[1/2] PR 92463 MPFR modernization in GFortran
diff mbox series

Message ID 20191118203450.8277-1-blomqvist.janne@gmail.com
State New
Headers show
Series
  • [1/2] PR 92463 MPFR modernization in GFortran
Related show

Commit Message

Janne Blomqvist Nov. 18, 2019, 8:34 p.m. UTC
Now that we require a minimum of MPFR 3.1.0+ to build GCC, we can do
some modernization of the MPFR usage in the GFortran frontend.

This patch replaces

1) GMP_RND* with MPFR_RND*

2) mp_exp_t with mpfr_exp_t

3) mp_prec_t with mpfr_prec_t

4) mp_rnd_t with mpfr_rnd_t

gcc/fortran/ChangeLog:

2019-11-18  Janne Blomqvist  <jb@gcc.gnu.org>

	PR fortran/92463
	* arith.c (gfc_mpfr_to_mpz): Change mp_exp_t to mpfr_exp_t.
	(gfc_check_real_range): Likewise.
	* gfortran.h (GFC_RND_MODE): Change GMP_RNDN to MPFR_RNDN.
	* module.c (mio_gmp_real): Change mp_exp_t to mpfr_exp_t.
	* simplify.c (degrees_f): Change mp_rnd_t to mpfr_rnd_t.
	(radians_f): Likewise.
	(fullprec_erfc_scaled): Change mp_prec_t to mpfr_prec_t.
	(asympt_erfc_scaled): Likewise.
	(gfc_simplify_nearest): Change mp_exp_t to mpfr_exp_t, and
	GMP_RND* to MPFR_RND*.
---
 gcc/fortran/arith.c    |  8 ++++----
 gcc/fortran/gfortran.h |  2 +-
 gcc/fortran/module.c   |  2 +-
 gcc/fortran/simplify.c | 20 ++++++++++----------
 4 files changed, 16 insertions(+), 16 deletions(-)

Comments

Tobias Burnus Nov. 20, 2019, 3:45 p.m. UTC | #1
Hi Janne,

On 11/18/19 9:34 PM, Janne Blomqvist wrote:
> Now that we require a minimum of MPFR 3.1.0+ to build GCC, we can do
> some modernization of the MPFR usage in the GFortran frontend.

OK – thanks for the cleanup.

[For reference, those changes were introduced in MPFR 3.0.0, cf. 
https://www.mpfr.org/mpfr-3.0.0/#changes ; 3.0 and 3.1 also added a 
bunch of new functions, MPFR_RNDA and did remove some undefined behaviour.]

Do you intent update the other places as well? Namely:

> This patch replaces
> 1) GMP_RND* with MPFR_RND*
Also used by: configure.ac, gcc/builtins.c, gcc/fold-const-call.c, 
gcc/gimple-ssa-sprintf.c, gcc/go/…, gcc/real.c, gcc/realmpfr.h, 
gcc/ubsan.c.

2) mp_exp_t with mpfr_exp_t

gcc/realmpfr.c + gcc/go/…

> 3) mp_prec_t with mpfr_prec_t
(only used by gcc/fortran)
> 4) mp_rnd_t with mpfr_rnd_t

