Re: [PATCH 4/4] UDP memory accounting and limitation(take 5): memory limitation

[Hideo AOKI <haoki@redhat.com>]

Herbert Xu wrote:
> On Fri, Oct 12, 2007 at 09:11:27PM +0900, Satoshi OSHIMA wrote:
>> +static inline int __ip_check_max_skb_pages(struct sock *sk, int size)
>> +{
>> +	switch(sk->sk_protocol) {
>> +	case IPPROTO_UDP:
>> +		if (sk->sk_prot->sysctl_mem[0] > UDP_MIN_SKB_PAGES)
>> +			if (atomic_read(sk->sk_prot->memory_allocated)+size
>> +			    >= sk->sk_prot->sysctl_mem[0])
>> +				return -ENOBUFS;
>> +		/* Fall through */	
>> +	default:
>> +		break;
>> +	}
>> +	return 0;
>> +}
> 
> Since you're not doing a per-user limit, you must give each
> socket a minimum even when the total exceeds your threshold.
> 
> Otherwise any local user can trivially DoS the whole system.
> 
>> +UDP variables:
>> +
>> +udp_mem - INTERGER
>> +	Number of pages allowed for queueing by all UDP sockets.
>> +	Minimal value is 4096. If 4096 is set, UDP memory will not
>> +	be limited.
>> +	Default: 4096
> 
> This seems to be a rather confusing setup.  We don't set a
> minimum for TCP so is this really necessary?

Hello Herbert,

Thank you so much for your comments.

I'm going to fix them in next take patch set.

Regards,
Hideo

--
Hitachi Computer Products (America) Inc.