* Re: [Patch net-next] net_sched: add reverse binding for tc class
From: Daniel Borkmann @ 2017-08-30 22:22 UTC (permalink / raw)
To: Cong Wang; +Cc: Linux Kernel Network Developers, Jamal Hadi Salim
In-Reply-To: <CAM_iQpU9B71C5cWs969k464T7HvNhcwps7NYNQx8Q3KCijOfnA@mail.gmail.com>
On 08/31/2017 12:01 AM, Cong Wang wrote:
> On Wed, Aug 30, 2017 at 2:48 PM, Daniel Borkmann <daniel@iogearbox.net> wrote:
>> On 08/30/2017 11:30 PM, Cong Wang wrote:
>> [...]
>>>
>>> Note, we still can NOT totally get rid of those class lookup in
>>> ->enqueue() because cgroup and flow filters have no way to determine
>>> the classid at setup time, they still have to go through dynamic lookup.
>>
>> [...]
>>>
>>> ---
>>> include/net/sch_generic.h | 1 +
>>> net/sched/cls_basic.c | 9 +++++++
>>> net/sched/cls_bpf.c | 9 +++++++
>>
>> Same is for cls_bpf as well, so bind_class wouldn't work there
>> either as we could return dynamic classids. bind_class cannot
>> be added here, too.
>
> I think you are probably right, but the following code is
> misleading there:
>
> if (tb[TCA_BPF_CLASSID]) {
> prog->res.classid = nla_get_u32(tb[TCA_BPF_CLASSID]);
> tcf_bind_filter(tp, &prog->res, base);
> }
>
> If the classid is dynamic, why this tb[TCA_BPF_CLASSID]?
The prog->res.classid is the default one, but can be overridden
later depending on the specified program. cls_bpf_classify() does
after prog return (filter_res holds return code):
[...]
if (filter_res == 0)
continue;
if (filter_res != -1) {
res->class = 0;
res->classid = filter_res;
} else {
*res = prog->res;
}
[...]
Meaning in case of a match (-1), we use the default bound one,
but prog may as well return an alternative found classid if it
wants to. So both versions are possible.
^ permalink raw reply
* [net 09/11] net/mlx5: E-Switch, Unload the representors in the correct order
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Shahar Klein, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Shahar Klein <shahark@mellanox.com>
When changing from switchdev to legacy mode, all the representor port
devices (uplink nic and reps) are cleaned up. Part of this cleaning
process is removing the neigh entries and the hash table containing them.
However, a representor neigh entry might be linked to the uplink port
hash table and if the uplink nic is cleaned first the cleaning of the
representor will end up in null deref.
Fix that by unloading the representors in the opposite order of load.
Fixes: cb67b832921c ("net/mlx5e: Introduce SRIOV VF representors")
Signed-off-by: Shahar Klein <shahark@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
index 95b64025ce36..5bc0593bd76e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c
@@ -815,7 +815,7 @@ void esw_offloads_cleanup(struct mlx5_eswitch *esw, int nvports)
struct mlx5_eswitch_rep *rep;
int vport;
- for (vport = 0; vport < nvports; vport++) {
+ for (vport = nvports - 1; vport >= 0; vport--) {
rep = &esw->offloads.vport_reps[vport];
if (!rep->valid)
continue;
--
2.13.0
^ permalink raw reply related
* [net 06/11] net/mlx5e: Fix dangling page pointer on DMA mapping error
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Eran Ben Elisha, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Eran Ben Elisha <eranbe@mellanox.com>
Function mlx5e_dealloc_rx_wqe is using page pointer value as an
indication to valid DMA mapping. In case that the mapping failed, we
released the page but kept the dangling pointer. Store the page pointer
only after the DMA mapping passed to avoid invalid page DMA unmap.
Fixes: bc77b240b3c5 ("net/mlx5e: Add fragmented memory support for RX multi packet WQE")
Signed-off-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index 325b2c8c1c6d..7344433259fc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -222,13 +222,13 @@ static inline int mlx5e_page_alloc_mapped(struct mlx5e_rq *rq,
if (unlikely(!page))
return -ENOMEM;
- dma_info->page = page;
dma_info->addr = dma_map_page(rq->pdev, page, 0,
RQ_PAGE_SIZE(rq), rq->buff.map_dir);
if (unlikely(dma_mapping_error(rq->pdev, dma_info->addr))) {
put_page(page);
return -ENOMEM;
}
+ dma_info->page = page;
return 0;
}
--
2.13.0
^ permalink raw reply related
* [net 11/11] net/mlx5e: Fix CQ moderation mode not set properly
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Tal Gilboa, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Tal Gilboa <talgi@mellanox.com>
cq_period_mode assignment was mistakenly removed so it was always set to "0",
which is EQE based moderation, regardless of the device CAPs and
requested value in ethtool.
