aboutsummaryrefslogtreecommitdiff
path: root/lib/msun
Commit message (Collapse)AuthorAgeFilesLines
* Prepare for mdoc(7)NG.Ruslan Ermilov2000-12-292-2/+1
| | | | Notes: svn path=/head/; revision=70481
* mdoc(7) police: use canonical form of .Dd macro.Ruslan Ermilov2000-12-111-1/+1
| | | | Notes: svn path=/head/; revision=69857
* mdoc(7) police: Er macro usage cleanup.Ruslan Ermilov2000-11-223-3/+3
| | | | Notes: svn path=/head/; revision=69051
* mdoc(7) police: Nm -> Fn where appropriate.Ruslan Ermilov2000-11-202-4/+4
| | | | Notes: svn path=/head/; revision=68946
* Fix #include orderBrian Somers2000-10-151-1/+1
| | | | | | | Spotted by: imura Notes: svn path=/head/; revision=67166
* Removed bogus 'l' suffixes in FP register to register instructions.Bruce Evans2000-06-066-8/+8
| | | | Notes: svn path=/head/; revision=61335
* MFS: Add a "magic" comment to help fixincludes realize it doesn't need toDavid E. O'Brien2000-03-271-0/+2
| | | | | | | | | modify this file when building GCC 2.96 [by hand or via the port]. Submitted by: Zack Weinberg <zack@wolery.cumb.org> Notes: svn path=/head/; revision=58647
* Remove more single-space hard sentence breaks.Sheldon Hearn2000-03-021-1/+2
| | | | Notes: svn path=/head/; revision=57695
* Remove single-space hard sentence breaks. These degrade the qualitySheldon Hearn2000-03-022-2/+4
| | | | | | | | of the typeset output, tend to make diffs harder to read and provide bad examples for new-comers to mdoc. Notes: svn path=/head/; revision=57686
* style fixAlexey Zelkin1999-11-091-5/+8
| | | | | | | | PR: docs/14737 Submitted by: Norihiro Kumagai <kuma@nk.rim.or.jp> Notes: svn path=/head/; revision=53036
* $Id$ -> $FreeBSD$Peter Wemm1999-08-28217-217/+217
| | | | Notes: svn path=/head/; revision=50476
* Disable building with alpha software completion options until we upgradeDoug Rabson1998-12-241-2/+3
| | | | | | | compilers. Notes: svn path=/head/; revision=42044
* Implement fpsetmask() and other fp*() functions. Programs should useDoug Rabson1998-12-231-1/+2
| | | | | | | | | | | | | #include <ieeefp.h> to access these functions instead of the i386 specific #include <machine/floatingpoint.h> Submitted by: Hidetoshi Shimokawa <simokawa@sat.t.u-tokyo.ac.jp> Notes: svn path=/head/; revision=42029
* There is no alpha asm code like on i386, so all the functions thatJohn Birrell1998-05-101-1/+28
| | | | | | | the i386 builds with a __generic prefix need to have that stripped. Notes: svn path=/head/; revision=35925
* Back out last changePaul Traina1998-04-231-2/+2
| | | | Notes: svn path=/head/; revision=35399
* Fix cabs and cabsf definitions to be prototypes.Paul Traina1998-04-221-3/+3
| | | | Notes: svn path=/head/; revision=35380
* Add alpha support. m68k crept in too. Oops. 8-)John Birrell1998-02-201-1/+16
| | | | Notes: svn path=/head/; revision=33662
* Document the fpgetprec/fpsetprec functions in their man page.Jonathan Lemon1998-02-041-1/+10
| | | | | | | | | Add cross-references to the elusive fpsetmask() function to various other man pages. Reviewed by: bde Notes: svn path=/head/; revision=33107
* PR: 5489Daniel O'Callaghan1998-01-151-5/+5
| | | | | | | | Submitted by: Steve G. Kargl <kargl@troutmask.apl.washington.edu> Repair corrupted text. Notes: svn path=/head/; revision=32534
* This is the only alpha math source that NetBSD has.John Birrell1998-01-102-0/+90
| | | | Notes: svn path=/head/; revision=32405
* Hide the declaration of `struct exception' from C++, since it conflictsBruce Evans1997-08-311-1/+5
| | | | | | | | with the standard C++ `class exception'. This makes matherr() difficult to use in C++. Small loss. Notes: svn path=/head/; revision=28971
* Fixed minor bugs related to the addition of gammaf.Bruce Evans1997-07-131-3/+3
| | | | | | | | The major bug, that gamma is documented as really being gamma, is still unfixed. Notes: svn path=/head/; revision=27371
