* [PATCH] Fix ldbl-128ibm fma (Inf, Inf, finite) (bug 23272)
@ 2018-06-13 22:01 Tulio Magno Quites Machado Filho
2018-06-13 23:07 ` Joseph Myers
0 siblings, 1 reply; 11+ messages in thread
From: Tulio Magno Quites Machado Filho @ 2018-06-13 22:01 UTC (permalink / raw
To: libc-alpha
This solution is mainly based on Joseph's fix at commit
ca121b117f2c9c97a4c121334481a96c94fef3a0.
It has extra differences because:
- part of the non-finite arguments were already being treated;
- when x and y are +-Inf and z if finite, an overflow can be
generated.
Tested on powerpc[|64|64le].
2018-06-13 Tulio Magno Quites Machado Filho <tuliom@linux.ibm.com>
[BZ 23272]
* sysdeps/ieee754/ldbl-128ibm/s_fmal.c (__fmal): Handle all
cases of non-finite arguments.
Signed-off-by: Tulio Magno Quites Machado Filho <tuliom@linux.ibm.com>
---
sysdeps/ieee754/ldbl-128ibm/s_fmal.c | 31 ++++++++++++-------------------
1 file changed, 12 insertions(+), 19 deletions(-)
diff --git a/sysdeps/ieee754/ldbl-128ibm/s_fmal.c b/sysdeps/ieee754/ldbl-128ibm/s_fmal.c
index e72a3e4d59..afba1ca727 100644
--- a/sysdeps/ieee754/ldbl-128ibm/s_fmal.c
+++ b/sysdeps/ieee754/ldbl-128ibm/s_fmal.c
@@ -127,37 +127,30 @@ long double
__fmal (long double x, long double y, long double z)
{
double xhi, xlo, yhi, ylo, zhi, zlo;
- int64_t hx, hy, hz;
- int xexp, yexp, zexp;
double scale_val;
int scale_exp;
ldbl_unpack (x, &xhi, &xlo);
- EXTRACT_WORDS64 (hx, xhi);
- xexp = (hx & 0x7ff0000000000000LL) >> 52;
ldbl_unpack (y, &yhi, &ylo);
- EXTRACT_WORDS64 (hy, yhi);
- yexp = (hy & 0x7ff0000000000000LL) >> 52;
ldbl_unpack (z, &zhi, &zlo);
- EXTRACT_WORDS64 (hz, zhi);
- zexp = (hz & 0x7ff0000000000000LL) >> 52;
- /* If z is Inf or NaN, but x and y are finite, avoid any exceptions
- from computing x * y. */
- if (zexp == 0x7ff && xexp != 0x7ff && yexp != 0x7ff)
+ if (__glibc_unlikely (!isfinite (x) || !isfinite (y) || x == 0 || y == 0))
+ if (!isfinite (x) && !isfinite (y) && isfinite(z))
+ /* Compute the result as x * y to avoid an overflow. */
+ return x * y;
+ else
+ /* If x or y is Inf, NaN or zero compute as x * y + z. */
+ return (x * y) + z;
+ else if (__glibc_unlikely (!isfinite (z)))
+ /* If z is Inf, but x and y are finite, the result should be z
+ rather than NaN. */
return (z + x) + y;
- /* If z is zero and x are y are nonzero, compute the result as x * y
+ /* If z is zero and x and y are nonzero, compute the result as x * y
to avoid the wrong sign of a zero result if x * y underflows to
0. */
- if (z == 0 && x != 0 && y != 0)
+ if (z == 0)
return x * y;
- /* If x or y or z is Inf/NaN, or if x * y is zero, compute as x * y
- + z. */
- if (xexp == 0x7ff || yexp == 0x7ff || zexp == 0x7ff
- || x == 0 || y == 0)
- return (x * y) + z;
-
{
SET_RESTORE_ROUND (FE_TONEAREST);
--
2.14.4
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH] Fix ldbl-128ibm fma (Inf, Inf, finite) (bug 23272)
2018-06-13 22:01 [PATCH] Fix ldbl-128ibm fma (Inf, Inf, finite) (bug 23272) Tulio Magno Quites Machado Filho
@ 2018-06-13 23:07 ` Joseph Myers
2018-09-21 11:59 ` Florian Weimer
2019-01-14 20:09 ` [PATCH] Add XFAIL_ROUNDING_IBM128_LIBGCC to more fma() tests Tulio Magno Quites Machado Filho
0 siblings, 2 replies; 11+ messages in thread
From: Joseph Myers @ 2018-06-13 23:07 UTC (permalink / raw
To: Tulio Magno Quites Machado Filho; +Cc: libc-alpha
On Wed, 13 Jun 2018, Tulio Magno Quites Machado Filho wrote:
> This solution is mainly based on Joseph's fix at commit
> ca121b117f2c9c97a4c121334481a96c94fef3a0.
