Re: [PATCH RFC V2 2/2] net: Optimize snmp stat aggregation by walking all the percpu data at once

[Joe Perches <joe@perches.com>]

On Fri, 2015-08-28 at 14:14 -0700, Eric Dumazet wrote:
> On Fri, 2015-08-28 at 14:09 -0700, Joe Perches wrote:
> > On Fri, 2015-08-28 at 13:55 -0700, Eric Dumazet wrote:
> > > On Fri, 2015-08-28 at 13:53 -0700, Joe Perches wrote:
> > > > It's 288 bytes on stack, maybe a kzalloc would be clearer too.
> > > 
> > > Could you read patch history and check why this has been rejected ?
> > 
> > I don't see a rejection, just that the initial
> > submission didn't check the allocation or add
> > an allocation buffer via kcalloc/kzalloc to the
> > inet6_fill_ifla6_attrs caller and change the
> > snmp6_fill_stats arguments.
> > 
> > It could also eliminate the put_unaligned calls.
> 
> Not really. You do not properly read this code.

Always a possibility, but I don't think so.

> put_unaligned is happening on a space allocated from rtnetlink skb, not
> the temp space needed to perform the per cpu folding.

That's why I suggested changing the snmp_fill_stats arguments.

If the naturally aligned allocated u64 array is used and then
copied as a block to the rtnetlink skb, I believe there's no
alignment issue that would require put_unaligned.

Do I still miss something?

> https://lkml.org/lkml/2015/8/25/476

I read that.