 13:57:15 PDT 2014


I'd have thought so, but it seems to be insufficient... the
instructions should be marked with FrameSetup though, would you like
me to commit my patch that did so?

-eric

On Fri, Apr 4, 2014 at 3:31 AM, Daniel Sanders
<Daniel.Sanders at imgtec.com> wrote:
> I've just had a quick look and it appears that the prologue_end marker is appearing at the start of the prologue in this case.
>         # BB#0:                                 # %entry
>                 .loc    1 1 0 prologue_end      # test.c:1:0
>                 addiu   $sp, $sp, -8
>         $tmp0:
>
> I'm not familiar with the debug info code but I think the problem is that we don't set the MachineInstr::FrameSetup on any instruction in the prologue. Does that sound right to you?
>
>> -----Original Message-----
>> From: Eric Christopher [mailto:echristo at gmail.com]
>> Sent: 03 April 2014 18:54
>> To: Daniel Sanders
>> Cc: llvm-commits at cs.uiuc.edu
>> Subject: Re: [llvm] r205529 - Fix for PR 19261:
>>
>> I've relaxed the testcase for now and I think it works for mips (I tested it with
>> -march=mips on llc by hand). I'll keep an eye out.
>>
>> FWIW there's something odd going on for mips here. It has two line table
>> entries for address 0. Probably worth taking a look at some point.
>>
>> -eric
>>
>> On Thu, Apr 3, 2014 at 10:08 AM, Eric Christopher <echristo at gmail.com>
>> wrote:
>> > I can see it on a couple of the others as well. I'll take a look. I
>> > think the testcase may just need to be relaxed.
>> >
>> > On Apr 3, 2014 7:03 AM, "Daniel Sanders" <Daniel.Sanders at imgtec.com>
>> wrote:
>> >>
>> >> Hi Eric,
>> >>
>> >> I think this commit caused llvm-mips-linux buildbot to turn red and I
>> >> don't think it has sent an email (I'm also on the blamelist and I
>> >> haven't received one).
>> >> The build is
>> >> http://lab.llvm.org:8011/builders/llvm-mips-linux/builds/8028
>> >>
>> >> I've re-run the test on an X86 with -march=mips added to the llc
>> >> command and it seems that an extra entry appears in the line
>> >> information between the CHECK and the CHECK-NEXT:
>> >> Address            Line   Column File   ISA Discriminator Flags
>> >> ------------------ ------ ------ ------ --- ------------- -------------
>> >> 0x0000000000000000      1      0      1   0             0  is_stmt
>> >> 0x0000000000000000      1      0      1   0             0  is_stmt
>> >> prologue_end   <--- extra line
>> >> 0x0000000000000010      2      0      1   0             0  is_stmt
>> >> 0x0000000000000018      4      0      1   0             0  is_stmt
>> >> 0x0000000000000020      6      0      1   0             0  is_stmt
>> >> 0x0000000000000034      6      0      1   0             0  is_stmt
>> >> end_sequence
>> >>
>> >> > -----Original Message-----
>> >> > From: llvm-commits-bounces at cs.uiuc.edu [mailto:llvm-commits-
>> >> > bounces at cs.uiuc.edu] On Behalf Of Eric Christopher
>> >> > Sent: 03 April 2014 13:12
>> >> > To: llvm-commits at cs.uiuc.edu
>> >> > Subject: [llvm] r205529 - Fix for PR 19261:
>> >> >
>> >> > Author: echristo
>> >> > Date: Thu Apr  3 07:11:51 2014
>> >> > New Revision: 205529
>> >> >
>> >> > URL: http://llvm.org/viewvc/llvm-project?rev=205529&view=rev
>> >> > Log:
>> >> > Fix for PR 19261:
>> >> >
>> >> > llc doesn't generate nodes for unconditional fall-through branches
>> >> > for targets without FastISel implementation (X86 has it, but can be
>> >> > disabled by
>> >> > "-fast-isel=false") in SelectionDAGBuilder::visitBr().
>> >> >
>> >> > So for line 4 in the following testcase
>> >> >
>> >> > 1: void foo(int i){
>> >> > 2:   switch(i){
>> >> > 3:   default:
>> >> > 4:     break;
>> >> > 5:   }
>> >> > 6:   return;
>> >> > 7: }
>> >> >
>> >> > there is no corresponding line in .debug_line section, and a
>> >> > debugger cannot set a breakpoint at line 4.
