RE: [RFC PATCH] iommu/dma/pci: account pci host bridge dma_mask for IOVA allocation

2017-03-24 Thread Oza Oza via iommu
-...@vger.kernel.org; iommu@lists.linux-foundation.org; linux-ker...@vger.kernel.org; linux-arm-ker...@lists.infradead.org; devicet...@vger.kernel.org; bcm-kernel-feedback-l...@broadcom.com Subject: Re: [RFC PATCH] iommu/dma/pci: account pci host bridge dma_mask for IOVA allocation On 20/03/17 08:57, Oza Oza

Re: [RFC PATCH] iommu/dma/pci: account pci host bridge dma_mask for IOVA allocation

2017-03-20 Thread Robin Murphy
; linux-ker...@vger.kernel.org; > linux-arm-ker...@lists.infradead.org; devicet...@vger.kernel.org; > bcm-kernel-feedback-l...@broadcom.com; Oza Pawandeep > Subject: [RFC PATCH] iommu/dma: account pci host bridge dma_mask for IOVA > allocation > > It is possible that PCI device

[RFC PATCH] iommu/dma/pci: account pci host bridge dma_mask for IOVA allocation

2017-03-20 Thread Oza Oza via iommu
...@vger.kernel.org; bcm-kernel-feedback-l...@broadcom.com; Oza Pawandeep Subject: [RFC PATCH] iommu/dma: account pci host bridge dma_mask for IOVA allocation It is possible that PCI device supports 64-bit DMA addressing, and thus it's driver sets device's dma_mask to DMA_BIT_MASK(64), however PCI host