Re: [Patch] SCSI I/O statistics

From: Mike Christie
Date: Tue Sep 26 2006 - 17:46:39 EST


Matthew Wilcox wrote:
> On Tue, Sep 26, 2006 at 10:39:30AM -0700, Andrew Morton wrote:
>> - Should it have been done at the block layer rather than at the scsi layer?
>
> This was already mentioned when he sent it to linux-scsi a few days ago.
> There are scsi commands which bypass the block layer, such as SG_IO, and


I do not think any command completely bypasses the block layer now.
SG_IO, tape, sd and scanning insertion go through blk_execute_rq_nowait
for insertion. For completion they go through blk_complete_request +
end_that_request_first/last in the normal path and blk_complete_request
+ blk_requeue_request in the retry path.


> block layer stats can make an extremely busy disc look not busy.
> -
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html

-
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/