diff options
| author | Aurelien Jarno <aurelien@aurel32.net> | 2016-07-31 23:17:47 +0200 |
|---|---|---|
| committer | Aurelien Jarno <aurelien@aurel32.net> | 2016-08-02 02:07:20 +0200 |
| commit | 33ae5b17cd9bc1bea1e35d9805875b10cb489571 (patch) | |
| tree | e8173781347efa062491126eae7748c8d043131c /sysdeps/sparc/sparc32/sparcv9/fpu/multiarch | |
| parent | fab382315ad3be7c773aaf7ca49c053cf91755fe (diff) | |
| download | glibc-33ae5b17cd9bc1bea1e35d9805875b10cb489571.tar.gz glibc-33ae5b17cd9bc1bea1e35d9805875b10cb489571.tar.xz glibc-33ae5b17cd9bc1bea1e35d9805875b10cb489571.zip | |
sparc: remove ceil, floor, trunc sparc specific implementations
The ceil, floor and trunc functions on sparc do not fully follow the
standard and trigger an inexact exception when presented a value which
is not an integer. Since glibc 2.24 this causes a few tests to fail,
for instance:
testing double (without inline functions)
Failure: ceil (lit_pi): Exception "Inexact" set
Failure: ceil (-lit_pi): Exception "Inexact" set
Failure: ceil (min_subnorm_value): Exception "Inexact" set
Failure: ceil (min_value): Exception "Inexact" set
Failure: ceil (0.1): Exception "Inexact" set
Failure: ceil (0.25): Exception "Inexact" set
Failure: ceil (0.625): Exception "Inexact" set
Failure: ceil (-min_subnorm_value): Exception "Inexact" set
Failure: ceil (-min_value): Exception "Inexact" set
Failure: ceil (-0.1): Exception "Inexact" set
Failure: ceil (-0.25): Exception "Inexact" set
Failure: ceil (-0.625): Exception "Inexact" set
I tried to fix that by using the same strategy than used on other
architectures, that is by saving the FSR register at the beginning
and restoring it at the end of the function. When doing so I noticed
a comment that this operation might be very costly, so I decided to
do some benchmarks.
The benchmarks below represent the time required to run each of the
function 60 millions of times with different input value. I have done
that in the basic V9 code, the VIS2 code, and using the default C
implementation of the libc, for both sparc32 and sparc64, on a Niagara
T1 based machine and an UltraSparc IIIi. Given I don't have access to a
more recent machine), I haven't been able to test the VIS3 version. Also
it should be noted that it doesn't make sense to do this benchmark for
V8 or earlier as in that case we use the default C implementation. The
results are available in the table below, the "+ fix" version correspond
to the one saving and restoring the FSR.
Niagara T1 / sparc32
--------------------
ceilf ceil floorf floor truncf trunc
V9 19.10 22.48 19.10 22.48 16.59 19.27
V9 + fix 19.77 23.34 19.77 23.33 17.27 20.12
VIS2 16.87 19.62 16.87 19.62
VIS2 + fix 17.55 20.47 17.55 20.47
C impl 11.39 13.80 11.40 13.80 10.88 10.84
Niagara T1 / sparc64
--------------------
ceilf ceil floorf floor truncf trunc
V9 18.14 22.23 18.14 22.23 15.64 19.02
V9 + fix 18.82 23.08 18.82 23.08 16.32 19.87
VIS2 15.92 19.37 15.92 19.37
VIS2 + fix 16.59 20.22 16.59 20.22
C impl 11.39 13.60 11.39 15.36 10.88 12.65
UltraSparc IIIi / sparc32
-------------------------
ceilf ceil floorf floor truncf trunc
V9 4.81 7.09 6.61 11.64 4.91 7.05
V9 + fix 7.20 10.42 7.14 10.54 6.76 9.47
VIS2 4.81 7.03 4.76 7.13
VIS2 + fix 6.76 9.51 6.71 9.63
C impl 3.88 8.62 3.90 9.45 3.57 6.62
UltraSparc IIIi / sparc64
-------------------------
ceilf ceil floorf floor truncf trunc
V9 3.48 4.39 3.48 4.41 3.01 3.85
V9 + fix 4.76 5.90 4.76 5.90 4.86 6.26
VIS2 2.95 3.61 2.95 3.61
VIS2 + fix 4.24 5.37 4.30 7.97
C impl 3.63 4.89 3.62 6.38 3.33 4.03
The first thing that should be noted is that the C implementation is
always faster on the Niagara T1 based machine. On the UltraSparc IIIi
the float version on sparc32 is also faster.
