[U-Boot] [PATCH] driver/net/fec: support fixed speed connection
If MAC is directly connected to another MAC (like a switch for example) we don't need to probe for a phy, autoneogation and so on. We simply have to setup speed.
Signed-off-by: Hannes Schmelzer oe5hpm@oevsv.at ---
doc/README.fec_mxc | 5 +++++ drivers/net/fec_mxc.c | 4 ++++ 2 files changed, 9 insertions(+)
diff --git a/doc/README.fec_mxc b/doc/README.fec_mxc index ed7e47d..9ca6ac2 100644 --- a/doc/README.fec_mxc +++ b/doc/README.fec_mxc @@ -27,6 +27,11 @@ CONFIG_FEC_MXC_PHYADDR Optional, selects the exact phy address that should be connected and function fecmxc_initialize will try to initialize it.
+CONFIG_FEC_FIXED_SPEED + Optional, selects a fixed speed on the MAC interface without asking some + phy. This is usefull if there is a direct MAC <-> MAC connection, for + example if the CPU is connected directly via the RGMII interface to a + ethernet-switch.
Reading the ethaddr from the SoC eFuses: if CONFIG_FEC_MXC is defined and the U-Boot environment does not contain the diff --git a/drivers/net/fec_mxc.c b/drivers/net/fec_mxc.c index 360f8e4..e871b3e 100644 --- a/drivers/net/fec_mxc.c +++ b/drivers/net/fec_mxc.c @@ -233,6 +233,7 @@ static int miiphy_restart_aneg(struct eth_device *dev) return ret; }
+#ifndef CONFIG_FEC_FIXED_SPEED static int miiphy_wait_aneg(struct eth_device *dev) { uint32_t start; @@ -260,6 +261,7 @@ static int miiphy_wait_aneg(struct eth_device *dev)
return 0; } +#endif /* CONFIG_FEC_FIXED_SPEED */ #endif
static int fec_rx_task_enable(struct fec_priv *fec) @@ -502,6 +504,8 @@ static int fec_open(struct eth_device *edev) } speed = fec->phydev->speed; } +#elif CONFIG_FEC_FIXED_SPEED + speed = CONFIG_FEC_FIXED_SPEED; #else miiphy_wait_aneg(edev); speed = miiphy_speed(edev->name, fec->phy_id);
On Wed, Jun 22, 2016 at 5:07 AM, Hannes Schmelzer oe5hpm@oevsv.at wrote:
If MAC is directly connected to another MAC (like a switch for example) we don't need to probe for a phy, autoneogation and so on. We simply have to setup speed.
Signed-off-by: Hannes Schmelzer oe5hpm@oevsv.at
You may want to add it to Kconfig, but that's up to Stefano.
Acked-by: Joe Hershberger joe.hershberger@ni.com
Hi Hannes,
On 22/06/2016 12:07, Hannes Schmelzer wrote:
If MAC is directly connected to another MAC (like a switch for example) we don't need to probe for a phy, autoneogation and so on. We simply have to setup speed.
Signed-off-by: Hannes Schmelzer oe5hpm@oevsv.at
doc/README.fec_mxc | 5 +++++ drivers/net/fec_mxc.c | 4 ++++ 2 files changed, 9 insertions(+)
diff --git a/doc/README.fec_mxc b/doc/README.fec_mxc index ed7e47d..9ca6ac2 100644 --- a/doc/README.fec_mxc +++ b/doc/README.fec_mxc @@ -27,6 +27,11 @@ CONFIG_FEC_MXC_PHYADDR Optional, selects the exact phy address that should be connected and function fecmxc_initialize will try to initialize it.
+CONFIG_FEC_FIXED_SPEED
- Optional, selects a fixed speed on the MAC interface without asking some
- phy. This is usefull if there is a direct MAC <-> MAC connection, for
- example if the CPU is connected directly via the RGMII interface to a
- ethernet-switch.
