Commit 35da1dfd authored by Tobias Waldekranz's avatar Tobias Waldekranz Committed by David S. Miller

net: dsa: mv88e6xxx: Improve performance of busy bit polling

Avoid a long delay when a busy bit is still set and has to be polled
again.

Measurements on a system with 2 Opals (6097F) and one Agate (6352)
show that even with this much tighter loop, we have about a 50% chance
of the bit being cleared on the first poll, all other accesses see the
bit being cleared on the second poll.

On a standard MDIO bus running MDC at 2.5MHz, a single access with 32
bits of preamble plus 32 bits of data takes 64*(1/2.5MHz) = 25.6us.

This means that mv88e6xxx_smi_direct_wait took 26us + CPU overhead in
the fast scenario, but 26us + 1500us + 26us + CPU overhead in the slow
case - bringing the average close to 1ms.

With this change in place, the slow case is closer to 2*26us + CPU
overhead, with the average well below 100us - a 10x improvement.

This translates to real-world winnings. On a 3-chip 20-port system,
the modprobe time drops by 88%:

Before:

root@coronet:~# time modprobe mv88e6xxx
real    0m 15.99s
user    0m 0.00s
sys     0m 1.52s

After:

root@coronet:~# time modprobe mv88e6xxx
real    0m 2.21s
user    0m 0.00s
sys     0m 1.54s
Signed-off-by: default avatarTobias Waldekranz <tobias@waldekranz.com>
Reviewed-by: default avatarAndrew Lunn <andrew@lunn.ch>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent 0da8aa00
...@@ -86,12 +86,16 @@ int mv88e6xxx_write(struct mv88e6xxx_chip *chip, int addr, int reg, u16 val) ...@@ -86,12 +86,16 @@ int mv88e6xxx_write(struct mv88e6xxx_chip *chip, int addr, int reg, u16 val)
int mv88e6xxx_wait_mask(struct mv88e6xxx_chip *chip, int addr, int reg, int mv88e6xxx_wait_mask(struct mv88e6xxx_chip *chip, int addr, int reg,
u16 mask, u16 val) u16 mask, u16 val)
{ {
const unsigned long timeout = jiffies + msecs_to_jiffies(50);
u16 data; u16 data;
int err; int err;
int i; int i;
/* There's no bus specific operation to wait for a mask */ /* There's no bus specific operation to wait for a mask. Even
for (i = 0; i < 16; i++) { * if the initial poll takes longer than 50ms, always do at
* least one more attempt.
*/
for (i = 0; time_before(jiffies, timeout) || (i < 2); i++) {
err = mv88e6xxx_read(chip, addr, reg, &data); err = mv88e6xxx_read(chip, addr, reg, &data);
if (err) if (err)
return err; return err;
...@@ -99,7 +103,10 @@ int mv88e6xxx_wait_mask(struct mv88e6xxx_chip *chip, int addr, int reg, ...@@ -99,7 +103,10 @@ int mv88e6xxx_wait_mask(struct mv88e6xxx_chip *chip, int addr, int reg,
if ((data & mask) == val) if ((data & mask) == val)
return 0; return 0;
usleep_range(1000, 2000); if (i < 2)
cpu_relax();
else
usleep_range(1000, 2000);
} }
dev_err(chip->dev, "Timeout while waiting for switch\n"); dev_err(chip->dev, "Timeout while waiting for switch\n");
......
...@@ -55,11 +55,15 @@ static int mv88e6xxx_smi_direct_write(struct mv88e6xxx_chip *chip, ...@@ -55,11 +55,15 @@ static int mv88e6xxx_smi_direct_write(struct mv88e6xxx_chip *chip,
static int mv88e6xxx_smi_direct_wait(struct mv88e6xxx_chip *chip, static int mv88e6xxx_smi_direct_wait(struct mv88e6xxx_chip *chip,
int dev, int reg, int bit, int val) int dev, int reg, int bit, int val)
{ {
const unsigned long timeout = jiffies + msecs_to_jiffies(50);
u16 data; u16 data;
int err; int err;
int i; int i;
for (i = 0; i < 16; i++) { /* Even if the initial poll takes longer than 50ms, always do
* at least one more attempt.
*/
for (i = 0; time_before(jiffies, timeout) || (i < 2); i++) {
err = mv88e6xxx_smi_direct_read(chip, dev, reg, &data); err = mv88e6xxx_smi_direct_read(chip, dev, reg, &data);
if (err) if (err)
return err; return err;
...@@ -67,7 +71,10 @@ static int mv88e6xxx_smi_direct_wait(struct mv88e6xxx_chip *chip, ...@@ -67,7 +71,10 @@ static int mv88e6xxx_smi_direct_wait(struct mv88e6xxx_chip *chip,
if (!!(data & BIT(bit)) == !!val) if (!!(data & BIT(bit)) == !!val)
return 0; return 0;
usleep_range(1000, 2000); if (i < 2)
cpu_relax();
else
usleep_range(1000, 2000);
} }
return -ETIMEDOUT; return -ETIMEDOUT;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment