Re: [PATCH] mtd: nand: MTD_NAND_MARVELL should depend on HAS_DMA

[Boris Brezillon]

On Tue, 30 Jan 2018 14:46:47 +0100
Miquel Raynal <miquel.raynal@xxxxxxxxxxxxxxxxxx> wrote:

> Hi Geert,
> 
> On Tue, 30 Jan 2018 14:23:21 +0100
> Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote:
> 
> > If NO_DMA=y:
> > 
> >     ERROR: "bad_dma_ops" [drivers/mtd/nand/marvell_nand.ko] undefined!
> > 
> > Add a dependency on HAS_DMA to fix this.
> > 
> > Fixes: 02f26ecf8c772751 ("mtd: nand: add reworked Marvell NAND controller driver")
> > Signed-off-by: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
> > ---
> >  drivers/mtd/nand/Kconfig | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/drivers/mtd/nand/Kconfig b/drivers/mtd/nand/Kconfig
> > index e6b8c59f2c0da7c0..736ac887303c88ba 100644
> > --- a/drivers/mtd/nand/Kconfig
> > +++ b/drivers/mtd/nand/Kconfig
> > @@ -328,7 +328,7 @@ config MTD_NAND_MARVELL
> >  	tristate "NAND controller support on Marvell boards"
> >  	depends on PXA3xx || ARCH_MMP || PLAT_ORION || ARCH_MVEBU || \
> >  		   COMPILE_TEST
> > -	depends on HAS_IOMEM
> > +	depends on HAS_IOMEM && HAS_DMA  
> 
> I think this is more a coding issue than a Kconfig issue. AFAIR, none
> of the ARCH_MVEBU SoCs use DMA in conjunction with the NAND controller.
> PXA SoCs may use DMA (NFCv1 only) but this is not a hard requirement and
> the driver is supposed to support PIO mode in all situations.
> 
> Can you please share your setup, the SoC, the configuration,
> maybe also the DT?

I don't think Geert is testing on a real board. He's just enabling
COMPILE_TEST and can then select your driver. Which means you really
need this depends on HAS_DMA.