Re: [PATCH] Detect escape of ErrorContextCallback stack pointers (and from PG_TRY() )

On Fri, 4 Sep 2020 at 14:55, Craig Ringer <craig(at)2ndquadrant(dot)com> wrote:

> Using a test program:
>
> return_stack_escape.c:14:3: warning: Address of stack memory associated
> with local variable 'g' is still referred to by the global variable
> 'guard_ptr' upon returning to the caller. This will be a dangling reference
> return do_fail;
> ^~~~~~~~~~~~~~
> 1 warning generated.
>
>
Example here:

https://github.com/ringerc/scrapcode/tree/master/c/clang_return_stack_checks

So I find that actually, the __attribute__((callback(fn)) approach is
unnecessary for the purpose proposed.

I still think we should be looking at tidying up the error contexts API,
but it's easy enough for extensions to provide a wrapper over it, so that's
not a big deal really.

I will submit my docs patches separately to ensure they get some attention,
and I'll experiment with llvm threadsafety annotations / capabilities.

--
Craig Ringer http://www.2ndQuadrant.com/
2ndQuadrant - PostgreSQL Solutions for the Enterprise