[dpdk-dev] [PATCH v8 10/10] app/testpmd:test VxLAN Tx checksum offload

[Liu, Jijiang]

> -----Original Message-----
> From: Olivier MATZ [mailto:olivier.matz at 6wind.com]
> Sent: Friday, November 14, 2014 5:10 PM
> To: Liu, Jijiang; Thomas Monjalon
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v8 10/10] app/testpmd:test VxLAN Tx checksum
> offload
> 
> Hi Jijiang,
> 
> On 11/14/2014 09:15 AM, Liu, Jijiang wrote:
> >
> > Thomas Monjalon wrote:
> >>
> >> You mean that PKT_TX_VXLAN_CKSUM request hardware checksumming of
> >> outer L3, outer L4, inner L3 and inner L4?
> >> So maybe the name and comments are not enough clear.
> >
> > Yes, PKT_TX_VXLAN_CKSUM request hardware checksum of outerL3, outer L4,
> inner L3 and inner L4.
> 
> I don't understand: it looks in contradiction with our previous
> discussion:
> 
> Olivier Matz wrote:
> >
> > Liu, Jijiang wrote:
> >>
> >> Olivier Matz wrote:
> >>> What is the
> >>> meaning of this flag? Is it enough to checksum outer L3, inner L3,
> >>> and inner L4 as specified in commit log? If yes, why are the other
> >>> flags PKT_TX_IPV4_CSUM, PKT_TX_UDP_CKSUM, (...) added in the mbuf
> later?
> >>> In my comprehension, these flags are needed in addition to
> >>> PKT_TX_VXLAN_CKSUM to do the checksum of the inner headers.
> >>
> >> Yes, these flags(PKT_TX_IPV4_CSUM, PKT_TX_UDP_CKSUM)  are needed by
> >> HW offload of non-tunneling and tunneling  packet.
> >
> > OK, so I understand that when PKT_TX_VXLAN_CKSUM is set, if the driver
> > supports it, it will process IP and UDP checksum of outer header,
> > using l2_len and l3_len.
> 
> So you say that PKT_TX_VXLAN_CKSUM is enough for all inner and outer headers,
> but also that PKT_TX_IPV4_CSUM, PKT_TX_UDP_CKSUM are needed.
> What occurs if we don't set them?
> 
> Now let's say you have an application that receives a TCP packet, then
> encaspulate it in vxlan, and forward it. You want to regenerate the checksum for
> the new outer headers, but you don't need to change the inner ones.
> You say that setting the PKT_TX_VXLAN_CKSUM will request the hw to process
> inner and outer checksum. This is not required in that case.
> Also, do you need to set the pseudo header checksum in the TCP inner header?

Anyway, I explain the checksum mechanism here again.

In my VXLAN patch set, for an VXLAN packet TX checksum offload,  the logics below:

1. only set outer L3/L4 header TX checksum
    tx_checksum set 0x3(0r 0x1) 0
  In this case, the PKT_TX_VXLAN_CKSUM flag is not set as we don't set inner flags(PKT_TX_IPV4_CSUM, PKT_TX_UDP_CKSUM), so we don't need to change inner ones, driver think it is the ordinary packet,  
HW will do outer L3/L4 checksum offload. 

2. only set inner L3/L4 header TX checksum
    tx_checksum set 0x30 0
  In this case, the PKT_TX_VXLAN_CKSUM flag is set, so driver think it is VXLAN packet, and we don't need to change outer ones because we don't set outer flags here (PKT_TX_IPV4_CSUM, PKT_TX_UDP_CKSUM).

3. set outer L3/L4 TX checksum and inner L3&L4 TX checksum 
tx_checksum set 0x31(0x33) 0
in this case, the PKT_TX_VXLAN_CKSUM flag is set, driver think it is VXLAN packet, and we need to change outer and inner as both outer and inner flags are set.

I'm reviewing your TSO patch to see if your logic is correct in the checksum engine.


> Regards,
> Olivier