[dpdk-stable] [dpdk-dev] [PATCH] power: remove duplicated symbols from map file

[Ferruh Yigit]

On 2/25/2021 9:11 AM, David Marchand wrote:
> Hello Ferruh,
> 
> On Wed, Feb 24, 2021 at 3:55 PM Ferruh Yigit <ferruh.yigit at intel.com> wrote:
>>
>> This is also causing build error, like:
>> https://travis-ci.com/github/ovsrobot/dpdk/jobs/482121104
>>
>> Also '@internal' marker removed from doxygen comment, since public API
>> should not be internal.
>>
>> Fixes: 4d3892dcd77b ("power: make channel message functions public")
>> Cc: stable at dpdk.org
>>
>> Signed-off-by: Ferruh Yigit <ferruh.yigit at intel.com>
>> ---
>> Cc: Aaron Conole <aconole at redhat.com>
>> Cc: Bruce Richardson <bruce.richardson at intel.com>
>> Cc: David Marchand <david.marchand at redhat.com>
>> Cc: Anatoly Burakov <anatoly.burakov at intel.com>
>> Cc: Juraj Linkeš <juraj.linkes at pantheon.tech>
>> ---
>>   lib/librte_power/rte_power_guest_channel.h | 4 ----
>>   lib/librte_power/version.map               | 2 --
>>   2 files changed, 6 deletions(-)
>>
>> diff --git a/lib/librte_power/rte_power_guest_channel.h b/lib/librte_power/rte_power_guest_channel.h
>> index ed4fbfdcd38a..7586e9491f11 100644
>> --- a/lib/librte_power/rte_power_guest_channel.h
>> +++ b/lib/librte_power/rte_power_guest_channel.h
>> @@ -119,8 +119,6 @@ struct rte_power_channel_packet_caps_list {
>>   };
>>
>>   /**
>> - * @internal
>> - *
>>    * @warning
>>    * @b EXPERIMENTAL: this API may change without prior notice.
>>    *
>> @@ -141,8 +139,6 @@ int rte_power_guest_channel_send_msg(struct rte_power_channel_packet *pkt,
>>                          unsigned int lcore_id);
>>
>>   /**
>> - * @internal
>> - *
>>    * @warning
>>    * @b EXPERIMENTAL: this API may change without prior notice.
>>    *
>> diff --git a/lib/librte_power/version.map b/lib/librte_power/version.map
>> index 3ba9390241d2..bce5110dd5b4 100644
>> --- a/lib/librte_power/version.map
>> +++ b/lib/librte_power/version.map
>> @@ -12,7 +12,6 @@ DPDK_21 {
>>          rte_power_get_capabilities;
>>          rte_power_get_env;
>>          rte_power_get_freq;
>> -       rte_power_guest_channel_send_msg;
> 
> This symbol was marked stable so far.
> Did I miss something?
> I would just remove the experimental tagging for this symbol which is
> incorrect in v21.02.
> 

You are right,

The header that has the declares the API was not public before the v21.02, so I 
assumed the symbol is internal and added to the stable section by mistake, but 
the intention seems to have it as public API.

I will keep the 'rte_power_guest_channel_send_msg' API, but 
'rte_power_guest_channel_receive_msg' is already experimental, so will keep it 
as it is.

> Something like:
> diff --git a/lib/librte_power/rte_power_guest_channel.h
> b/lib/librte_power/rte_power_guest_channel.h
> index ed4fbfdcd3..b5de1bd243 100644
> --- a/lib/librte_power/rte_power_guest_channel.h
> +++ b/lib/librte_power/rte_power_guest_channel.h
> @@ -119,11 +119,6 @@ struct rte_power_channel_packet_caps_list {
>   };
> 
>   /**
> - * @internal
> - *
> - * @warning
> - * @b EXPERIMENTAL: this API may change without prior notice.
> - *
>    * Send a message contained in pkt over the Virtio-Serial to the host endpoint.
>    *
>    * @param pkt
> @@ -136,13 +131,10 @@ struct rte_power_channel_packet_caps_list {
>    *  - 0 on success.
>    *  - Negative on error.
>    */
> -__rte_experimental
>   int rte_power_guest_channel_send_msg(struct rte_power_channel_packet *pkt,
>                          unsigned int lcore_id);
> 
>   /**
> - * @internal
> - *
>    * @warning
>    * @b EXPERIMENTAL: this API may change without prior notice.
>    *
> diff --git a/lib/librte_power/version.map b/lib/librte_power/version.map
> index 3ba9390241..b004e3e4a9 100644
> --- a/lib/librte_power/version.map
> +++ b/lib/librte_power/version.map
> @@ -38,6 +38,4 @@ EXPERIMENTAL {
>          # added in 21.02
>          rte_power_ethdev_pmgmt_queue_disable;
>          rte_power_ethdev_pmgmt_queue_enable;
> -       rte_power_guest_channel_receive_msg;
> -       rte_power_guest_channel_send_msg;
>   };
> 
> 
> 
> Sidenote, we can probably catch these duplicates during build.
> For duplicate entries with the same version, this is easy.
> I'll see if I can catch the other case too.
> 

Wasn't there a script for it already, I slightly remember it but not sure, cc'ed 
Thomas, perhaps he remembers.