[dpdk-dev,RFC,v3,1/2] vhost: Add callback and private data for vhost PMD
Commit Message
These variables are needed to be able to manage one of virtio devices
using both vhost library APIs and vhost PMD.
For example, if vhost PMD uses current callback handler and private data
provided by vhost library, A DPDK application that links vhost library
cannot use some of vhost library APIs. To avoid it, callback and private
data for vhost PMD are needed.
Signed-off-by: Tetsuya Mukawa <mukawa@igel.co.jp>
---
lib/librte_vhost/rte_virtio_net.h | 3 ++
lib/librte_vhost/vhost_user/virtio-net-user.c | 8 +++---
lib/librte_vhost/virtio-net.c | 40 +++++++++++++++++++++++++--
lib/librte_vhost/virtio-net.h | 3 +-
4 files changed, 47 insertions(+), 7 deletions(-)
Comments
The patch introduces a new PMD. This PMD is implemented as thin wrapper
of librte_vhost. The patch will work on below patch series.
- [PATCH v5 00/28] remove pci driver from vdevs
* Known issue.
We may see issues while handling RESET_OWNER message.
These handlings are done in vhost library, so not a part of vhost PMD.
So far, we are waiting for QEMU fixing.
PATCH v4 changes:
- Support vhost multiple queues.
- Rebase on "remove pci driver from vdevs".
- Optimize RX/TX functions.
- Fix resource leaks.
- Fix compile issue.
- Add patch to fix vhost library.
PATCH v3 changes:
- Optimize performance.
In RX/TX functions, change code to access only per core data.
- Add below API to allow user to use vhost library APIs for a port managed
by vhost PMD. There are a few limitations. See "rte_eth_vhost.h".
- rte_eth_vhost_portid2vdev()
To support this functionality, vhost library is also changed.
Anyway, if users doesn't use vhost PMD, can fully use vhost library APIs.
- Add code to support vhost multiple queues.
Actually, multiple queues functionality is not enabled so far.
PATCH v2 changes:
- Fix issues reported by checkpatch.pl
(Thanks to Stephen Hemminger)
Tetsuya Mukawa (3):
vhost: Fix wrong handling of virtqueue array index
vhost: Add callback and private data for vhost PMD
vhost: Add VHOST PMD
config/common_linuxapp | 6 +
doc/guides/nics/index.rst | 1 +
doc/guides/nics/vhost.rst | 82 +++
doc/guides/rel_notes/release_2_2.rst | 2 +
drivers/net/Makefile | 4 +
drivers/net/vhost/Makefile | 62 +++
drivers/net/vhost/rte_eth_vhost.c | 765 ++++++++++++++++++++++++++
drivers/net/vhost/rte_eth_vhost.h | 65 +++
drivers/net/vhost/rte_pmd_vhost_version.map | 8 +
lib/librte_vhost/rte_vhost_version.map | 6 +
lib/librte_vhost/rte_virtio_net.h | 3 +
lib/librte_vhost/vhost_user/virtio-net-user.c | 33 +-
lib/librte_vhost/virtio-net.c | 61 +-
lib/librte_vhost/virtio-net.h | 4 +-
mk/rte.app.mk | 8 +-
15 files changed, 1085 insertions(+), 25 deletions(-)
create mode 100644 doc/guides/nics/vhost.rst
create mode 100644 drivers/net/vhost/Makefile
create mode 100644 drivers/net/vhost/rte_eth_vhost.c
create mode 100644 drivers/net/vhost/rte_eth_vhost.h
create mode 100644 drivers/net/vhost/rte_pmd_vhost_version.map
Below patch has been submitted as a separate patch.
- [dpdk-dev,1/3] vhost: Fix wrong handling of virtqueue array index
(http://dpdk.org/dev/patchwork/patch/8038/)
Tetsuya
On 2015/10/27 15:12, Tetsuya Mukawa wrote:
> The patch introduces a new PMD. This PMD is implemented as thin wrapper
> of librte_vhost. The patch will work on below patch series.
> - [PATCH v5 00/28] remove pci driver from vdevs
>
> * Known issue.
> We may see issues while handling RESET_OWNER message.
> These handlings are done in vhost library, so not a part of vhost PMD.
> So far, we are waiting for QEMU fixing.
>
> PATCH v4 changes:
> - Support vhost multiple queues.
> - Rebase on "remove pci driver from vdevs".
> - Optimize RX/TX functions.
> - Fix resource leaks.
> - Fix compile issue.
> - Add patch to fix vhost library.
>
> PATCH v3 changes:
> - Optimize performance.
> In RX/TX functions, change code to access only per core data.
> - Add below API to allow user to use vhost library APIs for a port managed
> by vhost PMD. There are a few limitations. See "rte_eth_vhost.h".
