Hi Matteo, On 06/13/2014 08:48 PM, Matteo Fortini wrote:
Hi all, glad you found my patch useful. Sascha rejected it because he sees it more fit to separate the initialization of sama5d3 and sam9 since they are quite different. I started, as advised by Sascha, to create into sam9_smc.c the function void sama5d3_smc_configure(int id, int cs, struct sama5d3_smc_config *config) but this brings on some other changes to keep the same structure of functions, i.e. we would need to implement static void sama5d3_smc_cs_configure(void __iomem *base, struct sama5d3_smc_config *config) and all the related functions, since the argument changes from struct sam9_smc_config * to struct sama5d3_smc_config * Now I'm asking you all for a comment: should we go ahead and create a new sama5d3_smc.c file with all the functions (some will unfortunately be a duplicate of those present in sam9_smc.c), or should I do a partial hack to insert sama5d3 specific functions into sam9_smc.c (like, for example, playing with config structures so that the sam9 one is just the head of the sama5d3)?
I think we'd better to create a new sama5d3_smc.c. This will be more readable, and also benefit for the new coming SoC.
Thank you in advance for your comments, I ask Raphaël to wait until this patch is settled to send in his changes. They will be very useful for me, too (I had to deactivate PMECC to use my NAND...) M
Best Regards, Bo Shen _______________________________________________ barebox mailing list barebox@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/barebox