* Fixed wrong mnemonic `setnel' that gas happened to generate correct objectBruce Evans1997-04-301-2/+2
| | | | | | | | | code for. Obtained from: a slightly different fix in NetBSD Notes: svn path=/head/; revision=25322
* Added -D_ARCH_INDIRECT=i387_ to CFLAGS. _ARCH_INDIRECT will soon be usedBruce Evans1997-04-151-4/+5
| | | | | | | | | to control generation of indirections in ENTRY(). Only msun needs it. Use ${ARCH} consistently. Notes: svn path=/head/; revision=24964
* Fixed synopsis. Some float functions claimed to have the same name asBruce Evans1997-03-193-6/+6
| | | | | | | the double version. Notes: svn path=/head/; revision=24011
* Use __ieee754_sqrt() instead of sqrt() internally. Similarly for theBruce Evans1997-03-0912-33/+33
| | | | | | | | | float versions. Using sqrt() was inefficient. Obtained from: NetBSD Notes: svn path=/head/; revision=23579
* Include <machine/asm.h> instead of kernel-only <machine/asmacros.h>.Bruce Evans1997-03-0924-48/+48
| | | | Notes: svn path=/head/; revision=23577
* Fixed wrong magic numbers in scaling. hypotf() was very broken for largeBruce Evans1997-03-051-8/+8
| | | | | | | | | | | | | and small values: hypotf(2.3819765e+38, 2.0416943e+38) was NaN instead of 3.1372484e+38 hypotf(-3.4028235e+38, 3.3886450e+38) was NaN instead of Inf hypotf(-2.8025969e-45, -2.8025969e-45) was 0 instead of 4.2038954e-45 Found by: ucbtest Notes: svn path=/head/; revision=23397
* Revert $FreeBSD$ to $Id$Peter Wemm1997-02-22215-215/+215
| | | | Notes: svn path=/head/; revision=22993
* Split up the Bessel function wrapper files so that most wrapper functionsBruce Evans1997-02-2013-177/+315
| | | | | | | are in their own file. Notes: svn path=/head/; revision=22948
* Removed misplaced duplicate of comment about implementation details.Bruce Evans1997-02-201-22/+0
| | | | Notes: svn path=/head/; revision=22947
* Compute (1 - x^2) as ((1 - x) * (1 + x)) instead of as (1 - x * x) toBruce Evans1997-02-202-11/+23
| | | | | | | | | | | | | | | | | | avoid easily avoidable loss of precision when |x| is nearly 1. Extended (64-bit) precision only moves the meaning of "nearly" here. This probably could be done better by splitting up the range into |x| <= 0.5 and |x| > 0.5 like the C version. However, ucbtest does't report any errors in this version. Perhaps the C version should be used anyway. It's only 25% slower now on a P5, provided the C version of sqrt() isn't used, and the C version could be optimized better. Errors checked by: ucbtest Notes: svn path=/head/; revision=22946
* Select between the generic math functions and the i387-specific onesBruce Evans1997-02-1625-56/+110
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | at runtime. etc/make.conf: Nuked HAVE_FPU option. lib/msun/Makefile: Always build the i387 objects. Copy the i387 source files at build time so that the i387 objects have different names. This is simpler than renaming the files in the cvs repository or repeating half of bsd.lib.mk to add explicit rules. lib/msun/src/*.c: Renamed all functions that have an i387-specific version by adding `__generic_' to their names. lib/msun/src/get_hw_float.c: New file for getting machdep.hw_float from the kernel. sys/i386/include/asmacros.h: Abuse the ENTRY() macro to generate jump vectors and associated code. This works much like PIC PLT dynamic initialization. The PIC case is messy. The old i387 entry points are renamed. Renaming is easier here because the names are given by macro expansions. Notes: svn path=/head/; revision=22808
* Fixed the i87 version of exp(). It returned NaN for args +-Inf. It hadBruce Evans1997-02-161-0/+42
| | | | | | | | | some small (one or two ULP) inaccuracies. Found by: ucbtest Notes: svn path=/head/; revision=22802
* Put back .endif clobbered by the previous commit, breaking theJordan K. Hubbard1997-02-151-0/+1
