Re: [PATCH dwarves v2] btf_encoder: sanitize non-regular int base type

[Sedat Dilek <sedat.dilek@gmail.com>]

On Sun, Feb 7, 2021 at 3:18 PM Mark Wielaard <mark@klomp.org> wrote:
>
> Hi,
>
> On Sat, 2021-02-06 at 23:17 -0800, Yonghong Song wrote:
> > clang with dwarf5 may generate non-regular int base type,
> > i.e., not a signed/unsigned char/short/int/longlong/__int128.
> > Such base types are often used to describe
> > how an actual parameter or variable is generated. For example,
> >
> > 0x000015cf:   DW_TAG_base_type
> >                 DW_AT_name      ("DW_ATE_unsigned_1")
> >                 DW_AT_encoding  (DW_ATE_unsigned)
> >                 DW_AT_byte_size (0x00)
> >
> > 0x00010ed9:         DW_TAG_formal_parameter
> >                       DW_AT_location    (DW_OP_lit0,
> >                                          DW_OP_not,
> >                                          DW_OP_convert (0x000015cf) "DW_ATE_unsigned_1",
> >                                          DW_OP_convert (0x000015d4) "DW_ATE_unsigned_8",
> >                                          DW_OP_stack_value)
> >                       DW_AT_abstract_origin     (0x00013984 "branch")
> >
> > What it does is with a literal "0", did a "not" operation, and the converted to
> > one-bit unsigned int and then 8-bit unsigned int.
>
> Thanks for tracking this down. Do you have any idea why the clang
> compiler emits this? You might be right that it is intended to do what
> you describe it does (but then it would simply encode an unsigned
> constant 1 char in a very inefficient way). But as implemented it
> doesn't seem to make any sense. What would DW_OP_convert of an zero
> sized base type even mean (if it is intended as a 1 bit sized typed,
> then why is there no DW_AT_bit_size)?
>
> So I do think your patch makes sense. clang clearly is emitting
> something bogus. And so some fixup is needed. But maybe we should at
> least give a warning about it, otherwise it might never get fixed.
>
> BTW. If these bogus base types are only emitted as part of a location
> expression and not as part of an actual function or variable type
> description, then why are we even trying to encode it as a BTF type? It
> might be cheaper to just skip/drop it. But maybe the code setup makes
> it hard to know whether or not such a (bogus) type is actually
> referenced from a function or variable description?
>

As said this is with LLVM/Clang v12.0.0-rc1 and `make LLVM=1
LLVM_IAS=1` means use all LLVM tools (do not use GNU/binutils like
GNU/ld BFD) and Clang's Integrated ASsembler (means do not use GNU
AS).

When doing an indexed search for "DW_ATE_unsigned"...

...this points to changes recently done in places like
llvm/lib/CodeGen/AsmPrinter/.

I see 16 changes there touching DWARF-x area since llvmorg-12.0.0-rc1 release:

$ git log --oneline llvmorg-12.0.0-rc1.. llvm/lib/CodeGen/AsmPrinter/
e44a10094283 .gcc_except_table: Set SHF_LINK_ORDER if binutils>=2.36,
and drop unneeded unique ID for -fno-unique-section-names
853a2649160c [AsmPrinter] __patchable_function_entries: Set
SHF_LINK_ORDER for binutils 2.36 and above
e3c0b0fe0958 [WebAssembly] locals can now be indirect in DWARF
34f3249abdff [DebugInfo] Fix error from D95893, where I accidentally
used an unsigned int in a loop and it wraps around.
a740af4de970 [CodeView][DebugInfo] Update the code for removing
template arguments from the display name of a codeview function id.
56fa34ae3570 DebugInfo: Temporarily work around -gsplit-dwarf + LTO
.debug_gnu_pubnames regression after D94976
8998f5843503 Re-land D94976 after revert in e29552c5aff6
d32deaab4d53 Revert "[DWARF] Location-less inlined variables should
not have DW_TAG_variable"
ddc2f1e3fb4f [DWARF] Location-less inlined variables should not have
DW_TAG_variable
511c9a76fb98 [AsmPrinter] Use ListSeparator (NFC)
85b7b5625a00 Fix memory leak in 4318028cd2d7633a0cdeb0b5d4d2ed81fab87864
4318028cd2d7 DebugInfo: Add a DWARF FORM extension for addrx+offset
references to reduce relocations
e29552c5aff6 Revert "[DWARF] Create subprogram's DIE in DISubprogram's unit"
dd7297e1bffe DebugInfo: Fix bug in addr+offset exprloc to use DWARFv5
addrx op instead of DWARFv4 GNU extension
7e6c87ee0454 DebugInfo: Deduplicate addresses in debug_addr
ef0dcb506300 [DWARF] Create subprogram's DIE in DISubprogram's unit

What I try to say is with LLVM-13 (git) this might look different?

Here, I have a LLVM-12 ThinLTO+PGO optimized toolchain which saves 1/3
of Linux-kernel build-time.
So, I do not want to switch to Debian's or packages from <apt.llvm.org>.
These binaries take much much longer and I do not know if I get some
new issues with Linux v5.11-rc6+.

Again, I am not a LLVM toolchain expert.
Best is to ask on llvm-dev mailing-list?

Thanks.

- Sedat -



[1] https://github.com/llvm/llvm-project/search?o=desc&q=DW_ATE_unsigned&s=indexed