* [PATCH net 1/9] net/mlx5_core: Fix soft lockup in steering error flow
From: Saeed Mahameed @ 2016-04-19 12:33 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Or Gerlitz, Tal Alon, Eran Ben Elisha, Maor Gottlieb,
Saeed Mahameed
In-Reply-To: <1461069222-27076-1-git-send-email-saeedm@mellanox.com>
From: Maor Gottlieb <maorg@mellanox.com>
In the error flow of adding flow rule to auto-grouped flow
table, we call to tree_remove_node.
tree_remove_node locks the node's parent, however the node's parent
is already locked by mlx5_add_flow_rule and this causes a deadlock.
After this patch, if we failed to add the flow rule, we unlock the
flow table before calling to tree_remove_node.
fixes: f0d22d187473 ('net/mlx5_core: Introduce flow steering autogrouped
flow table')
Signed-off-by: Maor Gottlieb <maorg@mellanox.com>
Reported-by: Amir Vadai <amir@vadai.me>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 46 ++++++++-------------
1 files changed, 17 insertions(+), 29 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index 5121be4..3c7e3e5 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -1065,33 +1065,6 @@ unlock_fg:
return rule;
}
-static struct mlx5_flow_rule *add_rule_to_auto_fg(struct mlx5_flow_table *ft,
- u8 match_criteria_enable,
- u32 *match_criteria,
- u32 *match_value,
- u8 action,
- u32 flow_tag,
- struct mlx5_flow_destination *dest)
-{
- struct mlx5_flow_rule *rule;
- struct mlx5_flow_group *g;
-
- g = create_autogroup(ft, match_criteria_enable, match_criteria);
- if (IS_ERR(g))
- return (void *)g;
-
- rule = add_rule_fg(g, match_value,
- action, flow_tag, dest);
- if (IS_ERR(rule)) {
- /* Remove assumes refcount > 0 and autogroup creates a group
- * with a refcount = 0.
- */
- tree_get_node(&g->node);
- tree_remove_node(&g->node);
- }
- return rule;
-}
-
static struct mlx5_flow_rule *
_mlx5_add_flow_rule(struct mlx5_flow_table *ft,
u8 match_criteria_enable,
@@ -1119,8 +1092,23 @@ _mlx5_add_flow_rule(struct mlx5_flow_table *ft,
goto unlock;
}
- rule = add_rule_to_auto_fg(ft, match_criteria_enable, match_criteria,
- match_value, action, flow_tag, dest);
+ g = create_autogroup(ft, match_criteria_enable, match_criteria);
+ if (IS_ERR(g)) {
+ rule = (void *)g;
+ goto unlock;
+ }
+
+ rule = add_rule_fg(g, match_value,
+ action, flow_tag, dest);
+ if (IS_ERR(rule)) {
+ /* Remove assumes refcount > 0 and autogroup creates a group
+ * with a refcount = 0.
+ */
+ unlock_ref_node(&ft->node);
+ tree_get_node(&g->node);
+ tree_remove_node(&g->node);
+ return rule;
+ }
unlock:
unlock_ref_node(&ft->node);
return rule;
--
1.7.1
^ permalink raw reply related
* [PATCH net 4/9] net/mlx5e: Device's mtu field is u16 and not int
From: Saeed Mahameed @ 2016-04-19 12:33 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Or Gerlitz, Tal Alon, Eran Ben Elisha, Saeed Mahameed
In-Reply-To: <1461069222-27076-1-git-send-email-saeedm@mellanox.com>
For set/query MTU port firmware commands the MTU field
is 16 bits, here I changed all the "int mtu" parameters
of the functions wrapping those firmware commands to be u16.
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 4 ++--
drivers/net/ethernet/mellanox/mlx5/core/port.c | 10 +++++-----
include/linux/mlx5/port.h | 6 +++---
3 files changed, 10 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index e0adb60..2fbbc62 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -1408,7 +1408,7 @@ static int mlx5e_set_dev_port_mtu(struct net_device *netdev)
{
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
- int hw_mtu;
+ u16 hw_mtu;
int err;
err = mlx5_set_port_mtu(mdev, MLX5E_SW2HW_MTU(netdev->mtu), 1);
@@ -2004,7 +2004,7 @@ static int mlx5e_change_mtu(struct net_device *netdev, int new_mtu)
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
bool was_opened;
- int max_mtu;
+ u16 max_mtu;
int err = 0;
mlx5_query_port_max_mtu(mdev, &max_mtu, 1);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/port.c b/drivers/net/ethernet/mellanox/mlx5/core/port.c
index ae378c5..53cc1e2 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/port.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/port.c
@@ -247,8 +247,8 @@ int mlx5_query_port_admin_status(struct mlx5_core_dev *dev,
}
EXPORT_SYMBOL_GPL(mlx5_query_port_admin_status);
-static void mlx5_query_port_mtu(struct mlx5_core_dev *dev, int *admin_mtu,
- int *max_mtu, int *oper_mtu, u8 port)
+static void mlx5_query_port_mtu(struct mlx5_core_dev *dev, u16 *admin_mtu,
+ u16 *max_mtu, u16 *oper_mtu, u8 port)
{
u32 in[MLX5_ST_SZ_DW(pmtu_reg)];
u32 out[MLX5_ST_SZ_DW(pmtu_reg)];
@@ -268,7 +268,7 @@ static void mlx5_query_port_mtu(struct mlx5_core_dev *dev, int *admin_mtu,
*admin_mtu = MLX5_GET(pmtu_reg, out, admin_mtu);
}
-int mlx5_set_port_mtu(struct mlx5_core_dev *dev, int mtu, u8 port)
+int mlx5_set_port_mtu(struct mlx5_core_dev *dev, u16 mtu, u8 port)
{
u32 in[MLX5_ST_SZ_DW(pmtu_reg)];
u32 out[MLX5_ST_SZ_DW(pmtu_reg)];
@@ -283,14 +283,14 @@ int mlx5_set_port_mtu(struct mlx5_core_dev *dev, int mtu, u8 port)
}
EXPORT_SYMBOL_GPL(mlx5_set_port_mtu);
-void mlx5_query_port_max_mtu(struct mlx5_core_dev *dev, int *max_mtu,
+void mlx5_query_port_max_mtu(struct mlx5_core_dev *dev, u16 *max_mtu,
u8 port)
{
mlx5_query_port_mtu(dev, NULL, max_mtu, NULL, port);
}
EXPORT_SYMBOL_GPL(mlx5_query_port_max_mtu);
-void mlx5_query_port_oper_mtu(struct mlx5_core_dev *dev, int *oper_mtu,
+void mlx5_query_port_oper_mtu(struct mlx5_core_dev *dev, u16 *oper_mtu,
u8 port)
{
mlx5_query_port_mtu(dev, NULL, NULL, oper_mtu, port);
diff --git a/include/linux/mlx5/port.h b/include/linux/mlx5/port.h
index a1d145a..b30250a 100644
--- a/include/linux/mlx5/port.h
+++ b/include/linux/mlx5/port.h
@@ -54,9 +54,9 @@ int mlx5_set_port_admin_status(struct mlx5_core_dev *dev,
int mlx5_query_port_admin_status(struct mlx5_core_dev *dev,
enum mlx5_port_status *status);
-int mlx5_set_port_mtu(struct mlx5_core_dev *dev, int mtu, u8 port);
-void mlx5_query_port_max_mtu(struct mlx5_core_dev *dev, int *max_mtu, u8 port);
-void mlx5_query_port_oper_mtu(struct mlx5_core_dev *dev, int *oper_mtu,
+int mlx5_set_port_mtu(struct mlx5_core_dev *dev, u16 mtu, u8 port);
+void mlx5_query_port_max_mtu(struct mlx5_core_dev *dev, u16 *max_mtu, u8 port);
+void mlx5_query_port_oper_mtu(struct mlx5_core_dev *dev, u16 *oper_mtu,
u8 port);
int mlx5_query_port_vl_hw_cap(struct mlx5_core_dev *dev,
--
1.7.1
^ permalink raw reply related
* [PATCH net 2/9] net/mlx5e: Fix MLX5E_100BASE_T define
From: Saeed Mahameed @ 2016-04-19 12:33 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Or Gerlitz, Tal Alon, Eran Ben Elisha, Rana Shahout,
Saeed Mahameed
In-Reply-To: <1461069222-27076-1-git-send-email-saeedm@mellanox.com>
From: Rana Shahout <ranas@mellanox.com>
Bit 25 of eth_proto_capability in PTYS register is
1000Base-TT and not 100Base-T.
Fixes: f62b8bb8f2d3 ('net/mlx5: Extend mlx5_core to
support ConnectX-4 Ethernet functionality')
Signed-off-by: Rana Shahout <ranas@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en.h | 2 +-
.../net/ethernet/mellanox/mlx5/core/en_ethtool.c | 8 ++++----
2 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 879e627..e80ce94 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -609,7 +609,7 @@ enum mlx5e_link_mode {
MLX5E_100GBASE_KR4 = 22,
MLX5E_100GBASE_LR4 = 23,
MLX5E_100BASE_TX = 24,
- MLX5E_100BASE_T = 25,
+ MLX5E_1000BASE_T = 25,
MLX5E_10GBASE_T = 26,
MLX5E_25GBASE_CR = 27,
MLX5E_25GBASE_KR = 28,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index 68834b7..3476ab8 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -138,10 +138,10 @@ static const struct {
[MLX5E_100BASE_TX] = {
.speed = 100,
},
- [MLX5E_100BASE_T] = {
- .supported = SUPPORTED_100baseT_Full,
- .advertised = ADVERTISED_100baseT_Full,
- .speed = 100,
+ [MLX5E_1000BASE_T] = {
+ .supported = SUPPORTED_1000baseT_Full,
+ .advertised = ADVERTISED_1000baseT_Full,
+ .speed = 1000,
},
[MLX5E_10GBASE_T] = {
.supported = SUPPORTED_10000baseT_Full,
--
1.7.1
^ permalink raw reply related
* [PATCH net 9/9] net/mlx5e: Reset link modes upon setting speed to zero
From: Saeed Mahameed @ 2016-04-19 12:33 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Or Gerlitz, Tal Alon, Eran Ben Elisha, Saeed Mahameed
In-Reply-To: <1461069222-27076-1-git-send-email-saeedm@mellanox.com>
Upon ethtool request to set speed to 0 we handle it as a special request
to reset link modes to Device's defaults.
Fixes: f62b8bb8f2d3 ("net/mlx5: Extend mlx5_core to support ConnectX-4
Ethernet functionality")
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
.../net/ethernet/mellanox/mlx5/core/en_ethtool.c | 18 +++++++++++-------
1 files changed, 11 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index 3476ab8..206f7fc 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -640,17 +640,10 @@ static int mlx5e_set_settings(struct net_device *netdev,
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
u32 link_modes;
- u32 speed;
u32 eth_proto_cap, eth_proto_admin;
enum mlx5_port_status ps;
int err;
- speed = ethtool_cmd_speed(cmd);
-
- link_modes = cmd->autoneg == AUTONEG_ENABLE ?
- mlx5e_ethtool2ptys_adver_link(cmd->advertising) :
- mlx5e_ethtool2ptys_speed_link(speed);
-
err = mlx5_query_port_proto_cap(mdev, ð_proto_cap, MLX5_PTYS_EN);
if (err) {
netdev_err(netdev, "%s: query port eth proto cap failed: %d\n",
@@ -658,6 +651,17 @@ static int mlx5e_set_settings(struct net_device *netdev,
goto out;
}
+ if (cmd->autoneg == AUTONEG_ENABLE) {
+ link_modes = mlx5e_ethtool2ptys_adver_link(cmd->advertising);
+ } else {
+ u32 speed = ethtool_cmd_speed(cmd);
+
+ if (speed)
+ link_modes = mlx5e_ethtool2ptys_speed_link(speed);
+ else /* speed 0 means reset link_modes to Device's default */
+ link_modes = eth_proto_cap;
+ }
+
link_modes = link_modes & eth_proto_cap;
if (!link_modes) {
netdev_err(netdev, "%s: Not supported link mode(s) requested",
--
1.7.1
^ permalink raw reply related
* [PATCH net 5/9] net/mlx5e: Fix minimum MTU
From: Saeed Mahameed @ 2016-04-19 12:33 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Or Gerlitz, Tal Alon, Eran Ben Elisha, Saeed Mahameed
In-Reply-To: <1461069222-27076-1-git-send-email-saeedm@mellanox.com>
Minimum MTU that can be set in Connectx4 device is 68.
This fixes the case where a user wants to set invalid MTU,
the driver will fail to satisfy this request and the interface
will stay down.
It is better to report an error and continue working with old
mtu.
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 11 ++++++++---
1 files changed, 8 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 2fbbc62..93e4ef4 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -1999,22 +1999,27 @@ static int mlx5e_set_features(struct net_device *netdev,
return err;
}
+#define MXL5_HW_MIN_MTU 64
+#define MXL5E_MIN_MTU (MXL5_HW_MIN_MTU + ETH_FCS_LEN)
+
static int mlx5e_change_mtu(struct net_device *netdev, int new_mtu)
{
struct mlx5e_priv *priv = netdev_priv(netdev);
struct mlx5_core_dev *mdev = priv->mdev;
bool was_opened;
u16 max_mtu;
+ u16 min_mtu;
int err = 0;
mlx5_query_port_max_mtu(mdev, &max_mtu, 1);
max_mtu = MLX5E_HW2SW_MTU(max_mtu);
+ min_mtu = MLX5E_HW2SW_MTU(MXL5E_MIN_MTU);
- if (new_mtu > max_mtu) {
+ if (new_mtu > max_mtu || new_mtu < min_mtu) {
netdev_err(netdev,
- "%s: Bad MTU (%d) > (%d) Max\n",
- __func__, new_mtu, max_mtu);
+ "%s: Bad MTU (%d), valid range is: [%d..%d]\n",
+ __func__, new_mtu, min_mtu, max_mtu);
return -EINVAL;
}
--
1.7.1
^ permalink raw reply related
* Re: [PATCHv3 net-next] net: use jiffies_to_msecs to replace EXPIRES_IN_MS in inet/sctp_diag
From: Xin Long @ 2016-04-19 12:37 UTC (permalink / raw)
To: Eric Dumazet
Cc: network dev, linux-sctp, Marcelo Ricardo Leitner, Vlad Yasevich,
daniel, davem, jsitnick
In-Reply-To: <1461068732.10638.169.camel@edumazet-glaptop3.roam.corp.google.com>
On Tue, Apr 19, 2016 at 8:25 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> On Tue, 2016-04-19 at 15:10 +0800, Xin Long wrote:
>> EXPIRES_IN_MS macro comes from net/ipv4/inet_diag.c and dates
>> back to before jiffies_to_msecs() has been introduced.
