>Gadi Oxman wrote:
>>
>> I will try to create a similar patch for the 2.0.x kernels soon, for
>> potential integration into the upcoming 2.0.34 release. I am wondering
>> about the default /dev/sg? translation behavior, though -- should we
>> default to "translated" or "untranslated" access in the 2.0.x kernels?
>>
>Gadi,
>I understand that after your new patch enabling/disabling translation
>is done per device.
>Why not add an ioctl to get/set the transform bits? Then by default
>enable the translation in all cases. Programs that are aware that
>this is really an ATAPI device can use the ioctl to disable the
>translation (and restore previous settings when they exit).
>(for a program that accesses /dev/sgX it is much easier to use
>the ioctl). This ioctl can also be used by programs to distinguish
>between ide-scsi and 'real' scsi.
>Off topic: is there any plan (or work-in-progress) to have the
>similar of ide-scsi for paride? What else is needed to support
>parallel port cd-writers (like the HP7100e).
Solaris x86 supports a SCSI-parallel interface. Unfortunately, Sun
Violates it's own SCSA (Sun Common SCSI Architecture) specification
so even ATAPI CD-R does not work for now on Solaris x86. It could have
worked since 1991 on Solaris 2.1 x86. For nowm Linux is the only OS that
has a working SCSI<->ATAPI interworking that allows SCSI generic access.
But even the FreeBSD hackers seem to begin thinking on this problem
because of cdrecord.
Joerg
http://www.fokus.gmd.de/usr/schilling ftp://ftp.fokus.gmd.de/pub/unix