Problem in configurating eTSEC2 to RGMII mode

cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Problem in configurating eTSEC2 to RGMII mode

Jump to solution
1,677 Views
prasantahalder
Contributor II

Hi,

We are using both eTSECs of P1022 in RGMII mode in our custom board. The problem we are facing is with eTSEC2 which is connected to Marvell's RGMII to Copper PHY(88E1510). From PHY LEDs and MII registers, we have found the link is up and autonegotiation complete. But when we are trying to ping from U-BOOT, we are getting "eTSEC2 : tsec: tx error" and then "eTSEC2 : tsec: tx buffers full".

We have checked the POR value of cfg_tsec1_prtc, cfg_tsec2_prtc and cfg_serdes_ports are 1, 1 and "11111" respectively. We have the following questions on this -

1. What could be the possible reason for this ping failure?

2. We have only fed 125MHz clock from a oscillator in TSEC1_GTX_CLK125. Whereas TSEC2_GTX_CLK125 is kept floating. In GUTs_PMUXCR[1588_USB_PWRFIT] is set to GPIO_1 (i.e. 11). Do we need any other registers to be set  for TSEC1_GTX_CLK125 to feed both eTSEC1 and eTSEC2?

3. Another doubt is the autonegotiation result is 100BaseT FDX, whereas it should be 1000BaseT as the link partner is advertising all 10/100/1000BaseT.

Regards,

Prasanta

Labels (1)
0 Kudos
1 Solution
1,102 Views
prasantahalder
Contributor II

Thanks for your help. During debugging, we found that some of the "Misc Debug and Test Interface Pin" were set to low during reset in our board. After correcting this the eTSEC problem disappeared.

View solution in original post

0 Kudos
10 Replies
1,102 Views
shaileshpanchal
Contributor I

Hi , We are using eTSEC1 and eTSEC3 of P1020 in RGMII mode in custom board. We are facing similarly problem "eTSEC1 : tsec: tx error" and then "eTSEC1 : tsec: tx buffers full"

Please could you shared solution how you over come the above problem. 

0 Kudos
1,102 Views
prasantahalder
Contributor II

Hi Shailesh,

Check all the POR pins state during out of reset state. In our case, some  "Misc Debug and Test Interface Pin" were set to low during reset in our board. It was causing the problem.

Regards,

Prasanta

0 Kudos
1,102 Views
prasantahalder
Contributor II

Actually, we have taken the raw dump from uboot prompt after ping failed. During ping the MACCFG1=0x0000_0005 and MACCFG2=0000_7205.

0 Kudos
1,102 Views
ufedor
NXP Employee
NXP Employee

Absence of Syncd_Rx_EN and Syncd_Tx_EN means that corresponding Ethernet controller does not receive its RX and TX clocks.

What is PORDEVSR value?

0 Kudos
1,103 Views
prasantahalder
Contributor II

Thanks for your help. During debugging, we found that some of the "Misc Debug and Test Interface Pin" were set to low during reset in our board. After correcting this the eTSEC problem disappeared.

0 Kudos
1,102 Views
prasantahalder
Contributor II

Hi,

Please find attached the eTSEC1 dump. However we are not currently using eSEC1.

Regards,

Prasanta

0 Kudos
1,102 Views
ufedor
NXP Employee
NXP Employee

Why Rx_EN and Tx_EN are not set in the MACCFG1?

0 Kudos
1,102 Views
prasantahalder
Contributor II

Hi,

1. PMUXCR = 0x8000_1000 and PMUXCR2 = 0xFF00_0000

2. File attached

3. SDK version 1.8

4. We are running the ping test in uboot. So, we are currently using any DTS.

Regards,

Prasanta

0 Kudos
1,102 Views
ufedor
NXP Employee
NXP Employee

Please provide raw memory dump of the initialized eTSEC1 also.

0 Kudos
1,102 Views
ufedor
NXP Employee
NXP Employee

Please provide additional information:

1) PMUXCR value

2) complete raw memory dump of the initialized eTSEC2

3) which NXP Linux SDK version is used?

4) which DTS is used?

0 Kudos