[lld] r336631 - Report an error for an extremely large .gdb_index section.
David Blaikie via llvm-commits
llvm-commits at lists.llvm.org
Mon Jul 16 11:43:01 PDT 2018
Might be worth providing instructions (or a disabled test - if it can be
made to be just have too long running time, but not involve massive files)
for how to reproduce the test/failure, even if it's not part of the regular
test suite.
On Mon, Jul 9, 2018 at 6:27 PM Rui Ueyama via llvm-commits <
llvm-commits at lists.llvm.org> wrote:
> Author: ruiu
> Date: Mon Jul 9 18:22:25 2018
> New Revision: 336631
>
> URL: http://llvm.org/viewvc/llvm-project?rev=336631&view=rev
> Log:
> Report an error for an extremely large .gdb_index section.
>
> I believe the only way to test this functionality is to create extremely
> large object files and attempt to create a .gdb_index that is greater
> than 4 GiB. But I think that's too much for most environments and
> buildbots,
> so I'm commiting this without a test that actually triggers the new
> error condition.
>
> Modified:
> lld/trunk/ELF/SyntheticSections.cpp
> lld/trunk/ELF/SyntheticSections.h
>
> Modified: lld/trunk/ELF/SyntheticSections.cpp
> URL:
> http://llvm.org/viewvc/llvm-project/lld/trunk/ELF/SyntheticSections.cpp?rev=336631&r1=336630&r2=336631&view=diff
>
> ==============================================================================
> --- lld/trunk/ELF/SyntheticSections.cpp (original)
> +++ lld/trunk/ELF/SyntheticSections.cpp Mon Jul 9 18:22:25 2018
> @@ -2446,16 +2446,18 @@ GdbIndexSection::GdbIndexSection(std::ve
> SymtabOffset = CuTypesOffset + getAddressAreaSize(Chunks) * 20;
> ConstantPoolOffset = SymtabOffset + GdbSymtab.size() * 8;
>
> - size_t Off = 0;
> for (ArrayRef<uint32_t> Vec : CuVectors) {
> - CuVectorOffsets.push_back(Off);
> - Off += (Vec.size() + 1) * 4;
> + CuVectorOffsets.push_back(CuVectorsPoolSize);
> + CuVectorsPoolSize += (Vec.size() + 1) * 4;
> }
> - StringPoolOffset = ConstantPoolOffset + Off;
> -}
>
> -size_t GdbIndexSection::getSize() const {
> - return StringPoolOffset + StringPoolSize;
> + uint64_t PoolSize = CuVectorsPoolSize + StringPoolSize;
> + TotalSize = ConstantPoolOffset + PoolSize;
> +
> + // Length fields in the .gdb_index section are only 4 byte long,
> + // so the section cannot contain very large contents.
> + if (ConstantPoolOffset > UINT32_MAX || PoolSize > UINT32_MAX)
> + error(".gdb_index section too large");
> }
>
> void GdbIndexSection::writeTo(uint8_t *Buf) {
> @@ -2491,7 +2493,7 @@ void GdbIndexSection::writeTo(uint8_t *B
> // Write the symbol table.
> for (GdbSymbol *Sym : GdbSymtab) {
> if (Sym) {
> - write32le(Buf, Sym->NameOffset + StringPoolOffset -
> ConstantPoolOffset);
> + write32le(Buf, CuVectorsPoolSize + Sym->NameOffset);
> write32le(Buf + 4, CuVectorOffsets[Sym->CuVectorIndex]);
> }
> Buf += 8;
>
> Modified: lld/trunk/ELF/SyntheticSections.h
> URL:
> http://llvm.org/viewvc/llvm-project/lld/trunk/ELF/SyntheticSections.h?rev=336631&r1=336630&r2=336631&view=diff
>
> ==============================================================================
> --- lld/trunk/ELF/SyntheticSections.h (original)
> +++ lld/trunk/ELF/SyntheticSections.h Mon Jul 9 18:22:25 2018
> @@ -688,7 +688,7 @@ class GdbIndexSection final : public Syn
> public:
> GdbIndexSection(std::vector<GdbIndexChunk> &&Chunks);
> void writeTo(uint8_t *Buf) override;
> - size_t getSize() const override;
> + size_t getSize() const override { return TotalSize; }
> bool empty() const override;
>
> private:
> @@ -709,12 +709,13 @@ private:
> // object and used to build different areas of gdb index.
> std::vector<GdbIndexChunk> Chunks;
>
> - static constexpr uint32_t CuListOffset = 24;
> - uint32_t CuTypesOffset;
> - uint32_t SymtabOffset;
> - uint32_t ConstantPoolOffset;
> - uint32_t StringPoolOffset;
> - uint32_t StringPoolSize;
> + uint64_t CuListOffset = 24;
> + uint64_t CuTypesOffset;
> + uint64_t SymtabOffset;
> + uint64_t ConstantPoolOffset;
> + uint64_t CuVectorsPoolSize = 0;
> + uint64_t StringPoolSize;
> + uint64_t TotalSize;
>
> std::vector<size_t> CuVectorOffsets;
> };
>
>
> _______________________________________________
> 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/20180716/7dfef833/attachment.html>
More information about the llvm-commits
mailing list