Coming back about the fix saving and restoring the FSR, it appears
it has a big impact as expected. In that case the C implementation is
always faster than the fixed implementations.
This patch therefore removes the sparc specific implementations in
favor of the generic ones.
Changelog:
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/Makefile
[$(subdir) = math] (libm-sysdep_routines): Remove.
[$(subdir) = math && $(have-as-vis3) = yes] (libm-sysdep_routines):
Remove s_ceilf-vis3, s_ceil-vis3, s_floorf-vis3, s_floor-vis3,
s_truncf-vis3, s_trunc-vis3.
* sysdeps/sparc/sparc64/fpu/multiarch/Makefile: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil-vis2.S: Delete
file.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil-vis3.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf-vis2.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf-vis3.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor-vis2.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor-vis3.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf-vis2.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf-vis3.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_trunc-vis3.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_trunc.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_truncf-vis3.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_truncf.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/s_ceil.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/s_ceilf.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/s_floor.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/s_floorf.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/s_trunc.S: Likewise.
* sysdeps/sparc/sparc32/sparcv9/fpu/s_truncf.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_ceil-vis2.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_ceil-vis3.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_ceil.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_ceilf-vis2.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_ceilf-vis3.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_ceilf.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_floor-vis2.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_floor-vis3.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_floor.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_floorf-vis2.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_floorf-vis3.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_floorf.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_trunc-vis3.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_trunc.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_truncf-vis3.S: Likewise.
* sysdeps/sparc/sparc64/fpu/multiarch/s_truncf.S: Likewise.
* sysdeps/sparc/sparc64/fpu/s_ceil.S: Likewise.
* sysdeps/sparc/sparc64/fpu/s_ceilf.S: Likewise.
* sysdeps/sparc/sparc64/fpu/s_floor.S: Likewise.
* sysdeps/sparc/sparc64/fpu/s_floorf.S: Likewise.
* sysdeps/sparc/sparc64/fpu/s_trunc.S: Likewise.
* sysdeps/sparc/sparc64/fpu/s_truncf.S: Likewise.
Diffstat (limited to 'sysdeps/sparc/sparc32/sparcv9/fpu/multiarch')
17 files changed, 3 insertions, 690 deletions
diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/Makefile b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/Makefile index 0d92813d7b..ebbe28b07f 100644 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/Makefile +++ b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/Makefile @@ -1,15 +1,11 @@ ifeq ($(subdir),math) -libm-sysdep_routines += s_ceil-vis2 s_ceilf-vis2 \ - s_floor-vis2 s_floorf-vis2 ifeq ($(have-as-vis3),yes) -libm-sysdep_routines += m_copysignf-vis3 m_copysign-vis3 s_ceilf-vis3 \ - s_ceil-vis3 s_fabs-vis3 s_fabsf-vis3 s_floor-vis3 \ - s_floorf-vis3 s_llrintf-vis3 s_llrint-vis3 \ +libm-sysdep_routines += m_copysignf-vis3 m_copysign-vis3 s_fabs-vis3 \ + s_fabsf-vis3 s_llrintf-vis3 s_llrint-vis3 \ s_rintf-vis3 s_rint-vis3 w_sqrt-vis3 w_sqrtf-vis3 \ s_fminf-vis3 s_fmin-vis3 s_fmaxf-vis3 s_fmax-vis3 \ s_fmaf-vis3 s_fma-vis3 s_fdimf-vis3 s_fdim-vis3 \ - s_nearbyint-vis3 s_nearbyintf-vis3 s_truncf-vis3 \ - s_trunc-vis3 + s_nearbyint-vis3 s_nearbyintf-vis3 sysdep_routines += s_copysignf-vis3 s_copysign-vis3 endif endif diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil-vis2.