Re: [RFC PATCH 0/5] virtio: Add a new layer to abstract pci access method

On 2016/01/21 20:07, Tetsuya Mukawa wrote:

This patch series are not for upstreaming.

It describe how to use a new access method abstraction of "virtio-pci.c".
Because of this, some patches are not for upstreaming.

For example, below changes will be shared with Jianfeng's patches.
So these changes are just temporary.
 - "--shm" option to allocate EAL memory.
 - Some changes to access to EAL memory by virtual address.

Anyway, some changes are not for upstreaming, but virtual virtio-net PMD
should work with QEMU as described in commit log.

Tetsuya Mukawa (5):
  virtio: Change the parameter order of io_write8/16/32()
  virtio: move rte_eal_pci_unmap_device() to virtio_pci.c
  virtio: Add a new layer to abstract pci access method
  EAL: Add new EAL "--shm" option.
  virtio: Extend virtio-net PMD to support container environment

 config/common_linuxapp                     |    1 +
 drivers/net/virtio/Makefile                |    4 +
 drivers/net/virtio/qtest.c                 | 1237 ++++++++++++++++++++++++++++
 drivers/net/virtio/virtio_ethdev.c         |  454 ++++++++--
 drivers/net/virtio/virtio_ethdev.h         |   12 +
 drivers/net/virtio/virtio_pci.c            |  732 ++++++++++++----
 drivers/net/virtio/virtio_pci.h            |   39 +-
 drivers/net/virtio/virtio_rxtx.c           |    3 +-
 drivers/net/virtio/virtqueue.h             |    9 +-
 lib/librte_eal/common/eal_common_options.c |    5 +
 lib/librte_eal/common/eal_internal_cfg.h   |    1 +
 lib/librte_eal/common/eal_options.h        |    2 +
 lib/librte_eal/common/include/rte_memory.h |    5 +
 lib/librte_eal/linuxapp/eal/eal_memory.c   |   76 ++
 14 files changed, 2337 insertions(+), 243 deletions(-)
 create mode 100644 drivers/net/virtio/qtest.c

Hi Yuanhan and Jianfeng,

Here is example how to use this new abstraction.
Please check first 3 patches to know how to implement the abstraction.

Also please see changes for "virtio_pci.c" involved in last patch.
This is the example.

Thanks,
Tetsuya