[PATCH V10 04/13] PCI: imx6: Assert PERST# before enabling regulators

[Manivannan Sadhasivam]

On Tue, Apr 07, 2026 at 06:38:50AM +0000, Sherry Sun wrote:
> 
> > On Thu, Apr 02, 2026 at 05:50:58PM +0800, Sherry Sun wrote:
> > > According to the PCIe initialization requirements, PERST# signal
> > > should be asserted before applying power to the PCIe device, and
> > > deasserted after power and reference clock are stable.
> > >
> > 
> > Spec wording is not quite like this. Spec mandates asserting PERST# *before*
> > stopping refclk and powering down the device and deasserting it *after*
> > applying power and refclk stable.
> > 
> > I believe you want to assert PERST# before enabling regulator to prevent the
> > endpoint from functioning? If so, is it due to refclk not available yet or some
> > other reason?
> 
> You are right. My commit message wording was not that precise.
> The PCIe endpoint may start responding or driving signals as
> soon as its supply is enabled, even before the reference clock is stable.
> Asserting PERST# before enabling the regulator ensures that the endpoint
> remains in reset throughout the entire power-up sequence, until both
> power and refclk are known to be stable and link initialization can safely
> begin. This is mainly to avoid undefined behavior during early power-up.
> 
> I will update the commit message to better reflect this.
> 
> > 
> > > Currently, the driver enables the vpcie3v3aux regulator in
> > > imx_pcie_probe() before PERST# is asserted in imx_pcie_host_init(),
> > > which violates the PCIe power sequencing requirements. However, there
> > > is no issue so far because PERST# is requested as GPIOD_OUT_HIGH in
> > > imx_pcie_probe(), which guarantees that PERST# is asserted before
> > > enabling the vpcie3v3aux regulator.
> > >
> > > This is prepare for the upcoming changes that will parse the reset
> > > property using the new Root Port binding, which will use GPIOD_ASIS
> > > when requesting the reset GPIO. With GPIOD_ASIS, the GPIO state is not
> > > guaranteed, so explicit sequencing is required.
> > >
> > > Fix the power sequencing by:
> > > 1. Moving vpcie3v3aux regulator enable from probe to
> > >    imx_pcie_host_init(), where it can be properly sequenced with PERST#.
> > > 2. Moving imx_pcie_assert_perst() before regulator and clock enable to
> > >    ensure correct ordering.
> > >
> > > The vpcie3v3aux regulator is kept enabled for the entire PCIe
> > > controller lifecycle and automatically disabled on device removal via devm
> > cleanup.
> > >
> > 
> > vpcie3v3aux handling should be in a separate patch.
> 
> Actually the handling of vpcie3v3aux itself remains unchanged, I just adjust the
> sequence of regulator/clock enable and perst#.
> Previously, the imx driver enabled the vpcie3v3aux regulator in imx_pcie_probe()
> before PERST# is asserted in imx_pcie_host_init(), which violates the PCIe power
> sequencing requirements.
> This patch moves vpcie3v3aux regulator enable from probe to  imx_pcie_host_init(),
> where it can be properly sequenced with PERST#.
> Perhaps I should just remove this description to avoid confusion.
> 

Yeah.

- Mani

-- 
மணிவண்ணன் சதாசிவம்