5 Replies Latest reply on Jan 13, 2016 6:06 AM by Carlos_A

    1G operation of 82599

    PVWB Brown Belt

      We have a design using the 82599 that has errors such that we can only

      use one lane is each direction.  As we would like to test as much of the

      design as possible before doing the revision, are there EEPROM files

      available that allow 1G SERDES operation on lane 0 of the 82599?  I couldn't

      find any in Rev 4.25 version of 82599_EEPROM.zip

       

      Thanks, Paul.

        • Re: 1G operation of 82599
          Carlos_A Brown Belt

          Hello paulwilcox-baker,

           

          Thank you for contacting the Intel Embedded Community.

           

          The information that may help you is stated in sections 6.3.8.2 and 6.3.8.6, on pages 248 and 252 of the Intel(R) 82599 10 GbE Controller Datasheet


          In case that you have more questions related to this device, they should be further answered at the Forum: Wired Ethernet | Intel Communities.

           

          Please let us know if this information is useful to you.

           

          Best Regards,

          Carlos_A.

            • Re: 1G operation of 82599
              PVWB Brown Belt

              Dear Carlos,

               

              Thank you  for this information.  I now get 1G linkup

              shown at the switch connected to the 82599 and

              at the 82599.  However I cannot pass any data.

              The switch statistics show data going to the 82599.

              I can try "ping" to a device on the same network

              as that set using "ifconfig".   But I see no packets

              counted for Tx or Rx for the 82599.

               

              Another thing I have noticed is that the driver puts

              out an "Adapter Removed" message when I set

              the IP address using "ifconfig".  As there is no

              obvious way to attach a file to this message, I'll

              include the driver output and lspci output below.

              The lspci output is odd in that the BAR registers

              show as "disabled", rather than having an address

              set.

               

              Thanks, Paul.

               

              linux-oik7:~ # dmesg | grep ixgbe
              [    3.352762] ixgbe: Intel(R) 10 Gigabit PCI Express Network Driver - version 3.19.1-k
              [    3.352764] ixgbe: Copyright (c) 1999-2014 Intel Corporation.
              [    3.352789] ixgbe 0000:01:00.0: enabling device (0000 -> 0002)
              [    5.548402] ixgbe 0000:01:00.0: irq 62 for MSI/MSI-X
              [    5.548408] ixgbe 0000:01:00.0: irq 63 for MSI/MSI-X
              [    5.548414] ixgbe 0000:01:00.0: irq 64 for MSI/MSI-X
              [    5.548418] ixgbe 0000:01:00.0: irq 65 for MSI/MSI-X
              [    5.548422] ixgbe 0000:01:00.0: irq 66 for MSI/MSI-X
              [    5.548427] ixgbe 0000:01:00.0: irq 67 for MSI/MSI-X
              [    5.548431] ixgbe 0000:01:00.0: irq 68 for MSI/MSI-X
              [    5.548435] ixgbe 0000:01:00.0: irq 69 for MSI/MSI-X
              [    5.548440] ixgbe 0000:01:00.0: irq 70 for MSI/MSI-X
              [    5.548470] ixgbe 0000:01:00.0: Multiqueue Enabled: Rx Queue count = 8, Tx Queue count = 8
              [    5.548596] ixgbe 0000:01:00.0: PCI Express bandwidth of 32GT/s available
              [    5.548597] ixgbe 0000:01:00.0: (Speed:5.0GT/s, Width: x8, Encoding Loss:20%)
              [    5.548692] ixgbe 0000:01:00.0: MAC: 2, PHY: 1, PBA No: FFFFFF-0FF
              [    5.548694] ixgbe 0000:01:00.0: 00:07:10:bc:00:00
              [    5.582956] ixgbe 0000:01:00.0: Intel(R) 10 Gigabit Network Connection
              [  419.289278] ixgbe 0000:01:00.0: Adapter removed
              [  419.307108] ixgbe 0000:01:00.0: registered PHC device on enp1s0

               

              linux-oik7:~ # lspci -d 8086:10fc -v
              01:00.0 Ethernet controller: Intel Corporation 82599 10 Gigabit Dual Port Network Connection (rev 01)
                      Flags: fast devsel, IRQ 16
                      Memory at f7c00000 (64-bit, non-prefetchable) [disabled] [size=128K]
                      I/O ports at e000 [disabled] [size=32]
                      Memory at f7c20000 (64-bit, non-prefetchable) [disabled] [size=16K]
                      Capabilities: [40] Power Management version 3
                      Capabilities: [50] MSI: Enable- Count=1/1 Maskable+ 64bit+
                      Capabilities: [70] MSI-X: Enable- Count=64 Masked-
                      Capabilities: [a0] Express Endpoint, MSI 00
                      Capabilities: [e0] Vital Product Data
                      Capabilities: [100] Advanced Error Reporting
                      Capabilities: [140] Device Serial Number 02-00-d2-ff-ff-b5-54-07
                      Capabilities: [150] Alternative Routing-ID Interpretation (ARI)
                      Capabilities: [160] Single Root I/O Virtualization (SR-IOV)
                      Kernel driver in use: ixgbe
                      Kernel modules: ixgbe

               

              linux-oik7:~ #

                • Re: 1G operation of 82599
                  Carlos_A Brown Belt

                  Hello paulwilcox-baker,

                   

                  Thanks for your update.

                   

                  In order to better understand this situation, we would like to address the following questions:

                   

                  Could you please clarify us if the affected device is an Intel NIC or your own design?

                   

                  In case that it is your project:

                  Could you please let us know the used NVM image?

                  Could you please confirm if this design has been reviewed?

                  Could you please inform us the type of connection that you are using (if it is Backplane or SFP/SFP+)?

                  Could you please give us all the information of the devices that you are connecting to the affected units?

                   

                  In case that it is an external device, could please let us know if you have tried to connect it with a NIC?

                   

                  Please give us all the information that should answer these questions to have better idea of your problem.

                   

                  Thanks in advance for your cooperation.

                   

                  Best Regards,

                  Carlos_A.

                    • Re: 1G operation of 82599
                      PVWB Brown Belt

                      Dear Carlos,

                       

                      I did get data transfer in the end.

                       

                      The "Adapter removed" message

                      seems to be caused by a race

                      condition between the BIOS doing

                      PCIe setup and reset being removed

                      from the 82599. This caused the PCIe

                      BAR registers to be left at zero.  After doing a reboot,

                      which doesn't issue a hardware reset

                      to the 82599,  The BIOS loads the BAR

                      registers correctly and the driver loads properly,

                      and I can transfer data.

                       

                      I thought I had marked your post about the

                      register settings  as "Correct Answer"

                       

                      Thanks for the help, Paul.