On 6/12/20 2:46 PM, Gaëtan Rivet wrote:
> On 11/06/20 23:37 +0200, Maxime Coquelin wrote:
>> This patch makes the vDPA framework to no more
>> support only PCI devices, but any devices by relying
>> on the generic device name as identifier.
>>
>> Signed-off-by: Maxime Coquelin <maxime.coque...@redhat.com>
>> ---
>> drivers/vdpa/ifc/ifcvf_vdpa.c | 6 +-
>> drivers/vdpa/mlx5/mlx5_vdpa.c | 8 +--
>> drivers/vdpa/mlx5/mlx5_vdpa.h | 2 +-
>> examples/vdpa/main.c | 49 ++++++++--------
>> lib/librte_vhost/rte_vdpa.h | 42 +++++++-------
>> lib/librte_vhost/rte_vhost_version.map | 1 +
>> lib/librte_vhost/vdpa.c | 79 +++++++++++---------------
>> 7 files changed, 85 insertions(+), 102 deletions(-)
>>
>
> [...]
>
>> diff --git a/drivers/vdpa/mlx5/mlx5_vdpa.c b/drivers/vdpa/mlx5/mlx5_vdpa.c
>> index 1113d6cef0..e8255c7d7e 100644
>> --- a/drivers/vdpa/mlx5/mlx5_vdpa.c
>> +++ b/drivers/vdpa/mlx5/mlx5_vdpa.c
>> @@ -501,14 +501,13 @@ mlx5_vdpa_pci_probe(struct rte_pci_driver *pci_drv
>> __rte_unused,
>> priv->caps = attr.vdpa;
>> priv->log_max_rqt_size = attr.log_max_rqt_size;
>> priv->ctx = ctx;
>> - priv->dev_addr.pci_addr = pci_dev->addr;
>> - priv->dev_addr.type = VDPA_ADDR_PCI;
>> + priv->pci_dev = pci_dev;
>> priv->var = mlx5_glue->dv_alloc_var(ctx, 0);
>> if (!priv->var) {
>> DRV_LOG(ERR, "Failed to allocate VAR %u.\n", errno);
>> goto error;
>> }
>> - priv->id = rte_vdpa_register_device(&priv->dev_addr, &mlx5_vdpa_ops);
>> + priv->id = rte_vdpa_register_device(&pci_dev->device, &mlx5_vdpa_ops);
>> if (priv->id < 0) {
>> DRV_LOG(ERR, "Failed to register vDPA device.");
>> rte_errno = rte_errno ? rte_errno : EINVAL;
>> @@ -550,8 +549,7 @@ mlx5_vdpa_pci_remove(struct rte_pci_device *pci_dev)
>>
>> pthread_mutex_lock(&priv_list_lock);
>> TAILQ_FOREACH(priv, &priv_list, next) {
>> - if (memcmp(&priv->dev_addr.pci_addr, &pci_dev->addr,
>> - sizeof(pci_dev->addr)) == 0) {
>> + if (priv->pci_dev == pci_dev) {
>
> Assignment is ok I think but comparison is not, rte_pci_addr_cmp()
> should be used here instead. (memcmp was not correct either.)
>
> The structure can be padded depending on arch.
>
> Everything else seems ok to me.
>
Ok, I will change to use rte_pci_addr_cmp() in next revision.
Thanks,
Maxime
