On 06/26/2014 08:16 PM, Peter Maydell wrote:
Implement kvm_arm_vcpu_init() as a simple call to arm_arm_vcpu_init()
(which uses the KVM_ARM_VCPU_INIT vcpu ioctl to tell the kernel
to re-initialize the vCPU), rather than via the complicated code
which saves a copy of the register state on first init and then
writes it back to the kernel. This is much simpler and brings the
32-bit KVM code into line with the 64-bit code.
Signed-off-by: Peter Maydell <peter.mayd...@linaro.org>
---
The kernel has always supported being able to call VCPU_INIT
multiple times for this reset effect; I just didn't realize it
was possible when I wrote the original reset code.
When kvm64.c grows support for system registers we can probably
coalesce the two kvm_arm_reset_cpu() functions into one.
I also have a vague recollection that somebody reported that
we had an actual bug in this area that this patch would fix;
however I can't now find that in the mailing list archives :-(
I did: http://lists.gnu.org/archive/html/qemu-devel/2014-05/msg03131.html
Testing appreciated: my ARMv7 box is being a bit flaky at the
moment; I don't *think* the occasional weird stuff I see is
the effect of this patch but it's hard to be certain.
I will test your patch in the following days.
Diana
