From: =?iso-8859-2?Q?Horia_Geant=E3?= Subject: Re: [PATCH v4 1/5] powerpc: io.h: move iomap.h include so that it can use readq/writeq defs Date: Thu, 20 Jul 2017 12:36:07 +0000 Message-ID: References: <20170718165528.7383-1-logang@deltatee.com> <20170718165528.7383-2-logang@deltatee.com> <87iniphsug.fsf@concordia.ellerman.id.au> <15b6ea7d-5817-c0cd-aee4-c8a52d8808f9@deltatee.com> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-2" Content-Transfer-Encoding: quoted-printable Cc: Arnd Bergmann , Greg Kroah-Hartman , Stephen Bates , "Benjamin Herrenschmidt" , Paul Mackerras , Nicholas Piggin , Suresh Warrier , Oliver O'Halloran To: Logan Gunthorpe , Michael Ellerman , "linux-kernel@vger.kernel.org" , "linux-arch@vger.kernel.org" , "linux-ntb@googlegroups.com" , "linux-crypto@vger.kernel.org" Return-path: Received: from mail-eopbgr00061.outbound.protection.outlook.com ([40.107.0.61]:11388 "EHLO EUR02-AM5-obe.outbound.protection.outlook.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S934643AbdGTMgL (ORCPT ); Thu, 20 Jul 2017 08:36:11 -0400 Content-Language: en-US Sender: linux-crypto-owner@vger.kernel.org List-ID: On 7/20/2017 1:27 PM, Horia Geant=E3 wrote:=0A= > On 7/19/2017 7:04 PM, Logan Gunthorpe wrote:=0A= >>=0A= >>=0A= >> On 18/07/17 11:57 PM, Michael Ellerman wrote:=0A= >>> Seems fair enough, have you tested it at all?=0A= >>=0A= >> It's only been compile tested and the kbuild robot has beat up on it a b= it.=0A= >>=0A= > Looks like the patch set does not compile on PPC (.config generated=0A= > using make corenet64_smp_defconfig):=0A= > =0A= > [...]=0A= > LD vmlinux.o=0A= > MODPOST vmlinux.o=0A= > drivers/crypto/caam/jr.o: In function `rd_reg64':=0A= > /home/b05471/repos/cryptodev-2.6/drivers/crypto/caam/regs.h:154:=0A= > undefined reference to `.ioread64be_lo_hi'=0A= > /home/b05471/repos/cryptodev-2.6/drivers/crypto/caam/regs.h:151:=0A= > undefined reference to `.ioread64_lo_hi'=0A= > /home/b05471/repos/cryptodev-2.6/drivers/crypto/caam/regs.h:151:=0A= > undefined reference to `.ioread64_lo_hi'=0A= > /home/b05471/repos/cryptodev-2.6/drivers/crypto/caam/regs.h:154:=0A= > undefined reference to `.ioread64be_lo_hi'=0A= > drivers/crypto/caam/jr.o: In function `wr_reg64':=0A= > /home/b05471/repos/cryptodev-2.6/drivers/crypto/caam/regs.h:141:=0A= > undefined reference to `.iowrite64_lo_hi'=0A= > /home/b05471/repos/cryptodev-2.6/drivers/crypto/caam/regs.h:141:=0A= > undefined reference to `.iowrite64_lo_hi'=0A= > /home/b05471/repos/cryptodev-2.6/drivers/crypto/caam/regs.h:144:=0A= > undefined reference to `.iowrite64be_lo_hi'=0A= > /home/b05471/repos/cryptodev-2.6/drivers/crypto/caam/regs.h:144:=0A= > undefined reference to `.iowrite64be_lo_hi'=0A= > make: *** [vmlinux] Error 1=0A= > =0A= include/asm-generic/iomap.h is included before=0A= include/linux/io-64-nonatomic-lo-hi.h:=0A= =0A= include/linux/io.h -> arch/powerpc/include/asm/io.h ->=0A= include/asm-generic/iomap.h=0A= =0A= Thus, the "extern" version of ioread64_lo_hi and friends will be used=0A= (and not the inline version from io-64-nonatomic-lo-hi.h).=0A= =0A= But for this kernel .config:=0A= -CONFIG_GENERIC_IOMAP=3Dn and=0A= -there is no implementation of io{read|write}64[be]{_lo_hi|_hi_lo} in=0A= arch/powerpc/kernel/iomap.c=0A= =0A= Regards,=0A= Horia=0A=