Re: lib/46433: tests/lib/libm/t_exp should not use exp() itself

[Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>]

The following reply was made to PR lib/46433; it has been noted by GNATS.

From: Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>
To: jruohonen%iki.fi@localhost
Cc: gnats-bugs%NetBSD.org@localhost, isaki%NetBSD.org@localhost, 
tsutsui%ceres.dti.ne.jp@localhost
Subject: Re: lib/46433: tests/lib/libm/t_exp should not use exp() itself
Date: Thu, 31 May 2012 00:23:05 +0900

 >  To Tsutsui: by immediate fix I mean that exp(3) (and maybe others?) are
 >  still completely broken on this architecture, with or without a test. I am
 >  not sure why systems without FPU even allow libm(3) or expose functions from
 >  there...
 
 exp(3) in src/lib/libm/arch/mc68881 is not broken.
 FPU emulator in src/sys/arch/m68k/fpe is still in progress.
 Users who need immediate fix could use MKSOFTFLOAT environment.
 Current tests are not useful with or without correct FP implementation
 because they could return false positive, and skipping tests doesn't
 help users (and developers) at all.
 
 It's much better if libm tests can be used to check kernel
 FPE implementation, because m68k FPU has exponetial (and logarithm,
 hyperboric, trigonometric etc.) instructions and MD libm functions
 directly call them:
 http://nxr.netbsd.org/xref/src/lib/libm/arch/mc68881/e_exp.S#43
 That's the point of this PR.
 
 ---
 Izumi Tsutsui