From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Tue, 1 May 2001 23:29:43 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Tue, 1 May 2001 23:29:35 -0400 Received: from chac.inf.utfsm.cl ([200.1.19.54]:26628 "EHLO chac.inf.utfsm.cl") by vger.kernel.org with ESMTP id ; Tue, 1 May 2001 23:29:16 -0400 Message-Id: <200105020328.f423Ssvf011202@sleipnir.valparaiso.cl> To: sjhill@cotw.com cc: lkml Subject: Re: Question on including 'math.h' from C runtime... In-Reply-To: Message from "Steven J. Hill" of "Tue, 01 May 2001 22:17:23 EST." <3AEF7C43.9955C970@cotw.com> Date: Tue, 01 May 2001 23:28:54 -0400 From: Horst von Brand Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org "Steven J. Hill" said: > I checked in the archives and did not see a discussion of this > anywhere. I have received some Linux kernel code from a project > that I have inherited and a couple of the drivers are including > math.h from the C library. This being the header file from > '/usr/include/math.h' in most cases. There are only two places > in the kernel that also include this header file. They are: > > drivers/atm/iphase.c > drivers/net/hamradio/soundmodem/gentbl.c > > As far as I can tell '/usr/include/math.h' is just full of > defines and the header files it includes are also a bunch > of defines with a few macro functions sprinkled in. Can someone > shed light on if this is bad or not and why it would be done > or necessary? Thanks. Floating point in-kernel is a no-no-no. What you see is use in helper programs that are compiled as part of the kernel build. Check if that is your case too. -- Horst von Brand vonbrand@sleipnir.valparaiso.cl Casilla 9G, Vin~a del Mar, Chile +56 32 672616