Message-ID: <3F105B9A.7070803@pobox.com>
Date: Sat, 12 Jul 2003 15:03:54 -0400
From: Jeff Garzik <jgarzik@pobox.com>
Organization: none
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20021213 Debian/1.2.1-2.bunk
MIME-Version: 1.0
To: Jamie Lokier <jamie@shareable.org>
CC: Andrew Morton <akpm@osdl.org>, davej@codemonkey.org.uk,
       linux-kernel@vger.kernel.org
Subject: Re: 2.5 'what to expect'
References: <20030711140219.GB16433@suse.de> <20030712152406.GA9521@mail.jlokier.co.uk> <3F103018.6020008@pobox.com> <20030712112722.55f80b60.akpm@osdl.org> <20030712183929.GA10450@mail.jlokier.co.uk>
In-Reply-To: <20030712183929.GA10450@mail.jlokier.co.uk>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 953
Lines: 34

Jamie Lokier wrote:
> Andrew Morton wrote:
> 
>>>One problem is O_DIRECT should return an error on open(2) or fcntl(2), 
>>> not write(2).
>>
>>That is the 2.5 behaviour.
> 
> 
> What do you mean?
> 
> The problem with db4 is that operations on O_DIRECT handles now return
> EINVAL if the address isn't suitable aligned, and db4 is not expecting
> that - it aborts.  That was true for 2.5.74, at least.


These are separate db4+kernel problems.

The first is that there needs to be a reliable way to know if O_DIRECT 
writes are going to succeed or not.  The 2.4 kernel fails on write(2) 
but not open(2) or fcntl(2).

The second is that db4 doesn't appear to know about the alignment 
requirement.

	Jeff


-
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/