Re: [PATCH 2/3] mmc: mmci: stm32: do not use NEGEDGE in HS200 mode

[Christophe KERELLO]

Hello Mateusz,

On 11/3/26 12:40, mateusz.nowicki@xxxxxxxxxx wrote:
> Hello Christophe,
>
> > No, it is not what it is written in the RM. If clkdiv is equal to 0, 
> > NEGEDGE has no effect.
> > NEGEDGE has an effect only when clkdiv > 0, so it can be used with 
> > HS200.
>
> You're right, thanks for the correction. I read the NA in table 268 as
> "not allowed in HS200", but it is only there because CLKDIV is 0 in that
> configuration. I'll drop this patch.
>
> > On ST downstream Linux Kernel, you can find following patch not 
> > already upstream:
> > mmc: mmci: stm32: use the delay block above 100 MHz on stm32mp25 SoC 
> > (Commit ca3b1cf)
> > By applying previous downstream patch + patch 3 of this patchset + 
> > using  original slew rate that was upstream on
> > this board (slew rate set to 2 for the clock IO, 1 for the others 
> > IOs), HS200 should run on a ST MP25 DK board at 100 MHz.
>
> Thanks for the pointer, that was the missing piece. I went back to the
> DK and tested it with the SDMMC2 kernel clock at 200 MHz, so HS200 runs
> at 100 MHz like in mainline. Without ca3b1cf every write test failed,
> no matter whether NEGEDGE was set or which pin group I used. With
> ca3b1cf all runs were clean, with pins_a as well as with pins_b. I also
> checked 150 MHz (no divider, delay block in use), and it works with
> both pin groups too.
>
> So NEGEDGE was never the problem, it was the delay block running at
> 100 MHz. I'll include ca3b1cf and 01d3cac in v2, with you as the author.

Thank you very much for confirming that everything works correctly in 
both configurations.
I let you submit these two patches upstream; while the commit titles and 
messages can be adjusted,
we simply ask that you retain the original author information as well as 
the "Signed-off-by" line
when the patches originate from the ST downstream branch. Thank you in 
advance.

> One more thing about the pins. I added the pins_b groups (faster slew
> rate) to the DK device tree in 7.4 because pins_a seemed to corrupt data
> in HS200 at 100 MHz. Looking at it now, that was the same delay block
> issue. Would you like me to send a patch that switches the DK back to
> pins_a and drops pins_b?

Yes, using pins_a on the DK board is preferable to using pins_b.
This helps avoid overshoots on the SDMMC clock signal, which would 
otherwise occur with faster input/output (I/O) pins.
These overshoots can interfere with other functional blocks, such as the 
USB; therefore, it is better to use I/O pins
that provide a "smoother" signal. Since pins_b is not being used, it is 
best to remove the pins_b configurations for SDMMC2.

Regards,
Christophe Kerello.

> > This patch breaks also MP13 SoC that is supporting HS200.
>
> Understood, sorry about that.
>
> Regards,
> Mateusz