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil-vis2.S deleted file mode 100644 index edc24b3064..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil-vis2.S +++ /dev/null @@ -1,61 +0,0 @@ -/* ceil function, sparc32 v9 vis2 version. - Copyright (C) 2013-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by David S. Miller <davem@davemloft.net>, 2013. - - The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - The GNU C Library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with the GNU C Library; if not, see - <http://www.gnu.org/licenses/>. */ - -#include <sysdep.h> - - /* 'siam' (Set Interval Arithmetic Mode) is used to quickly override - the rounding mode during this routine. - - We add then subtract (or subtract than add if the initial - value was negative) 2**23 to the value, then subtract it - back out. - - This will clear out the fractional portion of the value and, - with suitable 'siam' initiated rouding mode settings, round - the final result in the proper direction. */ - -#define TWO_FIFTYTWO 0x43300000 /* 2**52 */ - -#define ZERO %f10 /* 0.0 */ -#define SIGN_BIT %f12 /* -0.0 */ - -ENTRY (__ceil_vis2) - sethi %hi(TWO_FIFTYTWO), %o2 - sllx %o0, 32, %o0 - or %o0, %o1, %o0 - stx %o0, [%sp + 72] - sllx %o2, 32, %o2 - fzero ZERO - ldd [%sp + 72], %f0 - fnegd ZERO, SIGN_BIT - stx %o2, [%sp + 72] - fabsd %f0, %f14 - ldd [%sp + 72], %f16 - fcmpd %fcc3, %f14, %f16 - fmovduge %fcc3, ZERO, %f16 - fand %f0, SIGN_BIT, SIGN_BIT - for %f16, SIGN_BIT, %f16 - siam (1 << 2) | 2 - faddd %f0, %f16, %f18 - siam (1 << 2) | 0 - fsubd %f18, %f16, %f18 - siam (0 << 2) - retl - for %f18, SIGN_BIT, %f0 -END (__ceil_vis2) diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil-vis3.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil-vis3.S deleted file mode 100644 index 8b14c6896a..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil-vis3.S +++ /dev/null @@ -1,62 +0,0 @@ -/* ceil function, sparc32 v9 vis3 version. - Copyright (C) 2012-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by David S. Miller <davem@davemloft.net>, 2012. - - The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - The GNU C Library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with the GNU C Library; if not, see - <http://www.gnu.org/licenses/>. */ - -#include <sysdep.h> - - /* 'siam' (Set Interval Arithmetic Mode) is used to quickly override - the rounding mode during this routine. - - We add then subtract (or subtract than add if the initial - value was negative) 2**23 to the value, then subtract it - back out. - - This will clear out the fractional portion of the value and, - with suitable 'siam' initiated rouding mode settings, round - the final result in the proper direction. - - We also use VIS3 moves to avoid using the stack to transfer - values between float and integer registers. */ - -#define TWO_FIFTYTWO 0x43300000 /* 2**52 */ - -#define ZERO %f10 /* 0.0 */ -#define SIGN_BIT %f12 /* -0.0 */ - -ENTRY (__ceil_vis3) - sethi %hi(TWO_FIFTYTWO), %o2 - sllx %o0, 32, %o0 - sllx %o2, 32, %o2 - or %o0, %o1, %o0 - movxtod %o0, %f0 - fzero ZERO - fnegd ZERO, SIGN_BIT - movxtod %o2, %f16 - fabsd %f0, %f14 - fcmpd %fcc3, %f14, %f16 - fmovduge %fcc3, ZERO, %f16 - fand %f0, SIGN_BIT, SIGN_BIT - for %f16, SIGN_BIT, %f16 - siam (1 << 2) | 2 - faddd %f0, %f16, %f18 - siam (1 << 2) | 0 - fsubd %f18, %f16, %f18 - siam (0 << 2) - retl - for %f18, SIGN_BIT, %f0 -END (__ceil_vis3) diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil.S deleted file mode 100644 index efc8d49365..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceil.S +++ /dev/null @@ -1,19 +0,0 @@ -#include <sparc-ifunc.h> -#include <math_ldbl_opt.h> - -SPARC_ASM_VIS3_VIS2_IFUNC(ceil) - -weak_alias (__ceil, ceil) - -#if LONG_DOUBLE_COMPAT(libm, GLIBC_2_0) -compat_symbol (libm, __ceil, ceill, GLIBC_2_0) -#endif - -# undef weak_alias -# define weak_alias(a, b) -# undef compat_symbol -# define compat_symbol(a, b, c, d) - -#define __ceil __ceil_generic - -#include "../s_ceil.S" diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf-vis2.