>
> It has extra differences because:
> - part of the non-finite arguments were already being treated;
> - when x and y are +-Inf and z if finite, an overflow can be
> generated.
I'm not clear what the observed issues you are fixing are. Is this patch
to avoid running into libgcc generating spurious overflows for the case
you mention, or is it intended to fix wrong results?
> + if (!isfinite (x) && !isfinite (y) && isfinite(z))
> + /* Compute the result as x * y to avoid an overflow. */
> + return x * y;
Why wouldn't this be needed if just one of x and y is not finite? If x *
y + z generates a spurious overflow in some cases where x * y is infinite,
I'd expect that to apply equally to the case where the infinity comes from
(Inf * finite), not just from (Inf * Inf).
--
Joseph S. Myers
joseph@codesourcery.com
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] Fix ldbl-128ibm fma (Inf, Inf, finite) (bug 23272)
2018-06-13 23:07 ` Joseph Myers
@ 2018-09-21 11:59 ` Florian Weimer
2018-09-21 12:46 ` Joseph Myers
2019-01-14 20:09 ` [PATCH] Add XFAIL_ROUNDING_IBM128_LIBGCC to more fma() tests Tulio Magno Quites Machado Filho
1 sibling, 1 reply; 11+ messages in thread
From: Florian Weimer @ 2018-09-21 11:59 UTC (permalink / raw
To: Joseph Myers; +Cc: Tulio Magno Quites Machado Filho, libc-alpha
* Joseph Myers:
> On Wed, 13 Jun 2018, Tulio Magno Quites Machado Filho wrote:
>
>> This solution is mainly based on Joseph's fix at commit
>> ca121b117f2c9c97a4c121334481a96c94fef3a0.
>>
>> It has extra differences because:
>> - part of the non-finite arguments were already being treated;
>> - when x and y are +-Inf and z if finite, an overflow can be
>> generated.
>
> I'm not clear what the observed issues you are fixing are. Is this patch
> to avoid running into libgcc generating spurious overflows for the case
> you mention, or is it intended to fix wrong results?
>
>> + if (!isfinite (x) && !isfinite (y) && isfinite(z))
>> + /* Compute the result as x * y to avoid an overflow. */
>> + return x * y;
>
> Why wouldn't this be needed if just one of x and y is not finite? If x *
> y + z generates a spurious overflow in some cases where x * y is infinite,
> I'd expect that to apply equally to the case where the infinity comes from
> (Inf * finite), not just from (Inf * Inf).
We still see a failures (see bug 23272). How can we move this forward?
Thanks,
Florian
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] Fix ldbl-128ibm fma (Inf, Inf, finite) (bug 23272)
2018-09-21 11:59 ` Florian Weimer
@ 2018-09-21 12:46 ` Joseph Myers
2018-09-21 12:51 ` Florian Weimer
0 siblings, 1 reply; 11+ messages in thread
From: Joseph Myers @ 2018-09-21 12:46 UTC (permalink / raw
To: Florian Weimer; +Cc: Tulio Magno Quites Machado Filho, libc-alpha
On Fri, 21 Sep 2018, Florian Weimer wrote:
> * Joseph Myers:
>
> > On Wed, 13 Jun 2018, Tulio Magno Quites Machado Filho wrote:
> >
> >> This solution is mainly based on Joseph's fix at commit
> >> ca121b117f2c9c97a4c121334481a96c94fef3a0.
> >>
> >> It has extra differences because:
> >> - part of the non-finite arguments were already being treated;
> >> - when x and y are +-Inf and z if finite, an overflow can be
> >> generated.
