Re: [dpdk-dev] [PATCH v4] ethdev: add IPv4 and L4 checksum RSS offload types

[Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>]

On 7/13/21 12:38 PM, Zhang, AlvinX wrote:
>> -----Original Message-----
>> From: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
>> Sent: Tuesday, July 13, 2021 3:55 PM
>> To: Zhang, AlvinX <alvinx.zhang@intel.com>; Zhang, Qi Z
>> <qi.z.zhang@intel.com>; ajit.khaparde@broadcom.com
>> Cc: dev@dpdk.org; Singh, Aman Deep <aman.deep.singh@intel.com>
>> Subject: Re: [PATCH v4] ethdev: add IPv4 and L4 checksum RSS offload types
>>
>> On 7/13/21 4:13 AM, Alvin Zhang wrote:
>>> This patch defines new RSS offload types for IPv4 and
>>> L4(TCP/UDP/SCTP) checksum, which are required when users want to
>>> distribute packets based on the IPv4 or L4 checksum field.
>>>
>>> For example "flow create 0 ingress pattern eth / ipv4 / end actions
>>> rss types ipv4-chksum end queues end / end", this flow causes all
>>> matching packets to be distributed to queues on basis of IPv4
>>> checksum.
>>>
>>> Signed-off-by: Alvin Zhang <alvinx.zhang@intel.com>
>>> Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
>>
>> I've failed to find mail where I've added the tag. I've replied with "LGTM" for v2
>> which has IPv4 checksum only and many comments on L4 checksum added in v3.
>> So, I think it is incorrect to inherit "LGTM" as Reviewed-by.
>>
>  
> I'll correct it at next version.         
>     
>>> Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
>>> Acked-by: Aman Deep Singh <aman.deep.singh@intel.com>
>>> ---
>>>
>>> v3: Add L4 checksum RSS offload type
>>> v4: Add doc and help string, update commit log
>>> ---
>>>  app/test-pmd/cmdline.c  | 12 +++++++++---
>>>  app/test-pmd/config.c   |  2 ++
>>>  lib/ethdev/rte_ethdev.h |  2 ++
>>>  3 files changed, 13 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c index
>>> 0268b18..93543d8 100644
>>> --- a/app/test-pmd/cmdline.c
>>> +++ b/app/test-pmd/cmdline.c
>>> @@ -2254,6 +2254,10 @@ struct cmd_config_rss {
>>>  		rss_conf.rss_hf = ETH_RSS_ECPRI;
>>>  	else if (!strcmp(res->value, "mpls"))
>>>  		rss_conf.rss_hf = ETH_RSS_MPLS;
>>> +	else if (!strcmp(res->value, "ipv4-chksum"))
>>> +		rss_conf.rss_hf = ETH_RSS_IPV4_CHKSUM;
>>> +	else if (!strcmp(res->value, "l4-chksum"))
>>> +		rss_conf.rss_hf = ETH_RSS_L4_CHKSUM;
>>>  	else if (!strcmp(res->value, "none"))
>>>  		rss_conf.rss_hf = 0;
>>>  	else if (!strcmp(res->value, "level-default")) { @@ -2325,7 +2329,7
>>> @@ struct cmd_config_rss {
>>>  	.help_str = "port config all rss "
>>>  		"all|default|eth|vlan|ip|tcp|udp|sctp|ether|port|vxlan|geneve|"
>>>
>> 	"nvgre|vxlan-gpe|l2tpv3|esp|ah|pfcp|ecpri|mpls|none|level-default|"
>>> -		"level-outer|level-inner|<flowtype_id>",
>>> +		"level-outer|level-inner|ipv4-chksum|l4-chksum|<flowtype_id>",
>>>  	.tokens = {
>>>  		(void *)&cmd_config_rss_port,
>>>  		(void *)&cmd_config_rss_keyword,
>>> @@ -2438,7 +2442,8 @@ struct cmd_config_rss_hash_key {
>>>  				 "ipv6-tcp-ex#ipv6-udp-ex#"
>>>  				 "l3-src-only#l3-dst-only#l4-src-only#l4-dst-only#"
>>>  				 "l2-src-only#l2-dst-only#s-vlan#c-vlan#"
>>> -				 "l2tpv3#esp#ah#pfcp#pppoe#gtpu#ecpri#mpls");
>>> +				 "l2tpv3#esp#ah#pfcp#pppoe#gtpu#ecpri#mpls#"
>>> +				 "ipv4-chksum#l4-chksum");
>>>  cmdline_parse_token_string_t cmd_config_rss_hash_key_value =
>>>  	TOKEN_STRING_INITIALIZER(struct cmd_config_rss_hash_key, key, NULL);
>>>
>>> @@ -2451,7 +2456,8 @@ struct cmd_config_rss_hash_key {
>>>  		"l2-payload|ipv6-ex|ipv6-tcp-ex|ipv6-udp-ex|"
>>>  		"l3-src-only|l3-dst-only|l4-src-only|l4-dst-only|"
>>>  		"l2-src-only|l2-dst-only|s-vlan|c-vlan|"
>>> -		"l2tpv3|esp|ah|pfcp|pppoe|gtpu|ecpri|mpls "
>>> +		"l2tpv3|esp|ah|pfcp|pppoe|gtpu|ecpri|mpls|"
>>> +		"ipv4-chksum|l4-chksum "
>>>  		"<string of hex digits (variable length, NIC dependent)>",
>>>  	.tokens = {
>>>  		(void *)&cmd_config_rss_hash_key_port, diff --git
>>> a/app/test-pmd/config.c b/app/test-pmd/config.c index 43c79b5..14968bf
>>> 100644
>>> --- a/app/test-pmd/config.c
>>> +++ b/app/test-pmd/config.c
>>> @@ -140,6 +140,8 @@
>>>  	{ "gtpu", ETH_RSS_GTPU },
>>>  	{ "ecpri", ETH_RSS_ECPRI },
>>>  	{ "mpls", ETH_RSS_MPLS },
>>> +	{ "ipv4-chksum", ETH_RSS_IPV4_CHKSUM },
>>> +	{ "l4-chksum", ETH_RSS_L4_CHKSUM },
>>>  	{ NULL, 0 },
>>>  };
>>>
>>> diff --git a/lib/ethdev/rte_ethdev.h b/lib/ethdev/rte_ethdev.h index
>>> faf3bd9..63b0321 100644
>>> --- a/lib/ethdev/rte_ethdev.h
>>> +++ b/lib/ethdev/rte_ethdev.h
>>> @@ -537,6 +537,8 @@ struct rte_eth_rss_conf {
>>>  #define ETH_RSS_PPPOE		   (1ULL << 31)
>>>  #define ETH_RSS_ECPRI		   (1ULL << 32)
>>>  #define ETH_RSS_MPLS		   (1ULL << 33)
>>> +#define ETH_RSS_IPV4_CHKSUM	   (1ULL << 34)
>>> +#define ETH_RSS_L4_CHKSUM	   (1ULL << 35)	/* TCP/UDP/SCTP */
>>
>> It does not reply on my questions at all.