Fixes: 6a9764efb255 ("net/mlx5e: Isolate open_channels from priv->params")
Signed-off-by: Tal Gilboa <talgi@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 57f31fa478ce..6ad7f07e7861 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -1969,6 +1969,7 @@ static void mlx5e_build_rx_cq_param(struct mlx5e_priv *priv,
}
mlx5e_build_common_cq_param(priv, param);
+ param->cq_period_mode = params->rx_cq_period_mode;
}
static void mlx5e_build_tx_cq_param(struct mlx5e_priv *priv,
--
2.13.0
^ permalink raw reply related
* [net 07/11] net/mlx5e: Don't override user RSS upon set channels
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Inbar Karmy, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Inbar Karmy <inbark@mellanox.com>
Currently, increasing the number of combined channels is changing
the RSS spread to use the new created channels.
Prevent the RSS spread change in case the user explicitly declare it,
to avoid overriding user configuration.
Tested:
when RSS default:
# ethtool -L ens8 combined 4
RSS spread will change and point to 4 channels.
# ethtool -X ens8 equal 4
# ethtool -L ens8 combined 6
RSS will not change after increasing the number of the channels.
Fixes: 8bf368620486 ('ethtool: ensure channel counts are within bounds during SCHANNELS')
Signed-off-by: Inbar Karmy <inbark@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index 917fade5f5d5..f5594014715b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -641,8 +641,10 @@ int mlx5e_ethtool_set_channels(struct mlx5e_priv *priv,
new_channels.params = priv->channels.params;
new_channels.params.num_channels = count;
- mlx5e_build_default_indir_rqt(priv->mdev, new_channels.params.indirection_rqt,
- MLX5E_INDIR_RQT_SIZE, count);
+ if (!netif_is_rxfh_configured(priv->netdev))
+ mlx5e_build_default_indir_rqt(priv->mdev,
+ new_channels.params.indirection_rqt,
+ MLX5E_INDIR_RQT_SIZE, count);
if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
priv->channels.params = new_channels.params;
--
2.13.0
^ permalink raw reply related
* [net 05/11] net/mlx5: Remove the flag MLX5_INTERFACE_STATE_SHUTDOWN
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Huy Nguyen, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Huy Nguyen <huyn@mellanox.com>
MLX5_INTERFACE_STATE_SHUTDOWN is not used in the code.
Fixes: 5fc7197d3a25 ("net/mlx5: Add pci shutdown callback")
Signed-off-by: Huy Nguyen <huyn@mellanox.com>
Reviewed-by: Daniel Jurgens <danielj@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/main.c | 2 --
include/linux/mlx5/driver.h | 1 -
2 files changed, 3 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index 4cdb414aa2d5..16885827367b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -1563,8 +1563,6 @@ static void shutdown(struct pci_dev *pdev)
int err;
dev_info(&pdev->dev, "Shutdown was called\n");
- /* Notify mlx5 clients that the kernel is being shut down */
- set_bit(MLX5_INTERFACE_STATE_SHUTDOWN, &dev->intf_state);
err = mlx5_try_fast_unload(dev);
if (err)
mlx5_unload_one(dev, priv, false);
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index 918f5e644506..205d82d4c468 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -674,7 +674,6 @@ enum mlx5_device_state {
enum mlx5_interface_state {
MLX5_INTERFACE_STATE_UP = BIT(0),
- MLX5_INTERFACE_STATE_SHUTDOWN = BIT(1),
};
enum mlx5_pci_status {
--
2.13.0
^ permalink raw reply related
* [net 10/11] net/mlx5e: Fix inline header size for small packets
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Moshe Shemesh, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Moshe Shemesh <moshe@mellanox.com>
Fix inline header size, make sure it is not greater than skb len.
This bug effects small packets, for example L2 packets with size < 18.
Fixes: ae76715d153e ("net/mlx5e: Check the minimum inline header mode before xmit")
Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_tx.c | 17 ++++++++++-------
1 file changed, 10 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
index aaa0f4ebba9a..31353e5c3c78 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
@@ -128,10 +128,10 @@ static inline int mlx5e_skb_l3_header_offset(struct sk_buff *skb)
return mlx5e_skb_l2_header_offset(skb);
}
-static inline unsigned int mlx5e_calc_min_inline(enum mlx5_inline_modes mode,
- struct sk_buff *skb)
+static inline u16 mlx5e_calc_min_inline(enum mlx5_inline_modes mode,
+ struct sk_buff *skb)
{
- int hlen;
+ u16 hlen;
switch (mode) {
case MLX5_INLINE_MODE_NONE:
@@ -140,19 +140,22 @@ static inline unsigned int mlx5e_calc_min_inline(enum mlx5_inline_modes mode,
hlen = eth_get_headlen(skb->data, skb_headlen(skb));
if (hlen == ETH_HLEN && !skb_vlan_tag_present(skb))
hlen += VLAN_HLEN;
- return hlen;
+ break;
case MLX5_INLINE_MODE_IP:
/* When transport header is set to zero, it means no transport
* header. When transport header is set to 0xff's, it means
* transport header wasn't set.
*/
- if (skb_transport_offset(skb))
- return mlx5e_skb_l3_header_offset(skb);
+ if (skb_transport_offset(skb)) {
+ hlen = mlx5e_skb_l3_header_offset(skb);
+ break;
+ }
/* fall through */
case MLX5_INLINE_MODE_L2:
default:
- return mlx5e_skb_l2_header_offset(skb);
+ hlen = mlx5e_skb_l2_header_offset(skb);
}
+ return min_t(u16, hlen, skb->len);
}
static inline void mlx5e_tx_skb_pull_inline(unsigned char **skb_data,
--
2.13.0
^ permalink raw reply related
* [net 08/11] net/mlx5e: Properly resolve TC offloaded ipv6 vxlan tunnel source address
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Paul Blakey, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Paul Blakey <paulb@mellanox.com>
Currently if vxlan tunnel ipv6 src isn't supplied the driver fails to
resolve it as part of the route lookup. The resulting encap header
is left with a zeroed out ipv6 src address so the packets are sent
with this src ip.
Use an appropriate route lookup API that also resolves the source
ipv6 address if it's not supplied.
Fixes: ce99f6b97fcd ('net/mlx5e: Support SRIOV TC encapsulation offloads for IPv6 tunnels')
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
index 3c536f560dd2..7f282e8f4e7f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tc.c
@@ -1443,12 +1443,10 @@ static int mlx5e_route_lookup_ipv6(struct mlx5e_priv *priv,
struct mlx5_eswitch *esw = priv->mdev->priv.eswitch;
int ret;
- dst = ip6_route_output(dev_net(mirred_dev), NULL, fl6);
- ret = dst->error;
- if (ret) {
- dst_release(dst);
+ ret = ipv6_stub->ipv6_dst_lookup(dev_net(mirred_dev), NULL, &dst,
+ fl6);
+ if (ret < 0)
return ret;
- }
*out_ttl = ip6_dst_hoplimit(dst);
--
2.13.0
^ permalink raw reply related
* [net 04/11] net/mlx5: Skip mlx5_unload_one if mlx5_load_one fails
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Huy Nguyen, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Huy Nguyen <huyn@mellanox.com>
There is an issue where the firmware fails during mlx5_load_one,
the health_care timer detects the issue and schedules a health_care call.
Then the mlx5_load_one detects the issue, cleans up and quits. Then
the health_care starts and calls mlx5_unload_one to clean up the resources
that no longer exist and causes kernel panic.
The root cause is that the bit MLX5_INTERFACE_STATE_DOWN is not set
after mlx5_load_one fails. The solution is removing the bit
MLX5_INTERFACE_STATE_DOWN and quit mlx5_unload_one if the
bit MLX5_INTERFACE_STATE_UP is not set. The bit MLX5_INTERFACE_STATE_DOWN
is redundant and we can use MLX5_INTERFACE_STATE_UP instead.
Fixes: 5fc7197d3a25 ("net/mlx5: Add pci shutdown callback")
Signed-off-by: Huy Nguyen <huyn@mellanox.com>
Reviewed-by: Daniel Jurgens <danielj@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/main.c | 4 +---
include/linux/mlx5/driver.h | 5 ++---
2 files changed, 3 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c
index c065132b956d..4cdb414aa2d5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c
@@ -1186,7 +1186,6 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
}
}
- clear_bit(MLX5_INTERFACE_STATE_DOWN, &dev->intf_state);
set_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state);
out:
mutex_unlock(&dev->intf_state_mutex);
@@ -1261,7 +1260,7 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
mlx5_drain_health_recovery(dev);
mutex_lock(&dev->intf_state_mutex);
- if (test_bit(MLX5_INTERFACE_STATE_DOWN, &dev->intf_state)) {
+ if (!test_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state)) {
dev_warn(&dev->pdev->dev, "%s: interface is down, NOP\n",
__func__);
if (cleanup)
@@ -1270,7 +1269,6 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv,
}
clear_bit(MLX5_INTERFACE_STATE_UP, &dev->intf_state);
- set_bit(MLX5_INTERFACE_STATE_DOWN, &dev->intf_state);
if (mlx5_device_registered(dev))
mlx5_detach_device(dev);
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index df6ce59a1f95..918f5e644506 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -673,9 +673,8 @@ enum mlx5_device_state {
};
enum mlx5_interface_state {
- MLX5_INTERFACE_STATE_DOWN = BIT(0),
- MLX5_INTERFACE_STATE_UP = BIT(1),
- MLX5_INTERFACE_STATE_SHUTDOWN = BIT(2),
+ MLX5_INTERFACE_STATE_UP = BIT(0),
+ MLX5_INTERFACE_STATE_SHUTDOWN = BIT(1),
};
enum mlx5_pci_status {
--
2.13.0
^ permalink raw reply related
* [net 03/11] net/mlx5: Fix arm SRQ command for ISSI version 0
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Noa Osherovich, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Noa Osherovich <noaos@mellanox.com>
Support for ISSI version 0 was recently broken as the arm_srq_cmd
command, which is used only for ISSI version 0, was given the opcode
for ISSI version 1 instead of ISSI version 0.
