RFC: Debug info for Cuda

Hi everybody,

As you know, Cuda/NVPTX target has very limited support of the debug info in Clang/LLVM. Currently, LLVM supports only emission of the line numbers debug info.

This is caused by limitations of the Cuda/NVPTX codegen. Clang/LLVM translates the source code to LLVM IR, which is then lowered to PTX (parallel thread execution) intermediate file. This PTX file represents special kind of the assembler code in text format, which contains the code itself + (possibly) debug info. Then this PTX file is compiled by ptxas tool into the CUDA binary representation.

Debug info representation in PTX file.

Hi everybody,
As you know, Cuda/NVPTX target has very limited support of the debug
info in Clang/LLVM. Currently, LLVM supports only emission of the line
numbers debug info.
This is caused by limitations of the Cuda/NVPTX codegen. Clang/LLVM
translates the source code to LLVM IR, which is then lowered to PTX
(parallel thread execution) intermediate file. This PTX file represents
special kind of the assembler code in text format, which contains the
code itself + (possibly) debug info. Then this PTX file is compiled by
ptxas tool into the CUDA binary representation.

Debug info representation in PTX file.

According to PTX Writer's Guide to Interoperability, Debug information
(PTX Interoperability :: CUDA Toolkit Documentation)
, debug information must be encoded in DWARF (Debug With Arbitrary
Record Format). The responsibility for generating debug information is
split between the PTX producer and the PTX-to-SASS backend. The PTX
producer is responsible for emitting binary DWARF into the PTX file,
using the .section and .b8-.b16-.b32-and-.b64 directives in PTX. This
should contain the .debug_info and .debug_abbrev sections, and possibly
optional sections .debug_pubnames and .debug_aranges. These sections
are standard DWARF2 sections that refer to labels and registers in the
PTX.

The PTX-to-SASS backend is responsible for generating the .debug_line
section from the .file and .loc directives in the PTX file. This
section maps source lines to SASS addresses. The PTX-to-SASS backend
also generates the .debug_frame section.

All this sounds like the standard division of responsibilities between
an LLVM code generator and the assembler.

LLVM is able to emit debug info in DWARF. But ptxas compiler has some
limitations, that make it hard to adapt LLVM for correct emission of
the debug info in PTX files.

Limitations/features of the PTX format/ptxas compiler.

a) Supports DWARF-2 only.

IIRC, Darwin had a similar restriction until recently.

b) Labels are allowed only in code section (only in functions).

If you have static/global variables, I guess their locations would
have to be described using a section+offset expression? Normally
we emit a location attribute that is just a reference to a label
for the variable.

c) Does not support label arithmetic in DWARF sections.
“.b32 L1 – L2” as the size of the section is not allowed, so the
sections sizes should be calculated explicitly.

MachO has a similar restriction, this should not be a problem if you
can do something like:
    L3 = L1 - L2
    .b32 L3

d) Debug info must point to the sections, not to labels inside these
sections.
“.b32 .debug_abbrevs”

Offhand for DWARF-2 I can't think of a reference that couldn't be done
this way.

e) Sections itself must be enclosed into braces
“.section .debug_info {…}”
f) Frame info is non-register based
Based on function local “__local_depot” array, that represents the
stack frame.
g) All variables must have non-standard DW_AT_address_class attribute
so the debuger had the info about address class of the variable -
global or local. DWARF standard does support this attribute, but it can
be appiled to pointer/reference types only, not variables.

For variables it would be more usual to use DW_AT_segment for this.
But that's an agreement that the compiler and debugger need to reach.

h) The first label in the function must follow the debug location macro.
In LLVM, it is followed by the debug location macro.

I am not 100% sure what you mean by this, but I think it has to do with
the fact that LLVM attaches locations to instructions, not labels. It
might or might not be easy to work around this; there might be an
unfortunate interaction with how emitting line-0 records works.

i) .debug_frame section is emitted by txas compiler.
DW_AT_frame_base must be set to dwarf::DW_FORM_data1
dwarf::DW_OP_call_frame_cfa value.

I doubt that's a problem.

j) Strings cannot be referenced by the labels, instead they must be
inlined in the sections in form of array of chars.

LLVM used to do inline strings, but switched to the .debug_str section
quite a while ago. On the other hand, I spent a little time maybe a
year ago looking into whether we could emit short strings inline as a
space-saving measure, and decided it was feasible. (I didn't do it
because the space savings was really trivial.) So I think doing this
would not be terribly hard.

Some changes in LLVM are required to support all these
limitation/features in the output PTX files.
Required changes in LLVM.

•include/llvm/CodeGen/AsmPrinter.h.
•Add “virtual MCSymbol *getFunctionFrameSymbol(const
MachineFunction *MF) const” for non-register-based frame info.
•Override “NVPTXMCAsmPrinter.cpp” to return the name of the
“__local_depot” frame storage.
•Add ”cuda-gdb” specific tuning.

