Skip to contentSkip to conversation

Problems with new ethernet init on mpc8360

U-Boot-Users Jerry Van Baren 4 participants 9 messages Feb 26, 2008

lore
Share & tools
lore
#0 lore raw

Hi Kim,

I have a mpc8360emds eval board. The ee62ed patch (below) speeds up the start up (good), but seems to cause the board to hang on the *second* ethernet operation.

I can cause this by doing two pings in a row: the first ping initializes the network and completes successfully. The second ping hangs the
board: it never completes. I cannot abort it with ^C. The only way out is to reset the board.

The same thing happens with two back-to-back tftp operations or a ping followed by a tftp operation.

If I back out the ee62ed patch (below), it "fixes" the problem.

I tried backing out some simple portions of the patch (added delays back in) with no success. I've given up chasing this for tonight. Any inspiration you can generate would be appreciated, I'll chase it some more tomorrow.

Thanks,
gvb

commit ee62ed3286f83b98b7785e0318dc6379e78f7ff6

Author: Kim Phillips <[email protected]>
Date:   Tue Jan 15 14:11:00 2008 -0600

net: reduce boot latency on QE UEC based boards

actually polling for PHY autonegotiation to finish enables us to
remove the
5 second boot prompt latency present on QE based boards.

call to qe_set_mii_clk_src in init_phy, and mv call to init_phy from uec_initialize to uec_init by Joakim Tjernlund; autonegotiation wait code shamelessly stolen from tsec driver.

also rm unused CONFIG_RMII_MODE code.

#1 lore raw
Quoted 14 lines · On Mon, 25 Feb 2008 22:55:37 -0500 Jerry Van Baren <[email protected]> wrote:

Hi Kim,

I have a mpc8360emds eval board. The ee62ed patch (below) speeds up the start up (good), but seems to cause the board to hang on the *second* ethernet operation.

I can cause this by doing two pings in a row: the first ping initializes the network and completes successfully. The second ping hangs the
board: it never completes. I cannot abort it with ^C. The only way out is to reset the board.

The same thing happens with two back-to-back tftp operations or a ping followed by a tftp operation.

Quoted 6 lines

If I back out the ee62ed patch (below), it "fixes" the problem.

I tried backing out some simple portions of the patch (added delays back in) with no success. I've given up chasing this for tonight. Any inspiration you can generate would be appreciated, I'll chase it some more tomorrow.

I verified rc2 on rev. 1.1 and 2.1 boards pings more than once.

can you provide more details about which u-boot version, what board rev you're on, what kind of connection you're using, speed, etc.?

Kim

#2 lore raw
Quoted 24 lines · Kim Phillips
Quoted 14 lines · On Mon, 25 Feb 2008 22:55:37 -0500 Jerry Van Baren <[email protected]> wrote:

Hi Kim,

I have a mpc8360emds eval board. The ee62ed patch (below) speeds up the start up (good), but seems to cause the board to hang on the *second* ethernet operation.

I can cause this by doing two pings in a row: the first ping initializes the network and completes successfully. The second ping hangs the
board: it never completes. I cannot abort it with ^C. The only way out is to reset the board.

The same thing happens with two back-to-back tftp operations or a ping followed by a tftp operation.

Quoted 6 lines

If I back out the ee62ed patch (below), it "fixes" the problem.

I tried backing out some simple portions of the patch (added delays back in) with no success. I've given up chasing this for tonight. Any inspiration you can generate would be appreciated, I'll chase it some more tomorrow.

Quoted 6 lines

I verified rc2 on rev. 1.1 and 2.1 boards pings more than once.

can you provide more details about which u-boot version, what board rev you're on, what kind of connection you're using, speed, etc.?

Kim

Hi Kim,

OK, probably operator error. I claim to be using 1.3.2-rc2 but my repository clone could be introducing problems. I'll do a clean clone and rebuild of u-boot, hopefully that is all it is.