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf-vis2.S deleted file mode 100644 index 019681b5f2..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf-vis2.S +++ /dev/null @@ -1,58 +0,0 @@ -/* Float ceil function, sparc32 v9 vis2 version. - Copyright (C) 2013-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by David S. Miller <davem@davemloft.net>, 2013. - - The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - The GNU C Library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with the GNU C Library; if not, see - <http://www.gnu.org/licenses/>. */ - -#include <sysdep.h> - - /* 'siam' (Set Interval Arithmetic Mode) is used to quickly override - the rounding mode during this routine. - - We add then subtract (or subtract than add if the initial - value was negative) 2**23 to the value, then subtract it - back out. - - This will clear out the fractional portion of the value and, - with suitable 'siam' initiated rouding mode settings, round - the final result in the proper direction. */ - -#define TWO_TWENTYTHREE 0x4b000000 /* 2**23 */ - -#define ZERO %f10 /* 0.0 */ -#define SIGN_BIT %f12 /* -0.0 */ - -ENTRY (__ceilf_vis2) - st %o0, [%sp + 68] - sethi %hi(TWO_TWENTYTHREE), %o2 - fzeros ZERO - ld [%sp + 68], %f0 - fnegs ZERO, SIGN_BIT - st %o2, [%sp + 68] - fabss %f0, %f14 - ld [%sp + 68], %f16 - fcmps %fcc3, %f14, %f16 - fmovsuge %fcc3, ZERO, %f16 - fands %f0, SIGN_BIT, SIGN_BIT - fors %f16, SIGN_BIT, %f16 - siam (1 << 2) | 2 - fadds %f0, %f16, %f1 - siam (1 << 2) | 0 - fsubs %f1, %f16, %f1 - siam (0 << 2) - retl - fors %f1, SIGN_BIT, %f0 -END (__ceilf_vis2) diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf-vis3.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf-vis3.S deleted file mode 100644 index 1bf306a115..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf-vis3.S +++ /dev/null @@ -1,59 +0,0 @@ -/* Float ceil function, sparc32 v9 vis3 version. - Copyright (C) 2012-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by David S. Miller <davem@davemloft.net>, 2012. - - The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - The GNU C Library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with the GNU C Library; if not, see - <http://www.gnu.org/licenses/>. */ - -#include <sysdep.h> - - /* 'siam' (Set Interval Arithmetic Mode) is used to quickly override - the rounding mode during this routine. - - We add then subtract (or subtract than add if the initial - value was negative) 2**23 to the value, then subtract it - back out. - - This will clear out the fractional portion of the value and, - with suitable 'siam' initiated rouding mode settings, round - the final result in the proper direction. - - We also use VIS3 moves to avoid using the stack to transfer - values between float and integer registers. */ - -#define TWO_TWENTYTHREE 0x4b000000 /* 2**23 */ - -#define ZERO %f10 /* 0.0 */ -#define SIGN_BIT %f12 /* -0.0 */ - -ENTRY (__ceilf_vis3) - movwtos %o0, %f0 - sethi %hi(TWO_TWENTYTHREE), %o2 - fzeros ZERO - fnegs ZERO, SIGN_BIT - movwtos %o2, %f16 - fabss %f0, %f14 - fcmps %fcc3, %f14, %f16 - fmovsuge %fcc3, ZERO, %f16 - fands %f0, SIGN_BIT, SIGN_BIT - fors %f16, SIGN_BIT, %f16 - siam (1 << 2) | 2 - fadds %f0, %f16, %f1 - siam (1 << 2) | 0 - fsubs %f1, %f16, %f1 - siam (0 << 2) - retl - fors %f1, SIGN_BIT, %f0 -END (__ceilf_vis3) diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf.S deleted file mode 100644 index 1c72a5728e..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_ceilf.S +++ /dev/null @@ -1,12 +0,0 @@ -#include <sparc-ifunc.h> - -SPARC_ASM_VIS3_VIS2_IFUNC(ceilf) - -weak_alias (__ceilf, ceilf) - -# undef weak_alias -# define weak_alias(a, b) - -#define __ceilf __ceilf_generic - -#include "../s_ceilf.S" diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor-vis2.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor-vis2.S deleted file mode 100644 index 0180d103b6..