From: Yi Min Zhao <zyi...@linux.vnet.ibm.com> On s390x, each pci device has its own iommu, which is only properly setup in qemu once the mpcifc instruction used to register the translation table has been intercepted. Therefore, for a pci device that is not configured or has not been initialized, proper translation is neither required nor possible. Moreover, we may not have a host bridge device ready yet.
This was exposed by a recent vfio change that triggers iommu translation during the initialization of the vfio pci device. Let's do an early exit in that case. Reviewed-by: Cornelia Huck <cornelia.h...@de.ibm.com> Reviewed-by: Jens Freimann <jf...@linux.vnet.ibm.com> Signed-off-by: Yi Min Zhao <zyi...@linux.vnet.ibm.com> Signed-off-by: Cornelia Huck <cornelia.h...@de.ibm.com> --- hw/s390x/s390-pci-bus.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/hw/s390x/s390-pci-bus.c b/hw/s390x/s390-pci-bus.c index 560b66a..d5712ae 100644 --- a/hw/s390x/s390-pci-bus.c +++ b/hw/s390x/s390-pci-bus.c @@ -309,8 +309,7 @@ static IOMMUTLBEntry s390_translate_iommu(MemoryRegion *iommu, hwaddr addr, uint64_t pte; uint32_t flags; S390PCIBusDevice *pbdev = container_of(iommu, S390PCIBusDevice, mr); - S390pciState *s = S390_PCI_HOST_BRIDGE(pci_device_root_bus(pbdev->pdev) - ->qbus.parent); + S390pciState *s; IOMMUTLBEntry ret = { .target_as = &address_space_memory, .iova = 0, @@ -319,8 +318,13 @@ static IOMMUTLBEntry s390_translate_iommu(MemoryRegion *iommu, hwaddr addr, .perm = IOMMU_NONE, }; + if (!pbdev->configured || !pbdev->pdev) { + return ret; + } + DPRINTF("iommu trans addr 0x%" PRIx64 "\n", addr); + s = S390_PCI_HOST_BRIDGE(pci_device_root_bus(pbdev->pdev)->qbus.parent); /* s390 does not have an APIC mapped to main storage so we use * a separate AddressSpace only for msix notifications */ -- 2.6.3