On Mon, 20 Mar 2023, Richard Biener wrote:
> On Fri, 17 Mar 2023, Jakub Jelinek wrote:
>
> > On Fri, Mar 17, 2023 at 02:18:52PM +0000, Richard Biener wrote:
> > > > And as you show on the testcases, it probably isn't a good idea for
> > > > BUILT_IN_EXPECT* either.
> > > >
> > > > So, perhaps use op_cfn_pass_through_arg1 for the ERF_RETURNS_ARG
> > > > functions
> > > > and BUILT_IN_EXPECT* ?
> > >
> > > But that already causes the problems (I didn't yet finish testing
> > > adding RET1 to BUILT_IN_EXPECT*). Note FRE currently does not use
> > > returns_arg but I have old patches that do - but those replace
> > > uses after a RET1 function with the return value to also reduce
> > > spilling around a call (they of course CSE equal calls).
> >
> > I meant in your patch drop the builtins.cc hunk and add from your
> > other patch
> > > > + case CFN_BUILT_IN_EXPECT:
> > > > + case CFN_BUILT_IN_EXPECT_WITH_PROBABILITY:
> > > > + m_valid = true;
> > > > + m_op1 = gimple_call_arg (call, 0);
> > > > + m_int = &op_cfn_pass_through_arg1;
> > > > + break;
> >
> > hunk to gimple_range_op_handler::maybe_builtin_call.
> > Does that already cause the problems?
>
> Yes, that's basically what the first variant of the patch did and
> the FAILs quoted are from testing that variant.
>
> There are no extra fails from the second more generic patch also
> touching builtins.cc
>
> > I mean, if we e.g. see that a range of the argument is singleton,
> > then it is fine to optimize the __builtin_expect away.
>
> Yes, it's somewhat odd that we handle the - case but not the +:
>
> @@ -204,6 +206,7 @@
> _2 = b.0_1 < 0;
> # RANGE [irange] long int [0, 1] NONZERO 0x1
> _3 = (long int) _2;
> + # RANGE [irange] long int [0, 1] NONZERO 0x1
> _4 = __builtin_expect (_3, 0);
> if (_4 != 0)
> goto <bb 5>; [10.00%]
> ...
> @@ -224,13 +228,14 @@
> goto <bb 3>; [100.00%]
>
> <bb 5> [local count: 977105059]:
> - # _9 = PHI <_4(3), 0(4)>
> + # RANGE [irange] long int [0, 1] NONZERO 0x1
> + # _9 = PHI <1(3), 0(4)>
> if (_9 != 0)
> - goto <bb 6>; [10.00%]
> + goto <bb 6>; [50.00%]
> else
> - goto <bb 7>; [90.00%]
> + goto <bb 7>; [50.00%]
>
>
> Predictions for bb 5
> - first match heuristics: 10.00%
> - combined heuristics: 10.00%
> - __builtin_expect heuristics of edge 5->6: 10.00%
> + no prediction heuristics: 50.00%
> + combined heuristics: 50.00%
>
> the dumps don't get any hints on where the first matchor
> __builtin_expect heuristic came from, but it looks like we
> run expr_expected_value_1 on _9 != 0 which recurses for _9
> and runs into the PHI handling which then looks for
> a common value into the PHI. In this case _4 is said
> to be zero by PRED_BUILTIN_EXPECT and probability 9000.
> But that seems wrong - it looks simply at the __builtin_expect
> def here, not taking into account the edge this flows through
> (the unlikely edge).
>
> If we look at the recorded edge predictions we instead see
>
> $28 = {ep_next = 0x43e4c80, ep_edge = <edge 0x7ffff67e0f00 (3 -> 5)>,
> ep_predictor = PRED_BUILTIN_EXPECT, ep_probability = 1000}
>
> so that's the same edge but unlikely now. Of course there's
> no value recorded for this. For the other edge into the PHI
> we get
>
> $31 = {ep_next = 0x43c1e90, ep_edge = <edge 0x7ffff67e0f60 (4 -> 5)>,
> ep_predictor = PRED_BUILTIN_EXPECT, ep_probability = 9000}
>
> so to me a more reasonable approach would be to record '0' from
> the 2nd edge with a probability of 9000 (or for the '+' case IL
> '1' with a probability of 1000). There's possibly a way to
> combine predictor + value (here we'd simply take the more
> probable value, or the constant for a PHI). I also see that
> we handle any PHI this way, not just PHIs defined in the same
> BB as the condition which op we're ultimatively interested in.
