[PATCH v2 09/11] m68k: stmark2: use ioport.h macros for resources

Angelo Dureghello adureghello at baylibre.com
Thu May 14 00:20:38 PDT 2026


Hi Andy,

On 13.05.2026 23:18, Andy Shevchenko wrote:
> On Wed, May 13, 2026 at 11:14:33AM +0200, Angelo Dureghello wrote:
> 
> > Align also other resource declaration using DEFINE_RES_.
> 
> DEFINE_RES_*()
> 
> ...
> 
> >  static struct resource dspi_spi0_resource[] = {
> > -	[0] = {
> > -		.start = MCFDSPI_BASE0,
> > -		.end   = MCFDSPI_BASE0 + 0xFF,
> > -		.flags = IORESOURCE_MEM,
> > -		},
> > -	[1] = {
> > -		.start = 12,
> > -		.end   = 13,
> > -		.flags = IORESOURCE_DMA,
> > -	},
> > -	[2] = {
> > -		.start = MCF_IRQ_DSPI0,
> > -		.end   = MCF_IRQ_DSPI0,
> > -		.flags = IORESOURCE_IRQ,
> > -	},
> > +	DEFINE_RES_MEM(MCFDSPI_BASE0, 0x100),
> > +	DEFINE_RES_DMA(12),
> > +	DEFINE_RES_DMA(13),
> > +	DEFINE_RES_IRQ(MCF_IRQ_DSPI0),
> >  };
> 
> This conversion drops hard indices, would it be a problem (for example,
> some code modifies the fields based on the [hard coded] index...)?
> 
> If not, perhaps good to mention in the commit message and move IRQ to be the
> second one?
>

i can't see any directly-related code accessing these resources by index,
so would not change this if there isn't any risk. 
Also i see DEFINE_RES_() stuff widely used in other arch, so thios make me
think accessing by index is something wrong.

> -- 
> With Best Regards,
> Andy Shevchenko

Regards,
angelo

> 
> 



More information about the linux-arm-kernel mailing list