* [PATCH] Fortran: improve bounds checking for DATA with implied-do [PR35095]
@ 2023-08-24 21:28 Harald Anlauf
2023-08-25 0:45 ` Jerry D
0 siblings, 1 reply; 2+ messages in thread
From: Harald Anlauf @ 2023-08-24 21:28 UTC (permalink / raw)
To: fortran, gcc-patches
[-- Attachment #1: Type: text/plain, Size: 505 bytes --]
Dear all,
the attached patch adds stricter bounds-checking for DATA statements
with implied-do. I chose to allow overindexing (for arrays of rank
greater than 1) for -std=legacy, as there might be codes in the wild
that need this (and this is accepted by some other compilers, while
NAG is strict here). We now get a warning with -std=gnu, and an
error with -std=fxxxx.
Regtested on x86_64-pc-linux-gnu. OK for mainline?
(The PR is over 15 years old, so no backport intended... ;-)
Thanks,
Harald
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: pr35095.diff --]
[-- Type: text/x-patch, Size: 4813 bytes --]
From 420804e7399dbc307a80f084cfb840444b8ebfe7 Mon Sep 17 00:00:00 2001
From: Harald Anlauf <anlauf@gmx.de>
Date: Thu, 24 Aug 2023 23:16:25 +0200
Subject: [PATCH] Fortran: improve bounds checking for DATA with implied-do
[PR35095]
gcc/fortran/ChangeLog:
PR fortran/35095
* data.cc (get_array_index): Add bounds-checking code and return error
status. Overindexing will be allowed as an extension for -std=legacy
and generate an error in standard-conforming mode.
(gfc_assign_data_value): Use error status from get_array_index for
graceful error recovery.
gcc/testsuite/ChangeLog:
PR fortran/35095
* gfortran.dg/data_bounds_1.f90: Adjust options to disable warnings.
* gfortran.dg/data_bounds_2.f90: New test.
---
gcc/fortran/data.cc | 47 ++++++++++++++++++---
gcc/testsuite/gfortran.dg/data_bounds_1.f90 | 2 +-
gcc/testsuite/gfortran.dg/data_bounds_2.f90 | 9 ++++
3 files changed, 51 insertions(+), 7 deletions(-)
create mode 100644 gcc/testsuite/gfortran.dg/data_bounds_2.f90
diff --git a/gcc/fortran/data.cc b/gcc/fortran/data.cc
index 7c2537dd3f0..0589fc3906f 100644
--- a/gcc/fortran/data.cc
+++ b/gcc/fortran/data.cc
@@ -43,13 +43,14 @@ static void formalize_init_expr (gfc_expr *);
/* Calculate the array element offset. */
-static void
+static bool
get_array_index (gfc_array_ref *ar, mpz_t *offset)
{
gfc_expr *e;
int i;
mpz_t delta;
mpz_t tmp;
+ bool ok = true;
mpz_init (tmp);
mpz_set_si (*offset, 0);
@@ -59,13 +60,42 @@ get_array_index (gfc_array_ref *ar, mpz_t *offset)
e = gfc_copy_expr (ar->start[i]);
gfc_simplify_expr (e, 1);
- if ((gfc_is_constant_expr (ar->as->lower[i]) == 0)
- || (gfc_is_constant_expr (ar->as->upper[i]) == 0)
- || (gfc_is_constant_expr (e) == 0))
- gfc_error ("non-constant array in DATA statement %L", &ar->where);
+ if (!gfc_is_constant_expr (ar->as->lower[i])
+ || !gfc_is_constant_expr (ar->as->upper[i])
+ || !gfc_is_constant_expr (e))
+ {
+ gfc_error ("non-constant array in DATA statement %L", &ar->where);
+ ok = false;
+ break;
+ }
mpz_set (tmp, e->value.integer);
gfc_free_expr (e);
+
+ /* Overindexing is only allowed as a legacy extension. */
+ if (mpz_cmp (tmp, ar->as->lower[i]->value.integer) < 0
+ && !gfc_notify_std (GFC_STD_LEGACY,
+ "Subscript at %L below array lower bound "
+ "(%ld < %ld) in dimension %d", &ar->c_where[i],
+ mpz_get_si (tmp),
+ mpz_get_si (ar->as->lower[i]->value.integer),
+ i+1))
+ {
+ ok = false;
+ break;
+ }
+ if (mpz_cmp (tmp, ar->as->upper[i]->value.integer) > 0
+ && !gfc_notify_std (GFC_STD_LEGACY,
+ "Subscript at %L above array upper bound "
+ "(%ld > %ld) in dimension %d", &ar->c_where[i],
+ mpz_get_si (tmp),
+ mpz_get_si (ar->as->upper[i]->value.integer),
+ i+1))
+ {
+ ok = false;
+ break;
+ }
+
mpz_sub (tmp, tmp, ar->as->lower[i]->value.integer);
mpz_mul (tmp, tmp, delta);
mpz_add (*offset, tmp, *offset);
@@ -77,6 +107,8 @@ get_array_index (gfc_array_ref *ar, mpz_t *offset)
}
mpz_clear (delta);
mpz_clear (tmp);
+
+ return ok;
}
/* Find if there is a constructor which component is equal to COM.
@@ -298,7 +330,10 @@ gfc_assign_data_value (gfc_expr *lvalue, gfc_expr *rvalue, mpz_t index,
}
if (ref->u.ar.type == AR_ELEMENT)
- get_array_index (&ref->u.ar, &offset);
+ {
+ if (!get_array_index (&ref->u.ar, &offset))
+ goto abort;
+ }
else
mpz_set (offset, index);
diff --git a/gcc/testsuite/gfortran.dg/data_bounds_1.f90 b/gcc/testsuite/gfortran.dg/data_bounds_1.f90
index 24cdc7c9815..1e6321a2884 100644
--- a/gcc/testsuite/gfortran.dg/data_bounds_1.f90
+++ b/gcc/testsuite/gfortran.dg/data_bounds_1.f90
@@ -1,5 +1,5 @@
! { dg-do compile }
-! { dg-options "-std=gnu" }
+! { dg-options "-std=gnu -w" }
! Checks the fix for PR32315, in which the bounds checks below were not being done.
!
! Contributed by Tobias Burnus <burnus@gcc.gnu.org>
diff --git a/gcc/testsuite/gfortran.dg/data_bounds_2.f90 b/gcc/testsuite/gfortran.dg/data_bounds_2.f90
new file mode 100644
index 00000000000..1aa9fd4c423
--- /dev/null
+++ b/gcc/testsuite/gfortran.dg/data_bounds_2.f90
@@ -0,0 +1,9 @@
+! { dg-do compile }
+! { dg-options "-std=f2018" }
+! PR fortran/35095 - Improve bounds checking for DATA with implied-do
+
+program chkdata
+ character(len=2), dimension(2,2) :: str
+ data (str(i,1),i=1,3) / 'A','B','C' / ! { dg-error "above array upper bound" }
+ data (str(j,2),j=0,2) / 'A','B','C' / ! { dg-error "below array lower bound" }
+end program chkdata
--
2.35.3
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] Fortran: improve bounds checking for DATA with implied-do [PR35095]
2023-08-24 21:28 [PATCH] Fortran: improve bounds checking for DATA with implied-do [PR35095] Harald Anlauf
@ 2023-08-25 0:45 ` Jerry D
0 siblings, 0 replies; 2+ messages in thread
From: Jerry D @ 2023-08-25 0:45 UTC (permalink / raw)
To: Harald Anlauf, fortran, gcc-patches
On 8/24/23 2:28 PM, Harald Anlauf via Fortran wrote:
> Dear all,
>
> the attached patch adds stricter bounds-checking for DATA statements
> with implied-do. I chose to allow overindexing (for arrays of rank
> greater than 1) for -std=legacy, as there might be codes in the wild
> that need this (and this is accepted by some other compilers, while
> NAG is strict here). We now get a warning with -std=gnu, and an
> error with -std=fxxxx.
>
> Regtested on x86_64-pc-linux-gnu. OK for mainline?
>
> (The PR is over 15 years old, so no backport intended... ;-)
>
> Thanks,
> Harald
>
Looks good Harold, OK for mainline.
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2023-08-25 0:45 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-08-24 21:28 [PATCH] Fortran: improve bounds checking for DATA with implied-do [PR35095] Harald Anlauf
2023-08-25 0:45 ` Jerry D
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).