>> >> >
>> >> > Fix this by always emitting a branch when we're not optimizing and
>> >> > add a testcase to ensure that there's code on every line we'd want to
>> break.
>> >> >
>> >> > Patch by Daniil Fukalov.
>> >> >
>> >> > Added:
>> >> >     llvm/trunk/test/DebugInfo/unconditional-branch.ll
>> >> > Modified:
>> >> >     llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
>> >> >
>> >> > Modified:
>> >> > llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
>> >> > URL: http://llvm.org/viewvc/llvm-
>> >> >
>> project/llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
>> >> > ?re v=205529&r1=205528&r2=205529&view=diff
>> >> >
>> ==========================================================
>> >> > ====================
>> >> > --- llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
>> >> > (original)
>> >> > +++ llvm/trunk/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
>> Thu
>> >> > Apr
>> >> > +++ 3 07:11:51 2014
>> >> > @@ -1571,8 +1571,9 @@ void SelectionDAGBuilder::visitBr(const
>> >> >      // Update machine-CFG edges.
>> >> >      BrMBB->addSuccessor(Succ0MBB);
>> >> >
>> >> > -    // If this is not a fall-through branch, emit the branch.
>> >> > -    if (Succ0MBB != NextBlock)
>> >> > +    // If this is not a fall-through branch or optimizations are
>> >> > switched off,
>> >> > +    // emit the branch.
>> >> > +    if (Succ0MBB != NextBlock || TM.getOptLevel() ==
>> >> > + CodeGenOpt::None)
>> >> >        DAG.setRoot(DAG.getNode(ISD::BR, getCurSDLoc(),
>> >> >                                MVT::Other, getControlRoot(),
>> >> >                                DAG.getBasicBlock(Succ0MBB)));
>> >> >
>> >> > Added: llvm/trunk/test/DebugInfo/unconditional-branch.ll
>> >> > URL: http://llvm.org/viewvc/llvm-
>> >> > project/llvm/trunk/test/DebugInfo/unconditional-
>> >> > branch.ll?rev=205529&view=auto
>> >> >
>> ==========================================================
>> >> > ====================
>> >> > --- llvm/trunk/test/DebugInfo/unconditional-branch.ll (added)
>> >> > +++ llvm/trunk/test/DebugInfo/unconditional-branch.ll Thu Apr  3
>> >> > +++ 07:11:51 2014
>> >> > @@ -0,0 +1,64 @@
>> >> > +; REQUIRES: object-emission
>> >> > +; PR 19261
>> >> > +
>> >> > +; RUN: %llc_dwarf -fast-isel=false -O0 -filetype=obj < %s > %t ; RUN:
>> >> > +llvm-dwarfdump %t | FileCheck %s
>> >> > +
>> >> > +; CHECK: {{0x[0-9a-f]+}}      1      0      1   0             0
>> >> > is_stmt
>> >> > +; CHECK-NEXT: {{0x[0-9a-f]+}}      2      0      1   0             0
>> >> > is_stmt
>> >> > +; CHECK-NEXT: {{0x[0-9a-f]+}}      4      0      1   0             0
>> >> > is_stmt
>> >> > +
>> >> > +; IR generated from clang -O0 -g with the following source:
>> >> > +;void foo(int i){
>> >> > +;  switch(i){
>> >> > +;  default:
>> >> > +;    break;
>> >> > +;  }
>> >> > +;  return;
>> >> > +;}
>> >> > +
>> >> > +; Function Attrs: nounwind
>> >> > +define void @foo(i32 %i) #0 {
>> >> > +entry:
>> >> > +  %i.addr = alloca i32, align 4
>> >> > +  store i32 %i, i32* %i.addr, align 4
>> >> > +  call void @llvm.dbg.declare(metadata !{i32* %i.addr}, metadata
>> >> > +!12), !dbg !13
>> >> > +  %0 = load i32* %i.addr, align 4, !dbg !14
