Re: [Qemu-devel] [RFC PATCH] spapr_vio/spapr_iommu: Move VIO bypass where it belongs

2015-01-13 Thread David Gibson
On Fri, Jan 09, 2015 at 12:02:51PM +1100, Alexey Kardashevskiy wrote: Instead of tweaking a TCE table device by adding there a bypass flag, let's add an alias to RAM and IOMMU memory region, and enable/disable those according to the selected bypass mode. This way IOMMU memory region can have

Re: [Qemu-devel] [RFC PATCH] spapr_vio/spapr_iommu: Move VIO bypass where it belongs

2015-01-13 Thread David Gibson
On Fri, Jan 09, 2015 at 01:59:25PM +0100, Paolo Bonzini wrote: On 09/01/2015 13:53, Alexey Kardashevskiy wrote: I think the minimum version should be 2 as well, because migrating from version 1 will not set the bypass field correctly. This why the patch is RFC :) I can keep the

Re: [Qemu-devel] [RFC PATCH] spapr_vio/spapr_iommu: Move VIO bypass where it belongs

2015-01-09 Thread Paolo Bonzini
On 09/01/2015 02:02, Alexey Kardashevskiy wrote: @@ -100,7 +98,7 @@ static const VMStateDescription vmstate_spapr_tce_table = { VMSTATE_UINT32_EQUAL(nb_table, sPAPRTCETable), /* IOMMU state */ -VMSTATE_BOOL(bypass, sPAPRTCETable), +

Re: [Qemu-devel] [RFC PATCH] spapr_vio/spapr_iommu: Move VIO bypass where it belongs

2015-01-09 Thread Paolo Bonzini
On 09/01/2015 13:53, Alexey Kardashevskiy wrote: I think the minimum version should be 2 as well, because migrating from version 1 will not set the bypass field correctly. This why the patch is RFC :) I can keep the flag in TCETable which would be a bit ugly but won't break migration.

Re: [Qemu-devel] [RFC PATCH] spapr_vio/spapr_iommu: Move VIO bypass where it belongs

2015-01-09 Thread Alexey Kardashevskiy
On 01/09/2015 09:05 PM, Paolo Bonzini wrote: On 09/01/2015 02:02, Alexey Kardashevskiy wrote: @@ -100,7 +98,7 @@ static const VMStateDescription vmstate_spapr_tce_table = { VMSTATE_UINT32_EQUAL(nb_table, sPAPRTCETable), /* IOMMU state */ -

[Qemu-devel] [RFC PATCH] spapr_vio/spapr_iommu: Move VIO bypass where it belongs

2015-01-08 Thread Alexey Kardashevskiy
Instead of tweaking a TCE table device by adding there a bypass flag, let's add an alias to RAM and IOMMU memory region, and enable/disable those according to the selected bypass mode. This way IOMMU memory region can have size of the actual window rather than ram_size which is essential for