[+ Christoffer, Will, Catalin - In the future, please Cc the relevant maintainers] On 09/12/16 20:59, Caraman Mihai Claudiu-B02008 wrote: > From: Mihai Caraman <mike.caraman@xxxxxxx> > > Stage 2 non-Gathering and non-Reordering memory attributes have precedence over stage 1 > Gathering and Reordering attributes. Honor guest device memory type mappings using GRE > memory type at stage 2. This feels incredibly scary. One you start enabling gathering and reordering, there is nothing that prevents a write to a device from a given guest to be merged or reordered with a write from another guest to the same device (think of the GIC). This would only require one of the guests to use GRE attributes itself. So as it stands, my feeling is that it is a big NO. > This is a preamble for a bigger change that will honor guest normal memory mappings for > targets connected to HN-F interfaces. Normal memory mappings for devices? Unless you can prove that this is a valid architectural configuration, this is unlikely to happen. And I'm not keen on taking any such change before seeing the whole thing. > > Signed-off-by: Mihai Caraman <mike.caraman@xxxxxxx> > --- > arch/arm64/include/asm/memory.h | 2 +- > arch/arm64/include/asm/pgtable-prot.h | 2 +- > 2 files changed, 2 insertions(+), 2 deletions(-) > > diff --git a/arch/arm64/include/asm/memory.h b/arch/arm64/include/asm/memory.h > index b71086d..d4834ba 100644 > --- a/arch/arm64/include/asm/memory.h > +++ b/arch/arm64/include/asm/memory.h > @@ -134,7 +134,7 @@ > * Memory types for Stage-2 translation > */ > #define MT_S2_NORMAL 0xf > -#define MT_S2_DEVICE_nGnRE 0x1 > +#define MT_S2_DEVICE_GRE 0x3 > > #ifdef CONFIG_ARM64_4K_PAGES > #define IOREMAP_MAX_ORDER (PUD_SHIFT) > diff --git a/arch/arm64/include/asm/pgtable-prot.h b/arch/arm64/include/asm/pgtable-prot.h > index 2142c77..11b43f7 100644 > --- a/arch/arm64/include/asm/pgtable-prot.h > +++ b/arch/arm64/include/asm/pgtable-prot.h > @@ -61,7 +61,7 @@ > #define PAGE_HYP_DEVICE __pgprot(PROT_DEVICE_nGnRE | PTE_HYP) > > #define PAGE_S2 __pgprot(PROT_DEFAULT | PTE_S2_MEMATTR(MT_S2_NORMAL) | PTE_S2_RDONLY) > -#define PAGE_S2_DEVICE __pgprot(PROT_DEFAULT | PTE_S2_MEMATTR(MT_S2_DEVICE_nGnRE) | PTE_S2_RDONLY | PTE_UXN) > +#define PAGE_S2_DEVICE __pgprot(PROT_DEFAULT | PTE_S2_MEMATTR(MT_S2_DEVICE_GRE) | PTE_S2_RDONLY | PTE_UXN) > > #define PAGE_NONE __pgprot(((_PAGE_DEFAULT) & ~PTE_VALID) | PTE_PROT_NONE | PTE_PXN | PTE_UXN) > #define PAGE_SHARED __pgprot(_PAGE_DEFAULT | PTE_USER | PTE_NG | PTE_PXN | PTE_UXN | PTE_WRITE) > Thanks, M. -- Jazz is not dead. It just smells funny... _______________________________________________ kvmarm mailing list kvmarm@xxxxxxxxxxxxxxxxxxxxx https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
