Re: [Qemu-devel] [PATCH 0/7] kvm: fix system reset & rework slot management

[Anthony Liguori <aliguori@us.ibm.com>]

Jan Kiszka wrote:
> This is the promised series to improve KVM's memory slot management.
> Things turned out to be more complicated and ugly than planned due to an
> unfortunate bug in KVM's kernel code, see [1] and patch 5 of this
> series.
>
> The most important impact of this series is that it finally enables
> support for resetting qemu guests in KVM mode. Path 1 and 2 lay the
> groundwork for this, and the slot management rework fixes the remaining
> issues around memory remapping via PAM.
>
> Find the patches also at git://git.kiszka.org/qemu.git queues/kvm
>
> Jan Kiszka (7):
>       kvm: Sync CPU state on reset
>       kvm: Apply SMM-already-initialized workaround on reset
>       kvm: Cleanup unmap condition in kvm_set_phys_mem
>       kvm: Add sanity checks to slot management
>       kvm: improve handling of overlapping slots
>       vga: Fix inconsistent tracking of map_addr
>       vga: Cleanup dirty logging
>
>  hw/acpi.c       |   17 +++--
>  hw/cirrus_vga.c |   17 ----
>  hw/vga.c        |   27 +------
>  hw/vga_int.h    |    1 -
>  kvm-all.c       |  219 +++++++++++++++++++++++++++++++++++++++----------------
>  kvm.h           |    7 +-
>  vl.c            |    2 +
>  7 files changed, 177 insertions(+), 113 deletions(-)
>
> [1] http://permalink.gmane.org/gmane.comp.emulators.kvm.devel/30680
>
>   

I've applied this series but please submit a cleanup to #1.

I decided to just apply this to trunk.  It's a lot of churn for stable 
and I consider KVM experimental for 0.10.x.  If you think it should go 
in stable though, I'm willing to reconsider.

Thanks,

Regards,

Anthony Liguori