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor-vis2.S +++ /dev/null @@ -1,61 +0,0 @@ -/* floor function, sparc32 v9 vis2 version. - Copyright (C) 2013-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by David S. Miller <davem@davemloft.net>, 2013. - - The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - The GNU C Library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with the GNU C Library; if not, see - <http://www.gnu.org/licenses/>. */ - -#include <sysdep.h> - - /* 'siam' (Set Interval Arithmetic Mode) is used to quickly override - the rounding mode during this routine. - - We add then subtract (or subtract than add if the initial - value was negative) 2**23 to the value, then subtract it - back out. - - This will clear out the fractional portion of the value and, - with suitable 'siam' initiated rouding mode settings, round - the final result in the proper direction. */ - -#define TWO_FIFTYTWO 0x43300000 /* 2**52 */ - -#define ZERO %f10 /* 0.0 */ -#define SIGN_BIT %f12 /* -0.0 */ - -ENTRY (__floor_vis2) - sethi %hi(TWO_FIFTYTWO), %o2 - sllx %o0, 32, %o0 - or %o0, %o1, %o0 - stx %o0, [%sp + 72] - sllx %o2, 32, %o2 - fzero ZERO - ldd [%sp + 72], %f0 - fnegd ZERO, SIGN_BIT - stx %o2, [%sp + 72] - fabsd %f0, %f14 - ldd [%sp + 72], %f16 - fcmpd %fcc3, %f14, %f16 - fmovduge %fcc3, ZERO, %f16 - fand %f0, SIGN_BIT, SIGN_BIT - for %f16, SIGN_BIT, %f16 - siam (1 << 2) | 3 - faddd %f0, %f16, %f18 - siam (1 << 2) | 0 - fsubd %f18, %f16, %f18 - siam (0 << 2) - retl - for %f18, SIGN_BIT, %f0 -END (__floor_vis2) diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor-vis3.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor-vis3.S deleted file mode 100644 index 7b91620a42..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor-vis3.S +++ /dev/null @@ -1,62 +0,0 @@ -/* floor function, sparc32 v9 vis3 version. - Copyright (C) 2012-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by David S. Miller <davem@davemloft.net>, 2012. - - The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - The GNU C Library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with the GNU C Library; if not, see - <http://www.gnu.org/licenses/>. */ - -#include <sysdep.h> - - /* 'siam' (Set Interval Arithmetic Mode) is used to quickly override - the rounding mode during this routine. - - We add then subtract (or subtract than add if the initial - value was negative) 2**23 to the value, then subtract it - back out. - - This will clear out the fractional portion of the value and, - with suitable 'siam' initiated rouding mode settings, round - the final result in the proper direction. - - We also use VIS3 moves to avoid using the stack to transfer - values between float and integer registers. */ - -#define TWO_FIFTYTWO 0x43300000 /* 2**52 */ - -#define ZERO %f10 /* 0.0 */ -#define SIGN_BIT %f12 /* -0.0 */ - -ENTRY (__floor_vis3) - sethi %hi(TWO_FIFTYTWO), %o2 - sllx %o0, 32, %o0 - sllx %o2, 32, %o2 - or %o0, %o1, %o0 - movxtod %o0, %f0 - fzero ZERO - fnegd ZERO, SIGN_BIT - movxtod %o2, %f16 - fabsd %f0, %f14 - fcmpd %fcc3, %f14, %f16 - fmovduge %fcc3, ZERO, %f16 - fand %f0, SIGN_BIT, SIGN_BIT - for %f16, SIGN_BIT, %f16 - siam (1 << 2) | 3 - faddd %f0, %f16, %f18 - siam (1 << 2) | 0 - fsubd %f18, %f16, %f18 - siam (0 << 2) - retl - for %f18, SIGN_BIT, %f0 -END (__floor_vis3) diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor.S deleted file mode 100644 index 1fe4b95ea6..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floor.S +++ /dev/null @@ -1,19 +0,0 @@ -#include <sparc-ifunc.h> -#include <math_ldbl_opt.h> - -SPARC_ASM_VIS3_VIS2_IFUNC(floor) - -weak_alias (__floor, floor) - -#if LONG_DOUBLE_COMPAT(libm, GLIBC_2_0) -compat_symbol (libm, __floor, floorl, GLIBC_2_0) -#endif - -# undef weak_alias -# define weak_alias(a, b) -# undef compat_symbol -# define compat_symbol(a, b, c, d) - -#define __floor __floor_generic - -#include "../s_floor.S" diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf-vis2.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf-vis2.S deleted file mode 100644 index 77e9f4bcc1..