>>
>> Now we can remove it and use jiffies_to_msecs().
>>
>> Signed-off-by: Xin Long <lucien.xin@gmail.com>
>> ---
>
> Ok. Note that you could have mentioned this was
>
> Suggested-by: Jakub Sitnicki <jkbs@redhat.com>
>
> Even coworkers deserve credits ;)
yeah, didn't know the tag "Suggested-by:" before.
Thanks. :D
>
> Acked-by: Eric Dumazet <edumazet@google.com>
>
> Thanks.
>
>
>
^ permalink raw reply
* [PATCH 1/4] net: thunderx: Introduce a mailbox message to reset VF counters
From: sunil.kovvuri @ 2016-04-19 13:33 UTC (permalink / raw)
To: netdev
Cc: linux-kernel, linux-arm-kernel, sgoutham, robert.richter,
Jerin Jacob
In-Reply-To: <1461072812-44239-1-git-send-email-sunil.kovvuri@gmail.com>
From: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Write access to VF statistics counter register is only allowed from PF.
Added a new mailbox message to reset VF's Rx/Tx counters, this is used
by userspace DPDK.
Signed-off-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Signed-off-by: Sunil Goutham <sgoutham@cavium.com>
---
drivers/net/ethernet/cavium/thunder/nic.h | 27 ++++++++++++++
drivers/net/ethernet/cavium/thunder/nic_main.c | 45 ++++++++++++++++++++++++
2 files changed, 72 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ethernet/cavium/thunder/nic.h b/drivers/net/ethernet/cavium/thunder/nic.h
index 83025bb..e2ac9bd 100644
--- a/drivers/net/ethernet/cavium/thunder/nic.h
+++ b/drivers/net/ethernet/cavium/thunder/nic.h
@@ -368,6 +368,7 @@ struct nicvf {
#define NIC_MBOX_MSG_PNICVF_PTR 0x14 /* Get primary qset nicvf ptr */
#define NIC_MBOX_MSG_SNICVF_PTR 0x15 /* Send sqet nicvf ptr to PVF */
#define NIC_MBOX_MSG_LOOPBACK 0x16 /* Set interface in loopback */
+#define NIC_MBOX_MSG_RESET_STAT_COUNTER 0x17 /* Reset statistics counters */
#define NIC_MBOX_MSG_CFG_DONE 0xF0 /* VF configuration done */
#define NIC_MBOX_MSG_SHUTDOWN 0xF1 /* VF is being shutdown */
@@ -484,6 +485,31 @@ struct set_loopback {
bool enable;
};
+/* Reset statistics counters */
+struct reset_stat_cfg {
+ u8 msg;
+ /* Bitmap to select NIC_PF_VNIC(vf_id)_RX_STAT(0..13) */
+ u16 rx_stat_mask;
+ /* Bitmap to select NIC_PF_VNIC(vf_id)_TX_STAT(0..4) */
+ u8 tx_stat_mask;
+ /* Bitmap to select NIC_PF_QS(0..127)_RQ(0..7)_STAT(0..1)
+ * bit14, bit15 NIC_PF_QS(vf_id)_RQ7_STAT(0..1)
+ * bit12, bit13 NIC_PF_QS(vf_id)_RQ6_STAT(0..1)
+ * ..
+ * bit2, bit3 NIC_PF_QS(vf_id)_RQ1_STAT(0..1)
+ * bit0, bit1 NIC_PF_QS(vf_id)_RQ0_STAT(0..1)
+ */
+ u16 rq_stat_mask;
+ /* Bitmap to select NIC_PF_QS(0..127)_SQ(0..7)_STAT(0..1)
+ * bit14, bit15 NIC_PF_QS(vf_id)_SQ7_STAT(0..1)
+ * bit12, bit13 NIC_PF_QS(vf_id)_SQ6_STAT(0..1)
+ * ..
+ * bit2, bit3 NIC_PF_QS(vf_id)_SQ1_STAT(0..1)
+ * bit0, bit1 NIC_PF_QS(vf_id)_SQ0_STAT(0..1)
+ */
+ u16 sq_stat_mask;
+};
+
/* 128 bit shared memory between PF and each VF */
union nic_mbx {
struct { u8 msg; } msg;
@@ -501,6 +527,7 @@ union nic_mbx {
struct sqs_alloc sqs_alloc;
struct nicvf_ptr nicvf;
struct set_loopback lbk;
+ struct reset_stat_cfg reset_stat;
};
#define NIC_NODE_ID_MASK 0x03
diff --git a/drivers/net/ethernet/cavium/thunder/nic_main.c b/drivers/net/ethernet/cavium/thunder/nic_main.c
index 95f17f8..77ee260 100644
--- a/drivers/net/ethernet/cavium/thunder/nic_main.c
+++ b/drivers/net/ethernet/cavium/thunder/nic_main.c
@@ -616,6 +616,48 @@ static int nic_config_loopback(struct nicpf *nic, struct set_loopback *lbk)
return 0;
}
+/* Reset statistics counters */
+static int nic_reset_stat_counters(struct nicpf *nic,
+ int vf, struct reset_stat_cfg *cfg)
+{
+ int i, stat, qnum;
+ u64 reg_addr;
+
+ for (i = 0; i < RX_STATS_ENUM_LAST; i++) {
+ if (cfg->rx_stat_mask & BIT(i)) {
+ reg_addr = NIC_PF_VNIC_0_127_RX_STAT_0_13 |
+ (vf << NIC_QS_ID_SHIFT) |
+ (i << 3);
+ nic_reg_write(nic, reg_addr, 0);
+ }
+ }
+
+ for (i = 0; i < TX_STATS_ENUM_LAST; i++) {
+ if (cfg->tx_stat_mask & BIT(i)) {
+ reg_addr = NIC_PF_VNIC_0_127_TX_STAT_0_4 |
+ (vf << NIC_QS_ID_SHIFT) |
+ (i << 3);
+ nic_reg_write(nic, reg_addr, 0);
+ }
+ }
+
+ for (i = 0; i <= 15; i++) {
+ qnum = i >> 1;
+ stat = i & 1 ? 1 : 0;
+ reg_addr = (vf << NIC_QS_ID_SHIFT) |
+ (qnum << NIC_Q_NUM_SHIFT) | (stat << 3);
+ if (cfg->rq_stat_mask & BIT(i)) {
+ reg_addr |= NIC_PF_QSET_0_127_RQ_0_7_STAT_0_1;
+ nic_reg_write(nic, reg_addr, 0);
+ }
+ if (cfg->sq_stat_mask & BIT(i)) {
+ reg_addr |= NIC_PF_QSET_0_127_SQ_0_7_STAT_0_1;
+ nic_reg_write(nic, reg_addr, 0);
+ }
+ }
+ return 0;
+}
+
static void nic_enable_vf(struct nicpf *nic, int vf, bool enable)
{
int bgx, lmac;
@@ -757,6 +799,9 @@ static void nic_handle_mbx_intr(struct nicpf *nic, int vf)
case NIC_MBOX_MSG_LOOPBACK:
ret = nic_config_loopback(nic, &mbx.lbk);
break;
+ case NIC_MBOX_MSG_RESET_STAT_COUNTER:
+ ret = nic_reset_stat_counters(nic, vf, &mbx.reset_stat);
+ break;
default:
dev_err(&nic->pdev->dev,
"Invalid msg from VF%d, msg 0x%x\n", vf, mbx.msg.msg);
--
1.7.1
^ permalink raw reply related
* [RFC PATCH net-next 0/8] Handle multiple received packets at each stage
From: Edward Cree @ 2016-04-19 13:33 UTC (permalink / raw)
To: netdev, David Miller; +Cc: Jesper Dangaard Brouer, linux-net-drivers
Earlier discussions on this list[1] suggested that having multiple packets
traverse the network stack together (rather than calling the stack for each
packet singly) could improve performance through better cache locality.
This patch series is an attempt to implement this by having drivers pass an
SKB list to the stack at the end of the NAPI poll. The stack then attempts
to keep the list together, only splitting it when either packets need to be
treated differently, or the next layer of the stack is not list-aware.
The first two patches simply place received packets on a list during the
event processing loop on the sfc EF10 architecture, then call the normal
stack for each packet singly at the end of the NAPI poll.
The remaining patches extend the 'listified' processing as far as the IP
receive handler.
Packet rate was tested with NetPerf UDP_STREAM, with 10 streams of 1-byte
packets, and the process and interrupt pinned to a single core on the RX
side.
The NIC was a 40G Solarflare 7x42Q; the CPU was a Xeon E3-1220V2 @ 3.10GHz.
Baseline: 5.07Mpps
after patch 2: 5.59Mpps (10.2% above baseline)
after patch 8: 6.44Mpps (25.6% above baseline)
I also attempted to measure the latency, but couldn't get reliable numbers;
my best estimate is that the series cost about 160ns if interrupt moderation
is disabled and busy-poll is enabled; about 60ns vice-versa.
I tried adding a check in the driver to only perform bundling if interrupt
moderation was active on the channel, but was unable to demonstrate any
latency gain from this, so I have omitted it from this series.
[1] http://thread.gmane.org/gmane.linux.network/395502
Edward Cree (8):
net: core: trivial netif_receive_skb_list() entry point
sfc: batch up RX delivery on EF10
net: core: unwrap skb list receive slightly further
net: core: Another step of skb receive list processing
net: core: another layer of lists, around PF_MEMALLOC skb handling
net: core: propagate SKB lists through packet_type lookup
net: ipv4: listified version of ip_rcv
net: ipv4: listify ip_rcv_finish
drivers/net/ethernet/sfc/ef10.c | 9 ++
drivers/net/ethernet/sfc/efx.c | 2 +
drivers/net/ethernet/sfc/net_driver.h | 3 +
drivers/net/ethernet/sfc/rx.c | 7 +-
include/linux/netdevice.h | 4 +
include/linux/netfilter.h | 27 ++++
include/linux/skbuff.h | 16 +++
include/net/ip.h | 2 +
include/trace/events/net.h | 14 ++
net/core/dev.c | 245 ++++++++++++++++++++++++++++------
net/ipv4/af_inet.c | 1 +
net/ipv4/ip_input.c | 127 ++++++++++++++++--
12 files changed, 409 insertions(+), 48 deletions(-)
^ permalink raw reply
* [PATCH 0/4] net: thunderx: Add multiqset support for DPDK
From: sunil.kovvuri @ 2016-04-19 13:33 UTC (permalink / raw)
To: netdev; +Cc: linux-kernel, linux-arm-kernel, sgoutham, robert.richter
From: Sunil Goutham <sgoutham@cavium.com>
This patch series mainly adds support for userspace application
like DPDK with a VNIC VF attached to request additional QSets
for having morethan the default 8 queues.
Jerin Jacob (1):
net: thunderx: Introduce a mailbox message to reset VF counters
Radoslaw Biernacki (3):
net: thunderx: Add multiqset support for dataplane apps
net: thunderx: add sysfs attribute for SQS/SVF assigments
net: thunderx: Improvement for MBX interface debug messages
drivers/net/ethernet/cavium/thunder/nic.h | 32 ++-
drivers/net/ethernet/cavium/thunder/nic_main.c | 321 +++++++++++++++++++---
drivers/net/ethernet/cavium/thunder/nicvf_main.c | 10 +-
3 files changed, 324 insertions(+), 39 deletions(-)
^ permalink raw reply
* [PATCH 2/4] net: thunderx: Add multiqset support for dataplane apps
From: sunil.kovvuri @ 2016-04-19 13:33 UTC (permalink / raw)
To: netdev
Cc: linux-kernel, linux-arm-kernel, sgoutham, robert.richter,
Radoslaw Biernacki
In-Reply-To: <1461072812-44239-1-git-send-email-sunil.kovvuri@gmail.com>
From: Radoslaw Biernacki <rad@semihalf.com>
This patch adds support to PF for allocating additional Qsets to
dataplane apps such as DPDK. Till now PF, upon host bound interface's
request it used to allocate Qsets from the free ones, but for dataplane
apps support has been added for it to request specific Qsets instead of
just PF's choice.
And also adds validation checks at different places, these are needed to
have proper secondary Qset allocation when interfaces in different domain
i.e Host, VFIO, DPDK e.t.c exist at the same time.
Some of the checks are
- Check if RSS indirection table has valid entries.
- When host bound interface requests additional Qsets, PF should
assign only those which in host domain i.e both primary VF and
secondary VFs should be using same driver. Hence added PCI driver
checks.
- If dataplane app terminates without proper shutdown then when
restarted it will request the same or different SQsets as were
assigned before. This is taken care of otherwise application
won't recover.
Removed 'sqs_used' which became redundant due to new SQset allocation scheme.
