kernel errors in messages about enclosure ...

[Scott Long]

Nate Lawson wrote:
> On Sat, 5 Apr 2003, Scott Long wrote:
> 
>>Marc G. Fournier wrote:
>>
>>>BTW, are there any plans to CAMify the aac driver?
>>
>>For your second question, the answer is 'no'.  The aac firmware talks in
>>terms of logical block commands.  'CAMifying' the driver would mean that
>>block requests from the filesystem would be translated into SCSI
>>commands by CAM and then converted back to block commands by the aac
>>driver.  Why bother?
> 
> 
> What is meant by "CAMify"?  aac(4) has a sim, an action routine and can
> parse XPT_SCSI_IO requests so isn't that "CAMified"?
> 
> -Nate

The block layer talks directly to the aac driver for normal array
access.  aac_cam.c provides 'passthrough' support that bypasses the
RAID layers in the firmware and allows CAM to talk directly to the
buses and devices on the back end.  It does not allow CAM to talk
to the logical arrays.

Scott