Change arm_srq_cmd to use the correct command opcode for ISSI version
0.
Fixes: af1ba291c5e4 ('{net, IB}/mlx5: Refactor internal SRQ API')
Signed-off-by: Noa Osherovich <noaos@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/srq.c | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/srq.c b/drivers/net/ethernet/mellanox/mlx5/core/srq.c
index f774de6f5fcb..520f6382dfde 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/srq.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/srq.c
@@ -201,13 +201,13 @@ static int destroy_srq_cmd(struct mlx5_core_dev *dev,
static int arm_srq_cmd(struct mlx5_core_dev *dev, struct mlx5_core_srq *srq,
u16 lwm, int is_srq)
{
- /* arm_srq structs missing using identical xrc ones */
- u32 srq_in[MLX5_ST_SZ_DW(arm_xrc_srq_in)] = {0};
- u32 srq_out[MLX5_ST_SZ_DW(arm_xrc_srq_out)] = {0};
+ u32 srq_in[MLX5_ST_SZ_DW(arm_rq_in)] = {0};
+ u32 srq_out[MLX5_ST_SZ_DW(arm_rq_out)] = {0};
- MLX5_SET(arm_xrc_srq_in, srq_in, opcode, MLX5_CMD_OP_ARM_XRC_SRQ);
- MLX5_SET(arm_xrc_srq_in, srq_in, xrc_srqn, srq->srqn);
- MLX5_SET(arm_xrc_srq_in, srq_in, lwm, lwm);
+ MLX5_SET(arm_rq_in, srq_in, opcode, MLX5_CMD_OP_ARM_RQ);
+ MLX5_SET(arm_rq_in, srq_in, op_mod, MLX5_ARM_RQ_IN_OP_MOD_SRQ);
+ MLX5_SET(arm_rq_in, srq_in, srq_number, srq->srqn);
+ MLX5_SET(arm_rq_in, srq_in, lwm, lwm);
return mlx5_cmd_exec(dev, srq_in, sizeof(srq_in),
srq_out, sizeof(srq_out));
--
2.13.0
^ permalink raw reply related
* [net 02/11] net/mlx5e: Fix DCB_CAP_ATTR_DCBX capability for DCBNL getcap.
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Huy Nguyen, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Huy Nguyen <huyn@mellanox.com>
Current code doesn't report DCB_CAP_DCBX_HOST capability when query
through getcap. User space lldptool expects capability to have HOST mode
set when it wants to configure DCBX CEE mode. In absence of HOST mode
capability, lldptool fails to switch to CEE mode.
This fix returns DCB_CAP_DCBX_HOST capability when port's DCBX
controlled mode is under software control.
Fixes: 3a6a931dfb8e ("net/mlx5e: Support DCBX CEE API")
Signed-off-by: Huy Nguyen <huyn@mellanox.com>
Reviewed-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en.h | 1 +
drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c | 21 ++++++++++++---------
2 files changed, 13 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 0039b4725405..2f26fb34d741 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -263,6 +263,7 @@ struct mlx5e_dcbx {
/* The only setting that cannot be read from FW */
u8 tc_tsa[IEEE_8021QAZ_MAX_TCS];
+ u8 cap;
};
#endif
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
index 810b51029c7f..c1d384fca4dc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
@@ -288,13 +288,8 @@ static int mlx5e_dcbnl_ieee_setpfc(struct net_device *dev,
static u8 mlx5e_dcbnl_getdcbx(struct net_device *dev)
{
struct mlx5e_priv *priv = netdev_priv(dev);
- struct mlx5e_dcbx *dcbx = &priv->dcbx;
- u8 mode = DCB_CAP_DCBX_VER_IEEE | DCB_CAP_DCBX_VER_CEE;
-
- if (dcbx->mode == MLX5E_DCBX_PARAM_VER_OPER_HOST)
- mode |= DCB_CAP_DCBX_HOST;
- return mode;
+ return priv->dcbx.cap;
}
static u8 mlx5e_dcbnl_setdcbx(struct net_device *dev, u8 mode)
@@ -312,6 +307,7 @@ static u8 mlx5e_dcbnl_setdcbx(struct net_device *dev, u8 mode)
/* set dcbx to fw controlled */
if (!mlx5e_dcbnl_set_dcbx_mode(priv, MLX5E_DCBX_PARAM_VER_OPER_AUTO)) {
dcbx->mode = MLX5E_DCBX_PARAM_VER_OPER_AUTO;
+ dcbx->cap &= ~DCB_CAP_DCBX_HOST;
return 0;
}
@@ -324,6 +320,8 @@ static u8 mlx5e_dcbnl_setdcbx(struct net_device *dev, u8 mode)
if (mlx5e_dcbnl_switch_to_host_mode(netdev_priv(dev)))
return 1;
+ dcbx->cap = mode;
+
return 0;
}
@@ -628,9 +626,9 @@ static u8 mlx5e_dcbnl_getcap(struct net_device *netdev,
*cap = false;
break;
case DCB_CAP_ATTR_DCBX:
- *cap = (DCB_CAP_DCBX_LLD_MANAGED |
- DCB_CAP_DCBX_VER_CEE |
- DCB_CAP_DCBX_STATIC);
+ *cap = priv->dcbx.cap |