It is connection is through a commercial switch, should be 100bT full duplex (not verified, cannot see the blinky lights from here). I don't know the board rev off-hand (I'm not on-site at the moment).

=> version
U-Boot 1.3.2-rc2-g330a77f6-dirty (Feb 25 2008 - 22:21:12) MPC83XX

root:~> cat /proc/cpuinfo
processor : 0
cpu : e300c1
clock : 528.000000MHz
revision : 2.0 (pvr 8083 0020)
bogomips : 131.58
timebase : 66000000
platform : MPC836x MDS

Thanks,
gvb

#3 lore raw
In reply to Kim Phillips
Quoted 24 lines · Kim Phillips
Quoted 14 lines · On Mon, 25 Feb 2008 22:55:37 -0500 Jerry Van Baren <[email protected]> wrote:

Hi Kim,

I have a mpc8360emds eval board. The ee62ed patch (below) speeds up the start up (good), but seems to cause the board to hang on the *second* ethernet operation.

I can cause this by doing two pings in a row: the first ping initializes the network and completes successfully. The second ping hangs the
board: it never completes. I cannot abort it with ^C. The only way out is to reset the board.

The same thing happens with two back-to-back tftp operations or a ping followed by a tftp operation.

Quoted 6 lines

If I back out the ee62ed patch (below), it "fixes" the problem.

I tried backing out some simple portions of the patch (added delays back in) with no success. I've given up chasing this for tonight. Any inspiration you can generate would be appreciated, I'll chase it some more tomorrow.

Quoted 6 lines

I verified rc2 on rev. 1.1 and 2.1 boards pings more than once.

can you provide more details about which u-boot version, what board rev you're on, what kind of connection you're using, speed, etc.?

Kim

OK, I confirmed it was operator error by doing a clean clone and rebuild of u-boot.

Sorry for the noise,
gvb

#4 lore raw
Quoted 29 lines · Jerry Van Baren
Quoted 28 lines · Kim Phillips
Quoted 14 lines · On Mon, 25 Feb 2008 22:55:37 -0500 Jerry Van Baren <[email protected]> wrote:

Hi Kim,

I have a mpc8360emds eval board. The ee62ed patch (below) speeds up the start up (good), but seems to cause the board to hang on the *second* ethernet operation.

I can cause this by doing two pings in a row: the first ping initializes the network and completes successfully. The second ping hangs the
board: it never completes. I cannot abort it with ^C. The only way out is to reset the board.

The same thing happens with two back-to-back tftp operations or a ping followed by a tftp operation.

Quoted 6 lines

If I back out the ee62ed patch (below), it "fixes" the problem.

I tried backing out some simple portions of the patch (added delays back in) with no success. I've given up chasing this for tonight. Any inspiration you can generate would be appreciated, I'll chase it some more tomorrow.

I verified rc2 on rev. 1.1 and 2.1 boards pings more than once.

can you provide more details about which u-boot version, what board rev you're on, what kind of connection you're using, speed, etc.?

Quoted 7 lines
Quoted 1 line

Kim

OK, I confirmed it was operator error by doing a clean clone and rebuild of u-boot.

Sorry for the noise,
gvb

My bad, the noise *wasn't* noise. What worked for me was 1.3.1 (above results), which initializes the ethernet on start up. When I cloned the u-boot repo (or u-boot-testing), the second ping hangs the board for me.

Is the fact that I don't have a valid env a clue (I have never bothered to save an env)? Hmmmm, what env variables could be messing me up? I would suspect the ethernet rotation one. I don't have the second etherspiggot plugged in, but that should simply fail and rotate back to the first etherspiggot (scratches head).