Note that our design philosophy for "tuning" is that a tuning option
unpacks into other separate flags. Not a problem, just an observation.

•Inlined strings must be used in sections, not string references\.
•Label arithmetic is replaced by the absolute section size

evaluation.

This one isn't a debug-info tuning decision, it's how your assembler works
and so is a target decision.

•Use “AsmPrinter::doInitialization\(\)” instead of NVPTX\-custom manual

initialization.
•Local variables address emitted as “__local_depot” + <var offset>.
•Add NVPTX specific “NVPTXMCAsmStreamer” class.
•Requires moving to includes of “MCAsmStreamer” class declaration.
•Overrides emission of the labels (names of the section are emitted
instead).
•Overrides emission of the sections (emit braces)
•Overrides string emission (as sequence of bytes, not as strings)
•Overrides emission of files/locations debug info
Required changes in Clang.

•Add option “-gcuda-gdb” to driver.
•Emit cuda-gdb compatible debug info (DWARF-2 by default + CudaGDB
tuning).
•Add options “-g --dont-merge-basicblocks --return-at-end” to “ptxas”
call.
•ptxas is able to translate debug information only if -O0
optimization level is used. It means, that we can use optimization
level in LLVM > O0, but still have to use O0 when calling ptxas
compiler.

This approach was implemented in https://github.com/clang-ykt to support
debug info emission for NVPTX target when generating code for OpenMP
offloading constructs. You can try to use it.

I haven't looked at your code but all the things you describe seem
reasonably feasible. Certainly the details of what you want to do
to the emitted DWARF are fine; I am less sure about the assembler
details, but if you have a worked example that makes it likely that
part is okay as well.
--paulr

06.11.2017 14:56, Robinson, Paul пишет:

Hi everybody,
As you know, Cuda/NVPTX target has very limited support of the debug
info in Clang/LLVM. Currently, LLVM supports only emission of the line
numbers debug info.
This is caused by limitations of the Cuda/NVPTX codegen. Clang/LLVM
translates the source code to LLVM IR, which is then lowered to PTX
(parallel thread execution) intermediate file. This PTX file represents
special kind of the assembler code in text format, which contains the
code itself + (possibly) debug info. Then this PTX file is compiled by
ptxas tool into the CUDA binary representation.

Debug info representation in PTX file.

According to PTX Writer's Guide to Interoperability, Debug information
(PTX Interoperability :: CUDA Toolkit Documentation)
, debug information must be encoded in DWARF (Debug With Arbitrary
Record Format). The responsibility for generating debug information is
split between the PTX producer and the PTX-to-SASS backend. The PTX
producer is responsible for emitting binary DWARF into the PTX file,
using the .section and .b8-.b16-.b32-and-.b64 directives in PTX. This
should contain the .debug_info and .debug_abbrev sections, and possibly
optional sections .debug_pubnames and .debug_aranges. These sections
are standard DWARF2 sections that refer to labels and registers in the
PTX.

The PTX-to-SASS backend is responsible for generating the .debug_line
section from the .file and .loc directives in the PTX file. This
section maps source lines to SASS addresses. The PTX-to-SASS backend
also generates the .debug_frame section.

All this sounds like the standard division of responsibilities between
an LLVM code generator and the assembler.

LLVM is able to emit debug info in DWARF. But ptxas compiler has some
limitations, that make it hard to adapt LLVM for correct emission of
the debug info in PTX files.

Limitations/features of the PTX format/ptxas compiler.

a) Supports DWARF-2 only.

IIRC, Darwin had a similar restriction until recently.

b) Labels are allowed only in code section (only in functions).

If you have static/global variables, I guess their locations would
have to be described using a section+offset expression? Normally
we emit a location attribute that is just a reference to a label
for the variable.

c) Does not support label arithmetic in DWARF sections.
“.b32 L1 – L2” as the size of the section is not allowed, so the
sections sizes should be calculated explicitly.

MachO has a similar restriction, this should not be a problem if you
can do something like:
    L3 = L1 - L2
    .b32 L3

Nope, it is not supported

(Reply-all’ing this time.)

This all seems pretty reasonable to me, although I guess as we say, the proof is in the patches. :slight_smile:

Hi, Alexey,

Thanks a bunch for working on this. A couple quick questions below…

Is this a change? It looks like this is already the behavior of the driver’s CUDA target code: Can you clarify what “unable to translate” mean? Does it refuse to compile the code, drop all debug info, drop all debug info except for line-table information, something else? Thanks again, Hal

Yes, but we need add the same behavior for OpenMP. It refuses to compile the code and emits error message.

I don’t understand the use case and reasons to blame PTXAS compiler here.

>>a) Supports DWARF-2 only.

What would you like to achieve with DWARF-3+ that you cannot do with DWARF2?

>> b) Labels are allowed only in code section (only in functions).
What is the use case here which needs labels outside functions?