> >
> > I'm not clear what the observed issues you are fixing are. Is this patch
> > to avoid running into libgcc generating spurious overflows for the case
> > you mention, or is it intended to fix wrong results?
> >
> >> + if (!isfinite (x) && !isfinite (y) && isfinite(z))
> >> + /* Compute the result as x * y to avoid an overflow. */
> >> + return x * y;
> >
> > Why wouldn't this be needed if just one of x and y is not finite? If x *
> > y + z generates a spurious overflow in some cases where x * y is infinite,
> > I'd expect that to apply equally to the case where the infinity comes from
> > (Inf * finite), not just from (Inf * Inf).
>
> We still see a failures (see bug 23272). How can we move this forward?
Do you mean bug 23584? That answers the first question about the symptoms
(both spurious overflow exceptions and bad results, apparently coming from
__gcc_qadd mishandling certain cases involving infinities), but doesn't
answer the second question about what happens when the infinite x * y
comes from (Inf * finite) (I'd guess the same problem, in which case the
proposed patch needs fixing further, but it looks like tests for that case
also need adding to libm-test-fma.inc).
--
Joseph S. Myers
joseph@codesourcery.com
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] Fix ldbl-128ibm fma (Inf, Inf, finite) (bug 23272)
2018-09-21 12:46 ` Joseph Myers
@ 2018-09-21 12:51 ` Florian Weimer
2018-09-21 17:27 ` Tulio Magno Quites Machado Filho
0 siblings, 1 reply; 11+ messages in thread
From: Florian Weimer @ 2018-09-21 12:51 UTC (permalink / raw
To: Joseph Myers; +Cc: Tulio Magno Quites Machado Filho, libc-alpha
* Joseph Myers:
> On Fri, 21 Sep 2018, Florian Weimer wrote:
>
>> * Joseph Myers:
>>
>> > On Wed, 13 Jun 2018, Tulio Magno Quites Machado Filho wrote:
>> >
>> >> This solution is mainly based on Joseph's fix at commit
>> >> ca121b117f2c9c97a4c121334481a96c94fef3a0.
>> >>
>> >> It has extra differences because:
>> >> - part of the non-finite arguments were already being treated;
>> >> - when x and y are +-Inf and z if finite, an overflow can be
>> >> generated.
>> >
>> > I'm not clear what the observed issues you are fixing are. Is this patch
>> > to avoid running into libgcc generating spurious overflows for the case
>> > you mention, or is it intended to fix wrong results?
>> >
>> >> + if (!isfinite (x) && !isfinite (y) && isfinite(z))
>> >> + /* Compute the result as x * y to avoid an overflow. */
>> >> + return x * y;
>> >
>> > Why wouldn't this be needed if just one of x and y is not finite? If x *
>> > y + z generates a spurious overflow in some cases where x * y is infinite,
>> > I'd expect that to apply equally to the case where the infinity comes from
>> > (Inf * finite), not just from (Inf * Inf).
>>
>> We still see a failures (see bug 23272). How can we move this forward?
>
> Do you mean bug 23584?
Yes, sorry I used the wrong bug number.
> That answers the first question about the symptoms (both spurious
> overflow exceptions and bad results, apparently coming from __gcc_qadd
> mishandling certain cases involving infinities), but doesn't answer
> the second question about what happens when the infinite x * y comes
> from (Inf * finite) (I'd guess the same problem, in which case the
> proposed patch needs fixing further, but it looks like tests for that
> case also need adding to libm-test-fma.inc).
Thanks. I have to admit this is a bit over my head.
Tulio, can you take care of this, or should someone else pick this up?
Thanks,
Florian
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] Fix ldbl-128ibm fma (Inf, Inf, finite) (bug 23272)
2018-09-21 12:51 ` Florian Weimer
@ 2018-09-21 17:27 ` Tulio Magno Quites Machado Filho
0 siblings, 0 replies; 11+ messages in thread
From: Tulio Magno Quites Machado Filho @ 2018-09-21 17:27 UTC (permalink / raw
To: Florian Weimer, Joseph Myers; +Cc: libc-alpha
Florian Weimer <fweimer@redhat.com> writes:
> * Joseph Myers:
>> That answers the first question about the symptoms (both spurious
>> overflow exceptions and bad results, apparently coming from __gcc_qadd
>> mishandling certain cases involving infinities), but doesn't answer
>> the second question about what happens when the infinite x * y comes
>> from (Inf * finite) (I'd guess the same problem, in which case the
>> proposed patch needs fixing further, but it looks like tests for that
>> case also need adding to libm-test-fma.inc).