Thanks,
gvb

U-Boot 1.3.2-rc2 (Feb 26 2008 - 13:21:20) MPC83XX

Reset Status: Software Hard, External/Internal Soft, External/Internal Hard

CPU:   e300c1, MPC8360E, Rev: 12 at 528 MHz, CSB:  264 MHz
Board: Freescale MPC8360EMDS
I2C:   ready
DRAM:  256 MB (DDR1, 64-bit, ECC off)
SDRAM: 64 MB (local bus)

PCI 32bit bus on PMC1 & PMC2 &PMC3
FLASH: 16 MB
*** Warning - bad CRC, using default environment

PCI clock is 66MHz

In:    serial
Out:   serial
Err:   serial
Net:   FSL UEC0, FSL UEC1
Hit any key to stop autoboot:  0

=> set serverip 192.168.47.8 ; set ipaddr 192.168.47.214

=> ping 192.168.47.8
UEC: PHY is Marvell 88E11x1 (1410cc2)
FSL UEC0: Full Duplex
switching to rgmii 100
FSL UEC0: Speed 100BT
FSL UEC0: Link is up
Using FSL UEC0 device
host 192.168.47.8 is alive

=> ping 192.168.47.8

#5 lore raw
Quoted 82 lines · Jerry Van Baren
Quoted 29 lines · Jerry Van Baren
Quoted 28 lines · Kim Phillips
Quoted 14 lines · On Mon, 25 Feb 2008 22:55:37 -0500 Jerry Van Baren <[email protected]> wrote:

Hi Kim,

I have a mpc8360emds eval board. The ee62ed patch (below) speeds up the start up (good), but seems to cause the board to hang on the *second* ethernet operation.

I can cause this by doing two pings in a row: the first ping initializes the network and completes successfully. The second ping hangs the
board: it never completes. I cannot abort it with ^C. The only way out is to reset the board.

The same thing happens with two back-to-back tftp operations or a ping followed by a tftp operation.

Quoted 6 lines

If I back out the ee62ed patch (below), it "fixes" the problem.

I tried backing out some simple portions of the patch (added delays back in) with no success. I've given up chasing this for tonight. Any inspiration you can generate would be appreciated, I'll chase it some more tomorrow.

I verified rc2 on rev. 1.1 and 2.1 boards pings more than once.

can you provide more details about which u-boot version, what board rev you're on, what kind of connection you're using, speed, etc.?

Quoted 6 lines
Quoted 1 line

Kim

OK, I confirmed it was operator error by doing a clean clone and rebuild of u-boot.

Sorry for the noise,
gvb

My bad, the noise *wasn't* noise. What worked for me was 1.3.1 (above results), which initializes the ethernet on start up. When I cloned the u-boot repo (or u-boot-testing), the second ping hangs the board for me.

Is the fact that I don't have a valid env a clue (I have never bothered to save an env)? Hmmmm, what env variables could be messing me up? I would suspect the ethernet rotation one. I don't have the second etherspiggot plugged in, but that should simply fail and rotate back to the first etherspiggot (scratches head).

Thanks,
gvb

U-Boot 1.3.2-rc2 (Feb 26 2008 - 13:21:20) MPC83XX

Reset Status: Software Hard, External/Internal Soft, External/Internal Hard

CPU:   e300c1, MPC8360E, Rev: 12 at 528 MHz, CSB:  264 MHz
Board: Freescale MPC8360EMDS
I2C:   ready
DRAM:  256 MB (DDR1, 64-bit, ECC off)
SDRAM: 64 MB (local bus)

PCI 32bit bus on PMC1 & PMC2 &PMC3
FLASH: 16 MB
*** Warning - bad CRC, using default environment

PCI clock is 66MHz

In:    serial
Out:   serial
Err:   serial
Net:   FSL UEC0, FSL UEC1
Hit any key to stop autoboot:  0

=> set serverip 192.168.47.8 ; set ipaddr 192.168.47.214

=> ping 192.168.47.8
UEC: PHY is Marvell 88E11x1 (1410cc2)
FSL UEC0: Full Duplex
switching to rgmii 100
FSL UEC0: Speed 100BT
FSL UEC0: Link is up
Using FSL UEC0 device
host 192.168.47.8 is alive

Quoted 1 line

=> ping 192.168.47.8

Sorry, I should have included the default environment contents for reference:

