mbox series

[v2,0/2] net: if_arp: use define instead of hard-coded value

Message ID 20181009142724.2213012-1-Haakon.Bugge@oracle.com
Headers show
Series net: if_arp: use define instead of hard-coded value | expand

Message

Haakon Bugge Oct. 9, 2018, 2:27 p.m. UTC
Struct arpreq contains the name of the device. All other places in the
kernel, the define IFNAMSIZ is used to designate its size. But in
if_arp.h, a literal constant is used.

Fixed by explicitly including uapi/linu/if.h, where IFNAMSIZ is defined.

The series also fixes some incorrect indents.


v1 -> v2:
   * Include uapi/linux/if.h from if_arp.h
   * Added Stephen's t-b

Håkon Bugge (2):
  net: if_arp: Fix incorrect indents
  net: if_arp: use define instead of hard-coded value

 include/uapi/linux/if_arp.h | 19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)

--
2.14.3



From a36b99413ea5beca1539f4faf065f33c7c19e6a5 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?H=C3=A5kon=20Bugge?= <Haakon.Bugge@oracle.com>
Date: Fri, 21 Sep 2018 11:28:36 +0200
Subject: [PATCH 0/2]
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Struct arpreq contains the name of the device. All other places in the
kernel, the define IFNAMSIZ is used to designate its size. But in
if_arp.h, a literal constant is used.

As it could be good reasons to use constants instead of the defines in
include files under uapi, it seems to be OK to use the define here,
without opening a can of worms in user-land.

This because if_arp.h includes netdevice.h, which also uses
IFNAMSIZ. For the distros I have checked, this also holds true for the
use-land side.

The series also fixes some incorrect indents.

Håkon Bugge (2):
  net: if_arp: Fix incorrect indents
  net: if_arp: use define instead of hard-coded value

 include/uapi/linux/if_arp.h | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

--
2.14.3

Comments

Haakon Bugge Oct. 23, 2018, 12:30 p.m. UTC | #1
> On 9 Oct 2018, at 16:27, Håkon Bugge <haakon.bugge@oracle.com> wrote:
> 
> Struct arpreq contains the name of the device. All other places in the
> kernel, the define IFNAMSIZ is used to designate its size. But in
> if_arp.h, a literal constant is used.
> 
> Fixed by explicitly including uapi/linu/if.h, where IFNAMSIZ is defined.
> 
> The series also fixes some incorrect indents.
> 
> 
> v1 -> v2:
>   * Include uapi/linux/if.h from if_arp.h
>   * Added Stephen's t-b
> 
> Håkon Bugge (2):
>  net: if_arp: Fix incorrect indents
>  net: if_arp: use define instead of hard-coded value
> 
> include/uapi/linux/if_arp.h | 19 ++++++++++---------
> 1 file changed, 10 insertions(+), 9 deletions(-)
> ---
> 2.14.3

Just a friendly reminder.


Thxs, Håkon


> 
> 
> From a36b99413ea5beca1539f4faf065f33c7c19e6a5 Mon Sep 17 00:00:00 2001
> From: =?UTF-8?q?H=C3=A5kon=20Bugge?= <Haakon.Bugge@oracle.com>
> Date: Fri, 21 Sep 2018 11:28:36 +0200
> Subject: [PATCH 0/2]
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
> 
> Struct arpreq contains the name of the device. All other places in the
> kernel, the define IFNAMSIZ is used to designate its size. But in
> if_arp.h, a literal constant is used.
> 
> As it could be good reasons to use constants instead of the defines in
> include files under uapi, it seems to be OK to use the define here,
> without opening a can of worms in user-land.
> 
> This because if_arp.h includes netdevice.h, which also uses
> IFNAMSIZ. For the distros I have checked, this also holds true for the
> use-land side.
> 
> The series also fixes some incorrect indents.
> 
> Håkon Bugge (2):
>  net: if_arp: Fix incorrect indents
>  net: if_arp: use define instead of hard-coded value
> 
> include/uapi/linux/if_arp.h | 18 +++++++++---------
> 1 file changed, 9 insertions(+), 9 deletions(-)
> 
> --
> 2.14.3
David Miller Oct. 23, 2018, 4:11 p.m. UTC | #2
From: Håkon Bugge <haakon.bugge@oracle.com>
Date: Tue, 23 Oct 2018 14:30:57 +0200

> Just a friendly reminder.

Reminder for what?

Your patch was applied to the net-next tree right after it was posted,
what more do you want?

I gather that you have no idea what tree was appropriate for your change
and therefore where your should check to see if it was applied or not.
Haakon Bugge Oct. 23, 2018, 4:31 p.m. UTC | #3
> On 23 Oct 2018, at 18:11, David Miller <davem@davemloft.net> wrote:
> 
> From: Håkon Bugge <haakon.bugge@oracle.com>
> Date: Tue, 23 Oct 2018 14:30:57 +0200
> 
>> Just a friendly reminder.
> 
> Reminder for what?
> 
> Your patch was applied to the net-next tree right after it was posted,
> what more do you want?

Oh, my bad then. Was expecting review comments or an "Applied, thanks". Will check the respective tree next time. No need for the v2 then I assume.


Thxs, Håkon