Date: Fri, 22 Mar 2024 15:44:28 -0400
From: Bruce Momjian <bruce@momjian.us>
To: walther@technowledgy.de
Cc: Thomas Munro <thomas.munro@gmail.com>,
	Peter Eisentraut <peter@eisentraut.org>,
	Christophe Pettus <xof@thebuild.com>,
	Andrew Dunstan <andrew@dunslane.net>,
	PostgreSQL Bugs <pgsql-bugs@lists.postgresql.org>,
	Tom Lane <tgl@sss.pgh.pa.us>, Robert Haas <robertmhaas@gmail.com>
Subject: Re: Regression tests fail with musl libc because libpq.so can't be
 loaded
Message-ID: <Zf3fnKpHIGpe5mOi@momjian.us>
References: 
 <CA+hUKGK2cSx3dveMqBD+E9N=zJ0WawBcbk_jnPqmdWBM0DT_3Q@mail.gmail.com>
 <477fa513-6bf0-456d-8d24-1e22a19df671@technowledgy.de>
 <ZfrmN4WHuJgD95fw@momjian.us>
 <CA+hUKGJzC4rr2Yr=QZyx9W83Wv=TVJHndKF1anjNLNXVUM7TqQ@mail.gmail.com>
 <f98cd8de-1c66-491a-8409-e62c09932080@technowledgy.de>
 <36690fcf-f0c8-4d60-af49-6a7efa0e2b14@technowledgy.de>
 <CA+hUKGJAkTZUa9JGr-zsadaqsN=K0P71t-1L2t2=MQFDMc4GuQ@mail.gmail.com>
 <ZfzLsiXbguTDatiZ@momjian.us>
 <a2463df1-be45-4a76-b90f-7676d778e7b1@technowledgy.de>
 <Zf2JU0_GDdbeBm5j@momjian.us>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="bGF/Zmuer4T+L7xb"
Content-Disposition: inline
In-Reply-To: <Zf2JU0_GDdbeBm5j@momjian.us>
Archived-At: 
 <https://www.postgresql.org/message-id/Zf3fnKpHIGpe5mOi%40momjian.us>
Precedence: bulk


--bGF/Zmuer4T+L7xb
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Fri, Mar 22, 2024 at 09:36:19AM -0400, Bruce Momjian wrote:
> On Fri, Mar 22, 2024 at 09:33:38AM +0100, walther@technowledgy.de wrote:
> > Bruce Momjian:
> > > I suggest we use the #ifdef test to continue our existing behavior for
> > > the libraries we know about, like glibc, and use the LD_* process title
> > > truncation hack for libc's we don't recognize.
> > > 
> > > Attached is a prototype patch which implements this based on previous
> > > patches.
> > 
> > The condition to check for linux/glibc in your patch is slightly off:
> > 
> >   #if ! defined(__linux__) || (! defined(__GLIBC__) && defined(__UCLIBC__ ))
> > 
> > should be
> > 
> >   #if defined(__linux__) && ! (defined(__GLIBC__) || defined(__UCLIBC__ ))
> > 
> > With the latter, it passes tests with musl.
> 
> Yes, my logic was wrong. Not sure what I was thinking, frankly.
> 
> I am not a big fan of negating a complex conditional, but would rather
> pass the negation into the conditional, new patch attached.

With no one "hoping this patch dies in a fire"*, I have updated it with
more details, which I now think is committable to master.  Is this
something to backpatch?  Seems too rare a bug to me.

* Robert Haas, https://www.postgresql.org/message-id/CA%2BTgmoYsyrCNmg%2BYh6rgP7K8r-bYPjCeF1tPxENRFwD4VZAZvw%40mail.gmail.com

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

  Only you can decide what is important to you.

--bGF/Zmuer4T+L7xb
Content-Type: text/x-diff; charset=us-ascii
Content-Disposition: attachment; filename="proctitle.diff"

diff --git a/src/backend/utils/misc/ps_status.c b/src/backend/utils/misc/ps_status.c
index 5d829e6e483..b4ed2b1a365 100644
--- a/src/backend/utils/misc/ps_status.c
+++ b/src/backend/utils/misc/ps_status.c
@@ -151,7 +151,37 @@ save_ps_display_args(int argc, char **argv)
 		for (i = 0; environ[i] != NULL; i++)
 		{
 			if (end_of_area + 1 == environ[i])
-				end_of_area = environ[i] + strlen(environ[i]);
+			{
+
+#if defined(__linux__) && (! defined(__GLIBC__) && ! defined(__UCLIBC__ ))
+				/*
+				 * The musl runtime linker stores pointers to variable
+				 * values which are defined in the process's environment.
+				 * Therefore, in these cases we cannot overwrite such
+				 * variable values when setting the process title or dynamic
+				 * linking (dlopen) might fail.  Here, we truncate the update
+				 * of the process title when either of two important dynamic
+				 * linking environment variables are set.  Musl does not
+				 * define any compiler symbols, so we have to do this for
+				 * any Linux libc we don't know is safe.
+				 */
+				if (strstr(environ[i], "LD_LIBRARY_PATH=") == environ[i] ||
+					strstr(environ[i], "LD_PRELOAD=") == environ[i])
+				{
+					/*
+					 * We can overwrite the name, but stop at the equals sign.
+					 * Future loops will not find contiguous space, but we
+					 * don't break early because we want to count the total
+					 * number.
+					 */
+					end_of_area = strchr(environ[i], '=');
+				}
+				else
+#endif
+				{
+					end_of_area = environ[i] + strlen(environ[i]);
+				}
+			}
 		}
 
 		ps_buffer = argv[0];

--bGF/Zmuer4T+L7xb--