[v2] net/iavf: fix negative GTP-U flow rules create successfully
Checks
Commit Message
Currently, when use 'flow' command to create a negative GTP-U rule, it
will be created successfully.
The list shows the impacted outer and inner 'ipv4' GTP-U patterns with
'ipv4' or 'gtpu' type:
- iavf_pattern_eth_ipv4_gtpu_ipv4_udp
- iavf_pattern_eth_ipv4_gtpu_eh_ipv4_udp
- iavf_pattern_eth_ipv4_gtpu_ipv4_tcp
- iavf_pattern_eth_ipv4_gtpu_eh_ipv4_tcp
- more impacted patterns with 'gtpu' type:
> iavf_pattern_eth_ipv4_gtpu_ipv4
> iavf_pattern_eth_ipv4_gtpu_eh_ipv4
Same as the outer and inner 'ipv6' GTP-U patterns.
So, this commit adds the invalid RSS combinations in 'invalid_rss_comb'
array to make result correct.
The list of added invalid RSS combinations:
- ETH_RSS_IPV4 | ETH_RSS_NONFRAG_IPV4_TCP
- ETH_RSS_IPV6 | ETH_RSS_NONFRAG_IPV6_TCP
- ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4
- ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4_UDP
- ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4_TCP
- ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6
- ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6_UDP
- ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6_TCP
Fixes: 91f27b2e39ab ("net/iavf: refactor RSS")
Signed-off-by: Murphy Yang <murphyx.yang@intel.com>
---
v2:
- add invalid RSS combinations
drivers/net/iavf/iavf_hash.c | 8 ++++++++
1 file changed, 8 insertions(+)
Comments
Hi, murphy
> -----Original Message-----
> From: Murphy Yang <murphyx.yang@intel.com>
> Sent: Friday, December 18, 2020 2:19 PM
> To: dev@dpdk.org
> Cc: Yang, Qiming <qiming.yang@intel.com>; Yang, SteveX
> <stevex.yang@intel.com>; Guo, Jia <jia.guo@intel.com>; Wu, Jingjing
> <jingjing.wu@intel.com>; Xing, Beilei <beilei.xing@intel.com>; Yang,
> MurphyX <murphyx.yang@intel.com>
> Subject: [PATCH v2] net/iavf: fix negative GTP-U flow rules create
> successfully
>
> Currently, when use 'flow' command to create a negative GTP-U rule, it will
> be created successfully.
>
> The list shows the impacted outer and inner 'ipv4' GTP-U patterns with 'ipv4'
> or 'gtpu' type:
> - iavf_pattern_eth_ipv4_gtpu_ipv4_udp
> - iavf_pattern_eth_ipv4_gtpu_eh_ipv4_udp
> - iavf_pattern_eth_ipv4_gtpu_ipv4_tcp
> - iavf_pattern_eth_ipv4_gtpu_eh_ipv4_tcp
> - more impacted patterns with 'gtpu' type:
> > iavf_pattern_eth_ipv4_gtpu_ipv4
> > iavf_pattern_eth_ipv4_gtpu_eh_ipv4
>
> Same as the outer and inner 'ipv6' GTP-U patterns.
>
> So, this commit adds the invalid RSS combinations in 'invalid_rss_comb'
> array to make result correct.
>
> The list of added invalid RSS combinations:
> - ETH_RSS_IPV4 | ETH_RSS_NONFRAG_IPV4_TCP
> - ETH_RSS_IPV6 | ETH_RSS_NONFRAG_IPV6_TCP
> - ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4
> - ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4_UDP
> - ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4_TCP
> - ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6
> - ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6_UDP
> - ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6_TCP
>
Seem that you add some specific negative flow case for the "ipv4(outer/inner) + gtpu" and "ipv4(outer/inner) + ipv4-udp(tcp)" in the currently invalid checking mode, so you commit log could be more clear and clean for that, thanks.
> Fixes: 91f27b2e39ab ("net/iavf: refactor RSS")
>
> Signed-off-by: Murphy Yang <murphyx.yang@intel.com>
> ---
> v2:
> - add invalid RSS combinations
>
> drivers/net/iavf/iavf_hash.c | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/drivers/net/iavf/iavf_hash.c b/drivers/net/iavf/iavf_hash.c index
> c4c73e6644..8393d8535b 100644
> --- a/drivers/net/iavf/iavf_hash.c
> +++ b/drivers/net/iavf/iavf_hash.c
> @@ -806,7 +806,15 @@ static void iavf_refine_proto_hdrs(struct
> virtchnl_proto_hdrs *proto_hdrs,
>
> static uint64_t invalid_rss_comb[] = {
> ETH_RSS_IPV4 | ETH_RSS_NONFRAG_IPV4_UDP,
> + ETH_RSS_IPV4 | ETH_RSS_NONFRAG_IPV4_TCP,
> ETH_RSS_IPV6 | ETH_RSS_NONFRAG_IPV6_UDP,
> + ETH_RSS_IPV6 | ETH_RSS_NONFRAG_IPV6_TCP,
> + ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4,
> + ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4_UDP,
> + ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4_TCP,
> + ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6,
> + ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6_UDP,
> + ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6_TCP,
Shouldn't it be " ETH_RSS_GTPU | ETH_RSS_IPV4" if outer ip case is the same? Whatever you should considerate both if you check one.
> RTE_ETH_RSS_L3_PRE32 | RTE_ETH_RSS_L3_PRE40 |
> RTE_ETH_RSS_L3_PRE48 | RTE_ETH_RSS_L3_PRE56 |
> RTE_ETH_RSS_L3_PRE96
> --
> 2.17.1
@@ -806,7 +806,15 @@ static void iavf_refine_proto_hdrs(struct virtchnl_proto_hdrs *proto_hdrs,
static uint64_t invalid_rss_comb[] = {
ETH_RSS_IPV4 | ETH_RSS_NONFRAG_IPV4_UDP,
+ ETH_RSS_IPV4 | ETH_RSS_NONFRAG_IPV4_TCP,
ETH_RSS_IPV6 | ETH_RSS_NONFRAG_IPV6_UDP,
+ ETH_RSS_IPV6 | ETH_RSS_NONFRAG_IPV6_TCP,
+ ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4,
+ ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4_UDP,
+ ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV4_TCP,
+ ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6,
+ ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6_UDP,
+ ETH_RSS_GTPU | IAVF_RSS_TYPE_INNER_IPV6_TCP,
RTE_ETH_RSS_L3_PRE32 | RTE_ETH_RSS_L3_PRE40 |
RTE_ETH_RSS_L3_PRE48 | RTE_ETH_RSS_L3_PRE56 |
RTE_ETH_RSS_L3_PRE96