From: "Chen, Kenneth W" <kenneth.w.chen@intel.com>
To: "'Benjamin LaHaise'" <bcrl@kvack.org>
Cc: "'Zach Brown'" <zach.brown@oracle.com>,
       "'Suparna Bhattacharya'" <suparna@in.ibm.com>,
       "Lahaise, Benjamin C" <benjamin.c.lahaise@intel.com>,
       <linux-kernel@vger.kernel.org>, "'linux-aio'" <linux-aio@kvack.org>
Subject: RE: [patch] clarify AIO_EVENTS_OFFSET constant
Date: Fri, 13 Oct 2006 13:38:01 -0700
Message-ID: <000501c6ef07$79c1c770$ff0da8c0@amr.corp.intel.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
Thread-Index: Acbu0x4q2TGEejUAQUCD+YCjSYY9ZAAM2hIQ
In-Reply-To: <20061013142257.GJ4141@kvack.org>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1247
Lines: 26

Benjamin LaHaise wrote on Friday, October 13, 2006 7:23 AM
> On Thu, Oct 12, 2006 at 05:00:24PM -0700, Chen, Kenneth W wrote:
> > A clean up patch: I think it is a lot easier to read AIO_EVENTS_OFFSET
> > as an offset because of aio_ring at the beginning of a head page, instead
> > of doing arithmetic of (event on 2nd page - event on 1st page).
> 
> Nak.  Your change fails if aio_ring is not an exact multiple of the 
> io_event size due to rounding errors, while the original code rounds 
> correctly.


Yeah, neither form is bullet proof, even the original one will break down
when io_event size is not in power of 2.  The good news is that no where
near in sight that these two data structures size will change.

To make it a real bullet proof, it should be something like the following:

#define AIO_EVENTS_OFFSET       ((sizeof(struct aio_ring) +             \
                                  sizeof(struct io_event) - 1 ) /       \
                                  sizeof(struct io_event))

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