qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Qemu-devel] Re: [PATCH v6 0/6] Inter-VM Shared Memory Device with migra


From: Cam Macdonell
Subject: [Qemu-devel] Re: [PATCH v6 0/6] Inter-VM Shared Memory Device with migration support
Date: Fri, 11 Jun 2010 16:03:47 -0600

Hi Anthony,

Is my implementation of master/peer roles acceptable?  I realize with
Alex's RAMList changes I may need to modify my patch, but is the
approach of marking memory non-migratable an acceptable
implementation?

Thanks,
Cam

On Fri, Jun 4, 2010 at 3:45 PM, Cam Macdonell <address@hidden> wrote:
> Latest patch for PCI shared memory device that maps a host shared memory 
> object
> to be shared between guests
>
> new in this series
>    - migration support with 'master' and 'peer' roles for guest to determine
>      who "owns" memory.  With 'master', the guest has the canonical copy of
>      the shared memory and will copy it with it on migration.  With 
> 'role=peer',
>      the guest will not copy the shared memory, but attach to what is on the
>      destination machine.
>    - modified phys_ram_dirty array for marking memory as not to be migrated
>    - add support for non-migrated memory regions
>
>    v5:
>    - fixed segfault for non-server case
>    - code style fixes
>    - removed limit on the number of guests
>    - shared memory server is now in qemu.git/contrib
>    - made ioeventfd setup function generic
>    - removed interrupts when guest joined (let application handle it)
>
>    v4:
>    - moved to single Doorbell register and use datamatch to trigger different
>      VMs rather than one register per eventfd
>    - remove writing arbitrary values to eventfds.  Only values of 1 are now
>      written to ensure correct usage
>
> Cam Macdonell (6):
>  Device specification for shared memory PCI device
>  Adds two new functions for assigning ioeventfd and irqfds.
>  Change phys_ram_dirty to phys_ram_status
>  Add support for marking memory to not be migrated.  On migration,
>    memory is checked for the NO_MIGRATION_FLAG.
>  Inter-VM shared memory PCI device
>  the stand-alone shared memory server for inter-VM shared memory
>
>  Makefile.target                         |    3 +
>  arch_init.c                             |   28 +-
>  contrib/ivshmem-server/Makefile         |   16 +
>  contrib/ivshmem-server/README           |   30 ++
>  contrib/ivshmem-server/ivshmem_server.c |  353 +++++++++++++
>  contrib/ivshmem-server/send_scm.c       |  208 ++++++++
>  contrib/ivshmem-server/send_scm.h       |   19 +
>  cpu-all.h                               |   18 +-
>  cpu-common.h                            |    2 +
>  docs/specs/ivshmem_device_spec.txt      |   96 ++++
>  exec.c                                  |   48 ++-
>  hw/ivshmem.c                            |  852 
> +++++++++++++++++++++++++++++++
>  kvm-all.c                               |   32 ++
>  kvm.h                                   |    1 +
>  qemu-char.c                             |    6 +
>  qemu-char.h                             |    3 +
>  qemu-doc.texi                           |   32 ++
>  17 files changed, 1710 insertions(+), 37 deletions(-)
>  create mode 100644 contrib/ivshmem-server/Makefile
>  create mode 100644 contrib/ivshmem-server/README
>  create mode 100644 contrib/ivshmem-server/ivshmem_server.c
>  create mode 100644 contrib/ivshmem-server/send_scm.c
>  create mode 100644 contrib/ivshmem-server/send_scm.h
>  create mode 100644 docs/specs/ivshmem_device_spec.txt
>  create mode 100644 hw/ivshmem.c
>
>



reply via email to

[Prev in Thread] Current Thread [Next in Thread]