> - rte_eth_vhost_portid2vdev()
> To support this functionality, vhost library is also changed.
> Anyway, if users doesn't use vhost PMD, can fully use vhost library APIs.
> - Add code to support vhost multiple queues.
> Actually, multiple queues functionality is not enabled so far.
>
> PATCH v2 changes:
> - Fix issues reported by checkpatch.pl
> (Thanks to Stephen Hemminger)
>
>
> Tetsuya Mukawa (3):
> vhost: Fix wrong handling of virtqueue array index
> vhost: Add callback and private data for vhost PMD
> vhost: Add VHOST PMD
>
> config/common_linuxapp | 6 +
> doc/guides/nics/index.rst | 1 +
> doc/guides/nics/vhost.rst | 82 +++
> doc/guides/rel_notes/release_2_2.rst | 2 +
> drivers/net/Makefile | 4 +
> drivers/net/vhost/Makefile | 62 +++
> drivers/net/vhost/rte_eth_vhost.c | 765 ++++++++++++++++++++++++++
> drivers/net/vhost/rte_eth_vhost.h | 65 +++
> drivers/net/vhost/rte_pmd_vhost_version.map | 8 +
> lib/librte_vhost/rte_vhost_version.map | 6 +
> lib/librte_vhost/rte_virtio_net.h | 3 +
> lib/librte_vhost/vhost_user/virtio-net-user.c | 33 +-
> lib/librte_vhost/virtio-net.c | 61 +-
> lib/librte_vhost/virtio-net.h | 4 +-
> mk/rte.app.mk | 8 +-
> 15 files changed, 1085 insertions(+), 25 deletions(-)
> create mode 100644 doc/guides/nics/vhost.rst
> create mode 100644 drivers/net/vhost/Makefile
> create mode 100644 drivers/net/vhost/rte_eth_vhost.c
> create mode 100644 drivers/net/vhost/rte_eth_vhost.h
> create mode 100644 drivers/net/vhost/rte_pmd_vhost_version.map
>
2015-10-27 16:54, Tetsuya Mukawa:
> Below patch has been submitted as a separate patch.
>
> - [dpdk-dev,1/3] vhost: Fix wrong handling of virtqueue array index
> (http://dpdk.org/dev/patchwork/patch/8038/)
Please could you rebase only the two last patches?
Thanks
PS:
WARNING:TYPO_SPELLING: 'failuer' may be misspelled - perhaps 'failure'?
#606: FILE: drivers/net/vhost/rte_eth_vhost.c:272:
+ RTE_LOG(INFO, PMD, "failuer to find a ethdev\n");
WARNING:TYPO_SPELLING: 'accesing' may be misspelled - perhaps 'accessing'?
#612: FILE: drivers/net/vhost/rte_eth_vhost.c:278:
+ /* Wait until rx/tx_pkt_burst stops accesing vhost device */
On 2015/10/31 3:30, Thomas Monjalon wrote:
> 2015-10-27 16:54, Tetsuya Mukawa:
>> Below patch has been submitted as a separate patch.
>>
>> - [dpdk-dev,1/3] vhost: Fix wrong handling of virtqueue array index
>> (http://dpdk.org/dev/patchwork/patch/8038/)
> Please could you rebase only the two last patches?
> Thanks
>
> PS:
> WARNING:TYPO_SPELLING: 'failuer' may be misspelled - perhaps 'failure'?
> #606: FILE: drivers/net/vhost/rte_eth_vhost.c:272:
> + RTE_LOG(INFO, PMD, "failuer to find a ethdev\n");
> WARNING:TYPO_SPELLING: 'accesing' may be misspelled - perhaps 'accessing'?
> #612: FILE: drivers/net/vhost/rte_eth_vhost.c:278:
> + /* Wait until rx/tx_pkt_burst stops accesing vhost device */
>
Hi Thomas,
Thank you so much for checking my patches.
I have fixed a few typos, and rebased on latest tree (with Bernard's patch).
I will submit again soon.