Signed-off-by: Radoslaw Biernacki <rad@semihalf.com>
Signed-off-by: Sunil Goutham <sgoutham@cavium.com>
---
drivers/net/ethernet/cavium/thunder/nic.h | 5 +-
drivers/net/ethernet/cavium/thunder/nic_main.c | 193 +++++++++++++++++++---
drivers/net/ethernet/cavium/thunder/nicvf_main.c | 2 +-
3 files changed, 170 insertions(+), 30 deletions(-)
diff --git a/drivers/net/ethernet/cavium/thunder/nic.h b/drivers/net/ethernet/cavium/thunder/nic.h
index e2ac9bd..b63278a 100644
--- a/drivers/net/ethernet/cavium/thunder/nic.h
+++ b/drivers/net/ethernet/cavium/thunder/nic.h
@@ -463,11 +463,12 @@ struct bgx_link_status {
u32 speed;
};
-/* Get Extra Qset IDs */
+/* Allocate additional SQS to VF */
struct sqs_alloc {
u8 msg;
- u8 vf_id;
+ u8 spec; /* 1 - For specific SQS allocation, 0 - For PF's choice */
u8 qs_count;
+ u8 svf[MAX_SQS_PER_VF]; /* SQS VF ids for specific allocation */
};
struct nicvf_ptr {
diff --git a/drivers/net/ethernet/cavium/thunder/nic_main.c b/drivers/net/ethernet/cavium/thunder/nic_main.c
index 77ee260..d6a6914 100644
--- a/drivers/net/ethernet/cavium/thunder/nic_main.c
+++ b/drivers/net/ethernet/cavium/thunder/nic_main.c
@@ -29,9 +29,9 @@ struct nicpf {
void __iomem *reg_base; /* Register start address */
u8 num_sqs_en; /* Secondary qsets enabled */
u64 nicvf[MAX_NUM_VFS_SUPPORTED];
+#define NIC_VF_UNASSIGNED ((u8)0xFF)
u8 vf_sqs[MAX_NUM_VFS_SUPPORTED][MAX_SQS_PER_VF];
u8 pqs_vf[MAX_NUM_VFS_SUPPORTED];
- bool sqs_used[MAX_NUM_VFS_SUPPORTED];
struct pkind_cfg pkind;
#define NIC_SET_VF_LMAC_MAP(bgx, lmac) (((bgx & 0xF) << 4) | (lmac & 0xF))
#define NIC_GET_BGX_FROM_VF_LMAC_MAP(map) ((map >> 4) & 0xF)
@@ -46,6 +46,7 @@ struct nicpf {
u16 rssi_base[MAX_NUM_VFS_SUPPORTED];
u16 rss_ind_tbl_size;
bool mbx_lock[MAX_NUM_VFS_SUPPORTED];
+ struct pci_dev *vf_pdev[MAX_NUM_VFS_SUPPORTED];
/* MSI-X */
bool msix_enabled;
@@ -458,10 +459,18 @@ static void nic_config_rss(struct nicpf *nic, struct rss_cfg_msg *cfg)
for (; rssi < (rssi_base + cfg->tbl_len); rssi++) {
u8 svf = cfg->ind_tbl[idx] >> 3;
- if (svf)
+ if (svf && (svf <= MAX_SQS_PER_VF)) {
qset = nic->vf_sqs[cfg->vf_id][svf - 1];
- else
+ if ((qset >= MAX_NUM_VFS_SUPPORTED) ||
+ (nic->pqs_vf[qset] != cfg->vf_id)) {
+ dev_err(&nic->pdev->dev,
+ "Invalid rss table entry %d from VF %d\n",
+ cfg->ind_tbl[idx], cfg->vf_id);
+ qset = cfg->vf_id;
+ }
+ } else {
qset = cfg->vf_id;
+ }
nic_reg_write(nic, NIC_PF_RSSI_0_4097_RQ | (rssi << 3),
(qset << 3) | (cfg->ind_tbl[idx] & 0x7));
idx++;
@@ -550,7 +559,19 @@ static void nic_send_pnicvf(struct nicpf *nic, int sqs)
static void nic_send_snicvf(struct nicpf *nic, struct nicvf_ptr *nicvf)
{
union nic_mbx mbx = {};
- int sqs_id = nic->vf_sqs[nicvf->vf_id][nicvf->sqs_id];
+ int sqs_id;
+
+ if (nicvf->sqs_id >= MAX_SQS_PER_VF) {
+ nic_mbx_send_nack(nic, nicvf->vf_id);
+ return;
+ }
+
+ sqs_id = nic->vf_sqs[nicvf->vf_id][nicvf->sqs_id];
+ if ((sqs_id < nic->num_vf_en) ||
+ (nic->pqs_vf[sqs_id] != nicvf->vf_id)) {
+ nic_mbx_send_nack(nic, nicvf->vf_id);
+ return;
+ }
mbx.nicvf.msg = NIC_MBOX_MSG_SNICVF_PTR;
mbx.nicvf.sqs_id = nicvf->sqs_id;
@@ -558,47 +579,150 @@ static void nic_send_snicvf(struct nicpf *nic, struct nicvf_ptr *nicvf)
nic_send_msg_to_vf(nic, nicvf->vf_id, &mbx);
}
+/* Find and take reference to all vf devices */
+static void nic_get_vf_pdev(struct nicpf *nic, int vf_en)
+{
+ struct pci_dev *pdev = nic->pdev;
+ struct pci_dev *vfdev;
+ u16 vid = pdev->vendor;
+ u16 devid;
+ int vf = 0, pos;
+
+ pos = pci_find_ext_capability(pdev, PCI_EXT_CAP_ID_SRIOV);
+ if (!pos)
+ return;
+ pci_read_config_word(pdev, pos + PCI_SRIOV_VF_DID, &devid);
+
+ vfdev = pci_get_device(vid, devid, NULL);
+ for (; vfdev; vfdev = pci_get_device(vid, devid, vfdev)) {
+ if (!vfdev->is_virtfn)
+ continue;
+ if (vfdev->physfn != pdev)
+ continue;
+ if (vf >= vf_en)
+ continue;
+ nic->vf_pdev[vf] = vfdev;
+ pci_dev_get(vfdev);
+ ++vf;
+ }
+}
+
+/* Release references to all vf devices */
+static void nic_put_vf_pdev(struct nicpf *nic)
+{
+ int vf;
+
+ for (vf = 0; vf < MAX_NUM_VFS_SUPPORTED; vf++) {
+ struct pci_dev *vfdev = nic->vf_pdev[vf];
+
+ nic->vf_pdev[vf] = NULL;
+ if (vfdev)
+ pci_dev_put(vfdev);
+ }
+}
+
+/* Check if pri.VF and sec.VF are in same domain i.e bound to same driver */
+static bool nic_check_svf_drv(struct nicpf *nic, u8 pvf, u8 svf)
+{
+ return pci_dev_driver(nic->vf_pdev[pvf]) ==
+ pci_dev_driver(nic->vf_pdev[svf]);
+}
+
/* Find next available Qset that can be assigned as a
* secondary Qset to a VF.
*/
-static int nic_nxt_avail_sqs(struct nicpf *nic)
+static int nic_nxt_avail_sqs(struct nicpf *nic, u8 pvf)
{
int sqs;
- for (sqs = 0; sqs < nic->num_sqs_en; sqs++) {
- if (!nic->sqs_used[sqs])
- nic->sqs_used[sqs] = true;
+ for (sqs = nic->num_vf_en;
+ sqs < (nic->num_vf_en + nic->num_sqs_en); sqs++) {
+ if ((nic->pqs_vf[sqs] == NIC_VF_UNASSIGNED) &&
+ nic_check_svf_drv(nic, pvf, sqs))
+ nic->pqs_vf[sqs] = pvf;
else
continue;
- return sqs + nic->num_vf_en;
+ return sqs;
}
return -1;
}
/* Allocate additional Qsets for requested VF */
-static void nic_alloc_sqs(struct nicpf *nic, struct sqs_alloc *sqs)
+static void nic_alloc_sqs(struct nicpf *nic, u8 pvf, struct sqs_alloc *sqs)
{
union nic_mbx mbx = {};
int idx, alloc_qs = 0;
int sqs_id;
- if (!nic->num_sqs_en)
+ if (!nic->num_sqs_en || (sqs->qs_count > MAX_SQS_PER_VF))
goto send_mbox;
- for (idx = 0; idx < sqs->qs_count; idx++) {
- sqs_id = nic_nxt_avail_sqs(nic);
- if (sqs_id < 0)
- break;
- nic->vf_sqs[sqs->vf_id][idx] = sqs_id;
- nic->pqs_vf[sqs_id] = sqs->vf_id;
- alloc_qs++;
+ if (sqs->spec) {
+ for (idx = 0; idx < sqs->qs_count; idx++) {
+ sqs_id = sqs->svf[idx];
+
+ /* Check if desired SQS is within the allowed range */
+ if (!((sqs_id >= nic->num_vf_en) &&
+ (sqs_id < (nic->num_vf_en + nic->num_sqs_en)))) {
+ dev_err(&nic->pdev->dev,
+ "Req SQS is invalid sqs->svf[%d]=%u",
+ idx, sqs_id);
+ break;
+ }
+
+ /* Check if desired SQS is free or assigned to a PVF */
+ if ((nic->pqs_vf[sqs_id] != NIC_VF_UNASSIGNED) &&
+ (nic->pqs_vf[sqs_id] != pvf)) {
+ dev_err(&nic->pdev->dev,
+ "SQS%d is already allocated to VF%u",
+ sqs_id, nic->pqs_vf[sqs_id]);
+ break;
+ }
+
+ /* Check if SQS is bound to the same driver as PVF */
+ if (!nic_check_svf_drv(nic, pvf, sqs_id)) {
+ dev_err(&nic->pdev->dev,
+ "SQS%d use different driver", sqs_id);
+ break;
+ }
+ }
+
+ if (idx != sqs->qs_count)
+ goto send_mbox;
+
+ /* Clear any existing assignments */
+ for (idx = 0; idx < MAX_SQS_PER_VF; idx++)
+ nic->vf_sqs[pvf][idx] = NIC_VF_UNASSIGNED;
+ for (idx = nic->num_vf_en;
+ idx < (nic->num_vf_en + nic->num_sqs_en); idx++) {
+ if (nic->pqs_vf[idx] == pvf)
+ nic->pqs_vf[idx] = NIC_VF_UNASSIGNED;
+ }
+
+ /* Populate VF's SQS table */
+ for (idx = 0; idx < sqs->qs_count; idx++) {
+ sqs_id = sqs->svf[idx];
+ nic->vf_sqs[pvf][idx] = sqs_id;
+ nic->pqs_vf[sqs_id] = pvf;
+ mbx.sqs_alloc.svf[idx] = sqs_id;
+ }
+ alloc_qs = idx;
+ } else {
+ for (idx = 0; idx < sqs->qs_count; idx++) {
+ sqs_id = nic_nxt_avail_sqs(nic, pvf);
+ if (sqs_id < 0)
+ break;
+ nic->vf_sqs[pvf][idx] = sqs_id;
+ nic->pqs_vf[sqs_id] = pvf;
+ mbx.sqs_alloc.svf[idx] = sqs_id;
+ }
+ alloc_qs = idx;
}
send_mbox:
mbx.sqs_alloc.msg = NIC_MBOX_MSG_ALLOC_SQS;
- mbx.sqs_alloc.vf_id = sqs->vf_id;
mbx.sqs_alloc.qs_count = alloc_qs;
- nic_send_msg_to_vf(nic, sqs->vf_id, &mbx);
+ nic_send_msg_to_vf(nic, pvf, &mbx);
}
static int nic_config_loopback(struct nicpf *nic, struct set_loopback *lbk)
@@ -776,13 +900,15 @@ static void nic_handle_mbx_intr(struct nicpf *nic, int vf)
goto unlock;
case NIC_MBOX_MSG_SHUTDOWN:
/* First msg in VF teardown sequence */
- if (vf >= nic->num_vf_en)
- nic->sqs_used[vf - nic->num_vf_en] = false;
- nic->pqs_vf[vf] = 0;
+ if (vf < nic->num_vf_en) {
+ for (i = 0; i < MAX_SQS_PER_VF; i++)
+ nic->vf_sqs[vf][i] = NIC_VF_UNASSIGNED;
+ }
+ nic->pqs_vf[vf] = NIC_VF_UNASSIGNED;
nic_enable_vf(nic, vf, false);
break;
case NIC_MBOX_MSG_ALLOC_SQS:
- nic_alloc_sqs(nic, &mbx.sqs_alloc);
+ nic_alloc_sqs(nic, vf, &mbx.sqs_alloc);
goto unlock;
case NIC_MBOX_MSG_NICVF_PTR:
nic->nicvf[vf] = mbx.nicvf.nicvf;
@@ -979,6 +1105,8 @@ static int nic_sriov_init(struct pci_dev *pdev, struct nicpf *nic)
return err;
}
+ nic_get_vf_pdev(nic, vf_en);
+
dev_info(&pdev->dev, "SRIOV enabled, number of VF available %d\n",
vf_en);
@@ -1035,7 +1163,7 @@ static int nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
{
struct device *dev = &pdev->dev;
struct nicpf *nic;
- int err;
+ int err, vf, sqs;
BUILD_BUG_ON(sizeof(union nic_mbx) > 16);
@@ -1090,6 +1218,13 @@ static int nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
/* Set RSS TBL size for each VF */
nic->rss_ind_tbl_size = NIC_MAX_RSS_IDR_TBL_SIZE;
+ /* Initialize all VF's primary Qset */
+ for (vf = 0; vf < MAX_NUM_VFS_SUPPORTED; vf++) {
+ nic->pqs_vf[vf] = NIC_VF_UNASSIGNED;
+ for (sqs = 0; sqs < MAX_SQS_PER_VF; sqs++)
+ nic->vf_sqs[vf][sqs] = NIC_VF_UNASSIGNED;
+ }
+
/* Register interrupts */
err = nic_register_interrupts(nic);
if (err)
@@ -1114,8 +1249,10 @@ static int nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
return 0;
err_disable_sriov:
- if (nic->flags & NIC_SRIOV_ENABLED)
+ if (nic->flags & NIC_SRIOV_ENABLED) {
+ nic_put_vf_pdev(nic);
pci_disable_sriov(pdev);
+ }
err_unregister_interrupts:
nic_unregister_interrupts(nic);
err_release_regions:
@@ -1130,8 +1267,10 @@ static void nic_remove(struct pci_dev *pdev)
{
struct nicpf *nic = pci_get_drvdata(pdev);
- if (nic->flags & NIC_SRIOV_ENABLED)
+ if (nic->flags & NIC_SRIOV_ENABLED) {
+ nic_put_vf_pdev(nic);
pci_disable_sriov(pdev);
+ }
if (nic->check_link) {
/* Destroy work Queue */
diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_main.c b/drivers/net/ethernet/cavium/thunder/nicvf_main.c
index bfee298..87d0f56 100644
--- a/drivers/net/ethernet/cavium/thunder/nicvf_main.c
+++ b/drivers/net/ethernet/cavium/thunder/nicvf_main.c