=> printenv
bootcmd=setenv bootargs root=/dev/nfs rw nfsroot=$serverip:$rootpath ip=$ipaddrr
ramboot=setenv bootargs root=/dev/ram rw console=$consoledev,$baudrate $othbootr
nfsboot=setenv bootargs root=/dev/nfs rw nfsroot=$serverip:$rootpath ip=$ipaddrr
bootdelay=6
baudrate=115200
loads_echo=1
ethaddr=00:04:9f:ef:01:01
eth1addr=00:04:9f:ef:01:02
loadaddr=200000
netdev=eth0
consoledev=ttyS0
ramdiskaddr=1000000
ramdiskfile=ramfs.83xx
fdtaddr=400000
fdtfile=mpc8360emds.dtb
stdin=serial
stdout=serial
stderr=serial
ethact=FSL UEC0

Environment size: 979/8188 bytes

#6 lore raw
Quoted 120 lines

-----Original Message-----
From: u-boot-users-bounces at lists.sourceforge.net [mailto:u-boot-users-bounces at lists.sourceforge.net] On Behalf Of Jerry Van Baren
Sent: den 26 februari 2008 19:41
To: Kim Phillips
Cc: U-Boot Users
Subject: Re: [U-Boot-Users] Problems with new ethernet init on mpc8360

Quoted 82 lines · Jerry Van Baren
Quoted 29 lines · Jerry Van Baren
Quoted 28 lines · Kim Phillips
Quoted 14 lines · On Mon, 25 Feb 2008 22:55:37 -0500 Jerry Van Baren <[email protected]> wrote:

Hi Kim,

I have a mpc8360emds eval board. The ee62ed patch (below) speeds up the start up (good), but seems to cause the board to hang on the *second* ethernet operation.

I can cause this by doing two pings in a row: the first ping initializes the network and completes successfully. The second ping hangs the
board: it never completes. I cannot abort it with ^C. The only way out is to reset the board.

The same thing happens with two back-to-back tftp operations or a ping followed by a tftp operation.

Quoted 6 lines

If I back out the ee62ed patch (below), it "fixes" the problem.

I tried backing out some simple portions of the patch (added delays back in) with no success. I've given up chasing this for tonight. Any inspiration you can generate would be appreciated, I'll chase it some more tomorrow.

I verified rc2 on rev. 1.1 and 2.1 boards pings more than once.

can you provide more details about which u-boot version, what board rev you're on, what kind of connection you're using, speed, etc.?

Quoted 6 lines
Quoted 1 line

Kim

OK, I confirmed it was operator error by doing a clean clone and rebuild of u-boot.

Sorry for the noise,
gvb

My bad, the noise *wasn't* noise. What worked for me was 1.3.1 (above results), which initializes the ethernet on start up. When I cloned the u-boot repo (or u-boot-testing), the second ping hangs the board for me.

Is the fact that I don't have a valid env a clue (I have never bothered to save an env)? Hmmmm, what env variables could be messing me up? I would suspect the ethernet rotation one. I don't have the second etherspiggot plugged in, but that should simply fail and rotate back to the first etherspiggot (scratches head).

Thanks,
gvb

U-Boot 1.3.2-rc2 (Feb 26 2008 - 13:21:20) MPC83XX

Reset Status: Software Hard, External/Internal Soft, External/Internal Hard

CPU:   e300c1, MPC8360E, Rev: 12 at 528 MHz, CSB:  264 MHz
Board: Freescale MPC8360EMDS
I2C:   ready
DRAM:  256 MB (DDR1, 64-bit, ECC off)
SDRAM: 64 MB (local bus)

PCI 32bit bus on PMC1 & PMC2 &PMC3
FLASH: 16 MB
*** Warning - bad CRC, using default environment

PCI clock is 66MHz

In:    serial
Out:   serial
Err:   serial
Net:   FSL UEC0, FSL UEC1
Hit any key to stop autoboot:  0

=> set serverip 192.168.47.8 ; set ipaddr 192.168.47.214

=> ping 192.168.47.8
UEC: PHY is Marvell 88E11x1 (1410cc2)
FSL UEC0: Full Duplex
switching to rgmii 100
FSL UEC0: Speed 100BT
FSL UEC0: Link is up
Using FSL UEC0 device
host 192.168.47.8 is alive

Quoted 1 line

=> ping 192.168.47.8

Sorry, I should have included the default environment contents for reference:

=> printenv
bootcmd=setenv bootargs root=/dev/nfs rw nfsroot=$serverip:$rootpath ip=$ipaddrr
ramboot=setenv bootargs root=/dev/ram rw console=$consoledev,$baudrate $othbootr
nfsboot=setenv bootargs root=/dev/nfs rw nfsroot=$serverip:$rootpath ip=$ipaddrr
bootdelay=6
baudrate=115200
loads_echo=1
ethaddr=00:04:9f:ef:01:01
eth1addr=00:04:9f:ef:01:02
loadaddr=200000
netdev=eth0
consoledev=ttyS0
ramdiskaddr=1000000
ramdiskfile=ramfs.83xx
fdtaddr=400000
fdtfile=mpc8360emds.dtb
stdin=serial
stdout=serial
stderr=serial
ethact=FSL UEC0

Quoted 1 line

Environment size: 979/8188 bytes

Maybe you suffer from something I noticed, sometimes the master for MDIO gets out of sync. Try this patch and see if it helps, not sure if it still applies but that should be easy to fix up.

#7 lore raw
Quoted 128 lines

-----Original Message-----
From: u-boot-users-bounces at lists.sourceforge.net [mailto:u-boot-users-bounces at lists.sourceforge.net] On Behalf Of Joakim Tjernlund
Sent: den 26 februari 2008 20:56
To: 'Jerry Van Baren'; 'Kim Phillips'
Cc: 'U-Boot Users'
Subject: Re: [U-Boot-Users] Problems with new ethernet init on mpc8360

Quoted 120 lines

-----Original Message-----
From: u-boot-users-bounces at lists.sourceforge.net [mailto:u-boot-users-bounces at lists.sourceforge.net] On Behalf Of Jerry Van Baren
Sent: den 26 februari 2008 19:41
To: Kim Phillips
Cc: U-Boot Users
Subject: Re: [U-Boot-Users] Problems with new ethernet init on mpc8360

Quoted 82 lines · Jerry Van Baren
Quoted 29 lines · Jerry Van Baren
Quoted 28 lines · Kim Phillips
Quoted 14 lines · On Mon, 25 Feb 2008 22:55:37 -0500 Jerry Van Baren <[email protected]> wrote:

Hi Kim,

I have a mpc8360emds eval board. The ee62ed patch (below) speeds up the start up (good), but seems to cause the board to hang on the *second* ethernet operation.

I can cause this by doing two pings in a row: the first ping initializes the network and completes successfully. The second ping hangs the
board: it never completes. I cannot abort it with ^C. The only way out is to reset the board.

The same thing happens with two back-to-back tftp operations or a ping followed by a tftp operation.

Quoted 6 lines

If I back out the ee62ed patch (below), it "fixes" the problem.

I tried backing out some simple portions of the patch (added delays back in) with no success. I've given up chasing this for tonight. Any inspiration you can generate would be appreciated, I'll chase it some more tomorrow.

I verified rc2 on rev. 1.1 and 2.1 boards pings more than once.

can you provide more details about which u-boot version, what board rev you're on, what kind of connection you're using, speed, etc.?

Quoted 6 lines
Quoted 1 line

Kim

OK, I confirmed it was operator error by doing a clean clone and rebuild of u-boot.

Sorry for the noise,
gvb

My bad, the noise *wasn't* noise. What worked for me was 1.3.1 (above results), which initializes the ethernet on start up. When I cloned the u-boot repo (or u-boot-testing), the second ping hangs the board for me.

