diff options
author | Peter Eisentraut | 2025-06-25 07:55:04 +0000 |
---|---|---|
committer | Peter Eisentraut | 2025-06-25 07:56:49 +0000 |
commit | 0cd69b3d7ef357f2b43258de5831c4de0bd51dec (patch) | |
tree | 58eb6ee394792c6a61b45828a2f14a4bebed8a3a /src/backend | |
parent | 69e5cdc47fa1fbd97c8b8c7abe6fb0b9a9822acb (diff) |
Restrict virtual columns to use built-in functions and types
Just like selecting from a view is exploitable (CVE-2024-7348),
selecting from a table with virtual generated columns is exploitable.
Users who are concerned about this can avoid selecting from views, but
telling them to avoid selecting from tables is less practical.
To address this, this changes it so that generation expressions for
virtual generated columns are restricted to using built-in functions
and types, and the columns are restricted to having a built-in type.
We assume that built-in functions and types cannot be exploited for
this purpose.
In the future, this could be expanded by some new mechanism to declare
other functions and types as safe or trusted for this purpose, but
that is to be designed.
(An alternative approach might have been to expand the
restrict_nonsystem_relation_kind GUC to handle this, like the fix for
CVE-2024-7348. But that is kind of an ugly approach. That fix had to
fit in the constraints of fixing an ancient vulnerability in all
branches. Since virtual generated columns are new, we're free from
the constraints of the past, and we can and should use cleaner
options.)
Reported-by: Feike Steenbergen <[email protected]>
Reviewed-by: jian he <[email protected]>
Discussion: https://www.postgresql.org/message-id/flat/CAK_s-G2Q7de8Q0qOYUR%3D_CTB5FzzVBm5iZjOp%2BmeVWpMpmfO0w%40mail.gmail.com
Diffstat (limited to 'src/backend')
-rw-r--r-- | src/backend/catalog/heap.c | 93 |
1 files changed, 93 insertions, 0 deletions
diff --git a/src/backend/catalog/heap.c b/src/backend/catalog/heap.c index 10f43c51c5a..649d3966e8e 100644 --- a/src/backend/catalog/heap.c +++ b/src/backend/catalog/heap.c @@ -665,6 +665,15 @@ CheckAttributeType(const char *attname, } /* + * For consistency with check_virtual_generated_security(). + */ + if ((flags & CHKATYPE_IS_VIRTUAL) && atttypid >= FirstUnpinnedObjectId) + ereport(ERROR, + errcode(ERRCODE_FEATURE_NOT_SUPPORTED), + errmsg("virtual generated column \"%s\" cannot have a user-defined type", attname), + errdetail("Virtual generated columns that make use of user-defined types are not yet supported.")); + + /* * This might not be strictly invalid per SQL standard, but it is pretty * useless, and it cannot be dumped, so we must disallow it. */ @@ -3216,6 +3225,86 @@ check_nested_generated(ParseState *pstate, Node *node) } /* + * Check security of virtual generated column expression. + * + * Just like selecting from a view is exploitable (CVE-2024-7348), selecting + * from a table with virtual generated columns is exploitable. Users who are + * concerned about this can avoid selecting from views, but telling them to + * avoid selecting from tables is less practical. + * + * To address this, this restricts generation expressions for virtual + * generated columns are restricted to using built-in functions and types. We + * assume that built-in functions and types cannot be exploited for this + * purpose. Note the overall security also requires that all functions in use + * a immutable. (For example, there are some built-in non-immutable functions + * that can run arbitrary SQL.) The immutability is checked elsewhere, since + * that is a property that needs to hold independent of security + * considerations. + * + * In the future, this could be expanded by some new mechanism to declare + * other functions and types as safe or trusted for this purpose, but that is + * to be designed. + */ + +/* + * Callback for check_functions_in_node() that determines whether a function + * is user-defined. + */ +static bool +contains_user_functions_checker(Oid func_id, void *context) +{ + return (func_id >= FirstUnpinnedObjectId); +} + +/* + * Checks for all the things we don't want in the generation expressions of + * virtual generated columns for security reasons. Errors out if it finds + * one. + */ +static bool +check_virtual_generated_security_walker(Node *node, void *context) +{ + ParseState *pstate = context; + + if (node == NULL) + return false; + + if (!IsA(node, List)) + { + if (check_functions_in_node(node, contains_user_functions_checker, NULL)) + ereport(ERROR, + errcode(ERRCODE_FEATURE_NOT_SUPPORTED), + errmsg("generation expression uses user-defined function"), + errdetail("Virtual generated columns that make use of user-defined functions are not yet supported."), + parser_errposition(pstate, exprLocation(node))); + + /* + * check_functions_in_node() doesn't check some node types (see + * comment there). We handle CoerceToDomain and MinMaxExpr by + * checking for built-in types. The other listed node types cannot + * call user-definable SQL-visible functions. + * + * We furthermore need this type check to handle built-in, immutable + * polymorphic functions such as array_eq(). + */ + if (exprType(node) >= FirstUnpinnedObjectId) + ereport(ERROR, + errcode(ERRCODE_FEATURE_NOT_SUPPORTED), + errmsg("generation expression uses user-defined type"), + errdetail("Virtual generated columns that make use of user-defined types are not yet supported."), + parser_errposition(pstate, exprLocation(node))); + } + + return expression_tree_walker(node, check_virtual_generated_security_walker, context); +} + +static void +check_virtual_generated_security(ParseState *pstate, Node *node) +{ + check_virtual_generated_security_walker(node, pstate); +} + +/* * Take a raw default and convert it to a cooked format ready for * storage. * @@ -3254,6 +3343,10 @@ cookDefault(ParseState *pstate, ereport(ERROR, (errcode(ERRCODE_INVALID_OBJECT_DEFINITION), errmsg("generation expression is not immutable"))); + + /* Check security of expressions for virtual generated column */ + if (attgenerated == ATTRIBUTE_GENERATED_VIRTUAL) + check_virtual_generated_security(pstate, expr); } else { |