@@ -386,7 +386,7 @@ static void nicvf_request_sqs(struct nicvf *nic)
return;
mbx.sqs_alloc.msg = NIC_MBOX_MSG_ALLOC_SQS;
- mbx.sqs_alloc.vf_id = nic->vf_id;
+ mbx.sqs_alloc.spec = 0; /* Let PF choose which SQS to alloc */
mbx.sqs_alloc.qs_count = nic->sqs_count;
if (nicvf_send_msg_to_pf(nic, &mbx)) {
/* No response from PF */
--
1.7.1
^ permalink raw reply related
* [PATCH 3/4] net: thunderx: add sysfs attribute for SQS/SVF assigments
From: sunil.kovvuri @ 2016-04-19 13:33 UTC (permalink / raw)
To: netdev
Cc: linux-kernel, linux-arm-kernel, sgoutham, robert.richter,
Radoslaw Biernacki
In-Reply-To: <1461072812-44239-1-git-send-email-sunil.kovvuri@gmail.com>
From: Radoslaw Biernacki <rad@semihalf.com>
With this sysfs attribute (sriov_sqs_assignment) administrator will be
able to read the current assigment of SQS/SVF for a given VF. This is
useful to decide which VFs needs to be attached to UIO for a successful
allocation of secondary Qsets
Signed-off-by: Radoslaw Biernacki <rad@semihalf.com>
Signed-off-by: Sunil Goutham <sgoutham@cavium.com>
---
drivers/net/ethernet/cavium/thunder/nic_main.c | 67 +++++++++++++++++++++++-
1 files changed, 66 insertions(+), 1 deletions(-)
diff --git a/drivers/net/ethernet/cavium/thunder/nic_main.c b/drivers/net/ethernet/cavium/thunder/nic_main.c
index d6a6914..e2d8db9 100644
--- a/drivers/net/ethernet/cavium/thunder/nic_main.c
+++ b/drivers/net/ethernet/cavium/thunder/nic_main.c
@@ -1159,6 +1159,60 @@ static void nic_poll_for_link(struct work_struct *work)
queue_delayed_work(nic->check_link, &nic->dwork, HZ * 2);
}
+ssize_t sriov_sqs_assignment_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct pci_dev *pdev = container_of(dev, struct pci_dev, dev);
+ struct pci_dev *vf_dev;
+ struct pci_driver *vf_drv;
+ struct nicpf *nic = pci_get_drvdata(pdev);
+ size_t vf, off, svf_idx;
+
+ off = scnprintf(buf, PAGE_SIZE, "%u\n", nic->num_vf_en);
+
+ for (vf = 0; vf < nic->num_vf_en; vf++) {
+ vf_dev = nic->vf_pdev[vf];
+ vf_drv = vf_dev ? pci_dev_driver(vf_dev) : NULL;
+ if (off >= PAGE_SIZE)
+ break;
+ off += scnprintf(&buf[off], PAGE_SIZE - off,
+ "%zu %04x:%02x:%02x.%d %s %c:",
+ vf, pci_domain_nr(vf_dev->bus),
+ vf_dev->bus->number, PCI_SLOT(vf_dev->devfn),
+ PCI_FUNC(vf_dev->devfn),
+ vf_drv ? vf_drv->name : "no-driver",
+ nic->vf_enabled[vf] ? '+' : '-');
+ for (svf_idx = 0; svf_idx < MAX_SQS_PER_VF; svf_idx++) {
+ if (off >= PAGE_SIZE)
+ break;
+ if (nic->vf_sqs[vf][svf_idx] == NIC_VF_UNASSIGNED)
+ break;
+ off += scnprintf(&buf[off], PAGE_SIZE - off, " %d",
+ nic->vf_sqs[vf][svf_idx]);
+ }
+ if (off >= PAGE_SIZE)
+ break;
+ off += scnprintf(&buf[off], PAGE_SIZE - off, "\n");
+ }
+
+ for (vf = nic->num_vf_en; vf < nic->num_vf_en + nic->num_sqs_en; vf++) {
+ vf_dev = nic->vf_pdev[vf];
+ vf_drv = vf_dev ? pci_dev_driver(vf_dev) : NULL;
+ if (off >= PAGE_SIZE)
+ break;
+ off += scnprintf(&buf[off], PAGE_SIZE - off,
+ "%zu %04x:%02x:%02x.%d %s: %u\n",
+ vf, pci_domain_nr(vf_dev->bus),
+ vf_dev->bus->number, PCI_SLOT(vf_dev->devfn),
+ PCI_FUNC(vf_dev->devfn),
+ vf_drv ? vf_drv->name : "no-driver",
+ nic->pqs_vf[vf]);
+ }
+
+ return off;
+}
+DEVICE_ATTR_RO(sriov_sqs_assignment);
+
static int nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
{
struct device *dev = &pdev->dev;
@@ -1235,12 +1289,18 @@ static int nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
if (err)
goto err_unregister_interrupts;
+ err = device_create_file(dev, &dev_attr_sriov_sqs_assignment);
+ if (err) {
+ err = -ENOMEM;
+ goto err_disable_sriov;
+ }
+
/* Register a physical link status poll fn() */
nic->check_link = alloc_workqueue("check_link_status",
WQ_UNBOUND | WQ_MEM_RECLAIM, 1);
if (!nic->check_link) {
err = -ENOMEM;
- goto err_disable_sriov;
+ goto err_remove_sysfs_attr;
}
INIT_DELAYED_WORK(&nic->dwork, nic_poll_for_link);
@@ -1248,6 +1308,8 @@ static int nic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
return 0;
+err_remove_sysfs_attr:
+ device_remove_file(dev, &dev_attr_sriov_sqs_assignment);
err_disable_sriov:
if (nic->flags & NIC_SRIOV_ENABLED) {
nic_put_vf_pdev(nic);
@@ -1266,6 +1328,9 @@ err_disable_device:
static void nic_remove(struct pci_dev *pdev)
{
struct nicpf *nic = pci_get_drvdata(pdev);
+ struct device *dev = &pdev->dev;
+
+ device_remove_file(dev, &dev_attr_sriov_sqs_assignment);
if (nic->flags & NIC_SRIOV_ENABLED) {
nic_put_vf_pdev(nic);
--
1.7.1
^ permalink raw reply related
* [PATCH 4/4] net: thunderx: Improvement for MBX interface debug messages
From: sunil.kovvuri @ 2016-04-19 13:33 UTC (permalink / raw)
To: netdev
Cc: linux-kernel, linux-arm-kernel, sgoutham, robert.richter,
Radoslaw Biernacki
In-Reply-To: <1461072812-44239-1-git-send-email-sunil.kovvuri@gmail.com>
From: Radoslaw Biernacki <rad@semihalf.com>
Adding debug messages in case of NACK for a mailbox message, also
did small cleanups.
Signed-off-by: Radoslaw Biernacki <rad@semihalf.com>
Signed-off-by: Sunil Goutham <sgoutham@cavium.com>
---
drivers/net/ethernet/cavium/thunder/nic_main.c | 16 ++++++++++------
drivers/net/ethernet/cavium/thunder/nicvf_main.c | 8 ++++++--
2 files changed, 16 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/cavium/thunder/nic_main.c b/drivers/net/ethernet/cavium/thunder/nic_main.c
index e2d8db9..550a950 100644
--- a/drivers/net/ethernet/cavium/thunder/nic_main.c
+++ b/drivers/net/ethernet/cavium/thunder/nic_main.c
@@ -820,7 +820,7 @@ static void nic_handle_mbx_intr(struct nicpf *nic, int vf)
mbx_addr += sizeof(u64);
}
- dev_dbg(&nic->pdev->dev, "%s: Mailbox msg %d from VF%d\n",
+ dev_dbg(&nic->pdev->dev, "%s: Mailbox msg 0x%02x from VF%d\n",
__func__, mbx.msg.msg, vf);
switch (mbx.msg.msg) {
case NIC_MBOX_MSG_READY:
@@ -830,8 +830,7 @@ static void nic_handle_mbx_intr(struct nicpf *nic, int vf)
nic->duplex[vf] = 0;
nic->speed[vf] = 0;
}
- ret = 1;
- break;
+ goto unlock;
case NIC_MBOX_MSG_QS_CFG:
reg_addr = NIC_PF_QSET_0_127_CFG |
(mbx.qs.num << NIC_QS_ID_SHIFT);
@@ -873,8 +872,10 @@ static void nic_handle_mbx_intr(struct nicpf *nic, int vf)
nic_tx_channel_cfg(nic, mbx.qs.num, &mbx.sq);
break;
case NIC_MBOX_MSG_SET_MAC:
- if (vf >= nic->num_vf_en)
+ if (vf >= nic->num_vf_en) {
+ ret = -1; /* NACK */
break;
+ }
lmac = mbx.mac.vf_id;
bgx = NIC_GET_BGX_FROM_VF_LMAC_MAP(nic->vf_lmac_map[lmac]);
lmac = NIC_GET_LMAC_FROM_VF_LMAC_MAP(nic->vf_lmac_map[lmac]);
@@ -934,10 +935,13 @@ static void nic_handle_mbx_intr(struct nicpf *nic, int vf)
break;
}
- if (!ret)
+ if (!ret) {
nic_mbx_send_ack(nic, vf);
- else if (mbx.msg.msg != NIC_MBOX_MSG_READY)
+ } else if (mbx.msg.msg != NIC_MBOX_MSG_READY) {
+ dev_err(&nic->pdev->dev, "NACK for MBOX 0x%02x from VF %d\n",
+ mbx.msg.msg, vf);
nic_mbx_send_nack(nic, vf);
+ }
unlock:
nic->mbx_lock[vf] = false;
}
diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_main.c b/drivers/net/ethernet/cavium/thunder/nicvf_main.c
index 87d0f56..12ea73a 100644
--- a/drivers/net/ethernet/cavium/thunder/nicvf_main.c
+++ b/drivers/net/ethernet/cavium/thunder/nicvf_main.c
@@ -134,15 +134,19 @@ int nicvf_send_msg_to_pf(struct nicvf *nic, union nic_mbx *mbx)
/* Wait for previous message to be acked, timeout 2sec */
while (!nic->pf_acked) {
- if (nic->pf_nacked)
+ if (nic->pf_nacked) {
+ netdev_err(nic->netdev,
+ "PF NACK to mbox msg 0x%02x from VF%d\n",
+ (mbx->msg.msg & 0xFF), nic->vf_id);
return -EINVAL;
+ }
msleep(sleep);
if (nic->pf_acked)
break;
timeout -= sleep;
if (!timeout) {
netdev_err(nic->netdev,
- "PF didn't ack to mbox msg %d from VF%d\n",
+ "PF didn't ACK to mbox msg 0x%02x from VF%d\n",
(mbx->msg.msg & 0xFF), nic->vf_id);
return -EBUSY;
}
--
1.7.1
^ permalink raw reply related
* [RFC PATCH net-next 1/8] net: core: trivial netif_receive_skb_list() entry point
From: Edward Cree @ 2016-04-19 13:34 UTC (permalink / raw)
To: netdev, David Miller; +Cc: Jesper Dangaard Brouer, linux-net-drivers
In-Reply-To: <5716338E.4050003@solarflare.com>
Just calls netif_receive_skb() in a loop.
Signed-off-by: Edward Cree <ecree@solarflare.com>
---
include/linux/netdevice.h | 1 +
net/core/dev.c | 20 ++++++++++++++++++++
2 files changed, 21 insertions(+)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index a3bb534..682d0ad 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -3218,6 +3218,7 @@ static inline void dev_consume_skb_any(struct sk_buff *skb)
int netif_rx(struct sk_buff *skb);
int netif_rx_ni(struct sk_buff *skb);
int netif_receive_skb(struct sk_buff *skb);
+void netif_receive_skb_list(struct sk_buff_head *list);
gro_result_t napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb);
void napi_gro_flush(struct napi_struct *napi, bool flush_old);
struct sk_buff *napi_get_frags(struct napi_struct *napi);
diff --git a/net/core/dev.c b/net/core/dev.c
index 52d446b..7abcb1d 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -4292,6 +4292,26 @@ int netif_receive_skb(struct sk_buff *skb)
}
EXPORT_SYMBOL(netif_receive_skb);
+/**
+ * netif_receive_skb_list - process many receive buffers from network
+ * @list: list of skbs to process. Must not be shareable (e.g. it may
+ * be on the stack)
+ *
+ * For now, just calls netif_receive_skb() in a loop, ignoring the
+ * return value.
+ *
+ * This function may only be called from softirq context and interrupts
+ * should be enabled.
+ */
+void netif_receive_skb_list(struct sk_buff_head *list)
+{
+ struct sk_buff *skb;
+
+ while ((skb = __skb_dequeue(list)) != NULL)
+ netif_receive_skb(skb);
+}
+EXPORT_SYMBOL(netif_receive_skb_list);
+
/* Network device is going away, flush any packets still pending
* Called with irqs disabled.
*/
^ permalink raw reply related
* [RFC PATCH net-next 2/8] sfc: batch up RX delivery on EF10
From: Edward Cree @ 2016-04-19 13:35 UTC (permalink / raw)
To: netdev, David Miller; +Cc: Jesper Dangaard Brouer, linux-net-drivers
In-Reply-To: <5716338E.4050003@solarflare.com>
Improves packet rate of 1-byte UDP receives by 10%.
