On Sat, Feb 12, 2011 at 12:40 AM, Michael Walle <mich...@walle.cc> wrote:
> Am Freitag 11 Februar 2011, 21:52:17 schrieb Blue Swirl:
>> > +static inline void hwsetup_add_uart(struct hwsetup *hw,
>> > + const char *name, uint32_t base, uint32_t irq)
>> > +{
>> > + hwsetup_add_u32(hw, 56); /* size */
>> > + hwsetup_add_tag(hw, HWSETUP_TAG_UART);
>> > + hwsetup_add_str(hw, name);
>> > + hwsetup_add_u32(hw, base);
>> > + hwsetup_add_u32(hw, 115200); /* baudrate */
>> > + hwsetup_add_u8(hw, 8); /* databits */
>> > + hwsetup_add_u8(hw, 1); /* stopbits */
>> > + hwsetup_add_u8(hw, 1); /* use_interrupt */
>> > + hwsetup_add_u8(hw, 1); /* block_on_transmit */
>> > + hwsetup_add_u8(hw, 1); /* block_on_receive */
>> > + hwsetup_add_u8(hw, 4); /* rx_buffer_size */
>> > + hwsetup_add_u8(hw, 4); /* tx_buffer_size */
>> > + hwsetup_add_u8(hw, irq);
>> > +}
>>
>> Overall, this seems to replicate fw_cfg functions, does it match real
>> HW or is this just for QEMU?
>
> It is some kind of (proprietary) device tree for the linux port done by
> Theobroma Systems. Maybe i should drop support for it and concentrate only on
> the non-linux BSP and on the milkymist platform...
No, it's OK then. I was just worried about reinventing the wheel.
