[PATCH V3 0/8] IOMMU probe deferral support

[Lorenzo Pieralisi]

Hi Sricharan,

On Wed, Nov 30, 2016 at 06:04:13AM +0530, Sricharan wrote:
> Hi Lorenzo,
> 
> >-----Original Message-----
> >From: linux-arm-kernel [mailto:linux-arm-kernel-bounces at lists.infradead.org] On Behalf Of Lorenzo Pieralisi
> >Sent: Monday, November 28, 2016 11:44 PM
> >To: Sricharan <sricharan at codeaurora.org>
> >Cc: linux-arm-msm at vger.kernel.org; joro at 8bytes.org; will.deacon at arm.com; tfiga at chromium.org; iommu at lists.linux-
> >foundation.org; srinivas.kandagatla at linaro.org; laurent.pinchart at ideasonboard.com; 'Robin Murphy' <robin.murphy at arm.com>;
> >linux-arm-kernel at lists.infradead.org; m.szyprowski at samsung.com
> >Subject: Re: [PATCH V3 0/8] IOMMU probe deferral support
> >
> >On Mon, Nov 28, 2016 at 11:12:08PM +0530, Sricharan wrote:
> >
> >[...]
> >
> >> >Cool. We're rather hoping that the ACPI stuff is good to go for 4.10
> >> >now, so it's probably worth pulling the rest of that in (beyond the one
> >> >patch I picked) to make sure the of_dma_configure/acpi_dma_configure
> >> >paths don't inadvertently diverge.
> >> >
> >>
> >> I rebased and was testing your branch with Lorenzo's series. One thing
> >> i am still trying to get right is the acpi_dma_configure call. With your
> >> series dma_configure calls pci_dma/of_dma configure, so i am just adding
> >> acpi_dma_configure call there for non-pci ACPI devices as well. I see that
> >> acpi_dma_configure right now is called from acpi_bind_one and
> >> iort_add_smmu_platform_device, both go through the really_probe function
> >> path, so moving acpi_dma_configure from the above the two functions
> >> to dma_configure. I remember we discussed this on another thread, so
> >> hopefully it is correct. I do not have an platform to test the ACPI though.
> >> I will take some testing help on V4 for this.
> >
> >I am happy to test it for you please just send me a pointer at your v4
> >code.
> >
>  I posted the v4 and CCed you there. So i am little skeptical about the acpi
> changes that i have posted. I was checking for a function equivalent 
> in acpi as of_match_node in DT, to figure out if the iommu_spec.np that
> the master device is pointing to is there in the iommu_of_table and based
> on that we can decide if to defer the probe. I was seeing iort_scan_node
> was its equivalent. But if that is not correct, then last patch has to be reworked.
> Anyways will be good to know your feedback on this.

Sure I will test it asap, thanks for putting it together.

Thanks,
Lorenzo