* [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4
@ 2020-11-09 14:22 Michael Pfeiffer
2020-11-10 14:46 ` Ferruh Yigit
2020-11-10 15:59 ` Ferruh Yigit
0 siblings, 2 replies; 12+ messages in thread
From: Michael Pfeiffer @ 2020-11-09 14:22 UTC (permalink / raw)
To: Keith Wiles; +Cc: dev, Michael Pfeiffer
Unlike TCP, UDP checksums are optional and may be zero to indicate "not
set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
this special case to the checksum offload emulation in net/tap.
Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
---
drivers/net/tap/rte_eth_tap.c | 13 +++++++++++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index 2f8abb12c..e486b41c5 100644
--- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c
@@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
uint16_t cksum = 0;
void *l3_hdr;
void *l4_hdr;
+ struct rte_udp_hdr *udp_hdr;
if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
l2_len += 4;
@@ -349,10 +350,18 @@ tap_verify_csum(struct rte_mbuf *mbuf)
/* Don't verify checksum for multi-segment packets. */
if (mbuf->nb_segs > 1)
return;
- if (l3 == RTE_PTYPE_L3_IPV4)
+ if (l3 == RTE_PTYPE_L3_IPV4) {
+ if (l4 == RTE_PTYPE_L4_UDP) {
+ udp_hdr = (struct rte_udp_hdr *)l4_hdr;
+ if (udp_hdr->dgram_cksum == 0) {
+ mbuf->ol_flags |= PKT_RX_L4_CKSUM_NONE;
+ return;
+ }
+ }
cksum = ~rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr);
- else if (l3 == RTE_PTYPE_L3_IPV6)
+ } else if (l3 == RTE_PTYPE_L3_IPV6) {
cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
+ }
mbuf->ol_flags |= cksum ?
PKT_RX_L4_CKSUM_BAD :
PKT_RX_L4_CKSUM_GOOD;
--
2.29.2
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-09 14:22 [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4 Michael Pfeiffer
@ 2020-11-10 14:46 ` Ferruh Yigit
2020-11-10 15:56 ` Ferruh Yigit
2020-11-10 16:01 ` Morten Brørup
2020-11-10 15:59 ` Ferruh Yigit
1 sibling, 2 replies; 12+ messages in thread
From: Ferruh Yigit @ 2020-11-10 14:46 UTC (permalink / raw)
To: Michael Pfeiffer, Keith Wiles; +Cc: dev
On 11/9/2020 2:22 PM, Michael Pfeiffer wrote:
> Unlike TCP, UDP checksums are optional and may be zero to indicate "not
> set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
> this special case to the checksum offload emulation in net/tap.
>
> Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
> ---
> drivers/net/tap/rte_eth_tap.c | 13 +++++++++++--
> 1 file changed, 11 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
> index 2f8abb12c..e486b41c5 100644
> --- a/drivers/net/tap/rte_eth_tap.c
> +++ b/drivers/net/tap/rte_eth_tap.c
> @@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
> uint16_t cksum = 0;
> void *l3_hdr;
> void *l4_hdr;
> + struct rte_udp_hdr *udp_hdr;
>
> if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
> l2_len += 4;
> @@ -349,10 +350,18 @@ tap_verify_csum(struct rte_mbuf *mbuf)
> /* Don't verify checksum for multi-segment packets. */
> if (mbuf->nb_segs > 1)
> return;
> - if (l3 == RTE_PTYPE_L3_IPV4)
> + if (l3 == RTE_PTYPE_L3_IPV4) {
> + if (l4 == RTE_PTYPE_L4_UDP) {
> + udp_hdr = (struct rte_udp_hdr *)l4_hdr;
> + if (udp_hdr->dgram_cksum == 0) {
> + mbuf->ol_flags |= PKT_RX_L4_CKSUM_NONE;
> + return;
> + }
> + }
> cksum = ~rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr);
> - else if (l3 == RTE_PTYPE_L3_IPV6)
> + } else if (l3 == RTE_PTYPE_L3_IPV6) {
> cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
> + }
> mbuf->ol_flags |= cksum ?
> PKT_RX_L4_CKSUM_BAD :
> PKT_RX_L4_CKSUM_GOOD;
>
While checking this I stuck with following part:
cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
...
mbuf->ol_flags |= cksum ?
PKT_RX_L4_CKSUM_BAD :
PKT_RX_L4_CKSUM_GOOD;
Is this correct, or am I missing something, can intention be '!' here instead of
'~' ?
