Re: [PATCH] phydev: Add sysctl variable for polling interval of phy
From: 송은봉
Date: Mon Mar 11 2013 - 22:22:04 EST
Hello.
>Hello,
>On 03/11/2013 12:44 AM, EUNBONG SONG wrote:
>This patch should be submitted according to the rules described in:
>https://www.kernel.org/doc/Documentation/SubmittingPatches
Sorry, i made a mistake. so i rewrite the patches here.
---
drivers/net/phy/phy.c | 4 +++-
drivers/net/phy/phy_device.c | 43 ++++++++++++++++++++++++++++++++++++++++++
include/linux/phy.h | 1 -
3 files changed, 46 insertions(+), 2 deletions(-)
diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c
index ef9ea92..11b773d 100644
--- a/drivers/net/phy/phy.c
+++ b/drivers/net/phy/phy.c
@@ -42,6 +42,8 @@
#include <asm/irq.h>
#include <asm/uaccess.h>
+extern unsigned int sysctl_phy_poll_interval;
+
/**
* phy_print_status - Convenience function to print out the current phy status
* @phydev: the phy_device struct
@@ -966,7 +968,7 @@ void phy_state_machine(struct work_struct *work)
if (err < 0)
phy_error(phydev);
- schedule_delayed_work(&phydev->state_queue, PHY_STATE_TIME * HZ);
+ schedule_delayed_work(&phydev->state_queue, msecs_to_jiffies(sysctl_phy_poll_interval));
}
static inline void mmd_phy_indirect(struct mii_bus *bus, int prtad, int devad,
diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
index 3657b4a..4ed5368 100644
--- a/drivers/net/phy/phy_device.c
+++ b/drivers/net/phy/phy_device.c
@@ -33,6 +33,7 @@
#include <linux/mii.h>
#include <linux/ethtool.h>
#include <linux/phy.h>
+#include <linux/sysctl.h>
#include <asm/io.h>
#include <asm/irq.h>
@@ -42,6 +43,45 @@ MODULE_DESCRIPTION("PHY library");
MODULE_AUTHOR("Andy Fleming");
MODULE_LICENSE("GPL");
+unsigned int sysctl_phy_poll_interval = 1000;
+static unsigned int min_phy_poll_interval = 1;
+static unsigned int max_phy_poll_interval = 10000;
+
+static struct ctl_table_header *phy_table_header;
+
+static ctl_table phy_table[] = {
+ {
+ .procname = "phy_poll_interval",
+ .data = &sysctl_phy_poll_interval,
+ .maxlen = sizeof(unsigned int),
+ .mode = 0644,
+ .proc_handler = &proc_dointvec_minmax,
+ .extra1 = &min_phy_poll_interval,
+ .extra2 = &max_phy_poll_interval,
+ },
+ { }
+};
+
+static ctl_table phy_dir_table[] = {
+ {
+ .procname = "phy",
+ .maxlen = 0,
+ .mode = 0555,
+ .child = phy_table,
+ },
+ {}
+};
+
+static ctl_table phy_root_table[] = {
+ {
+ .procname = "dev",
+ .maxlen = 0,
+ .mode = 0555,
+ .child = phy_dir_table,
+ },
+ {}
+};
+
void phy_device_free(struct phy_device *phydev)
{
put_device(&phydev->dev);
@@ -1134,6 +1174,8 @@ static int __init phy_init(void)
if (rc)
mdio_bus_exit();
+ phy_table_header = register_sysctl_table(phy_root_table);
+
return rc;
}
@@ -1141,6 +1183,7 @@ static void __exit phy_exit(void)
{
phy_driver_unregister(&genphy_driver);
mdio_bus_exit();
+ unregister_sysctl_table(phy_table_header);
}
subsys_initcall(phy_init);
diff --git a/include/linux/phy.h b/include/linux/phy.h
index 33999ad..0cbc1fe 100644
--- a/include/linux/phy.h
+++ b/include/linux/phy.h
@@ -68,7 +68,6 @@ typedef enum {
#define PHY_INIT_TIMEOUT 100000
-#define PHY_STATE_TIME 1
#define PHY_FORCE_TIMEOUT 10
#define PHY_AN_TIMEOUT 10
>Besides that, I do not think that a system-wide knob here is
>appropriate, you may rather introduce a new ethtool ioctl() to change
>the PHY device polling interval on a per-PHY device basis.
This patch follows current phy driver flow, "PHY_STATE_TIME" is system-wide variable for phy polling
interval. So i just replace "PHY_STATE_TIME" variable to sysctl variable.
