From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:47894)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <ehabkost@redhat.com>) id 1c1HBx-0008VS-HO
	for qemu-devel@nongnu.org; Mon, 31 Oct 2016 14:20:18 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <ehabkost@redhat.com>) id 1c1HBt-0007W0-FW
	for qemu-devel@nongnu.org; Mon, 31 Oct 2016 14:20:17 -0400
Received: from mx1.redhat.com ([209.132.183.28]:46660)
	by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32)
	(Exim 4.71) (envelope-from <ehabkost@redhat.com>) id 1c1HBt-0007Vp-B3
	for qemu-devel@nongnu.org; Mon, 31 Oct 2016 14:20:13 -0400
Date: Mon, 31 Oct 2016 16:20:10 -0200
From: Eduardo Habkost <ehabkost@redhat.com>
Message-ID: <20161031182010.GE2919@thinpad.lan.raisama.net>
References: <1477924663-30950-1-git-send-email-pbonzini@redhat.com>
	<1477924663-30950-9-git-send-email-pbonzini@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1477924663-30950-9-git-send-email-pbonzini@redhat.com>
Subject: Re: [Qemu-devel] [PULL 08/27] hostmem-file: make option 'size'
 optional
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: Paolo Bonzini <pbonzini@redhat.com>
Cc: qemu-devel@nongnu.org, Haozhong Zhang <haozhong.zhang@intel.com>

On Mon, Oct 31, 2016 at 03:37:24PM +0100, Paolo Bonzini wrote:
[...]
> @@ -1309,21 +1317,27 @@ static void *file_ram_alloc(RAMBlock *block,
>  
>      file_size = get_file_size(fd);
>  
> -    if (memory < block->page_size) {
> +    if (!mem_size && file_size > 0) {
> +        mem_size = file_size;
> +        memory_region_set_size(block->mr, mem_size);
> +    }
> +
> +    if (mem_size < block->page_size) {
>          error_setg(errp, "memory size 0x" RAM_ADDR_FMT " must be equal to "
>                     "or larger than page size 0x%zx",
> -                   memory, block->page_size);
> +                   mem_size, block->page_size);
>          goto error;
>      }
>  
> -    if (file_size > 0 && file_size < memory) {
> +    if (file_size > 0 && file_size < mem_size) {
>          error_setg(errp, "backing store %s size %"PRId64
>                     " does not match 'size' option %"PRIu64,
> -                   path, file_size, memory);
> +                   path, file_size, mem_size);
>          goto error;
>      }
>  
> -    memory = ROUND_UP(memory, block->page_size);
> +    mem_size = ROUND_UP(mem_size, block->page_size);
> +    *memory = mem_size;

I suggested not touching *memory unless it was zero, and setting
it to the memory region size, not the rounded-up size. Haozhong
said this was going to be changed.

This will have the side-effect of setting block->used_length and
block->max_length to the rounded up size in
qemu_ram_alloc_from_file() (instead of the original memory region
size). I don't know what could be the consequences of that.

This patch also skip HOST_PAGE_ALIGN-ing mem_size after getting
the file size, which would be different from the behavior when
size is specified explicitly. (And I also don't know the
consequences of that)

Considering that this pull request failed to build, I suggest
waiting for a new version from Haozhong.

-- 
Eduardo