Horton

Re: remaining sql/json patches

amit <amitlangote09@gmail.com>

View thread
From: Amit Langote <amitlangote09@gmail.com>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
Cc: Andrew Dunstan <andrew@dunslane.net>, Erik Rijkers <er@xs4all.nl>, PostgreSQL-development <pgsql-hackers@postgresql.org>, jian he <jian.universality@gmail.com>
Date: 2023-10-02T05:26:49Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. SQL/JSON: Various improvements to SQL/JSON query function docs

  2. SQL/JSON: Fix some obsolete comments.

  3. SQL/JSON: Fix issues with DEFAULT .. ON ERROR / EMPTY

  4. JSON_TABLE: Add support for NESTED paths and columns

  5. Fix JsonExpr deparsing to emit QUOTES and WRAPPER correctly

  6. Fix typo introduced in 6185c9737

  7. Add basic JSON_TABLE() functionality

  8. Avoid splitting errmsg string to span multiple lines

  9. Add SQL/JSON query functions

  10. Implement various jsonpath methods

  11. Add soft error handling to some expression nodes

  12. Adjust populate_record_field() to handle errors softly

  13. Refactor code used by jsonpath executor to fetch variables

  14. Test EXPLAIN (FORMAT JSON) ... XMLTABLE

  15. Simplify productions for FORMAT JSON [ ENCODING name ]

  16. Add trailing commas to enum definitions

  17. doc: add missing <returnvalue> and whitespace

  18. Add more SQL/JSON constructor functions

  19. Rename a nonterminal used in SQL/JSON grammar

  20. Some refactoring to export json(b) conversion functions

  21. Don't include CaseTestExpr in JsonValueExpr.formatted_expr

  22. Code review for commit b6e1157e7d

  23. Pass constructName to transformJsonValueExpr()

  24. Unify JSON categorize type API and export for external use

  25. Make some indentation in gram.y consistent

  26. Allow most keywords to be used as column labels without requiring AS.

  27. Reduce size of backend scanner's tables.

  28. Use perfect hashing, instead of binary search, for keyword lookup.

On Mon, Oct 2, 2023 at 1:24 PM Amit Langote <amitlangote09@gmail.com> wrote:
> Pushed this 30 min ago (no email on -committers yet!) and am looking
> at the following llvm crash reported by buildfarm animal pogona [1]:
>
> #4  0x00007f5bceb673d5 in __assert_fail_base (fmt=0x7f5bcecdbdc8
> "%s%s%s:%u: %s%sAssertion `%s' failed.\\n%n",
> assertion=assertion@entry=0x7f5bc1336419 "(i >= FTy->getNumParams() ||
> FTy->getParamType(i) == Args[i]->getType()) && \\"Calling a function
> with a bad signature!\\"", file=file@entry=0x7f5bc1336051
> "/home/bf/src/llvm-project-5/llvm/lib/IR/Instructions.cpp",
> line=line@entry=299, function=function@entry=0x7f5bc13362af "void
> llvm::CallInst::init(llvm::FunctionType *, llvm::Value *,
> ArrayRef<llvm::Value *>, ArrayRef<llvm::OperandBundleDef>, const
> llvm::Twine &)") at ./assert/assert.c:92
> #5  0x00007f5bceb763a2 in __assert_fail (assertion=0x7f5bc1336419 "(i
> >= FTy->getNumParams() || FTy->getParamType(i) == Args[i]->getType())
> && \\"Calling a function with a bad signature!\\"",
> file=0x7f5bc1336051
> "/home/bf/src/llvm-project-5/llvm/lib/IR/Instructions.cpp", line=299,
> function=0x7f5bc13362af "void llvm::CallInst::init(llvm::FunctionType
> *, llvm::Value *, ArrayRef<llvm::Value *>,
> ArrayRef<llvm::OperandBundleDef>, const llvm::Twine &)") at
> ./assert/assert.c:101
> #6  0x00007f5bc110f138 in llvm::CallInst::init (this=0x557a91f3e508,
> FTy=0x557a91ed9ae0, Func=0x557a91f8be88, Args=..., Bundles=...,
> NameStr=...) at
> /home/bf/src/llvm-project-5/llvm/lib/IR/Instructions.cpp:297
> #7  0x00007f5bc0fa579d in llvm::CallInst::CallInst
> (this=0x557a91f3e508, Ty=0x557a91ed9ae0, Func=0x557a91f8be88,
> Args=..., Bundles=..., NameStr=..., InsertBefore=0x0) at
> /home/bf/src/llvm-project-5/llvm/include/llvm/IR/Instructions.h:1934
