ok for me

On 5 January 2017 at 13:19, Francois Ozog <francois.o...@linaro.org> wrote:

> Monday is my busiest day (last call ends 10pm) but we can do a morning
> call: 11amCET?
>
> On 5 January 2017 at 12:28, Christophe Milard <
> christophe.mil...@linaro.org> wrote:
>
>> Hi Francois,
>>
>> Can we have a HO on monday ( I understood today was busy for you, FF and
>> I tomorrow is a swedish national holiday). If not, we can of course take in
>> on tuesday.
>>
>> The topics I'd like to discuss are:
>> 1) change of interface: pass a struct <object>_param_t at register time
>> and get a <object_t> handle as a result.
>> I Hoped we could simplify that, but Petri has now enforced this on the
>> south interface as well (cf buddy allocator patch) and now we have an
>> inconsistant situation: I hence suggest to take this same approach at
>> register time as well.
>>
>> 2) The callback functions dicussed above (I am not 100% following you,
>> FF).
>>
>> 3) what should be done on HW removal: do we expect enumerators to
>> "disregister"/ be removed. I am assuming a case where we have device
>> dependant enumerator
>>
>> Christophe.
>>
>> On 5 January 2017 at 11:43, Francois Ozog <francois.o...@linaro.org>
>> wrote:
>>
>>> Hi Christophe,
>>>
>>> event are not just "new device" (in that case I would agree with you).
>>> Events can be:
>>> - "attention" (before "poweroff") for a specific device to orderly
>>> shutdown before poweroff (or remove if virtio-net device in a VM).
>>> - SFP insertion/removal of a specific port
>>> - Error condition (an internal watchdog may detect packet stall or I
>>> don't know)
>>>
>>> FF
>>>
>>> On 5 January 2017 at 09:13, Christophe Milard <
>>> christophe.mil...@linaro.org> wrote:
>>>
>>>> Hi Francois,
>>>>
>>>> Following your suggestion, the driver API contains:
>>>>      /** Register event notifier function for hotplug events:
>>>>          */
>>>>         int (*register_notifier)(void (*event_handler) (void));
>>>>
>>>> But I an now wondering: Why this double callback?
>>>>
>>>> Wouldn't it be simpler to have the functions:
>>>>
>>>> odpdrv_enum_class_probe_request(): called by a enum class when it
>>>> needs to refresh its enumerator list.
>>>> (ODP would eventually call the enumerator class probe function after
>>>> recieving the probe request), ... and...
>>>>
>>>> odpdrv_enum_probe_request(): called by an enumerator when it needs to
>>>> refresh its device list. (ODP would eventually call the enumerator
>>>> probe function after recieving the probe request)
>>>>
>>>> Christophe
>>>>
>>>
>>>
>>>
>>> --
>>> [image: Linaro] <http://www.linaro.org/>
>>> François-Frédéric Ozog | *Director Linaro Networking Group*
>>> T: +33.67221.6485
>>> francois.o...@linaro.org | Skype: ffozog
>>>
>>>
>>
>
>
> --
> [image: Linaro] <http://www.linaro.org/>
> François-Frédéric Ozog | *Director Linaro Networking Group*
> T: +33.67221.6485
> francois.o...@linaro.org | Skype: ffozog
>
>

Reply via email to