Re: [PATCH] i.MX6 PCIe: Fix imx6_pcie_deassert_core_reset() polarity

[Marcel Ziswiler]

Hi Petr

On Wed, 2016-03-30 at 14:06 +0200, Petr Åtetiar wrote:
> Krzysztof HaÅasa <khalasa@xxxxxxx> [2016-03-25 14:32:35]:
> 
> CzeÅÄ,
> 
> > 
> > I wonder if all boards (except maybe that Toradex set) use an
> > active-low
> > PCIe reset and are now broken. Perhaps Toradex uses active-high and
> > thus
> > works.
> I'm really puzzled by this :-) With your patch applied I get
> following on
> Toradex Apalis modules:
> 
> ÂDTS:ÂÂÂÂÂÂÂÂÂreset-gpio = <&gpio1 28 GPIO_ACTIVE_LOW>;
> Âdmesg:ÂÂÂÂÂÂÂimx6q-pcie 1ffc000.pcie: phy link never came up
> Âgpio:ÂÂÂÂÂÂÂÂgpio-28ÂÂ(ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ|resetÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ)
> out hi
> Âpin voltage: 0V
> 
> ÂDTS:ÂÂÂÂÂÂÂÂÂreset-gpio = <&gpio1 28 GPIO_ACTIVE_HIGH>;
> Âdmesg:ÂÂÂÂÂÂÂath9k 0000:01:00.0: enabling device (0140 -> 0142)
> Âgpio:ÂÂÂÂÂÂÂÂgpio-28ÂÂ(ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ|resetÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ)
> out lo
> Âpin voltage: 3V3
> 
> So Toradex Apalis is actually active-high?

Yes, I actually explained this in detail in my cover letter:

http://article.gmane.org/gmane.linux.drivers.devicetree/154829

>  Thanks.
> 
> -- ynezz


Cheers

Marcel