Re: linux-next: manual merge of the kvm-ppc tree with the powerpc-merge tree

On 10/11/2012 08:04:58 AM, Alexander Graf wrote:

=20
On 11.10.2012, at 05:32, Tabi Timur-B04825 wrote:
=20

quoted

On Wed, Oct 10, 2012 at 9:47 PM, Stephen Rothwell =20

[off-list ref] wrote:

quoted

quoted

Commit 549d62d889b4 ("KVM: PPC: use definitions in epapr header
for hcalls") from the kvm-ppc tree added an include of =20

asm/epapr_hcall.h

quoted

quoted

to the user visible part of asm/kvm_para.h so asm/epapr_hcall.h =20

became a

quoted

quoted

user visible header file.

Any real user-space code that tries to call any of the functions in
epapr_hcall.h will cause an exception.

Actually it'll cause a linker error, as of the "Don't use hardcoded =20
opcode for ePAPR hcall invocation" patch.

quoted

Claiming that kernel header files that KVM needs are suddenly
user-space header files doesn't make much sense to me, but I guess
it's not my decision.

=20
This is about kvm_host.h, which is the part that is exported to user =20
space. It usually contains constants and structs that are required =20
for the API.

s/kvm_host/kvm_para/

I'm not sure anything in kvm_para.h or epapr_hcalls.h would be of use =20
to userspace, but it would be of use to the guest OS, so it's still not =20
kernel-internal.

The variable externs should be #ifdef __KERNEL__ as David points out.  =20
The asm stuff could still be useful in a guest OS if the guest provides =20
epapr_hypercall_start.

-Scott=

On 11.10.2012, at 17:50, Scott Wood wrote:

On 10/11/2012 08:04:58 AM, Alexander Graf wrote:

quoted

On 11.10.2012, at 05:32, Tabi Timur-B04825 wrote:

quoted

On Wed, Oct 10, 2012 at 9:47 PM, Stephen Rothwell =

[off-list ref] wrote:

quoted

quoted

quoted

Commit 549d62d889b4 ("KVM: PPC: use definitions in epapr header
for hcalls") from the kvm-ppc tree added an include of =

asm/epapr_hcall.h

quoted

quoted

quoted

to the user visible part of asm/kvm_para.h so asm/epapr_hcall.h =

became a

quoted

quoted

quoted

user visible header file.

Any real user-space code that tries to call any of the functions in
epapr_hcall.h will cause an exception.

=20
Actually it'll cause a linker error, as of the "Don't use hardcoded =

opcode for ePAPR hcall invocation" patch.

=20

quoted

quoted

Claiming that kernel header files that KVM needs are suddenly
user-space header files doesn't make much sense to me, but I guess
it's not my decision.

This is about kvm_host.h, which is the part that is exported to user =

space. It usually contains constants and structs that are required for =
the API.

=20
s/kvm_host/kvm_para/
=20
I'm not sure anything in kvm_para.h or epapr_hcalls.h would be of use =

to userspace, but it would be of use to the guest OS, so it's still not =
kernel-internal.

=20
The variable externs should be #ifdef __KERNEL__ as David points out.  =

The asm stuff could still be useful in a guest OS if the guest provides =
epapr_hypercall_start.

I don't think we should worry about guest kernel code potentially having =
use for this. Let's make the user space headers be great for user space =
usage, not for guest code usage :).

Alex=