Re: [PATCH v16 4/7] s390/pci: Store PCI error information for passthrough devices

[Farhan Ali]

On 5/18/2026 3:07 AM, Niklas Schnelle wrote:
> On Thu, 2026-05-14 at 11:00 -0700, Farhan Ali wrote:
> > For a passthrough device we need co-operation from user space to recover
> > the device. This would require to bubble up any error information to user
> > space.  Let's store this error information for passthrough devices, so it
> > can be retrieved later.
> >
> > We can now have userspace drivers (vfio-pci based) on s390x. The userspace
> > drivers will not have any KVM fd and so no kzdev associated with them. So
> > we need to update the logic for detecting passthrough devices to not depend
> > on struct kvm_zdev.
> >
> > Reviewed-by: Matthew Rosato <mjrosato@xxxxxxxxxxxxx>
> > Signed-off-by: Farhan Ali <alifm@xxxxxxxxxxxxx>
> > ---
> >   arch/s390/include/asm/pci.h      |  30 ++++++++
> >   arch/s390/pci/pci.c              |   1 +
> >   arch/s390/pci/pci_event.c        | 114 +++++++++++++++++--------------
> >   drivers/vfio/pci/vfio_pci_zdev.c |  18 ++++-
> >   4 files changed, 109 insertions(+), 54 deletions(-)
> --- snip ---
> > @@ -216,12 +223,23 @@ static pci_ers_result_t zpci_event_attempt_error_recovery(struct pci_dev *pdev)
> >   		goto out_unlock;
> >   	}
> >   
> > +	mediated_recovery = zpci_store_pci_error(pdev, ccdf) == 0 ? true : false;
> > +
> This incorrectly treats it as not-mediated recovery when the queue runs
> full and zpci_store_pci_error() returns -ENOMEM (Sashiko also noticed
> this though the explanation is a bit hard to read).

Yes, will fix in next revision.


> >   	ers_res = zpci_event_notify_error_detected(pdev, driver);
> >   	if (ers_result_indicates_abort(ers_res)) {
> >   		status_str = "failed (abort on detection)";
> >   		goto out_unlock;
> >   	}
> >   
> > +	if (mediated_recovery) {
> > +		pr_info("%s: Leaving recovery of pass-through device to user-space\n",
> > +			pci_name(pdev));
> > +		ers_res = PCI_ERS_RESULT_RECOVERED;
> > +		status_str = "in progress";
> > +		mutex_unlock(&zdev->pending_errs_lock);
> The unlock here seems like a bad leftover from the previous code.
> You're not holding that lock here. (Also found by Sashiko).

Ugh yes, left this from last version. Will fix.

Thanks

Farhan

> > +		goto out_unlock;
> > +	}
> > +
> --- snip ---