[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4EF1C80B.4060701@monstr.eu>
Date: Wed, 21 Dec 2011 12:50:35 +0100
From: Michal Simek <monstr@...str.eu>
To: Eric Dumazet <eric.dumazet@...il.com>
CC: David Miller <davem@...emloft.net>,
John Williams <john.williams@...alogix.com>,
netdev@...r.kernel.org
Subject: Re: ICMP packets - ll_temac with Microblaze
Eric Dumazet wrote:
> Le mercredi 21 décembre 2011 à 12:10 +0100, Michal Simek a écrit :
>> Eric Dumazet wrote:
>>> Le mercredi 21 décembre 2011 à 11:45 +0100, Eric Dumazet a écrit :
>>>> Le mercredi 21 décembre 2011 à 11:41 +0100, Michal Simek a écrit :
>>>>
>>>>> yes, page size is 4k.
>>>>>
>>>>> Here is the log for allocation - origin size for MTU 9000 is 0x2420(9248) and size after
>>>>> overhead is 0x3f40(16192)
>>>>>
>>>>> kmalloc_node_track_caller size 00002420, SKB_WITH_OVERHEAD(ksize(data)) 00003f40
>>>>> skb->head c7870000
>>>>> skb->data c7870000
>>>>> skb->tail c7870000
>>>>> skb->truesize 000040c0
>>>>> skb->end c7873f40
>>>>>
>>>> So basically truesize is bigger than previous kernel, and you hit a
>>>> socket limit. (RCVBUF ?)
>>>>
>>>> truesize is more correct, so we uncover prior bugs.
>>>>
>>>> We should allow at least one packet, regardless of its size :(
>>>>
>>>>
>>> I would try the following patch :
>>>
>>> diff --git a/include/net/sock.h b/include/net/sock.h
>>> index bf6b9fd..21bb3b5 100644
>>> --- a/include/net/sock.h
>>> +++ b/include/net/sock.h
>>> @@ -662,12 +662,14 @@ static inline void __sk_add_backlog(struct sock *sk, struct sk_buff *skb)
>>>
>>> /*
>>> * Take into account size of receive queue and backlog queue
>>> + * Do not take into account this skb truesize,
>>> + * to allow even a single big packet to come.
>>> */
>>> static inline bool sk_rcvqueues_full(const struct sock *sk, const struct sk_buff *skb)
>>> {
>>> unsigned int qsize = sk->sk_backlog.len + atomic_read(&sk->sk_rmem_alloc);
>>>
>>> - return qsize + skb->truesize > sk->sk_rcvbuf;
>>> + return qsize > sk->sk_rcvbuf;
>>> }
>>>
>>> /* The per-socket spinlock must be held here. */
>>>
>> It doesn't fix the problem. Do you want to add some printk to code?
>>
>> Michal
>>
>>
>>
>
> try to change SOCK_MIN_RCVBUF
>
> include/net/sock.h
>
> #define SOCK_MIN_RCVBUF (2048 + sizeof(struct sk_buff))
>
> ->
>
> #define SOCK_MIN_RCVBUF (16384 + sizeof(struct sk_buff))
>
Doesn't work. :-(
Michal
--
Michal Simek, Ing. (M.Eng)
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel 2.6 Microblaze Linux - http://www.monstr.eu/fdt/
Microblaze U-BOOT custodian
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists