Hi Intel-Support,
We were using the i210 as Network-Controller on the 1st PCIe Lane of the Cyclone V SoC with kernel 4.0 successfully.
Right now we are trying to start with Kernel 4.14. the (meanwhile mainline) Intel-PCIe-Driver and the Intel i210 driver are not working somehow.
Can you provide any help to us?
I also would like to ask you for the eepromARM-Tool.
kind regards,
Sebastian
链接已复制
Hello AdolfoS,
thank you for your reply.
We tried embedded i210 devices (with and without Flash) and also a LAN Card with an i210.
We are building a platform with a Cyclone V SoC (Altera/ Intel PSG) and an i210 flashless.
It seems to us as if we have an address-offset of "Minus 2000".
Example:
Kernel 4.0: Register 0x5ba0 = 0x1551
This setup is running.
Kernel 4.14: Register 0x5ba0 is found at 0x3ba0. (0x5ba0 - 0x2000 = 0x3ba0)
This matches all registers we looked at.
Do you have any idea why we see the Offset -2000?
The DMESG is consisting of a lot of our own printouts. I can provide a clear one later.
regards,
Sebastian
Hello PXCSEB
Do you observe that offset with all PCI devices or just with the I210 LAN Card?
Have you tried with the latest Intel Driver: https://downloadcenter.intel.com/download/22283/Ethernet-Intel-Ethernet-Adapter-Complete-Driver-Pack?product=64399 Download Intel® Ethernet Adapter Complete Driver Pack ?
Best Regards,
Adolfo Sanchez
Hello AdolfoS!
Thank you for your reply.
>> Do you observe that offset with all PCI devices or just with the I210 LAN Card?
We observe the offset with all i210 devices. Other PCIe Devices work well, even an intel network card with another chipset.
>> Have you tried with the latest Intel Driver: https://downloadcenter.intel.com/download/22283/Ethernet-Intel-Ethernet-Adapter-Complete-Driver-Pack%3Fproduct%3D64399 Download Intel® Ethernet Adapter Complete Driver Pack ?
Yes
Thank you!
Sebastian
Hello, PXCSEB:
Thanks for your update.
We suggest you verify that the affected Kernel is compatible with all devices affected by the cited situation and it has all the patches. You can verify this using the channels listed at https://www.kernel.org/category/contact-us.html The Linux Kernel Contact Us website.
We hope that this information is useful to you.
Best regards,
Carlos_A.
Hello, PXCSEB :
Thanks for your update.
Please let us clarify our previous suggestion. Based on your previous communication your implementation works perfectly with a previous Kernel version and the only change is a Kernel update.
Due to this fact, the issue is related to the Kernel and it has its support channel that has been mentioned in our previous communication.
We hope that this message helps you to understand our recommendation.
Best regards,
Carlos_A.
Hello Carlos,
we suspect the PCIe-Driver vom Altera/Intel right now in our setup. No Kernel between 4.5 and 4.14 is working with Cyclone V and i210 in our System Environment.
I got a premier support now and I hope there we can find a solution.
Can you supply me the eepromARM-Tool?
Best regards,
Sebastian
Hello, PXCSEB :
Thanks for your reply.
The https://cdrdv2.intel.com/v1/dl/getContent/572162 Intel I210, I211, and X550 EEPROM Access Tool (EAT) document # 572162 is the requested tool.
It is accessible when you are logged into your Resource & Design Center (RDC) privileged account. It can be requested by filling out the https://www.intel.com/content/www/us/en/forms/design/contact-support.html RDC Account Support form.
We hope that this information may help you.
Best regards,
Carlos_A.
Hello, PXCSEB :
Thanks for your update.
This problem is a kernel version issue which is unrelated to hardware. Due to this fact, we suggest you address this consultation as a reference to the following channel:
https://www.linux.com/forum https://www.linux.com/forum
We hope that this information may help you.
Best regards,
Carlos_A.
Hi Carlos!
Thank you for your update.
Why do you think that this is a kernel version issue? We tried several Kernels from 4.0 to 4.14 and all have the same behaviour...
I will try the linux-forum. But if you have further more specific information why you suspect the kernel this would be very helpful. Did you observe this behaviour already somewhere else?
best regards,
Sebastian
