[8876] in linux-scsi channel archive
Re: Request splits
daemon@ATHENA.MIT.EDU (Chaitanya Tumuluri)
Tue May 23 17:38:34 2000
Message-Id: <200005232039.NAA79665@getafix.engr.sgi.com>
To: "Stephen C. Tweedie" <sct@redhat.com>
Cc: linux-scsi@vger.rutgers.edu
In-reply-to: Your message of "Tue, 23 May 2000 20:37:09 BST."
<20000523203709.A759@redhat.com>
Date: Tue, 23 May 2000 13:39:38 -0700
From: Chaitanya Tumuluri <chait@getafix.engr.sgi.com>
On Tue, 23 May 2000 20:37:09 BST, "Stephen C. Tweedie" <sct@redhat.com> wrote:
>Hi,
>
>On Thu, May 18, 2000 at 01:48:14PM -0700, Brian Pomerantz wrote:
>>
>> I've attached a plot of the performance curve for the Ciprico 7000
>> attached to the QLogic 2200. The dd I have been using aligns the
>> buffer to a page, which as far as I know should work with Stephen
>> Tweedie's raw I/O. Is there anything else that needs to be done to
>> use the raw device once it has been attached to a block device?
>
>Nope, there shouldn't be.
>
>To allow testing of larger blocksizes, I'll add a tunable sysctl
>limit to the 2.3 raw device code to support larger blocksizes. I'm
Is this tunable for the 2.2 series kernels? I guess it makes sense there.
For the 2.3 kernels, my patch removes any such limitation and sends
down the entire I/O request at whatever blocksize is requested.
>curious, though --- does anybody know what the typical blocksize is
>when applications are using raw I/O, as opposed to benchmark tests?
I've seen databases typically request in the range of 4k - 16M; the 16M
being _really_ _rare_ events. I'd be hardpressed for a typical value but
I think allowing for upto 1M request size should be good.
Stephen, I'll be sending out the patch against the 2.3.99.pre9-2 diffs.
at the rawio site shortly.
Thanks,
-Chait.
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
