[Ksummit-discuss] [CORE TOPIC] (group) maintainership models

[Laurent Pinchart]

Hi Russell,

On Tuesday 02 Aug 2016 09:41:36 Russell King - ARM Linux wrote:
> On Tue, Aug 02, 2016 at 09:48:53AM +0300, Peter Ujfalusi wrote:
> > I'm not going to move the legacy API under dmaengine, it makes no sense.
> > 
> > My plan is to do what I did with the eDMA driver stack:
> > - convert drivers using the legacy API to dmaengine
> > - 'merge' the code from plat-omap to the dmaengine driver
> >   - while doing this the legacy API will vanish along with the
> >   plat-omap/dma.c
> >
> > As with the eDMA, the sDMA stack will need some cleanup, but the good
> > thing is that the dmaengine driver have minimal dependency on the legacy
> > plat-omap/dma.c API.
> 
> I totally oppose moving arch/arm/plat-omap/dma.c into drivers/dma.
> The omap-dma driver in drivers/dma is a replacement modern driver
> for the legacy crap, and most users of the legacy crap have been
> converted.
> 
> Why we still have users of the legacy crap is because:
> 
> (a) they don't fit into the DMA engine API very well
> (b) I don't have hardware to be able to test them
> 
> and, because of that, I'm not going to start proposing DMA engine API
> extensions to support something that I've no way to even test.
> 
> I've asked for help from TI on this when I was woring on the DMA
> engine implementation, and there was no interest - it kept being
> put back on me as something for me to sort.  It became an impossible
> situation.
> 
> So, how things were left was that I'd introduce a warning into the
> legacy driver, and in a few years time we'd strip it out of users
> and delete the legacy driver.
> 
> The warning (which is rather large, because it's a WARN(), so
> involves a kernel stack dump) has not produced (afaik) any complaints
> from anyone.  So, we can probably conclude that any of these drivers
> which are using the legacy DMA code are probably not being used.

I know of devices that ship with a recent kernel (v4.4 last time I checked) 
patched to remove the warning. We could argue that the vendor should submit 
patches to fix the offending driver(s), but the reality is that they don't 
have enough kernel development experience to do so. This doesn't mean we have 
to fix the problem for them, but I'm afraid we can't conclude this easily that 
those drivers are not used.

> This means we can save ourselves a lot of work and just delete all
> these drivers using the legacy OMAP DMA along with the legacy OMAP
> DMA driver.

-- 
Regards,

Laurent Pinchart