RTEMS | Use FDT to define memory mappings (#5510)

[Chris Johns (@chris)]

Chris Johns commented: https://gitlab.rtems.org/rtems/rtos/rtems/-/issues/5510#note_145102


The ability to use FDT in a BSP like is would be nice however we lack an FDT framework in RTEMS to sit this on. My concern is anything we do here becomes yet another approach.

The [default FDT](https://gitlab.rtems.org/rtems/rtos/rtems/-/tree/main/bsps/aarch64/xilinx-zynqmp/fdt) for the BSP could be used if updated to reflect these changes so a user does not see any breakage. I am not sure a user can override the default FDT blobs? Being able to do that starts to get into the ability to overlay blobs or even using other techniques. I feel we should first develop an approach we can look to work across RTEMS and then we align all BSPs to it we can. Some BSPs may never be aligned and so they can stay as isolated cases.

As a specific example libBSD with FreBSD PCIe support uses a new interrupt structure to define the interrupt mappings. This is a perfect case for FDT but the DT may be specific to a BSP or piece of hardware. PCIe is being used as a chip-to-chip bus these days so we will see the emergence of custom PCI end points.

-- 
View it on GitLab: https://gitlab.rtems.org/rtems/rtos/rtems/-/issues/5510#note_145102
You're receiving this email because of your account on gitlab.rtems.org.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/bugs/attachments/20260312/2f07e973/attachment.htm>