Skip to contentSkip to conversation

arm: socfpga: set skew settings for ethernet phy

U-Boot dinguyen at opensource.altera.com 5 participants 9 messages Nov 11, 2014 – Nov 25, 2014

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

From: Dinh Nguyen <[email protected]>

Set the PHY skew settings for the ethernet phy on the SOCFPGA devkit.

Cc Vince Bridgers <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Marek Vasut <[email protected]>
Cc Tom Rini <[email protected]>
Cc Albert Aribaud <[email protected]>
Cc Wolfgang Denk <[email protected]>
---
 board/altera/socfpga/socfpga_cyclone5.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
board/altera/socfpga/socfpga_cyclone5.c +12 -0
index 0f81d89..a2831c8 100644
@@ -8,7 +8,9 @@
 #include <asm/arch/reset_manager.h>
 #include <asm/io.h>
 
+#include <micrel.h>
 #include <netdev.h>
+#include <phy.h>
 
 DECLARE_GLOBAL_DATA_PTR;
 
@@ -39,3 +41,13 @@ int board_init(void)
 
 	return 0;
 }
+
+int board_phy_config(struct phy_device *phydev)
+{
+	ksz9021_phy_extended_write(phydev, MII_KSZ9021_EXT_RGMII_RX_DATA_SKEW,
+				   0x0);
+	ksz9021_phy_extended_write(phydev, MII_KSZ9021_EXT_RGMII_TX_DATA_SKEW,
+				   0x0);
+	ksz9021_phy_extended_write(phydev, MII_KSZ9021_EXT_RGMII_CLOCK_SKEW,
+				   0xf0f0);
+}
Signature · 1 line
2.0.3
#1 lore raw

Hi!

Quoted 4 lines

From: Dinh Nguyen <[email protected]>

Set the PHY skew settings for the ethernet phy on the SOCFPGA devkit.

Should the timing configuration be done in the config file...?

Marek disliked getting the values from the environment, and I'm not sure if that is neccessary. It can be removed if needed.

drivers/net/phy/micrel.c +31 -2
index 507b9a3..06a31b0 100644
@@ -112,17 +112,46 @@ static int ksz9021_phy_extwrite(struct phy_device *phydev, int addr,
 	return ksz9021_phy_extended_write(phydev, regnum, val);
 }
 
+
+
 /* Micrel ksz9021 */
 static int ksz9021_config(struct phy_device *phydev)
 {
 	unsigned ctrl1000 = 0;
 	const unsigned master = CTRL1000_PREFER_MASTER |
-			CTRL1000_CONFIG_MASTER | CTRL1000_MANUAL_CONFIG;
+		CTRL1000_CONFIG_MASTER | CTRL1000_MANUAL_CONFIG;
 	unsigned features = phydev->drv->features;
 
+	printf("ksz9021: configuring\n");
+
+	printf("Configuring PHY skew timing for %s\n",
+	       phydev->drv->name);
+
+	/* min rx data delay */
+	if (ksz9021_phy_extended_write(phydev,
+				       MII_KSZ9021_EXT_RGMII_RX_DATA_SKEW,
+				       getenv_ulong(CONFIG_KSZ9021_DATA_SKEW_ENV, 16,
+						    CONFIG_KSZ9021_DATA_SKEW_VAL)) < 0)
+		return -1;
+	/* min tx data delay */
+	if (ksz9021_phy_extended_write(phydev,
+				       MII_KSZ9021_EXT_RGMII_TX_DATA_SKEW,
+				       getenv_ulong(CONFIG_KSZ9021_DATA_SKEW_ENV, 16,
+						    CONFIG_KSZ9021_DATA_SKEW_VAL)) < 0)
+		return -1;
+	/* max rx/tx clock delay, min rx/tx control */
+	if (ksz9021_phy_extended_write(phydev,
+				       MII_KSZ9021_EXT_RGMII_CLOCK_SKEW,
+				       getenv_ulong(CONFIG_KSZ9021_CLK_SKEW_ENV, 16,
+						    CONFIG_KSZ9021_CLK_SKEW_VAL)) < 0)
+		return -1;
+
+
+
+
 	if (getenv("disable_giga"))
 		features &= ~(SUPPORTED_1000baseT_Half |
-				SUPPORTED_1000baseT_Full);
+			      SUPPORTED_1000baseT_Full);
 	/* force master mode for 1000BaseT due to chip errata */
 	if (features & SUPPORTED_1000baseT_Half)
 		ctrl1000 |= ADVERTISE_1000HALF | master;
include/configs/socfpga_cyclone5.h +18 -8
index 8d54bf89..53a3d71 100644
@@ -53,19 +56,29 @@
 
