Re: [PATCH v17 5/7] vfio-pci/zdev: Add a device feature for error information

[Farhan Ali <alifm@linux.ibm.com>]

On 6/2/2026 1:32 PM, Alex Williamson wrote:
> On Wed, 20 May 2026 10:11:11 -0700
> Farhan Ali <alifm@linux.ibm.com> wrote:
>
>> For zPCI devices, we have platform specific error information. The platform
>> firmware provides this error information to the operating system in an
>> architecture specific mechanism. To enable recovery from userspace for
>> these devices, we want to expose this error information to userspace. Add a
>> new device feature to expose this information.
>>
>> Reviewed-by: Matthew Rosato <mjrosato@linux.ibm.com>
>> Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com>
>> Signed-off-by: Farhan Ali <alifm@linux.ibm.com>
>> ---
>>   arch/s390/include/asm/pci.h      |  2 ++
>>   arch/s390/pci/pci_event.c        | 19 ++++++++++++++++
>>   drivers/vfio/pci/vfio_pci_core.c |  2 ++
>>   drivers/vfio/pci/vfio_pci_priv.h |  9 ++++++++
>>   drivers/vfio/pci/vfio_pci_zdev.c | 39 ++++++++++++++++++++++++++++++++
>>   include/uapi/linux/vfio.h        | 30 ++++++++++++++++++++++++
>>   6 files changed, 101 insertions(+)
>>
>> diff --git a/arch/s390/include/asm/pci.h b/arch/s390/include/asm/pci.h
>> index 016386f7ef4a..88a125b92bdd 100644
>> --- a/arch/s390/include/asm/pci.h
>> +++ b/arch/s390/include/asm/pci.h
>> @@ -364,6 +364,8 @@ int zpci_clear_error_state(struct zpci_dev *zdev);
>>   int zpci_reset_load_store_blocked(struct zpci_dev *zdev);
>>   void zpci_start_mediated_recovery(struct zpci_dev *zdev);
>>   void zpci_stop_mediated_recovery(struct zpci_dev *zdev);
>> +int zpci_get_pending_error(struct zpci_dev *zdev,
>> +			   struct zpci_ccdf_err *ccdf);
>>   
>>   #ifdef CONFIG_NUMA
>>   
>> diff --git a/arch/s390/pci/pci_event.c b/arch/s390/pci/pci_event.c
>> index cf2ffa21ab8c..db1b44baf8fa 100644
>> --- a/arch/s390/pci/pci_event.c
>> +++ b/arch/s390/pci/pci_event.c
>> @@ -75,6 +75,25 @@ static int zpci_store_pci_error(struct pci_dev *pdev,
>>   	return 0;
>>   }
>>   
>> +int zpci_get_pending_error(struct zpci_dev *zdev,
>> +			   struct zpci_ccdf_err *ccdf)
>> +{
>> +	int head = 0;
>> +
>> +	guard(mutex)(&zdev->pending_errs_lock);
>> +
>> +	if (!zdev->pending_errs.count)
>> +		return -ENOMSG;
>> +
>> +	head = zdev->pending_errs.head % ZPCI_ERR_PENDING_MAX;
>> +	memcpy(ccdf, &zdev->pending_errs.err[head],
>> +	       sizeof(struct zpci_ccdf_err));
>> +	zdev->pending_errs.head++;
>> +	zdev->pending_errs.count--;
>> +	return 0;
>> +}
>> +EXPORT_SYMBOL_GPL(zpci_get_pending_error);
>> +
>>   void zpci_start_mediated_recovery(struct zpci_dev *zdev)
>>   {
>>   	guard(mutex)(&zdev->pending_errs_lock);
>> diff --git a/drivers/vfio/pci/vfio_pci_core.c b/drivers/vfio/pci/vfio_pci_core.c
>> index 050e7542952e..27642f10fe97 100644
>> --- a/drivers/vfio/pci/vfio_pci_core.c
>> +++ b/drivers/vfio/pci/vfio_pci_core.c
>> @@ -1569,6 +1569,8 @@ int vfio_pci_core_ioctl_feature(struct vfio_device *device, u32 flags,
>>   		return vfio_pci_core_feature_token(vdev, flags, arg, argsz);
>>   	case VFIO_DEVICE_FEATURE_DMA_BUF:
>>   		return vfio_pci_core_feature_dma_buf(vdev, flags, arg, argsz);
>> +	case VFIO_DEVICE_FEATURE_ZPCI_ERROR:
>> +		return vfio_pci_zdev_feature_err(device, flags, arg, argsz);
>>   	default:
>>   		return -ENOTTY;
>>   	}
>> diff --git a/drivers/vfio/pci/vfio_pci_priv.h b/drivers/vfio/pci/vfio_pci_priv.h
>> index fca9d0dfac90..4e7162234a2e 100644
>> --- a/drivers/vfio/pci/vfio_pci_priv.h
>> +++ b/drivers/vfio/pci/vfio_pci_priv.h
>> @@ -93,6 +93,8 @@ int vfio_pci_info_zdev_add_caps(struct vfio_pci_core_device *vdev,
>>   				struct vfio_info_cap *caps);
>>   int vfio_pci_zdev_open_device(struct vfio_pci_core_device *vdev);
>>   void vfio_pci_zdev_close_device(struct vfio_pci_core_device *vdev);
>> +int vfio_pci_zdev_feature_err(struct vfio_device *device, u32 flags,
>> +			      void __user *arg, size_t argsz);
>>   #else
>>   static inline int vfio_pci_info_zdev_add_caps(struct vfio_pci_core_device *vdev,
>>   					      struct vfio_info_cap *caps)
>> @@ -107,6 +109,13 @@ static inline int vfio_pci_zdev_open_device(struct vfio_pci_core_device *vdev)
>>   
>>   static inline void vfio_pci_zdev_close_device(struct vfio_pci_core_device *vdev)
>>   {}
>> +
>> +static inline int vfio_pci_zdev_feature_err(struct vfio_device *device,
>> +					    u32 flags, void __user *arg,
>> +					    size_t argsz)
>> +{
>> +	return -ENOTTY;
>> +}
>>   #endif
>>   
>>   static inline bool vfio_pci_is_vga(struct pci_dev *pdev)
>> diff --git a/drivers/vfio/pci/vfio_pci_zdev.c b/drivers/vfio/pci/vfio_pci_zdev.c
>> index 78a28db00c6d..cc148123a97b 100644
>> --- a/drivers/vfio/pci/vfio_pci_zdev.c
>> +++ b/drivers/vfio/pci/vfio_pci_zdev.c
>> @@ -141,6 +141,45 @@ int vfio_pci_info_zdev_add_caps(struct vfio_pci_core_device *vdev,
>>   	return ret;
>>   }
>>   
>> +int vfio_pci_zdev_feature_err(struct vfio_device *device, u32 flags,
>> +			      void __user *arg, size_t argsz)
>> +{
>> +	struct vfio_device_feature_zpci_err err = {};
>> +	struct vfio_pci_core_device *vdev;
>> +	struct zpci_ccdf_err ccdf = {};
>> +	struct zpci_dev *zdev;
>> +	int ret;
>> +
>> +	vdev = container_of(device, struct vfio_pci_core_device, vdev);
>> +	zdev = to_zpci(vdev->pdev);
>> +	if (!zdev)
>> +		return -ENODEV;
>> +
>> +	ret = vfio_check_feature(flags, argsz, VFIO_DEVICE_FEATURE_GET,
>> +				 sizeof(err));
>> +	if (ret != 1)
>> +		return ret;
>> +
>> +	ret = zpci_get_pending_error(zdev, &ccdf);
>> +	if (ret)
>> +		return ret;
>> +
>> +	err.fh = ccdf.fh;
>> +	err.fid = ccdf.fid;
>> +	err.ett = ccdf.ett;
>> +	err.mvn = ccdf.mvn;
>> +	err.dmaas = ccdf.dmaas;
>> +	err.q = ccdf.q;
>> +	err.rw = ccdf.rw;
>> +	err.faddr = ccdf.faddr;
>> +	err.pec = ccdf.pec;
>> +
>> +	if (copy_to_user(arg, &err, sizeof(err)))
>> +		return -EFAULT;
>> +
>> +	return 0;
>> +}
> There's a concern here that the error is dequeued but we can still fail
> resulting in a lost error.  Is that a sufficiently unlikely scenario to
> ignore or should it at least be documented?