Signed-off-by: Edward Cree <ecree@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 9 +++++++++
drivers/net/ethernet/sfc/efx.c | 2 ++
drivers/net/ethernet/sfc/net_driver.h | 3 +++
drivers/net/ethernet/sfc/rx.c | 7 ++++++-
4 files changed, 20 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 98d33d4..e348f8f 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -2656,6 +2656,7 @@ static int efx_ef10_ev_process(struct efx_channel *channel, int quota)
{
struct efx_nic *efx = channel->efx;
efx_qword_t event, *p_event;
+ struct sk_buff_head rx_list;
unsigned int read_ptr;
int ev_code;
int tx_descs = 0;
@@ -2664,6 +2665,11 @@ static int efx_ef10_ev_process(struct efx_channel *channel, int quota)
if (quota <= 0)
return spent;
+ /* Prepare the batch receive list */
+ EFX_BUG_ON_PARANOID(channel->rx_list != NULL);
+ channel->rx_list = &rx_list;
+ __skb_queue_head_init(channel->rx_list);
+
read_ptr = channel->eventq_read_ptr;
for (;;) {
@@ -2724,6 +2730,9 @@ static int efx_ef10_ev_process(struct efx_channel *channel, int quota)
}
out:
+ /* Receive any packets we queued up */
+ netif_receive_skb_list(channel->rx_list);
+ channel->rx_list = NULL;
channel->eventq_read_ptr = read_ptr;
return spent;
}
diff --git a/drivers/net/ethernet/sfc/efx.c b/drivers/net/ethernet/sfc/efx.c
index 0705ec86..e004c0b 100644
--- a/drivers/net/ethernet/sfc/efx.c
+++ b/drivers/net/ethernet/sfc/efx.c
@@ -527,6 +527,8 @@ static int efx_probe_channel(struct efx_channel *channel)
goto fail;
}
+ channel->rx_list = NULL;
+
return 0;
fail:
diff --git a/drivers/net/ethernet/sfc/net_driver.h b/drivers/net/ethernet/sfc/net_driver.h
index 38c4223..d969c85 100644
--- a/drivers/net/ethernet/sfc/net_driver.h
+++ b/drivers/net/ethernet/sfc/net_driver.h
@@ -418,6 +418,7 @@ enum efx_sync_events_state {
* __efx_rx_packet(), or zero if there is none
* @rx_pkt_index: Ring index of first buffer for next packet to be delivered
* by __efx_rx_packet(), if @rx_pkt_n_frags != 0
+ * @rx_list: list of SKBs from current RX, awaiting processing
* @rx_queue: RX queue for this channel
* @tx_queue: TX queues for this channel
* @sync_events_state: Current state of sync events on this channel
@@ -462,6 +463,8 @@ struct efx_channel {
unsigned int rx_pkt_n_frags;
unsigned int rx_pkt_index;
+ struct sk_buff_head *rx_list;
+
struct efx_rx_queue rx_queue;
struct efx_tx_queue tx_queue[EFX_TXQ_TYPES];
diff --git a/drivers/net/ethernet/sfc/rx.c b/drivers/net/ethernet/sfc/rx.c
index 8956995..025e387 100644
--- a/drivers/net/ethernet/sfc/rx.c
+++ b/drivers/net/ethernet/sfc/rx.c
@@ -642,7 +642,12 @@ static void efx_rx_deliver(struct efx_channel *channel, u8 *eh,
return;
/* Pass the packet up */
- netif_receive_skb(skb);
+ if (channel->rx_list != NULL)
+ /* Add to list, will pass up later */
+ __skb_queue_tail(channel->rx_list, skb);
+ else
+ /* No list, so pass it up now */
+ netif_receive_skb(skb);
}
/* Handle a received packet. Second half: Touches packet payload. */
^ permalink raw reply related
* [RFC PATCH net-next 3/8] net: core: unwrap skb list receive slightly further
From: Edward Cree @ 2016-04-19 13:35 UTC (permalink / raw)
To: netdev, David Miller; +Cc: Jesper Dangaard Brouer, linux-net-drivers
In-Reply-To: <5716338E.4050003@solarflare.com>
Adds iterator skb_queue_for_each() to run over a list without modifying it.
Signed-off-by: Edward Cree <ecree@solarflare.com>
---
include/linux/skbuff.h | 16 ++++++++++++++++
include/trace/events/net.h | 7 +++++++
net/core/dev.c | 4 +++-
3 files changed, 26 insertions(+), 1 deletion(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index da0ace3..2851c38 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -1501,6 +1501,22 @@ static inline struct sk_buff *skb_peek_tail(const struct sk_buff_head *list_)
}
/**
+ * skb_queue_for_each - iterate over an skb queue
+ * @pos: the &struct sk_buff to use as a loop cursor.
+ * @head: the &struct sk_buff_head for your list.
+ *
+ * The reference count is not incremented and the reference is therefore
+ * volatile; the list lock is not taken either. Use with caution.
+ *
+ * The list must not be modified (though the individual skbs can be)
+ * within the loop body.
+ *
+ * After loop completion, @pos will be %NULL.
+ */
+#define skb_queue_for_each(pos, head) \
+ for (pos = skb_peek(head); pos != NULL; pos = skb_peek_next(pos, head))
+
+/**
* skb_queue_len - get queue length
* @list_: list to measure
*
diff --git a/include/trace/events/net.h b/include/trace/events/net.h
index 49cc7c3..30f359c 100644
--- a/include/trace/events/net.h
+++ b/include/trace/events/net.h
@@ -222,6 +222,13 @@ DEFINE_EVENT(net_dev_rx_verbose_template, netif_receive_skb_entry,
TP_ARGS(skb)
);
+DEFINE_EVENT(net_dev_rx_verbose_template, netif_receive_skb_list_entry,
+
+ TP_PROTO(const struct sk_buff *skb),
+
+ TP_ARGS(skb)
+);
+
DEFINE_EVENT(net_dev_rx_verbose_template, netif_rx_entry,
TP_PROTO(const struct sk_buff *skb),
diff --git a/net/core/dev.c b/net/core/dev.c
index 7abcb1d..4bb6724 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -4307,8 +4307,10 @@ void netif_receive_skb_list(struct sk_buff_head *list)
{
struct sk_buff *skb;
+ skb_queue_for_each(skb, list)
+ trace_netif_receive_skb_list_entry(skb);
while ((skb = __skb_dequeue(list)) != NULL)
- netif_receive_skb(skb);
+ netif_receive_skb_internal(skb);
}
EXPORT_SYMBOL(netif_receive_skb_list);
^ permalink raw reply related
* [RFC PATCH net-next 4/8] net: core: Another step of skb receive list processing
From: Edward Cree @ 2016-04-19 13:35 UTC (permalink / raw)
To: netdev, David Miller; +Cc: Jesper Dangaard Brouer, linux-net-drivers
In-Reply-To: <5716338E.4050003@solarflare.com>
netif_receive_skb_list_internal() now processes a list and hands it
on to the next function.
The code duplication is unfortunate, but the common part between the list
and non-list versions of the function takes a lock (rcu_read_lock()), so
factoring it out would be a little ugly.
Signed-off-by: Edward Cree <ecree@solarflare.com>
---
net/core/dev.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++----
1 file changed, 46 insertions(+), 4 deletions(-)
diff --git a/net/core/dev.c b/net/core/dev.c
index 4bb6724..586807d 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -4241,6 +4241,14 @@ static int __netif_receive_skb(struct sk_buff *skb)
return ret;
}
+static void __netif_receive_skb_list(struct sk_buff_head *list)
+{
+ struct sk_buff *skb;
+
+ while ((skb = __skb_dequeue(list)) != NULL)
+ __netif_receive_skb(skb);
+}
+
static int netif_receive_skb_internal(struct sk_buff *skb)
{
int ret;
@@ -4269,6 +4277,41 @@ static int netif_receive_skb_internal(struct sk_buff *skb)
return ret;
}
+static void netif_receive_skb_list_internal(struct sk_buff_head *list)
+{
+ struct sk_buff_head sublist;
+ struct sk_buff *skb;
+
+ __skb_queue_head_init(&sublist);
+
+ rcu_read_lock();
+ while ((skb = __skb_dequeue(list)) != NULL) {
+ net_timestamp_check(netdev_tstamp_prequeue, skb);
+ if (skb_defer_rx_timestamp(skb)) {
+ /* Handled, don't add to sublist */
+ continue;
+ }
+
+#ifdef CONFIG_RPS
+ if (static_key_false(&rps_needed)) {
+ struct rps_dev_flow voidflow, *rflow = &voidflow;
+ int cpu = get_rps_cpu(skb->dev, skb, &rflow);
+
+ if (cpu >= 0) {
+ enqueue_to_backlog(skb, cpu, &rflow->last_qtail);
+ /* Handled, don't add to sublist */
+ continue;
+ }
+ }
+#endif
+ __skb_queue_tail(&sublist, skb);
+ }
+
+ __netif_receive_skb_list(&sublist);
+ rcu_read_unlock();
+ return;
+}
+
/**
* netif_receive_skb - process receive buffer from network
* @skb: buffer to process
@@ -4297,8 +4340,8 @@ EXPORT_SYMBOL(netif_receive_skb);
* @list: list of skbs to process. Must not be shareable (e.g. it may
* be on the stack)
*
- * For now, just calls netif_receive_skb() in a loop, ignoring the
- * return value.
+ * Since return value of netif_receive_skb() is normally ignored, and
+ * wouldn't be meaningful for a list, this function returns void.
*
* This function may only be called from softirq context and interrupts
* should be enabled.
@@ -4309,8 +4352,7 @@ void netif_receive_skb_list(struct sk_buff_head *list)
skb_queue_for_each(skb, list)
trace_netif_receive_skb_list_entry(skb);
- while ((skb = __skb_dequeue(list)) != NULL)
- netif_receive_skb_internal(skb);
+ netif_receive_skb_list_internal(list);
}
EXPORT_SYMBOL(netif_receive_skb_list);
^ permalink raw reply related
* [RFC PATCH net-next 5/8] net: core: another layer of lists, around PF_MEMALLOC skb handling
From: Edward Cree @ 2016-04-19 13:36 UTC (permalink / raw)
To: netdev, David Miller; +Cc: Jesper Dangaard Brouer, linux-net-drivers
In-Reply-To: <5716338E.4050003@solarflare.com>
First example of a layer splitting the list (rather than merely taking
individual packets off it).
Again, trying to factor the common parts wouldn't make this any nicer.
Signed-off-by: Edward Cree <ecree@solarflare.com>
---
net/core/dev.c | 36 ++++++++++++++++++++++++++++++++++--
1 file changed, 34 insertions(+), 2 deletions(-)
diff --git a/net/core/dev.c b/net/core/dev.c
index 586807d..0f914bf 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -4216,6 +4216,14 @@ out:
return ret;
}
+static void __netif_receive_skb_list_core(struct sk_buff_head *list, bool pfmemalloc)
+{
+ struct sk_buff *skb;
+
+ while ((skb = __skb_dequeue(list)) != NULL)
+ __netif_receive_skb_core(skb, pfmemalloc);
+}
+
static int __netif_receive_skb(struct sk_buff *skb)
{
int ret;
@@ -4243,10 +4251,34 @@ static int __netif_receive_skb(struct sk_buff *skb)
static void __netif_receive_skb_list(struct sk_buff_head *list)
{
+ struct sk_buff_head sublist;
+ bool pfmemalloc = false; /* Is current sublist PF_MEMALLOC? */
+ unsigned long pflags;
struct sk_buff *skb;
- while ((skb = __skb_dequeue(list)) != NULL)
- __netif_receive_skb(skb);
+ __skb_queue_head_init(&sublist);
+
+ while ((skb = __skb_dequeue(list)) != NULL) {
+ if ((sk_memalloc_socks() && skb_pfmemalloc(skb)) != pfmemalloc) {
+ /* Handle the previous sublist */
+ __netif_receive_skb_list_core(&sublist, pfmemalloc);
+ pfmemalloc = !pfmemalloc;
+ /* See comments in __netif_receive_skb */
+ if (pfmemalloc) {
+ pflags = current->flags;
+ current->flags |= PF_MEMALLOC;
+ } else {
+ tsk_restore_flags(current, pflags, PF_MEMALLOC);
+ }
+ __skb_queue_head_init(&sublist);
+ }
+ __skb_queue_tail(&sublist, skb);
+ }
+ /* Handle the last sublist */
+ __netif_receive_skb_list_core(&sublist, pfmemalloc);
+ /* Restore pflags */
+ if (pfmemalloc)
+ tsk_restore_flags(current, pflags, PF_MEMALLOC);
}
static int netif_receive_skb_internal(struct sk_buff *skb)
^ permalink raw reply related
* [RFC PATCH net-next 6/8] net: core: propagate SKB lists through packet_type lookup
From: Edward Cree @ 2016-04-19 13:36 UTC (permalink / raw)
To: netdev, David Miller; +Cc: Jesper Dangaard Brouer, linux-net-drivers
In-Reply-To: <5716338E.4050003@solarflare.com>
This could maybe be made more efficient if we first split the list based on
skb->protocol, and then did ptype lookup for each sublist. Unfortunately,
there are things liks sch_handle_ingress and the rx_handlers that can
produce different results per packet.
