Re: [Qemu-devel] [PATCH v4 3/8] intel_iommu: pass whole remapped addresses to apic

2016-10-10 Thread Igor Mammedov
On Sat, 8 Oct 2016 14:14:09 +0800 Peter Xu wrote: > On Fri, Oct 07, 2016 at 06:24:15PM +0200, Radim Krčmář wrote: > > [...] > > > KVM accepts the address in host endianess and QEMU/VTD code also uses > > host endianess for internal representation of memory addresses, so this > > hunk should be

Re: [Qemu-devel] [PATCH v4 3/8] intel_iommu: pass whole remapped addresses to apic

2016-10-07 Thread Peter Xu
On Wed, Oct 05, 2016 at 03:06:52PM +0200, Radim Krčmář wrote: > The MMIO interface to APIC only allowed 8 bit addresses, which is not > enough for 32 bit addresses from EIM remapping. > Intel stored upper 24 bits in the high MSI address, so use the same > technique. The technique is also used in KV

Re: [Qemu-devel] [PATCH v4 3/8] intel_iommu: pass whole remapped addresses to apic

2016-10-07 Thread Peter Xu
On Sat, Oct 08, 2016 at 02:14:09PM +0800, Peter Xu wrote: > On Fri, Oct 07, 2016 at 06:24:15PM +0200, Radim Krčmář wrote: > > [...] > > > KVM accepts the address in host endianess and QEMU/VTD code also uses > > host endianess for internal representation of memory addresses, so this > > hunk shou

Re: [Qemu-devel] [PATCH v4 3/8] intel_iommu: pass whole remapped addresses to apic

2016-10-07 Thread Peter Xu
On Fri, Oct 07, 2016 at 06:24:15PM +0200, Radim Krčmář wrote: [...] > KVM accepts the address in host endianess and QEMU/VTD code also uses > host endianess for internal representation of memory addresses, so this > hunk should be fine. > > It is confusing, because the VTD is definitely broken w

Re: [Qemu-devel] [PATCH v4 3/8] intel_iommu: pass whole remapped addresses to apic

2016-10-07 Thread Radim Krčmář
2016-10-07 15:05+0200, Igor Mammedov: > On Wed, 5 Oct 2016 15:06:52 +0200 > Radim Krčmář wrote: > >> The MMIO interface to APIC only allowed 8 bit addresses, which is not >> enough for 32 bit addresses from EIM remapping. >> Intel stored upper 24 bits in the high MSI address, so use the same >>

Re: [Qemu-devel] [PATCH v4 3/8] intel_iommu: pass whole remapped addresses to apic

2016-10-07 Thread Igor Mammedov
On Wed, 5 Oct 2016 15:06:52 +0200 Radim Krčmář wrote: > The MMIO interface to APIC only allowed 8 bit addresses, which is not > enough for 32 bit addresses from EIM remapping. > Intel stored upper 24 bits in the high MSI address, so use the same > technique. The technique is also used in KVM MSI

[Qemu-devel] [PATCH v4 3/8] intel_iommu: pass whole remapped addresses to apic

2016-10-05 Thread Radim Krčmář
The MMIO interface to APIC only allowed 8 bit addresses, which is not enough for 32 bit addresses from EIM remapping. Intel stored upper 24 bits in the high MSI address, so use the same technique. The technique is also used in KVM MSI interface. Other APICs are unlikely to handle those upper bits.