And I think it's better as a system-wide variable, because we usually expect all phys has same reponse-time
to the change of state.
>Having said that don't your devices support a dedicated PHY interrupt
>line? This would definitively be the way to get better latency with
>respect to PHY events reported back to the host.
I usually do not interrupt because if phy state goes down, there are so many jobs for this.
For example, remove arp table entries for this port, remove all routing table entries for this port .
If we plug and unplug the physical port in very rapid speed, there could be so many interrupt events.
And this is not what we want.
>1 millisecond sounds like we are going to eat up a lot of CPU time
>polling PHY registers. Do you need that much reactivity?
I agree with you, 1 milliseconds is very harmful to cpu. But it's just a minimum value not a default value.
Actually, i'm not sure which minimum and maximum value is best. So i just take wide range.
Thanks
------- Original Message -------
Sender : Florian Fainelli<florian@xxxxxxxxxxx>
Date : 2013-03-11 19:45 (GMT+09:00)
Title : Re: Fwd: [PATCH] phydev: Add sysctl variable for polling interval of phy
Hello,
On 03/11/2013 12:44 AM, EUNBONG SONG wrote:
>
> From d55a22be52e5a768409aa0999d6636cdfc369676 Mon Sep 17 00:00:00 2001
> From: eunbonsong
> Date: Sun, 10 Mar 2013 04:57:39 -0700
> Subject: [PATCH] phydev: Add sysctl variable for polling interval of phy state
>
> This adds a dev.phy.phy_poll_interval sysctl variable. This value is represented in milliseconds.
> And phy_state_machine() is scheduled as this variable.
> I think HZ is enough for PC. But sometimes especially in network devices
> such as switches,routers, needs more granularity for detecting phy state change.
This patch should be submitted according to the rules described in:
https://www.kernel.org/doc/Documentation/SubmittingPatches
Besides that, I do not think that a system-wide knob here is
appropriate, you may rather introduce a new ethtool ioctl() to change
the PHY device polling interval on a per-PHY device basis.
Having said that don't your devices support a dedicated PHY interrupt
line? This would definitively be the way to get better latency with
respect to PHY events reported back to the host.
>
>
> ---
> drivers/net/phy/phy.c | 4 +++-
> drivers/net/phy/phy_device.c | 43 ++++++++++++++++++++++++++++++++++++++++++
> include/linux/phy.h | 1 -
> 3 files changed, 46 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c
> index ef9ea92..126a69f 100644
> --- a/drivers/net/phy/phy.c
> +++ b/drivers/net/phy/phy.c
> @@ -42,6 +42,8 @@
> #include
> #include
>
> +extern unsigned long sysctl_phy_poll_interval;
> +
> /**
> * phy_print_status - Convenience function to print out the current phy status
> * @phydev: the phy_device struct
> @@ -966,7 +968,7 @@ void phy_state_machine(struct work_struct *work)
> if (err < 0)
> phy_error(phydev);
>
> - schedule_delayed_work(&phydev->state_queue, PHY_STATE_TIME * HZ);
> + schedule_delayed_work(&phydev->state_queue, msecs_to_jiffies(sysctl_phy_poll_interval));
> }
>
> static inline void mmd_phy_indirect(struct mii_bus *bus, int prtad, int devad,
> diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
> index 3657b4a..c2697e2 100644
> --- a/drivers/net/phy/phy_device.c
> +++ b/drivers/net/phy/phy_device.c
> @@ -33,6 +33,7 @@
> #include
> #include
> #include
> +#include
>
> #include
> #include
> @@ -42,6 +43,45 @@ MODULE_DESCRIPTION("PHY library");
> MODULE_AUTHOR("Andy Fleming");
> MODULE_LICENSE("GPL");
>
> +unsigned long sysctl_phy_poll_interval = 1000;
> +static unsigned long min_phy_poll_interval = 1;
> +static unsigned long max_phy_poll_interval = 10000;
1 millisecond sounds like we are going to eat up a lot of CPU time
polling PHY registers. Do you need that much reactivity?
--
Florian翁{.nÇ+돴윯돪†+%듚lzwm낂b앸㎠咽r¸›zX㎉ ®w¥Š{ayº Ê뉅숇,j?f＂·hš뗠z¹ ®wⅱ¸ ◁¦j:+v돣ŠwèjØm¶Ÿÿ¾«묎çzZ+껠šŽ듶¢j"얎!¶iO뺞¬z·švØ^¶m§ÿ操 nÆ듺þY&—