Re: struct ifnet and ifaddr handling [was: Re: Making global variables of if.c MPSAFE]

[Ryota Ozaki <ozaki-r%netbsd.org@localhost>]

On Wed, Nov 26, 2014 at 11:58 AM, Ryota Ozaki <ozaki-r%netbsd.org@localhost> wrote:
> On Wed, Nov 26, 2014 at 11:24 AM, Masao Uebayashi <uebayasi%gmail.com@localhost> wrote:
>> On Wed, Nov 26, 2014 at 1:31 AM, Ryota Ozaki <ozaki-r%netbsd.org@localhost> wrote:
>>>> I'd change if__pad1 to if_flags2 and define IFF2_SLOWTIMO_MPSAFE etc.
>>>
>>> Hmm, I think it's a last resort. I don't want to do something like
>>> (ifp->if_flags & IFF_XXX) && (ifp->if_flags2 & IFF_XXX2) if possible.
>>>
>>> I'm thinking more to avoid the separation.
>>
>> It's a temporary measure until a good struct ifnet will be designed.
>
> If it's a temporal solution, I have no objection :)
>
>>
>> I propose to keep ABI by reusing if__pad1.  If you propose ABI
>> breakage, you should understand it and estimate the impact by
>> yourself. :)
>
> Of course, I'm thinking a way without ABI breakage.
>
>   ozaki-r

Anyway MPSAFE-ification is a separate task that we can do later.

Let's finish callout per interface first :)

Here is a latest patch:
http://www.netbsd.org/~ozaki-r/watchdog-callout-per-if.diff

So my remaining concern is whether we can embed callout_t with _KERNEL
which I proposed earlier. If the answer is no, the above patch that
uses a pointer of struct callout is okay for me.

  ozaki-r