[llvm] r315279 - Ignore all duplicate frame index expression
Reid Kleckner via llvm-commits
llvm-commits at lists.llvm.org
Thu Oct 12 14:16:52 PDT 2017
Thanks, I was able to reproduce this and am getting a test case soon. I
haven't checked if it's fragment related yet, but I suspect not since it's
-O0.
On Thu, Oct 12, 2017 at 9:52 AM, Björn Steinbrink <bsteinbr at gmail.com>
wrote:
> Can you check what kind of conflict is this? Two non-fragment
> expressions, or a non-fragment expression conflicting with a fragment
> expression?
>
> 2017-10-12 15:25 GMT+02:00 Daniel Jasper <djasper at google.com>:
> > With this change, we are actually running into the assertion "conflicting
> > locations for variable". I have re-instantiated the previous detection
> logic
> > in r315576.
> > What would lead to such two variables conflicting?
> > Reid, can you help generate a test case? I'll forward repro instructions
> of
> > internal tests to you.
> >
> > On Tue, Oct 10, 2017 at 8:46 AM, Bjorn Steinbrink via llvm-commits
> > <llvm-commits at lists.llvm.org> wrote:
> >>
> >> Author: bsteinbr
> >> Date: Tue Oct 10 00:46:17 2017
> >> New Revision: 315279
> >>
> >> URL: http://llvm.org/viewvc/llvm-project?rev=315279&view=rev
> >> Log:
> >> Ignore all duplicate frame index expression
> >>
> >> Some passes might duplicate calls to llvm.dbg.declare creating
> >> duplicate frame index expression which currently trigger an assertion
> >> which is meant to catch erroneous, overlapping fragment declarations.
> >> But identical frame index expressions are just redundant and don't
> >> actually conflict with each other, so we can be more lenient and just
> >> ignore the duplicates.
> >>
> >> Reviewers: aprantl, rnk
> >>
> >> Subscribers: llvm-commits, JDevlieghere
> >>
> >> Differential Revision: https://reviews.llvm.org/D38540
> >>
> >> Modified:
> >> llvm/trunk/lib/CodeGen/AsmPrinter/DwarfDebug.cpp
> >> llvm/trunk/lib/CodeGen/AsmPrinter/DwarfDebug.h
> >>
> >> Modified: llvm/trunk/lib/CodeGen/AsmPrinter/DwarfDebug.cpp
> >> URL:
> >> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/
> CodeGen/AsmPrinter/DwarfDebug.cpp?rev=315279&r1=315278&r2=315279&view=diff
> >>
> >> ============================================================
> ==================
> >> --- llvm/trunk/lib/CodeGen/AsmPrinter/DwarfDebug.cpp (original)
> >> +++ llvm/trunk/lib/CodeGen/AsmPrinter/DwarfDebug.cpp Tue Oct 10
> 00:46:17
> >> 2017
> >> @@ -220,9 +220,34 @@ ArrayRef<DbgVariable::FrameIndexExpr> Db
> >> return A.Expr->getFragmentInfo()->OffsetInBits <
> >> B.Expr->getFragmentInfo()->OffsetInBits;
> >> });
> >> +
> >> return FrameIndexExprs;
> >> }
> >>
> >> +void DbgVariable::addMMIEntry(const DbgVariable &V) {
> >> + assert(DebugLocListIndex == ~0U && !MInsn && "not an MMI entry");
> >> + assert(V.DebugLocListIndex == ~0U && !V.MInsn && "not an MMI entry");
> >> + assert(V.Var == Var && "conflicting variable");
> >> + assert(V.IA == IA && "conflicting inlined-at location");
> >> +
> >> + assert(!FrameIndexExprs.empty() && "Expected an MMI entry");
> >> + assert(!V.FrameIndexExprs.empty() && "Expected an MMI entry");
> >> +
> >> + for (const auto &FIE : V.FrameIndexExprs)
> >> + // Ignore duplicate entries.
> >> + if (llvm::none_of(FrameIndexExprs, [&](const FrameIndexExpr
> &Other) {
> >> + return FIE.FI == Other.FI && FIE.Expr == Other.Expr;
> >> + }))
> >> + FrameIndexExprs.push_back(FIE);
> >> +
> >> + assert((FrameIndexExprs.size() == 1 ||
> >> + llvm::all_of(FrameIndexExprs,
> >> + [](FrameIndexExpr &FIE) {
> >> + return FIE.Expr && FIE.Expr->isFragment();
> >> + })) &&
> >> + "conflicting locations for variable");
> >> +}
> >> +
> >> static const DwarfAccelTable::Atom TypeAtoms[] = {
> >> DwarfAccelTable::Atom(dwarf::DW_ATOM_die_offset,
> >> dwarf::DW_FORM_data4),
> >> DwarfAccelTable::Atom(dwarf::DW_ATOM_die_tag,
> dwarf::DW_FORM_data2),
> >>
> >> Modified: llvm/trunk/lib/CodeGen/AsmPrinter/DwarfDebug.h
> >> URL:
> >> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/
> CodeGen/AsmPrinter/DwarfDebug.h?rev=315279&r1=315278&r2=315279&view=diff
> >>
> >> ============================================================
> ==================
> >> --- llvm/trunk/lib/CodeGen/AsmPrinter/DwarfDebug.h (original)
> >> +++ llvm/trunk/lib/CodeGen/AsmPrinter/DwarfDebug.h Tue Oct 10 00:46:17
> >> 2017
> >> @@ -138,30 +138,7 @@ public:
> >> /// Get the FI entries, sorted by fragment offset.
> >> ArrayRef<FrameIndexExpr> getFrameIndexExprs() const;
> >> bool hasFrameIndexExprs() const { return !FrameIndexExprs.empty(); }
> >> -
> >> - void addMMIEntry(const DbgVariable &V) {
> >> - assert(DebugLocListIndex == ~0U && !MInsn && "not an MMI entry");
> >> - assert(V.DebugLocListIndex == ~0U && !V.MInsn && "not an MMI
> entry");
> >> - assert(V.Var == Var && "conflicting variable");
> >> - assert(V.IA == IA && "conflicting inlined-at location");
> >> -
> >> - assert(!FrameIndexExprs.empty() && "Expected an MMI entry");
> >> - assert(!V.FrameIndexExprs.empty() && "Expected an MMI entry");
> >> -
> >> - if (FrameIndexExprs.size()) {
> >> - auto *Expr = FrameIndexExprs.back().Expr;
> >> - // Get rid of duplicate non-fragment entries. More than one
> >> non-fragment
> >> - // dbg.declare makes no sense so ignore all but the first.
> >> - if (!Expr || !Expr->isFragment())
> >> - return;
> >> - }
> >> - FrameIndexExprs.append(V.FrameIndexExprs.begin(),
> >> V.FrameIndexExprs.end());
> >> - assert(llvm::all_of(FrameIndexExprs,
> >> - [](FrameIndexExpr &FIE) {
> >> - return FIE.Expr && FIE.Expr->isFragment();
> >> - }) &&
> >> - "conflicting locations for variable");
> >> - }
> >> + void addMMIEntry(const DbgVariable &V);
> >>
> >> // Translate tag to proper Dwarf tag.
> >> dwarf::Tag getTag() const {
> >>
> >>
> >> _______________________________________________
> >> 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/20171012/9791dfb1/attachment.html>
More information about the llvm-commits
mailing list