https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101675
Bug ID: 101675
Summary: analyzer/pr94851-2.c marked XFAIL because it fails
with new jump threader
Product: gcc
Version: 12.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: middle-end
Assignee: unassigned at gcc dot gnu.org
Reporter: aldyh at gcc dot gnu.org
CC: dmalcolm at gcc dot gnu.org
Target Milestone: ---
By *.033t.forwprop1, we have the following:
<bb 4> :
# p_9 = PHI <p_16(2), p_22(3)>
# last_11 = PHI <last_17(2), p_9(3)>
if (p_9 != 0B)
goto <bb 5>; [INV]
else
goto <bb 6>; [INV]
<bb 5> :
_3 = p_9->m_name;
_4 = (char) c_20;
if (_3 != _4)
goto <bb 3>; [INV]
else
goto <bb 6>; [INV]
<bb 6> :
if (p_9 != 0B)
goto <bb 7>; [INV]
else
goto <bb 8>; [INV]
<bb 7> :
printf ("over writing mark %c\n", c_20);
goto <bb 13>; [INV]
<bb 8> :
p_24 = malloc (16);
if (p_24 == 0B)
goto <bb 9>; [INV]
else
goto <bb 10>; [INV]
After the jump threader rewrite (commit
2e96b5f14e4025691b57d2301d71aa6092ed44bc), we thread the 4->6->8 path in the
early threader pass:
Registering FSM jump thread: (4, 6) incoming edge; (6, 8) nocopy; (6, 8)
This causes the false edge on the first conditional to jump directly to BB8.
Presumably this is confusing the analyzer, which is no longer warning here:
if (curbp->b_amark == (AMARK *)NULL)
curbp->b_amark = p;
else
last->m_next = p; /* { dg-warning "dereference of NULL 'last'" "deref" {
xfail *-*-* } } */