From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:54127)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <ehabkost@redhat.com>) id 1bxYR0-00045q-Ua
	for qemu-devel@nongnu.org; Fri, 21 Oct 2016 07:56:28 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <ehabkost@redhat.com>) id 1bxYQw-00011M-W4
	for qemu-devel@nongnu.org; Fri, 21 Oct 2016 07:56:27 -0400
Received: from mx1.redhat.com ([209.132.183.28]:44098)
	by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32)
	(Exim 4.71) (envelope-from <ehabkost@redhat.com>) id 1bxYQw-00010P-RW
	for qemu-devel@nongnu.org; Fri, 21 Oct 2016 07:56:22 -0400
Date: Fri, 21 Oct 2016 09:56:20 -0200
From: Eduardo Habkost <ehabkost@redhat.com>
Message-ID: <20161021115620.GN5057@thinpad.lan.raisama.net>
References: <20161020061301.31372-1-haozhong.zhang@intel.com>
	<20161020143412.5ea6b564@nial.brq.redhat.com>
	<20161020131138.4gzxk5ekltkiqtq2@hz-desktop>
	<20161020154215.4b765284@nial.brq.redhat.com>
	<20161020135610.GY5057@thinpad.lan.raisama.net>
	<20161020161521.6e20ce2a@nial.brq.redhat.com>
	<20161020144734.GA5057@thinpad.lan.raisama.net>
	<20161020173538.1ee9a556@nial.brq.redhat.com>
	<20161021072210.sxfwixky4idlmzni@hz-desktop>
	<20161021130700.7beff8d3@nial.brq.redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20161021130700.7beff8d3@nial.brq.redhat.com>
Subject: Re: [Qemu-devel] [PATCH] hostmem-file: add a property 'notrunc' to
 avoid data corruption
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel/>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Igor Mammedov <imammedo@redhat.com>
Cc: Haozhong Zhang <haozhong.zhang@intel.com>, kwolf@redhat.com, Xiao Guangrong <guangrong.xiao@linux.intel.com>, Peter Crosthwaite <crosthwaite.peter@gmail.com>, qemu-devel@nongnu.org, mreitz@redhat.com, Paolo Bonzini <pbonzini@redhat.com>, Richard Henderson <rth@twiddle.net>

On Fri, Oct 21, 2016 at 01:07:00PM +0200, Igor Mammedov wrote:
[...]
> How about following behavior:
> 
>  1) memory-backend-file,mem-path=/some_dir,size=2G
>     - uses truncate to extend temporary file created in "mem-path" to 'size'
>     for this case to work 'size' is mandatory
> 
>  2) memory-backend-file,mem-path=/existing_file,size=2G
>     - uses truncate to extend/shrink file "mem-path" to 'size'
>     for this case 'size' could be made optional,
>     if we take in account that backend could be used as persistent
>     storage then shrinking or extending "mem-path" would be corruption
>     as backend has no idea about internal layout if mapped file.
>     We can do something like this here:
> 
>     if (is_size_opt_provided and size_of(mem-path) != 0) {
>        error_out with "mem-path=foo size XXX doesn't match 'size=xxx' option"
>     } else if (is_size_opt_provided and size_of(mem-path) == 0) {
>        // may be we don't need this case and
>        // just fold this in above error case
>        truncate(mem-path) // extend/shrink
>     } else {
>       set_size_opt(size_of(mem-path))
>     }

Agreed.

-- 
Eduardo