| | | | | | | build. Notes: svn path=/head/; revision=22748
* Disabled the i387 version if log1p(). It just evaluates log(1 + x).Bruce Evans1997-02-151-2/+4
| | | | | | | | | | | | This defeats the point of log1p(). ucbtest reports errors of +-5e+15 ULPs. A correct version would use the i387 fyl2xp1 instruction for small x and maybe scale to small x. The C version does the scaling reasonably efficiently, and fyl2px1 is slow (at least on P5s), so not much is lost by always using the C version (only 25% for small x even with the broken i387 version; 50% for large x). Notes: svn path=/head/; revision=22731
* Sort cross references.Wolfram Schneider1997-01-2013-33/+33
| | | | Notes: svn path=/head/; revision=21907
* Make the long-awaited change from $Id$ to $FreeBSD$Jordan K. Hubbard1997-01-14208-208/+208
| | | | | | | | | | | This will make a number of things easier in the future, as well as (finally!) avoiding the Id-smashing problem which has plagued developers for so long. Boy, I'm glad we're not using sup anymore. This update would have been insane otherwise. Notes: svn path=/head/; revision=21673
* Delete -D_POSIX_MODE and -D_MULTI_LIBM from CFLAGS. They never had any effectGarrett Wollman1997-01-081-2/+2
| | | | | | | | | | because _IEEE_LIBM always takes priority, so the definition just served to confuse. Reviewed by: bde Notes: svn path=/head/; revision=21435
* comma typosWolfram Schneider1996-12-231-3/+3
| | | | Notes: svn path=/head/; revision=20888
* Removed references to nonexistent functions log2() and log2f().Bruce Evans1996-12-181-3/+1
| | | | Notes: svn path=/head/; revision=20648
* Fixed fiddling with the control word. Use the stack space reserved forBruce Evans1996-12-142-12/+12
| | | | | | | that purpose instead of space below the stack. Notes: svn path=/head/; revision=20447
* Clean up the FPU stack before returning. One stack slot was leaked onBruce Evans1996-12-142-2/+4
| | | | | | | | every call. The damage was sometimes limited by other routines using and freeing stack slots that should have been empty to being with. Notes: svn path=/head/; revision=20445
* consistancy fixupPeter Wemm1996-08-301-2/+2
| | | | | | | Submitted by: "Philippe Charnier" <charnier@xp11.frmug.org> Notes: svn path=/head/; revision=17958
* cmp -s || install -c ==> install -CPeter Wemm1996-08-301-5/+3
| | | | Notes: svn path=/head/; revision=17935
* Made rintf() actually work. It was completely broken (when s_rint.cBruce Evans1996-08-282-6/+19
| | | | | | | | | | | | | | was compiled with -O) by the precision bug in the i386 version of gcc (assignments and casts don't clip the precision). E.g., rintf(12.3456789) was 12.125. Avoid the same bug in rint(). It was only broken for the unusual case when the i387 precision is 64 bits. FreeBSD defaults to 53 bit precision to avoid problems like this, but the standard math emulator always uses 64 bit precision. Notes: svn path=/head/; revision=17860
* Fix up the NAME lines forthe ceil and floor man pages to beMike Pritchard1996-08-212-4/+4
| | | | | | | | | | less confusing. Reviewed by: bde Partially obtained from: NetBSD-bugs Notes: svn path=/head/; revision=17756
* General -Wall warning cleanup, part I.Jordan K. Hubbard1996-07-127-16/+16
| | | | | | | Submitted-By: Kent Vander Velden <graphix@iastate.edu> Notes: svn path=/head/; revision=17141
* Clean up the FP stack before returning. The i387 exp() leaked an FPBruce Evans1996-06-011-1/+2
| | | | | | | | register on its first call. Subsequent calls reused the register so the leak didn't accumulate. Fixes PR 1275. Notes: svn path=/head/; revision=16054
* Don't trash %ebp.Bruce Evans1996-02-251-2/+2
| | | | | | | Obtained from: NetBSD Notes: svn path=/head/; revision=14251
generated by cgit v1.2.3 (git 2.25.1) at 2024-10-13 20:27:00 +0000