>
> So my conclusion is that currently it's pure luck the testcase
> "works", and thus adjusting it and opening a bug with the
> above findings would be appropriate?
>
> Honza?
I filed PR109210 and updated the patch with adjustment of the
two failing testcases.
Bootstrapped and tested on x86_64-unknown-linux-gnu.
OK?
Thanks,
Richard.
>From aac0ed731c49f74d7f363b80745150943b75bd4c Mon Sep 17 00:00:00 2001
From: Richard Biener <rguent...@suse.de>
Date: Fri, 17 Mar 2023 13:14:49 +0100
Subject: [PATCH] tree-optimization/109170 - bogus use-after-free with
__builtin_expect
To: gcc-patches@gcc.gnu.org
The following adds a missing range-op for __builtin_expect which
helps -Wuse-after-free to detect the case a realloc original
pointer is used when the result was NULL. The implementation
should handle all argument one pass-through builtins we handle
in the fnspec machinery.
The gcc.dg/tree-ssa/ssa-lim-21.c testcase needs adjustment because
for (int j = 0; j < m; j++)
if (__builtin_expect (m, 0))
for (int i = 0; i < m; i++)
is now correctly optimized to a unconditional jump by EVRP - m
cannot be zero when the outer loop is entered. I've adjusted
the outer loop to iterate 'n' times which makes us apply store-motion
to 'count' and 'q->data1' but only out of the inner loop and
as expected not apply store motion to 'q->data' at all.
The gcc.dg/predict-20.c testcase relies on broken behavior of
profile estimation when trying to handle __builtin_expect values
flowing into PHI nodes. I have opened PR109210 and removed
the expected matching from the testcase.
PR tree-optimization/109170
* gimple-range-op.cc (cfn_pass_through_arg1): New.
(gimple_range_op_handler::maybe_builtin_call): Handle
__builtin_expect and similar via cfn_pass_through_arg1
and inspecting the calls fnspec.
* builtins.cc (builtin_fnspec): Handle BUILT_IN_EXPECT
and BUILT_IN_EXPECT_WITH_PROBABILITY.
* gcc.dg/Wuse-after-free-pr109170.c: New testcase.
* gcc.dg/tree-ssa/ssa-lim-21.c: Adjust.
* gcc.dg/predict-20.c: Likewise.
---
gcc/builtins.cc | 2 ++
gcc/gimple-range-op.cc | 32 ++++++++++++++++++-
.../gcc.dg/Wuse-after-free-pr109170.c | 15 +++++++++
gcc/testsuite/gcc.dg/predict-20.c | 3 +-
gcc/testsuite/gcc.dg/tree-ssa/ssa-lim-21.c | 7 ++--
5 files changed, 54 insertions(+), 5 deletions(-)
create mode 100644 gcc/testsuite/gcc.dg/Wuse-after-free-pr109170.c
diff --git a/gcc/builtins.cc b/gcc/builtins.cc
index 90246e214d6..56545027297 100644
--- a/gcc/builtins.cc
+++ b/gcc/builtins.cc
@@ -11715,6 +11715,8 @@ builtin_fnspec (tree callee)
case BUILT_IN_RETURN_ADDRESS:
return ".c";
case BUILT_IN_ASSUME_ALIGNED:
+ case BUILT_IN_EXPECT:
+ case BUILT_IN_EXPECT_WITH_PROBABILITY:
return "1cX ";
/* But posix_memalign stores a pointer into the memory pointed to
by its first argument. */
diff --git a/gcc/gimple-range-op.cc b/gcc/gimple-range-op.cc
index a5d625387e7..1a00f1690e5 100644
--- a/gcc/gimple-range-op.cc
+++ b/gcc/gimple-range-op.cc
@@ -43,6 +43,7 @@ along with GCC; see the file COPYING3. If not see
#include "range.h"
#include "value-query.h"
#include "gimple-range.h"
+#include "attr-fnspec.h"
// Given stmt S, fill VEC, up to VEC_SIZE elements, with relevant ssa-names
// on the statement. For efficiency, it is an error to not pass in enough
@@ -309,6 +310,26 @@ public:
}
} op_cfn_constant_p;
+// Implement range operator for integral/pointer functions returning
+// the first argument.