gcc/builtins.c, gcc/fold-const-call.c, gcc/realmpfr.c

Cheers,

Tobias

> gcc/fortran/ChangeLog:
>
> 2019-11-18  Janne Blomqvist  <jb@gcc.gnu.org>
>
> 	PR fortran/92463
> 	* arith.c (gfc_mpfr_to_mpz): Change mp_exp_t to mpfr_exp_t.
> 	(gfc_check_real_range): Likewise.
> 	* gfortran.h (GFC_RND_MODE): Change GMP_RNDN to MPFR_RNDN.
> 	* module.c (mio_gmp_real): Change mp_exp_t to mpfr_exp_t.
> 	* simplify.c (degrees_f): Change mp_rnd_t to mpfr_rnd_t.
> 	(radians_f): Likewise.
> 	(fullprec_erfc_scaled): Change mp_prec_t to mpfr_prec_t.
> 	(asympt_erfc_scaled): Likewise.
> 	(gfc_simplify_nearest): Change mp_exp_t to mpfr_exp_t, and
> 	GMP_RND* to MPFR_RND*.
> ---
>   gcc/fortran/arith.c    |  8 ++++----
>   gcc/fortran/gfortran.h |  2 +-
>   gcc/fortran/module.c   |  2 +-
>   gcc/fortran/simplify.c | 20 ++++++++++----------
>   4 files changed, 16 insertions(+), 16 deletions(-)
Janne Blomqvist Nov. 20, 2019, 8:04 p.m. UTC | #2
On Wed, Nov 20, 2019 at 5:46 PM Tobias Burnus <Tobias_Burnus@mentor.com> wrote:
>
> Hi Janne,
>
> On 11/18/19 9:34 PM, Janne Blomqvist wrote:
> > Now that we require a minimum of MPFR 3.1.0+ to build GCC, we can do
> > some modernization of the MPFR usage in the GFortran frontend.
>
> OK – thanks for the cleanup.
>
> [For reference, those changes were introduced in MPFR 3.0.0, cf.
> https://www.mpfr.org/mpfr-3.0.0/#changes ; 3.0 and 3.1 also added a
> bunch of new functions, MPFR_RNDA and did remove some undefined behaviour.]
>
> Do you intent update the other places as well? Namely:
>
> > This patch replaces
> > 1) GMP_RND* with MPFR_RND*
> Also used by: configure.ac, gcc/builtins.c, gcc/fold-const-call.c,
> gcc/gimple-ssa-sprintf.c, gcc/go/…, gcc/real.c, gcc/realmpfr.h,
> gcc/ubsan.c.
>
> 2) mp_exp_t with mpfr_exp_t
>
> gcc/realmpfr.c + gcc/go/…
>
> > 3) mp_prec_t with mpfr_prec_t
> (only used by gcc/fortran)
> > 4) mp_rnd_t with mpfr_rnd_t
>
> gcc/builtins.c, gcc/fold-const-call.c, gcc/realmpfr.c
>

Thanks for the review, committed as r278523. My plan was that once the
fortran-specific parts are done, I'll reassign the PR to the
middle-end, and then me or somebody else can make a separate patch for
that. However, since we're in stage3 now I believe this will have to
wait until after the GCC 10 release as the middle-end people seem to
take these rules more seriously.

Patch
diff mbox series

diff --git a/gcc/fortran/arith.c b/gcc/fortran/arith.c
index a40163e8a2d..cf480779709 100644
--- a/gcc/fortran/arith.c
+++ b/gcc/fortran/arith.c
@@ -38,7 +38,7 @@  along with GCC; see the file COPYING3.  If not see
 void
 gfc_mpfr_to_mpz (mpz_t z, mpfr_t x, locus *where)
 {
-  mp_exp_t e;
+  mpfr_exp_t e;
 
   if (mpfr_inf_p (x) || mpfr_nan_p (x))
     {
@@ -376,7 +376,7 @@  gfc_check_real_range (mpfr_t p, int kind)
     }
   else if (mpfr_cmp (q, gfc_real_kinds[i].tiny) < 0)
     {
-      mp_exp_t emin, emax;
+      mpfr_exp_t emin, emax;
       int en;
 
       /* Save current values of emin and emax.  */
@@ -396,9 +396,9 @@  gfc_check_real_range (mpfr_t p, int kind)
 
       /* Copy sign if needed.  */
       if (mpfr_sgn (p) < 0)
-	mpfr_neg (p, q, GMP_RNDN);
+	mpfr_neg (p, q, MPFR_RNDN);
       else
-	mpfr_set (p, q, GMP_RNDN);
+	mpfr_set (p, q, MPFR_RNDN);
     }
 
   mpfr_clear (q);
diff --git a/gcc/fortran/gfortran.h b/gcc/fortran/gfortran.h
index e962db59bc5..f4a2b99bdc4 100644
--- a/gcc/fortran/gfortran.h
+++ b/gcc/fortran/gfortran.h
@@ -2130,7 +2130,7 @@  gfc_intrinsic_sym;
 