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf-vis2.S +++ /dev/null @@ -1,58 +0,0 @@ -/* Float floor function, sparc32 v9 vis2 version. - Copyright (C) 2013-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by David S. Miller <davem@davemloft.net>, 2013. - - The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - The GNU C Library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with the GNU C Library; if not, see - <http://www.gnu.org/licenses/>. */ - -#include <sysdep.h> - - /* 'siam' (Set Interval Arithmetic Mode) is used to quickly override - the rounding mode during this routine. - - We add then subtract (or subtract than add if the initial - value was negative) 2**23 to the value, then subtract it - back out. - - This will clear out the fractional portion of the value and, - with suitable 'siam' initiated rouding mode settings, round - the final result in the proper direction. */ - -#define TWO_TWENTYTHREE 0x4b000000 /* 2**23 */ - -#define ZERO %f10 /* 0.0 */ -#define SIGN_BIT %f12 /* -0.0 */ - -ENTRY (__floorf_vis2) - st %o0, [%sp + 68] - sethi %hi(TWO_TWENTYTHREE), %o2 - fzeros ZERO - ld [%sp + 68], %f0 - fnegs ZERO, SIGN_BIT - st %o2, [%sp + 68] - fabss %f0, %f14 - ld [%sp + 68], %f16 - fcmps %fcc3, %f14, %f16 - fmovsuge %fcc3, ZERO, %f16 - fands %f0, SIGN_BIT, SIGN_BIT - fors %f16, SIGN_BIT, %f16 - siam (1 << 2) | 3 - fadds %f0, %f16, %f1 - siam (1 << 2) | 0 - fsubs %f1, %f16, %f1 - siam (0 << 2) - retl - fors %f1, SIGN_BIT, %f0 -END (__floorf_vis2) diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf-vis3.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf-vis3.S deleted file mode 100644 index bf7da6fbfa..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf-vis3.S +++ /dev/null @@ -1,59 +0,0 @@ -/* Float floor function, sparc32 v9 vis3 version. - Copyright (C) 2012-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by David S. Miller <davem@davemloft.net>, 2012. - - The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - The GNU C Library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with the GNU C Library; if not, see - <http://www.gnu.org/licenses/>. */ - -#include <sysdep.h> - - /* 'siam' (Set Interval Arithmetic Mode) is used to quickly override - the rounding mode during this routine. - - We add then subtract (or subtract than add if the initial - value was negative) 2**23 to the value, then subtract it - back out. - - This will clear out the fractional portion of the value and, - with suitable 'siam' initiated rouding mode settings, round - the final result in the proper direction. - - We also use VIS3 moves to avoid using the stack to transfer - values between float and integer registers. */ - -#define TWO_TWENTYTHREE 0x4b000000 /* 2**23 */ - -#define ZERO %f10 /* 0.0 */ -#define SIGN_BIT %f12 /* -0.0 */ - -ENTRY (__floorf_vis3) - movwtos %o0, %f0 - sethi %hi(TWO_TWENTYTHREE), %o2 - fzeros ZERO - fnegs ZERO, SIGN_BIT - movwtos %o2, %f16 - fabss %f0, %f14 - fcmps %fcc3, %f14, %f16 - fmovsuge %fcc3, ZERO, %f16 - fands %f0, SIGN_BIT, SIGN_BIT - fors %f16, SIGN_BIT, %f16 - siam (1 << 2) | 3 - fadds %f0, %f16, %f1 - siam (1 << 2) | 0 - fsubs %f1, %f16, %f1 - siam (0 << 2) - retl - fors %f1, SIGN_BIT, %f0 -END (__floorf_vis3) diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf.S deleted file mode 100644 index d2a83cb9b8..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_floorf.S +++ /dev/null @@ -1,12 +0,0 @@ -#include <sparc-ifunc.h> - -SPARC_ASM_VIS3_VIS2_IFUNC(floorf) - -weak_alias (__floorf, floorf) - -# undef weak_alias -# define weak_alias(a, b) - -#define __floorf __floorf_generic - -#include "../s_floorf.S" diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_trunc-vis3.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_trunc-vis3.S deleted file mode 100644 index 7f212d927e..