Re: [PATCH] fold-const-call.cc: add strnlen in fold_const_call [PR86937]

[Daniel Henrique Barboza]

On 8/14/2026 9:08 AM, Daniel Henrique Barboza wrote:

Hi,


On 8/12/2026 8:48 PM, Andrea Pinski wrote:

On Wed, Aug 12, 2026 at 11:01 AM Daniel Barboza
<daniel.barb...@oss.qualcomm.com> wrote:

We're not trying to fold builtin_strnlen calls into constants like we do
with builtin_strlen.  This results in missed optimizations where we
have a strnlen that uses a PHI as input:

const char a[4] = "123";

int g (int i)
{
   return __builtin_strnlen (i ? a : "", 4);
}

The strnlen call could be folded into a single PHI <0, 3> but instead
we're calling strnlen with the PHI result:

   # iftmp.1_2 = PHI <&aD.4472(3), ""(2)>
   # VUSE <.MEM_4(D)>
   # RANGE [irange] long unsigned int [0, 4] MASK 0x7 VALUE 0x0
   # USE = nonlocal escaped const-pool { D.4472 } (nonlocal)
   _1 = strnlenD.1862 (iftmp.1_2, 4);
---
  gcc/fold-const-call.cc                     | 11 ++++++++++
  gcc/testsuite/gcc.dg/tree-ssa/pr86937.c    | 15 +++++++++++++
  gcc/testsuite/gcc.dg/warn-strnlen-no-nul.c | 25 +++++++++++-----------
  3 files changed, 39 insertions(+), 12 deletions(-)
  create mode 100644 gcc/testsuite/gcc.dg/tree-ssa/pr86937.c

diff --git a/gcc/fold-const-call.cc b/gcc/fold-const-call.cc
index b3a9d7715d6..c483b80e704 100644
--- a/gcc/fold-const-call.cc
+++ b/gcc/fold-const-call.cc
@@ -1911,6 +1911,17 @@ fold_const_call (combined_fn fn, tree type, tree arg0, 
tree arg1)
         }
        return NULL_TREE;

+    case CFN_BUILT_IN_STRNLEN:
+      if ((p0 = c_getstr (arg0)))
+       {
+         unsigned HOST_WIDE_INT s1 = 0;
+         if (!size_t_cst_p (arg1, &s1))
+           return NULL_TREE;
+
+         return build_int_cst (type, strnlen (p0, s1));

+       }
+      return NULL_TREE;
+
      case CFN_FOLD_LEFT_PLUS:
        return fold_const_fold_left (type, arg0, arg1, PLUS_EXPR);

diff --git a/gcc/testsuite/gcc.dg/tree-ssa/pr86937.c 
b/gcc/testsuite/gcc.dg/tree-ssa/pr86937.c
new file mode 100644
index 00000000000..fcd1c9d0eb7
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/tree-ssa/pr86937.c
@@ -0,0 +1,15 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -fdump-tree-pre" } */
+const char a[4] = "123";
+
+int f (int i)
+{
+  return __builtin_strlen (i ? a : "");
+}
+
+int g (int i)
+{
+  return __builtin_strnlen (i ? a : "", 4);
+}
+
+/* { dg-final { scan-tree-dump-times "strnlen" 0 "pre" } } */

Please add a space after strnlen just in case it matches something
else. Basically there should be no calls left to strnlen. The dumps
are always `functionname (...)`.

The output this test produces is :

_1 = strnlenD.1862 (iftmp.1_2, 4);

So adding a space after won't match the dump.

Grepping the existing files I see that gcc.dg/strlenopt-46.c does:

{ dg-final { scan-tree-dump-times "= strnlen" 7 "optimized" } } */


Thus I think we can do "= strnlen" here too.  Thanks,

Nevermind that :D using the actual testsuite the output will be

  _1 = __builtin_strnlen (iftmp.1_2, 4);


Your suggestion was correct.  I'll make the change for v2.


Cheers,
Daniel

Daniel

diff --git a/gcc/testsuite/gcc.dg/warn-strnlen-no-nul.c 
b/gcc/testsuite/gcc.dg/warn-strnlen-no-nul.c
index 70f6a432b97..fde2fa0e189 100644
--- a/gcc/testsuite/gcc.dg/warn-strnlen-no-nul.c
+++ b/gcc/testsuite/gcc.dg/warn-strnlen-no-nul.c
@@ -147,14 +147,14 @@ T (v0 ? b[3] : "", bsz);
     the strnlen calls are safe because the reads are bounded by
     the length of the constant arguments.  Most of the calls are
     not diagnosed anymore as a result of the fix for PR 103215.  */
-T (v0 ? "" : b[0], bsz + 1);      /* { dg-warning "bound 6 exceeds source size 
5" } */
+T (v0 ? "" : b[0], bsz + 1);      /* { dg-warning "bound 6 exceeds source size 5" 
"pr86937" { xfail *-*-* } } */

So if I read the comment, I think the dg-warning can just be removed.

