Block environment variable mutations from trusted PL/Perl.

Many process environment variables (e.g. PATH), bypass the containment
expected of a trusted PL.  Hence, trusted PLs must not offer features
that achieve setenv().  Otherwise, an attacker having USAGE privilege on
the language often can achieve arbitrary code execution, even if the
attacker lacks a database server operating system user.

To fix PL/Perl, replace trusted PL/Perl %ENV with a tied hash that just
replaces each modification attempt with a warning.  Sites that reach
these warnings should evaluate the application-specific implications of
proceeding without the environment modification:

  Can the application reasonably proceed without the modification?

    If no, switch to plperlu or another approach.

    If yes, the application should change the code to stop attempting
    environment modifications.  If that's too difficult, add "untie
    %main::ENV" in any code executed before the warning.  For example,
    one might add it to the start of the affected function or even to
    the plperl.on_plperl_init setting.

In passing, link to Perl's guidance about the Perl features behind the
security posture of PL/Perl.

Back-patch to v12 (all supported versions).

Andrew Dunstan and Noah Misch

Security: CVE-2024-10979

5 files changed, 151 insertions, 2 deletions

diff --git a/src/pl/plperl/GNUmakefile b/src/pl/plperl/GNUmakefile
index 3a6954ce60e..01588d016a0 100644
--- a/src/pl/plperl/GNUmakefile
+++ b/src/pl/plperl/GNUmakefile
@@ -55,10 +55,10 @@ endif # win32
 
 SHLIB_LINK = $(perl_embed_ldflags)
 
