AuroraMiddleware/glibc
1
0
Fork 0
mirror of https://sourceware.org/git/glibc.git synced 2025-01-07 18:10:07 +00:00
Code Issues Packages Projects Releases Wiki Activity
34,306 Commits 335 Branches 1,646 Tags 310 MiB
80190d2b0e
Commit Graph

3 Commits

Author SHA1 Message Date
Joseph Myers
e44acb2063 Use floor functions not __floor functions in glibc libm. 
Similar to the changes that were made to call sqrt functions directly
in glibc, instead of __ieee754_sqrt variants, so that the compiler
could inline them automatically without needing special inline
definitions in lots of math_private.h headers, this patch makes libm
code call floor functions directly instead of __floor variants,
removing the inlines / macros for x86_64 (SSE4.1) and powerpc
(POWER5).

The redirection used to ensure that __ieee754_sqrt does still get
called when the compiler doesn't inline a built-in function expansion
is refactored so it can be applied to other functions; the refactoring
is arranged so it's not limited to unary functions either (it would be
reasonable to use this mechanism for copysign - removing the inline in
math_private_calls.h but also eliminating unnecessary local PLT entry
use in the cases (powerpc soft-float and e500v1, for IBM long double)
where copysign calls don't get inlined).

The point of this change is that more architectures can get floor
calls inlined where they weren't previously (AArch64, for example),
without needing special inline definitions in their math_private.h,
and existing such definitions in math_private.h headers can be
removed.

Note that it's possible that in some cases an inline may be used where
an IFUNC call was previously used - this is the case on x86_64, for
example.  I think the direct calls to floor are still appropriate; if
there's any significant performance cost from inline SSE2 floor
instead of an IFUNC call ending up with SSE4.1 floor, that indicates
that either the function should be doing something else that's faster
than using floor at all, or it should itself have IFUNC variants, or
that the compiler choice of inlining for generic tuning should change
to allow for the possibility that, by not inlining, an SSE4.1 IFUNC
might be called at runtime - but not that glibc should avoid calling
floor internally.  (After all, all the same considerations would apply
to any user program calling floor, where it might either be inlined or
left as an out-of-line call allowing for a possible IFUNC.)

Tested for x86_64, and with build-many-glibcs.py.