>
> Thanks. I have to admit this is a bit over my head.
>
> Tulio, can you take care of this, or should someone else pick this up?
Yes, it's in my task list.
I still have to confirm Joseph's points and will get back to you.
I still need more time, though.
--
Tulio Magno
^ permalink raw reply [flat|nested] 11+ messages in thread
* [PATCH] Add XFAIL_ROUNDING_IBM128_LIBGCC to more fma() tests
2018-06-13 23:07 ` Joseph Myers
2018-09-21 11:59 ` Florian Weimer
@ 2019-01-14 20:09 ` Tulio Magno Quites Machado Filho
2019-01-14 21:33 ` Joseph Myers
1 sibling, 1 reply; 11+ messages in thread
From: Tulio Magno Quites Machado Filho @ 2019-01-14 20:09 UTC (permalink / raw
To: Joseph Myers, libc-alpha; +Cc: fweimer, carlos
Joseph Myers <joseph@codesourcery.com> writes:
> On Wed, 13 Jun 2018, Tulio Magno Quites Machado Filho wrote:
>
>> This solution is mainly based on Joseph's fix at commit
>> ca121b117f2c9c97a4c121334481a96c94fef3a0.
>>
>> It has extra differences because:
>> - part of the non-finite arguments were already being treated;
>> - when x and y are +-Inf and z if finite, an overflow can be
>> generated.
>
> I'm not clear what the observed issues you are fixing are. Is this patch
> to avoid running into libgcc generating spurious overflows for the case
> you mention, or is it intended to fix wrong results?
Contrary to what I had thought: this is indeed caused by spurious
overflows from libgcc (confirmed with the patch applied to libgcc).
We can suppress the all errors with XFAIL_ROUNDING_IBM128_LIBGCC.
---8<---
Ignore 112 errors in math/test-ldouble-fma and math/test-ildouble-fma
when IBM 128-bit long double used.
These errors are caused by spurious overflows from libgcc.
2019-01-14 Tulio Magno Quites Machado Filho <tuliom@linux.ibm.com>
* math/libm-test-fma.inc: Set XFAIL_ROUNDING_IBM128_LIBGCC to
more tests.
Signed-off-by: Tulio Magno Quites Machado Filho <tuliom@linux.ibm.com>
---
math/libm-test-fma.inc | 112 ++++++++++++++++++++++++-------------------------
1 file changed, 56 insertions(+), 56 deletions(-)
diff --git a/math/libm-test-fma.inc b/math/libm-test-fma.inc
index c5a97eae32..0e0e4794c7 100644
--- a/math/libm-test-fma.inc
+++ b/math/libm-test-fma.inc
@@ -119,32 +119,32 @@ static const struct test_fff_f_data fma_test_data[] =
TEST_fff_f (fma, plus_infty, plus_infty, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, plus_infty, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, plus_infty, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, plus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, plus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, plus_infty, plus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, plus_infty, plus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, plus_infty, minus_infty, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, minus_infty, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, minus_infty, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, minus_infty, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, minus_infty, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, minus_infty, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, minus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, minus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, plus_infty, minus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, plus_infty, minus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, minus_infty, plus_infty, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, plus_infty, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, plus_infty, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, plus_infty, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, plus_infty, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, plus_infty, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, plus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, plus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, minus_infty, plus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, minus_infty, plus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, minus_infty, minus_infty, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, minus_infty, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, minus_infty, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, minus_infty, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, minus_infty, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, minus_infty, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, minus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, minus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, minus_infty, minus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, minus_infty, minus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, plus_infty, max_value, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, max_value, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
@@ -152,192 +152,192 @@ static const struct test_fff_f_data fma_test_data[] =
TEST_fff_f (fma, plus_infty, max_value, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, max_value, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, max_value, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, max_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, max_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, plus_infty, max_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, plus_infty, max_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, plus_infty, min_value, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, min_value, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, min_value, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, min_value, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, min_value, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, min_value, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, min_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, min_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, plus_infty, min_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, plus_infty, min_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, plus_infty, min_subnorm_value, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, min_subnorm_value, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, min_subnorm_value, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, min_subnorm_value, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, min_subnorm_value, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, min_subnorm_value, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, min_subnorm_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, min_subnorm_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, plus_infty, min_subnorm_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, plus_infty, min_subnorm_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, plus_infty, -max_value, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -max_value, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -max_value, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -max_value, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -max_value, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -max_value, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, -max_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, -max_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, plus_infty, -max_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, plus_infty, -max_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, plus_infty, -min_value, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -min_value, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -min_value, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -min_value, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -min_value, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -min_value, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, -min_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, -min_