(didn't mean for this to be off-list sorry about that) On Sun, 16 Aug 2026 10:42, Lucas Tanure <[email protected] mailto:tanure@l= inux.com > wrote: >On 10/08/2026 15:01, Ferass El Hafidi wrote: >> On Thu, 06 Aug 2026 10:18, "Lucas Tanure via groups.io"=C2=A0 >> <[email protected] mailto:[email protected] > wrote: >>> The S4-generation UART (S4, T7, ...) derives its baud rate from the >>> 24 MHz crystal divided by 2, selected via the XTAL_DIV2 bit in the >>> baud rate register, while older SoCs divide the crystal by 3. Add the >>> "amlogic,meson-s4-uart" compatible with driver data selecting the >>> div-by-2 scheme, matching the Linux driver (meson_s4_uart_data) and >>> the configuration the vendor BL2 programs on T7 hardware. >>> >>> Older SoCs keep the existing div-by-3 behaviour: Linux also uses >>> div-by-2 on G12A, but both modes work there and switching would risk >>> regressing boards that have shipped with div-by-3 for years. >>> >>> Also drain the transmitter before changing the baud rate so in-flight >>> characters are not garbled, as the Linux driver does. >>> >>> Assisted-by: Claude:claude-fable-5 >>=C2=A0 >> As I understand it, LLM-assisted patches are not welcome in U-Boot rig= ht >> now: https://lore.kernel.org/u-boot/20260515220758.GM1858239@bill-the-= cat/ >>=C2=A0 >> AFAIK there isn't an official AI policy yet, but it seems the concensu= s >> on that matter is currently "please don't". Sorry. > >Ok, I will drop this line. >
Please do not drop this line. I appreciate the transparency towards your use of AI. I would wait for someone more knowlegable about AI use discussions to reply. Best regards, Ferass
