svn commit: r290003 - head/sys/ofed/include/linux

[Bryan Drewery]

> On Oct 30, 2015, at 03:15, Gary Jennejohn <gljennjohn at gmail.com> wrote:
> 
> On Fri, 30 Oct 2015 09:04:36 +0100
> Hans Petter Selasky <hps at selasky.org> wrote:
> 
>>> On 10/30/15 08:41, Gary Jennejohn wrote:
>>> On Thu, 29 Oct 2015 16:28:51 +0100
>>> Hans Petter Selasky <hps at selasky.org> wrote:
>>> 
>>>> On 10/29/15 15:36, Gleb Smirnoff wrote:  
>>>>>> The LinuxKPI is not a binary compatibility module, and will at some
>>>>> H> point have API's diverging from Linux, to fit BSD API's better.  
>>>>> 
>>>>> This statement makes the name of LinuxKPI quite pointless, as well
>>>>> as the whole idea of the KPI unclear.  
>>>> 
>>>> Hi,
>>>> 
>>>> To be more clear. Adding bind_irq_to_cpu() is more an exception than the
>>>> default. A the moment I think Linux doesn't have an equivalent of this
>>>> function, because of Linux's interrupt model.
>>> 
>>> My question is whether a "normal" FreeBSD user has any reason to
>>> enable LinuxKPI now or in the future.
>> 
>> Hi,
>> 
>> If drivers which depend on this feature are KLD's there's no reason to 
>> enable this by default in GENERIC. The current and future clients of 
>> LINUXKPI will possibly be KLD's and then MODULE_DEPEND() will do the 
>> magic behind the scenes. Was this your question?
> 
> Well, I guess the answer is "no", if I understand you correctly.
> 
> 

Only device driver developers need to think about this. Users don't. It's purely to make development simpler. If a user uses such a driver it will load this module automatically. It won't let anyone load a Linux driver, it is only for FreeBSD drivers using Linux-like code to make porting and maintenance simpler.