Re: [PATCH v3 0/4] arm64: dts: exynos2200: introduce serial busses, except spi

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 8/16/25 20:29, Sam Protsenko wrote:
> On Fri, Aug 15, 2025 at 2:05 AM Ivaylo Ivanov
> <ivo.ivanov.ivanov1@xxxxxxxxx> wrote:
>> Hey, folks!
>>
>> This patchset adds serial busses, implemented in usi, for exynos2200.
>> It's missing spi, due to me having troubles with reads when testing.
>> Serial_0/1 have not been included in the patchset, as it seems like
>> they're encapsulated in usi blocks, but are the only implemented
>> protocol and/or do not have a dedicated register for setting other
>> protocols in a sysreg. That'd at least require patches in the usi
>> driver and bindings to add support for.
>>
>> About the naming convention for usi nodes, I've chosen to keep the
>> downstream one instead of relabelling all to avoid confusion when
>> cross-referencing the vendor DT and to keep consistency with clock
>> names. They're labelled the same in the bootloader too.
>>
>> Best regards,
>> Ivaylo
>>
>> Changes in v3:
>> - drop the serial_0/1 patch
> Why did you decide to drop it? As I understand, having a serial
> console enabled can be quite valuable for this platform bring up.

On the only board that is supported (g0s), serial can either be accessed
via a special usb-c cable, whose way of working only Samsung knows,
or soldering to uart pads on a tiny masked jtag (?) connector on the motherboard.
The latter I'm not sure exists, as I haven't checked on g0s. (it does on dreamlte).
Using uart is not really viable for debugging, framebuffer/usb work too :)

> If you don't know how to add the related USI node -- it's ok, can be done
> later, as soon as the serial works without it.

I can't test that. On dreamlte a similar scenario exists, however usiv1
does not require the clock gating code, so not encapsulating it in
a USI node is fine. Not sure about g0s, however.

Best regards,
Ivaylo

>
>> - add r-b tags from Sam
>> - increase the size of all syscon to 0x10000 and not 0x3000
>> - change description of last patch to be more meaningful regarding the
>> usiN and usiN_i2c mess
>> - s/usi6_i2c_cmgp/usi_i2c_cmgp6, following the TRM naming convention
>>
>> Changes in v2:
>> - add a patch that switches address and size cells to 1 in /soc
>> - adjust all new nodes to define reg props with 2 cells in total instead of 4
>>
>> Ivaylo Ivanov (4):
>>   arm64: dts: exynos2200: fix typo in hsi2c23 bus pins label
>>   arm64: dts: exynos2200: use 32-bit address space for /soc
>>   arm64: dts: exynos2200: increase the size of all syscons
>>   arm64: dts: exynos2200: define all usi nodes
>>
>>  .../boot/dts/exynos/exynos2200-pinctrl.dtsi   |    2 +-
>>  arch/arm64/boot/dts/exynos/exynos2200.dtsi    | 1433 ++++++++++++++++-
>>  2 files changed, 1398 insertions(+), 37 deletions(-)
>>
>> --
>> 2.43.0
>>
>>





[Index of Archives]     [Linux SoC Development]     [Linux Rockchip Development]     [Linux for Synopsys ARC Processors]    
  • [Linux on Unisoc (RDA Micro) SoCs]     [Linux Actions SoC]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Linux SCSI]     [Yosemite News]

  •   Powered by Linux