Regards,
Tetsuya
@@ -108,6 +108,7 @@ struct virtio_net {
uint32_t virt_qp_nb;
uint32_t mem_idx; /** Used in set memory layout, unique for each queue within virtio device. */
void *priv; /**< private context */
+ void *pmd_priv; /**< private context for vhost PMD */
} __rte_cache_aligned;
/**
@@ -198,6 +199,8 @@ int rte_vhost_driver_unregister(const char *dev_name);
/* Register callbacks. */
int rte_vhost_driver_callback_register(struct virtio_net_device_ops const * const);
+/* Register callbacks for vhost PMD (Only for internal). */
+int rte_vhost_driver_pmd_callback_register(struct virtio_net_device_ops const * const);
/* Start vhost driver session blocking loop. */
int rte_vhost_driver_session_start(void);
@@ -288,7 +288,7 @@ user_set_vring_kick(struct vhost_device_ctx ctx, struct VhostUserMsg *pmsg)
if (virtio_is_ready(dev) &&
!(dev->flags & VIRTIO_DEV_RUNNING))
- notify_ops->new_device(dev);
+ notify_new_device(dev);
}
/*
@@ -302,7 +302,7 @@ user_get_vring_base(struct vhost_device_ctx ctx,
/* We have to stop the queue (virtio) if it is running. */
if (dev->flags & VIRTIO_DEV_RUNNING)
- notify_ops->destroy_device(dev);
+ notify_destroy_device(dev);
/* Here we are safe to get the last used index */
ops->get_vring_base(ctx, state->index, state);
@@ -333,7 +333,7 @@ user_reset_owner(struct vhost_device_ctx ctx,
/* We have to stop the queue (virtio) if it is running. */
if (dev->flags & VIRTIO_DEV_RUNNING)
- notify_ops->destroy_device(dev);
+ notify_destroy_device(dev);
RTE_LOG(INFO, VHOST_CONFIG,
"reset owner --- state idx:%d state num:%d\n", state->index, state->num);
@@ -379,7 +379,7 @@ user_destroy_device(struct vhost_device_ctx ctx)
uint32_t i;
if (dev && (dev->flags & VIRTIO_DEV_RUNNING))
- notify_ops->destroy_device(dev);
+ notify_destroy_device(dev);
for (i = 0; i < dev->virt_qp_nb; i++)
if (dev && dev->mem_arr[i]) {
@@ -64,6 +64,8 @@ struct virtio_net_config_ll {
/* device ops to add/remove device to/from data core. */
struct virtio_net_device_ops const *notify_ops;
+/* device ops for vhost PMD to add/remove device to/from data core. */
+struct virtio_net_device_ops const *pmd_notify_ops;
/* root address of the linked list of managed virtio devices */
static struct virtio_net_config_ll *ll_root;
@@ -84,6 +86,29 @@ static uint64_t VHOST_FEATURES = VHOST_SUPPORTED_FEATURES;
static uint64_t VHOST_PROTOCOL_FEATURES = VHOST_SUPPORTED_PROTOCOL_FEATURES;
+int
+notify_new_device(struct virtio_net *dev)
+{
+ if ((pmd_notify_ops != NULL) && (pmd_notify_ops->new_device != NULL)) {
+ int ret = pmd_notify_ops->new_device(dev);
+ if (ret != 0)
+ return ret;
+ }
+ if ((notify_ops != NULL) && (notify_ops->new_device != NULL))
+ return notify_ops->new_device(dev);
+
+ return 0;
+}
+
+void
+notify_destroy_device(volatile struct virtio_net *dev)
+{
+ if ((pmd_notify_ops != NULL) && (pmd_notify_ops->destroy_device != NULL))
+ pmd_notify_ops->destroy_device(dev);
+ if ((notify_ops != NULL) && (notify_ops->destroy_device != NULL))
+ notify_ops->destroy_device(dev);
+}
+
/*
* Converts QEMU virtual address to Vhost virtual address. This function is
* used to convert the ring addresses to our address space.
@@ -421,7 +446,7 @@ destroy_device(struct vhost_device_ctx ctx)
* the function to remove it from the data core.
*/
if ((ll_dev_cur->dev.flags & VIRTIO_DEV_RUNNING))
- notify_ops->destroy_device(&(ll_dev_cur->dev));
+ notify_destroy_device(&(ll_dev_cur->dev));
ll_dev_cur = rm_config_ll_entry(ll_dev_cur,
ll_dev_last);
} else {
@@ -884,7 +909,7 @@ set_backend(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
dev->virtqueue[VIRTIO_RXQ]->enabled = 1;
dev->virtqueue[VIRTIO_TXQ]->enabled = 1;
}
- return notify_ops->new_device(dev);
+ return notify_new_device(dev);
}
/* Otherwise we remove it. */
} else
@@ -1006,3 +1031,14 @@ rte_vhost_driver_callback_register(struct virtio_net_device_ops const * const op
return 0;
}
+
+/*
+ * Register ops so that we can add/remove device to data core.
+ */
+int
+rte_vhost_driver_pmd_callback_register(struct virtio_net_device_ops const * const ops)
+{
+ pmd_notify_ops = ops;
+
+ return 0;
+}
@@ -39,7 +39,8 @@
#define VHOST_USER_PROTOCOL_F_VRING_FLAG 2
-struct virtio_net_device_ops const *notify_ops;
struct virtio_net *get_device(struct vhost_device_ctx ctx);
+int notify_new_device(struct virtio_net *dev);
+void notify_destroy_device(volatile struct virtio_net *dev);
#endif