Thread: BUG #16666: Slight memory leak when running pg_ctl reload

The following bug has been logged on the website:

Bug reference:      16666
Logged by:          BoChen
Email address:      bchen90@163.com
PostgreSQL version: 11.8
Operating system:   SUSE Linux Enterprise Server 11 (x86_64)
Description:

I found slighting memory leakage occured when running pg_ctl reload command
for the following code segment(guc.c:6726-6740). When runing 'pg_ctl
reload', for these string type PGC_POSTMASTER GUC variables like
'unix_socket_directories', the following code segment will leak newval's
memery.

if (prohibitValueChange)
{
    /* newval shouldn't be NULL, so we're a bit sloppy here */
    if (*conf->variable == NULL || newval == NULL ||
        strcmp(*conf->variable, newval) != 0)
    {
        record->status |= GUC_PENDING_RESTART;
        ereport(elevel,
                (errcode(ERRCODE_CANT_CHANGE_RUNTIME_PARAM),
                 errmsg("parameter \"%s\" cannot be changed without restarting the
server",
                        name)));
        return 0;
    }
    record->status &= ~GUC_PENDING_RESTART;
    return -1;
}

On Mon, Oct 12, 2020 at 01:32:13PM +0000, PG Bug reporting form wrote:
> The following bug has been logged on the website:
> 
> Bug reference:      16666
> Logged by:          BoChen
> Email address:      bchen90@163.com
> PostgreSQL version: 11.8
> Operating system:   SUSE Linux Enterprise Server 11 (x86_64)
> Description:        
> 
> I found slighting memory leakage occured when running pg_ctl reload command
> for the following code segment(guc.c:6726-6740). When runing 'pg_ctl
> reload', for these string type PGC_POSTMASTER GUC variables like
> 'unix_socket_directories', the following code segment will leak newval's
> memery.

Uh, you are saying pg_ctl leaks memory or the database server leaks
memory?  I don't see any memory allocation in the code you posted.

-- 
  Bruce Momjian  <bruce@momjian.us>        https://momjian.us
  EnterpriseDB                             https://enterprisedb.com

  The usefulness of a cup is in its emptiness, Bruce Lee

Bruce Momjian <bruce@momjian.us> writes:
> Uh, you are saying pg_ctl leaks memory or the database server leaks
> memory?  I don't see any memory allocation in the code you posted.

I've not looked at the code yet, but I think the OP is worried that
we need to free(newval) before these "return"s.

            regards, tom lane

PG Bug reporting form <noreply@postgresql.org> writes:
> I found slighting memory leakage occured when running pg_ctl reload command
> for the following code segment(guc.c:6726-6740). When runing 'pg_ctl
> reload', for these string type PGC_POSTMASTER GUC variables like
> 'unix_socket_directories', the following code segment will leak newval's
> memery.

Yeah, you're right.  If there's a "newextra" chunk, that'll get
leaked too.

Even counting that, it's not very plausible that one would SIGHUP a
server often enough that the bloat would become obvious, since these
values typically aren't very large.  But it's clearly a bug.

Patch applied, thanks for the report!

            regards, tom lane