 #include <mpfr.h>
 #include <mpc.h>
-#define GFC_RND_MODE GMP_RNDN
+#define GFC_RND_MODE MPFR_RNDN
 #define GFC_MPC_RND_MODE MPC_RNDNN
 
 typedef splay_tree gfc_constructor_base;
diff --git a/gcc/fortran/module.c b/gcc/fortran/module.c
index 755f237a0e7..b11a30bea73 100644
--- a/gcc/fortran/module.c
+++ b/gcc/fortran/module.c
@@ -3343,7 +3343,7 @@  mio_gmp_integer (mpz_t *integer)
 static void
 mio_gmp_real (mpfr_t *real)
 {
-  mp_exp_t exponent;
+  mpfr_exp_t exponent;
   char *p;
 
   if (iomode == IO_INPUT)
diff --git a/gcc/fortran/simplify.c b/gcc/fortran/simplify.c
index 0461d31cd88..a5c940ca2d5 100644
--- a/gcc/fortran/simplify.c
+++ b/gcc/fortran/simplify.c
@@ -1768,7 +1768,7 @@  simplify_trig_call (gfc_expr *icall)
 /* Convert a floating-point number from radians to degrees.  */
 
 static void
-degrees_f (mpfr_t x, mp_rnd_t rnd_mode)
+degrees_f (mpfr_t x, mpfr_rnd_t rnd_mode)
 {
   mpfr_t tmp;
   mpfr_init (tmp);
@@ -1791,7 +1791,7 @@  degrees_f (mpfr_t x, mp_rnd_t rnd_mode)
 /* Convert a floating-point number from degrees to radians.  */
 
 static void
-radians_f (mpfr_t x, mp_rnd_t rnd_mode)
+radians_f (mpfr_t x, mpfr_rnd_t rnd_mode)
 {
   mpfr_t tmp;
   mpfr_init (tmp);
@@ -2681,7 +2681,7 @@  gfc_simplify_erfc (gfc_expr *x)
 static void
 fullprec_erfc_scaled (mpfr_t res, mpfr_t arg)
 {
-  mp_prec_t prec;
+  mpfr_prec_t prec;
   mpfr_t a, b;
 
   prec = mpfr_get_default_prec ();
@@ -2718,7 +2718,7 @@  asympt_erfc_scaled (mpfr_t res, mpfr_t arg)
 {
   mpfr_t sum, x, u, v, w, oldsum, sumtrunc;
   mpz_t num;
-  mp_prec_t prec;
+  mpfr_prec_t prec;
   unsigned i;
 
   prec = mpfr_get_default_prec ();
@@ -5914,7 +5914,7 @@  gfc_expr *
 gfc_simplify_nearest (gfc_expr *x, gfc_expr *s)
 {
   gfc_expr *result;
-  mp_exp_t emin, emax;
+  mpfr_exp_t emin, emax;
   int kind;
 
   if (x->expr_type != EXPR_CONSTANT || s->expr_type != EXPR_CONSTANT)
@@ -5928,20 +5928,20 @@  gfc_simplify_nearest (gfc_expr *x, gfc_expr *s)
 
   /* Set emin and emax for the current model number.  */
   kind = gfc_validate_kind (BT_REAL, x->ts.kind, 0);
-  mpfr_set_emin ((mp_exp_t) gfc_real_kinds[kind].min_exponent -
+  mpfr_set_emin ((mpfr_exp_t) gfc_real_kinds[kind].min_exponent -
 		mpfr_get_prec(result->value.real) + 1);
-  mpfr_set_emax ((mp_exp_t) gfc_real_kinds[kind].max_exponent - 1);
-  mpfr_check_range (result->value.real, 0, GMP_RNDU);
+  mpfr_set_emax ((mpfr_exp_t) gfc_real_kinds[kind].max_exponent - 1);
+  mpfr_check_range (result->value.real, 0, MPFR_RNDU);
 
   if (mpfr_sgn (s->value.real) > 0)
     {
       mpfr_nextabove (result->value.real);
-      mpfr_subnormalize (result->value.real, 0, GMP_RNDU);
+      mpfr_subnormalize (result->value.real, 0, MPFR_RNDU);
     }
   else
     {
       mpfr_nextbelow (result->value.real);
-      mpfr_subnormalize (result->value.real, 0, GMP_RNDD);
+      mpfr_subnormalize (result->value.real, 0, MPFR_RNDD);
     }
 
   mpfr_set_emin (emin);