>> >> > +  switch i32 %0, label %sw.default [
>> >> > +  ], !dbg !14
>> >> > +
>> >> > +sw.default:                                       ; preds = %entry
>> >> > +  br label %sw.epilog, !dbg !15
>> >> > +
>> >> > +sw.epilog:                                        ; preds = %sw.default
>> >> > +  ret void, !dbg !17
>> >> > +}
>> >> > +
>> >> > +; Function Attrs: nounwind readnone declare void
>> >> > + at llvm.dbg.declare(metadata, metadata) #1
>> >> > +
>> >> > +attributes #0 = { nounwind "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 = { nounwind readnone }
>> >> > +
>> >> > +!llvm.dbg.cu = !{!0}
>> >> > +!llvm.module.flags = !{!9, !10}
>> >> > +!llvm.ident = !{!11}
>> >> > +
>> >> > +!0 = metadata !{i32 786449, metadata !1, i32 12, metadata !"clang
>> >> > +version 3.5.0 (204712)", i1 false, metadata !"", i32 0, metadata
>> >> > +!2, metadata !2, metadata !3, metadata !2, metadata !2, metadata
>> >> > +!"", i32 1} ; [ DW_TAG_compile_unit ] [D:\work\EPRs\396363/test.c]
>> >> > [DW_LANG_C99]
>> >> > +!1 = metadata !{metadata !"test.c", metadata
>> >> > +!"D:\5Cwork\5CEPRs\5C396363"}
>> >> > +!2 = metadata !{}
>> >> > +!3 = metadata !{metadata !4}
>> >> > +!4 = metadata !{i32 786478, metadata !1, metadata !5, metadata
>> >> > +!"foo",
>> >> > metadata !"foo", metadata !"", i32 1, metadata !6, i1 false, i1
>> >> > true,
>> >> > i32 0, i32 0,
>> >> > null, i32 256, i1 false, void (i32)* @foo, null, null, metadata !2,
>> >> > i32 1} ; [ DW_TAG_subprogram ] [line 1] [def] [foo]
>> >> > +!5 = metadata !{i32 786473, metadata !1}          ; [ DW_TAG_file_type
>> >> > ]
>> >> > [D:\work\EPRs\396363/test.c]
>> >> > +!6 = metadata !{i32 786453, i32 0, null, metadata !"", i32 0, i64
>> >> > +0,
>> >> > +i64 0, i64 0, i32 0, null, metadata !7, i32 0, null, null, null} ;
>> >> > +[ DW_TAG_subroutine_type ] [line 0, size 0, align 0, offset 0]
>> >> > +[from ]
>> >> > +!7 = metadata !{null, metadata !8}
>> >> > +!8 = metadata !{i32 786468, null, null, metadata !"int", i32 0,
>> >> > +i64 32,
>> >> > +i64 32, i64 0, i32 0, i32 5} ; [ DW_TAG_base_type ] [int] [line 0,
>> >> > +size 32, align 32, offset 0, enc DW_ATE_signed]
>> >> > +!9 = metadata !{i32 2, metadata !"Dwarf Version", i32 4}
>> >> > +!10 = metadata !{i32 1, metadata !"Debug Info Version", i32 1}
>> >> > +!11 = metadata !{metadata !"clang version 3.5.0 (204712)"}
>> >> > +!12 = metadata !{i32 786689, metadata !4, metadata !"i", metadata
>> >> > +!5,
>> >> > +i32 16777217, metadata !8, i32 0, i32 0} ; [ DW_TAG_arg_variable ]
>> >> > +[i] [line 1]
>> >> > +!13 = metadata !{i32 1, i32 0, metadata !4, null}
>> >> > +!14 = metadata !{i32 2, i32 0, metadata !4, null}
>> >> > +!15 = metadata !{i32 4, i32 0, metadata !16, null}
>> >> > +!16 = metadata !{i32 786443, metadata !1, metadata !4, i32 2, i32
>> >> > +0,
>> >> > +i32 0, i32 0} ; [ DW_TAG_lexical_block ]
>> >> > +[D:\work\EPRs\396363/test.c]
>> >> > +!17 = metadata !{i32 6, i32 0, metadata !4, null}
>> >> >
>> >> >
>> >> > _______________________________________________
>> >> > llvm-commits mailing list
>> >> > llvm-commits at cs.uiuc.edu
>> >> > http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits



More information about the llvm-commits mailing list