+ DCB_CAP_DCBX_VER_CEE |
+ DCB_CAP_DCBX_VER_IEEE;
break;
default:
*cap = 0;
@@ -760,5 +758,10 @@ void mlx5e_dcbnl_initialize(struct mlx5e_priv *priv)
if (MLX5_CAP_GEN(priv->mdev, dcbx))
mlx5e_dcbnl_query_dcbx_mode(priv, &dcbx->mode);
+ priv->dcbx.cap = DCB_CAP_DCBX_VER_CEE |
+ DCB_CAP_DCBX_VER_IEEE;
+ if (priv->dcbx.mode == MLX5E_DCBX_PARAM_VER_OPER_HOST)
+ priv->dcbx.cap |= DCB_CAP_DCBX_HOST;
+
mlx5e_ets_init(priv);
}
--
2.13.0
^ permalink raw reply related
* [pull request][net 00/11] Mellanox, mlx5 fixes 2017-08-30
From: Saeed Mahameed @ 2017-08-30 22:20 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Saeed Mahameed
Hi Dave,
This series contains some misc fixes to the mlx5 driver.
Please pull and let me know if there's any problem.
For -stable:
Kernels >= 4.12
net/mlx5e: Fix CQ moderation mode not set properly
net/mlx5e: Don't override user RSS upon set channels
Kernels >= 4.11
net/mlx5e: Properly resolve TC offloaded ipv6 vxlan tunnel source address
Kernels >= 4.10
net/mlx5e: Fix DCB_CAP_ATTR_DCBX capability for DCBNL getcap
net/mlx5e: Check for qos capability in dcbnl_initialize
Kernels >= 4.9
net/mlx5e: Fix dangling page pointer on DMA mapping error
Kernels >= 4.8
net/mlx5e: Fix inline header size for small packets
net/mlx5: E-Switch, Unload the representors in the correct order
net/mlx5: Fix arm SRQ command for ISSI version 0
Thanks,
Saeed.
---
The following changes since commit 183db481279437590f75a8a0479d512e5dd597de:
drivers: net: xgene: Correct probe sequence handling (2017-08-29 16:13:08 -0700)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux.git tags/mlx5-fixes-2017-08-30
for you to fetch changes up to 1213ad28f9595a08e3877248bbba1a25c40225d6:
net/mlx5e: Fix CQ moderation mode not set properly (2017-08-30 21:20:43 +0300)
----------------------------------------------------------------
mlx5-fixes-2017-08-30
----------------------------------------------------------------
Eran Ben Elisha (1):
net/mlx5e: Fix dangling page pointer on DMA mapping error
Huy Nguyen (4):
net/mlx5e: Check for qos capability in dcbnl_initialize
net/mlx5e: Fix DCB_CAP_ATTR_DCBX capability for DCBNL getcap.
net/mlx5: Skip mlx5_unload_one if mlx5_load_one fails
net/mlx5: Remove the flag MLX5_INTERFACE_STATE_SHUTDOWN
Inbar Karmy (1):
net/mlx5e: Don't override user RSS upon set channels
Moshe Shemesh (1):
net/mlx5e: Fix inline header size for small packets
Noa Osherovich (1):
net/mlx5: Fix arm SRQ command for ISSI version 0
Paul Blakey (1):
net/mlx5e: Properly resolve TC offloaded ipv6 vxlan tunnel source address
Shahar Klein (1):
net/mlx5: E-Switch, Unload the representors in the correct order
Tal Gilboa (1):
net/mlx5e: Fix CQ moderation mode not set properly
drivers/net/ethernet/mellanox/mlx5/core/en.h | 1 +
drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c | 24 ++++++++++++++--------
.../net/ethernet/mellanox/mlx5/core/en_ethtool.c | 6 ++++--
drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 1 +
drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 2 +-
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c | 8 +++-----
drivers/net/ethernet/mellanox/mlx5/core/en_tx.c | 17 ++++++++-------
.../ethernet/mellanox/mlx5/core/eswitch_offloads.c | 2 +-
drivers/net/ethernet/mellanox/mlx5/core/main.c | 6 +-----
drivers/net/ethernet/mellanox/mlx5/core/srq.c | 12 +++++------
include/linux/mlx5/driver.h | 4 +---
11 files changed, 44 insertions(+), 39 deletions(-)
^ permalink raw reply
* [net 01/11] net/mlx5e: Check for qos capability in dcbnl_initialize
From: Saeed Mahameed @ 2017-08-30 22:21 UTC (permalink / raw)
To: David S. Miller; +Cc: netdev, Huy Nguyen, Saeed Mahameed
In-Reply-To: <20170830222110.15737-1-saeedm@mellanox.com>
From: Huy Nguyen <huyn@mellanox.com>
qos capability is the master capability bit that determines
if the DCBX is supported for the PCI function. If this bit is off,
driver cannot run any dcbx code.
