diff options
author | Joseph Myers <joseph@codesourcery.com> | 2017-08-21 17:45:10 +0000 |
---|---|---|
committer | Joseph Myers <joseph@codesourcery.com> | 2017-08-21 17:45:10 +0000 |
commit | 813378e9fe17e029caf627cab76fe23eb46815fa (patch) | |
tree | d5dff36004e083dee3b77b25d93ade581ca89e6d /sysdeps/ia64/fpu | |
parent | 4504783c0f65b7074204c6126c6255ed89d6594e (diff) | |
download | glibc-813378e9fe17e029caf627cab76fe23eb46815fa.tar.gz glibc-813378e9fe17e029caf627cab76fe23eb46815fa.tar.xz glibc-813378e9fe17e029caf627cab76fe23eb46815fa.zip |
Obsolete matherr, _LIB_VERSION, libieee.a.
This patch obsoletes support for SVID libm error handling (the system where a user-defined function matherr is called on a libm function error; only enabled if you also set _LIB_VERSION = _SVID_ or _LIB_VERSION = _XOPEN_) and the use of the _LIB_VERSION global variable to control libm error handling. matherr and _LIB_VERSION are made into compat symbols, not supported for new ports or for static linking. The libieee.a object file (which sets _LIB_VERSION = _IEEE_, so disabling errno setting for some functions) is also removed, and all the related definitions are removed from math.h. The manual already recommends against using matherr, and it's already not supported for _Float128 functions (those use new wrappers that don't support matherr, only errno) - this patch means that it becomes possible to e.g. add sinf32 as an alias to sinf without that resulting in undesired matherr support in sinf32 for existing glibc ports. matherr support is not part of any standard supported by glibc (it was removed in XPG4). Because matherr is a function to be defined by the user, of course user programs defining such a function will still continue to link; it just quietly won't be used. If they try to write to the library's copy of _LIB_VERSION to enable SVID error handling, however, they will get a link error (but if they define their own _LIB_VERSION variable, they won't). I expect the most likely case of build failures from this patch to be programs with unconditional cargo-culted uses of -lieee (based on a notion of "I want IEEE floating point", not any actual requirement for that library). Ideally, the new-port-or-static-linking case would use the new wrappers used for _Float128. This is not implemented in this patch, because of the complication of architecture-specific (powerpc32 and sparc) sqrt wrappers that use _LIB_VERSION and __kernel_standard directly. Thus, the old wrappers and __kernel_standard are still built unconditionally, and _LIB_VERSION still exists in static libm. But when the old wrappers and __kernel_standard are built in the non-compat case, _LIB_VERSION and matherr are defined as macros so code to support those features isn't actually built into static libm or new ports' shared libm after this patch. I intend to move to the new wrappers for static libm and new ports in followup patches. I believe the sqrt wrappers for powerpc32 and sparc can reasonably be removed. GCC already optimizes the normal case of sqrt by generating code that uses a hardware instruction and only calls the sqrt function if the argument was negative (if -fno-math-errno, of course, it just uses the hardware instruction without any check for negative argument being needed). Thus those wrappers will only actually get called in the case of negative arguments, which is not a case it makes sense to optimize for. But even without removing the powerpc32 and sparc wrappers it should still be possible to move to the new wrappers for static libm and new ports, just without having those dubious architecture-specific optimizations in static