On Fri, Sep 23, 2022 at 4:19 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> That's a fairly scary observation. What other places are silently
> expecting such memory to be zeroed? Do we need to fix things so
> that min_dynamic_shared_memory doesn't break this API?
Reminds me of the recent parallel VACUUM bug, fixed by commit
662ba729. That involved unintialized DSM, too.
As I said at the time: we should be using Valgrind for this.
--
Peter Geoghegan