Re: [dpdk-dev] [PATCH] librte_net: define TCP flags in rte_tcp.h instead of in rte_eth_ctrl.h

[Olivier Matz <olivier.matz@6wind.com>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="UTF-8", Size: 5471 bytes --]

Hi Morten,

On Fri, May 10, 2019 at 07:34:14PM +0200, Morten Brørup wrote:
> TCP flags were moved to the TCP header file from the Ethernet control
> header file, keeping their names intact.
> 
> Missing TCP ECN flags were added.
> 
> The ALL mask did not include TCP ECN flags, so it was renamed to reflect
> that it applies to N-tuple filtering only.
> A driver using the ALL mask was updated accordingly.
> 
> Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
> ---
>  app/test-pmd/cmdline.c           |    4 ++--
>  drivers/net/e1000/igb_ethdev.c   |    8 ++++----
>  lib/librte_ethdev/rte_eth_ctrl.h |    8 +-------
>  lib/librte_net/rte_tcp.h         |   13 +++++++++++++
>  4 files changed, 20 insertions(+), 13 deletions(-)
> 
> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
> index ee50e45..59a45c1 100644
> --- a/app/test-pmd/cmdline.c
> +++ b/app/test-pmd/cmdline.c
> @@ -9983,7 +9983,7 @@ cmd_2tuple_filter_parsed(void *parsed_result,
>  			" when protocol is TCP.\n");
>  		return;
>  	}
> -	if (res->tcp_flags_value > TCP_FLAG_ALL) {
> +	if (res->tcp_flags_value > RTE_NTUPLE_TCP_FLAGS_MASK) {
>  		printf("invalid TCP flags.\n");
>  		return;
>  	}
> @@ -10141,7 +10141,7 @@ cmd_5tuple_filter_parsed(void *parsed_result,
>  			" when protocol is TCP.\n");
>  		return;
>  	}
> -	if (res->tcp_flags_value > TCP_FLAG_ALL) {
> +	if (res->tcp_flags_value > RTE_NTUPLE_TCP_FLAGS_MASK) {
>  		printf("invalid TCP flags.\n");
>  		return;
>  	}
> diff --git a/drivers/net/e1000/igb_ethdev.c b/drivers/net/e1000/igb_ethdev.c
> index b897e8a..c2fe19a 100644
> --- a/drivers/net/e1000/igb_ethdev.c
> +++ b/drivers/net/e1000/igb_ethdev.c
> @@ -3702,7 +3702,7 @@ ntuple_filter_to_2tuple(struct rte_eth_ntuple_filter *filter,
>  		return -EINVAL;
>  	if (filter->priority > E1000_2TUPLE_MAX_PRI)
>  		return -EINVAL;  /* filter index is out of range. */
> -	if (filter->tcp_flags > TCP_FLAG_ALL)
> +	if (filter->tcp_flags > RTE_NTUPLE_TCP_FLAGS_MASK)
>  		return -EINVAL;  /* flags is invalid. */
>  
>  	switch (filter->dst_port_mask) {
> @@ -3782,7 +3782,7 @@ igb_inject_2uple_filter(struct rte_eth_dev *dev,
>  		ttqf &= ~E1000_TTQF_MASK_ENABLE;
>  
>  	/* tcp flags bits setting. */
> -	if (filter->filter_info.tcp_flags & TCP_FLAG_ALL) {
> +	if (filter->filter_info.tcp_flags & RTE_NTUPLE_TCP_FLAGS_MASK) {
>  		if (filter->filter_info.tcp_flags & TCP_URG_FLAG)
>  			imir_ext |= E1000_IMIREXT_CTRL_URG;
>  		if (filter->filter_info.tcp_flags & TCP_ACK_FLAG)
> @@ -4188,7 +4188,7 @@ ntuple_filter_to_5tuple_82576(struct rte_eth_ntuple_filter *filter,
>  		return -EINVAL;
>  	if (filter->priority > E1000_2TUPLE_MAX_PRI)
>  		return -EINVAL;  /* filter index is out of range. */
> -	if (filter->tcp_flags > TCP_FLAG_ALL)
> +	if (filter->tcp_flags > RTE_NTUPLE_TCP_FLAGS_MASK)
>  		return -EINVAL;  /* flags is invalid. */
>  
>  	switch (filter->dst_ip_mask) {
> @@ -4318,7 +4318,7 @@ igb_inject_5tuple_filter_82576(struct rte_eth_dev *dev,
>  	imir |= filter->filter_info.priority << E1000_IMIR_PRIORITY_SHIFT;
>  
>  	/* tcp flags bits setting. */
> -	if (filter->filter_info.tcp_flags & TCP_FLAG_ALL) {
> +	if (filter->filter_info.tcp_flags & RTE_NTUPLE_TCP_FLAGS_MASK) {
>  		if (filter->filter_info.tcp_flags & TCP_URG_FLAG)
>  			imir_ext |= E1000_IMIREXT_CTRL_URG;
>  		if (filter->filter_info.tcp_flags & TCP_ACK_FLAG)
> diff --git a/lib/librte_ethdev/rte_eth_ctrl.h b/lib/librte_ethdev/rte_eth_ctrl.h
> index 5ea8ae2..f21e84b 100644
> --- a/lib/librte_ethdev/rte_eth_ctrl.h
> +++ b/lib/librte_ethdev/rte_eth_ctrl.h
> @@ -184,13 +184,7 @@ struct rte_eth_syn_filter {
>  		RTE_NTUPLE_FLAGS_DST_PORT | \
>  		RTE_NTUPLE_FLAGS_PROTO)
>  
> -#define TCP_URG_FLAG 0x20
> -#define TCP_ACK_FLAG 0x10
> -#define TCP_PSH_FLAG 0x08
> -#define TCP_RST_FLAG 0x04
> -#define TCP_SYN_FLAG 0x02
> -#define TCP_FIN_FLAG 0x01
> -#define TCP_FLAG_ALL 0x3F
> +#define RTE_NTUPLE_TCP_FLAGS_MASK 0x3F /**< TCP flags filter can match. */
>  
>  /**
>   * A structure used to define the ntuple filter entry
> diff --git a/lib/librte_net/rte_tcp.h b/lib/librte_net/rte_tcp.h
> index 91f5898..2e89b5e 100644
> --- a/lib/librte_net/rte_tcp.h
> +++ b/lib/librte_net/rte_tcp.h
> @@ -2,6 +2,7 @@
>   * Copyright(c) 1982, 1986, 1990, 1993
>   *      The Regents of the University of California.
>   * Copyright(c) 2010-2014 Intel Corporation.
> + * Copyright(c) 2019 SmartShare Systems.
>   * All rights reserved.
>   */
>  

