Date: Thu, 8 Nov 2001 08:15:11 -0700 From: Scott Long <scott_long@btc.adaptec.com> To: Thomas Quinot <quinot@inf.enst.fr> Cc: Nick Hibma <n_hibma@qubesoft.com>, freebsd-scsi@freebsd.org, freebsd-current@freebsd.org Subject: Re: SCSI->IDE Message-ID: <20011108081511.A81770@hollin.btc.adaptec.com> In-Reply-To: <20011108104243.A30845@cuivre.fr.eu.org> References: <20011107155735.B30053@cuivre.fr.eu.org> <CEEKLNFIGKODPCCPEKLDMEOACDAA.n_hibma@qubesoft.com> <20011108104243.A30845@cuivre.fr.eu.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Nov 08, 2001 at 10:42:43AM +0100, Thomas Quinot wrote:
> Le 2001-11-07, Nick Hibma écrivait :
>
> > You are not doing any conversion of commands from SCSI to ATAPI. Did you
> > verify that all SCSI commands are valid ATAPI commands?
>
> Actually the current version of the patch does do some translation
> (MODE_{SENSE,SELECT}_6 are mapped to their _10 equivalent, because
> the _6 variants are not implemented by ATAPI/MMC devices), but a
> better solution might be to move these translations up to transport
> independant layers, and be handled through quirk entries, or some
> similar flagging mechanism. After all, ATAPI really is a /transport/
> mechanism for SCSI commands: it is up to the initiator of a request
> to know which command set to use for a given device.
Adding even more quirk entries is not the right way to go. One, or
both, of the following things should happen.
1. The SIM should inform the upper layers that 6 bytes commands are
not allowed.
2. If a 6 byte command is issued and fails, the failure code should be
analyzed and the command retried in its 10 byte version. If that
succeeds, then the upper layers should be flagged to only use 10 byte
commands.
Scott
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20011108081511.A81770>