Above list of Layer 4 protocols does not say if the flag may
be reported when all above protocols are supported or some
protocols are supported only.

Also I had a question about UDP checksum 0 since it could be
treated in a different ways since logically it is no checksum
at all.

> You have said:       
>>>> I don't think so. IMHO, it should report all RSS capabilities      
>>>> regardless generic vs flow API RSS action.       
>>>      
>> May be it is OK to report subset in      
>> dev_info->flow_type_rss_offloads, but I'm very uncomfortable with the       
>> approach. Superset sounds more logical to me, but has drawbacks as        
>> well.         
> 
> Here I have another question:              
> There are flow type definition and RSS offload type definition,         
> #define RTE_ETH_FLOW_RAW                 1       
> #define RTE_ETH_FLOW_IPV4                2     
> #define RTE_ETH_FLOW_FRAG_IPV4           3       
> #define RTE_ETH_FLOW_NONFRAG_IPV4_TCP    4        
> 
> #define ETH_RSS_IPV4               (1ULL << 2)      
> #define ETH_RSS_FRAG_IPV4          (1ULL << 3)         
> #define ETH_RSS_NONFRAG_IPV4_TCP   (1ULL << 4)        
>   
> are they the different expressions of the same concept?

Sorry, but I don't understand the question.

> 
> If yes, why they have been decoupled by the commit: fce6b66893.

Do you mean that motivation in the changeset description is not
clear?

> Then what the flow type of ETH_RSS_PORT, it's UDP, TCP, or SCTP?      

I have no answer to the question as well. That's why I'm trying
to avoid it in the patch.

> 
> If not, the PMDs can report supported RSS flow type by dev_info->flow_type_rss_offloads,   
> but have no way to report supported RSS offload types.          

Sorry, don't understand.

>>
>> Nack