libm. Everything said about matherr equally applies to matherrf and matherrl (IA64-specific, undocumented), except that the structure of IA64 libm means it won't be converted to using the new wrappers (it doesn't use the old ones either, but its own error-handling code instead). As with other tests of compat symbols, I expect test-matherr and test-matherr-2 to need to become appropriately conditional once we have a system for disabling such tests for ports too new to have the relevant symbols. Tested for x86_64 and x86, and with build-many-glibcs.py. * math/math.h [__USE_MISC] (_LIB_VERSION_TYPE): Remove. [__USE_MISC] (_LIB_VERSION): Likewise. [__USE_MISC] (struct exception): Likewise. [__USE_MISC] (matherr): Likewise. [__USE_MISC] (DOMAIN): Likewise. [__USE_MISC] (SING): Likewise. [__USE_MISC] (OVERFLOW): Likewise. [__USE_MISC] (UNDERFLOW): Likewise. [__USE_MISC] (TLOSS): Likewise. [__USE_MISC] (PLOSS): Likewise. [__USE_MISC] (HUGE): Likewise. [__USE_XOPEN] (MAXFLOAT): Define even if [__USE_MISC]. * math/math-svid-compat.h: New file. * conform/linknamespace.pl (@whitelist): Remove matherr, matherrf and matherrl. * include/math.h [!_ISOMAC] (__matherr): Remove. * manual/arith.texi (FP Exceptions): Do not document matherr. * math/Makefile (tests): Change test-matherr to test-matherr-3. (tests-internal): New variable. (install-lib): Do not add libieee.a. (non-lib.a): Likewise. (extra-objs): Do not add libieee.a and ieee-math.o. (CPPFLAGS-s_lib_version.c): Remove variable. ($(objpfx)libieee.a): Remove rule. ($(addprefix $(objpfx), $(tests-internal)): Depend on $(libm). * math/ieee-math.c: Remove. * math/libm-test-support.c (matherr): Remove. * math/test-matherr.c: Use <support/test-driver.c>. Add copyright and license notices. Include <math-svid-compat.h> and <shlib-compat.h>. (matherr): Undefine as macro. Use compat_symbol_reference. (_LIB_VERSION): Likewise. * math/test-matherr-2.c: New file. * math/test-matherr-3.c: Likewise. * sysdeps/generic/math_private.h (__kernel_standard): Remove declaration. (__kernel_standard_f): Likewise. (__kernel_standard_l): Likewise. * sysdeps/ieee754/s_lib_version.c: Do not include <math.h> or <math_private.h>. Include <math-svid-compat.h>. (_LIB_VERSION): Undefine as macro. (_LIB_VERSION_INTERNAL): Always initialize to _POSIX_. Define only if [LIBM_SVID_COMPAT || !defined SHARED]. If [LIBM_SVID_COMPAT], use compat_symbol. * sysdeps/ieee754/s_matherr.c: Do not include <math.h> or <math_private.h>. Include <math-svid-compat.h>. (matherr): Undefine as macro. (__matherr): Define only if [LIBM_SVID_COMPAT]. Use compat_symbol. * sysdeps/ia64/fpu/libm_error.c: Include <math-svid-compat.h>. [_LIBC && LIBM_SVID_COMPAT] (matherrf): Use compat_symbol_reference. [_LIBC && LIBM_SVID_COMPAT] (matherrl): Likewise. [_LIBC && !LIBM_SVID_COMPAT] (matherrf): Define as macro. [_LIBC && !LIBM_SVID_COMPAT] (matherrl): Likewise. * sysdeps/ia64/fpu/libm_support.h: Include <math-svid-compat.h>. (MATHERR_D): Remove declaration. [!_LIBC] (_LIB_VERSION_TYPE): Likewise [!LIBM_BUILD] (_LIB_VERSIONIMF): Likewise. [LIBM_BUILD] (pmatherrf): Likewise. [LIBM_BUILD] (pmatherr): Likewise. [LIBM_BUILD] (pmatherrl): Likewise. (DOMAIN): Likewise. (SING): Likewise. (OVERFLOW): Likewise. (UNDERFLOW): Likewise. (TLOSS): Likewise. (PLOSS): Likewise. * sysdeps/ia64/fpu/s_matherrf.c: Include <math-svid-compat.h>. (__matherrf): Define only if [LIBM_SVID_COMPAT]. Use compat_symbol. * sysdeps/ia64/fpu/s_matherrl.c: Include <math-svid-compat.h>. (__matherrl): Define only if [LIBM_SVID_COMPAT]. Use compat_symbol. * math/lgamma-compat.h: Include <math-svid-compat.h>. * math/w_acos_compat.c: Likewise. * math/w_acosf_compat.c: Likewise. * math/w_acosh_compat.c: Likewise. * math/w_acoshf_compat.c: Likewise. * math/w_acoshl_compat.c: Likewise. * math/w_acosl_compat.c: Likewise. * math/w_asin_compat.c: Likewise. * math/w_asinf_compat.c: Likewise. * math/w_asinl_compat.c: Likewise. * math/w_atan2_compat.c: Likewise. * math/w_atan2f_compat.c: Likewise. * math/w_atan2l_compat.c: Likewise. * math/w_atanh_compat.c: Likewise. * math/w_atanhf_compat.c: Likewise. * math/w_atanhl_compat.c: Likewise. * math/w_cosh_compat.c: Likewise. * math/w_coshf_compat.c: Likewise. * math/w_coshl_compat.c: Likewise. * math/w_exp10_compat.c: Likewise. * math/w_exp10f_compat.c: Likewise. * math/w_exp10l_compat.c: Likewise. * math/w_exp2_compat.c: Likewise. * math/w_exp2f_compat.c: Likewise. * math/w_exp2l_compat.c: Likewise. * math/w_fmod_compat.c: Likewise. * math/w_fmodf_compat.c: Likewise. * math/w_fmodl_compat.c: Likewise. * math/w_hypot_compat.c: Likewise. * math/w_hypotf_compat.c: Likewise. * math/w_hypotl_compat.c: Likewise. * math/w_j0_compat.c: Likewise. * math/w_j0f_compat.c: Likewise. * math/w_j0l_compat.c: Likewise. * math/w_j1_compat.c: Likewise. * math/w_j1f_compat.c: Likewise. * math/w_j1l_compat.c: Likewise. * math/w_jn_compat.c: Likewise. * math/w_jnf_compat.c: Likewise. * math/w_jnl_compat.c: Likewise. * math/w_lgamma_main.c: Likewise. * math/w_lgamma_r_compat.c: Likewise. * math/w_lgammaf_main.c: Likewise. * math/w_lgammaf_r_compat.c: Likewise. * math/w_lgammal_main.c: Likewise. * math/w_lgammal_r_compat.c: Likewise. * math/w_log10_compat.c: Likewise. * math/w_log10f_compat.c: Likewise. * math/w_log10l_compat.c: Likewise. * math/w_log2_compat.c: Likewise. * math/w_log2f_compat.c: Likewise. * math/w_log2l_compat.c: Likewise. * math/w_log_compat.c: Likewise. * math/w_logf_compat.c: Likewise. * math/w_logl_compat.c: Likewise. * math/w_pow_compat.c: Likewise. * math/w_powf_compat.c: Likewise. * math/w_powl_compat.c: Likewise. * math/w_remainder_compat.c: Likewise. * math/w_remainderf_compat.c: Likewise. * math/w_remainderl_compat.c: Likewise. * math/w_scalb_compat.c: Likewise. * math/w_scalbf_compat.c: Likewise. * math/w_scalbl_compat.c: Likewise. * math/w_sinh_compat.c: Likewise. * math/w_sinhf_compat.c: Likewise. * math/w_sinhl_compat.c: Likewise. * math/w_sqrt_compat.c: Likewise. * math/w_sqrtf_compat.c: Likewise. * math/w_sqrtl_compat.c: Likewise. * math/w_tgamma_compat.c: Likewise. * math/w_tgammaf_compat.c: Likewise. * math/w_tgammal_compat.c: Likewise. * sysdeps/ieee754/dbl-64/w_exp_compat.c: Likewise. * sysdeps/ieee754/flt-32/w_expf_compat.c: Likewise. * sysdeps/ieee754/k_standard.c: Likewise. * sysdeps/ieee754/k_standardf.c: Likewise. * sysdeps/ieee754/k_standardl.c: Likewise. * sysdeps/ieee754/ldbl-128/w_expl_compat.c: Likewise. * sysdeps/ieee754/ldbl-128ibm/w_expl_compat.c: Likewise. * sysdeps/ieee754/ldbl-96/w_expl_compat.c: Likewise. * sysdeps/powerpc/powerpc32/power4/fpu/w_sqrt_compat.S: Likewise. * sysdeps/powerpc/powerpc32/power4/fpu/w_sqrtf_compat.S: Likewise. * sysdeps/powerpc/powerpc32/power5/fpu/w_sqrt_compat.S: Likewise. * sysdeps/powerpc/powerpc32/power5/fpu/w_sqrtf_compat.S: Likewise. * sysdeps/sparc/sparc32/fpu/w_sqrt_compat.S: Likewise. * sysdeps/sparc/sparc32/fpu/w_sqrtf_compat.S: Likewise. * sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/w_sqrt_compat-vis3.S: Likewise. * sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/w_sqrtf_compat-vis3.S: Likewise. * sysdeps/sparc/sparc32/sparcv9/fpu/w_sqrt_compat.S: Likewise. * sysdeps/sparc/sparc32/sparcv9/fpu/w_sqrtf_compat.S: Likewise. * sysdeps/sparc/sparc64/fpu/w_sqrt_compat.S: Likewise. * sysdeps/sparc/sparc64/fpu/w_sqrtf_compat.S: Likewise.
