Setting domain->iommufd_hwpt in iommufd_hwpt_alloc() only covers the HWPT
allocations from user space, but not for an auto domain. This resulted in
a NULL pointer access in the auto domain pathway:
Unable to handle kernel NULL pointer dereference at
virtual address
0000000000000008
pc : iommufd_sw_msi+0x54/0x2b0
lr : iommufd_sw_msi+0x40/0x2b0
Call trace:
iommufd_sw_msi+0x54/0x2b0 (P)
iommu_dma_prepare_msi+0x64/0xa8
its_irq_domain_alloc+0xf0/0x2c0
irq_domain_alloc_irqs_parent+0x2c/0xa8
msi_domain_alloc+0xa0/0x1a8
Since iommufd_sw_msi() requires to access the domain->iommufd_hwpt, it is
better to set that explicitly prior to calling iommu_domain_set_sw_msi().
Fixes:
748706d7ca06 ("iommu: Turn fault_data to iommufd private pointer")
Link: https://patch.msgid.link/r/20250305211800.229465-1-nicolinc@nvidia.com
Reported-by: Ankit Agrawal <ankita@nvidia.com>
Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Tested-by: Ankit Agrawal <ankita@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
goto out_abort;
}
}
+ hwpt->domain->iommufd_hwpt = hwpt;
iommu_domain_set_sw_msi(hwpt->domain, iommufd_sw_msi);
/*
goto out_abort;
}
hwpt->domain->owner = ops;
+ hwpt->domain->iommufd_hwpt = hwpt;
iommu_domain_set_sw_msi(hwpt->domain, iommufd_sw_msi);
if (WARN_ON_ONCE(hwpt->domain->type != IOMMU_DOMAIN_NESTED)) {
hwpt->domain = NULL;
goto out_abort;
}
+ hwpt->domain->iommufd_hwpt = hwpt;
hwpt->domain->owner = viommu->iommu_dev->ops;
iommu_domain_set_sw_msi(hwpt->domain, iommufd_sw_msi);
refcount_inc(&fault->obj.users);
iommufd_put_object(ucmd->ictx, &fault->obj);
}
- hwpt->domain->iommufd_hwpt = hwpt;
cmd->out_hwpt_id = hwpt->obj.id;
rc = iommufd_ucmd_respond(ucmd, sizeof(*cmd));