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, plus_infty, -min_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, plus_infty, -min_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, plus_infty, -min_subnorm_value, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -min_subnorm_value, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -min_subnorm_value, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -min_subnorm_value, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -min_subnorm_value, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, plus_infty, -min_subnorm_value, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, -min_subnorm_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, plus_infty, -min_subnorm_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, plus_infty, -min_subnorm_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, plus_infty, -min_subnorm_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, minus_infty, max_value, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, max_value, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, max_value, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, max_value, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, max_value, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, max_value, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, max_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, max_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, minus_infty, max_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, minus_infty, max_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, minus_infty, min_value, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, min_value, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, min_value, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, min_value, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, min_value, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, min_value, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, min_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, min_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, minus_infty, min_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, minus_infty, min_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, minus_infty, min_subnorm_value, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, min_subnorm_value, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, min_subnorm_value, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, min_subnorm_value, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, min_subnorm_value, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, min_subnorm_value, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, min_subnorm_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, min_subnorm_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, minus_infty, min_subnorm_value, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, minus_infty, min_subnorm_value, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, minus_infty, -max_value, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -max_value, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -max_value, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -max_value, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -max_value, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -max_value, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, -max_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, -max_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, minus_infty, -max_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, minus_infty, -max_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, minus_infty, -min_value, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -min_value, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -min_value, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -min_value, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -min_value, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -min_value, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, -min_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, -min_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, minus_infty, -min_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, minus_infty, -min_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, minus_infty, -min_subnorm_value, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -min_subnorm_value, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -min_subnorm_value, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -min_subnorm_value, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -min_subnorm_value, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, minus_infty, -min_subnorm_value, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, -min_subnorm_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, minus_infty, -min_subnorm_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, minus_infty, -min_subnorm_value, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, minus_infty, -min_subnorm_value, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, max_value, plus_infty, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, max_value, plus_infty, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, max_value, plus_infty, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, max_value, plus_infty, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, max_value, plus_infty, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, max_value, plus_infty, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, max_value, plus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, max_value, plus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, max_value, plus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, max_value, plus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, min_value, plus_infty, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_value, plus_infty, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_value, plus_infty, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_value, plus_infty, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_value, plus_infty, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_value, plus_infty, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, min_value, plus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, min_value, plus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, min_value, plus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, min_value, plus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, min_subnorm_value, plus_infty, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_subnorm_value, plus_infty, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_subnorm_value, plus_infty, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_subnorm_value, plus_infty, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_subnorm_value, plus_infty, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_subnorm_value, plus_infty, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, min_subnorm_value, plus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, min_subnorm_value, plus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, min_subnorm_value, plus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, min_subnorm_value, plus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, -max_value, plus_infty, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -max_value, plus_infty, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -max_value, plus_infty, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -max_value, plus_infty, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -max_value, plus_infty, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -max_value, plus_infty, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -max_value, plus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -max_value, plus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, -max_value, plus