-REGRESS_OPTS = --dbname=$(PL_TESTDB)
+REGRESS_OPTS = --dbname=$(PL_TESTDB) --dlpath=$(top_builddir)/src/test/regress
 REGRESS = plperl_setup plperl plperl_lc plperl_trigger plperl_shared \
 	plperl_elog plperl_util plperl_init plperlu plperl_array \
-	plperl_call plperl_transaction
+	plperl_call plperl_transaction plperl_env
 # if Perl can support two interpreters in one backend,
 # test plperl-and-plperlu cases
 ifneq ($(PERL),)
diff --git a/src/pl/plperl/input/plperl_env.source b/src/pl/plperl/input/plperl_env.source
new file mode 100644
index 00000000000..8fe526e1b8b
--- /dev/null
+++ b/src/pl/plperl/input/plperl_env.source
@@ -0,0 +1,52 @@
+--
+-- Test the environment setting
+--
+
+CREATE FUNCTION get_environ()
+   RETURNS text[]
+   AS '@libdir@/regress@DLSUFFIX@', 'get_environ'
+   LANGUAGE C STRICT;
+
+-- fetch the process environment
+
+CREATE FUNCTION process_env () RETURNS text[]
+LANGUAGE plpgsql AS
+$$
+
+declare
+   res text[];
+   tmp text[];
+   f record;
+begin
+    for f in select unnest(get_environ()) as t loop
+         tmp := regexp_split_to_array(f.t, '=');
+         if array_length(tmp, 1) = 2 then
+            res := res || tmp;
+         end if;
+    end loop;
+    return res;
+end
+
+$$;
+
+-- plperl should not be able to affect the process environment
+
+DO
+$$
+   $ENV{TEST_PLPERL_ENV_FOO} = "shouldfail";
+   untie %ENV;
+   $ENV{TEST_PLPERL_ENV_FOO} = "testval";
+   my $penv = spi_exec_query("select unnest(process_env()) as pe");
+   my %received;
+   for (my $f = 0; $f < $penv->{processed}; $f += 2)
+   {
+      my $k = $penv->{rows}[$f]->{pe};
+      my $v = $penv->{rows}[$f+1]->{pe};
+      $received{$k} = $v;
+   }
+   unless (exists $received{TEST_PLPERL_ENV_FOO})
+   {
+      elog(NOTICE, "environ unaffected")
+   }
+
+$$ LANGUAGE plperl;
diff --git a/src/pl/plperl/output/plperl_env.source b/src/pl/plperl/output/plperl_env.source
new file mode 100644
index 00000000000..37b7e23d5ce
--- /dev/null
+++ b/src/pl/plperl/output/plperl_env.source
@@ -0,0 +1,49 @@
+--
+-- Test the environment setting
+--
+CREATE FUNCTION get_environ()
+   RETURNS text[]
+   AS '@libdir@/regress@DLSUFFIX@', 'get_environ'
+   LANGUAGE C STRICT;
+-- fetch the process environment
+CREATE FUNCTION process_env () RETURNS text[]
+LANGUAGE plpgsql AS
+$$
+
+declare
+   res text[];
+   tmp text[];
+   f record;
+begin
+    for f in select unnest(get_environ()) as t loop
+         tmp := regexp_split_to_array(f.t, '=');
+         if array_length(tmp, 1) = 2 then
+            res := res || tmp;
+         end if;
+    end loop;
+    return res;
+end
+
+$$;
+-- plperl should not be able to affect the process environment
+DO
+$$
+   $ENV{TEST_PLPERL_ENV_FOO} = "shouldfail";
+   untie %ENV;
+   $ENV{TEST_PLPERL_ENV_FOO} = "testval";
+   my $penv = spi_exec_query("select unnest(process_env()) as pe");
+   my %received;
+   for (my $f = 0; $f < $penv->{processed}; $f += 2)
+   {
+      my $k = $penv->{rows}[$f]->{pe};
+      my $v = $penv->{rows}[$f+1]->{pe};
+      $received{$k} = $v;
+   }
+   unless (exists $received{TEST_PLPERL_ENV_FOO})
+   {
+      elog(NOTICE, "environ unaffected")
+   }
+
+$$ LANGUAGE plperl;
+WARNING:  attempted alteration of $ENV{TEST_PLPERL_ENV_FOO} at line 12.
+NOTICE:  environ unaffected
diff --git a/src/pl/plperl/plc_trusted.pl b/src/pl/plperl/plc_trusted.pl
index dea3727682c..5854661fff5 100644
--- a/src/pl/plperl/plc_trusted.pl
+++ b/src/pl/plperl/plc_trusted.pl
@@ -27,3 +27,27 @@ require Carp;
 require Carp::Heavy;
 require warnings;
 require feature if $] >= 5.010000;
+
+#<<< protect next line from perltidy so perlcritic annotation works
+package PostgreSQL::InServer::WarnEnv; ## no critic (RequireFilenameMatchesPackage)
+#>>>
+
+use strict;
+use warnings;
+use Tie::Hash;
+our @ISA = qw(Tie::StdHash);
+
+sub STORE  { warn "attempted alteration of \$ENV{$_[1]}"; }
+sub DELETE { warn "attempted deletion of \$ENV{$_[1]}"; }
+sub CLEAR  { warn "attempted clearance of ENV hash"; }
+
+# Remove magic property of %ENV. Changes to this will now not be reflected in
+# the process environment.
+*main::ENV = {%ENV};
+
+# Block %ENV changes from trusted PL/Perl, and warn. We changed %ENV to just a
+# normal hash, yet the application may be expecting the usual Perl %ENV
+# magic. Blocking and warning avoids silent application breakage. The user can
+# untie or otherwise disable this, e.g. if the lost mutation is unimportant
+# and modifying the code to stop that mutation would be onerous.
+tie %main::ENV, 'PostgreSQL::InServer::WarnEnv', %ENV or die $!;
diff --git a/src/test/regress/regress.c b/src/test/regress/regress.c
index 09bc42a8c0f..aa9fef866fa 100644
--- a/src/test/regress/regress.c
+++ b/src/test/regress/regress.c
@@ -35,6 +35,7 @@
 #include "optimizer/plancat.h"
 #include "port/atomics.h"
 #include "storage/spin.h"
+#include "utils/array.h"
 #include "utils/builtins.h"
 #include "utils/geo_decls.h"
 #include "utils/memutils.h"
@@ -624,6 +625,29 @@ make_tuple_indirect(PG_FUNCTION_ARGS)
 	PG_RETURN_POINTER(newtup->t_data);
 }
 
+PG_FUNCTION_INFO_V1(get_environ);
+
+Datum
+get_environ(PG_FUNCTION_ARGS)
+{
+	extern char **environ;
+	int			nvals = 0;
+	ArrayType  *result;
+	Datum	   *env;
+
+	for (char **s = environ; *s; s++)
+		nvals++;
+
+	env = palloc(nvals * sizeof(Datum));
+
+	for (int i = 0; i < nvals; i++)
+		env[i] = CStringGetTextDatum(environ[i]);
+
+	result = construct_array(env, nvals, TEXTOID, -1, false, TYPALIGN_INT);
+
+	PG_RETURN_POINTER(result);
+}
+
 PG_FUNCTION_INFO_V1(regress_putenv);
 
 Datum