[PATCH v2 1/3] dmaengine: add new dma API for max_segment_number

On Wed, Jun 8, 2011 at 12:10 AM, FUJITA Tomonori
[off-list ref] wrote:

On Tue, 7 Jun 2011 23:56:21 -0700
Dan Williams [off-list ref] wrote:

quoted

quoted

The dma parameter restriction could be due to software (HBA drivers,
or subsystem). The value should be whatever the dma device driver says
it is in such case?

I'm assuming that the dma driver is taking responsibility for setting
this correctly. ?How would this work otherwise... HBA driver or
subsystem queries the dmaengine device and then sets this parameter on
its behalf? ?In other words dmanengine *is* the subsystem, if I am
understanding your definition.

Oops, I meant that the subsystem is software layer above
dmaengine. For example, SCSI subsystem sets the limit of max number of
sglist entries. That is, it is possible that software layer above
dmaengine could set dma limit, which is smaller than the limit of
dmaengine?

Perhaps, but this sounds like the reverse of what happens today where
scsi device drivers with knowledge of their own hardware will tell the
midlayer/subsystem the restriction.  The change with regard to this
patch is that the scsi device driver (for example) will recognize that
the device it is driving will not be a bus master and will arrange to
allocate a dma channel from dmaengine.  When said scsi driver reports
the dma restrictions to the subsystem it will borrow the parameters
from the dma channel, not the scsi device.  So yes, I still think it
should be whatever the dma channel says.

Although, you've been doing scsi work longer than I, so maybe I'm
overlooking something...?

Are there any cases today where the subsystem imposes tighter
restrictions on the dma geometry than what the device reports?  Even
if that were the case it would be same situation that the scsi device
driver reports maximum parameters, but the subsystem opts for
something tighter.  Whether the maximal parameters come from the scsi
device or the dma channel is moot.