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-10 14:46 ` Ferruh Yigit
@ 2020-11-10 15:56 ` Ferruh Yigit
2020-11-10 16:01 ` Morten Brørup
1 sibling, 0 replies; 12+ messages in thread
From: Ferruh Yigit @ 2020-11-10 15:56 UTC (permalink / raw)
To: Michael Pfeiffer, Keith Wiles; +Cc: dev
On 11/10/2020 2:46 PM, Ferruh Yigit wrote:
> On 11/9/2020 2:22 PM, Michael Pfeiffer wrote:
>> Unlike TCP, UDP checksums are optional and may be zero to indicate "not
>> set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
>> this special case to the checksum offload emulation in net/tap.
>>
>> Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
>> ---
>> drivers/net/tap/rte_eth_tap.c | 13 +++++++++++--
>> 1 file changed, 11 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
>> index 2f8abb12c..e486b41c5 100644
>> --- a/drivers/net/tap/rte_eth_tap.c
>> +++ b/drivers/net/tap/rte_eth_tap.c
>> @@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
>> uint16_t cksum = 0;
>> void *l3_hdr;
>> void *l4_hdr;
>> + struct rte_udp_hdr *udp_hdr;
>> if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
>> l2_len += 4;
>> @@ -349,10 +350,18 @@ tap_verify_csum(struct rte_mbuf *mbuf)
>> /* Don't verify checksum for multi-segment packets. */
>> if (mbuf->nb_segs > 1)
>> return;
>> - if (l3 == RTE_PTYPE_L3_IPV4)
>> + if (l3 == RTE_PTYPE_L3_IPV4) {
>> + if (l4 == RTE_PTYPE_L4_UDP) {
>> + udp_hdr = (struct rte_udp_hdr *)l4_hdr;
>> + if (udp_hdr->dgram_cksum == 0) {
>> + mbuf->ol_flags |= PKT_RX_L4_CKSUM_NONE;
>> + return;
>> + }
>> + }
>> cksum = ~rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr);
>> - else if (l3 == RTE_PTYPE_L3_IPV6)
>> + } else if (l3 == RTE_PTYPE_L3_IPV6) {
>> cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
>> + }
>> mbuf->ol_flags |= cksum ?
>> PKT_RX_L4_CKSUM_BAD :
>> PKT_RX_L4_CKSUM_GOOD;
>>
>
> While checking this I stuck with following part:
>
> cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
> ...
> mbuf->ol_flags |= cksum ?
> PKT_RX_L4_CKSUM_BAD :
> PKT_RX_L4_CKSUM_GOOD;
>
>
> Is this correct, or am I missing something, can intention be '!' here instead of
> '~' ?
This seems correct, this is related to what cksum functions return and how to
verify cksum, I will proceed with the patch.
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-09 14:22 [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4 Michael Pfeiffer
2020-11-10 14:46 ` Ferruh Yigit
@ 2020-11-10 15:59 ` Ferruh Yigit
2020-11-11 7:23 ` Michael Pfeiffer
1 sibling, 1 reply; 12+ messages in thread
From: Ferruh Yigit @ 2020-11-10 15:59 UTC (permalink / raw)
To: Michael Pfeiffer, Keith Wiles; +Cc: dev
On 11/9/2020 2:22 PM, Michael Pfeiffer wrote:
> Unlike TCP, UDP checksums are optional and may be zero to indicate "not
> set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
> this special case to the checksum offload emulation in net/tap.
>
> Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
> ---
> drivers/net/tap/rte_eth_tap.c | 13 +++++++++++--
> 1 file changed, 11 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
> index 2f8abb12c..e486b41c5 100644
> --- a/drivers/net/tap/rte_eth_tap.c
> +++ b/drivers/net/tap/rte_eth_tap.c
> @@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
> uint16_t cksum = 0;
> void *l3_hdr;
> void *l4_hdr;
> + struct rte_udp_hdr *udp_hdr;
>
> if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
> l2_len += 4;
> @@ -349,10 +350,18 @@ tap_verify_csum(struct rte_mbuf *mbuf)
> /* Don't verify checksum for multi-segment packets. */
> if (mbuf->nb_segs > 1)
> return;
> - if (l3 == RTE_PTYPE_L3_IPV4)
> + if (l3 == RTE_PTYPE_L3_IPV4) {
> + if (l4 == RTE_PTYPE_L4_UDP) {
> + udp_hdr = (struct rte_udp_hdr *)l4_hdr;
> + if (udp_hdr->dgram_cksum == 0) {
Overall patch looks good to me, but can you please add a comment on top of above
check to describe why checksum can be zero, as done in the commit log.
