Hi Mason, On Thu, Jan 17, 2019 at 7:32 AM <masonccyang@xxxxxxxxxxx> wrote: > > "Sergei Shtylyov" <sergei.shtylyov@xxxxxxxxxxxxxxxxxx> > > 2019/01/17 上午 03:36 > > >> > > >> [...] > > >> > +#define RPC_WBUF 0x8000 // Write Buffer > > >> > +#define RPC_WBUF_SIZE 256 // Write Buffer size > > >> > > >> I wonder if the write buffer should be in a separate "reg" prop tuple... > > >> Have you considered that? > > >> > > > > > > I don't get your point! > > > > I mean that the write buffer should be a separate "reg" property > > address/size tuple, > > so that the RPC device has 3 memory resources. Our SPI driver used > > this scheme, and I > > like it. > > > > I got your point but I think this RPC_WBUF@0xEE20_8000 is in the same group > and it's size only 256 bytes, not like external address space read mode > @0x0800_0000 w/ 64K bytes. > > Maybe Geert or Marek could comment on it, thanks. Given the writer buffer is separate from the other registers (it's in a different 4 KiB page, and thus may be at a different offset on future SoCs), and not present on RZ/A1, I think it's a good idea to use a separate reg entry for it. Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds