Re: [E1000-devel] Packet Drop in Intel Fortville Cards

[Jeff K. <jef...@in...>]

On Wed, 2016-07-20 at 14:17 -0700, Jeff Kirsher wrote:
> On Thu, 2016-07-14 at 10:52 +0000, 남효창 wrote:
> > 
> > Dear Anjali Singhai Jain,
> >  
> > I am Hyochang Nam in SECUI, Korea.
> > In these days, we found an odd symptom in Intel Fortville cards.
> > When we execute tcpdump command in linux shell,
> > packet drop happened even though the traffic was not heavy.
> > We think this symptom was caused by interface reset codes in your
> > previous patch:
> > http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/d
> > ri
> > vers/net/ethernet/intel/i40e/i40e_main.c?id=92faef85613d6d56803ad6130f6
> > da
> > f4a803deddc 
> > Whenever we execute tcpdump, network interfaces are reset,
> > and during the reset timing packets are always dropped in NIC side.
> > Moreover, we had not met this issue in old device drivers of i40e.
> >  
> > We had reported this issue to Intel IPS ( Case #0008532, Click Here ),
> > and SK San in Intel had tried to reproduce it, but failed.
> >  
> > Anyway, in our side, we are struggling this issue still.
> > Our test environments are as follows:
> >   - OS: CentOS 7.2
> >   - Device Driver: i40e-1.5.18
> >   - Card: X710 DA4
> >  
> > Would you let us know this symptom is related to your patch?
> > or would you give us any comment for this symptom?
> Please CC e1000-devel and intel-wired-lan mailing lists, so that all the
> appropriate developers have a chance to review/respond.  Especially if
> the
> person you are emailing is out or on vacation, you can get a quicker
> response by CC'ing the mailing lists as well.
> 
> I have gone ahead and CC'd the appropriate mailing lists for you.

CC'd the correct intel-wired-lan mailing list this time... :-)