Hi,
On 2021-04-20 00:58:21 -0500, Justin Pryzby wrote:
> On Tue, Apr 20, 2021 at 12:38:26AM -0500, Justin Pryzby wrote:
> > I don't know if this is related to the other issues, but this seems leaky.
>
> And it explains how the context use counter can exceed its threshold.
>
> create or replace function fn() returns void language plpgsql as $$ declare rec int; begin SELECT relpages INTO rec
FROMpg_class LIMIT 1; end $$;
> explain analyze
> SELECT fn()
> FROM generate_series(1,99);
>
> SELECT SUM(a) FROM (VALUES(1))a(a);
>
> time PGOPTIONS='-cclient_min_messages=debug -clog_executor_stats=off -clog_min_duration_statement=-1 -cjit=on
-cjit_above_cost=0-cjit_inline_above_cost=0' psql ts -f jitleak.sql
> ...
> psql:jitleak.sql:6: DEBUG: recreating LLVM context after 100 uses
Right - at the moment the context can only be recreated when there's no
JITed queries ongoing. That's why I asked whether your "real" query
contains function calls, and whether those functions are expensive
enough to be JITed.
> Question: does the jit context need to be recreated only when inlining is
> enabled? Or maybe it's better if it's not conditionalized like that..
It'd be sufficient to do it when doing inlining, but currently that's
not tracked separately.
Greetings,
Andres Freund