> + mbuf->ol_flags |= PKT_RX_L4_CKSUM_NONE;
> + return;
> + }
> + }
> cksum = ~rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr);
> - else if (l3 == RTE_PTYPE_L3_IPV6)
> + } else if (l3 == RTE_PTYPE_L3_IPV6) {
> cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
> + }
> mbuf->ol_flags |= cksum ?
> PKT_RX_L4_CKSUM_BAD :
> PKT_RX_L4_CKSUM_GOOD;
>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-10 14:46 ` Ferruh Yigit
2020-11-10 15:56 ` Ferruh Yigit
@ 2020-11-10 16:01 ` Morten Brørup
2020-11-10 17:42 ` Ferruh Yigit
1 sibling, 1 reply; 12+ messages in thread
From: Morten Brørup @ 2020-11-10 16:01 UTC (permalink / raw)
To: Ferruh Yigit, Michael Pfeiffer, Keith Wiles; +Cc: dev
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Ferruh Yigit
> Sent: Tuesday, November 10, 2020 3:47 PM
>
> On 11/9/2020 2:22 PM, Michael Pfeiffer wrote:
> > Unlike TCP, UDP checksums are optional and may be zero to indicate "not
> > set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
> > this special case to the checksum offload emulation in net/tap.
> >
> > Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
> > ---
> > drivers/net/tap/rte_eth_tap.c | 13 +++++++++++--
> > 1 file changed, 11 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/net/tap/rte_eth_tap.c
> b/drivers/net/tap/rte_eth_tap.c
> > index 2f8abb12c..e486b41c5 100644
> > --- a/drivers/net/tap/rte_eth_tap.c
> > +++ b/drivers/net/tap/rte_eth_tap.c
> > @@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
> > uint16_t cksum = 0;
> > void *l3_hdr;
> > void *l4_hdr;
> > + struct rte_udp_hdr *udp_hdr;
> >
> > if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
> > l2_len += 4;
> > @@ -349,10 +350,18 @@ tap_verify_csum(struct rte_mbuf *mbuf)
> > /* Don't verify checksum for multi-segment packets. */
> > if (mbuf->nb_segs > 1)
> > return;
> > - if (l3 == RTE_PTYPE_L3_IPV4)
> > + if (l3 == RTE_PTYPE_L3_IPV4) {
> > + if (l4 == RTE_PTYPE_L4_UDP) {
> > + udp_hdr = (struct rte_udp_hdr *)l4_hdr;
> > + if (udp_hdr->dgram_cksum == 0) {
> > + mbuf->ol_flags |= PKT_RX_L4_CKSUM_NONE;
> > + return;
> > + }
> > + }
> > cksum = ~rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr);
> > - else if (l3 == RTE_PTYPE_L3_IPV6)
> > + } else if (l3 == RTE_PTYPE_L3_IPV6) {
> > cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
> > + }
> > mbuf->ol_flags |= cksum ?
> > PKT_RX_L4_CKSUM_BAD :
> > PKT_RX_L4_CKSUM_GOOD;
> >
>
> While checking this I stuck with following part:
>
> cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
> ...
> mbuf->ol_flags |= cksum ?
> PKT_RX_L4_CKSUM_BAD :
> PKT_RX_L4_CKSUM_GOOD;
>
>
> Is this correct, or am I missing something, can intention be '!' here
> instead of
> '~' ?
It is correct. The packet's checksum is calculated by rte_ipv6_udptcp_cksum(), and it should be 0xFFFF. The '~' operation makes cksum 0 iff the calculated checksum is 0xFFFF.
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-10 16:01 ` Morten Brørup
@ 2020-11-10 17:42 ` Ferruh Yigit
2020-11-11 7:06 ` Morten Brørup
0 siblings, 1 reply; 12+ messages in thread
From: Ferruh Yigit @ 2020-11-10 17:42 UTC (permalink / raw)
To: Morten Brørup, Michael Pfeiffer, Keith Wiles; +Cc: dev
On 11/10/2020 4:01 PM, Morten Brørup wrote:
>> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Ferruh Yigit
>> Sent: Tuesday, November 10, 2020 3:47 PM
>>
>> On 11/9/2020 2:22 PM, Michael Pfeiffer wrote:
>>> Unlike TCP, UDP checksums are optional and may be zero to indicate "not
>>> set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
>>> this special case to the checksum offload emulation in net/tap.
>>>
>>> Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
>>> ---
>>> drivers/net/tap/rte_eth_tap.c | 13 +++++++++++--
>>> 1 file changed, 11 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/net/tap/rte_eth_tap.c
>> b/drivers/net/tap/rte_eth_tap.c
>>> index 2f8abb12c..e486b41c5 100644
>>> --- a/drivers/net/tap/rte_eth_tap.c
>>> +++ b/drivers/net/tap/rte_eth_tap.c
>>> @@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
>>> uint16_t cksum = 0;
>>> void *l3_hdr;
>>> void *l4_hdr;
>>> + struct rte_udp_hdr *udp_hdr;
>>>
>>> if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
>>> l2_len += 4;
>>> @@ -349,10 +350,18 @@ tap_verify_csum(struct rte_mbuf *mbuf)
>>> /* Don't verify checksum for multi-segment packets. */
>>> if (mbuf->nb_segs > 1)
>>> return;
>>> - if (l3 == RTE_PTYPE_L3_IPV4)
>>> + if (l3 == RTE_PTYPE_L3_IPV4) {
>>> + if (l4 == RTE_PTYPE_L4_UDP) {
>>> + udp_hdr = (struct rte_udp_hdr *)l4_hdr;
>>> + if (udp_hdr->dgram_cksum == 0) {
>>> + mbuf->ol_flags |= PKT_RX_L4_CKSUM_NONE;
>>> + return;
>>> + }
>>> + }
>>> cksum = ~rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr);
>>> - else if (l3 == RTE_PTYPE_L3_IPV6)
>>> + } else if (l3 == RTE_PTYPE_L3_IPV6) {
>>> cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
>>> + }
>>> mbuf->ol_flags |= cksum ?
>>> PKT_RX_L4_CKSUM_BAD :
>>> PKT_RX_L4_CKSUM_GOOD;
>>>
>>
>> While checking this I stuck with following part:
>>
>> cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
>> ...
>> mbuf->ol_flags |= cksum ?
>> PKT_RX_L4_CKSUM_BAD :
>> PKT_RX_L4_CKSUM_GOOD;
>>
>>
>> Is this correct, or am I missing something, can intention be '!' here
>> instead of
>> '~' ?
>
> It is correct. The packet's checksum is calculated by rte_ipv6_udptcp_cksum(), and it should be 0xFFFF. The '~' operation makes cksum 0 iff the calculated checksum is 0xFFFF.
>
Yep, figure that out late,
as far as I understand when the checksum value is zero,
'rte_ipv6_udptcp_cksum()' will return the checksum value and when checksum is
correct in the packet, function will return 0xFFFF, this is based on checksum
calculation, is this right?
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-10 17:42 ` Ferruh Yigit
@ 2020-11-11 7:06 ` Morten Brørup
0 siblings, 0 replies; 12+ messages in thread
From: Morten Brørup @ 2020-11-11 7:06 UTC (permalink / raw)
To: Ferruh Yigit, Michael Pfeiffer, Keith Wiles; +Cc: dev
> From: Ferruh Yigit [mailto:ferruh.yigit@intel.com]
> Sent: Tuesday, November 10, 2020 6:43 PM
>
> On 11/10/2020 4:01 PM, Morten Brørup wrote:
> >> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Ferruh Yigit
> >> Sent: Tuesday, November 10, 2020 3:47 PM
> >>
> >> On 11/9/2020 2:22 PM, Michael Pfeiffer wrote:
> >>> Unlike TCP, UDP checksums are optional and may be zero to indicate
> "not
> >>> set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]).
> Add
> >>> this special case to the checksum offload emulation in net/tap.
> >>>
> >>> Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
> >>> ---
> >>> drivers/net/tap/rte_eth_tap.c | 13 +++++++++++--
> >>> 1 file changed, 11 insertions(+), 2 deletions(-)
> >>>
> >>> diff --git a/drivers/net/tap/rte_eth_tap.c
> >> b/drivers/net/tap/rte_eth_tap.c
> >>> index 2f8abb12c..e486b41c5 100644
> >>> --- a/drivers/net/tap/rte_eth_tap.c
> >>> +++ b/drivers/net/tap/rte_eth_tap.c
> >>> @@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
> >>> uint16_t cksum = 0;
> >>> void *l3_hdr;
> >>> void *l4_hdr;
> >>> + struct rte_udp_hdr *udp_hdr;
> >>>
> >>> if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
> >>> l2_len += 4;
> >>> @@ -349,10 +350,18 @@ tap_verify_csum(struct rte_mbuf *mbuf)
> >>> /* Don't verify checksum for multi-segment packets.
> */
> >>> if (mbuf->nb_segs > 1)
> >>> return;
> >>> - if (l3 == RTE_PTYPE_L3_IPV4)
> >>> + if (l3 == RTE_PTYPE_L3_IPV4) {
> >>> + if (l4 == RTE_PTYPE_L4_UDP) {
> >>> + udp_hdr = (struct rte_udp_hdr *)l4_hdr;
> >>> + if (udp_hdr->dgram_cksum == 0) {
> >>> + mbuf->ol_flags |= PKT_RX_L4_CKSUM_NONE;
> >>> + return;
> >>> + }
> >>> + }
> >>> cksum = ~rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr);
> >>> - else if (l3 == RTE_PTYPE_L3_IPV6)
> >>> + } else if (l3 == RTE_PTYPE_L3_IPV6) {
> >>> cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
> >>> + }
> >>> mbuf->ol_flags |= cksum ?
> >>> PKT_RX_L4_CKSUM_BAD :
> >>> PKT_RX_L4_CKSUM_GOOD;
> >>>
> >>
> >> While checking this I stuck with following part:
> >>
> >> cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
> >> ...
> >> mbuf->ol_flags |= cksum ?
> >> PKT_RX_L4_CKSUM_BAD :
> >> PKT_RX_L4_CKSUM_GOOD;
> >>
> >>
> >> Is this correct, or am I missing something, can intention be '!'
> here
> >> instead of
> >> '~' ?
> >
> > It is correct. The packet's checksum is calculated by
> rte_ipv6_udptcp_cksum(), and it should be 0xFFFF. The '~' operation
> makes cksum 0 iff the calculated checksum is 0xFFFF.
> >
>
> Yep, figure that out late,
> as far as I understand when the checksum value is zero,
> 'rte_ipv6_udptcp_cksum()' will return the checksum value and when
> checksum is
> correct in the packet, function will return 0xFFFF, this is based on
> checksum
> calculation, is this right?
Exactly.
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-10 15:59 ` Ferruh Yigit
@ 2020-11-11 7:23 ` Michael Pfeiffer
2020-11-11 9:31 ` Ferruh Yigit
0 siblings, 1 reply; 12+ messages in thread
From: Michael Pfeiffer @ 2020-11-11 7:23 UTC (permalink / raw)
To: Ferruh Yigit, Keith Wiles; +Cc: dev
Hi,
On Tue, 2020-11-10 at 15:59 +0000, Ferruh Yigit wrote:
> On 11/9/2020 2:22 PM, Michael Pfeiffer wrote:
> > Unlike TCP, UDP checksums are optional and may be zero to indicate "not
> > set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
> > this special case to the checksum offload emulation in net/tap.
> >
> > Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
> > ---
> > drivers/net/tap/rte_eth_tap.c | 13 +++++++++++--
> > 1 file changed, 11 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
> > index 2f8abb12c..e486b41c5 100644
> > --- a/drivers/net/tap/rte_eth_tap.c
> > +++ b/drivers/net/tap/rte_eth_tap.c
> > @@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
> > uint16_t cksum = 0;
> > void *l3_hdr;
> > void *l4_hdr;
> > + struct rte_udp_hdr *udp_hdr;
> >
> > if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
> > l2_len += 4;
> > @@ -349,10 +350,18 @@ tap_verify_csum(struct rte_mbuf *mbuf)
> > /* Don't verify checksum for multi-segment packets. */
> > if (mbuf->nb_segs > 1)
> > return;
> > - if (l3 == RTE_PTYPE_L3_IPV4)
> > + if (l3 == RTE_PTYPE_L3_IPV4) {
> > + if (l4 == RTE_PTYPE_L4_UDP) {
> > + udp_hdr = (struct rte_udp_hdr *)l4_hdr;
> > + if (udp_hdr->dgram_cksum == 0) {
>
> Overall patch looks good to me, but can you please add a comment on top of
> above
> check to describe why checksum can be zero, as done in the commit log.
