https://www.postgresql.org/message-id/canncwrjtse6wkkus_y8wj2phvrvaqpxmk_qtepsf_+nvpyx...@mail.gmail.com
As the problem discussed in the above thread, I also run into that. Besides
updating the doc, I would like to report a error for it.
If the code in PG_TRY contains any non local control flow other than
ereport(ERROR) like goto, break etc., the PG_CATCH or PG_END_TRY cannot
be called, then the PG_exception_stack will point to the memory whose
stack frame has been released. So after that, when the pg_re_throw
called, __longjmp() will crash and report Segmentation fault error.
In that case, to help developers to figure out the root cause easily, it is
better to report that 'the sigjmp_buf is invalid' rather than letting
the __longjmp report any error.
Addition to sigjmp_buf, add another field 'int magic' which is next to
the sigjum_buf in the local stack frame memory. The magic's value is always
'PG_exception_magic 0x12345678'. And in 'pg_re_throw' routine, check if
the magic's value is still '0x12345678', if not, that means the memory
where the 'PG_exception_stack' points to has been released, and the
'sigbuf'
must be invalid.
The related code is in patch 0001
------------------------------
I'm not sure if it is necessary to add a regress test for it. In patch
0002, to test the
patch can work correctly, I have added a function 'pg_re_throw_crash' in
regress.c
create function pg_re_throw_crash()
RETURNS void
AS :'regresslib', 'pg_re_throw_crash'
LANGUAGE C STRICT STABLE PARALLEL SAFE;
create above function and run 'select pg_re_throw_crash()', then will get
the error
'FATAL: Invalid sigjum_buf, code in PG_TRY cannot contain any non local
control flow other than ereport'
--
Best regards !
Xiaoran Wang
0001-Imporve-pg_re_throw-check-if-sigjmp_buf-is-valid-and.patch
Description: Binary data
0002-Test-pg_re_throw-checking-invalid-sigjmp_buf.patch
Description: Binary data
