18 Feb
2014
18 Feb
'14
8:41 p.m.
From: Antonio Quartulli antonio@meshcoding.com Date: Tue, 18 Feb 2014 07:44:53 +0100
On 17/02/14 22:13, David Miller wrote:
From: Antonio Quartulli antonio@meshcoding.com Date: Mon, 17 Feb 2014 21:48:40 +0100
- atomic_set(&bat_priv->packet_size_max, min_mtu);
Please fix this.
The only operations performed on packet_size_max are 'set' and 'read'. This is not what one uses atomic_t's for.
The use of an atomic_t in this context is a NOP. You aren't getting any kind of synchronization at all.
True. Thanks for the suggestion. Unfortunately this is not the only "fake-atomic" variable we have.
So I pulled this, but I just want you to know that you should really start to bear down and minimize the amount of changes you are submitting for 'net' as we're starting to get deep into -rc territory.
Thanks.