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, -max_value, plus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, -min_value, plus_infty, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_value, plus_infty, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_value, plus_infty, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_value, plus_infty, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_value, plus_infty, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_value, plus_infty, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -min_value, plus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -min_value, plus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, -min_value, plus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, -min_value, plus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, -min_subnorm_value, plus_infty, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_subnorm_value, plus_infty, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_subnorm_value, plus_infty, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_subnorm_value, plus_infty, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_subnorm_value, plus_infty, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_subnorm_value, plus_infty, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -min_subnorm_value, plus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -min_subnorm_value, plus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, -min_subnorm_value, plus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, -min_subnorm_value, plus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, max_value, minus_infty, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, max_value, minus_infty, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, max_value, minus_infty, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, max_value, minus_infty, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, max_value, minus_infty, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, max_value, minus_infty, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, max_value, minus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, max_value, minus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, max_value, minus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, max_value, minus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, min_value, minus_infty, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_value, minus_infty, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_value, minus_infty, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_value, minus_infty, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_value, minus_infty, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_value, minus_infty, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, min_value, minus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, min_value, minus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, min_value, minus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, min_value, minus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, min_subnorm_value, minus_infty, plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_subnorm_value, minus_infty, minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_subnorm_value, minus_infty, min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_subnorm_value, minus_infty, -min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_subnorm_value, minus_infty, min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, min_subnorm_value, minus_infty, -min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, min_subnorm_value, minus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, min_subnorm_value, minus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, min_subnorm_value, minus_infty, max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, min_subnorm_value, minus_infty, -max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, -max_value, minus_infty, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -max_value, minus_infty, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -max_value, minus_infty, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -max_value, minus_infty, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -max_value, minus_infty, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -max_value, minus_infty, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -max_value, minus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -max_value, minus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, -max_value, minus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, -max_value, minus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, -min_value, minus_infty, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_value, minus_infty, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_value, minus_infty, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_value, minus_infty, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_value, minus_infty, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_value, minus_infty, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -min_value, minus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -min_value, minus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, -min_value, minus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, -min_value, minus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
TEST_fff_f (fma, -min_subnorm_value, minus_infty, plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_subnorm_value, minus_infty, minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_subnorm_value, minus_infty, min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_subnorm_value, minus_infty, -min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_subnorm_value, minus_infty, min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
TEST_fff_f (fma, -min_subnorm_value, minus_infty, -min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -min_subnorm_value, minus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
- TEST_fff_f (fma, -min_subnorm_value, minus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+ TEST_fff_f (fma, -min_subnorm_value, minus_infty, max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+ TEST_fff_f (fma, -min_subnorm_value, minus_infty, -max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
AUTO_TESTS_fff_f (fma),
};
--
2.14.5
^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [PATCH] Add XFAIL_ROUNDING_IBM128_LIBGCC to more fma() tests
2019-01-14 20:09 ` [PATCH] Add XFAIL_ROUNDING_IBM128_LIBGCC to more fma() tests Tulio Magno Quites Machado Filho
@ 2019-01-14 21:33 ` Joseph Myers
2019-01-15 18:55 ` Tulio Magno Quites Machado Filho
0 siblings, 1 reply; 11+ messages in thread
From: Joseph Myers @ 2019-01-14 21:33 UTC (permalink / raw
To: Tulio Magno Quites Machado Filho; +Cc: libc-alpha, fweimer, carlos
On Mon, 14 Jan 2019, Tulio Magno Quites Machado Filho wrote:
> Contrary to what I had thought: this is indeed caused by spurious
> overflows from libgcc (confirmed with the patch applied to libgcc).
> We can suppress the all errors with XFAIL_ROUNDING_IBM128_LIBGCC.
In that case - if it's only an issue in non-default rounding modes - this
patch is OK.
--
Joseph S. Myers
joseph@codesourcery.com
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] Add XFAIL_ROUNDING_IBM128_LIBGCC to more fma() tests
2019-01-14 21:33 ` Joseph Myers
@ 2019-01-15 18:55 ` Tulio Magno Quites Machado Filho
2019-01-15 20:52 ` Joseph Myers
0 siblings, 1 reply; 11+ messages in thread
From: Tulio Magno Quites Machado Filho @ 2019-01-15 18:55 UTC (permalink / raw
To: Joseph Myers; +Cc: libc-alpha, fweimer, carlos
Joseph Myers <joseph@codesourcery.com> writes:
> On Mon, 14 Jan 2019, Tulio Magno Quites Machado Filho wrote:
>
>> Contrary to what I had thought: this is indeed caused by spurious
>> overflows from libgcc (confirmed with the patch applied to libgcc).
>> We can suppress the all errors with XFAIL_ROUNDING_IBM128_LIBGCC.
