Re: [PATCH 2/2] mmc: dw_mmc: add resets support to dw_mmc

[Shawn Lin]

å 2016/3/29 16:23, zhangfei åé:
> On 03/29/2016 10:22 AM, Shawn Lin wrote:
>
> > >         +               else if (IS_ERR(host->pdata)) {
> > >                                  dev_err(host->dev, "platform data not
> > >         available\n");
> > >                                  return -EINVAL;
> > >                          }
> > >         @@ -3012,6 +3022,9 @@ int dw_mci_probe(struct dw_mci *host)
> > >                          }
> > >                  }
> > >
> > >         +       if (host->pdata->rstc != NULL)
> > >         +               reset_control_deassert(host->pdata->rstc);
> > >         +
> > >
> > >
> > >     sorry, I can't follow your intention here. Shouldn't it be something
> > >     like "assert mmc -> may need delay -> deassert mmc". As your current
> > >     code, nothing happend right?
> Should be abstracted in reset driver.
>
> > > The chip exits from bootloader with this bit asserted. And when entering
> > > kernel, we only need to deassert.
> > >
> > > In my current code, the driver deassert mmc in _probe(), and assert mmc
> > > in _remove().
> >
> > I catch your point. From the previous discussion, we add it to make sure
> > dw_mmc in good state after leaving bootloader to kernel. But My real
> > question is that you can assert it in  bootloader, so you can also
> > dessert it in bootloaer to make sure dw_mmc work fine when probing
> > in kernel. In that way, we don't need this patch?
>
> uefi does not have exit point, and kernel may not assume mmc controller
> state is always correct when boot.
> If Uefi need copy Image from mmc, mmc controller is in working state.
> When jump to kernel, uefi mmc driver can not recover itself.
> If kernel assume mmc controller state is clean, mmc will be in abnormal
> state.
> Some controller will clear itself when set clock, however, hip660 does
> not, it need special register to access.

yep, I understand your requirement.

> > More to think, Is it ok to match the behaviour of bootloader stage?
> > My bootloader doesn't assert the reset pin of dw_mmc, so it seams if
> > I want to fix you issue on kernel stage, I need a new round of
> > assert->delay->deassert.
>
> The process like delay (if required) should be abstracted in reset driver.
> reset framework just export reset_control_assert/reset_control_deassert
> API.
> Unfortunately not find clear description in Documentation/.
> Suppose deassert is like start, while assert is like stop.

yes, no docs except dt-bindings..... So seems the usage of these two
APIs depends on the implementation of reset controller driver

> drivers/reset/core.c
> reset_control_deassert - deasserts the reset line
> reset_control_assert - asserts the reset line
>
> More example:
> drivers/mmc/host/sdhci-st.c
> drivers/mmc/host/sunxi-mmc.c
> drivers/usb/host/ohci-platform.c
> drivers/i2c/busses/i2c-mv64xxx.c

But I'm afraid I have to nack here....

Looking at these files:
drivers/dma/stm32-dma.c
drivers/net/ethernet/mediatek/mtk_eth_soc.c
drivers/devfreq/tegra-devfreq.c
drivers/crypto/rockchip/rk3288_crypto.c
drivers/thermal/rockchip_thermal.c
drivers/thermal/tegra_soctherm.c
drivers/i2c/busses/i2c-tegra.c
....

they just do the way like: assert->[delay](maybe abstracted)->deassert
I think these drivers are vendor specific, so they can do
the reset operations in consistent with the implementation of
their platforms' reset controller drivers.

But, dw_mmc is shared by many Socs. So It's not good to do it in
dw_mci_probe, otherwise you force my reset controller driver to be
implemented in the same way as yours..... Right?

How about move it to your drv_data->init callback?


> Thanks


--
Best Regards
Shawn Lin