Signed-off-by: Edward Cree <ecree@solarflare.com>
---
include/trace/events/net.h | 7 +++
net/core/dev.c | 146 ++++++++++++++++++++++++++++++++-------------
2 files changed, 113 insertions(+), 40 deletions(-)
diff --git a/include/trace/events/net.h b/include/trace/events/net.h
index 30f359c..7a17a31 100644
--- a/include/trace/events/net.h
+++ b/include/trace/events/net.h
@@ -130,6 +130,13 @@ DEFINE_EVENT(net_dev_template, netif_receive_skb,
TP_ARGS(skb)
);
+DEFINE_EVENT(net_dev_template, netif_receive_skb_list,
+
+ TP_PROTO(struct sk_buff *skb),
+
+ TP_ARGS(skb)
+);
+
DEFINE_EVENT(net_dev_template, netif_rx,
TP_PROTO(struct sk_buff *skb),
diff --git a/net/core/dev.c b/net/core/dev.c
index 0f914bf..db1d16a 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -4061,12 +4061,13 @@ static inline int nf_ingress(struct sk_buff *skb, struct packet_type **pt_prev,
return 0;
}
-static int __netif_receive_skb_core(struct sk_buff *skb, bool pfmemalloc)
+static int __netif_receive_skb_taps(struct sk_buff *skb, bool pfmemalloc,
+ struct packet_type **pt_prev)
{
- struct packet_type *ptype, *pt_prev;
rx_handler_func_t *rx_handler;
struct net_device *orig_dev;
bool deliver_exact = false;
+ struct packet_type *ptype;
int ret = NET_RX_DROP;
__be16 type;
@@ -4081,7 +4082,7 @@ static int __netif_receive_skb_core(struct sk_buff *skb, bool pfmemalloc)
skb_reset_transport_header(skb);
skb_reset_mac_len(skb);
- pt_prev = NULL;
+ *pt_prev = NULL;
another_round:
skb->skb_iif = skb->dev->ifindex;
@@ -4106,25 +4107,25 @@ another_round:
goto skip_taps;
list_for_each_entry_rcu(ptype, &ptype_all, list) {
- if (pt_prev)
- ret = deliver_skb(skb, pt_prev, orig_dev);
- pt_prev = ptype;
+ if (*pt_prev)
+ ret = deliver_skb(skb, *pt_prev, orig_dev);
+ *pt_prev = ptype;
}
list_for_each_entry_rcu(ptype, &skb->dev->ptype_all, list) {
- if (pt_prev)
- ret = deliver_skb(skb, pt_prev, orig_dev);
- pt_prev = ptype;
+ if (*pt_prev)
+ ret = deliver_skb(skb, *pt_prev, orig_dev);
+ *pt_prev = ptype;
}
skip_taps:
#ifdef CONFIG_NET_INGRESS
if (static_key_false(&ingress_needed)) {
- skb = sch_handle_ingress(skb, &pt_prev, &ret, orig_dev);
+ skb = sch_handle_ingress(skb, pt_prev, &ret, orig_dev);
if (!skb)
goto out;
- if (nf_ingress(skb, &pt_prev, &ret, orig_dev) < 0)
+ if (nf_ingress(skb, pt_prev, &ret, orig_dev) < 0)
goto out;
}
#endif
@@ -4136,9 +4137,9 @@ ncls:
goto drop;
if (skb_vlan_tag_present(skb)) {
- if (pt_prev) {
- ret = deliver_skb(skb, pt_prev, orig_dev);
- pt_prev = NULL;
+ if (*pt_prev) {
+ ret = deliver_skb(skb, *pt_prev, orig_dev);
+ *pt_prev = NULL;
}
if (vlan_do_receive(&skb))
goto another_round;
@@ -4148,9 +4149,9 @@ ncls:
rx_handler = rcu_dereference(skb->dev->rx_handler);
if (rx_handler) {
- if (pt_prev) {
- ret = deliver_skb(skb, pt_prev, orig_dev);
- pt_prev = NULL;
+ if (*pt_prev) {
+ ret = deliver_skb(skb, *pt_prev, orig_dev);
+ *pt_prev = NULL;
}
switch (rx_handler(&skb)) {
case RX_HANDLER_CONSUMED:
@@ -4181,47 +4182,112 @@ ncls:
/* deliver only exact match when indicated */
if (likely(!deliver_exact)) {
- deliver_ptype_list_skb(skb, &pt_prev, orig_dev, type,
+ deliver_ptype_list_skb(skb, pt_prev, orig_dev, type,
&ptype_base[ntohs(type) &
PTYPE_HASH_MASK]);
}
- deliver_ptype_list_skb(skb, &pt_prev, orig_dev, type,
+ deliver_ptype_list_skb(skb, pt_prev, orig_dev, type,
&orig_dev->ptype_specific);
if (unlikely(skb->dev != orig_dev)) {
- deliver_ptype_list_skb(skb, &pt_prev, orig_dev, type,
+ deliver_ptype_list_skb(skb, pt_prev, orig_dev, type,
&skb->dev->ptype_specific);
}
-
- if (pt_prev) {
- if (unlikely(skb_orphan_frags(skb, GFP_ATOMIC)))
- goto drop;
- else
- ret = pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
- } else {
+ if (*pt_prev && unlikely(skb_orphan_frags(skb, GFP_ATOMIC)))
+ goto drop;
+ return ret;
drop:
- if (!deliver_exact)
- atomic_long_inc(&skb->dev->rx_dropped);
- else
- atomic_long_inc(&skb->dev->rx_nohandler);
- kfree_skb(skb);
- /* Jamal, now you will not able to escape explaining
- * me how you were going to use this. :-)
- */
- ret = NET_RX_DROP;
- }
-
+ if (!deliver_exact)
+ atomic_long_inc(&skb->dev->rx_dropped);
+ else
+ atomic_long_inc(&skb->dev->rx_nohandler);
+ kfree_skb(skb);
+ /* Jamal, now you will not able to escape explaining
+ * me how you were going to use this. :-)
+ */
+ ret = NET_RX_DROP;
out:
+ *pt_prev = NULL;
return ret;
}
-static void __netif_receive_skb_list_core(struct sk_buff_head *list, bool pfmemalloc)
+static int __netif_receive_skb_core(struct sk_buff *skb, bool pfmemalloc)
+{
+ struct net_device *orig_dev = skb->dev;
+ struct packet_type *pt_prev;
+ int ret;
+
+ ret = __netif_receive_skb_taps(skb, pfmemalloc, &pt_prev);
+ if (pt_prev)
+ ret = pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
+ return ret;
+}
+
+static inline void __netif_receive_skb_list_ptype(struct sk_buff_head *list,
+ struct packet_type *pt_prev,
+ struct net_device *orig_dev)
{
struct sk_buff *skb;
while ((skb = __skb_dequeue(list)) != NULL)
- __netif_receive_skb_core(skb, pfmemalloc);
+ pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
+}
+
+static void __netif_receive_skb_list_core(struct sk_buff_head *list, bool pfmemalloc)
+{
+ /* Fast-path assumptions:
+ * - There is no RX handler.
+ * - Only one packet_type matches.
+ * If either of these fails, we will end up doing some per-packet
+ * processing in-line, then handling the 'last ptype' for the whole
+ * sublist. This can't cause out-of-order delivery to any single ptype,
+ * because the 'last ptype' must be constant across the sublist, and all
+ * other ptypes are handled per-packet. Unless, that is, a ptype can
+ * be delivered to more than once for a single packet - but that seems
+ * like it would be a bad idea anyway.
+ * So it should be fine (at least, I think so), but you'll lose the
+ * (putative) performance benefits of batching.
+ */
+ /* Current (common) ptype of sublist */
+ struct packet_type *pt_curr = NULL;
+ /* In the normal (device RX) case, orig_dev should be the same for
+ * every skb in the list. But as I'm not certain of this, I check
+ * it's constant and split the list if not.
+ * So, od_curr is the current (common) orig_dev of sublist.
+ */
+ struct net_device *od_curr = NULL;
+ struct sk_buff_head sublist;
+ struct sk_buff *skb;
+
+ __skb_queue_head_init(&sublist);
+
+ while ((skb = __skb_dequeue(list)) != NULL) {
+ struct packet_type *pt_prev;
+ struct net_device *orig_dev = skb->dev;
+
+ __netif_receive_skb_taps(skb, pfmemalloc, &pt_prev);
+ if (pt_prev) {
+ if (skb_queue_empty(&sublist)) {
+ pt_curr = pt_prev;
+ od_curr = orig_dev;
+ } else if (!(pt_curr == pt_prev &&
+ od_curr == orig_dev)) {
+ /* dispatch old sublist */
+ __netif_receive_skb_list_ptype(&sublist,
+ pt_curr,
+ od_curr);
+ /* start new sublist */
+ __skb_queue_head_init(&sublist);
+ pt_curr = pt_prev;
+ od_curr = orig_dev;
+ }
+ __skb_queue_tail(&sublist, skb);
+ }
+ }
+
+ /* dispatch final sublist */
+ __netif_receive_skb_list_ptype(&sublist, pt_curr, od_curr);
}
static int __netif_receive_skb(struct sk_buff *skb)
^ permalink raw reply related
* [RFC PATCH net-next 7/8] net: ipv4: listified version of ip_rcv
From: Edward Cree @ 2016-04-19 13:37 UTC (permalink / raw)
To: netdev, David Miller; +Cc: Jesper Dangaard Brouer, linux-net-drivers
In-Reply-To: <5716338E.4050003@solarflare.com>
Also involved adding a way to run a netfilter hook over a list of packets.
Rather than attempting to make netfilter know about lists (which would be
horrendous) we just let it call the regular okfn (in this case
ip_rcv_finish()) for any packets it steals, and have it give us back a list
of packets it's synchronously accepted (which normally NF_HOOK would
automatically call okfn() on, but we want to be able to potentially pass
the list to a listified version of okfn().)
There is potential for out-of-order receives if the netfilter hook ends up
synchronously stealing packets, as they will be processed before any accepts
earlier in the list. However, it was already possible for an asynchronous
accept to cause out-of-order receives, so hopefully I haven't broken
anything that wasn't broken already.
Signed-off-by: Edward Cree <ecree@solarflare.com>
---
include/linux/netdevice.h | 3 ++
include/linux/netfilter.h | 27 +++++++++++++++++
include/net/ip.h | 2 ++
net/core/dev.c | 11 +++++--
net/ipv4/af_inet.c | 1 +
net/ipv4/ip_input.c | 75 ++++++++++++++++++++++++++++++++++++++++++-----
6 files changed, 110 insertions(+), 9 deletions(-)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 682d0ad..292f2d5 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -2143,6 +2143,9 @@ struct packet_type {
struct net_device *,
struct packet_type *,
struct net_device *);
+ void (*list_func) (struct sk_buff_head *,
+ struct packet_type *,
+ struct net_device *);
bool (*id_match)(struct packet_type *ptype,
struct sock *sk);
void *af_packet_priv;
diff --git a/include/linux/netfilter.h b/include/linux/netfilter.h
index 9230f9a..e18e91b 100644
--- a/include/linux/netfilter.h
+++ b/include/linux/netfilter.h
@@ -220,6 +220,24 @@ NF_HOOK_THRESH(uint8_t pf, unsigned int hook, struct net *net, struct sock *sk,
return ret;
}
+static inline void
+NF_HOOK_LIST_THRESH(uint8_t pf, unsigned int hook, struct net *net, struct sock *sk,
+ struct sk_buff_head *list, struct sk_buff_head *sublist,
+ struct net_device *in, struct net_device *out,
+ int (*okfn)(struct net *, struct sock *, struct sk_buff *),
+ int thresh)
+{
+ struct sk_buff *skb;
+
+ __skb_queue_head_init(sublist); /* list of synchronously ACCEPTed skbs */
+ while ((skb = __skb_dequeue(list)) != NULL) {
+ int ret = nf_hook_thresh(pf, hook, net, sk, skb, in, out, okfn,
+ thresh);
+ if (ret == 1)
+ __skb_queue_tail(sublist, skb);
+ }
+}
+
static inline int
NF_HOOK_COND(uint8_t pf, unsigned int hook, struct net *net, struct sock *sk,
struct sk_buff *skb, struct net_device *in, struct net_device *out,
@@ -242,6 +260,15 @@ NF_HOOK(uint8_t pf, unsigned int hook, struct net *net, struct sock *sk, struct
return NF_HOOK_THRESH(pf, hook, net, sk, skb, in, out, okfn, INT_MIN);
}
+static inline void
+NF_HOOK_LIST(uint8_t pf, unsigned int hook, struct net *net, struct sock *sk,
+ struct sk_buff_head *list, struct sk_buff_head *sublist,
+ struct net_device *in, struct net_device *out,
+ int (*okfn)(struct net *, struct sock *, struct sk_buff *))
+{
+ NF_HOOK_LIST_THRESH(pf, hook, net, sk, list, sublist, in, out, okfn, INT_MIN);
+}
+
/* Call setsockopt() */
int nf_setsockopt(struct sock *sk, u_int8_t pf, int optval, char __user *opt,
unsigned int len);
diff --git a/include/net/ip.h b/include/net/ip.h
index 93725e5..c994c44 100644
--- a/include/net/ip.h
+++ b/include/net/ip.h
@@ -106,6 +106,8 @@ int ip_build_and_send_pkt(struct sk_buff *skb, const struct sock *sk,
struct ip_options_rcu *opt);
int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
struct net_device *orig_dev);
+void ip_list_rcv(struct sk_buff_head *list, struct packet_type *pt,
+ struct net_device *orig_dev);
int ip_local_deliver(struct sk_buff *skb);
int ip_mr_input(struct sk_buff *skb);
int ip_output(struct net *net, struct sock *sk, struct sk_buff *skb);
diff --git a/net/core/dev.c b/net/core/dev.c
index db1d16a..da768e2 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -4230,8 +4230,15 @@ static inline void __netif_receive_skb_list_ptype(struct sk_buff_head *list,
{
struct sk_buff *skb;
- while ((skb = __skb_dequeue(list)) != NULL)
- pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
+ if (!pt_prev)
+ return;
+ if (skb_queue_empty(list))
+ return;
+ if (pt_prev->list_func != NULL)
+ pt_prev->list_func(list, pt_prev, orig_dev);
+ else
+ while ((skb = __skb_dequeue(list)) != NULL)
+ pt_prev->func(skb, skb->dev, pt_prev, orig_dev);
}
static void __netif_receive_skb_list_core(struct sk_buff_head *list, bool pfmemalloc)
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 2e6e65f..1424147 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -1757,6 +1757,7 @@ fs_initcall(ipv4_offload_init);
static struct packet_type ip_packet_type __read_mostly = {
.type = cpu_to_be16(ETH_P_IP),
.func = ip_rcv,
+ .list_func = ip_list_rcv,
};
static int __init inet_init(void)
diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c
index e3d7827..e7d0d85 100644
--- a/net/ipv4/ip_input.c
+++ b/net/ipv4/ip_input.c
@@ -395,10 +395,9 @@ drop:
/*
* Main IP Receive routine.
