RE: Non systematic PCIe error on resume on i.MX8QP on Linux 7.2

> -----Original Message-----
> From: Francesco Dolcini <[email protected]>
> Sent: Monday, July 20, 2026 5:39 PM
> To: Hongxing Zhu <[email protected]>
> Cc: Francesco Dolcini <[email protected]>; Frank Li <[email protected]>; linux-
> [email protected]; [email protected]; [email protected];
> [email protected]; [email protected]; [email protected];
> [email protected]; [email protected]; [email protected];
> [email protected]; [email protected]
> Subject: Re: Non systematic PCIe error on resume on i.MX8QP on Linux 7.2
> 
> On Mon, Jul 20, 2026 at 09:16:27AM +0000, Hongxing Zhu wrote:
> > > -----Original Message-----
> > > From: Francesco Dolcini <[email protected]>
> > > Sent: Monday, July 20, 2026 1:28 PM
> > > To: Hongxing Zhu <[email protected]>; Frank Li
> > > <[email protected]>; linux- [email protected]; [email protected]
> > > Cc: [email protected]; [email protected];
> > > [email protected]; [email protected]; [email protected];
> > > [email protected]; [email protected]; [email protected];
> > > [email protected]; linux- [email protected]
> > > Subject: Non systematic PCIe error on resume on i.MX8QP on Linux 7.2
> > >
> > > Hello,
> > > our CI spotted this error on resume. It is not systematic, and it is
> > > possible that this is not a new issue from 7.2.
> > >
> > > Does it rings any bell?
> > >
> > > [  745.849237] imx8qm-hsio-phy 5f180000.phy: HSIO_PM_REQ_CORE_RST is
> > > set [  745.855800] phy phy-5f180000.phy.1: phy poweron failed -->
> > > -110
> > Based on the log, the PHY initialization is failing during resume.
> > I'll set up a platform on my end to see if I can reproduce this issue consistently.
> 
> Thanks for the support, appreciated.
> 
> > BTW, is this occurring on the i.MX8QM MEK board?
> The issue is on apalis-imx8 (you can find the DT in the usual place in the linux
> kernel tree).
> 
> > What's the frequency of occurrence in CI (failure rate)?
> The last batch was 3 failure in 10 executions, so 30%.
Hi Francesco:
I tried the stress PM tests on both i.MX8QM MEK and i.MX8QXP MEK boards.
Unfortunately, I can't reproduce this random problem.

Best Regards
Richard Zhu
> 
> Francesco
>