Re: [PATCH v2 3/3] dwc: PCI: intel: Intel PCIe RC controller driver

From: Martin Blumenstingl
Date: Thu Aug 29 2019 - 17:01:36 EST


Hi Kishon,

On Thu, Aug 29, 2019 at 7:10 AM Kishon Vijay Abraham I <kishon@xxxxxx> wrote:
[...]
> The PCI EXPRESS CARD ELECTROMECHANICAL SPECIFICATION defines the Power
> Sequencing and Reset Signal Timings in Table 2-4. Please also refer Figure
> 2-10: Power Up of the CEM.
>
> ââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââ
> â Symbol â Parameter â Min â Max â Units â
> ââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââ
> â T PVPERL â Power stable to PERST# inactive â 100 â â ms â
> ââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââ
> â T PERST-CLK â REFCLK stable before PERST# inactive â 100 â â Îs â
> ââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââ
> â T PERST â PERST# active time â 100 â â Îs â
> ââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââ
> â T FAIL â Power level invalid to PERST# active â â 500 â ns â
> ââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââ
> â T WKRF â WAKE# rise â fall time â â 100 â ns â
> ââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââââ
>
> In my code I used T PERST-CLK (i.e REFCLK stable before PERST# inactive).
> REFCLK to the card is enabled as part of PHY enable and then wait for 100Îs
> before making PERST# inactive.
>
> Power to the device is given during board power up and the assumption here is
> it will take more the 100ms for the probe to be invoked after board power up
> (i.e after ROM, bootloaders and linux kernel). But if you have a regulator that
> is enabled in PCI probe, then T PVPERL (100ms) should also used in probe.
thank you for this detailed overview and for the explanation about the
assumptions you made (and why)


Martin