 /* Ethernet on SoC (EMAC) */
 #if defined(CONFIG_CMD_NET)
-#define CONFIG_EMAC_BASE		SOCFPGA_EMAC0_ADDRESS
+#define CONFIG_EMAC_BASE		SOCFPGA_EMAC1_ADDRESS
 #define CONFIG_PHY_INTERFACE_MODE	PHY_INTERFACE_MODE_RGMII
-#define CONFIG_EPHY0_PHY_ADDR		0
-
-/* PHY */
 #define CONFIG_EPHY1_PHY_ADDR		4
+
 #define CONFIG_PHY_MICREL
 #define CONFIG_PHY_MICREL_KSZ9021
-#define CONFIG_KSZ9021_CLK_SKEW_ENV	"micrel-ksz9021-clk-skew"
-#define CONFIG_KSZ9021_CLK_SKEW_VAL	0xf0f0
-#define CONFIG_KSZ9021_DATA_SKEW_ENV	"micrel-ksz9021-data-skew"
-#define CONFIG_KSZ9021_DATA_SKEW_VAL	0x0
 
+/* phy */
+#define CONFIG_EPHY0_PHY_ADDR          0
+#define CONFIG_EPHY1_PHY_ADDR          4
+#define CONFIG_KSZ9021_CLK_SKEW_ENV    "micrel-ksz9021-clk-skew"
+#define CONFIG_KSZ9021_CLK_SKEW_VAL    0xf0f0
+#define CONFIG_KSZ9021_DATA_SKEW_ENV   "micrel-ksz9021-data-skew"
+#define CONFIG_KSZ9021_DATA_SKEW_VAL   0x0
+/* Type of PHY available */
+#define SOCFPGA_PHYSEL_ENUM_GMII       0x0
+#define SOCFPGA_PHYSEL_ENUM_MII                0x1
+#define SOCFPGA_PHYSEL_ENUM_RGMII      0x2
+#define SOCFPGA_PHYSEL_ENUM_RMII       0x3
+
+
+#else
+#error Youll need ethernet :-)
 #endif
 
 /* Extra Environment */
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#2 lore raw

Dear Pavel Machek,

drivers/net/phy/micrel.c static int ksz9021_phy_extwrite(struct phy_device *phydev, int addr, · line 116

Quoted 11 lines · In message <20141113120956.GB30779@amd> you wrote:
drivers/net/phy/micrel.c +2 -0
index 507b9a3..06a31b0 100644
@@ -112,17 +112,46 @@ static int ksz9021_phy_extwrite(struct phy_device *phydev, int addr,
 	return ksz9021_phy_extended_write(phydev, regnum, val);
 }
 
+
+

Please don't add excessive white space / blank lines.

drivers/net/phy/micrel.c static int ksz9021_phy_extwrite(struct phy_device *phydev, int addr, · line 128

Quoted 4 lines
drivers/net/phy/micrel.c +4 -0
+	printf("ksz9021: configuring\n");
+
+	printf("Configuring PHY skew timing for %s\n",
+	       phydev->drv->name);

Should these eventually be changed into debug() ?

Output gets really noisy otherwise.

drivers/net/phy/micrel.c static int ksz9021_phy_extwrite(struct phy_device *phydev, int addr, · line 152

Quoted 1 line
drivers/net/phy/micrel.c +0 -0
 	if (getenv("disable_giga"))

Should we not rather use the standard "hwconf" approach here?

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
There is a time in the tides of men, Which, taken at its flood, leads
on to success. On the other hand, don't count on it.   - T. K. Lawson
#3 lore raw

Hi!

Quoted 24 lines · Wolfgang Denk

Dear Pavel Machek,

drivers/net/phy/micrel.c static int ksz9021_phy_extwrite(struct phy_device *phydev, int addr, · line 116

Quoted 11 lines · In message <20141113120956.GB30779@amd> you wrote:
drivers/net/phy/micrel.c +2 -0
index 507b9a3..06a31b0 100644
@@ -112,17 +112,46 @@ static int ksz9021_phy_extwrite(struct phy_device *phydev, int addr,
 	return ksz9021_phy_extended_write(phydev, regnum, val);
 }
 
+
+

Please don't add excessive white space / blank lines.

drivers/net/phy/micrel.c static int ksz9021_phy_extwrite(struct phy_device *phydev, int addr, · line 128

Quoted 4 lines
drivers/net/phy/micrel.c +4 -0
+	printf("ksz9021: configuring\n");
+
+	printf("Configuring PHY skew timing for %s\n",
+	       phydev->drv->name);

Should these eventually be changed into debug() ?

