On 1/12/2019 4:49 AM, Hyong Youb Kim wrote:
> On Fri, Jan 11, 2019 at 03:46:47PM +0000, Ferruh Yigit wrote:
>> On 12/10/2018 6:28 PM, Hyong Youb Kim wrote:
>>> The VIC hardware has 64 MAC filters per vNIC, which can be either
>>> unicast or multicast. Use one half for unicast and the other half for
>>> multicast, as the VIC kernel drivers for Linux and Windows do.
>>>
>>> Signed-off-by: Hyong Youb Kim <hyon...@cisco.com>
>>> Reviewed-by: John Daley <johnd...@cisco.com>
>>
>> <...>
>>
>>> +static void debug_log_add_del_addr(struct ether_addr *addr, bool add)
>>> +{
>>> + char mac_str[ETHER_ADDR_FMT_SIZE];
>>> + if (rte_log_get_level(enicpmd_logtype_init) == RTE_LOG_DEBUG) {
>>> + ether_format_addr(mac_str, ETHER_ADDR_FMT_SIZE, addr);
>>> + PMD_INIT_LOG(ERR, " %s address %s\n",
>>> + add ? "add" : "remove", mac_str);
>>> + }
>>> +}
>>
>> Why logging with 'ERR' level after checking if 'DEBUG' level is set?
>
> Thanks for pointing this out. Our mistake. Should be DEBUG.
>
>> And why need that check at all?
>
> The outer check ("is log level debug?") is there to make this function
> to do nothing (including ether_format_addr) if log level > debug. I
> could not find a good way to combine the level test,
> ether_format_addr, and rte_log into one clean macro..
>
> Since this function is on a slow path, I think I can just remove the
> level check and do something like the following. Would this satisfy
> your concern?
>
> char mac_str[ETHER_ADDR_FMT_SIZE];
> ether_format_addr(mac_str, ETHER_ADDR_FMT_SIZE, addr);
> PMD_INIT_LOG(DEBUG, " %s address %s\n", add ? "add" : "remove", mac_str);
Yes, I think this is better, thanks.