Is the fact that I don't have a valid env a clue (I have never bothered to save an env)? Hmmmm, what env variables could be messing me up? I would suspect the ethernet rotation one. I don't have the second etherspiggot plugged in, but that should simply fail and rotate back to the first etherspiggot (scratches head).

Thanks,
gvb

U-Boot 1.3.2-rc2 (Feb 26 2008 - 13:21:20) MPC83XX

Reset Status: Software Hard, External/Internal Soft, External/Internal Hard

CPU:   e300c1, MPC8360E, Rev: 12 at 528 MHz, CSB:  264 MHz
Board: Freescale MPC8360EMDS
I2C:   ready
DRAM:  256 MB (DDR1, 64-bit, ECC off)
SDRAM: 64 MB (local bus)

PCI 32bit bus on PMC1 & PMC2 &PMC3
FLASH: 16 MB
*** Warning - bad CRC, using default environment

PCI clock is 66MHz

In:    serial
Out:   serial
Err:   serial
Net:   FSL UEC0, FSL UEC1
Hit any key to stop autoboot:  0

=> set serverip 192.168.47.8 ; set ipaddr 192.168.47.214

=> ping 192.168.47.8
UEC: PHY is Marvell 88E11x1 (1410cc2)
FSL UEC0: Full Duplex
switching to rgmii 100
FSL UEC0: Speed 100BT
FSL UEC0: Link is up
Using FSL UEC0 device
host 192.168.47.8 is alive

Quoted 1 line

=> ping 192.168.47.8

Sorry, I should have included the default environment contents for reference:

=> printenv
bootcmd=setenv bootargs root=/dev/nfs rw nfsroot=$serverip:$rootpath ip=$ipaddrr
ramboot=setenv bootargs root=/dev/ram rw console=$consoledev,$baudrate $othbootr
nfsboot=setenv bootargs root=/dev/nfs rw nfsroot=$serverip:$rootpath ip=$ipaddrr
bootdelay=6
baudrate=115200
loads_echo=1
ethaddr=00:04:9f:ef:01:01
eth1addr=00:04:9f:ef:01:02
loadaddr=200000
netdev=eth0
consoledev=ttyS0
ramdiskaddr=1000000
ramdiskfile=ramfs.83xx
fdtaddr=400000
fdtfile=mpc8360emds.dtb
stdin=serial
stdout=serial
stderr=serial
ethact=FSL UEC0

Quoted 6 lines
Quoted 1 line

Environment size: 979/8188 bytes

Maybe you suffer from something I noticed, sometimes the master for MDIO gets out of sync. Try this patch and see if it helps, not sure if it still applies but that should be easy to fix up.

[SNIP]

Maybe you need to define netmask also? Or possibly enable full u-boot relocation?

Jocke

#8 lore raw
In reply to Joakim Tjernlund

Joakim Tjernlund wrote:

[snip]

Quoted 3 lines

Maybe you suffer from something I noticed, sometimes the master for MDIO gets out of sync. Try this patch and see if it helps, not sure if it still applies but that should be easy to fix up.

It applied cleanly with some line offsets, but didn't fix my problem. :-(

Quoted 11 lines

From 241d7e4f4fb26b2424311a6012c845bd97dfdafa Mon Sep 17 00:00:00 2001
From: Joakim Tjernlund <[email protected]>
Date: Tue, 18 Dec 2007 14:33:18 +0100
Subject: [PATCH] Move call to qe_set_mii_clk_src() into uec_read_phy_reg() and
uec_write_phy_reg() to make sure MII master clock source is correct.

---
 drivers/qe/uec.c     |    5 -----
 drivers/qe/uec_phy.c |    6 ++++++
 2 files changed, 6 insertions(+), 5 deletions(-)

Defining the netmask didn't help either. Since it works the first time, including bootp with a full tftp load (lots of packets), it implies the network is fine but someone is shutting down the network (improperly?) after the ping/tftp/bootp command completes and not starting it up again on the next network command.

Thanks,
gvb