Sure, I will update the patch. I am also not completely sure whether
PKT_RX_L4_CKSUM_NONE is the right flag for this case (rather than _UNKNOWN).
From rte_core_mbuf.h:
* - PKT_RX_L4_CKSUM_UNKNOWN: no information about the RX L4 checksum
* - PKT_RX_L4_CKSUM_NONE: the L4 checksum is not correct in the packet
* data, but the integrity of the L4 data is verified.
The second part after the "but" is not really the case here. I don't know how
relevant the distinction is, as most application side code will probably only
do something like
if ((mbuf->ol_flags & PKT_RX_L4_CKSUM_MASK) == PKT_RX_L4_CKSUM_BAD)
rte_pktmbuf_free(mbuf);
anyway. Do you have any opinions on that?
> > + mbuf->ol_flags |=
> > PKT_RX_L4_CKSUM_NONE;
> > + return;
> > + }
> > + }
> > cksum = ~rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr);
> > - else if (l3 == RTE_PTYPE_L3_IPV6)
> > + } else if (l3 == RTE_PTYPE_L3_IPV6) {
> > cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
> > + }
> > mbuf->ol_flags |= cksum ?
> > PKT_RX_L4_CKSUM_BAD :
> > PKT_RX_L4_CKSUM_GOOD;
> >
>
Regards
Michael
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-11 7:23 ` Michael Pfeiffer
@ 2020-11-11 9:31 ` Ferruh Yigit
2020-11-13 13:02 ` Ferruh Yigit
0 siblings, 1 reply; 12+ messages in thread
From: Ferruh Yigit @ 2020-11-11 9:31 UTC (permalink / raw)
To: Michael Pfeiffer, Keith Wiles, Olivier Matz; +Cc: dev
On 11/11/2020 7:23 AM, Michael Pfeiffer wrote:
> Hi,
>
> On Tue, 2020-11-10 at 15:59 +0000, Ferruh Yigit wrote:
>> On 11/9/2020 2:22 PM, Michael Pfeiffer wrote:
>>> Unlike TCP, UDP checksums are optional and may be zero to indicate "not
>>> set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
>>> this special case to the checksum offload emulation in net/tap.
>>>
>>> Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
>>> ---
>>> drivers/net/tap/rte_eth_tap.c | 13 +++++++++++--
>>> 1 file changed, 11 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
>>> index 2f8abb12c..e486b41c5 100644
>>> --- a/drivers/net/tap/rte_eth_tap.c
>>> +++ b/drivers/net/tap/rte_eth_tap.c
>>> @@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
>>> uint16_t cksum = 0;
>>> void *l3_hdr;
>>> void *l4_hdr;
>>> + struct rte_udp_hdr *udp_hdr;
>>>
>>> if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
>>> l2_len += 4;
>>> @@ -349,10 +350,18 @@ tap_verify_csum(struct rte_mbuf *mbuf)
>>> /* Don't verify checksum for multi-segment packets. */
>>> if (mbuf->nb_segs > 1)
>>> return;
>>> - if (l3 == RTE_PTYPE_L3_IPV4)
>>> + if (l3 == RTE_PTYPE_L3_IPV4) {
>>> + if (l4 == RTE_PTYPE_L4_UDP) {
>>> + udp_hdr = (struct rte_udp_hdr *)l4_hdr;
>>> + if (udp_hdr->dgram_cksum == 0) {
>>
>> Overall patch looks good to me, but can you please add a comment on top of
>> above
>> check to describe why checksum can be zero, as done in the commit log.
>
> Sure, I will update the patch. I am also not completely sure whether
> PKT_RX_L4_CKSUM_NONE is the right flag for this case (rather than _UNKNOWN).
> From rte_core_mbuf.h:
>
> * - PKT_RX_L4_CKSUM_UNKNOWN: no information about the RX L4 checksum
> * - PKT_RX_L4_CKSUM_NONE: the L4 checksum is not correct in the packet
> * data, but the integrity of the L4 data is verified.
>
> The second part after the "but" is not really the case here. I don't know how
> relevant the distinction is, as most application side code will probably only
> do something like
>
> if ((mbuf->ol_flags & PKT_RX_L4_CKSUM_MASK) == PKT_RX_L4_CKSUM_BAD)
> rte_pktmbuf_free(mbuf);
>
> anyway. Do you have any opinions on that?