Quoted 5 lines

Output gets really noisy otherwise.

drivers/net/phy/micrel.c static int ksz9021_phy_extwrite(struct phy_device *phydev, int addr, · line 152

Quoted 1 line
drivers/net/phy/micrel.c +0 -0
 	if (getenv("disable_giga"))

Should we not rather use the standard "hwconf" approach here?

Comments accepted (this was just a test patch I hand edited it, thats why it was so ugly), but we need to decide, first, which approach to take. (Configuration in environment vs. configuration in config file vs. hardcoded configuration in .c file.). Then, clean patch can be prepared.

I wanted to sign it off, so that it is clear it can be used as a basis for better patch, but I was aware it is not clean enough for merging -- how do I do that?

Thanks,
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#4 lore raw

Hello Pavel,

Quoted 3 lines

I wanted to sign it off, so that it is clear it can be used as a basis for better patch, but I was aware it is not clean enough for merging -- how do I do that?

Why not simply tag it [RFC]?

Quoted 5 lines

Thanks,
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Amicalement,

Signature · 1 line
Albert.
#5 lore raw
Quoted 7 lines · Albert ARIBAUD

Hello Pavel,

Quoted 3 lines

I wanted to sign it off, so that it is clear it can be used as a basis for better patch, but I was aware it is not clean enough for merging -- how do I do that?

Why not simply tag it [RFC]?

Well, dunno, I was not asking for code review at that point. There was a patch submitted, and I was asking if maybe different approach is better.

Subject already said "Re: ...", that is usually disussion, not a patch that needs to be reviewed.

Best regards,
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#6 lore raw
Quoted 7 lines · Pavel Machek
Quoted 6 lines · Albert ARIBAUD

Hello Pavel,

Quoted 3 lines

I wanted to sign it off, so that it is clear it can be used as a basis for better patch, but I was aware it is not clean enough for merging -- how do I do that?

Quoted 8 lines
Quoted 1 line

Why not simply tag it [RFC]?

Well, dunno, I was not asking for code review at that point. There was a patch submitted, and I was asking if maybe different approach is better.

Subject already said "Re: ...", that is usually disussion, not a patch that needs to be reviewed.

This discussion is moot, let's stick to the subject guys, thanks!

Best regards,
Marek Vasut

#7 lore raw

[U-Boot] arm: socfpga: script for installing u-boot / u-boot-SPL

It is very easy to overwrite partition on your harddrive when all you want is installing new u-boot(-SPL). Plus... figuring out needed steps for installation is not trivial, either. This automates installation, and provides safety net by checking signature on existing partition.

[I managed to overwrite partition once, and in second incident second mistake -- -EPERM -- saved my data]

tools/socfpga_install new file +39 -0
new file mode 100755
index 0000000..b62d3e77
@@ -0,0 +1,39 @@
+#!/usr/bin/python3
+
+import subprocess
+import sys
+import os
+
+print("install device ", sys.argv[1])
+device = open(sys.argv[1], "r+b")
+
+sig = device.read(0x44)
+if sig[0x40] != 0x41 or sig[0x41] != 0x53 or \
+   sig[0x42] != 0x30 or sig[0x43] != 0x31:
+    print("Do not see AS01 signature, do you have right partition?")
+    sys.exit(1)
+
+print("u-boot-spl.bin at ", sys.argv[2])
+spl = sys.argv[2]
+
+res = subprocess.check_call(["tools/mkimage", "-T", "socfpgaimage", "-d", spl, "img.tmp"])
+if res != 0:
+   print("mkimage failed.")
+   sys.exit(1)
+device.seek(0)
+
+spl_bin = open("img.tmp", "rb").read()
+if len(spl_bin) != 65536:
+   print("Preloader image has wrong length.")
+   sys.exit(1)
+ 
+device.write(spl_bin)
+device.write(spl_bin)
+device.write(spl_bin)
+device.write(spl_bin)
+
+print("u-boot.img at ", sys.argv[3])
+uboot_bin = open(sys.argv[3], "rb").read()
+print("Have u-boot, %d bytes" % len(uboot_bin))
+device.write(uboot_bin)
+print("All done.")
\ No newline at end of file
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#8 lore raw

[U-Boot] arm: socfpga: script for installing u-boot / u-boot-SPL

Quoted 9 lines · Pavel Machek

It is very easy to overwrite partition on your harddrive when all you want is installing new u-boot(-SPL). Plus... figuring out needed steps for installation is not trivial, either. This automates installation, and provides safety net by checking signature on existing partition.

[I managed to overwrite partition once, and in second incident second mistake -- -EPERM -- saved my data]

+CC Stefano, he has that SWUP tool so he might chime in.

Best regards,
Marek Vasut