[dpdk-dev] dpdk-1.6.0r1 problem with Intel 82599ES

Jun Xiao xiaojuntime at gmail.com
Tue May 6 05:06:36 CEST 2014


It's not really CPU relevance but IOMMU fault, iommu=pt kernel option
solves the problem.

On Tue, May 6, 2014 at 10:55 AM, Jun Xiao <xiaojuntime at gmail.com> wrote:
> Sorry for the spam, I just noticed it's AMD cpu! Is it possible to
> make it fail quickly if dpdk cannot really work with AMD cpu?
>
> On Tue, May 6, 2014 at 10:26 AM, Jun Xiao <xiaojuntime at gmail.com> wrote:
>> Hi,
>>
>> I am running dpdk-1.6.0r1 with Intel 82599ES with some problems:
>> - The test driver is l2fwd, and it is started successfully without any
>> problem (all DEBUG build flags are enabled, but I didn't find anything
>> abnormal)
>> - l2fwd cannot see any packet generated by pktgen on another machine.
>> - The connectivity between l2fwd machine and pktgen machine is fine
>> (verified by ping before igb_uio is loaded)
>> - # tools/pci_unbind.py --status
>>
>> Network devices using IGB_UIO driver
>> ====================================
>> 0000:22:00.0 '82599ES 10-Gigabit SFI/SFP+ Network Connection'
>> drv=igb_uio unused=
>> 0000:22:00.1 '82599ES 10-Gigabit SFI/SFP+ Network Connection'
>> drv=igb_uio unused=
>>
>> Network devices using kernel driver
>> ===================================
>> 0000:01:00.0 'NetXtreme II BCM5709 Gigabit Ethernet' if=em1 drv=bnx2
>> unused=<none> *Active*
>> 0000:01:00.1 'NetXtreme II BCM5709 Gigabit Ethernet' if=em2 drv=bnx2
>> unused=<none>
>>
>> I am pretty confident on the procedures since I tried it out on
>> another setup with Intel I350 Gigabit NIC.
>>
>> Thanks,
>> Jun


More information about the dev mailing list