On 16 February 2018 at 02:29, Philippe Mathieu-Daudé <f4...@amsat.org> wrote:
> using the sdbus_*() API.
> Signed-off-by: Philippe Mathieu-Daudé <f4...@amsat.org>
> ---
> RFC because how pl181_sdbus_create_inplace() doing class_init(SDBus) in
> realize(pl181) seems weird...

In the two other places where we set the set_inserted and set_readonly
methods for the SDBus class (pxa2xx_mmci.c and sdhci.c), we do it by
defining a subclass of TYPE_SD_BUS, whose class init function can then
set up the method pointers. Is there a reason not to do pl181 the
same way as those two examples?

-- PMM

Reply via email to