[PATCH 1/2] Revert "x86: Do not use __builtin_isinf_sign for _Float64x/long double"

[Andreas K. Huettel]

Am Dienstag, 13. Januar 2026, 20:19:06 Mitteleuropäische Normalzeit schrieb Adhemerval Zanella Netto:
> 
> On 13/01/26 12:04, Carlos O'Donell wrote:
> > On 1/13/26 9:19 AM, Adhemerval Zanella wrote:
> >> This reverts commit 999cd617cb7e40a2fa719e91fe1028c853ae14d5.  It
> >> breaks libstdc++ and other languages bootstrap.
> > 
> > I think we should continue to move this forward to improve the build
> > support, so while I'm approving the revert, I think we should fix this
> > and backport after due consideration.
> > 
> > Revert matches.
> > 
> > Reviewed-by: Carlos O'Donell <carlos@redhat.com>
> 
> Ack, I will wait Andreas give the OK before install it.

Sure. Right now this seems to be the safest option.

> 
> > 
> >> ---
> >>   bits/fp-builtin-denormal.h              | 4 ----
> >>   math/math.h                             | 5 +----
> >>   sysdeps/x86/bits/fp-builtin-denormal.h  | 4 ----
> >>   sysdeps/x86/fpu/test-builtin-denormal.c | 4 ----
> >>   4 files changed, 1 insertion(+), 16 deletions(-)
> >>
> >> diff --git a/bits/fp-builtin-denormal.h b/bits/fp-builtin-denormal.h
> >> index 0f24de149a..457ab0c97a 100644
> >> --- a/bits/fp-builtin-denormal.h
> >> +++ b/bits/fp-builtin-denormal.h
> >> @@ -26,7 +26,3 @@
> >>       Intel double extended-precision (long double).  By default assume 1 to
> >>       enable the usage of compiler builtin on math.h.  */
> >>   #define __FP_BUILTIN_FPCLASSIFY_DENORMAL 1
> >> -
> >> -/* __FP_BUILTIN_ISINF_SIGN_DENORMAL is defined to 1 if compiler supports
> >> -   handling pseudo-dernormal numbers with isinf_sign builtin.  */
> >> -#define __FP_BUILTIN_ISINF_SIGN_DENORMAL 1
> >> diff --git a/math/math.h b/math/math.h
> >> index 1066113ca6..77dfcdc36b 100644
> >> --- a/math/math.h
> >> +++ b/math/math.h
> >> @@ -1249,10 +1249,7 @@ enum
> >>        ? __isinff128 (x) : __builtin_isinf_sign (x))
> >>   # elif (__GNUC_PREREQ (4,4) && !defined __SUPPORT_SNAN__) \
> >>          || __glibc_clang_prereq (3,7)
> >> -#  define isinf(x) __MATH_TG_BUILTIN_CLASSIFY ((x),                  \
> >> -                           __builtin_isinf_sign, (x),     \
> >> -                           isinf, (x),              \
> >> -                           __FP_BUILTIN_ISINF_SIGN_DENORMAL)
> >> +#  define isinf(x) __builtin_isinf_sign (x)
> >>   # else
> >>   #  define isinf(x) __MATH_TG ((x), __isinf, (x))
> >>   # endif
> >> diff --git a/sysdeps/x86/bits/fp-builtin-denormal.h b/sysdeps/x86/bits/fp-builtin-denormal.h
> >> index 30ed7c899a..33549c876a 100644
> >> --- a/sysdeps/x86/bits/fp-builtin-denormal.h
> >> +++ b/sysdeps/x86/bits/fp-builtin-denormal.h
> >> @@ -23,7 +23,3 @@
> >>   /* Neither GCC (bug 123161) nor clang (issue 172533) handles pseudo-normal
> >>      numbers correctly with fpclassify builtin.  */
> >>   #define __FP_BUILTIN_FPCLASSIFY_DENORMAL 0
> >> -
> >> -/* Neither GCC (bug 123173) nor clang (issue 172651) handles pseudo-normal
> >> -   numbers correctly with isinf_sign builtin.  */
> >> -#define __FP_BUILTIN_ISINF_SIGN_DENORMAL 0
> >> diff --git a/sysdeps/x86/fpu/test-builtin-denormal.c b/sysdeps/x86/fpu/test-builtin-denormal.c
> >> index 8856ebce80..6584ec339f 100644
> >> --- a/sysdeps/x86/fpu/test-builtin-denormal.c
> >> +++ b/sysdeps/x86/fpu/test-builtin-denormal.c
> >> @@ -45,10 +45,6 @@ do_test (void)
> >>         TEST_COMPARE (feclearexcept (FE_INVALID), 0);
> >>         TEST_COMPARE (fpclassify (inputs[i].value), FP_NAN);
> >>         TEST_COMPARE (fetestexcept (FE_INVALID), 0);
> >> -
> >> -      TEST_COMPARE (feclearexcept (FE_INVALID), 0);
> >> -      TEST_COMPARE (isinf (inputs[i].value), 0);
> >> -      TEST_COMPARE (fetestexcept (FE_INVALID), 0);
> >>       }
> >>       return 0;
> > 
> > 
> 
> 


-- 
PD Dr. Andreas K. Huettel
Institute for Experimental and Applied Physics
University of Regensburg
93040 Regensburg
Germany

tel. +49 151 241 67748 (mobile)
tel. +49 941 943 1618 (office)
e-mail andreas.huettel@ur.de
https://www.akhuettel.de/
https://www.akhuettel.de/group/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1018 bytes
Desc: This is a digitally signed message part.
URL: <https://sourceware.org/pipermail/libc-alpha/attachments/20260114/0a9a7ac2/attachment.sig>