On 10/29/20 3:18 PM, wangyunjian wrote:
>> -----Original Message-----
>> From: Thomas Monjalon [mailto:tho...@monjalon.net]
>> Sent: Tuesday, October 27, 2020 10:29 PM
>> To: wangyunjian <wangyunj...@huawei.com>
>> Cc: dev@dpdk.org; ferruh.yi...@intel.com; andrew.rybche...@oktetlabs.ru;
>> Lilijun (Jerry) <jerry.lili...@huawei.com>; xudingke <xudin...@huawei.com>;
>> wangyunjian <wangyunj...@huawei.com>; sta...@dpdk.org
>> Subject: Re: [dpdk-dev] [PATCH v2] ethdev: fix data type for port id
>>
>> 27/10/2020 14:41, wangyunjian:
>>> From: Yunjian Wang <wangyunj...@huawei.com>
>>>
>>> The ethdev port id should be 16 bits now. This patch fixes the data
>>> type of the variable for 'pid', changing from uint32_t to uint16_t.
>>>
>>> We also need use RTE_BUILD_BUG_ON() to ensure that
>> RTE_MAX_ETHPORTS is
>>> less or equal to UINT16_MAX.
>>
>> Actually the need is to check that we have room for an increment after
>> RTE_MAX_ETHPORTS, meaning RTE_MAX_ETHPORTS < UINT16_MAX.
> 
> OK.
> 
>>
>>> Fixes: 5b7ba31148a8 ("ethdev: add port ownership")
>>> Cc: sta...@dpdk.org
>>>
>>> Signed-off-by: Yunjian Wang <wangyunj...@huawei.com>
>>> ---
>>> v2:
>>>    add RTE_BUILD_BUG_ON() check for RTE_MAX_ETHPORTS validity
>>> ---
>>>  rte_eth_dev_get_port_by_name(const char *name, uint16_t *port_id)  {
>>> -   uint32_t pid;
>>> +   uint16_t pid;
>>>
>>>     if (name == NULL) {
>>>             RTE_ETHDEV_LOG(ERR, "Null pointer is specified\n"); @@ -4292,6
>>> +4292,8 @@ RTE_INIT(eth_dev_init_cb_lists)  {
>>>     int i;
>>>
>>> +   RTE_BUILD_BUG_ON(RTE_MAX_ETHPORTS > UINT16_MAX);
>>
>> As explained above, should check >=
>>
>> This check is global for ethdev.
>> Please could you move it somewhere else?
>>
>>
> 
> What about following:
> 
> # git diff
> diff --git a/lib/librte_eal/freebsd/eal.c b/lib/librte_eal/freebsd/eal.c
> index ccea60afe7..ff878f4d4d 100644
> --- a/lib/librte_eal/freebsd/eal.c
> +++ b/lib/librte_eal/freebsd/eal.c
> @@ -672,6 +672,8 @@ rte_eal_init(int argc, char **argv)
>         struct internal_config *internal_conf =
>                 eal_get_internal_configuration();
>  
> +       RTE_BUILD_BUG_ON(RTE_MAX_ETHPORTS >= UINT16_MAX);
> +
>         /* checks if the machine is adequate */
>         if (!rte_cpu_is_supported()) {
>                 rte_eal_init_alert("unsupported cpu type.");
> diff --git a/lib/librte_eal/linux/eal.c b/lib/librte_eal/linux/eal.c
> index 9b579b8200..bd7fb7b5f1 100644
> --- a/lib/librte_eal/linux/eal.c
> +++ b/lib/librte_eal/linux/eal.c
> @@ -970,6 +970,8 @@ rte_eal_init(int argc, char **argv)
>         struct internal_config *internal_conf =
>                 eal_get_internal_configuration();
>  
> +       RTE_BUILD_BUG_ON(RTE_MAX_ETHPORTS >= UINT16_MAX);
> +
>         /* checks if the machine is adequate */
>         if (!rte_cpu_is_supported()) {
>                 rte_eal_init_alert("unsupported cpu type.");
> diff --git a/lib/librte_eal/windows/eal.c b/lib/librte_eal/windows/eal.c
> index 141f22adb7..27c2ebd114 100644
> --- a/lib/librte_eal/windows/eal.c
> +++ b/lib/librte_eal/windows/eal.c
> @@ -265,6 +265,8 @@ rte_eal_init(int argc, char **argv)
>         struct internal_config *internal_conf =
>                 eal_get_internal_configuration();
>  
> +       RTE_BUILD_BUG_ON(RTE_MAX_ETHPORTS >= UINT16_MAX);
> +
>         rte_eal_log_init(NULL, 0);
> 

If RTE_MAX_ETHPORTS is a maximum number of ports starting from
0 and loops are port_id < RTE_MAX_ETHPORTS, RTE_MAX_ETHPORTS
may be UINT16_MAX and it avoid 16-bit unsigned integer
overflow, but may be it is safer to cut one more port from
space do checks as above.

Reply via email to