[dpdk-dev] 回复: ixgbe and UDP with zero checksum

[Feifei Wang]

Hi, Haiyue

We will release the patch to support zero checksum  in arm neon path later.

Best Regards
Feifei

> -----邮件原件-----
> 发件人: Wang, Haiyue <haiyue.wang at intel.com>
> 发送时间: 2021年2月2日 15:42
> 收件人: Feifei Wang <Feifei.Wang2 at arm.com>; Ruifeng Wang
> <Ruifeng.Wang at arm.com>
> 抄送: Guo, Jia <jia.guo at intel.com>; Aaron Conole <aconole at redhat.com>;
> dev at dpdk.org; Yang, Qiming <qiming.yang at intel.com>; Paolo Valerio
> <pvalerio at redhat.com>
> 主题: RE: ixgbe and UDP with zero checksum
> 
> + ARM experts, Feifei and Ruifeng.
> 
> Need your further support for ARM NEON path.
> 
> https://patchwork.dpdk.org/patch/87617/
> 
> > -----Original Message-----
> > From: Paolo Valerio <pvalerio at redhat.com>
> > Sent: Friday, January 29, 2021 22:20
> > To: Wang, Haiyue <haiyue.wang at intel.com>; Yang, Qiming
> > <qiming.yang at intel.com>
> > Cc: Guo, Jia <jia.guo at intel.com>; Aaron Conole <aconole at redhat.com>;
> > dev at dpdk.org
> > Subject: RE: ixgbe and UDP with zero checksum
> >
> > "Wang, Haiyue" <haiyue.wang at intel.com> writes:
> >
> > > Hi Paolo,
> > >
> > >> -----Original Message-----
> > >> From: Wang, Haiyue
> > >> Sent: Friday, January 29, 2021 10:02
> > >> To: Paolo Valerio <pvalerio at redhat.com>
> > >> Cc: Guo, Jia <jia.guo at intel.com>; Aaron Conole
> > >> <aconole at redhat.com>; dev at dpdk.org
> > >> Subject: RE: ixgbe and UDP with zero checksum
> > >>
> > >> > -----Original Message-----
> > >> > From: Paolo Valerio <pvalerio at redhat.com>
> > >> > Sent: Thursday, January 28, 2021 05:35
> > >> > To: Wang, Haiyue <haiyue.wang at intel.com>
> > >> > Cc: Guo, Jia <jia.guo at intel.com>; Aaron Conole
> > >> > <aconole at redhat.com>; dev at dpdk.org
> > >> > Subject: RE: ixgbe and UDP with zero checksum
> > >> >
> > >> > "Wang, Haiyue" <haiyue.wang at intel.com> writes:
> > >> >
> > >> > > Hi Paolo,
> > >> > >
> > >> > >> -----Original Message-----
> > >> > >> From: Paolo Valerio <pvalerio at redhat.com>
> > >> > >> Sent: Wednesday, January 27, 2021 21:50
> > >> > >> To: dev at dpdk.org
> > >> > >> Cc: Guo, Jia <jia.guo at intel.com>; Wang, Haiyue
> > >> > >> <haiyue.wang at intel.com>; Aaron Conole <aconole at redhat.com>
> > >> > >> Subject: ixgbe and UDP with zero checksum
> > >> > >>
> > >> > >> Hi,
> > >> > >>
> > >> > >> performing some tests, I noticed that on ixgbe when receiving
> > >> > >> UDP packets with zero checksum (no checksum) over IPv4, the
> > >> > >> corresponding ol_flag for the l4 checksum is set to
> PKT_RX_L4_CKSUM_BAD.
> > >> > >>
> > >> > >> In particular, this apparently has an impact on OvS using ct()
> > >> > >> action where UDP packets with zero checksum are not tracked
> because of that.
> > >> > >
> > >> > >
> > >> > >>
> > >> > >> [1]
> > >> > >>
> > >> >
> > >>
> >
> https://patchwork.ozlabs.org/project/netdev/patch/20090724040031.30202
> > .1531.stgit at localhost.localdomai
> > >> > >> n/
> > >> > >
> > >> > > About 12 years old patch, it is hardware errata. For fixing
> > >> > > this, have to always disable vector Rx path for 82599, it seems
> > >> > > not a good idea to bring in this workaround. :(
> > >> > >
> > >> >
> > >> > Thanks for the answer.
> > >> > Yes, as I mentioned, the patch is old although still meaningful.
> > >> > I linked it mostly because it mentions the hw errata.
> > >> >
> > >>
> > >> What's your PCI device ID ? My worked ixgbe:
> > >>
> > >
> > > Sorry, I missed the PKT_RX_L4_CKSUM_BAD information, yes, my NIC
> have the issue.
> > >
> > >> 86:00.0 Ethernet controller [0200]: Intel Corporation 82599ES
> > >> 10-Gigabit SFI/SFP+ Network
> > Connection
> > >> [8086:10fb] (rev 01)
> > >>
> > >> I'm wondering if people will complain that the patch will mark the
> > >> real bad checksum UDP as
> > >
> > > Zero checksum is more popular case, please file a bug on
> https://bugs.dpdk.org/ to trace the fix.
> > >
> > > Thanks for pointing it out.
> > >
> >
> > ack, I'm going to file it.
> >
> > Thanks,
> > Paolo
> >
> > >> GOOD. For handling this correctly, looks like driver needs to check
> > >> the UDP's checksum value, if zero, then skip the error information, but
> this makes driver do the network stack things ...
> > >>
> > >>