[dpdk-dev,v2] net/mlx5: fix memory region cache init
Checks
Commit Message
This patch moved MR cache init from device configuration function to
probe function to make sure init only once.
Fixes: 974f1e7ef146 ("net/mlx5: add new memory region support")
Cc: yskoh@mellanox.com
Signed-off-by: Xueming Li <xuemingl@mellanox.com>
---
drivers/net/mlx5/mlx5.c | 13 +++++++++++++
drivers/net/mlx5/mlx5_ethdev.c | 11 -----------
drivers/net/mlx5/mlx5_mr.c | 1 +
3 files changed, 14 insertions(+), 11 deletions(-)
Comments
> On May 26, 2018, at 6:28 AM, Xueming Li <xuemingl@mellanox.com> wrote:
>
> This patch moved MR cache init from device configuration function to
> probe function to make sure init only once.
>
> Fixes: 974f1e7ef146 ("net/mlx5: add new memory region support")
> Cc: yskoh@mellanox.com
>
> Signed-off-by: Xueming Li <xuemingl@mellanox.com>
> ---
Acked-by: Yongseok Koh <yskoh@mellanox.com>
Thanks
Saturday, May 26, 2018 8:08 PM, Yongseok Koh:
> Subject: Re: [PATCH v2] net/mlx5: fix memory region cache init
>
>
> > On May 26, 2018, at 6:28 AM, Xueming Li <xuemingl@mellanox.com>
> wrote:
> >
> > This patch moved MR cache init from device configuration function to
> > probe function to make sure init only once.
> >
> > Fixes: 974f1e7ef146 ("net/mlx5: add new memory region support")
> > Cc: yskoh@mellanox.com
> >
> > Signed-off-by: Xueming Li <xuemingl@mellanox.com>
> > ---
> Acked-by: Yongseok Koh <yskoh@mellanox.com>
Applied to next-net-mlx, thanks.
>
> Thanks
Ferruh,
Sunday, May 27, 2018 8:05 AM, Shahaf Shuler:
> Subject: Re: [dpdk-dev] [PATCH v2] net/mlx5: fix memory region cache init
>
> Saturday, May 26, 2018 8:08 PM, Yongseok Koh:
> > Subject: Re: [PATCH v2] net/mlx5: fix memory region cache init
> >
> >
> > > On May 26, 2018, at 6:28 AM, Xueming Li <xuemingl@mellanox.com>
> > wrote:
> > >
> > > This patch moved MR cache init from device configuration function to
> > > probe function to make sure init only once.
> > >
> > > Fixes: 974f1e7ef146 ("net/mlx5: add new memory region support")
> > > Cc: yskoh@mellanox.com
> > >
> > > Signed-off-by: Xueming Li <xuemingl@mellanox.com>
> > > ---
> > Acked-by: Yongseok Koh <yskoh@mellanox.com>
>
> Applied to next-net-mlx, thanks.
This is yet another critical fix for a bug caught recently.
This patch is to prevent deadlock when restarting the port.
Hope you will be able to pull this one in.
>
> >
> > Thanks
@@ -1193,6 +1193,19 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
goto port_error;
}
priv->config.max_verbs_prio = verb_priorities;
+ /*
+ * Once the device is added to the list of memory event
+ * callback, its global MR cache table cannot be expanded
+ * on the fly because of deadlock. If it overflows, lookup
+ * should be done by searching MR list linearly, which is slow.
+ */
+ err = mlx5_mr_btree_init(&priv->mr.cache,
+ MLX5_MR_BTREE_CACHE_N * 2,
+ eth_dev->device->numa_node);
+ if (err) {
+ err = rte_errno;
+ goto port_error;
+ }
/* Add device to memory callback list. */
rte_rwlock_write_lock(&mlx5_shared_data->mem_event_rwlock);
LIST_INSERT_HEAD(&mlx5_shared_data->mem_event_cb_list,
@@ -392,17 +392,6 @@ mlx5_dev_configure(struct rte_eth_dev *dev)
if (++j == rxqs_n)
j = 0;
}
- /*
- * Once the device is added to the list of memory event callback, its
- * global MR cache table cannot be expanded on the fly because of
- * deadlock. If it overflows, lookup should be done by searching MR list
- * linearly, which is slow.
- */
- if (mlx5_mr_btree_init(&priv->mr.cache, MLX5_MR_BTREE_CACHE_N * 2,
- dev->device->numa_node)) {
- /* rte_errno is already set. */
- return -rte_errno;
- }
return 0;
}
@@ -191,6 +191,7 @@ mlx5_mr_btree_init(struct mlx5_mr_btree *bt, int n, int socket)
rte_errno = EINVAL;
return -rte_errno;
}
+ assert(!bt->table && !bt->size);
memset(bt, 0, sizeof(*bt));
bt->table = rte_calloc_socket("B-tree table",
n, sizeof(struct mlx5_mr_cache),