Fixes: e207b7e99176 ("net/mlx5e: ConnectX-4 firmware support for DCBX")
Signed-off-by: Huy Nguyen <huyn@mellanox.com>
Reviewed-by: Parav Pandit <parav@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
index 2eb54d36e16e..810b51029c7f 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c
@@ -754,6 +754,9 @@ void mlx5e_dcbnl_initialize(struct mlx5e_priv *priv)
{
struct mlx5e_dcbx *dcbx = &priv->dcbx;
+ if (!MLX5_CAP_GEN(priv->mdev, qos))
+ return;
+
if (MLX5_CAP_GEN(priv->mdev, dcbx))
mlx5e_dcbnl_query_dcbx_mode(priv, &dcbx->mode);
--
2.13.0
^ permalink raw reply related
* Re: [PATCH] net: bcm63xx_enet: make bcm_enetsw_ethtool_ops const
From: David Miller @ 2017-08-30 22:20 UTC (permalink / raw)
To: bhumirks
Cc: julia.lawall, f.fainelli, bcm-kernel-feedback-list, netdev,
linux-arm-kernel, linux-kernel
In-Reply-To: <1504085108-23495-1-git-send-email-bhumirks@gmail.com>
From: Bhumika Goyal <bhumirks@gmail.com>
Date: Wed, 30 Aug 2017 14:55:08 +0530
> Make this const as it is never modified.
>
> Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Applied to net-next.
^ permalink raw reply
* Re: [PATCH v2] ipv6: sr: fix get_srh() to comply with IPv6 standard "RFC 8200"
From: David Miller @ 2017-08-30 22:19 UTC (permalink / raw)
To: amsalam20; +Cc: netdev, david.lebrun
In-Reply-To: <1504083037-1605-1-git-send-email-amsalam20@gmail.com>
From: Ahmed Abdelsalam <amsalam20@gmail.com>
Date: Wed, 30 Aug 2017 10:50:37 +0200
> IPv6 packet may carry more than one extension header, and IPv6 nodes must
> accept and attempt to process extension headers in any order and occurring
> any number of times in the same packet. Hence, there should be no
> assumption that Segment Routing extension header is to appear immediately
> after the IPv6 header.
>
> Moreover, section 4.1 of RFC 8200 gives a recommendation on the order of
> appearance of those extension headers within an IPv6 packet. According to
> this recommendation, Segment Routing extension header should appear after
> Hop-by-Hop and Destination Options headers (if they present).
>
> This patch fixes the get_srh(), so it gets the segment routing header
> regardless of its position in the chain of the extension headers in IPv6
> packet, and makes sure that the IPv6 routing extension header is of Type 4.
>
> Signed-off-by: Ahmed Abdelsalam <amsalam20@gmail.com>
Applied.
^ permalink raw reply
* Re: [PATCH net-next v4 00/13] net: mvpp2: comphy configuration
From: David Miller @ 2017-08-30 22:18 UTC (permalink / raw)
To: antoine.tenart
Cc: kishon, andrew, jason, sebastian.hesselbarth, gregory.clement,
thomas.petazzoni, nadavh, linux, linux-kernel, mw, stefanc,
miquel.raynal, netdev
In-Reply-To: <20170830082924.3180-1-antoine.tenart@free-electrons.com>
From: Antoine Tenart <antoine.tenart@free-electrons.com>
Date: Wed, 30 Aug 2017 10:29:11 +0200
> This series, following up the one one the GoP/MAC configuration, aims at
> stopping to depend on the firmware/bootloader configuration when using
> the PPv2 engine. With this series the PPv2 driver does not need to rely
> on a previous configuration, and dynamic reconfiguration while the
> kernel is running can be done (i.e. switch one port from SGMII to 10G,
> or the opposite). A port can now be configured in a different mode than
> what's done in the firmware/bootloader as well.
>
> The series first contain patches in the generic PHY framework to support
> what is called the comphy (common PHYs), which is an h/w block providing
> PHYs that can be configured in various modes ranging from SGMII, 10G
> to SATA and others. As of now only the SGMII and 10G modes are
> supported by the comphy driver.
>
> Then patches are modifying the PPv2 driver to first add the comphy
> initialization sequence (i.e. calls to the generic PHY framework) and to
> then take advantage of this to allow dynamic reconfiguration (i.e.
> configuring the mode of a port given what's connected, between sgmii and
> 10G). Note the use of the comphy in the PPv2 driver is kept optional
> (i.e. if not described in dt the driver still as before an relies on the
> firmware/bootloader configuration).
>
> Finally there are dt/defconfig patches to describe and take advantage of
> this.
>
> This was tested on a range of devices: 8040-db, 8040-mcbin and 7040-db.
