[next,S80-V3,07/11] i40e/i40evf: bundle more descriptors when allocating buffers

Message ID 20170907120556.45699-7-alice.michael@intel.com
State Under Review
Delegated to: Jeff Kirsher
Headers show
Series
  • [next,S80-V3,01/11] i40e: use the safe hash table iterator when deleting mac filters
Related show

Commit Message

Alice Michael Sept. 7, 2017, 12:05 p.m.
From: Jacob Keller <jacob.e.keller@intel.com>

Double the number of descriptors we'll bundle into one tail bump when
receiving. Empirical testing has shown that we reduce CPU utilization
and don't appear to reduce throughput or packet rate. 32 seems to be the
sweet spot, as it's half the default polling budget, so we'd essentially
reduce from 4 tail writes when polling down to 2. Increasing this up to
64 appears to have negative impacts as it may become possible that we
don't bump the tail each time we get polled, which could cause a long
delay between returning descriptors to the hardware.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
---
 drivers/net/ethernet/intel/i40e/i40e_txrx.h   | 2 +-
 drivers/net/ethernet/intel/i40evf/i40e_txrx.h | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

Comments

Bowers, AndrewX Sept. 12, 2017, 9:41 p.m. | #1
> -----Original Message-----
> From: Intel-wired-lan [mailto:intel-wired-lan-bounces@osuosl.org] On
> Behalf Of Alice Michael
> Sent: Thursday, September 7, 2017 5:06 AM
> To: Michael, Alice <alice.michael@intel.com>; intel-wired-
> lan@lists.osuosl.org
> Subject: [Intel-wired-lan] [next PATCH S80-V3 07/11] i40e/i40evf: bundle
> more descriptors when allocating buffers
> 
> From: Jacob Keller <jacob.e.keller@intel.com>
> 
> Double the number of descriptors we'll bundle into one tail bump when
> receiving. Empirical testing has shown that we reduce CPU utilization and
> don't appear to reduce throughput or packet rate. 32 seems to be the sweet
> spot, as it's half the default polling budget, so we'd essentially reduce from 4
> tail writes when polling down to 2. Increasing this up to
> 64 appears to have negative impacts as it may become possible that we don't
> bump the tail each time we get polled, which could cause a long delay
> between returning descriptors to the hardware.
> 
> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
> ---
>  drivers/net/ethernet/intel/i40e/i40e_txrx.h   | 2 +-
>  drivers/net/ethernet/intel/i40evf/i40e_txrx.h | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)

Tested-by: Andrew Bowers <andrewx.bowers@intel.com>

Patch

diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.h b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
index c3156aa..ff57ae4 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
@@ -208,7 +208,7 @@  static inline bool i40e_test_staterr(union i40e_rx_desc *rx_desc,
 }
 
 /* How many Rx Buffers do we bundle into one write to the hardware ? */
-#define I40E_RX_BUFFER_WRITE	16	/* Must be power of 2 */
+#define I40E_RX_BUFFER_WRITE	32	/* Must be power of 2 */
 #define I40E_RX_INCREMENT(r, i) \
 	do {					\
 		(i)++;				\
diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
index 8f9830d..8d26c85 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.h
@@ -191,7 +191,7 @@  static inline bool i40e_test_staterr(union i40e_rx_desc *rx_desc,
 }
 
 /* How many Rx Buffers do we bundle into one write to the hardware ? */
-#define I40E_RX_BUFFER_WRITE	16	/* Must be power of 2 */
+#define I40E_RX_BUFFER_WRITE	32	/* Must be power of 2 */
 #define I40E_RX_INCREMENT(r, i) \
 	do {					\
 		(i)++;				\