On Sun, May 26, 2019 at 11:50 PM Andrey Smirnov <andrew.smirnov@xxxxxxxxx> wrote: > > On Sun, May 26, 2019 at 11:29 PM Ahmad Fatoum <a.fatoum@xxxxxxxxxxxxxx> wrote: > > > > Hello Andrey, > > > > On 26/5/19 23:55, Andrey Smirnov wrote: > > > On Tue, May 21, 2019 at 8:56 AM Ahmad Fatoum <a.fatoum@xxxxxxxxxxxxxx> wrote: > > >> > > > This particular patch does break a non 6Q+ version of RDU2, but the > > > follow up ones in the series fix it, so it seems that no action is > > > really necessary on my part. > > > > The reparenting removed in this patch isn't reinstated by the rest of the series. > > They merely apply parentage expressed in the device tree in a glitch-free manner. > > > > As both barebox and kernel imx6qdl-zii-rdu2.dtsi lack the relevant > > assigned-clock-parents snippet, I am not sure what it is this patch broke that the > > follow-up ones fixed? > > Not sure, will investigate. > Ah, I see, so with re-parenting to PLL5 removed ldb_* clocks default to MMDC_CH1 and for that to work "clk: imx6: Mask mmdc_ch1 handshake for periph2_sel and mmdc_ch1_axi_podf" is needed. After that commit MMDC_CH1 seems to work well enough for frame buffer to come up and function. Thanks, Andrey Smirnov _______________________________________________ barebox mailing list barebox@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/barebox