IMHO this scenario would be unlikely and could be ignored. I think to 
handle this we would need something like a peek() API and only remove 
and update the pending_errs if copy_to_user succeeds. This would be 
adding more complexity for a relatively low possibility of failure.


>
>> +
>>   int vfio_pci_zdev_open_device(struct vfio_pci_core_device *vdev)
>>   {
>>   	struct zpci_dev *zdev = to_zpci(vdev->pdev);
>> diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
>> index 5de618a3a5ee..3eead86a00ab 100644
>> --- a/include/uapi/linux/vfio.h
>> +++ b/include/uapi/linux/vfio.h
>> @@ -1534,6 +1534,36 @@ struct vfio_device_feature_dma_buf {
>>    */
>>   #define VFIO_DEVICE_FEATURE_MIG_PRECOPY_INFOv2  12
>>   
>> +/**
>> + * VFIO_DEVICE_FEATURE_ZPCI_ERROR feature provides PCI error information to
>> + * userspace for vfio-pci devices on s390. On s390, PCI error recovery
>> + * involves platform firmware and notification to operating systems is done
>> + * by architecture specific mechanism. Exposing this information to
>> + * userspace allows it to take appropriate actions to handle an
>> + * error on the device. The ioctl returns -ENOMSG if there are no pending
>> + * PCI errors.
>> + */
>> +
>> +struct vfio_device_feature_zpci_err {
>> +	__u32 feature_flags;		/* Indicate future features */
>> +	__u32 reserved1;
>> +	__u32 fh;			/* function handle */
>> +	__u32 fid;			/* function id */
>> +	__u32 ett		:  4;	/* expected table type */
>> +	__u32 mvn		: 12;	/* MSI vector number */
>> +	__u32 dmaas		:  8;	/* DMA address space */
>> +	__u32 reserved2		:  6;
>> +	__u32 q			:  1;	/* event qualifier */
>> +	__u32 rw		:  1;	/* read/write */
>> +	__u64 faddr;			/* failing address */
>> +	__u32 reserved3;
>> +	__u16 reserved4;
>> +	__u16 pec;			/* PCI event code */
>> +	__u8 reserved5[28];		/* Allow for future expansion */
>> +};
>> +
>> +#define VFIO_DEVICE_FEATURE_ZPCI_ERROR 13
>> +
> There are currently at least 4 different patch series, including this
> one, that are vying for device feature allocations.  This series spans
> PCI, s390, and vfio, and I'm not entirely sure which maintainer should
> take it.  With Niklas' review on patch 4, should we ask Bjorn to take
> 1-3 through PCI, share a branch, and I could merge 4-7 through vfio?
> Thanks,
>
> Alex

If it helps I can break the series into PCI only and vfio only patches. 
Please let me know if there is anything I can help with to make the 
merge easier.

Thanks

Farhan