>>>c) Does not support label arithmetic in DWARF sections.
Same. Please explain use case.

d) Debug info must point to the sections, not to labels inside these sections.

e) Sections itself must be enclosed into braces

> “.section .debug_info {…}”

Again, why is this a limitation?

>>>> i) .debug_frame section is emitted by txas compiler.
> DW_AT_frame_base must be set to dwarf::DW_FORM_data1
> dwarf::DW_OP_call_frame_cfa value.
I doubt that’s a problem.

Why is this a problem?

Nobody blames ptxas. I’m not saying that these are the troubles, I’m just saying that it has some features and we have some problems to be solved.
But lack of labels, label arithmetics in DWARF sections is the real problem, because LLVM actively uses it in DWARF sections

Best regards,
Alexey Bataev

But lack of labels, label arithmetics in DWARF sections is the real problem, because LLVM actively uses it in DWARF sections

Can you please elaborate more on this, how would it help?
May be an example would help.

Sure, for example it used to evaluate the size of debug section like this:
.section xxxx {
Lsection_start:
.b32 Lsection_end-Lsection_start
Lsection_end:
}

Labels are used to reference strings.
Labels are used to reference non-contiguous address ranges

08.11.2017 7:00, Madhur Amilkanthwar пишет:

But lack of labels, label arithmetics in DWARF sections is the real problem, because LLVM actively uses it in DWARF sections

Can you please elaborate more on this, how would it help?
May be an example would help.

Nobody blames ptxas. I'm not saying that these are the troubles, I'm just saying that it has some features and we have some problems to be solved.
But lack of labels, label arithmetics in DWARF sections is the real problem, because LLVM actively uses it in DWARF sections

Best regards,
Alexey Bataev

8 нояб. 2017 г., в 5:35, Madhur Amilkanthwar <madhur13490@gmail.com<mailto:madhur13490@gmail.com>> написал(а):

I don't understand the use case and reasons to blame PTXAS compiler here.

a) Supports DWARF-2 only.

What would you like to achieve with DWARF-3+ that you cannot do with DWARF2?

b) Labels are allowed only in code section (only in functions).

What is the use case here which needs labels outside functions?

c) Does not support label arithmetic in DWARF sections.

Same. Please explain use case.

d) Debug info must point to the sections, not to labels inside these sections.
e) Sections itself must be enclosed into braces

    “.section .debug_info {…}”

Again, why is this a limitation?

i) .debug_frame section is emitted by txas compiler.

    DW_AT_frame_base must be set to dwarf::DW_FORM_data1
dwarf::DW_OP_call_frame_cfa value.

I doubt that's a problem.

Why is this a problem?

06.11.2017 14:56, Robinson, Paul пишет:

Hi everybody,
As you know, Cuda/NVPTX target has very limited support of the debug
info in Clang/LLVM. Currently, LLVM supports only emission of the line
numbers debug info.
This is caused by limitations of the Cuda/NVPTX codegen. Clang/LLVM
translates the source code to LLVM IR, which is then lowered to PTX
(parallel thread execution) intermediate file. This PTX file represents
special kind of the assembler code in text format, which contains the
code itself + (possibly) debug info. Then this PTX file is compiled by
ptxas tool into the CUDA binary representation.

Debug info representation in PTX file.

According to PTX Writer's Guide to Interoperability, Debug information
(PTX Interoperability :: CUDA Toolkit Documentation)
, debug information must be encoded in DWARF (Debug With Arbitrary
Record Format). The responsibility for generating debug information is
split between the PTX producer and the PTX-to-SASS backend. The PTX
producer is responsible for emitting binary DWARF into the PTX file,
using the .section and .b8-.b16-.b32-and-.b64 directives in PTX. This
should contain the .debug_info and .debug_abbrev sections, and possibly
optional sections .debug_pubnames and .debug_aranges. These sections
are standard DWARF2 sections that refer to labels and registers in the
PTX.

The PTX-to-SASS backend is responsible for generating the .debug_line
section from the .file and .loc directives in the PTX file. This
section maps source lines to SASS addresses. The PTX-to-SASS backend
also generates the .debug_frame section.

All this sounds like the standard division of responsibilities between
an LLVM code generator and the assembler.

LLVM is able to emit debug info in DWARF. But ptxas compiler has some
limitations, that make it hard to adapt LLVM for correct emission of
the debug info in PTX files.

Limitations/features of the PTX format/ptxas compiler.

a) Supports DWARF-2 only.

IIRC, Darwin had a similar restriction until recently.

b) Labels are allowed only in code section (only in functions).

If you have static/global variables, I guess their locations would
have to be described using a section+offset expression? Normally
we emit a location attribute that is just a reference to a label
for the variable.

c) Does not support label arithmetic in DWARF sections.
    “.b32 L1 – L2” as the size of the section is not allowed, so the
sections sizes should be calculated explicitly.

MachO has a similar restriction, this should not be a problem if you
can do something like:
    L3 = L1 - L2
    .b32 L3

Nope, it is not supported