Re: New TRIM/UNMAP tree published (2009-05-02)

[Jeff Garzik]

Matthew Wilcox wrote:
> > (3) when creating a discard request, use block helpers w/ queue-specific  
> > knowledge to create either
> > 	(a) one request, REQ_TYPE_FS, with discard flag or
> > 	(b) two requests, REQ_TYPE_FS followed by REQ_TYPE_DISCARD
>
> I'm not sure we need option 3b.


Well -- it is a hard requirement to map 1:1 struct request with the 
underlying hardware device's command set.

If a device command set lacks a READ-and-DISCARD operation, then you 
_must_ create two struct request.  Otherwise you break block layer 
tagging and other 1:1-based assumptions that exist today.	

Thus, given that all of ATA|SCSI|NVMHCI have a DISCARD operation 
distinct from other commands...

option 3b is the overwhelming common case.

	Jeff



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/