On Sun, Mar 23, 2014 at 12:44 PM, Christoffer Dall
[off-list ref] wrote:

On Sat, Mar 22, 2014 at 09:54:55AM +0530, Anup Patel wrote:

quoted

On 22 March 2014 04:57, Christoffer Dall [off-list ref] wrote:

quoted

On Fri, Mar 21, 2014 at 06:23:25PM +0530, Anup Patel wrote:

quoted

Currently, the in-kernel PSCI emulation provides PSCI v0.1 interface to
VCPUs. This patch extends current in-kernel PSCI emulation to provide
PSCI v0.2 interface to VCPUs.

By default, ARM/ARM64 KVM will always provide PSCI v0.1 interface for
keeping the ABI backward-compatible.

To select PSCI v0.2 interface for VCPUs, the user space (i.e. QEMU or
KVMTOOL) will have to set KVM_ARM_VCPU_PSCI_0_2 feature when doing VCPU
init using KVM_ARM_VCPU_INIT ioctl.

Signed-off-by: Anup Patel <redacted>
Signed-off-by: Pranavkumar Sawargaonkar <redacted>

The content here can carry my Reviewed-by tag, but I think Rob's point
about sharing this code between KVM arm/arm64 and the PSCI code itself
it important, see:

http://lists.infradead.org/pipermail/linux-arm-kernel/2014-March/241792.html

My apologies, I totally missed Rob's comment.

No worries at all.

quoted

My suggestion would be to have:
arch/arm/include/uapi/asm/psci.h
arch/arm64/include/uapi/asm/psci.h

I think the suggestion was to move all this to include/uapi/linux/psci.h
or something equivalent so also arm and arm64 can share these defines,
but maybe I read too much into it.

Yes. It is perfectly fine for a somewhat architecture specific header
to live in include/linux (or include/uapi/linux).

Rob