From mboxrd@z Thu Jan 1 00:00:00 1970 From: Etienne Lorrain Subject: Re: [RFC] ATA host-protected area (HPA) device mapper? Date: Fri, 9 Jun 2006 22:41:47 +0200 (CEST) Message-ID: <20060609204147.61760.qmail@web26912.mail.ukl.yahoo.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Received: from web26912.mail.ukl.yahoo.com ([217.146.177.79]:2927 "HELO web26912.mail.ukl.yahoo.com") by vger.kernel.org with SMTP id S965156AbWFIUls (ORCPT ); Fri, 9 Jun 2006 16:41:48 -0400 Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: linux-kernel@vger.kernel.org, linux-ide@vger.kernel.org --- XXXXXXXX wrote: > It tends to be preferred around here to implement hardware features s= o=20 > that hardware actually works, rather than to implement spec and blame= =20 > the hardware for failing to work to the spec. Well, the IDE hardware works, and conform to the ATA specs. Gujin is using those ATA command for quite a long time. I just have a problem with SATA chipset not being ATA1-7 compatible. > If, as you say, most BIOS=20 > don't obey the spec, then implementing a spec that bears little=20 > resemblance to actual behavior only helps the small number of=20 > spec-compliant implementations. The main problem for BIOS manufacturer is the keyboard mapping recognition to enter the password. And there is no real point in freezing the password system of a locked drive. Gujin try to solve most of those problems, but because it is located on the hard disk itself limit the choice - it would be so nice to have a big enough BIOS FLASH and put Gujin there. Second best choice is having two hard disks (compact flash as first HD with compact flash <-> IDE adapter). > > The complete ATA specification is describing a register interface= since ATA1, > > SATA chipsets should be software compatible, even if they add comm= ands/interfaces: > > http://www.sata-io.org/interopfaq.asp say: > > Are there any known interoperability issues with SATA? > > One of the primary requirements of the SATA 1.0 specification w= as to > > maintain backward compatibility with existing operating system = drivers > > to eliminate incompatibility issues. > >=20 >=20 > Again with the "The spec says it works this way, so I can do it this=20 > way." I realize that you can be perfectly, legally, safe implementin= g=20 > specified behavior, but it leaves you doing what you are doing here,=20 > which is blaming the hardware for not living up to the specification.= =20 Well the ATA specs are quite old, Gujin still work with the BIOS hard= disk so still act as a bootloader, but it can not protecting you against a 1 milisecond attack setting password. By the way the SATA specs are not free. > It is generally the choice in the linux kernel to make the hardware=20 > work, and ignore the specification where it is irrelevant to actual=20 > functionality. Read up on the SAS Transport Layer. Oh yes, Gujin has some strange code to make things work at the end, look for ATAPI stuff - I can and do ignore the specs to make the softw= are work on real hardware. Cheers, Etienne. __________________________________________________ Do You Yahoo!? En finir avec le spam? Yahoo! Mail vous offre la meilleure protection p= ossible contre les messages non sollicit=E9s=20 http://mail.yahoo.fr Yahoo! Mail=20