Re: [Libguestfs] [PATCH 1/3] New API: mdadm-create for creating MD devices.
On 11/16/2011 06:07 AM, Richard W.M. Jones wrote:
On Mon, Nov 14, 2011 at 03:29:45PM +0800, Wanlong Gao wrote:
> On 11/11/2011 08:58 PM, Richard W.M. Jones wrote:
>
>> From: "Richard W.M. Jones" <rjones(a)redhat.com>
>
>
> Reviewed-by: Wanlong Gao <gaowanlong(a)cn.fujitsu.com>
Thanks for looking at these patches. If you (or Fujitsu) have any
other needs for libguestfs, I'd be more than happy to discuss them.
Yeah, Thanks a lot Richard.
[...]
>> + if ((chunk & 1023) != 0) {
>> + reply_with_error ("chunk size must be a multiple of 1024
bytes");
>
> I'd like to make the 1024 to a macro, what do you thinks so?
I wasn't sure if you meant (a) remove the use of the constant here and
replace it with a macro, or (b) allow users to specify other chunk
sizes (eg. less than 1024 bytes, or not a multiple of 1024 bytes).
As background: The reason we do this check is because the 'mdadm
--chunk' option only allows the size to be specified in K. We like to
use bytes in the libguestfs API [1]. We can only let users specify
multiples of 1024 here. It may be in future that mdadm would allow
chunk sizes which are not a multiple of 1024 bytes, but it doesn't
right now.
(a) is possible, but (b) is not possible right now.
I meant (a) just for the code readable and the future maintenance?
-Wanlong Gao
Rich.
[1] That is, we like to use bytes in new APIs, but we weren't very
consistent on this point in the early days.