On 05.01.2012, at 10:54, Christian Borntraeger wrote:
> On some cpus the overhead for virtualization instructions is in the same
> range as a system call. Having to call multiple ioctls to get set registers
> will make certain userspace handled exits more expensive than necessary.
> Lets provide two sections in kvm_run to have a shared save area for
> guest registers.
> 1. the first section is read-only, to handle registers that have side-effects
> 2. the second section is read/write, e.g. for general purpose registers.
> We also provide two 64bit flags fields (architecture specific), that will
> specify which parts of these fields are valid. Each bit will define that
> a group of registers (like general purpose) or a single register is valid.
> In that way we can extend and shrink the interface. (The structure definition
> itself can only grow of course).
>
> Signed-off-by: Christian Borntraeger <borntraeger@xxxxxxxxxx>
> ---
> Documentation/virtual/kvm/api.txt | 28 ++++++++++++++++++++++++++++
> arch/ia64/include/asm/kvm.h | 7 +++++++
> arch/powerpc/include/asm/kvm.h | 7 +++++++
> arch/s390/include/asm/kvm.h | 6 ++++++
> arch/x86/include/asm/kvm.h | 7 +++++++
> include/linux/kvm.h | 20 ++++++++++++++++++++
> 6 files changed, 75 insertions(+), 0 deletions(-)
>
> diff --git a/Documentation/virtual/kvm/api.txt b/Documentation/virtual/kvm/api.txt
> index da1f8fd..a149e22 100644
> --- a/Documentation/virtual/kvm/api.txt
> +++ b/Documentation/virtual/kvm/api.txt
> @@ -1693,6 +1693,34 @@ developer registration required to access it).
> /* Fix the size of the union. */
> char padding[256];
> };
> +
> + /*
> + * Here are two fields that allow to access often used registers
> + * directly, to avoid the overhead of the ioctl system call. Each
> + * register can be extended or reduced by having flag bits specifying
> + * if a group of registers is valid.
> + */
> + __u64 kvm_valid_sync_ro_regs;
> + __u64 kvm_valid_sync_rw_regs;
> + union {
> + /* registers which can be only read */
> + struct kvm_sync_ro_regs sync_ro;
> + char padding[1024];
> + };
> + union {
> + /* read/write guest registers */
> + struct kvm_sync_rw_regs sync_rw;
> + char padding[1024];
> + };
Ah, new patch set. Same comment here. Please give the union a name. Also, I still don't understand why we need 2 structs. All we need is a bitmap that says "this field is dirty" and then kernel space can decide to ignore it.
Alex
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