>
I also checked for that and wasn't sure about it :) cc'ed Olivier too for comment.
I think it is NOT 'PKT_RX_L4_CKSUM_UNKNOWN', since we know that checksum value
is 0x0000 which means it is not provided.
'PKT_RX_L4_CKSUM_NONE' suits better but not sure about the expectation on
"integrity of the L4 data is verified" part, I assume that explanation is just
to differentiate between 'CKSUM_BAD'.
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-11 9:31 ` Ferruh Yigit
@ 2020-11-13 13:02 ` Ferruh Yigit
2020-11-13 14:03 ` [dpdk-dev] [PATCH v2] " Michael Pfeiffer
0 siblings, 1 reply; 12+ messages in thread
From: Ferruh Yigit @ 2020-11-13 13:02 UTC (permalink / raw)
To: Michael Pfeiffer, Keith Wiles, Olivier Matz; +Cc: dev
On 11/11/2020 9:31 AM, Ferruh Yigit wrote:
> On 11/11/2020 7:23 AM, Michael Pfeiffer wrote:
>> Hi,
>>
>> On Tue, 2020-11-10 at 15:59 +0000, Ferruh Yigit wrote:
>>> On 11/9/2020 2:22 PM, Michael Pfeiffer wrote:
>>>> Unlike TCP, UDP checksums are optional and may be zero to indicate "not
>>>> set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
>>>> this special case to the checksum offload emulation in net/tap.
>>>>
>>>> Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
>>>> ---
>>>> drivers/net/tap/rte_eth_tap.c | 13 +++++++++++--
>>>> 1 file changed, 11 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
>>>> index 2f8abb12c..e486b41c5 100644
>>>> --- a/drivers/net/tap/rte_eth_tap.c
>>>> +++ b/drivers/net/tap/rte_eth_tap.c
>>>> @@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
>>>> uint16_t cksum = 0;
>>>> void *l3_hdr;
>>>> void *l4_hdr;
>>>> + struct rte_udp_hdr *udp_hdr;
>>>> if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
>>>> l2_len += 4;
>>>> @@ -349,10 +350,18 @@ tap_verify_csum(struct rte_mbuf *mbuf)
>>>> /* Don't verify checksum for multi-segment packets. */
>>>> if (mbuf->nb_segs > 1)
>>>> return;
>>>> - if (l3 == RTE_PTYPE_L3_IPV4)
>>>> + if (l3 == RTE_PTYPE_L3_IPV4) {
>>>> + if (l4 == RTE_PTYPE_L4_UDP) {
>>>> + udp_hdr = (struct rte_udp_hdr *)l4_hdr;
>>>> + if (udp_hdr->dgram_cksum == 0) {
>>>
>>> Overall patch looks good to me, but can you please add a comment on top of
>>> above
>>> check to describe why checksum can be zero, as done in the commit log.
>>
>> Sure, I will update the patch. I am also not completely sure whether
>> PKT_RX_L4_CKSUM_NONE is the right flag for this case (rather than _UNKNOWN).
>> From rte_core_mbuf.h:
>>
>> * - PKT_RX_L4_CKSUM_UNKNOWN: no information about the RX L4 checksum
>> * - PKT_RX_L4_CKSUM_NONE: the L4 checksum is not correct in the packet
>> * data, but the integrity of the L4 data is verified.
>>
>> The second part after the "but" is not really the case here. I don't know how
>> relevant the distinction is, as most application side code will probably only
>> do something like
>>
>> if ((mbuf->ol_flags & PKT_RX_L4_CKSUM_MASK) == PKT_RX_L4_CKSUM_BAD)
>> rte_pktmbuf_free(mbuf);
>>
>> anyway. Do you have any opinions on that?
>>
>
> I also checked for that and wasn't sure about it :) cc'ed Olivier too for comment.
>
> I think it is NOT 'PKT_RX_L4_CKSUM_UNKNOWN', since we know that checksum value
> is 0x0000 which means it is not provided.
>
> 'PKT_RX_L4_CKSUM_NONE' suits better but not sure about the expectation on
> "integrity of the L4 data is verified" part, I assume that explanation is just
> to differentiate between 'CKSUM_BAD'.
I suggest to continue with 'PKT_RX_L4_CKSUM_NONE'.
Can it be possible to get the new version today, so we can include this to the -rc4?