>
> @Dave: the dt patches should go through the mvebu tree (patches 9-13).
Ok, patches 1-8 applied to net-next, thanks!
^ permalink raw reply
* Re: DSA mv88e6xxx RX frame errors and TCP/IP RX failure
From: Andrew Lunn @ 2017-08-30 22:06 UTC (permalink / raw)
To: Tim Harvey; +Cc: netdev, Vivien Didelot, linux-kernel@vger.kernel.org
In-Reply-To: <CAJ+vNU0VzgDkH-ksHgaA0cn3nbCOfcUs_BU90v6xw_QappssgA@mail.gmail.com>
On Wed, Aug 30, 2017 at 12:53:56PM -0700, Tim Harvey wrote:
> Greetings,
>
> I'm seeing RX frame errors when using the mv88e6xxx DSA driver on
> 4.13-rc7. The board I'm using is a GW5904 [1] which has an IMX6 FEC
> MAC (eth0) connected via RGMII to a MV88E6176 with its downstream
> P0/P1/P2/P3 to front panel RJ45's (lan1-lan4).
Hi Tim
Can you confirm the counter is this one:
/* Report late collisions as a frame error. */
if (status & (BD_ENET_RX_NO | BD_ENET_RX_CL))
ndev->stats.rx_frame_errors++;
I don't see anywhere else frame errors are counted, but it would be
good to prove we are looking in the right place.
Andrew
^ permalink raw reply
* Re: [Patch net-next] net_sched: add reverse binding for tc class
From: Cong Wang @ 2017-08-30 22:01 UTC (permalink / raw)
To: Daniel Borkmann; +Cc: Linux Kernel Network Developers, Jamal Hadi Salim
In-Reply-To: <59A732B8.6020405@iogearbox.net>
On Wed, Aug 30, 2017 at 2:48 PM, Daniel Borkmann <daniel@iogearbox.net> wrote:
> On 08/30/2017 11:30 PM, Cong Wang wrote:
> [...]
>>
>> Note, we still can NOT totally get rid of those class lookup in
>> ->enqueue() because cgroup and flow filters have no way to determine
>> the classid at setup time, they still have to go through dynamic lookup.
>
> [...]
>>
>> ---
>> include/net/sch_generic.h | 1 +
>> net/sched/cls_basic.c | 9 +++++++
>> net/sched/cls_bpf.c | 9 +++++++
>
>
> Same is for cls_bpf as well, so bind_class wouldn't work there
> either as we could return dynamic classids. bind_class cannot
> be added here, too.
I think you are probably right, but the following code is
misleading there:
if (tb[TCA_BPF_CLASSID]) {
prog->res.classid = nla_get_u32(tb[TCA_BPF_CLASSID]);
tcf_bind_filter(tp, &prog->res, base);
}
If the classid is dynamic, why this tb[TCA_BPF_CLASSID]?
^ permalink raw reply
* Re: [PATCH] DSA support for Micrel KSZ8895
From: Andrew Lunn @ 2017-08-30 22:00 UTC (permalink / raw)
To: Tristram.Ha
Cc: pavel, Woojung.Huh, nathan.leigh.conrad, vivien.didelot,
f.fainelli, netdev, linux-kernel
In-Reply-To: <93AF473E2DA327428DE3D46B72B1E9FD4111FA48@CHN-SV-EXMX02.mchp-main.com>
> The KSZ8795 driver will be submitted after Labor Day (9/4) if
> testing reveals no problem. The KSZ8895 driver will be submitted
> right after that. You should have no problem using the driver right
> away.
Hi Tristram
Release early, release often. It stops people wasting time....
Also, we are likely to give you feedback, asking you to make
changes. Testing is important, but you are probably going to have to
do it a number of times. So it is not everything passes now, don't
worry, you will have time to fix things up as you go through review
cycles.
Andrew
^ permalink raw reply
* Re: [PATCH net-next] dp83640: don't hold spinlock while calling netif_rx_ni
From: David Miller @ 2017-08-30 21:50 UTC (permalink / raw)
To: stefan.sorensen; +Cc: richardcochran, netdev
In-Reply-To: <20170830065847.2444-1-stefan.sorensen@spectralink.com>
From: Stefan Sørensen <stefan.sorensen@spectralink.com>
Date: Wed, 30 Aug 2017 08:58:47 +0200
> We should not hold a spinlock while pushing the skb into the networking
> stack, so move the call to netif_rx_ni out of the critical region to where
> we have dropped the spinlock.
>
> Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Looks good, applied, thanks.
^ permalink raw reply
* Re: [Patch net-next] net_sched: add reverse binding for tc class
From: Daniel Borkmann @ 2017-08-30 21:48 UTC (permalink / raw)
To: Cong Wang, netdev; +Cc: Jamal Hadi Salim
In-Reply-To: <20170830213036.24250-1-xiyou.wangcong@gmail.com>
On 08/30/2017 11:30 PM, Cong Wang wrote:
[...]
> Note, we still can NOT totally get rid of those class lookup in
> ->enqueue() because cgroup and flow filters have no way to determine
> the classid at setup time, they still have to go through dynamic lookup.
[...]
> ---
> include/net/sch_generic.h | 1 +
> net/sched/cls_basic.c | 9 +++++++
> net/sched/cls_bpf.c | 9 +++++++
Same is for cls_bpf as well, so bind_class wouldn't work there
either as we could return dynamic classids. bind_class cannot
be added here, too.
> net/sched/cls_flower.c | 9 +++++++
> net/sched/cls_fw.c | 9 +++++++
> net/sched/cls_matchall.c | 9 +++++++
> net/sched/cls_route.c | 9 +++++++
> net/sched/cls_rsvp.h | 9 +++++++
> net/sched/cls_tcindex.c | 9 +++++++
> net/sched/cls_u32.c | 9 +++++++
> net/sched/sch_api.c | 68 +++++++++++++++++++++++++++++++++++++++++++++--
> 11 files changed, 148 insertions(+), 2 deletions(-)
^ permalink raw reply
* Re: [PATCH net-next] net: cpsw: Don't handle SIOC[GS]HWTSTAMP when CPTS is disabled
From: David Miller @ 2017-08-30 21:47 UTC (permalink / raw)
To: stefan.sorensen; +Cc: grygorii.strashko, netdev, linux-omap
In-Reply-To: <20170830065055.419-1-stefan.sorensen@spectralink.com>
From: Stefan Sørensen <stefan.sorensen@spectralink.com>
Date: Wed, 30 Aug 2017 08:50:55 +0200
> There is no reason to handle SIOC[GS]HWTSTAMP and return -EOPNOTSUPP when
> CPTS is disabled, so just pass them on to the phy. This will allow PTP
> timestamping on a capable phy by disabling CPTS.
>
> Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
It should not be required to disable a Kconfig option just to get PHY
timestamping to work properly.
Rather, if the CPTS code returns -EOPNOTSUPP we should try to
fallthrough to the PHY library based methods.
Thanks.
^ permalink raw reply
* Re: [PATCH net-next 3/3 v11] drivers: net: ethernet: qualcomm: rmnet: Initial implementation
From: Subash Abhinov Kasiviswanathan @ 2017-08-30 21:40 UTC (permalink / raw)
To: David Miller
Cc: dcbw, netdev, fengguang.wu, jiri, stephen, David.Laight, marcel,
andrew
In-Reply-To: <20170830.142744.324884406082731585.davem@davemloft.net>
> Subash, keep in mind that since I applied your v11 patches already you
> will need to send me relative fixes and changes at this point, rather
> than resubmit the series.
>
> Thank you.
Thanks for the heads up David. Will do.
--
Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a
Linux Foundation Collaborative Project
^ permalink raw reply
* Re: [patch net-next v2 0/3] net/sched: Improve getting objects by indexes
From: David Miller @ 2017-08-30 21:39 UTC (permalink / raw)
To: chrism; +Cc: netdev, jhs, xiyou.wangcong, jiri, mawilcox
In-Reply-To: <1504074719-15147-1-git-send-email-chrism@mellanox.com>
From: Chris Mi <chrism@mellanox.com>
Date: Wed, 30 Aug 2017 02:31:56 -0400
> Using current TC code, it is very slow to insert a lot of rules.
>
> In order to improve the rules update rate in TC,
> we introduced the following two changes:
> 1) changed cls_flower to use IDR to manage the filters.
> 2) changed all act_xxx modules to use IDR instead of
> a small hash table
>
> But IDR has a limitation that it uses int. TC handle uses u32.
> To make sure there is no regression, we add several new IDR APIs
> to support unsigned long.
>
> v2
> ==
>
> Addressed Hannes's comment:
> express idr_alloc in terms of idr_alloc_ext and most of the other functions
Series applied, thanks.
^ permalink raw reply
* Re: [PATCH net 9/9] sch_tbf: fix two null pointer dereferences on init failure
From: Nikolay Aleksandrov @ 2017-08-30 21:39 UTC (permalink / raw)
To: Cong Wang
Cc: Linux Kernel Network Developers, Eric Dumazet, Jamal Hadi Salim,
Jiri Pirko, Roopa Prabhu
In-Reply-To: <CAM_iQpWn_q4CaOKhw7nxFR40V2JaVwr_5kx-ZUpdFxPa4hxKyA@mail.gmail.com>
On 30/08/17 20:37, Cong Wang wrote:
> On Wed, Aug 30, 2017 at 2:49 AM, Nikolay Aleksandrov
> <nikolay@cumulusnetworks.com> wrote:
>> Reproduce:
>> $ sysctl net.core.default_qdisc=tbf
>> $ ip l set ethX up
>
> I once upon a time had a patch to disallow those qdisc's
> to be default. Probably I should resend it.
>
That sounds good. A lot of them can't be default anyway, they need
some options to be set, so it's definitely worth looking into.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox