RSU, Factory and application in CyclonV

[NAdel1]

I have factory image and application image and it works properly.

When the application image corrupt at the beginning the "pof_error" signal is high. but....

When the application image is corrupt in the middle, the FPGA start configuring the application and the error cause to reconfiguing from the factory (when the "Auto restart configuration after error" is 'tick') and it run this way contiuesly.

So my question is:

1. Is the factory image suppose to know that it is not the first time it trying to configure the application image and how does is know that (the AnF?...), or there is not any mechanism for this scenario and it really suppose to enter to a deadlock ?
2. If it doesn't know that, what design can i do for this scenario? read AnF? read the registers?

thank a lot.

hope for your quick answer.

[YuanLi_S_Intel]

Hi Naftali,

Yes sir, you need to tick the "Auto restart configuration after error" in quartus in order for the quartus to configure to factory image when there is an error in configuring application image.

To boot into application / factory image at first, you will need to set the AnF bit.

For more information about remote update, you may refer to link below:

https://www.intel.com/content/dam/www/programmable/us/en/pdfs/literature/ug/ug_altremote.pdf

Thank You.

Regards,

Bruce

[NAdel1]

Hi Bruce,

Your answer is nice, but not to my question.

I understood that tick it cause the factory image to reconfigure when the application image is corrupt.

My question is what will stop the factory image from trying again and again after it reconfigured?

The flow now is this:

1. factory image
2. trying to configure the application image
3. because of the error, the FPGA reconfigure the factory image
4. trying again to configure the application image
5. because of the error, the FPGA reconfigure the factory image

and so...

How can i stop it in stage 3, after it tried to configure the application image and failed and configure the factory image?

Thanx.

[NAdel1]

Hi,

Eventually in my fsm that runs the RSU block, before the reset state, i read the register "000" from the RSU, and i checked the bit 0 (CRC_ERROR), if its '1' - turn up ERROR flag and stay in the factory image, else, keep moving to application image.

I tried to read AnF and if it '1' then stay in factory image, but apparently this register erased before it back to factory image after the application error occurred.

Thanx for your help.

regards.

[YuanLi_S_Intel]

HI Naftali,

Usually, when appliaction image is having issue, the RSU will try to reload the application image 2/3 times. After that, it will fall back to factory image.

What you are seeing is normal, the bit 0 is 1 as your RSU fails to load application image. It will turn to 0 (normal) when factory image is loaded.

Thank You