Commit Graph

142 Commits

Author SHA1 Message Date
thorpej caae5cd93d Use <machine/asm.h> 1996-08-13 17:00:28 +00:00
explorer a9347d2469 Fix description of ceil(3). Closes pr lib/2690 1996-08-12 22:39:34 +00:00
thorpej 688ff1ddf9 Correct think-o; this is an assembler source file (and the Makefile
expected as such).
1996-07-08 03:27:25 +00:00
jtc ca7561ba59 Fix typo 1996-07-06 00:15:45 +00:00
explorer 7e4136b303 Add a continuation slash in the commented-out i387 section 1996-07-05 22:26:51 +00:00
jtc 9fc92b384e Add new i387 math functions to ARCH_SRCS 1996-07-03 18:02:43 +00:00
jtc a5e2dfa95f use one less fp stack slot 1996-07-03 17:31:28 +00:00
jtc 0ab3a0b137 It's necessary to use 64 bit precision to avoid errors in the least
significant bit(s) (for some operands).
1996-07-03 17:27:34 +00:00
jtc 88420f2e4d this simple adaptation exp() does not work 1996-07-03 17:20:16 +00:00
jtc 5f9accdf53 Assembly versions of expf(), logf(),log10f(), scalbf(), expm1() and expm1f(). 1996-07-03 16:50:20 +00:00
jtc 991d9c139a we don't need to branch 1996-06-26 14:44:05 +00:00
jtc 6f5348dbbd Enable mc68881 version of fmod 1996-06-26 14:39:49 +00:00
jtc 3a2de776bf add m68881 specific fmod() 1996-06-26 14:36:14 +00:00
jtc 6655b8bb1f enable alpha-specific libm files 1996-06-26 07:38:42 +00:00
jtc c13e010e3c Alpha versions of copysign() and copysignf() 1996-06-26 07:29:59 +00:00
jtc 5743dc5d7a Fix from Bruce Evans. Commited to FreeBSD earlier this month, but was
never forwarded to me :-(.

> Clean up the FP stack before returning.  The i387 exp() leaked an FP
> register on its first call.  Subsequent calls reused the register so
> the leak didn't accumulate.
1996-06-25 02:07:09 +00:00
jtc 84fc678011 Use setne instead of setnel, new versions of gas balk at the bad mnemonic.
Wrong mask was used in s_finitef.S.
1996-06-04 18:00:34 +00:00
mark ce67b636ad Get the word order of doubles correct for the arm32 port.
Note: The arm32 port is little endian except for the FP word order.
1996-05-07 20:26:07 +00:00
phil d9f99fc936 Removing a "volatile" put there for the ns32k which is no longer needed. 1996-04-08 15:43:41 +00:00
jtc 3bf9d01bf7 infnan(3) no longer exists. 1996-03-30 01:33:59 +00:00
jtc ee341062c6 Changed 3m to 3. There is no 3m manpage section. 1996-03-30 01:28:55 +00:00
jtc 977400602b Changed w_gamma.c, w_gamma_r.c, w_gammaf.c, w_gammaf_r.c to call
__ieee754_lgamma_r or __ieee754_lgammaf_r.
Removed e_gamma.c, e_gamma_r.c, e_gammaf.c, e_gammaf_r.c, e_lgamma.c,
e_lgamma_f.c.
1995-11-20 22:06:19 +00:00
pk 26646c7969 Setup stack-frame correctly (same as s_ilogb.S). 1995-10-22 20:32:43 +00:00
jtc e2143cf975 Fix from Jochen Pohl, i387 version of ilogb didn't set up the stack
frame correctly (PR #1590).
1995-10-12 15:53:09 +00:00
ragge a6a78ddeb9 Got Makefile to understand to use non-IEEE files on vax systems. 1995-10-10 23:43:31 +00:00
ragge d76e5e0af8 Machine specific math library files for vax. 1995-10-10 23:40:21 +00:00
ragge 25d43745af Machine independent math library files for non-IEEE architectures. 1995-10-10 23:36:31 +00:00
jtc c1c8f42080 Changed core routines to call __ieee754_sqrt() instead of sqrt(). I
reported this enhancement to fdlibm-comments, and got the following
reply:

Date: Thu, 11 May 1995 14:35:25 -0700
From: Kwok.Ng@Eng.Sun.COM (KC Ng)
To: fdlibm-comments@sunpro.Eng.Sun.COM, jtc@cygnus.com
Subject: Re: fdlibm 5.2: why do core functions use sqrt?

