On Mon, Mar 09, 2015 at 12:22:09PM +0300, Dan Carpenter wrote:
> On Sun, Mar 08, 2015 at 06:01:23PM +0530, Sudip Mukherjee wrote:
> > diff --git a/drivers/staging/sm750fb/sm750.h b/drivers/staging/sm750fb/sm750.h
> > index 711676c..2ab7b74 100644
> > --- a/drivers/staging/sm750fb/sm750.h
> > +++ b/drivers/staging/sm750fb/sm750.h
> > @@ -59,10 +59,10 @@ struct lynx_share{
> > }mtrr;
> > #endif
> > /* all smi graphic adaptor got below attributes */
> > - resource_size_t vidmem_start;
> > - resource_size_t vidreg_start;
> > - resource_size_t vidmem_size;
> > - resource_size_t vidreg_size;
> > + unsigned long vidmem_start;
> > + unsigned long vidreg_start;
> > + unsigned long vidmem_size;
> > + unsigned long vidreg_size;
> > volatile unsigned char __iomem * pvReg;
> > unsigned char __iomem * pvMem;
> > /* locks*/
>
> This seems like a very risky sort of change. It's not explained very
> well in the changelog. What's the deal?
it mainly started with the build warnings of incorrect format specifier.
v2 gives a little more details in the comments section. copying that here for your convenience:
"V2: Giedrius commented resource_size_t can be either u64 or u32
depending on if CONFIG_PHYS_ADDR_T_64BIT. based on his comments i
should have kept the datatype as resource_size_t and used %pa as the
format specifier. But since this is a framebuffer device and it
should follow what the framebuffer layer is suggesting in
struct fb_fix_screeninfo at smem_start and mmio_start."
so accoringly, like all other framebuffer devices, vidmem_start and vidreg_start should be unsigned long and vidmem_size and vidreg_size should be __u32.
regards
sudip
>
> regards,
> dan carpenter
>
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