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_trunc-vis3.S +++ /dev/null @@ -1,57 +0,0 @@ -/* Truncate argument to nearest integral value not larger than - the argument, sparc32 v9 vis3 version. - - Copyright (C) 2013-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by David S. Miller <davem@davemloft.net>, 2013. - - The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - The GNU C Library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with the GNU C Library; if not, see - <http://www.gnu.org/licenses/>. */ - -#include <sysdep.h> -#include <math_ldbl_opt.h> - - /* We pop constants into the FPU registers using the incoming - argument stack slots, since this avoid having to use any PIC - references. We also thus avoid having to allocate a register - window. - - VIS instructions are used to facilitate the formation of - easier constants, and the propagation of the sign bit. */ - -#define TWO_FIFTYTWO 0x43300000 /* 2**52 */ - -#define ZERO %f10 /* 0.0 */ -#define SIGN_BIT %f12 /* -0.0 */ - -ENTRY (__trunc_vis3) - sethi %hi(TWO_FIFTYTWO), %o2 - sllx %o0, 32, %o0 - or %o0, %o1, %o0 - fzero ZERO - movxtod %o0, %f0 - sllx %o2, 32, %o2 - fnegd ZERO, SIGN_BIT - movxtod %o2, %f16 - fabsd %f0, %f14 - fcmpd %fcc3, %f14, %f16 - fmovduge %fcc3, ZERO, %f14 - fand %f0, SIGN_BIT, SIGN_BIT - fdtox %f14, %f14 - fxtod %f14, %f14 - faddd %f0, ZERO, %f18 - fmovduge %fcc3, %f18, %f14 - retl - for %f14, SIGN_BIT, %f0 -END (__trunc_vis3) diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_trunc.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_trunc.S deleted file mode 100644 index 3787fa1f11..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_trunc.S +++ /dev/null @@ -1,19 +0,0 @@ -#include <sparc-ifunc.h> -#include <math_ldbl_opt.h> - -SPARC_ASM_VIS3_IFUNC(trunc) - -weak_alias (__trunc, trunc) - -#if LONG_DOUBLE_COMPAT(libm, GLIBC_2_1) -compat_symbol (libm, __trunc, truncl, GLIBC_2_1) -#endif - -# undef weak_alias -# define weak_alias(a, b) -# undef compat_symbol -# define compat_symbol(a, b, c, d) - -#define __trunc __trunc_generic - -#include "../s_trunc.S" diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_truncf-vis3.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_truncf-vis3.S deleted file mode 100644 index 2acd04dae8..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_truncf-vis3.S +++ /dev/null @@ -1,53 +0,0 @@ -/* Truncate argument to nearest integral value not larger than - the argument, sparc32 v9 vis3 version. - - Copyright (C) 2013-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by David S. Miller <davem@davemloft.net>, 2013. - - The GNU C Library is free software; you can redistribute it and/or - modify it under the terms of the GNU Lesser General Public - License as published by the Free Software Foundation; either - version 2.1 of the License, or (at your option) any later version. - - The GNU C Library is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU - Lesser General Public License for more details. - - You should have received a copy of the GNU Lesser General Public - License along with the GNU C Library; if not, see - <http://www.gnu.org/licenses/>. */ - -#include <sysdep.h> - - /* We pop constants into the FPU registers using the incoming - argument stack slots, since this avoid having to use any PIC - references. We also thus avoid having to allocate a register - window. - - VIS instructions are used to facilitate the formation of - easier constants, and the propagation of the sign bit. */ - -#define TWO_TWENTYTHREE 0x4b000000 /* 2**23 */ - -#define ZERO %f10 /* 0.0 */ -#define SIGN_BIT %f12 /* -0.0 */ - -ENTRY (__truncf_vis3) - movwtos %o0, %f1 - sethi %hi(TWO_TWENTYTHREE), %o2 - fzeros ZERO - fnegs ZERO, SIGN_BIT - movwtos %o2, %f16 - fabss %f1, %f14 - fcmps %fcc3, %f14, %f16 - fmovsuge %fcc3, ZERO, %f14 - fands %f1, SIGN_BIT, SIGN_BIT - fstoi %f14, %f14 - fitos %f14, %f14 - fadds %f1, ZERO, %f18 - fmovsuge %fcc3, %f18, %f14 - retl - fors %f14, SIGN_BIT, %f0 -END (__truncf_vis3) diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_truncf.S b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_truncf.S deleted file mode 100644 index 2ca2517334..0000000000 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_truncf.S +++ /dev/null @@ -1,12 +0,0 @@ -#include <sparc-ifunc.h> - -SPARC_ASM_VIS3_IFUNC(truncf) - -weak_alias (__truncf, truncf) - -# undef weak_alias -# define weak_alias(a, b) - -#define __truncf __truncf_generic - -#include "../s_truncf.S" |