Re: [PATCH net-next v2 9/9] ibmvnic: Do not replenish RX buffers after every polling loop

[Thomas Falcon <tlfalcon@linux.ibm.com>]

On 11/19/20 2:38 PM, ljp wrote:
> On 2020-11-19 14:26, Thomas Falcon wrote:
>> On 11/19/20 3:43 AM, ljp wrote:
>>
>>> On 2020-11-18 19:12, Thomas Falcon wrote:
>>>
>>>> From: "Dwip N. Banerjee" <dnbanerg@us.ibm.com>
>>>>
>>>> Reduce the amount of time spent replenishing RX buffers by
>>>> only doing so once available buffers has fallen under a certain
>>>> threshold, in this case half of the total number of buffers, or
>>>> if the polling loop exits before the packets processed is less
>>>> than its budget.
>>>>
>>>> Signed-off-by: Dwip N. Banerjee <dnbanerg@us.ibm.com>
>>>> ---
>>>> drivers/net/ethernet/ibm/ibmvnic.c | 5 ++++-
>>>> 1 file changed, 4 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/drivers/net/ethernet/ibm/ibmvnic.c
>>>> b/drivers/net/ethernet/ibm/ibmvnic.c
>>>> index 96df6d8fa277..9fe43ab0496d 100644
>>>> --- a/drivers/net/ethernet/ibm/ibmvnic.c
>>>> +++ b/drivers/net/ethernet/ibm/ibmvnic.c
>>>> @@ -2537,7 +2537,10 @@ static int ibmvnic_poll(struct napi_struct
>>>>
>>>> *napi, int budget)
>>>> frames_processed++;
>>>> }
>>>>
>>>> - if (adapter->state != VNIC_CLOSING)
>>>> + if (adapter->state != VNIC_CLOSING &&
>>>> + ((atomic_read(&adapter->rx_pool[scrq_num].available) <
>>>> + adapter->req_rx_add_entries_per_subcrq / 2) ||
>>>> + frames_processed < budget))
>>>
>>> 1/2 seems a simple and good algorithm.
>>> Explaining why "frames_process < budget" is necessary in the commit
>>> message
>>> or source code also helps.
>>
>> Hello, Lijun. The patch author, Dwip Banerjee, suggested the modified
>> commit message below:
>>
>> Reduce the amount of time spent replenishing RX buffers by
>>  only doing so once available buffers has fallen under a certain
>>  threshold, in this case half of the total number of buffers, or
>>  if the polling loop exits before the packets processed is less
>>  than its budget. Non-exhaustion of NAPI budget implies lower
>>  incoming packet pressure, allowing the leeway to refill the buffers
>>  in preparation for any impending burst.
>
> It looks good to me.
>
>>
>> Would such an update require a v3?
>
> I assume you ask Jakub, right?
>
>
Yes. There was an issue with my mail client in my earlier response, so I 
am posting Dwip's modified commit message again below.

Reduce the amount of time spent replenishing RX buffers by only doing so 
once available buffers has fallen under a certain threshold, in this 
case half of the total number of buffers, or if the polling loop exits 
before the packets processed is less than its budget. Non-exhaustion of 
NAPI budget implies lower incoming packet pressure, allowing the leeway 
to refill the buffers in preparation for any impending burst.

>>>> replenish_rx_pool(adapter, &adapter->rx_pool[scrq_num]);
>>>> if (frames_processed < budget) {
>>>> if (napi_complete_done(napi, frames_processed)) {