Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-5589-g5f1d99180a6ab9.
gcc/analyzer/ChangeLog:
* sm-malloc.cc (deref_before_check::emit): Add logging of the
various conditions for late-rejection of a
-Wanalyzer-deref-before-check warning.
Signed-off-by: David Malcolm <[email protected]>
---
gcc/analyzer/sm-malloc.cc | 52 +++++++++++++++++++++++++++++++++------
1 file changed, 44 insertions(+), 8 deletions(-)
diff --git a/gcc/analyzer/sm-malloc.cc b/gcc/analyzer/sm-malloc.cc
index 8ce771062ca6a..88bfc74888c63 100644
--- a/gcc/analyzer/sm-malloc.cc
+++ b/gcc/analyzer/sm-malloc.cc
@@ -1612,19 +1612,34 @@ public:
bool emit (diagnostic_emission_context &ctxt) final override
{
+ LOG_SCOPE (ctxt.get_logger ());
+ logger *logger = ctxt.get_logger ();
+
/* Don't emit the warning if we can't show where the deref
and the check occur. */
if (!m_deref_enode)
- return false;
+ {
+ if (logger)
+ logger->log ("rejecting: no deref enode");
+ return false;
+ }
if (!m_check_enode)
- return false;
+ {
+ if (logger)
+ logger->log ("rejecting: no check enode");
+ return false;
+ }
/* Only emit the warning for intraprocedural cases. */
const program_point &deref_point = m_deref_enode->get_point ();
const program_point &check_point = m_check_enode->get_point ();
if (!program_point::effectively_intraprocedural_p (deref_point,
check_point))
- return false;
+ {
+ if (logger)
+ logger->log ("rejecting: not effectively intraprocedural");
+ return false;
+ }
/* Reject the warning if the check occurs within a macro defintion.
This avoids false positives for such code as:
@@ -1661,7 +1676,11 @@ public:
a source of real bugs; see e.g. PR 77425. */
location_t check_loc = m_check_enode->get_point ().get_location ();
if (linemap_location_from_macro_definition_p (line_table, check_loc))
- return false;
+ {
+ if (logger)
+ logger->log ("rejecting: check occurs within macro definition");
+ return false;
+ }
/* Reject warning if the check is in a loop header within a
macro expansion. This rejects cases like:
@@ -1676,16 +1695,29 @@ public:
would just be noise if we reported it. */
if (loop_header_p (m_check_enode->get_point ())
&& linemap_location_from_macro_expansion_p (line_table, check_loc))
- return false;
+ {
+ if (logger)
+ logger->log
+ ("rejecting: check occurs in loop header macro expansion");
+ return false;
+ }
/* Reject if m_deref_expr is sufficiently different from m_arg
for cases where the dereference is spelled differently from
the check, which is probably two different ways to get the
same svalue, and thus not worth reporting. */
if (!m_deref_expr)
- return false;
+ {
+ if (logger)
+ logger->log ("rejecting: no deref_expr");
+ return false;
+ }
if (!sufficiently_similar_p (m_deref_expr, m_arg))
- return false;
+ {
+ if (logger)
+ logger->log ("rejecting: not sufficiently similar to arg");
+ return false;
+ }
/* Reject the warning if the deref's BB doesn't dominate that
of the check, so that we don't warn e.g. for shared cleanup
@@ -1697,7 +1729,11 @@ public:
if (!dominated_by_p (CDI_DOMINATORS,
m_check_enode->get_supernode ()->m_bb,
m_deref_enode->get_supernode ()->m_bb))
- return false;
+ {
+ if (logger)
+ logger->log ("rejecting: deref doesn't dominate the check");
+ return false;
+ }
return ctxt.warn ("check of %qE for NULL after already"
" dereferencing it",
--
2.26.3