*/
-int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, struct net_device *orig_dev)
+static struct sk_buff *ip_rcv_core(struct sk_buff *skb, struct net *net)
{
const struct iphdr *iph;
- struct net *net;
u32 len;
/* When the interface is in promisc. mode, drop all the crap
@@ -408,7 +407,6 @@ int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
goto drop;
- net = dev_net(dev);
IP_UPD_PO_STATS_BH(net, IPSTATS_MIB_IN, skb->len);
skb = skb_share_check(skb, GFP_ATOMIC);
@@ -475,9 +473,7 @@ int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
/* Must drop socket now because of tproxy. */
skb_orphan(skb);
- return NF_HOOK(NFPROTO_IPV4, NF_INET_PRE_ROUTING,
- net, NULL, skb, dev, NULL,
- ip_rcv_finish);
+ return skb;
csum_error:
IP_INC_STATS_BH(net, IPSTATS_MIB_CSUMERRORS);
@@ -486,5 +482,70 @@ inhdr_error:
drop:
kfree_skb(skb);
out:
- return NET_RX_DROP;
+ return NULL;
+}
+
+/*
+ * IP receive entry point
+ */
+int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
+ struct net_device *orig_dev)
+{
+ struct net *net = dev_net(dev);
+
+ skb = ip_rcv_core(skb, net);
+ if (skb == NULL)
+ return NET_RX_DROP;
+ return NF_HOOK(NFPROTO_IPV4, NF_INET_PRE_ROUTING,
+ net, NULL, skb, dev, NULL,
+ ip_rcv_finish);
+}
+
+static void ip_sublist_rcv(struct sk_buff_head *list, struct net_device *dev,
+ struct net *net)
+{
+ struct sk_buff_head sublist;
+ struct sk_buff *skb;
+
+ NF_HOOK_LIST(NFPROTO_IPV4, NF_INET_PRE_ROUTING, net, NULL,
+ list, &sublist, dev, NULL, ip_rcv_finish);
+ while ((skb = __skb_dequeue(&sublist)) != NULL)
+ ip_rcv_finish(net, NULL, skb);
+}
+
+/* Receive a list of IP packets */
+void ip_list_rcv(struct sk_buff_head *list, struct packet_type *pt,
+ struct net_device *orig_dev)
+{
+ struct net_device *curr_dev = NULL;
+ struct net *curr_net = NULL;
+ struct sk_buff_head sublist;
+ struct sk_buff *skb;
+
+ __skb_queue_head_init(&sublist);
+
+ while ((skb = __skb_dequeue(list)) != NULL) {
+ struct net_device *dev = skb->dev;
+ struct net *net = dev_net(dev);
+
+ skb = ip_rcv_core(skb, net);
+ if (skb == NULL)
+ continue;
+
+ if (skb_queue_empty(&sublist)) {
+ curr_dev = dev;
+ curr_net = net;
+ } else if (curr_dev != dev || curr_net != net) {
+ /* dispatch old sublist */
+ ip_sublist_rcv(&sublist, dev, net);
+ /* start new sublist */
+ __skb_queue_head_init(&sublist);
+ curr_dev = dev;
+ curr_net = net;
+ }
+ /* add to current sublist */
+ __skb_queue_tail(&sublist, skb);
+ }
+ /* dispatch final sublist */
+ ip_sublist_rcv(&sublist, curr_dev, curr_net);
}
^ permalink raw reply related
* [RFC PATCH net-next 8/8] net: ipv4: listify ip_rcv_finish
From: Edward Cree @ 2016-04-19 13:37 UTC (permalink / raw)
To: netdev, David Miller; +Cc: Jesper Dangaard Brouer, linux-net-drivers
In-Reply-To: <5716338E.4050003@solarflare.com>
Signed-off-by: Edward Cree <ecree@solarflare.com>
---
net/ipv4/ip_input.c | 58 ++++++++++++++++++++++++++++++++++++++++++++++++-----
1 file changed, 53 insertions(+), 5 deletions(-)
diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c
index e7d0d85..5bbc409 100644
--- a/net/ipv4/ip_input.c
+++ b/net/ipv4/ip_input.c
@@ -308,7 +308,8 @@ drop:
return true;
}
-static int ip_rcv_finish(struct net *net, struct sock *sk, struct sk_buff *skb)
+static int ip_rcv_finish_core(struct net *net, struct sock *sk,
+ struct sk_buff *skb)
{
const struct iphdr *iph = ip_hdr(skb);
struct rtable *rt;
@@ -385,13 +386,22 @@ static int ip_rcv_finish(struct net *net, struct sock *sk, struct sk_buff *skb)
goto drop;
}
- return dst_input(skb);
+ return NET_RX_SUCCESS;
drop:
kfree_skb(skb);
return NET_RX_DROP;
}
+static int ip_rcv_finish(struct net *net, struct sock *sk, struct sk_buff *skb)
+{
+ int ret = ip_rcv_finish_core(net, sk, skb);
+
+ if (ret != NET_RX_DROP)
+ ret = dst_input(skb);
+ return ret;
+}
+
/*
* Main IP Receive routine.
*/
@@ -501,16 +511,54 @@ int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
ip_rcv_finish);
}
+static void ip_sublist_rcv_finish(struct sk_buff_head *list)
+{
+ struct sk_buff *skb;
+
+ while ((skb = __skb_dequeue(list)) != NULL)
+ dst_input(skb);
+}
+
+static void ip_list_rcv_finish(struct net *net, struct sock *sk,
+ struct sk_buff_head *list)
+{
+ struct dst_entry *curr_dst = NULL;
+ struct sk_buff_head sublist;
+ struct sk_buff *skb;
+
+ __skb_queue_head_init(&sublist);
+
+ while ((skb = __skb_dequeue(list)) != NULL) {
+ struct dst_entry *dst;
+
+ if (ip_rcv_finish_core(net, sk, skb) == NET_RX_DROP)
+ continue;
+
+ dst = skb_dst(skb);
+ if (skb_queue_empty(&sublist)) {
+ curr_dst = dst;
+ } else if (curr_dst != dst) {
+ /* dispatch old sublist */
+ ip_sublist_rcv_finish(&sublist);
+ /* start new sublist */
+ __skb_queue_head_init(&sublist);
+ curr_dst = dst;
+ }
+ /* add to current sublist */
+ __skb_queue_tail(&sublist, skb);
+ }
+ /* dispatch final sublist */
+ ip_sublist_rcv_finish(&sublist);
+}
+
static void ip_sublist_rcv(struct sk_buff_head *list, struct net_device *dev,
struct net *net)
{
struct sk_buff_head sublist;
- struct sk_buff *skb;
NF_HOOK_LIST(NFPROTO_IPV4, NF_INET_PRE_ROUTING, net, NULL,
list, &sublist, dev, NULL, ip_rcv_finish);
- while ((skb = __skb_dequeue(&sublist)) != NULL)
- ip_rcv_finish(net, NULL, skb);
+ ip_list_rcv_finish(net, NULL, &sublist);
}
/* Receive a list of IP packets */
^ permalink raw reply related
* Re: [PATCH net] tcp: Fix SOF_TIMESTAMPING_TX_ACK when handling dup acks
From: Soheil Hassas Yeganeh @ 2016-04-19 13:54 UTC (permalink / raw)
To: Martin KaFai Lau
Cc: netdev, Kernel Team, Eric Dumazet, Neal Cardwell,
Soheil Hassas Yeganeh, Willem de Bruijn, Yuchung Cheng
In-Reply-To: <1461019193-3034571-1-git-send-email-kafai@fb.com>
On Mon, Apr 18, 2016 at 6:39 PM, Martin KaFai Lau <kafai@fb.com> wrote:
> Assuming SOF_TIMESTAMPING_TX_ACK is on. When dup acks are received,
> it could incorrectly think that a skb has already
> been acked and queue a SCM_TSTAMP_ACK cmsg to the
> sk->sk_error_queue.
>
> In tcp_ack_tstamp(), it checks
> 'between(shinfo->tskey, prior_snd_una, tcp_sk(sk)->snd_una - 1)'.
> If prior_snd_una == tcp_sk(sk)->snd_una like the following packetdrill
> script, between() returns true but the tskey is actually not acked.
> e.g. try between(3, 2, 1).
>
> The fix is to replace between() with one before() and one !before().
> By doing this, the -1 offset on the tcp_sk(sk)->snd_una can also be
> removed.
>
> A packetdrill script is used to reproduce the dup ack scenario.
> Due to the lacking cmsg support in packetdrill (may be I
> cannot find it), a BPF prog is used to kprobe to
> sock_queue_err_skb() and print out the value of
> serr->ee.ee_data.
>
> Both the packetdrill and the bcc BPF script is attached at the end of
> this commit message.
>
> BPF Output Before Fix:
> ~~~~~~
> <...>-2056 [001] d.s. 433.927987: : ee_data:1459 #incorrect
> packetdrill-2056 [001] d.s. 433.929563: : ee_data:1459 #incorrect
> packetdrill-2056 [001] d.s. 433.930765: : ee_data:1459 #incorrect
> packetdrill-2056 [001] d.s. 434.028177: : ee_data:1459
> packetdrill-2056 [001] d.s. 434.029686: : ee_data:14599
>
> BPF Output After Fix:
> ~~~~~~
> <...>-2049 [000] d.s. 113.517039: : ee_data:1459
> <...>-2049 [000] d.s. 113.517253: : ee_data:14599
>
> BCC BPF Script:
> ~~~~~~
> #!/usr/bin/env python
>
> from __future__ import print_function
> from bcc import BPF
>
> bpf_text = """
> #include <uapi/linux/ptrace.h>
> #include <net/sock.h>
> #include <bcc/proto.h>
> #include <linux/errqueue.h>
>
> #ifdef memset
> #undef memset
> #endif
>
> int trace_err_skb(struct pt_regs *ctx)
> {
> struct sk_buff *skb = (struct sk_buff *)ctx->si;
> struct sock *sk = (struct sock *)ctx->di;
> struct sock_exterr_skb *serr;
> u32 ee_data = 0;
>
> if (!sk || !skb)
> return 0;
>
> serr = SKB_EXT_ERR(skb);
> bpf_probe_read(&ee_data, sizeof(ee_data), &serr->ee.ee_data);
> bpf_trace_printk("ee_data:%u\\n", ee_data);
>
> return 0;
> };
> """
>
> b = BPF(text=bpf_text)
> b.attach_kprobe(event="sock_queue_err_skb", fn_name="trace_err_skb")
> print("Attached to kprobe")
> b.trace_print()
>
> Packetdrill Script:
> ~~~~~~
> +0 `sysctl -q -w net.ipv4.tcp_min_tso_segs=10`
> +0 `sysctl -q -w net.ipv4.tcp_no_metrics_save=1`
> +0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
> +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
> +0 bind(3, ..., ...) = 0
> +0 listen(3, 1) = 0
>
> 0.100 < S 0:0(0) win 32792 <mss 1460,sackOK,nop,nop,nop,wscale 7>
> 0.100 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 7>
> 0.200 < . 1:1(0) ack 1 win 257
> 0.200 accept(3, ..., ...) = 4
> +0 setsockopt(4, SOL_TCP, TCP_NODELAY, [1], 4) = 0
>
> +0 setsockopt(4, SOL_SOCKET, 37, [2688], 4) = 0
> 0.200 write(4, ..., 1460) = 1460
> 0.200 write(4, ..., 13140) = 13140
>
> 0.200 > P. 1:1461(1460) ack 1
> 0.200 > . 1461:8761(7300) ack 1
> 0.200 > P. 8761:14601(5840) ack 1
>
> 0.300 < . 1:1(0) ack 1 win 257 <sack 1461:2921,nop,nop>
> 0.300 < . 1:1(0) ack 1 win 257 <sack 1461:4381,nop,nop>
> 0.300 < . 1:1(0) ack 1 win 257 <sack 1461:5841,nop,nop>
> 0.300 > P. 1:1461(1460) ack 1
> 0.400 < . 1:1(0) ack 14601 win 257
>
> 0.400 close(4) = 0
> 0.400 > F. 14601:14601(0) ack 1
> 0.500 < F. 1:1(0) ack 14602 win 257
> 0.500 > . 14602:14602(0) ack 2
>
> Signed-off-by: Martin KaFai Lau <kafai@fb.com>
> Cc: Eric Dumazet <edumazet@google.com>
> Cc: Neal Cardwell <ncardwell@google.com>
> Cc: Soheil Hassas Yeganeh <soheil.kdev@gmail.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
> Cc: Willem de Bruijn <willemb@google.com>
> Cc: Yuchung Cheng <ycheng@google.com>
> ---
> net/ipv4/tcp_input.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
> index e6e65f7..0edb071 100644
> --- a/net/ipv4/tcp_input.c
> +++ b/net/ipv4/tcp_input.c
> @@ -3098,7 +3098,8 @@ static void tcp_ack_tstamp(struct sock *sk, struct sk_buff *skb,
>
> shinfo = skb_shinfo(skb);
> if ((shinfo->tx_flags & SKBTX_ACK_TSTAMP) &&
> - between(shinfo->tskey, prior_snd_una, tcp_sk(sk)->snd_una - 1))
> + !before(shinfo->tskey, prior_snd_una) &&
> + before(shinfo->tskey, tcp_sk(sk)->snd_una))
> __skb_tstamp_tx(skb, NULL, sk, SCM_TSTAMP_ACK);
> }
Nice catch! Thanks.
> --
> 2.5.1
>
^ permalink raw reply
* [PATCH v2 0/1] drivers: net: cpsw: Fix NULL pointer dereference with two slave PHYs
From: Andrew Goodbody @ 2016-04-19 13:56 UTC (permalink / raw)
To: netdev
Cc: linux-kernel, linux-omap, mugunthanvnm, grygorii.strashko, tony,
Andrew Goodbody
Resend to add more people on Cc: as requested by Grygrii Strashko.
This is a fix for a NULL pointer dereference from cpsw which is triggered
by having two slave PHYs attached to a cpsw network device. The problem is
due to only maintaining a single reference to a PHY node in the prive data
which gets overwritten by the second PHY probe. So move the PHY node
reference to the individual slave data so that there is now one per slave.
v1 had a problem that data->slaves was used before it had been filled in
Andrew Goodbody (1):
Prevent NUll pointer dereference with two PHYs on cpsw
drivers/net/ethernet/ti/cpsw.c | 30 +++++++++++++++---------------
1 file changed, 15 insertions(+), 15 deletions(-)
--
2.5.0
^ permalink raw reply
* [PATCH v2 1/1] drivers: net: cpsw: Prevent NUll pointer dereference with two PHYs
From: Andrew Goodbody @ 2016-04-19 13:56 UTC (permalink / raw)
To: netdev
Cc: linux-kernel, linux-omap, mugunthanvnm, grygorii.strashko, tony,
Andrew Goodbody
In-Reply-To: <1461074186-25535-1-git-send-email-andrew.goodbody@cambrionix.com>
Adding a 2nd PHY to cpsw results in a NULL pointer dereference
as below. Fix by maintaining a reference to each PHY node in slave
struct instead of a single reference in the priv struct which was
overwritten by the 2nd PHY.
