Re: [PATCH] net: Prevent multiple NAPI instances co-existing in the list

[Dennis Chen <kernel.org.gnu@gmail.com>]

On Fri, Jan 9, 2015 at 10:27 AM, Herbert Xu <herbert@gondor.apana.org.au> wrote:
> On Fri, Jan 09, 2015 at 10:24:18AM +0800, Dennis Chen wrote:
>>
>> Hi Herbert, please see this code piece in napi_poll:
>>
>>      /* Some drivers may have called napi_schedule
>>      * prior to exhausting their budget.
>>      */
>>     if (unlikely(!list_empty(&n->poll_list))) {
>>         pr_warn_once("%s: Budget exhausted after napi rescheduled\n",
>>                  n->dev ? n->dev->name : "backlog");
>>         goto out_unlock;
>>     }
>>
>> Here "Some drivers" may have called  napi_schedule to make
>> n->poll_list is not empty, does that mean "Some drivers" will clear
>> NAPI_STATE_SCHED bit, otherwise the napi_schedule() will do nothing,
>> does that make sense for you question? ;-)
>
> No it tells me that you don't understand the problem at all.
> Those drivers will end up resetting the NAPI_STATE_SCHED bit
> after clearing it.
>
> Cheers,
> --
> Email: Herbert Xu <herbert@gondor.apana.org.au>
> Home Page: http://gondor.apana.org.au/~herbert/
> PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Thanks, would you pls give me an example of those drivers? I'll study
it further...

-- 
Den