  T (v0 ? "" : b[1], bsz + 1);
  T (v0 ? "" : b[2], bsz + 1);
-T (v0 ? "" : b[3], bsz + 1);
-T (v0 ? b[0] : "", bsz + 1);      /* { dg-warning "bound 6 exceeds source size 
5" } */

Remove this warning rather than changing xfailing.

+T (v0 ? "" : b[3], bsz + 1);      /* { dg-warning "unterminated" } */

This new warning is correct as b[3] is "54321" in a array size of 5.

+T (v0 ? b[0] : "", bsz + 1);      /* { dg-warning "bound 6 exceeds source size 5" 
"pr86937" { xfail *-*-* } } */
  T (v0 ? b[1] : "", bsz + 1);
  T (v0 ? b[2] : "", bsz + 1);
-T (v0 ? b[3] : "", bsz + 1);
+T (v0 ? b[3] : "", bsz + 1);      /* { dg-warning "unterminated" } */

Likewise.

  T (v0 ? "" : b[i0], bsz);
  T (v0 ? "" : b[i1], bsz);
@@ -168,11 +168,11 @@ T (v0 ? b[i3] : "", bsz);
  T (v0 ? "" : b[i0], bsz + 1);
  T (v0 ? "" : b[i1], bsz + 1);
  T (v0 ? "" : b[i2], bsz + 1);
-T (v0 ? "" : b[i3], bsz + 1);
+T (v0 ? "" : b[i3], bsz + 1);     /* { dg-warning "unterminated" } */

Likewise.

  T (v0 ? b[i0] : "", bsz + 1);
  T (v0 ? b[i1] : "", bsz + 1);
  T (v0 ? b[i2] : "", bsz + 1);
-T (v0 ? b[i3] : "", bsz + 1);
+T (v0 ? b[i3] : "", bsz + 1);     /* { dg-warning "unterminated" } */

Likewise.

  T (v0 ? "1234" : b[3], bsz);
  T (v0 ? "1234" : b[i3], bsz);
@@ -184,17 +184,18 @@ T (v0 ? b[0] : b[2], bsz);
  T (v0 ? b[2] : b[3], bsz);
  T (v0 ? b[3] : b[2], bsz);

-T (v0 ? "1234" : b[3], bsz + 1);
-T (v0 ? "1234" : b[i3], bsz + 1);
-T (v0 ? b[3] : "1234", bsz + 1);
-T (v0 ? b[i3] : "1234", bsz + 1);
+/* New warnings being thrown after PR86937.  */
+T (v0 ? "1234" : b[3], bsz + 1);     /* { dg-warning "unterminated" } */
+T (v0 ? "1234" : b[i3], bsz + 1);    /* { dg-warning "unterminated" } */
+T (v0 ? b[3] : "1234", bsz + 1);     /* { dg-warning "unterminated" } */
+T (v0 ? b[i3] : "1234", bsz + 1);    /* { dg-warning "unterminated" } */

These are all valid warnings.

  /* That the following are not diagnosed is a bug/limitation resulting from
     the fix for PR 103215.  */
  T (v0 ? a : b[3], bsz + 1);       /* { dg-warning "bound 6 exceeds source size 5" 
"pr103215" { xfail *-*-* } } */
  T (v0 ? b[0] : b[2], bsz + 1);    /* { dg-warning "bound 6 exceeds source size 5" 
"pr103215" { xfail *-*-* } } */
-T (v0 ? b[2] : b[3], bsz + 1);    /* { dg-warning "bound 6 exceeds source size 5" 
"pr103215" { xfail *-*-* } } */
-T (v0 ? b[3] : b[2], bsz + 1);    /* { dg-warning "bound 6 exceeds source size 5" 
"pr103215" { xfail *-*-* } } */
+T (v0 ? b[2] : b[3], bsz + 1);    /* { dg-warning "unterminated" } */
+T (v0 ? b[3] : b[2], bsz + 1);    /* { dg-warning "unterminated" } */

Yes the same new valid warning b[3] is unterminated.

So ok with the removal of the newly xfailed dg-warning since those
warnings won't show up ever since the values are all calculated at
compile time.

Thanks,
Andrea

  struct A { char a[5], b[5]; };

--
2.43.0