From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gate.crashing.org (gate.crashing.org [63.228.1.57]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTPS id C61F8B6EDF for ; Sat, 26 Jun 2010 00:55:33 +1000 (EST) From: Kumar Gala Content-Type: text/plain; charset=us-ascii Subject: HWCAP for fsqrt vs PPC64 Date: Fri, 25 Jun 2010 09:51:42 -0500 Message-Id: To: Alan Modra , Paul Mackerras , Wienskoski Edmar-RA8797 , Stewart Katherine-RA4863 , "Joseph S. Myers" , Benjamin Herrenschmidt Mime-Version: 1.0 (Apple Message framework v1078) Cc: "linuxppc-dev@ozlabs.org list" , libc-alpha@sourceware.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Guys, We seem to be using the PPC_FEATURE_64 HWCAP to imply fsqrt support. = However, the new e5500 core from Freescale that is 64-bit does NOT = implement fsqrt so this assumption is wrong: sysdeps/powerpc/fpu/math_private.h # if __WORDSIZE =3D=3D 64 || defined _ARCH_PWR4 # define __CPU_HAS_FSQRT 1 # else # define __CPU_HAS_FSQRT ((GLRO(dl_hwcap) & PPC_FEATURE_64) !=3D 0) # endif Open to suggestions on how to fix this as its a kernel/libc interaction = point. - k=