Messages in this thread |  | | Subject | Re: [PATCH 00/11] RFC: KBUS messaging subsystem | From | Tony Ibbs <> | Date | Wed, 3 Aug 2011 21:23:19 +0100 |
| |
On 17 May 2011, at 09:50, Florian Fainelli wrote:
> Sorry for this late answer.
And apologies for my own late response. I'll try to keep this short as I hope the "Restatement" side-thread will address some of it.
> On Tuesday 22 March 2011 20:36:40 Jonathan Corbet wrote: >> >> - Why kbus over, say, a user-space daemon and unix-domain sockets? I'm >> not sure I see the advantage that comes with putting this into kernel >> space. > > I also fail to see why this would be required. In my opininon you are trading > the reliability over complexity by putting this in the kernel.
I hope that's addressed in the "So why did we write it as a kernel module?" section of the "Restatement" message thread. Basically, I believe a kernel module is smaller and "steals" reliability from code written and tested by others. That doesn't mean it's a good solution "in the wild", of course (privately we can add whatever we want to the kernel, but in public it is and must be controlled).
> Indeed, I would also suggest having a look at what generic netlink already > provides like messages per application PID, multicasting and marshaling. If > you intend to keep a part of it in the kernel, you should have a look at this, > because from my experience with generic netlink, most of the hard job you are > re-doing here, has already been done in a generic manner.
If we do end up heading that way, I hope you won't mind if I ask you for advice!
All the best, Tibs
|  |