Reading the ethaddr from the SoC eFuses: if CONFIG_FEC_MXC is defined and the U-Boot environment does not contain the diff --git a/drivers/net/fec_mxc.c b/drivers/net/fec_mxc.c index 360f8e4..e871b3e 100644 --- a/drivers/net/fec_mxc.c +++ b/drivers/net/fec_mxc.c @@ -233,6 +233,7 @@ static int miiphy_restart_aneg(struct eth_device *dev) return ret; }
+#ifndef CONFIG_FEC_FIXED_SPEED static int miiphy_wait_aneg(struct eth_device *dev) { uint32_t start; @@ -260,6 +261,7 @@ static int miiphy_wait_aneg(struct eth_device *dev)
return 0; } +#endif /* CONFIG_FEC_FIXED_SPEED */ #endif
static int fec_rx_task_enable(struct fec_priv *fec) @@ -502,6 +504,8 @@ static int fec_open(struct eth_device *edev) } speed = fec->phydev->speed; } +#elif CONFIG_FEC_FIXED_SPEED
- speed = CONFIG_FEC_FIXED_SPEED;
#else miiphy_wait_aneg(edev); speed = miiphy_speed(edev->name, fec->phy_id);
Really I had already sent a patch for this issue, and I merged this into u-boot-imx, - next
http://git.denx.de/?p=u-boot/u-boot-imx.git;a=commit;h=57b7371e24e429ee37403...
Anyway, your patch is more complete as mine. I will drop my patch and substitute with your.
Best regards, Stefano Babic
On 07/11/2016 04:23 PM, Stefano Babic wrote:
Hi Hannes,
Hi Stefano,
On 22/06/2016 12:07, Hannes Schmelzer wrote:
If MAC is directly connected to another MAC (like a switch for example) we don't need to probe for a phy, autoneogation and so on. We simply have to setup speed.
Signed-off-by: Hannes Schmelzer oe5hpm@oevsv.at
doc/README.fec_mxc | 5 +++++ drivers/net/fec_mxc.c | 4 ++++ 2 files changed, 9 insertions(+)
diff --git a/doc/README.fec_mxc b/doc/README.fec_mxc index ed7e47d..9ca6ac2 100644 --- a/doc/README.fec_mxc +++ b/doc/README.fec_mxc @@ -27,6 +27,11 @@ CONFIG_FEC_MXC_PHYADDR Optional, selects the exact phy address that should be connected and function fecmxc_initialize will try to initialize it.
+CONFIG_FEC_FIXED_SPEED
Optional, selects a fixed speed on the MAC interface without asking some
phy. This is usefull if there is a direct MAC <-> MAC connection, for
example if the CPU is connected directly via the RGMII interface to a
ethernet-switch.
Reading the ethaddr from the SoC eFuses: if CONFIG_FEC_MXC is defined and the U-Boot environment does not contain the
diff --git a/drivers/net/fec_mxc.c b/drivers/net/fec_mxc.c index 360f8e4..e871b3e 100644 --- a/drivers/net/fec_mxc.c +++ b/drivers/net/fec_mxc.c @@ -233,6 +233,7 @@ static int miiphy_restart_aneg(struct eth_device *dev) return ret; }
+#ifndef CONFIG_FEC_FIXED_SPEED static int miiphy_wait_aneg(struct eth_device *dev) { uint32_t start; @@ -260,6 +261,7 @@ static int miiphy_wait_aneg(struct eth_device *dev)
return 0; } +#endif /* CONFIG_FEC_FIXED_SPEED */ #endif
static int fec_rx_task_enable(struct fec_priv *fec) @@ -502,6 +504,8 @@ static int fec_open(struct eth_device *edev) } speed = fec->phydev->speed; } +#elif CONFIG_FEC_FIXED_SPEED
- speed = CONFIG_FEC_FIXED_SPEED; #else miiphy_wait_aneg(edev); speed = miiphy_speed(edev->name, fec->phy_id);
Really I had already sent a patch for this issue, and I merged this into u-boot-imx, - next
http://git.denx.de/?p=u-boot/u-boot-imx.git;a=commit;h=57b7371e24e429ee37403...
Anyway, your patch is more complete as mine. I will drop my patch and substitute with your.
many thanks, looking forward for getting merged into u-boot master.
Best regards, Stefano Babic
cheers, Hannes
participants (4)
-
Hannes Schmelzer -
Hannes Schmelzer -
Joe Hershberger -
Stefano Babic