On 22-01-10 19:18:58, Jesse Taube wrote: > > > On 1/10/22 17:48, Abel Vesa wrote: > > On 22-01-10 12:12:00, Stephen Boyd wrote: > > > Quoting Jesse Taube (2022-01-09 11:07:42) > > > > > > + > > > > > > + clk_hw_data->num = IMXRT1050_CLK_END; > > > > > > + hws = clk_hw_data->hws; > > > > > > + > > > > > > + hws[IMXRT1050_CLK_OSC] = __clk_get_hw(of_clk_get_by_name(np, "osc")); > > > > > > > > > > Use clk_parent_data instead and reference the binding name with .fw_name > > > > > > > > Hi Stephen, sorry to bother you again. > > > > I'm wondering if adding `clk_parent_data` is necessary as it seems like > > > > it wold be a big change and the other IMX boards haven't implemented > > > > this yet would it be okay if I don't do this, or if it is necessary > > > > could you link to a patch set to change it. > > > > > > > > > > Is it a big change because the imx_clk_hw*() APIs aren't prepared for > > > non-string parents? Can you make some clk_parent_data based APIs and > > > migrate things over gradually? This is really a question for i.MX folks > > > to see if anyone is working on moving away from the string names. > > > > I'm currently looking into it. I would suggest we take this patch as is > > and I'll switch it later on to clk_parent_data. > Thanks so much that would be great! > > Should I sent the other changes requested I have them ready? Or should i > wait. Sure, send it. If Stephen is OK with my suggestion, I will pick it. > > Best regards > Jesse Taube
