[dpdk-dev] [PATCH] eventdev: fix build related to port id storage size

[dpdk-dev] [PATCH] eventdev: fix build related to port id storage size

[Ferruh Yigit]

When CONFIG_RTE_MAX_ETHPORTS more than 255 getting following build
error with gcc, gcc (GCC) 7.3.1

.../lib/librte_eventdev/rte_event_eth_rx_adapter.c:
   In function ‘eth_poll_wrr_calc’:
.../build/include/rte_ethdev.h:1328:23:
   error: comparison is always true due to limited range of data type
          [-Werror=type-limits]
       (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \
                       ^
.../build/include/rte_ethdev.h:1345:2:
   note: in expansion of macro ‘RTE_ETH_FOREACH_DEV_OWNED_BY’
  RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER)
  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
.../lib/librte_eventdev/rte_event_eth_rx_adapter.c:220:3:
   note: in expansion of macro ‘RTE_ETH_FOREACH_DEV’
   RTE_ETH_FOREACH_DEV(d) {
   ^~~~~~~~~~~~~~~~~~~

Same for
rte_event_eth_rx_adapter.c:220:3 & rte_event_eth_rx_adapter.c:921:2

The storage type of the variable used in RTE_ETH_FOREACH_DEV is uint8_t
which is causing the build error.

port_id storage type is increased to uint16_t, fixing storage type fixes
the build error.

Fixes: 9c38b704d280 ("eventdev: add eth Rx adapter implementation")
Cc: stable@dpdk.org
Cc: nikhil.rao@intel.com

Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
---
Patch rebased on next-eventdev.

Cc: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Cc: Nikhil Rao <nikhil.rao@intel.com>

There are a few uint8_t storage type on suspicious variable names like
uint8_t eth_dev_id, not sure if this is ethdev port_id. It can be good
to double check storage size.
---
 lib/librte_eventdev/rte_event_eth_rx_adapter.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/librte_eventdev/rte_event_eth_rx_adapter.c b/lib/librte_eventdev/rte_event_eth_rx_adapter.c
index 4c0c02530..9016bb950 100644
--- a/lib/librte_eventdev/rte_event_eth_rx_adapter.c
+++ b/lib/librte_eventdev/rte_event_eth_rx_adapter.c
@@ -190,7 +190,7 @@ wrr_next(struct rte_event_eth_rx_adapter *rx_adapter,
 static int
 eth_poll_wrr_calc(struct rte_event_eth_rx_adapter *rx_adapter)
 {
-	uint8_t d;
+	uint16_t d;
 	uint16_t q;
 	unsigned int i;
 
@@ -859,7 +859,7 @@ rte_event_eth_rx_adapter_create_ext(uint8_t id, uint8_t dev_id,
 	struct rte_event_eth_rx_adapter *rx_adapter;
 	int ret;
 	int socket_id;
-	uint8_t i;
+	uint16_t i;
 	char mem_name[ETH_RX_ADAPTER_SERVICE_NAME_LEN];
 	const uint8_t default_rss_key[] = {
 		0x6d, 0x5a, 0x56, 0xda, 0x25, 0x5b, 0x0e, 0xc2,
-- 
2.14.3

Re: [dpdk-dev] [PATCH] eventdev: fix build related to port id storage size

[Ferruh Yigit]

On 5/10/2018 1:45 AM, Ferruh Yigit wrote:
> When CONFIG_RTE_MAX_ETHPORTS more than 255 getting following build
> error with gcc, gcc (GCC) 7.3.1
> 
> .../lib/librte_eventdev/rte_event_eth_rx_adapter.c:
>    In function ‘eth_poll_wrr_calc’:
> .../build/include/rte_ethdev.h:1328:23:
>    error: comparison is always true due to limited range of data type
>           [-Werror=type-limits]
>        (unsigned int)p < (unsigned int)RTE_MAX_ETHPORTS; \
>                        ^
> .../build/include/rte_ethdev.h:1345:2:
>    note: in expansion of macro ‘RTE_ETH_FOREACH_DEV_OWNED_BY’
>   RTE_ETH_FOREACH_DEV_OWNED_BY(p, RTE_ETH_DEV_NO_OWNER)
>   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
> .../lib/librte_eventdev/rte_event_eth_rx_adapter.c:220:3:
>    note: in expansion of macro ‘RTE_ETH_FOREACH_DEV’
>    RTE_ETH_FOREACH_DEV(d) {
>    ^~~~~~~~~~~~~~~~~~~
> 
> Same for
> rte_event_eth_rx_adapter.c:220:3 & rte_event_eth_rx_adapter.c:921:2
> 
> The storage type of the variable used in RTE_ETH_FOREACH_DEV is uint8_t
> which is causing the build error.
> 
> port_id storage type is increased to uint16_t, fixing storage type fixes
> the build error.
> 
> Fixes: 9c38b704d280 ("eventdev: add eth Rx adapter implementation")
> Cc: stable@dpdk.org
> Cc: nikhil.rao@intel.com
> 
> Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>

Self NAK.

Seems duplication of:
eventdev: change port_id to uint16_t
https://dpdk.org/dev/patchwork/patch/39449/

> ---
> Patch rebased on next-eventdev.
> 
> Cc: Jerin Jacob <jerin.jacob@caviumnetworks.com>
> Cc: Nikhil Rao <nikhil.rao@intel.com>
> 
> There are a few uint8_t storage type on suspicious variable names like
> uint8_t eth_dev_id, not sure if this is ethdev port_id. It can be good
> to double check storage size.
> ---
>  lib/librte_eventdev/rte_event_eth_rx_adapter.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/lib/librte_eventdev/rte_event_eth_rx_adapter.c b/lib/librte_eventdev/rte_event_eth_rx_adapter.c
> index 4c0c02530..9016bb950 100644
> --- a/lib/librte_eventdev/rte_event_eth_rx_adapter.c
> +++ b/lib/librte_eventdev/rte_event_eth_rx_adapter.c
> @@ -190,7 +190,7 @@ wrr_next(struct rte_event_eth_rx_adapter *rx_adapter,
>  static int
>  eth_poll_wrr_calc(struct rte_event_eth_rx_adapter *rx_adapter)
>  {
> -	uint8_t d;
> +	uint16_t d;
>  	uint16_t q;
>  	unsigned int i;
>  
> @@ -859,7 +859,7 @@ rte_event_eth_rx_adapter_create_ext(uint8_t id, uint8_t dev_id,
>  	struct rte_event_eth_rx_adapter *rx_adapter;
>  	int ret;
>  	int socket_id;
> -	uint8_t i;
> +	uint16_t i;
>  	char mem_name[ETH_RX_ADAPTER_SERVICE_NAME_LEN];
>  	const uint8_t default_rss_key[] = {
>  		0x6d, 0x5a, 0x56, 0xda, 0x25, 0x5b, 0x0e, 0xc2,
>