>
> In that case - if it's only an issue in non-default rounding modes - this
> patch is OK.
They appear only when rounding upward and downward.
Pushed as ecdacd34a2ac3b6d5a529ff218b29261d9d98a7a.
Thanks!
--
Tulio Magno
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] Add XFAIL_ROUNDING_IBM128_LIBGCC to more fma() tests
2019-01-15 18:55 ` Tulio Magno Quites Machado Filho
@ 2019-01-15 20:52 ` Joseph Myers
2019-01-16 12:44 ` Tulio Magno Quites Machado Filho
0 siblings, 1 reply; 11+ messages in thread
From: Joseph Myers @ 2019-01-15 20:52 UTC (permalink / raw
To: Tulio Magno Quites Machado Filho; +Cc: libc-alpha, fweimer, carlos
On Tue, 15 Jan 2019, Tulio Magno Quites Machado Filho wrote:
> Joseph Myers <joseph@codesourcery.com> writes:
>
> > On Mon, 14 Jan 2019, Tulio Magno Quites Machado Filho wrote:
> >
> >> Contrary to what I had thought: this is indeed caused by spurious
> >> overflows from libgcc (confirmed with the patch applied to libgcc).
> >> We can suppress the all errors with XFAIL_ROUNDING_IBM128_LIBGCC.
> >
> > In that case - if it's only an issue in non-default rounding modes - this
> > patch is OK.
>
> They appear only when rounding upward and downward.
>
> Pushed as ecdacd34a2ac3b6d5a529ff218b29261d9d98a7a.
Will you be closing bug 23584 (given the conclusion it's actually a libgcc
issue, addressed via the XFAILing, not a glibc bug)?
--
Joseph S. Myers
joseph@codesourcery.com
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [PATCH] Add XFAIL_ROUNDING_IBM128_LIBGCC to more fma() tests
2019-01-15 20:52 ` Joseph Myers
@ 2019-01-16 12:44 ` Tulio Magno Quites Machado Filho
0 siblings, 0 replies; 11+ messages in thread
From: Tulio Magno Quites Machado Filho @ 2019-01-16 12:44 UTC (permalink / raw
To: Joseph Myers; +Cc: libc-alpha, fweimer, carlos
Joseph Myers <joseph@codesourcery.com> writes:
> [ text/plain ]
> On Tue, 15 Jan 2019, Tulio Magno Quites Machado Filho wrote:
>
>> Joseph Myers <joseph@codesourcery.com> writes:
>>
>> > On Mon, 14 Jan 2019, Tulio Magno Quites Machado Filho wrote:
>> >
>> >> Contrary to what I had thought: this is indeed caused by spurious
>> >> overflows from libgcc (confirmed with the patch applied to libgcc).
>> >> We can suppress the all errors with XFAIL_ROUNDING_IBM128_LIBGCC.
>> >
>> > In that case - if it's only an issue in non-default rounding modes - this
>> > patch is OK.
>>
>> They appear only when rounding upward and downward.
>>
>> Pushed as ecdacd34a2ac3b6d5a529ff218b29261d9d98a7a.
>
> Will you be closing bug 23584 (given the conclusion it's actually a libgcc
> issue, addressed via the XFAILing, not a glibc bug)?
Done.
Thanks! I was not aware of this bug report. ;-)
--
Tulio Magno
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2019-01-16 12:44 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2018-06-13 22:01 [PATCH] Fix ldbl-128ibm fma (Inf, Inf, finite) (bug 23272) Tulio Magno Quites Machado Filho
2018-06-13 23:07 ` Joseph Myers
2018-09-21 11:59 ` Florian Weimer
2018-09-21 12:46 ` Joseph Myers
2018-09-21 12:51 ` Florian Weimer
2018-09-21 17:27 ` Tulio Magno Quites Machado Filho
2019-01-14 20:09 ` [PATCH] Add XFAIL_ROUNDING_IBM128_LIBGCC to more fma() tests Tulio Magno Quites Machado Filho
2019-01-14 21:33 ` Joseph Myers
2019-01-15 18:55 ` Tulio Magno Quites Machado Filho
2019-01-15 20:52 ` Joseph Myers
2019-01-16 12:44 ` Tulio Magno Quites Machado Filho
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).