[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 9 Sep 2016 10:11:58 -0600
From: Jens Axboe <axboe@...com>
To: Joe Perches <joe@...ches.com>, Josef Bacik <jbacik@...com>,
<linux-block@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
<kernel-team@...com>, <mpa@...gutronix.de>,
<nbd-general@...ts.sourceforge.net>
Subject: Re: [PATCH 3/5] nbd: use flags instead of bool
On 09/09/2016 10:04 AM, Joe Perches wrote:
> On Fri, 2016-09-09 at 07:55 -0600, Jens Axboe wrote:
>> On 09/08/2016 07:20 PM, Joe Perches wrote:
>>> On Thu, 2016-09-08 at 17:12 -0400, Josef Bacik wrote:
>>>> In preparation for some future changes, change a few of the state bools over to
>>>> normal bits to set/clear properly.
>>> []
>>>> diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c
>>> []
>>>> @@ -41,8 +41,12 @@
>>>>
>>>> #include <linux/nbd.h>
>>>>
>>>> +#define NBD_TIMEDOUT 0
>>>> +#define NBD_DISCONNECT_REQUESTED 1
>>>> +
>>>> struct nbd_device {
>>>> u32 flags;
>>> + unsigned long runtime_flags;
>>> Better to use DECLARE_BITMAP
>
>> It's a few flags, we know it fits in a long. There's no point to using
>> anything but that, and set/test/clear_bit().
>
> It lets the reader know how it's used.
The variable is called 'runtime_flags' - if that doesn't already tell
the reader how it's used, then I'd suggest the reader go read something
else.
I'm all for using established APIs where it makes sense. Declaring a
bitmap for a few fields isn't that.
--
Jens Axboe
Powered by blists - more mailing lists