> #8  0x00007f5bc0fa538c in llvm::CallInst::Create (Ty=0x557a91ed9ae0,
> Func=0x557a91f8be88, Args=..., Bundles=..., NameStr=...,
> InsertBefore=0x0) at
> /home/bf/src/llvm-project-5/llvm/include/llvm/IR/Instructions.h:1444
> #9  0x00007f5bc0fa51f9 in llvm::IRBuilder<llvm::ConstantFolder,
> llvm::IRBuilderDefaultInserter>::CreateCall (this=0x557a91f9c6a0,
> FTy=0x557a91ed9ae0, Callee=0x557a91f8be88, Args=..., Name=...,
> FPMathTag=0x0) at
> /home/bf/src/llvm-project-5/llvm/include/llvm/IR/IRBuilder.h:1669
> #10 0x00007f5bc100edda in llvm::IRBuilder<llvm::ConstantFolder,
> llvm::IRBuilderDefaultInserter>::CreateCall (this=0x557a91f9c6a0,
> Callee=0x557a91f8be88, Args=..., Name=..., FPMathTag=0x0) at
> /home/bf/src/llvm-project-5/llvm/include/llvm/IR/IRBuilder.h:1663
> #11 0x00007f5bc100714e in LLVMBuildCall (B=0x557a91f9c6a0,
> Fn=0x557a91f8be88, Args=0x7ffde6fa0b50, NumArgs=6, Name=0x7f5bc30b648c
> "funccall_iocoerce_in_safe") at
> /home/bf/src/llvm-project-5/llvm/lib/IR/Core.cpp:2964
> #12 0x00007f5bc30af861 in llvm_compile_expr (state=0x557a91fbeac0) at
>
/home/bf/bf-build/pogona/HEAD/pgsql.build/../pgsql/src/backend/jit/llvm/llvmjit_expr.c:1373
>
> This seems to me to be complaining about the following addition:
>
> +                   {
> +                       Oid         ioparam = op->d.iocoerce.typioparam;
> +                       LLVMValueRef v_params[6];
> +                       LLVMValueRef v_success;
> +
> +                       v_params[0] = l_ptr_const(op->d.iocoerce.finfo_in,
> +                                                 l_ptr(StructFmgrInfo));
> +                       v_params[1] = v_output;
> +                       v_params[2] = l_oid_const(lc, ioparam);
> +                       v_params[3] = l_int32_const(lc, -1);
> +                       v_params[4] =
l_ptr_const(op->d.iocoerce.escontext,
> +
> l_ptr(StructErrorSaveContext));
>
> -                   LLVMBuildStore(b, v_retval, v_resvaluep);
> +                       /*
> +                        * InputFunctionCallSafe() will write directly
into
> +                        * *op->resvalue.
> +                        */
> +                       v_params[5] = v_resvaluep;
> +
> +                       v_success = LLVMBuildCall(b, llvm_pg_func(mod,
> "InputFunctionCallSafe"),
> +                                                 v_params,
lengthof(v_params),
> +
 "funccall_iocoerce_in_safe");
> +
> +                       /*
> +                        * Return null if InputFunctionCallSafe()
encountered
> +                        * an error.
> +                        */
> +                       v_resnullp = LLVMBuildICmp(b, LLVMIntEQ,
v_success,
> +                                                  l_sbool_const(0), "");
> +                   }

Although most animals except pogona looked fine, I've decided to revert the
patch for now.

IIUC, LLVM is complaining that the code in the above block is not passing
the arguments of InputFunctionCallSafe() using the correct types.  I'm not
exactly sure which particular argument is not handled correctly in the
above code, but perhaps it's:


+                       v_params[1] = v_output;

which maps to char *str argument of InputFunctionCallSafe().  v_output is
set in the code preceding the above block as follows:

                    /* and call output function (can never return NULL) */
                    v_output = LLVMBuildCall(b, v_fn_out, &v_fcinfo_out,
                                             1, "funccall_coerce_out");

I thought that it would be fine to pass it as-is to the call of
InputFunctionCallSafe() given that v_fn_out is a call to a function that
returns char *, but perhaps not.


-- 
Thanks, Amit Langote
EDB: http://www.enterprisedb.com