Re: Win32 open items

From: "Magnus Hagander" <mha(at)sollentuna(dot)net>
To: "Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: <pgsql-hackers-win32(at)postgresql(dot)org>
Subject: Re: Win32 open items
Date: 2004-10-26 15:07:45
Message-ID: 6BCB9D8A16AC4241919521715F4D8BCE476012@algol.sollentuna.se
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers-win32

> > The problem with this is that PQrequestCancel() is not thread-safe.
>
> What is your basis for asserting that?

Looking at the source code?

We are talking about two different threads *accessing the same PGconn*.
For example:
if (conn->sock < 0)
{
strcpy(conn->errorMessage.data,
"PQrequestCancel() -- connection is not
open\n");
conn->errorMessage.len =
strlen(conn->errorMessage.data);
#ifdef WIN32
WSASetLastError(save_errno);
#else
errno = save_errno;
#endif
return FALSE;
}

What if another thread is modifying or reading the errorMessage variable
at the same time? Or what if another thread just PQfinished() the entire
conenction, which causes a free() call? None of this stuff is
thread-protected...

It may be re-entrant-safe, but I don't beleive it's thread safe. libpq
in general is only thread safe as long as a single connection is only
used on a single thread.

//Magnus

Browse pgsql-hackers-win32 by date

  From Date Subject
Next Message Andrew Dunstan 2004-10-26 15:34:46 Re: pg_ctl strangeness under msys
Previous Message Tom Lane 2004-10-26 14:41:09 Re: Win32 open items