> I noticed that core (e_*.c) fdlibm functions like __ieee754_acos()
> ensure that they call sqrt() with arguments in range (x > 0), when
> they could call __ieee754_sqrt() directly.
>
> Since sqrt() does a lot more work (verifies x is in range, etc.) is
> there any reason for this?  I'd think that calling __iee754_sqrt()
> would be more appropriate.  ....

You are right. __ieee754_sqrt should be in use with e_*.c.
1995-05-12 04:57:13 +00:00
jtc 9ae90685f2 Winning Strategies has placed this code in the public domain. 1995-05-11 23:03:44 +00:00
jtc d042a3eb20 Id -> NetBSD 1995-05-10 20:44:22 +00:00
jtc 5abf82e236 Id -> NetBSD 1995-05-08 23:55:16 +00:00
jtc 0898887210 Id -> NetBSD 1995-05-08 23:44:37 +00:00
jtc 750e07c91c i387 float version of atan2() 1995-05-08 23:35:10 +00:00
jtc 11fa45b258 Add i387 "float" functions to the makefile 1995-05-03 20:42:23 +00:00
jtc 684161edae Added i387 remainderf and sqrtf 1995-05-03 14:47:35 +00:00
jtc 5c7208348c Winning Strategies has placed this code into the Public Domain 1995-04-28 22:43:31 +00:00
jtc 5e5fe216ab removed log2 from manpage, libm doesn't provide that function. 1995-04-27 18:21:43 +00:00
jtc df8b005f6e First cut at adding float versions of math functions to the manpages. 1995-04-14 18:39:53 +00:00
jtc d6c3ebedb4 Fix overlooked FIXME: unsigned int -> u_int32_t. 1995-03-25 01:48:53 +00:00
jtc fd8d17a56d First pass at documenting "float" versions of these functions. 1995-03-04 01:36:29 +00:00
mycroft 4c34be6b52 Clean up deleted files. 1995-03-01 00:00:00 +00:00
jtc ea66efc784 Enable mc68881 versions of remainder() and scalb(). 1995-02-18 01:46:41 +00:00
jtc 8961dd234b Special case 0, as was done in s_scalbn.S.
I'm only doing this for consistancy --- someone really needs to measure
whether this optimization is justified.
1995-02-18 01:39:51 +00:00
jtc 324e516335 68881-assisted versions of these functions. 1995-02-18 01:24:42 +00:00
jtc 6a189fc179 Remove unused static constant "one". 1995-01-25 01:07:23 +00:00
jtc 3d9c0cdc87 Integrate last of the relavant changes from fdlibm 5.2:
4. Performance issue on rem_pio2
       An attempt to speed up the argument reduction in the trig function is to
       consider pi/4 < x < 3pi/4 a special case. This was done in the file
       e_rem_pio2.c
1995-01-25 01:06:35 +00:00
jtc 10a7ebda19 Bug fix from fdlibm 5.2. 1995-01-16 20:05:43 +00:00
cgd 3a370daf86 specify man pages the new way. 1994-12-22 09:44:45 +00:00
jtc 600d3f84d5 Add volatile qualifier to huge and tiny, so that huge*huge and tiny*tiny
are performed at run-time.
1994-12-16 08:38:20 +00:00
jtc c446bbd004 Added s_isinff.c 1994-09-25 17:31:51 +00:00