+class cfn_pass_through_arg1 : public range_operator
+{
+public:
+ using range_operator::fold_range;
+ virtual bool fold_range (irange &r, tree, const irange &lh,
+ const irange &, relation_trio) const
+ {
+ r = lh;
+ return true;
+ }
+ virtual bool op1_range (irange &r, tree, const irange &lhs,
+ const irange &, relation_trio) const
+ {
+ r = lhs;
+ return true;
+ }
+} op_cfn_pass_through_arg1;
+
// Implement range operator for CFN_BUILT_IN_SIGNBIT.
class cfn_signbit : public range_operator_float
{
@@ -967,6 +988,15 @@ gimple_range_op_handler::maybe_builtin_call ()
break;
default:
- break;
+ {
+ unsigned arg;
+ if (gimple_call_fnspec (call).returns_arg (&arg) && arg == 0)
+ {
+ m_valid = true;
+ m_op1 = gimple_call_arg (call, 0);
+ m_int = &op_cfn_pass_through_arg1;
+ }
+ break;
+ }
}
}
diff --git a/gcc/testsuite/gcc.dg/Wuse-after-free-pr109170.c
b/gcc/testsuite/gcc.dg/Wuse-after-free-pr109170.c
new file mode 100644
index 00000000000..fa7dc66d66c
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/Wuse-after-free-pr109170.c
@@ -0,0 +1,15 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -Wuse-after-free" } */
+
+unsigned long bufmax = 0;
+unsigned long __open_catalog_bufmax;
+void *realloc(void *, __SIZE_TYPE__);
+void free(void *);
+
+void __open_catalog(char *buf)
+{
+ char *old_buf = buf;
+ buf = realloc (buf, bufmax);
+ if (__builtin_expect ((buf == ((void *)0)), 0))
+ free (old_buf); /* { dg-bogus "used after" } */
+}
diff --git a/gcc/testsuite/gcc.dg/predict-20.c
b/gcc/testsuite/gcc.dg/predict-20.c
index 31d01835b80..7bb0d411f88 100644
--- a/gcc/testsuite/gcc.dg/predict-20.c
+++ b/gcc/testsuite/gcc.dg/predict-20.c
@@ -16,8 +16,9 @@ c ()
break;
}
int d = b < 0;
+ /* We fail to apply __builtin_expect heuristics here. Se PR109210. */
if (__builtin_expect (d, 0))
asm("");
}
-/* { dg-final { scan-tree-dump-times "__builtin_expect heuristics of edge" 3
"profile_estimate"} } */
+/* { dg-final { scan-tree-dump-times "__builtin_expect heuristics of edge" 2
"profile_estimate" } } */
diff --git a/gcc/testsuite/gcc.dg/tree-ssa/ssa-lim-21.c
b/gcc/testsuite/gcc.dg/tree-ssa/ssa-lim-21.c
index ffe6f8f699d..fe29e841f28 100644
--- a/gcc/testsuite/gcc.dg/tree-ssa/ssa-lim-21.c
+++ b/gcc/testsuite/gcc.dg/tree-ssa/ssa-lim-21.c
@@ -17,7 +17,7 @@ void
func (int m, int n, int k, struct obj *a)
{
struct obj *q = a;
- for (int j = 0; j < m; j++)
+ for (int j = 0; j < n; j++)
if (__builtin_expect (m, 0))
for (int i = 0; i < m; i++)
{
@@ -31,5 +31,6 @@ func (int m, int n, int k, struct obj *a)
}
}
-/* { dg-final { scan-tree-dump-not "Executing store motion of" "lim2" } } */
-
+/* { dg-final { scan-tree-dump "Executing store motion of count from loop 2"
"lim2" } } */
+/* { dg-final { scan-tree-dump "Executing store motion of \[^ \]*data1 from
loop 2" "lim2" } } */
+/* { dg-final { scan-tree-dump-times "Executing store motion of" 2 "lim2" }
} */
--
2.35.3
