[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20180305.162040.1738604452489890352.davem@davemloft.net>
Date: Mon, 05 Mar 2018 16:20:40 -0500 (EST)
From: David Miller <davem@...emloft.net>
To: jeffrey.t.kirsher@...el.com
Cc: eric.dumazet@...il.com, pkerbrat@...ray.eu, netdev@...r.kernel.org,
nhorman@...hat.com, sassmann@...hat.com, jogreene@...hat.com,
mgligor@...ray.eu
Subject: Re: [net-next 4/4] e1000e: allocate ring descriptors with
dma_zalloc_coherent
From: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
Date: Mon, 05 Mar 2018 11:09:29 -0800
> On Mon, 2018-03-05 at 10:23 -0800, Eric Dumazet wrote:
>> On Mon, 2018-03-05 at 10:16 -0800, Jeff Kirsher wrote:
>> > From: Pierre-Yves Kerbrat <pkerbrat@...ray.eu>
>> >
>> > Descriptor rings were not initialized at zero when allocated
>> > When area contained garbage data, it caused skb_over_panic in
>> > e1000_clean_rx_irq (if data had E1000_RXD_STAT_DD bit set)
>> >
>> > This patch makes use of dma_zalloc_coherent to make sure the
>> > ring is memset at 0 to prevent the area from containing garbage.
>> >
>>
>> This looks like a net candidate, fixing a bug, with 0 chance adding a
>> regression IMO.
>
> I am fine with that. Dave, let me know if you want me to re-submit
> this change for net/stable.
Yes, please add this patch to the net-queue pull request you also
sent today.
Thanks.
Powered by blists - more mailing lists