Re: [PATCH v3 1/6] arm64: KVM: Add support for Stage-2 control of memory types and cacheability

2018-06-29 Thread Christoffer Dall
On Fri, Jun 29, 2018 at 10:21:27AM +0100, Marc Zyngier wrote: > On Fri, 29 Jun 2018 10:07:50 +0100, > Christoffer Dall wrote: > > > > On Fri, Jun 29, 2018 at 09:09:47AM +0100, Marc Zyngier wrote: > > > On Thu, 28 Jun 2018 21:56:38 +0100, > > > Christoffer Dall wrote: > > > > > > > > On Wed, Jun

Re: [PATCH v3 1/6] arm64: KVM: Add support for Stage-2 control of memory types and cacheability

2018-06-29 Thread Marc Zyngier
On Fri, 29 Jun 2018 10:07:50 +0100, Christoffer Dall wrote: > > On Fri, Jun 29, 2018 at 09:09:47AM +0100, Marc Zyngier wrote: > > On Thu, 28 Jun 2018 21:56:38 +0100, > > Christoffer Dall wrote: > > > > > > On Wed, Jun 27, 2018 at 01:20:53PM +0100, Marc Zyngier wrote: > > > > Up to ARMv8.3, the

Re: [PATCH v3 1/6] arm64: KVM: Add support for Stage-2 control of memory types and cacheability

2018-06-29 Thread Christoffer Dall
On Fri, Jun 29, 2018 at 09:09:47AM +0100, Marc Zyngier wrote: > On Thu, 28 Jun 2018 21:56:38 +0100, > Christoffer Dall wrote: > > > > On Wed, Jun 27, 2018 at 01:20:53PM +0100, Marc Zyngier wrote: > > > Up to ARMv8.3, the combinaison of Stage-1 and Stage-2 attributes > > > results in the strongest

Re: [PATCH v3 1/6] arm64: KVM: Add support for Stage-2 control of memory types and cacheability

2018-06-29 Thread Marc Zyngier
On Thu, 28 Jun 2018 21:56:38 +0100, Christoffer Dall wrote: > > On Wed, Jun 27, 2018 at 01:20:53PM +0100, Marc Zyngier wrote: > > Up to ARMv8.3, the combinaison of Stage-1 and Stage-2 attributes > > results in the strongest attribute of the two stages. This means > > that the hypervisor has to

Re: [PATCH v3 1/6] arm64: KVM: Add support for Stage-2 control of memory types and cacheability

2018-06-28 Thread Christoffer Dall
On Wed, Jun 27, 2018 at 01:20:53PM +0100, Marc Zyngier wrote: > Up to ARMv8.3, the combinaison of Stage-1 and Stage-2 attributes > results in the strongest attribute of the two stages. This means > that the hypervisor has to perform quite a lot of cache maintenance > just in case the guest has

Re: [PATCH v3 1/6] arm64: KVM: Add support for Stage-2 control of memory types and cacheability

2018-06-27 Thread Marc Zyngier
On 27/06/18 17:20, Suzuki K Poulose wrote: > Marc, > > On 27/06/18 13:20, Marc Zyngier wrote: >> Up to ARMv8.3, the combinaison of Stage-1 and Stage-2 attributes >> results in the strongest attribute of the two stages. This means >> that the hypervisor has to perform quite a lot of cache

Re: [PATCH v3 1/6] arm64: KVM: Add support for Stage-2 control of memory types and cacheability

2018-06-27 Thread Suzuki K Poulose
Marc, On 27/06/18 13:20, Marc Zyngier wrote: Up to ARMv8.3, the combinaison of Stage-1 and Stage-2 attributes results in the strongest attribute of the two stages. This means that the hypervisor has to perform quite a lot of cache maintenance just in case the guest has some non-cacheable

[PATCH v3 1/6] arm64: KVM: Add support for Stage-2 control of memory types and cacheability

2018-06-27 Thread Marc Zyngier
Up to ARMv8.3, the combinaison of Stage-1 and Stage-2 attributes results in the strongest attribute of the two stages. This means that the hypervisor has to perform quite a lot of cache maintenance just in case the guest has some non-cacheable mappings around. ARMv8.4 solves this problem by