net: ethernet: fec: Replace interrupt driven MDIO with polled IO

Measurements of the MDIO bus have shown that driving the MDIO bus
using interrupts is slow. Back to back MDIO transactions take about
90us, with 25us spent performing the transaction, and the remainder of
the time the bus is idle.

Replacing the completion interrupt with polled IO results in back to
back transactions of 40us. The polling loop waiting for the hardware
to complete the transaction takes around 28us. Which suggests
interrupt handling has an overhead of 50us, and polled IO nearly
halves this overhead, and doubles the MDIO performance.

Suggested-by: Chris Heally <cphealy@gmail.com>
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Andrew Lunn
2020-04-20 00:04:00 +02:00
committed by David S. Miller
parent 9d3679fe0f
commit 29ae6bd1b0
2 changed files with 35 additions and 36 deletions

View File

@@ -376,8 +376,7 @@ struct bufdesc_ex {
#define FEC_ENET_TS_AVAIL ((uint)0x00010000)
#define FEC_ENET_TS_TIMER ((uint)0x00008000)
#define FEC_DEFAULT_IMASK (FEC_ENET_TXF | FEC_ENET_RXF | FEC_ENET_MII)
#define FEC_NAPI_IMASK FEC_ENET_MII
#define FEC_DEFAULT_IMASK (FEC_ENET_TXF | FEC_ENET_RXF)
#define FEC_RX_DISABLED_IMASK (FEC_DEFAULT_IMASK & (~FEC_ENET_RXF))
/* ENET interrupt coalescing macro define */
@@ -543,7 +542,6 @@ struct fec_enet_private {
int link;
int full_duplex;
int speed;
struct completion mdio_done;
int irq[FEC_IRQ_NUM];
bool bufdesc_ex;
int pause_flag;