IMHO, adding the copyright for flags move is a bit overkill.

> @@ -35,6 +36,18 @@ struct tcp_hdr {
>  	uint16_t tcp_urp;   /**< TCP urgent pointer, if any. */
>  } __attribute__((__packed__));
>  
> +/**
> + * TCP Flags
> + */
> +#define TCP_CWR_FLAG 0x80 /**< Congestion Window Reduced */
> +#define TCP_ECE_FLAG 0x40 /**< ECN-Echo */
> +#define TCP_URG_FLAG 0x20 /**< Urgent Pointer field significant */
> +#define TCP_ACK_FLAG 0x10 /**< Acknowledgment field significant */
> +#define TCP_PSH_FLAG 0x08 /**< Push Function */
> +#define TCP_RST_FLAG 0x04 /**< Reset the connection */
> +#define TCP_SYN_FLAG 0x02 /**< Synchronize sequence numbers */
> +#define TCP_FIN_FLAG 0x01 /**< No more data from sender */
> +
>  #ifdef __cplusplus
>  }
>  #endif

Moving these flags in rte_tcp.h makes sense.

But I think they will need to be renamed with a RTE_ prefix inside this
patchset: https://patches.dpdk.org/patch/52556/
So we need to take care to the dependency between these 2 patchsets.

Acked-by: Olivier Matz <olivier.matz@6wind.com>