	* include/math.h [!_ISOMAC && !(__FINITE_MATH_ONLY__ &&
	__FINITE_MATH_ONLY__ > 0) && !NO_MATH_REDIRECT] (MATH_REDIRECT):
	New macro.
	[!_ISOMAC && !(__FINITE_MATH_ONLY__ && __FINITE_MATH_ONLY__ > 0)
	&& !NO_MATH_REDIRECT] (MATH_REDIRECT_LDBL): Likewise.
	[!_ISOMAC && !(__FINITE_MATH_ONLY__ && __FINITE_MATH_ONLY__ > 0)
	&& !NO_MATH_REDIRECT] (MATH_REDIRECT_F128): Likewise.
	[!_ISOMAC && !(__FINITE_MATH_ONLY__ && __FINITE_MATH_ONLY__ > 0)
	&& !NO_MATH_REDIRECT] (MATH_REDIRECT_UNARY_ARGS): Likewise.
	[!_ISOMAC && !(__FINITE_MATH_ONLY__ && __FINITE_MATH_ONLY__ > 0)
	&& !NO_MATH_REDIRECT] (sqrt): Redirect using MATH_REDIRECT.
	[!_ISOMAC && !(__FINITE_MATH_ONLY__ && __FINITE_MATH_ONLY__ > 0)
	&& !NO_MATH_REDIRECT] (floor): Likewise.
	* sysdeps/aarch64/fpu/s_floor.c: Define NO_MATH_REDIRECT before
	header inclusion.
	* sysdeps/aarch64/fpu/s_floorf.c: Likewise.
	* sysdeps/ieee754/dbl-64/s_floor.c: Likewise.
	* sysdeps/ieee754/dbl-64/wordsize-64/s_floor.c: Likewise.
	* sysdeps/ieee754/float128/s_floorf128.c: Likewise.
	* sysdeps/ieee754/flt-32/s_floorf.c: Likewise.
	* sysdeps/ieee754/ldbl-128/s_floorl.c: Likewise.
	* sysdeps/ieee754/ldbl-128ibm/s_floorl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_floor_template.c: Likewise.
	* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_floor.c: Likewise.
	* sysdeps/powerpc/powerpc32/power4/fpu/multiarch/s_floorf.c: Likewise.
	* sysdeps/powerpc/powerpc64/fpu/multiarch/s_floor.c: Likewise.
	* sysdeps/powerpc/powerpc64/fpu/multiarch/s_floorf.c: Likewise.
	* sysdeps/riscv/rv64/rvd/s_floor.c: Likewise.
	* sysdeps/riscv/rvf/s_floorf.c: Likewise.
	* sysdeps/sparc/sparc64/fpu/multiarch/s_floor.c: Likewise.
	* sysdeps/sparc/sparc64/fpu/multiarch/s_floorf.c: Likewise.
	* sysdeps/x86_64/fpu/multiarch/s_floor.c: Likewise.
	* sysdeps/x86_64/fpu/multiarch/s_floorf.c: Likewise.
	* sysdeps/powerpc/fpu/math_private.h [_ARCH_PWR5X] (__floor):
	Remove macro.
	[_ARCH_PWR5X] (__floorf): Likewise.
	* sysdeps/x86_64/fpu/math_private.h [__SSE4_1__] (__floor): Remove
	inline function.
	[__SSE4_1__] (__floorf): Likewise.
	* math/w_lgamma_main.c (LGFUNC (__lgamma)): Use floor functions
	instead of __floor variants.
	* math/w_lgamma_r_compat.c (__lgamma_r): Likewise.
	* math/w_lgammaf_main.c (LGFUNC (__lgammaf)): Likewise.
	* math/w_lgammaf_r_compat.c (__lgammaf_r): Likewise.
	* math/w_lgammal_main.c (LGFUNC (__lgammal)): Likewise.
	* math/w_lgammal_r_compat.c (__lgammal_r): Likewise.
	* math/w_tgamma_compat.c (__tgamma): Likewise.
	* math/w_tgamma_template.c (M_DECL_FUNC (__tgamma)): Likewise.
	* math/w_tgammaf_compat.c (__tgammaf): Likewise.
	* math/w_tgammal_compat.c (__tgammal): Likewise.
	* sysdeps/ieee754/dbl-64/e_lgamma_r.c (sin_pi): Likewise.
	* sysdeps/ieee754/dbl-64/k_rem_pio2.c (__kernel_rem_pio2):
	Likewise.
	* sysdeps/ieee754/dbl-64/lgamma_neg.c (__lgamma_neg): Likewise.
	* sysdeps/ieee754/flt-32/e_lgammaf_r.c (sin_pif): Likewise.
	* sysdeps/ieee754/flt-32/lgamma_negf.c (__lgamma_negf): Likewise.
	* sysdeps/ieee754/ldbl-128/e_lgammal_r.c (__ieee754_lgammal_r):
	Likewise.
	* sysdeps/ieee754/ldbl-128/e_powl.c (__ieee754_powl): Likewise.
	* sysdeps/ieee754/ldbl-128/lgamma_negl.c (__lgamma_negl):
	Likewise.
	* sysdeps/ieee754/ldbl-128/s_expm1l.c (__expm1l): Likewise.
	* sysdeps/ieee754/ldbl-128ibm/e_lgammal_r.c (__ieee754_lgammal_r):
	Likewise.
	* sysdeps/ieee754/ldbl-128ibm/e_powl.c (__ieee754_powl): Likewise.
	* sysdeps/ieee754/ldbl-128ibm/lgamma_negl.c (__lgamma_negl):
	Likewise.
	* sysdeps/ieee754/ldbl-128ibm/s_expm1l.c (__expm1l): Likewise.
	* sysdeps/ieee754/ldbl-128ibm/s_truncl.c (__truncl): Likewise.
	* sysdeps/ieee754/ldbl-96/e_lgammal_r.c (sin_pi): Likewise.
	* sysdeps/ieee754/ldbl-96/lgamma_negl.c (__lgamma_negl): Likewise.
	* sysdeps/powerpc/power5+/fpu/s_modf.c (__modf): Likewise.
	* sysdeps/powerpc/power5+/fpu/s_modff.c (__modff): Likewise.
 2018-09-14 13:09:01 +00:00
Joseph Myers
688903eb3e Update copyright dates with scripts/update-copyrights. 
* All files with FSF copyright notices: Update copyright dates
	using scripts/update-copyrights.
	* locale/programs/charmap-kw.h: Regenerated.
	* locale/programs/locfile-kw.h: Likewise.
 2018-01-01 00:32:25 +00:00
Joseph Myers
e53df1dee8 Rework m68k libm functions to use declare_mgen_alias. 
Many m68k libm functions use their own system to share code between
different types and functions, involving defining macros before
including code for another function (for example, s_atan.c also acts
as a template that can define other functions).

Thes files serving as templates generate function aliases directly
with e.g. "weak_alias (__CONCATX(__,FUNC), FUNC)" in s_atan.c.  To be
prepared to generate _Float32, _Float64 and _Float32x function
aliases, this needs changing so that the libm_alias_* macros get used
instead.  As the macro to use varies depending on the type, that would
mean additional macros to define in several different places to get
the appropriate alias-generation macro used in each case.

Rather than adding to the m68k-specific mechanisms, this patch
converts the functions in question to use something closer to the
math/ type-generic template mechanism.  After this patch, these
functions have m68k-specific templates such as s_atan_template.c, but
those templates use all the same macros as in the math/ templates,
such as FLOAT, M_DECL_FUNC, M_SUF and declare_mgen_alias.  There is no
automatic generation of the files such as s_atan.c that include the
appropriate math-type-macros-*.h header and the template file (the
existing automatic generation logic is only applicable for the fixed
set of templates listed in math/ - and sysdeps sources always override
files generated that way), so those files are still checked in, but
they are all the obvious two-line files (with one additional
definition in the case of the expm1 implementations), rather than
making e.g. s_atan.c special.

