On Wed, Aug 8, 2018 at 10:46 AM, Julia Suvorova <jus...@mail.ru> wrote: > On 08.08.2018 12:09, Stefan Hajnoczi wrote: >> >> On Mon, Aug 6, 2018 at 11:01 AM, Steffen Görtz >> <cont...@steffen-goertz.de> wrote: >>> >>> +#define PAGE_SIZE 1024 >>> +#define FLASH_SIZE (256 * PAGE_SIZE) >>> +#define FLASH_BASE 0x00000000 >>> +#define UICR_BASE 0x10001000 >>> +#define UICR_SIZE 0x100 >>> +#define NVMC_BASE 0x4001E000UL >>> +#define NVMC_READY 0x400 >>> +#define NVMC_CONFIG 0x504 >>> +#define NVMC_ERASEPAGE 0x508 >>> +#define NVMC_ERASEPCR1 0x508 >>> +#define NVMC_ERASEALL 0x50C >>> +#define NVMC_ERASEPCR0 0x510 >>> +#define NVMC_ERASEUICR 0x514 >> >> >> All these constants could live in include/hw/arm/nrf51.h. That way >> tests do not need to duplicate them. As more devices are implemented >> we can expect this list to grow. > > > Maybe it's better to move them to device headers? > include/hw/arm/nrf51.h has to include them in any case to use device > types.
That would be fine. Each device (uart, rng, etc) would have it's own include/hw/... header file and test cases can include that file to get the constants/structs. Stefan
