Messages in this thread | | | Date | Fri, 22 Oct 2010 19:03:17 -0400 | From | "Emilio G. Cota" <> | Subject | Re: [-next] staging/vme: various fixes + new driver model for VME |
| |
On Fri, Oct 22, 2010 at 10:26:11 +0100, Martyn Welch wrote: > On 22/10/10 07:36, Emilio G. Cota wrote: > > Hi Emilio, > > Thank you for the fixes. After a quick glance, there seem to be a number > of valid fixes here, but I'm very concerned by the patches that change > the driver model. We discussed this approach in August last year, I am > still yet to be convinced by the approach you wish to take.
Yes I remember we discussed it privately. I didn't provide a solution at that point, I just told you that I believed the model could be improved--and this is my attempt.
> > > > - Make DMA work on the tsi148 (it's the only bridge I've got). > > This will probably involve changing or extending the > > current API. > > The DMA is already working on the tsi148. In what way do you feel that > the current API needs changing or extending for DMA?
Sorry I may be confused here. That's the impression I got after reading the DMA code a few months ago.
Anyway forget about this for the time being.
> > - Test the whole thing with real hardware and a real VME driver > > (currently out of tree as well, see [2]), which I'll try > > to get merged, too--currently we just have vme_user.c which > > really isn't a kosher driver. > > > > The current API has been tested with real hardware, for both supported > vme bridges, on multiple cards, by multiple people. If the changes to > the API are to be applied, they would need to be throughly tested before > they are applied. As I've said above - I am still not convinced by the > change in approach.
Well what we have merged is just vme_user, which is an unusual driver that is not even finished yet. I can't speak of what's not merged.
Don't get me wrong, I didn't mean to say that the current code has never been tested.
What I wanted to say is that I could eventually merge a driver, and that could validate the whole thing. However I wouldn't like to port everything, test it, and then get a NAK because of the changed driver model. So that's why I'm sending this first--I don't mind if it's not yet merged, but at least I'd like to know if I'm on the right track.
Thanks
Emilio
| |