2021-11-23 17:24:22

by Alexander Lobakin

Subject: [PATCH net-next 2/9] e1000: switch to napi_build_skb()

napi_build_skb() reuses per-cpu NAPI skbuff_head cache in order
to save some cycles on freeing/allocating skbuff_heads on every
new Rx or completed Tx element.
e1000 driver runs Tx completion polling cycle right before the Rx
one. Now that e1000 uses napi_consume_skb() to put skbuff_heads of
completed entries into the cache, it will never empty and always
warm at that moment. Switch to the napi_build_skb() to relax mm
pressure on heavy Rx and increase throughput.

Signed-off-by: Alexander Lobakin <[email protected]>
Reviewed-by: Michal Swiatkowski <[email protected]>
---
drivers/net/ethernet/intel/e1000/e1000_main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/e1000/e1000_main.c b/drivers/net/ethernet/intel/e1000/e1000_main.c
index 975a145d48ef..3f5feb55cfba 100644
--- a/drivers/net/ethernet/intel/e1000/e1000_main.c
+++ b/drivers/net/ethernet/intel/e1000/e1000_main.c
@@ -4384,7 +4384,7 @@ static bool e1000_clean_rx_irq(struct e1000_adapter *adapter,
if (!skb) {
unsigned int frag_len = e1000_frag_len(adapter);

- skb = build_skb(data - E1000_HEADROOM, frag_len);
+ skb = napi_build_skb(data - E1000_HEADROOM, frag_len);
if (!skb) {
adapter->alloc_rx_buff_failed++;
break;
--
2.33.1

2021-12-28 00:07:32

by Brelinski, Tony

Subject: RE: [Intel-wired-lan] [PATCH net-next 2/9] e1000: switch to napi_build_skb()

> -----Original Message-----
> From: Intel-wired-lan <[email protected]> On Behalf Of
> Alexander Lobakin
> Sent: Tuesday, November 23, 2021 9:19 AM
> To: [email protected]
> Cc: [email protected]; [email protected]; Jakub Kicinski
> <[email protected]>; David S. Miller <[email protected]>
> Subject: [Intel-wired-lan] [PATCH net-next 2/9] e1000: switch to
> napi_build_skb()
>
> napi_build_skb() reuses per-cpu NAPI skbuff_head cache in order to save
> some cycles on freeing/allocating skbuff_heads on every new Rx or
> completed Tx element.
> e1000 driver runs Tx completion polling cycle right before the Rx one. Now
> that e1000 uses napi_consume_skb() to put skbuff_heads of completed
> entries into the cache, it will never empty and always warm at that moment.
> Switch to the napi_build_skb() to relax mm pressure on heavy Rx and
> increase throughput.
>
> Signed-off-by: Alexander Lobakin <[email protected]>
> Reviewed-by: Michal Swiatkowski <[email protected]>
> ---
> drivers/net/ethernet/intel/e1000/e1000_main.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)

Tested-by: Tony Brelinski <[email protected]>