Thanks
^ permalink raw reply [flat|nested] 12+ messages in thread
* [dpdk-dev] [PATCH v2] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-13 13:02 ` Ferruh Yigit
@ 2020-11-13 14:03 ` Michael Pfeiffer
2020-11-13 14:49 ` Ferruh Yigit
0 siblings, 1 reply; 12+ messages in thread
From: Michael Pfeiffer @ 2020-11-13 14:03 UTC (permalink / raw)
To: Ferruh Yigit; +Cc: Keith Wiles, dev, Michael Pfeiffer
Unlike TCP, UDP checksums are optional and may be zero to indicate "not
set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
this special case to the checksum offload emulation in net/tap.
Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
---
v2:
* Added comment.
drivers/net/tap/rte_eth_tap.c | 18 ++++++++++++++++--
1 file changed, 16 insertions(+), 2 deletions(-)
diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index 2f8abb12c..2542de306 100644
--- a/drivers/net/tap/rte_eth_tap.c
+++ b/drivers/net/tap/rte_eth_tap.c
@@ -303,6 +303,7 @@ tap_verify_csum(struct rte_mbuf *mbuf)
uint16_t cksum = 0;
void *l3_hdr;
void *l4_hdr;
+ struct rte_udp_hdr *udp_hdr;
if (l2 == RTE_PTYPE_L2_ETHER_VLAN)
l2_len += 4;
@@ -349,10 +350,23 @@ tap_verify_csum(struct rte_mbuf *mbuf)
/* Don't verify checksum for multi-segment packets. */
if (mbuf->nb_segs > 1)
return;
- if (l3 == RTE_PTYPE_L3_IPV4)
+ if (l3 == RTE_PTYPE_L3_IPV4) {
+ if (l4 == RTE_PTYPE_L4_UDP) {
+ udp_hdr = (struct rte_udp_hdr *)l4_hdr;
+ if (udp_hdr->dgram_cksum == 0) {
+ /*
+ * For IPv4, a zero UDP checksum
+ * indicates that the sender did not
+ * generate one [RFC 768].
+ */
+ mbuf->ol_flags |= PKT_RX_L4_CKSUM_NONE;
+ return;
+ }
+ }
cksum = ~rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr);
- else if (l3 == RTE_PTYPE_L3_IPV6)
+ } else if (l3 == RTE_PTYPE_L3_IPV6) {
cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr);
+ }
mbuf->ol_flags |= cksum ?
PKT_RX_L4_CKSUM_BAD :
PKT_RX_L4_CKSUM_GOOD;
--
2.20.1
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [dpdk-dev] [PATCH v2] net/tap: Allow all-zero checksum for UDP over IPv4
2020-11-13 14:03 ` [dpdk-dev] [PATCH v2] " Michael Pfeiffer
@ 2020-11-13 14:49 ` Ferruh Yigit
0 siblings, 0 replies; 12+ messages in thread
From: Ferruh Yigit @ 2020-11-13 14:49 UTC (permalink / raw)
To: Michael Pfeiffer; +Cc: Keith Wiles, dev
On 11/13/2020 2:03 PM, Michael Pfeiffer wrote:
> Unlike TCP, UDP checksums are optional and may be zero to indicate "not
> set" [RFC 768] (except for IPv6, where this prohibited [RFC 8200]). Add
> this special case to the checksum offload emulation in net/tap.
>
> Signed-off-by: Michael Pfeiffer <michael.pfeiffer@tu-ilmenau.de>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Applied to dpdk-next-net/main, thanks.
^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2020-11-13 14:49 UTC | newest]
Thread overview: 12+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-11-09 14:22 [dpdk-dev] [PATCH] net/tap: Allow all-zero checksum for UDP over IPv4 Michael Pfeiffer
2020-11-10 14:46 ` Ferruh Yigit
2020-11-10 15:56 ` Ferruh Yigit
2020-11-10 16:01 ` Morten Brørup
2020-11-10 17:42 ` Ferruh Yigit
2020-11-11 7:06 ` Morten Brørup
2020-11-10 15:59 ` Ferruh Yigit
2020-11-11 7:23 ` Michael Pfeiffer
2020-11-11 9:31 ` Ferruh Yigit
2020-11-13 13:02 ` Ferruh Yigit
2020-11-13 14:03 ` [dpdk-dev] [PATCH v2] " Michael Pfeiffer
2020-11-13 14:49 ` Ferruh Yigit
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).