RTOS/AM3359: FOE download update program

Cazaban Laurence

Part Number: AM3359

Tool/software: TI-RTOS

Hello everyone,

I have an Ethercat network which is composed by an Ethercat master with a real time Core Intime and a stack provided by KPA and 6 Ethercat slaves ARM3359.

The software in my Texas slave is base on the one provided by Texas with the Beckofff patch: ethercat_slave_full_AM335x_arm.pjt

I have a new problem since I have update my stack master server provided by KPA. I have updated my version of KPA with MDK_INtime6_v1.6_December_2018

In fact, with these versions, the transfer by FOE is not functioning anymore:

I try to transfert an update file:

In KPA Studio, I ask the Bootstrap Mode,
Then I try to transfer an update file by FOE and, at the moment on the transfer begins, all the network is broken.

If I make the same operation with another version of KPA (KPA MDK 1.5 x86 [INtime 6.x]-MDK_INtime6.2_v1.5_March_2017) everything is ok !

In fact, after contacting KPA, it seems that the problem comes from that in BOOTSTRAP Mode, at the beginning of FOE downloading, the master try to read the slave EEprom and in the BOOTSTRAP Mode it seems impossible to read the slave EEprom.

Is there anyone who can help me on this subject ?

Thanks a lot

Laurence

over 7 years ago

0 Paula Carrillo over 7 years ago

TI__Mastermind 40580 points

Hi Laurence, maybe ignorance from my side, but why the master try to read slave EEPROM when updating application (new FW)?. My understanding is that after requesting BOOT mode, new app is written in Flash.

thank you,

Paula

0 Cazaban Laurence over 7 years ago in reply to Paula Carrillo

Expert 1500 points

Hello Paula,

I don't know why master try to read the slave EEPROM when the file is transferred by FOE. It's the way KPA had decided to do.... In fact, it is when the file is transferred, after requestinq BOOT to INIT that the app is written in Flash.

But the real question is why it is impossible to read EEPROM in BOOTSTRAP mode ?

Thanks

Laurence

0 Paula Carrillo over 7 years ago in reply to Cazaban Laurence

TI__Mastermind 40580 points

Laurence, I consulted with one of our developers. Our implementation starts writing application in BOOTSTRAP mode and completes (at least the last block) when it is BOOT_TO_INIT transition.

EEPROM is stored in Flash. If the Master is updating the application using FoE, the flash will be busy in writing the application and this might be the reason KPA master is not able to read the EEPROM at the same time.

thank you,
Paula

0 Paula Carrillo over 7 years ago in reply to Paula Carrillo

TI__Mastermind 40580 points

Laurence, just to clarify, EEPROM is stored in Flash however for reading it there is a local copy in RAM. In any case, EEPROM follows defined protocol and rejects access while busy. If you have some Wireshark logs could you please share then with us?

thank you,
Paula

0 Cazaban Laurence over 7 years ago in reply to Paula Carrillo

Expert 1500 points

Hello Paula,

Thanks a lot for your help.

I have no idea what I can do.

I join you a wireshark log: is it ok for you ?

Thanks a lot

Laurence

4846.dump.zip

0 Paula Carrillo over 7 years ago in reply to Cazaban Laurence

TI__Mastermind 40580 points

Hi Laurence, let me confirm and do a request.. I guess above wireshark is for No_OK scenario (latest KPA doing FoE and not able to read EEPROM).

If so, can you also give me a wireshark which shows OK scenario (KPA, doing FoE and reading EEPROM)?

I would be easier to compare for me.

thanks for your help,
Paula

0 Cazaban Laurence over 7 years ago in reply to Paula Carrillo

Expert 1500 points

Hello Paula,

I confirm: the log is for a no-ok situation !

I will make logs for a ok situation today and send it to you as soon as possible.

Thanks for your help

Laurence

0 Cazaban Laurence over 7 years ago in reply to Paula Carrillo

Expert 1500 points

dump_ok.zipThis is a log for an ok situation.

Thanks a lot for your help

0 Paula Carrillo over 7 years ago in reply to Cazaban Laurence

TI__Mastermind 40580 points

Laurence, I missed your reply, my apology. Let me check and come back to you soon.
thank you,

Paula

0 Cazaban Laurence over 7 years ago in reply to Paula Carrillo

Expert 1500 points

Hello Paula,

Sorry to ask you again but Have you any news for me concerning my problem on FOE download.

I'm still in convesration with KPA Koenig and they wait me to give them more informations on the subject.

Thanks for your help

LAurence

0 Paula Carrillo over 7 years ago in reply to Cazaban Laurence

TI__Mastermind 40580 points

Hi Laurence, from your wiresharks I filtered by EEPROM control/status bit 15 == busy (ecat.reg.ctrlstat.busy == 1) both wireshark OK and NOK show telegrams with busy status and next frame was OK (no busy). So, I couldn't get any good conclusion there.. I will ping our developers for further ideas.

The main idea was to check when we see EEPROM interface busy and firmware being updated. If you have a chance please also take a look, maybe I missed something..

Wireshark filter for EEPROM Control/Status register => ecat.ado == 0x502.

thank you,

Paula

0 Cazaban Laurence over 7 years ago in reply to Paula Carrillo

Expert 1500 points

Hello PAula,

Thanks for your help. I understand that my problem is a KPA Problem : it is not normal to try to access the eeprom in order to check flags while FOE transfer is on going.

I send them a message to give them this information. I will give you their feedback.

Thanks,

Laurence

0 Paula Carrillo over 7 years ago in reply to Cazaban Laurence

TI__Mastermind 40580 points

Hi Laurence, let me mark this thread as resolve for the moment. We can re-open it or start another one later if it is the case.

thank you,

Paula

0 Cazaban Laurence over 7 years ago in reply to Paula Carrillo

Expert 1500 points

Ok Paula: let's do like this !
Thanks for all

Processors

Processors forum

RTOS/AM3359: FOE download update program