Functions are only converted where they should have aliases for
_FloatN / _FloatNx types.  Those m68k functions that do not generate
public names (those that only generate __ieee754_*, with wrappers
generating the public names, and classification functions that only
exist once per format not once per type so don't get aliases) are
unchanged.  However, log1p (public names generated by wrapper) and
significand (not provided for new types so no new aliases needed)
needed changing because they previously included the atan
implementations.  Now, s_significand.c is the main implementation for
functions with that prototype and using the old implementation
approach, while log1p includes it in place of atan.

Any further cleanups in this area (which preserve the proper set of
functions getting aliases defined by libm_alias_float and
libm_alias_double) are of course welcome, just not needed for the
goals of this patch.

Tested with build-many-glibcs.py that installed stripped shared
libraries are unchanged by the patch.

	* sysdeps/m68k/m680x0/fpu/s_atan_template.c: New file.
	* sysdeps/m68k/m680x0/fpu/s_ceil_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_cos_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_expm1_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_fabs_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_floor_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_frexp_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_lrint_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_modf_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_nearbyint_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_remquo_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_rint_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_sin_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_sincos_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_tan_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_tanh_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_trunc_template.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_atan.c: Reimplement to use
	s_atan_template.c.
	* sysdeps/m68k/m680x0/fpu/s_atanf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_atanl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_ceil.c: Reimplement to use
	s_ceil_template.c.
	* sysdeps/m68k/m680x0/fpu/s_ceilf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_ceill.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_cos.c: Reimplement to use
	s_cos_template.c.
	* sysdeps/m68k/m680x0/fpu/s_cosf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_cosl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_expm1.c: Reimplement to use
	s_expm1_template.c.
	* sysdeps/m68k/m680x0/fpu/s_expm1f.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_expm1l.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_fabs.c: Reimplement to use
	s_fabs_template.c.
	* sysdeps/m68k/m680x0/fpu/s_fabsf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_fabsl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_floor.c: Reimplement to use
	s_floor_template.c.
	* sysdeps/m68k/m680x0/fpu/s_floorf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_floorl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_frexp.c: Reimplement to use
	s_frexp_template.c.
	* sysdeps/m68k/m680x0/fpu/s_frexpf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_lrint.c: Reimplement to use
	s_lrint_template.c.
	* sysdeps/m68k/m680x0/fpu/s_lrintf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_lrintl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_modf.c: Reimplement to use
	s_modf_template.c.
	* sysdeps/m68k/m680x0/fpu/s_modff.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_modfl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_nearbyint.c: Reimplement to use
	s_nearbyint_template.c.
	* sysdeps/m68k/m680x0/fpu/s_nearbyintf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_nearbyintl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_remquo.c: Reimplement to use
	s_remquo_template.c.
	* sysdeps/m68k/m680x0/fpu/s_remquof.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_remquol.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_rint.c: Reimplement to use
	s_rint_template.c.
	* sysdeps/m68k/m680x0/fpu/s_rintf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_rintl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_sin.c: Reimplement to use
	s_sin_template.c.
	* sysdeps/m68k/m680x0/fpu/s_sinf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_sinl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_sincos.c: Reimplement to use
	s_sincos_template.c.
	* sysdeps/m68k/m680x0/fpu/s_sincosf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_sincosl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_tan.c: Reimplement to use
	s_tan_template.c.
	* sysdeps/m68k/m680x0/fpu/s_tanf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_tanl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_tanh.c: Reimplement to use
	s_tanh_template.c.
	* sysdeps/m68k/m680x0/fpu/s_tanhf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_tanhl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_trunc.c: Reimplement to use
	s_trunc_template.c.
	* sysdeps/m68k/m680x0/fpu/s_truncf.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_truncl.c: Likewise.
	* sysdeps/m68k/m680x0/fpu/s_significand.c: Reimplement based on
	s_atan.c instead of including s_atan.c.
	* sysdeps/m68k/m680x0/fpu/s_significandf.c: Reimplement based on
	s_atanf.c instead of including s_atanf.c.
	* sysdeps/m68k/m680x0/fpu/s_significandl.c: Reimplement based on
	s_atanl.c instead of including s_atanl.c.
	* sysdeps/m68k/m680x0/fpu/s_log1p.c: Include s_significand.c
	instead of s_atan.c.
	* sysdeps/m68k/m680x0/fpu/s_log1pf.c: Include s_significandf.c
	instead of s_atanf.c.
	* sysdeps/m68k/m680x0/fpu/s_log1pl.c: Include s_significandl.c
	instead of s_atanl.c.
 2017-11-30 22:39:07 +00:00