public inbox for fortran@gcc.gnu.org
 help / color / mirror / Atom feed
* Re: [patch, fortran] PR66310 Problems with intrinsic repeat for large number of copies
@ 2016-07-12 13:26 Dominique d'Humières
  2016-07-12 18:34 ` Jerry DeLisle
  2016-07-18  0:03 ` Jerry DeLisle
  0 siblings, 2 replies; 9+ messages in thread
From: Dominique d'Humières @ 2016-07-12 13:26 UTC (permalink / raw)
  To: jvdelisle; +Cc: FX Coudert, fortran, gcc-patches

> > 2016-07-11  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
> > 
> > 	PR fortran/66310
> > 	* simplify.c (gfc_simplify_repeat): Set max repeat to huge - 1 to allow
> > 	one byte for null terminating the resulting string constant.
>
> OK, thanks
Please hold on. I still see several problem with the patch applied. One is

   program p
      character :: z = 'z'
      print *, repeat(z, huge(1)-2**9)
   end

a.out(67209,0x7fff77e0b000) malloc: *** mach_vm_map(size=18446744071562067968) failed (error code=3)
*** error: can't allocate region
*** set a breakpoint in malloc_error_break to debug
Operating system error: Cannot allocate memory
Memory allocation failure in realloc

on x86_64-apple-darwin15.5 with/without the patch. print *, repeat(z, huge(1)-2**9-1) "works".

I’ll try to report the other problems in the PR later today.

Dominique


^ permalink raw reply	[flat|nested] 9+ messages in thread
* [patch, fortran] PR66310 Problems with intrinsic repeat for large number of copies
@ 2016-07-12  0:54 Jerry DeLisle
  2016-07-12 12:57 ` FX
  2016-07-13  1:54 ` William Clodius
  0 siblings, 2 replies; 9+ messages in thread
From: Jerry DeLisle @ 2016-07-12  0:54 UTC (permalink / raw)
  To: gfortran, gcc patches

Attached patch sets a limit of huge-1 on the character count to avoid the
integer wrap.

Regression tested with adjustment to test case.

OK for trunk?

Regards,

Jerry

2016-07-11  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

	PR fortran/66310
	* simplify.c (gfc_simplify_repeat): Set max repeat to huge - 1 to allow
	one byte for null terminating the resulting string constant.

diff --git a/gcc/fortran/simplify.c b/gcc/fortran/simplify.c
index 95a8d10..03257ec 100644
--- a/gcc/fortran/simplify.c
+++ b/gcc/fortran/simplify.c
@@ -5081,22 +5081,27 @@ gfc_simplify_repeat (gfc_expr *e, gfc_expr *n)
   /* Check that NCOPIES isn't too large.  */
   if (len)
     {
-      mpz_t max, mlen;
+      mpz_t max, mlen, limit;
       int i;

-      /* Compute the maximum value allowed for NCOPIES: huge(cl) / len.  */
+      /* Compute the maximum value allowed for NCOPIES:
+           huge(cl) - 1 / len.  */
       mpz_init (max);
+      mpz_init (limit);
       i = gfc_validate_kind (BT_INTEGER, gfc_charlen_int_kind, false);

+      /* Set the limit on size to huge-1 to avoid unsigned integer
+        wrapping in gfc_get_character_expr.  */
+      mpz_sub_ui (limit, gfc_integer_kinds[i].huge, 1);
+
       if (have_length)
        {
-         mpz_tdiv_q (max, gfc_integer_kinds[i].huge,
-                     e->ts.u.cl->length->value.integer);
+         mpz_tdiv_q (max, limit, e->ts.u.cl->length->value.integer);
        }
       else
        {
          mpz_init_set_si (mlen, len);
-         mpz_tdiv_q (max, gfc_integer_kinds[i].huge, mlen);
+         mpz_tdiv_q (max, limit, mlen);
          mpz_clear (mlen);
        }

@@ -5104,6 +5109,7 @@ gfc_simplify_repeat (gfc_expr *e, gfc_expr *n)
       if (mpz_cmp (ncopies, max) > 0)
        {
          mpz_clear (max);
+         mpz_clear (limit);
          mpz_clear (ncopies);
          gfc_error ("Argument NCOPIES of REPEAT intrinsic is too large at %L",
                     &n->where);
@@ -5111,6 +5117,7 @@ gfc_simplify_repeat (gfc_expr *e, gfc_expr *n)
        }

       mpz_clear (max);
+      mpz_clear (limit);
     }
   mpz_clear (ncopies);

diff --git a/gcc/testsuite/gfortran.dg/repeat_4.f90
b/gcc/testsuite/gfortran.dg/repeat_4.f90
index e5b5acc..77483a1 100644
--- a/gcc/testsuite/gfortran.dg/repeat_4.f90
+++ b/gcc/testsuite/gfortran.dg/repeat_4.f90
@@ -22,7 +22,8 @@ program test

   ! Check for too large NCOPIES argument and limit cases
   print *, repeat(t0, huge(0))
-  print *, repeat(t1, huge(0))
+  print *, repeat(t1, huge(0)) ! { dg-error "Argument NCOPIES of REPEAT
intrinsic is too large " }
+  print *, repeat(t1, huge(0) - 1)
   print *, repeat(t2, huge(0)) ! { dg-error "Argument NCOPIES of REPEAT
intrinsic is too large " }
   print *, repeat(s2, huge(0)) ! { dg-error "Argument NCOPIES of REPEAT
intrinsic is too large " }

^ permalink raw reply	[flat|nested] 9+ messages in thread
end of thread, other threads:[~2016-07-19 16:43 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-07-12 13:26 [patch, fortran] PR66310 Problems with intrinsic repeat for large number of copies Dominique d'Humières
2016-07-12 18:34 ` Jerry DeLisle
2016-07-18  0:03 ` Jerry DeLisle
2016-07-19  9:56   ` Dominique d'Humières
2016-07-19 16:43     ` Jerry DeLisle
  -- strict thread matches above, loose matches on Subject: below --
2016-07-12  0:54 Jerry DeLisle
2016-07-12 12:57 ` FX
2016-07-13  1:54 ` William Clodius
2016-07-13  3:22   ` Jerry DeLisle

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).