Re: [PATCH 1/2] intel-iommu: PCIe hot plug for Intel iommu - IOMMU API

[David Woodhouse <dwmw2@infradead.org>]

On Fri, 2009-10-02 at 14:18 -0700, Fenghua Yu wrote:
> We expand existing IOMMU API iommu_detach_device for PCIe hot remove. We don't
> need IOMMU API for hot add.


I think this could do with more explanation. A hot remove will cause
detach_device() to be called with a NULL domain parameter? Surely we
shouldn't be printing an error with KERN_ERR just because such a device
hasn't been used for DMA before it was removed?

> Signed-off-by: Fenghua Yu <fenghua.yu@intel.com>
> ---
> 
>  drivers/pci/intel-iommu.c |   13 ++++++++++++-
>  1 files changed, 12 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/pci/intel-iommu.c b/drivers/pci/intel-iommu.c
> index 855dd7c..5c12283 100644
> --- a/drivers/pci/intel-iommu.c
> +++ b/drivers/pci/intel-iommu.c
> @@ -3559,9 +3559,20 @@ static int intel_iommu_attach_device(struct iommu_domain *domain,
>  static void intel_iommu_detach_device(struct iommu_domain *domain,
>  				      struct device *dev)
>  {
> -	struct dmar_domain *dmar_domain = domain->priv;
> +	struct dmar_domain *dmar_domain;
>  	struct pci_dev *pdev = to_pci_dev(dev);
>  
> +	/* If domain=NULL, we'll find a valid domain for the device.*/
> +	if (!domain) {
> +		dmar_domain = find_domain(pdev);
> +		if (!dmar_domain) {
> +			printk(KERN_ERR "Can not find IOMMU domain for %s\n",
> +			       pci_name(pdev));
> +			return;
> +		}
> +	} else
> +		dmar_domain = domain->priv;
> +
>  	domain_remove_one_dev_info(dmar_domain, pdev);
>  }
>  
-- 
dwmw2