Re: Two issues with pwrseq_sd8787

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

 



On Wed, Mar 13, 2019 at 01:34:38PM +0100, Ulf Hansson wrote:
> On Mon, 11 Mar 2019 at 16:22, Adrian Bunk <bunk@xxxxxxxxxx> wrote:
> >
> > Hi,
> >
> > I am facing two issues with pwrseq_sd8787:
> >
> > 1. Usage with modular mwifiex but built-in MMC
> > I have built-in MMC (also for the root filesystem) but modular
> > pwrseq_sd8787. Are the Kconfig dependencies incorrect and I should
> > be able to set PWRSEQ_SD8787=y, or how is this supposed to work?
> 
> The Kconfig currently depends on MWIFIEX, so if you have that set to M
> you can only set PWRSEQ_SD8787 to M as well.
> 
> This should work!
> 
> When the corresponding mmc host driver probes, the DT parsing via
> mmc_of_parse() should finds the pwrseq node, but then we fail to
> lookup the corresponding mmc pwrseq driver, simply because it has not
> been registred yet (built as module). If everything works as expected,
> this should lead to that -EPROBE_DEFER becomes propagated, such that
> the driver core re-try to prob later.
> 
> Are you saying this doesn't work, no?

Not in my case, but that might be a special case:

mmc0 is using pwrseq_sd8787.
mmc1 contains the root partition where the pwrseq_sd8787 module is.

This seems to create a chicken&egg problem by also deferring detection 
of the MMC root partition?

> > 2. GPIO_ACTIVE_LOW reset-gpios
> > With reset-gpios active low it is initially low,
> > mmc_pwrseq_sd8787_probe() sets that to high for a very short
> > amount of time before mmc_pwrseq_sd8787_pre_power_on() sets
> > it back to low.
> > What is the proper way to get rid of this sub-ms low-high-low?
> 
> Is that not according to the spec? In such case, I suggest you post a
> patch changing the code so confirms to the spec.
>...

The spec says to keep powerdown-gpios low for at least 300ms after
enabling power.

reset-gpios controls whatever circuitry you have built for supplying 
power to the chip.

Whether you make the GPIO controlling the latter active high or active 
low (or not a GPIO at all) is nothing the spec cares about.

> Hope this helps!
> 
> Kind regards
> Uffe

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed




[Index of Archives]     [Linux Memonry Technology]     [Linux USB Devel]     [Linux Media]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux