| Message ID | 1469093572-6793-1-git-send-email-yuval.shaia@oracle.com (mailing list archive) |
|---|---|
| State | RFC |
| Headers | show |
On 7/21/2016 2:32 AM, Yuval Shaia wrote: > Adding sysfs entries to read/write VF's Node GUID. > This abbility will enable sys-admin to configure node GUID for VFs the same > way it is done for VF's port GUIDs. > > Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com> > --- This is inline with port GUID for paravirtualized HCAs and seems reasonable to me. > drivers/infiniband/hw/mlx4/sysfs.c | 72 ++++++++++++++++++++++++++++++++++++ > 1 files changed, 72 insertions(+), 0 deletions(-) > > diff --git a/drivers/infiniband/hw/mlx4/sysfs.c b/drivers/infiniband/hw/mlx4/sysfs.c > index 69fb5ba..6fc4366 100644 > --- a/drivers/infiniband/hw/mlx4/sysfs.c > +++ b/drivers/infiniband/hw/mlx4/sysfs.c > @@ -371,6 +371,7 @@ struct mlx4_port { > struct attribute_group gid_group; > struct device_attribute enable_smi_admin; > struct device_attribute smi_enabled; > + struct device_attribute node_guid; > int slave; > u8 port_num; > }; > @@ -620,6 +621,7 @@ static int add_vf_smi_entries(struct mlx4_port *p) > sysfs_remove_file(&p->kobj, &p->smi_enabled.attr); > return ret; > } > + Stray change. > return 0; > } > > @@ -635,6 +637,69 @@ static void remove_vf_smi_entries(struct mlx4_port *p) > sysfs_remove_file(&p->kobj, &p->enable_smi_admin.attr); > } > > +static ssize_t sysfs_show_node_guid(struct device *dev, > + struct device_attribute *attr, > + char *buf) > +{ > + struct mlx4_port *p = container_of(attr, struct mlx4_port, node_guid); > + ssize_t len = 0; > + __be64 guid = mlx4_get_slave_node_guid(p->dev->dev, p->slave); > + > + len = sprintf(buf, "slave %d 0x%llx\n", p->slave, be64_to_cpu(guid)); > + > + return len; > +} > + > +static ssize_t sysfs_store_node_guid(struct device *dev, > + struct device_attribute *attr, > + const char *buf, size_t count) > +{ > + struct mlx4_port *p = container_of(attr, struct mlx4_port, node_guid); > + __be64 guid; > + Shouldn't we check only master can set this up ? > + if (sscanf(buf, "0x%llx", &guid) != 1) > + return -EINVAL; > + > + mlx4_put_slave_node_guid(p->dev->dev, p->slave, cpu_to_be64(guid)); > + > + return count; > +} > + > +static int add_vf_admin_guid_entry(struct mlx4_port *p) > +{ > + int is_eth = rdma_port_get_link_layer(&p->dev->ib_dev, p->port_num) == > + IB_LINK_LAYER_ETHERNET; > + int ret; > + > + /* do not display entries if eth transport, or if master */ > + if (is_eth || p->slave == mlx4_master_func_num(p->dev->dev)) > + return 0; > + > + sysfs_attr_init(&p->node_guid.attr); > + p->node_guid.show = sysfs_show_node_guid; > + p->node_guid.store = sysfs_store_node_guid; > + p->node_guid.attr.name = "node_guid"; > + p->node_guid.attr.mode = 0644; > + ret = sysfs_create_file(&p->kobj, &p->node_guid.attr); > + if (ret) { > + pr_err("failed to create sysfs entry for node_guid\n"); Use existing driver macro or append print with "mlx4_core:" > + return ret; > + } > + > + return 0; > +} > + > +static void remove_vf_node_guid_entry(struct mlx4_port *p) > +{ > + int is_eth = rdma_port_get_link_layer(&p->dev->ib_dev, p->port_num) == > + IB_LINK_LAYER_ETHERNET; > + > + if (is_eth || p->slave == mlx4_master_func_num(p->dev->dev)) > + return; > + > + sysfs_remove_file(&p->kobj, &p->node_guid.attr); > +} > + > static int add_port(struct mlx4_ib_dev *dev, int port_num, int slave) > { > struct mlx4_port *p; > @@ -686,6 +751,10 @@ static int add_port(struct mlx4_ib_dev *dev, int port_num, int slave) > if (ret) > goto err_free_gid; > > + ret = add_vf_admin_guid_entry(p); > + if (ret) > + goto err_free_gid; > + > list_add_tail(&p->kobj.entry, &dev->pkeys.pkey_port_list[slave]); > return 0; > > @@ -743,6 +812,7 @@ static int register_one_pkey_tree(struct mlx4_ib_dev *dev, int slave) > if (err) > goto err_add; > } > + Stray change. Regards, Santosh -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 7/21/2016 6:12 PM, Santosh Shilimkar wrote: > On 7/21/2016 2:32 AM, Yuval Shaia wrote: >> Adding sysfs entries to read/write VF's Node GUID. >> This abbility will enable sys-admin to configure node GUID for VFs the typo >> same >> way it is done for VF's port GUIDs. >> Hi Yuval, The following commit in iproute2 by Eli Cohen added a standard way to configure node/port GUID using IPoIB netdev: d91fb3f4c7e4 Add support for configuring Infiniband GUIDs The configuration will be done in the following way: ip link set dev ib0 vf 0 node_guid 00:02:c9:03:00:21:6e:70 ip link set dev ib0 vf 0 port_guid 00:02:c9:03:00:21:6e:78 Can't this be used here as well? Thanks >> Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com> >> --- > This is inline with port GUID for paravirtualized HCAs and seems > reasonable to me. > >> drivers/infiniband/hw/mlx4/sysfs.c | 72 >> ++++++++++++++++++++++++++++++++++++ >> 1 files changed, 72 insertions(+), 0 deletions(-) >> >> diff --git a/drivers/infiniband/hw/mlx4/sysfs.c >> b/drivers/infiniband/hw/mlx4/sysfs.c >> index 69fb5ba..6fc4366 100644 >> --- a/drivers/infiniband/hw/mlx4/sysfs.c >> +++ b/drivers/infiniband/hw/mlx4/sysfs.c >> @@ -371,6 +371,7 @@ struct mlx4_port { >> struct attribute_group gid_group; >> struct device_attribute enable_smi_admin; >> struct device_attribute smi_enabled; >> + struct device_attribute node_guid; >> int slave; >> u8 port_num; >> }; >> @@ -620,6 +621,7 @@ static int add_vf_smi_entries(struct mlx4_port *p) >> sysfs_remove_file(&p->kobj, &p->smi_enabled.attr); >> return ret; >> } >> + > Stray change. > >> return 0; >> } >> >> @@ -635,6 +637,69 @@ static void remove_vf_smi_entries(struct >> mlx4_port *p) >> sysfs_remove_file(&p->kobj, &p->enable_smi_admin.attr); >> } >> >> +static ssize_t sysfs_show_node_guid(struct device *dev, >> + struct device_attribute *attr, >> + char *buf) >> +{ >> + struct mlx4_port *p = container_of(attr, struct mlx4_port, >> node_guid); >> + ssize_t len = 0; >> + __be64 guid = mlx4_get_slave_node_guid(p->dev->dev, p->slave); >> + >> + len = sprintf(buf, "slave %d 0x%llx\n", p->slave, >> be64_to_cpu(guid)); >> + >> + return len; >> +} >> + >> +static ssize_t sysfs_store_node_guid(struct device *dev, >> + struct device_attribute *attr, >> + const char *buf, size_t count) >> +{ >> + struct mlx4_port *p = container_of(attr, struct mlx4_port, >> node_guid); >> + __be64 guid; >> + > Shouldn't we check only master can set this up ? > >> + if (sscanf(buf, "0x%llx", &guid) != 1) >> + return -EINVAL; >> + >> + mlx4_put_slave_node_guid(p->dev->dev, p->slave, cpu_to_be64(guid)); >> + >> + return count; >> +} >> + >> +static int add_vf_admin_guid_entry(struct mlx4_port *p) >> +{ >> + int is_eth = rdma_port_get_link_layer(&p->dev->ib_dev, >> p->port_num) == >> + IB_LINK_LAYER_ETHERNET; >> + int ret; >> + >> + /* do not display entries if eth transport, or if master */ >> + if (is_eth || p->slave == mlx4_master_func_num(p->dev->dev)) >> + return 0; >> + >> + sysfs_attr_init(&p->node_guid.attr); >> + p->node_guid.show = sysfs_show_node_guid; >> + p->node_guid.store = sysfs_store_node_guid; >> + p->node_guid.attr.name = "node_guid"; >> + p->node_guid.attr.mode = 0644; >> + ret = sysfs_create_file(&p->kobj, &p->node_guid.attr); >> + if (ret) { >> + pr_err("failed to create sysfs entry for node_guid\n"); > Use existing driver macro or append print with "mlx4_core:" >> + return ret; >> + } >> + >> + return 0; >> +} >> + >> +static void remove_vf_node_guid_entry(struct mlx4_port *p) >> +{ >> + int is_eth = rdma_port_get_link_layer(&p->dev->ib_dev, >> p->port_num) == >> + IB_LINK_LAYER_ETHERNET; >> + >> + if (is_eth || p->slave == mlx4_master_func_num(p->dev->dev)) >> + return; >> + >> + sysfs_remove_file(&p->kobj, &p->node_guid.attr); >> +} >> + >> static int add_port(struct mlx4_ib_dev *dev, int port_num, int slave) >> { >> struct mlx4_port *p; >> @@ -686,6 +751,10 @@ static int add_port(struct mlx4_ib_dev *dev, int >> port_num, int slave) >> if (ret) >> goto err_free_gid; >> >> + ret = add_vf_admin_guid_entry(p); >> + if (ret) >> + goto err_free_gid; >> + >> list_add_tail(&p->kobj.entry, &dev->pkeys.pkey_port_list[slave]); >> return 0; >> >> @@ -743,6 +812,7 @@ static int register_one_pkey_tree(struct >> mlx4_ib_dev *dev, int slave) >> if (err) >> goto err_add; >> } >> + > Stray change. > > Regards, > Santosh > -- > To unsubscribe from this list: send the line "unsubscribe linux-rdma" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 7/21/2016 8:53 AM, Majd Dibbiny wrote: > On 7/21/2016 6:12 PM, Santosh Shilimkar wrote: >> On 7/21/2016 2:32 AM, Yuval Shaia wrote: >>> Adding sysfs entries to read/write VF's Node GUID. >>> This abbility will enable sys-admin to configure node GUID for VFs the > typo >>> same >>> way it is done for VF's port GUIDs. >>> > > Hi Yuval, > > The following commit in iproute2 by Eli Cohen added a standard way to > configure node/port GUID using IPoIB netdev: > > d91fb3f4c7e4 Add support for configuring Infiniband GUIDs > > The configuration will be done in the following way: > ip link set dev ib0 vf 0 node_guid 00:02:c9:03:00:21:6e:70 > ip link set dev ib0 vf 0 port_guid 00:02:c9:03:00:21:6e:78 > > Can't this be used here as well? > This could work as well. Does this has any dependency with 'iproute' version ? In other words does this kernel patch needs any iproute patch ? -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, Jul 21, 2016 at 6:53 PM, Majd Dibbiny <majd@mellanox.com> wrote: > On 7/21/2016 6:12 PM, Santosh Shilimkar wrote: >> On 7/21/2016 2:32 AM, Yuval Shaia wrote: >>> Adding sysfs entries to read/write VF's Node GUID. >>> This ability will enable sys-admin to configure node GUID for VFs the > Hi Yuval, > The following commit in iproute2 by Eli Cohen added a standard way to > configure node/port GUID using IPoIB netdev: > > d91fb3f4c7e4 Add support for configuring Infiniband GUIDs > > The configuration will be done in the following way: > ip link set dev ib0 vf 0 node_guid 00:02:c9:03:00:21:6e:70 > ip link set dev ib0 vf 0 port_guid 00:02:c9:03:00:21:6e:78 > > Can't this be used here as well? correct, NAK for adding more proprietary sysfs entries here, use the standard rtnl way as in mlx5 Or. -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/infiniband/hw/mlx4/sysfs.c b/drivers/infiniband/hw/mlx4/sysfs.c index 69fb5ba..6fc4366 100644 --- a/drivers/infiniband/hw/mlx4/sysfs.c +++ b/drivers/infiniband/hw/mlx4/sysfs.c @@ -371,6 +371,7 @@ struct mlx4_port { struct attribute_group gid_group; struct device_attribute enable_smi_admin; struct device_attribute smi_enabled; + struct device_attribute node_guid; int slave; u8 port_num; }; @@ -620,6 +621,7 @@ static int add_vf_smi_entries(struct mlx4_port *p) sysfs_remove_file(&p->kobj, &p->smi_enabled.attr); return ret; } + return 0; } @@ -635,6 +637,69 @@ static void remove_vf_smi_entries(struct mlx4_port *p) sysfs_remove_file(&p->kobj, &p->enable_smi_admin.attr); } +static ssize_t sysfs_show_node_guid(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct mlx4_port *p = container_of(attr, struct mlx4_port, node_guid); + ssize_t len = 0; + __be64 guid = mlx4_get_slave_node_guid(p->dev->dev, p->slave); + + len = sprintf(buf, "slave %d 0x%llx\n", p->slave, be64_to_cpu(guid)); + + return len; +} + +static ssize_t sysfs_store_node_guid(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct mlx4_port *p = container_of(attr, struct mlx4_port, node_guid); + __be64 guid; + + if (sscanf(buf, "0x%llx", &guid) != 1) + return -EINVAL; + + mlx4_put_slave_node_guid(p->dev->dev, p->slave, cpu_to_be64(guid)); + + return count; +} + +static int add_vf_admin_guid_entry(struct mlx4_port *p) +{ + int is_eth = rdma_port_get_link_layer(&p->dev->ib_dev, p->port_num) == + IB_LINK_LAYER_ETHERNET; + int ret; + + /* do not display entries if eth transport, or if master */ + if (is_eth || p->slave == mlx4_master_func_num(p->dev->dev)) + return 0; + + sysfs_attr_init(&p->node_guid.attr); + p->node_guid.show = sysfs_show_node_guid; + p->node_guid.store = sysfs_store_node_guid; + p->node_guid.attr.name = "node_guid"; + p->node_guid.attr.mode = 0644; + ret = sysfs_create_file(&p->kobj, &p->node_guid.attr); + if (ret) { + pr_err("failed to create sysfs entry for node_guid\n"); + return ret; + } + + return 0; +} + +static void remove_vf_node_guid_entry(struct mlx4_port *p) +{ + int is_eth = rdma_port_get_link_layer(&p->dev->ib_dev, p->port_num) == + IB_LINK_LAYER_ETHERNET; + + if (is_eth || p->slave == mlx4_master_func_num(p->dev->dev)) + return; + + sysfs_remove_file(&p->kobj, &p->node_guid.attr); +} + static int add_port(struct mlx4_ib_dev *dev, int port_num, int slave) { struct mlx4_port *p; @@ -686,6 +751,10 @@ static int add_port(struct mlx4_ib_dev *dev, int port_num, int slave) if (ret) goto err_free_gid; + ret = add_vf_admin_guid_entry(p); + if (ret) + goto err_free_gid; + list_add_tail(&p->kobj.entry, &dev->pkeys.pkey_port_list[slave]); return 0; @@ -743,6 +812,7 @@ static int register_one_pkey_tree(struct mlx4_ib_dev *dev, int slave) if (err) goto err_add; } + return 0; err_add: @@ -754,6 +824,7 @@ err_add: sysfs_remove_group(p, &mport->pkey_group); sysfs_remove_group(p, &mport->gid_group); remove_vf_smi_entries(mport); + remove_vf_node_guid_entry(mport); kobject_put(p); } kobject_put(dev->dev_ports_parent[slave]); @@ -799,6 +870,7 @@ static void unregister_pkey_tree(struct mlx4_ib_dev *device) sysfs_remove_group(p, &port->pkey_group); sysfs_remove_group(p, &port->gid_group); remove_vf_smi_entries(port); + remove_vf_node_guid_entry(port); kobject_put(p); kobject_put(device->dev_ports_parent[slave]); }
Adding sysfs entries to read/write VF's Node GUID. This abbility will enable sys-admin to configure node GUID for VFs the same way it is done for VF's port GUIDs. Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com> --- drivers/infiniband/hw/mlx4/sysfs.c | 72 ++++++++++++++++++++++++++++++++++++ 1 files changed, 72 insertions(+), 0 deletions(-)