Diffstat (limited to 'sysdeps/ia64/fpu')
-rw-r--r-- | sysdeps/ia64/fpu/libm_error.c | 8 | ||||
-rw-r--r-- | sysdeps/ia64/fpu/libm_support.h | 39 | ||||
-rw-r--r-- | sysdeps/ia64/fpu/s_matherrf.c | 5 | ||||
-rw-r--r-- | sysdeps/ia64/fpu/s_matherrl.c | 5 |
4 files changed, 18 insertions, 39 deletions
diff --git a/sysdeps/ia64/fpu/libm_error.c b/sysdeps/ia64/fpu/libm_error.c index a9307dd4b5..34d1466e09 100644 --- a/sysdeps/ia64/fpu/libm_error.c +++ b/sysdeps/ia64/fpu/libm_error.c @@ -91,11 +91,19 @@ #include <stdio.h> #include <stdlib.h> #include "libm_support.h" +#include <math-svid-compat.h> #ifdef _LIBC # define pmatherr matherr # define pmatherrf matherrf # define pmatherrl matherrl +# if LIBM_SVID_COMPAT +compat_symbol_reference (libm, matherrf, matherrf, GLIBC_2_2_3); +compat_symbol_reference (libm, matherrl, matherrl, GLIBC_2_2_3); +# else +# define matherrf(EXC) ((void) (EXC), 0) +# define matherrl(EXC) ((void) (EXC), 0) +# endif #else _LIB_VERSION_TYPE #if defined( __POSIX__ ) diff --git a/sysdeps/ia64/fpu/libm_support.h b/sysdeps/ia64/fpu/libm_support.h index dc9c0a2648..4e7a66799d 100644 --- a/sysdeps/ia64/fpu/libm_support.h +++ b/sysdeps/ia64/fpu/libm_support.h @@ -83,6 +83,8 @@ #ifndef __LIBM_SUPPORT_H_INCLUDED__ #define __LIBM_SUPPORT_H_INCLUDED__ +#include <math-svid-compat.h> + #ifndef _LIBC #if !(defined(_WIN32) || defined(_WIN64)) # pragma const_seg(".rodata") /* place constant data in text (code) section */ @@ -156,36 +158,8 @@ struct exceptionl #endif extern int MATHERR_F(struct exceptionf*); -extern int MATHERR_D(struct EXC_DECL_D*); extern int matherrl(struct exceptionl*); -#ifndef _LIBC -// Add code to support _LIB_VERSIONIMF -typedef enum -{ - _IEEE_ = -1, // IEEE-like behavior - _SVID_, // SysV, Rel. 4 behavior - _XOPEN_, // Unix98 - _POSIX_, // Posix - _ISOC_ // ISO C9X -} _LIB_VERSION_TYPE; -#endif - -// This is a run-time variable and may affect -// floating point behavior of the libm functions - -#if !defined( LIBM_BUILD ) -#if defined( _DLL ) -extern _LIB_VERSION_TYPE __declspec(dllimport) _LIB_VERSIONIMF; -#else -extern _LIB_VERSION_TYPE _LIB_VERSIONIMF; -#endif /* _DLL */ -#else -extern int (*pmatherrf)(struct exceptionf*); -extern int (*pmatherr)(struct EXC_DECL_D*); -extern int (*pmatherrl)(struct exceptionl*); -#endif /* LIBM_BUILD */ - /* memory format definitions (LITTLE_ENDIAN only) */ #if !(defined(SIZE_INT_32) || defined(SIZE_INT_64)) @@ -1031,15 +1005,6 @@ struct ker80 { #endif -/* error codes */ - -#define DOMAIN 1 /* argument domain error */ -#define SING 2 /* argument singularity */ -#define OVERFLOW 3 /* overflow range error */ -#define UNDERFLOW 4 /* underflow range error */ -#define TLOSS 5 /* total loss of precision */ -#define PLOSS 6 /* partial loss of precision */ - /* */ #define VOLATILE_32 /*volatile*/ diff --git a/sysdeps/ia64/fpu/s_matherrf.c b/sysdeps/ia64/fpu/s_matherrf.c index a47d4b89a2..5808c04573 100644 --- a/sysdeps/ia64/fpu/s_matherrf.c +++ b/sysdeps/ia64/fpu/s_matherrf.c @@ -13,8 +13,10 @@ #include <math.h> #include <math_private.h> +#include <math-svid-compat.h> #include "libm_support.h" +#if LIBM_SVID_COMPAT int weak_function __matherrf(struct exceptionf *x) @@ -23,4 +25,5 @@ __matherrf(struct exceptionf *x) if(x->arg1!=x->arg1) return 0; return n; } -weak_alias (__matherrf, matherrf) +compat_symbol (libm, __matherrf, matherrf, GLIBC_2_2_3); +#endif diff --git a/sysdeps/ia64/fpu/s_matherrl.c b/sysdeps/ia64/fpu/s_matherrl.c index 6fe92d7739..dc64b31e50 100644 --- a/sysdeps/ia64/fpu/s_matherrl.c +++ b/sysdeps/ia64/fpu/s_matherrl.c @@ -13,8 +13,10 @@ #include <math.h> #include <math_private.h> +#include <math-svid-compat.h> #include "libm_support.h" +#if LIBM_SVID_COMPAT int weak_function __matherrl(struct exceptionl *x) @@ -23,4 +25,5 @@ __matherrl(struct exceptionl *x) if(x->arg1!=x->arg1) return 0; return n; } -weak_alias (__matherrl, matherrl) +compat_symbol (libm, __matherrl, matherrl, GLIBC_2_2_3); +#endif |