On Sat, Sep 17, 2016 at 09:47:12PM +0200, Arnd Bergmann wrote: > On Friday, September 16, 2016 8:15:05 PM CEST Vinod Koul wrote: > > > > > > IMHO it should be as simple as: > > > > > > bool "Foo DMA support" > > > - depends on ARCH_FOO > > > + depends on ARCH_FOO || COMPILE_TEST > > > > > > and allmodconfig will enable the driver for compile testing... > > > > Yeah that was my first thought, but we got a better suggestion from Arnd and > > I do consider DMAengines to be core which should be selected by default. > > > > Yes I should have _explictly_ mentioned that in the cover letter, my bad. > > I noticed now that some of the drivers are 'tristate' and currently > can be built as loadable modules, which makes sense if all drivers > using the dma engine are also modules. In those cases, the method > that Geert suggested might be better, to avoid forcing the symbol > to be built-in. > > There may also be cases of drivers that are only used on one platform > and that are completely optional, i.e. all drivers that can use it > are for stuff that either has an MMIO mode fallback or that users > can also reasonably disable. I don't know of any such examples right > now, so not sure if that is actually the case or just in theory. Yeah even I was thinking about changing it. I will update the patchset and post updated one in a day or two Thanks -- ~Vinod -- To unsubscribe from this list: send the line "unsubscribe dmaengine" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
