Kumar Gala wrote:
Jerry,
I was wondering if you wouldn't mind summarizing the state of the fdt/libfdt work. I haven't been following what's been going on here in u-boot.
thanks
- k
Hi Kumar,
I try to keep the wiki page up to date, I've updated it in response to your query: http://www.denx.de/wiki/UBoot/UBootFdtInfo (it is linked off the Custodian page:) http://www.denx.de/wiki/UBoot/Custodians
LIBFDT is in good shape in general, quite usable (send patches if I lie...). The #1 outstanding task as I see it is to convert the board-, CPU-, and PCI-specific ft_*() fdt initialization/update routines over to to use libfdt. Patches for these improvements would go to the appropriate 8Yxx maintainer's repositories {Kim|Andy|Jon}.
Outstanding Patches: -------------------- * Andy Fleming's patch WRT making sure the blob is in the lower portion (16GB at the moment) of memory. Discussion has been resolved, it will be applied and a pull request will go out tonight (Wolfgang's tomorrow ;-).
* Unpublished patch of JerryVanBaren, implementing a suggestion by Scott Wood to make the /chosen handling finer grained: if the /chosen node exists, currently u-boot-fdt bails out. A better methodology is to not overwrite pre-existing properties on a per-property basis, so if /chosen exists but a necessary /chosen/property doesn't, it gets created. Good idea, but not published and poorly tested at the moment.
CPU sub-architectures: ---------------------- See the table(and update or send me updates): http://www.denx.de/wiki/UBoot/UBootFdtInfo#AdoptionTable
83xx: mpc8360emds is converted (my eval board). Kim has done work on 834x board(s) but I'm not sure which ones have been converted off-hand.
86xx: Jon has said he has a note on his whiteboard to convert the 86xx over to using LIBFDT.
85xx: Unknown. Andy is the maintainer and has submitted patch(es) so there obviously is some activity there.
Best regards, gvb