[ 17.870933] Unable to handle kernel NULL pointer dereference at virtual address 00000180
[ 17.879557] pgd = dc8bc000
[ 17.882514] [00000180] *pgd=9c882831, *pte=00000000, *ppte=00000000
[ 17.889213] Internal error: Oops: 17 [#1] ARM
[ 17.893838] Modules linked in:
[ 17.897102] CPU: 0 PID: 1657 Comm: connmand Not tainted 4.5.0-ge463dfb-dirty #11
[ 17.904947] Hardware name: Cambrionix whippet
[ 17.909576] task: dc859240 ti: dc968000 task.ti: dc968000
[ 17.915339] PC is at phy_attached_print+0x18/0x8c
[ 17.920339] LR is at phy_attached_info+0x14/0x18
[ 17.925247] pc : [<c042baec>] lr : [<c042bb74>] psr: 600f0113
[ 17.925247] sp : dc969cf8 ip : dc969d28 fp : dc969d18
[ 17.937425] r10: dda7a400 r9 : 00000000 r8 : 00000000
[ 17.942971] r7 : 00000001 r6 : ddb00480 r5 : ddb8cb34 r4 : 00000000
[ 17.949898] r3 : c0954cc0 r2 : c09562b0 r1 : 00000000 r0 : 00000000
[ 17.956829] Flags: nZCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none
[ 17.964401] Control: 10c5387d Table: 9c8bc019 DAC: 00000051
[ 17.970500] Process connmand (pid: 1657, stack limit = 0xdc968210)
[ 17.977059] Stack: (0xdc969cf8 to 0xdc96a000)
[ 17.981692] 9ce0: dc969d28 dc969d08
[ 17.990386] 9d00: c038f9bc c038f6b4 ddb00480 dc969d34 dc969d28 c042bb74 c042bae4 00000000
[ 17.999080] 9d20: c09562b0 c0954cc0 dc969d5c dc969d38 c043ebfc c042bb6c 00000007 00000003
[ 18.007773] 9d40: ddb00000 ddb8cb58 ddb00480 00000001 dc969dec dc969d60 c0441614 c043ea68
[ 18.016465] 9d60: 00000000 00000003 00000000 fffffff4 dc969df4 0000000d 00000000 00000000
[ 18.025159] 9d80: dc969db4 dc969d90 c005dc08 c05839e0 dc969df4 0000000d ddb00000 00001002
[ 18.033851] 9da0: 00000000 00000000 dc969dcc dc969db8 c005ddf4 c005dbc8 00000000 00000118
[ 18.042544] 9dc0: dc969dec dc969dd0 ddb00000 c06db27c ffff9003 00001002 00000000 00000000
[ 18.051237] 9de0: dc969e0c dc969df0 c057c88c c04410dc dc969e0c ddb00000 ddb00000 00000001
[ 18.059930] 9e00: dc969e34 dc969e10 c057cb44 c057c7d8 ddb00000 ddb00138 00001002 beaeda20
[ 18.068622] 9e20: 00000000 00000000 dc969e5c dc969e38 c057cc28 c057cac0 00000000 dc969e80
[ 18.077315] 9e40: dda7a40c beaeda20 00000000 00000000 dc969ecc dc969e60 c05e36d0 c057cc14
[ 18.086007] 9e60: dc969e84 00000051 beaeda20 00000000 dda7a40c 00000014 ddb00000 00008914
[ 18.094699] 9e80: 30687465 00000000 00000000 00000000 00009003 00000000 00000000 00000000
[ 18.103391] 9ea0: 00001002 00008914 dd257ae0 beaeda20 c098a428 beaeda20 00000011 00000000
[ 18.112084] 9ec0: dc969edc dc969ed0 c05e4e54 c05e3030 dc969efc dc969ee0 c055f5ac c05e4cc4
[ 18.120777] 9ee0: beaeda20 dd257ae0 dc8ab4c0 00008914 dc969f7c dc969f00 c010b388 c055f45c
[ 18.129471] 9f00: c071ca40 dd257ac0 c00165e8 dc968000 dc969f3c dc969f20 dc969f64 dc969f28
[ 18.138164] 9f20: c0115708 c0683ec8 dd257ac0 dd257ac0 dc969f74 dc969f40 c055f350 c00fc66c
[ 18.146857] 9f40: dd82e4d0 00000011 00000000 00080000 dd257ac0 00000000 dc8ab4c0 dc8ab4c0
[ 18.155550] 9f60: 00008914 beaeda20 00000011 00000000 dc969fa4 dc969f80 c010bc34 c010b2fc
[ 18.164242] 9f80: 00000000 00000011 00000002 00000036 c00165e8 dc968000 00000000 dc969fa8
[ 18.172935] 9fa0: c00163e0 c010bbcc 00000000 00000011 00000011 00008914 beaeda20 00009003
[ 18.181628] 9fc0: 00000000 00000011 00000002 00000036 00081018 00000001 00000000 beaedc10
[ 18.190320] 9fe0: 00083188 beaeda1c 00043a5d b6d29c0c 600b0010 00000011 00000000 00000000
[ 18.198989] Backtrace:
[ 18.201621] [<c042bad8>] (phy_attached_print) from [<c042bb74>] (phy_attached_info+0x14/0x18)
[ 18.210664] r3:c0954cc0 r2:c09562b0 r1:00000000
[ 18.215588] r4:ddb00480
[ 18.218322] [<c042bb60>] (phy_attached_info) from [<c043ebfc>] (cpsw_slave_open+0x1a0/0x280)
[ 18.227293] [<c043ea5c>] (cpsw_slave_open) from [<c0441614>] (cpsw_ndo_open+0x544/0x674)
[ 18.235874] r7:00000001 r6:ddb00480 r5:ddb8cb58 r4:ddb00000
[ 18.241944] [<c04410d0>] (cpsw_ndo_open) from [<c057c88c>] (__dev_open+0xc0/0x128)
[ 18.249972] r9:00000000 r8:00000000 r7:00001002 r6:ffff9003 r5:c06db27c r4:ddb00000
[ 18.258255] [<c057c7cc>] (__dev_open) from [<c057cb44>] (__dev_change_flags+0x90/0x154)
[ 18.266745] r5:00000001 r4:ddb00000
[ 18.270575] [<c057cab4>] (__dev_change_flags) from [<c057cc28>] (dev_change_flags+0x20/0x50)
[ 18.279523] r9:00000000 r8:00000000 r7:beaeda20 r6:00001002 r5:ddb00138 r4:ddb00000
[ 18.287811] [<c057cc08>] (dev_change_flags) from [<c05e36d0>] (devinet_ioctl+0x6ac/0x76c)
[ 18.296483] r9:00000000 r8:00000000 r7:beaeda20 r6:dda7a40c r5:dc969e80 r4:00000000
[ 18.304762] [<c05e3024>] (devinet_ioctl) from [<c05e4e54>] (inet_ioctl+0x19c/0x1c8)
[ 18.312882] r10:00000000 r9:00000011 r8:beaeda20 r7:c098a428 r6:beaeda20 r5:dd257ae0
[ 18.321235] r4:00008914
[ 18.323956] [<c05e4cb8>] (inet_ioctl) from [<c055f5ac>] (sock_ioctl+0x15c/0x2d8)
[ 18.331829] [<c055f450>] (sock_ioctl) from [<c010b388>] (do_vfs_ioctl+0x98/0x8d0)
[ 18.339765] r7:00008914 r6:dc8ab4c0 r5:dd257ae0 r4:beaeda20
[ 18.345822] [<c010b2f0>] (do_vfs_ioctl) from [<c010bc34>] (SyS_ioctl+0x74/0x84)
[ 18.353573] r10:00000000 r9:00000011 r8:beaeda20 r7:00008914 r6:dc8ab4c0 r5:dc8ab4c0
[ 18.361924] r4:00000000
[ 18.364653] [<c010bbc0>] (SyS_ioctl) from [<c00163e0>] (ret_fast_syscall+0x0/0x3c)
[ 18.372682] r9:dc968000 r8:c00165e8 r7:00000036 r6:00000002 r5:00000011 r4:00000000
[ 18.380960] Code: e92dd810 e24cb010 e24dd010 e59b4004 (e5902180)
[ 18.387580] ---[ end trace c80529466223f3f3 ]---
Signed-off-by: Andrew Goodbody <andrew.goodbody@cambrionix.com>
---
v2 - Move allocation of memory for priv->slaves to inside cpsw_probe_dt so it
has data->slaves initialised first which is needed to calculate size
drivers/net/ethernet/ti/cpsw.c | 30 +++++++++++++++---------------
1 file changed, 15 insertions(+), 15 deletions(-)
diff --git a/drivers/net/ethernet/ti/cpsw.c b/drivers/net/ethernet/ti/cpsw.c
index 42fdfd4..e62909c 100644
--- a/drivers/net/ethernet/ti/cpsw.c
+++ b/drivers/net/ethernet/ti/cpsw.c
@@ -349,6 +349,7 @@ struct cpsw_slave {
struct cpsw_slave_data *data;
struct phy_device *phy;
struct net_device *ndev;
+ struct device_node *phy_node;
u32 port_vlan;
u32 open_stat;
};
@@ -367,7 +368,6 @@ struct cpsw_priv {
spinlock_t lock;
struct platform_device *pdev;
struct net_device *ndev;
- struct device_node *phy_node;
struct napi_struct napi_rx;
struct napi_struct napi_tx;
struct device *dev;
@@ -1148,8 +1148,8 @@ static void cpsw_slave_open(struct cpsw_slave *slave, struct cpsw_priv *priv)
cpsw_ale_add_mcast(priv->ale, priv->ndev->broadcast,
1 << slave_port, 0, 0, ALE_MCAST_FWD_2);
- if (priv->phy_node)
- slave->phy = of_phy_connect(priv->ndev, priv->phy_node,
+ if (slave->phy_node)
+ slave->phy = of_phy_connect(priv->ndev, slave->phy_node,
&cpsw_adjust_link, 0, slave->data->phy_if);
else
slave->phy = phy_connect(priv->ndev, slave->data->phy_id,
@@ -1946,7 +1946,7 @@ static int cpsw_probe_dt(struct cpsw_priv *priv,
struct device_node *node = pdev->dev.of_node;
struct device_node *slave_node;
struct cpsw_platform_data *data = &priv->data;
- int i = 0, ret;
+ int i, ret;
u32 prop;
if (!node)
@@ -1958,6 +1958,14 @@ static int cpsw_probe_dt(struct cpsw_priv *priv,
}
data->slaves = prop;
+ priv->slaves = devm_kzalloc(&pdev->dev,
+ sizeof(struct cpsw_slave) * data->slaves,
+ GFP_KERNEL);
+ if (!priv->slaves)
+ return -ENOMEM;
+ for (i = 0; i < data->slaves; i++)
+ priv->slaves[i].slave_num = i;
+
if (of_property_read_u32(node, "active_slave", &prop)) {
dev_err(&pdev->dev, "Missing active_slave property in the DT.\n");
return -EINVAL;
@@ -2023,6 +2031,7 @@ static int cpsw_probe_dt(struct cpsw_priv *priv,
if (ret)
dev_warn(&pdev->dev, "Doesn't have any child node\n");
+ i = 0;
for_each_child_of_node(node, slave_node) {
struct cpsw_slave_data *slave_data = data->slave_data + i;
const void *mac_addr = NULL;
@@ -2033,7 +2042,8 @@ static int cpsw_probe_dt(struct cpsw_priv *priv,
if (strcmp(slave_node->name, "slave"))
continue;
- priv->phy_node = of_parse_phandle(slave_node, "phy-handle", 0);
+ priv->slaves[i].phy_node =
+ of_parse_phandle(slave_node, "phy-handle", 0);
parp = of_get_property(slave_node, "phy_id", &lenp);
if (of_phy_is_fixed_link(slave_node)) {
struct device_node *phy_node;
@@ -2292,16 +2302,6 @@ static int cpsw_probe(struct platform_device *pdev)
memcpy(ndev->dev_addr, priv->mac_addr, ETH_ALEN);
- priv->slaves = devm_kzalloc(&pdev->dev,
- sizeof(struct cpsw_slave) * data->slaves,
- GFP_KERNEL);
- if (!priv->slaves) {
- ret = -ENOMEM;
- goto clean_runtime_disable_ret;
- }
- for (i = 0; i < data->slaves; i++)
- priv->slaves[i].slave_num = i;
-
priv->slaves[0].ndev = ndev;
priv->emac_port = 0;
--
2.5.0
^ permalink raw reply related
* Re: [PATCHv2] wlcore: spi: add wl18xx support
From: Arnd Bergmann @ 2016-04-19 14:21 UTC (permalink / raw)
To: Reizer, Eyal
Cc: Kalle Valo, Eyal Reizer, linux-wireless@vger.kernel.org,
netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
devicetree@vger.kernel.org, linux-spi@vger.kernel.org
In-Reply-To: <8665E2433BC68541A24DFFCA87B70F5B360C0745@DFRE01.ent.ti.com>
On Tuesday 19 April 2016 09:05:45 Reizer, Eyal wrote:
> > > It is also part of the generic spi.h (include/Linux/spi/spi.h),
> > > already part of " struct spi_device" So it seemed redundant adding
> > > another mechanism for implementing the same.
> > > Platform that interact with a wilink need to use it, and platforms
> > > that don't have this capability will probably not interact with a wilink device
> > using SPI.
> >
> > The cs_gpio field in spi_device belongs to the spi host controller, no other
> > slave driver uses it.
> >
> > I wasn't asking for a duplication of this mechanism, but an interface to use it
> > properly. Internally, the spi core uses the spi_set_cs() function to pick a CS.
> > Find a way to use that rather than reimplementing it incorrectly.
> >
>
> Understood. As this special CS manipulation is unique to wspi (wilink spi) I think the
> best option is to move this gpio allocation into wlcore_spi as a new device tree entry
> used only by this driver.
> If you agree I will submit a v3.
I don't think that can work either: aside of not solving the problem
of wilink devices on spi controllers that don't use gpio, it also doesn't
solve the problem of what happens when the driver manually triggers the
gpio to hold the CS signal while another driver talks to a different
device using another CS on the same controller.
Arnd
^ permalink raw reply
* [patch -next] geneve: testing the wrong variable in geneve6_build_skb()
From: Dan Carpenter @ 2016-04-19 14:30 UTC (permalink / raw)
To: David S. Miller, Alexander Duyck
Cc: Jesse Gross, John W. Linville, Pravin B Shelar, Jiri Benc,
Daniel Borkmann, Tom Herbert, netdev, kernel-janitors
We intended to test "err" and not "skb".
Fixes: aed069df099c ('ip_tunnel_core: iptunnel_handle_offloads returns int and doesn't free skb')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index efbc7ce..512dbe0 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -733,7 +733,7 @@ static int geneve6_build_skb(struct dst_entry *dst, struct sk_buff *skb,
goto free_dst;
err = udp_tunnel_handle_offloads(skb, udp_sum);
- if (IS_ERR(skb))
+ if (err)
goto free_dst;
gnvh = (struct genevehdr *)__skb_push(skb, sizeof(*gnvh) + opt_len);
^ permalink raw reply related
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