* [next-queue v4 PATCH 5/7] i40e: Add TX and RX support in switchdev mode.
From: Sridhar Samudrala @ 2017-01-12 22:35 UTC (permalink / raw)
To: alexander.h.duyck, john.r.fastabend, anjali.singhai,
jakub.kicinski, davem, scott.d.peterson, gerlitz.or, jiri,
intel-wired-lan, netdev
In-Reply-To: <1484260510-9162-1-git-send-email-sridhar.samudrala@intel.com>
In switchdev mode, broadcast filter is not enabled on VFs. The broadcasts
and unknown frames from VFs are received by the PF and passed to
corresponding VF port representator netdev.
A host based switching entity like a linux bridge or OVS redirects these
frames to the right VFs via VFPR netdevs. Any frames sent via VFPR netdevs
are sent as directed transmits to the corresponding VFs. To enable directed
transmit, skb metadata dst is used to pass the VF id and the frame is
requeued to call the PFs transmit routine.
Small script to demonstrate inter VF pings in switchdev mode.
PF: enp5s0f0, VFs: enp5s2,enp5s2f1 VFPRs:enp5s0f0-vf0, enp5s0f0-vf1
# rmmod i40e; modprobe i40e
# devlink dev eswitch set pci/0000:05:00.0 mode switchdev
# echo 2 > /sys/class/net/enp5s0f0/device/sriov_numvfs
# ip link set enp5s0f0 vf 0 mac 00:11:22:33:44:55
# ip link set enp5s0f0 vf 1 mac 00:11:22:33:44:56
# rmmod i40evf; modprobe i40evf
/* Create 2 namespaces and move the VFs to the corresponding ns. */
# ip netns add ns0
# ip link set enp5s2 netns ns0
# ip netns exec ns0 ip addr add 192.168.1.10/24 dev enp5s2
# ip netns exec ns0 ip link set enp5s2 up
# ip netns add ns1
# ip link set enp5s2f1 netns ns1
# ip netns exec ns1 ip addr add 192.168.1.11/24 dev enp5s2f1
# ip netns exec ns1 ip link set enp5s2f1 up
/* bring up pf and vfpr netdevs */
# ip link set enp5s0f0 up
# ip link set enp5s0f0-vf0 up
# ip link set enp5s0f0-vf1 up
/* Create a linux bridge and add vfpr netdevs to it. */
# ip link add vfpr-br type bridge
# ip link set enp5s0f0-vf0 master vfpr-br
# ip link set enp5s0f0-vf1 master vfpr-br
# ip addr add 192.168.1.1/24 dev vfpr-br
# ip link set vfpr-br up
# ip netns exec ns0 ping -c3 192.168.1.11
# ip netns exec ns1 ping -c3 192.168.1.10
Signed-off-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e.h | 1 +
drivers/net/ethernet/intel/i40e/i40e_main.c | 4 +
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 106 ++++++++++++++++++++-
drivers/net/ethernet/intel/i40e/i40e_txrx.h | 2 +
drivers/net/ethernet/intel/i40e/i40e_type.h | 3 +
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 17 +++-
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h | 1 +
7 files changed, 127 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 331a1e0..95cfe46 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -55,6 +55,7 @@
#include <linux/net_tstamp.h>
#include <linux/ptp_clock_kernel.h>
#include <net/devlink.h>
+#include <net/dst_metadata.h>
#include "i40e_type.h"
#include "i40e_prototype.h"
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 53a0618..9212656 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -11322,6 +11322,7 @@ static int i40e_devlink_eswitch_mode_get(struct devlink *devlink, u16 *mode)
static int i40e_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode)
{
struct i40e_pf *pf = devlink_priv(devlink);
+ struct i40e_vsi *vsi = pf->vsi[pf->lan_vsi];
struct i40e_vf *vf;
int i, j, err = 0;
@@ -11335,6 +11336,8 @@ static int i40e_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode)
i40e_free_vfpr_netdev(vf);
}
pf->eswitch_mode = mode;
+ vsi->netdev->priv_flags |=
+ (IFF_XMIT_DST_RELEASE | IFF_XMIT_DST_RELEASE_PERM);
break;
case DEVLINK_ESWITCH_MODE_SWITCHDEV:
for (i = 0; i < pf->num_alloc_vfs; i++) {
@@ -11349,6 +11352,7 @@ static int i40e_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode)
}
}
pf->eswitch_mode = mode;
+ netif_keep_dst(vsi->netdev);
break;
default:
err = -EOPNOTSUPP;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index 0291ed4..f43d1df 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -1283,16 +1283,39 @@ static bool i40e_alloc_mapped_page(struct i40e_ring *rx_ring,
* @rx_ring: rx ring in play
* @skb: packet to send up
* @vlan_tag: vlan tag for packet
+ * @lpbk: is it a loopback frame?
**/
static void i40e_receive_skb(struct i40e_ring *rx_ring,
- struct sk_buff *skb, u16 vlan_tag)
+ struct sk_buff *skb, u16 vlan_tag, bool lpbk)
{
struct i40e_q_vector *q_vector = rx_ring->q_vector;
+ struct i40e_pf *pf = rx_ring->vsi->back;
+ struct i40e_vf *vf;
+ struct ethhdr *eth;
+ int vf_id;
if ((rx_ring->netdev->features & NETIF_F_HW_VLAN_CTAG_RX) &&
(vlan_tag & VLAN_VID_MASK))
__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), vlan_tag);
+ if ((pf->eswitch_mode == DEVLINK_ESWITCH_MODE_LEGACY) || !lpbk)
+ goto gro_receive;
+
+ /* If a loopback packet is received from a VF in switchdev mode, pass
+ * the frame to the corresponding VFPR netdev based on the source MAC
+ * in the frame.
+ */
+ eth = (struct ethhdr *)skb_mac_header(skb);
+ for (vf_id = 0; vf_id < pf->num_alloc_vfs; vf_id++) {
+ vf = &pf->vf[vf_id];
+ if (ether_addr_equal(eth->h_source,
+ vf->default_lan_addr.addr)) {
+ skb->dev = vf->vfpr_netdev;
+ break;
+ }
+ }
+
+gro_receive:
napi_gro_receive(&q_vector->napi, skb);
}
@@ -1501,6 +1524,7 @@ static inline void i40e_rx_hash(struct i40e_ring *ring,
* @rx_desc: pointer to the EOP Rx descriptor
* @skb: pointer to current skb being populated
* @rx_ptype: the packet type decoded by hardware
+ * @lpbk: is it a loopback frame?
*
* This function checks the ring, descriptor, and packet information in
* order to populate the hash, checksum, VLAN, protocol, and
@@ -1509,7 +1533,7 @@ static inline void i40e_rx_hash(struct i40e_ring *ring,
static inline
void i40e_process_skb_fields(struct i40e_ring *rx_ring,
union i40e_rx_desc *rx_desc, struct sk_buff *skb,
- u8 rx_ptype)
+ u8 rx_ptype, bool *lpbk)
{
u64 qword = le64_to_cpu(rx_desc->wb.qword1.status_error_len);
u32 rx_status = (qword & I40E_RXD_QW1_STATUS_MASK) >>
@@ -1518,6 +1542,9 @@ void i40e_process_skb_fields(struct i40e_ring *rx_ring,
u32 tsyn = (rx_status & I40E_RXD_QW1_STATUS_TSYNINDX_MASK) >>
I40E_RXD_QW1_STATUS_TSYNINDX_SHIFT;
+ *lpbk = !!((rx_status & I40E_RXD_QW1_STATUS_LPBK_MASK) >>
+ I40E_RXD_QW1_STATUS_LPBK_SHIFT);
+
if (unlikely(tsynvalid))
i40e_ptp_rx_hwtstamp(rx_ring->vsi->back, skb, tsyn);
@@ -2045,6 +2072,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
u8 rx_ptype;
u64 qword;
unsigned int xdp_consumed_bytes = 0;
+ bool lpbk;
/* return some buffers to hardware, one at a time is too slow */
if (cleaned_count >= I40E_RX_BUFFER_WRITE) {
@@ -2113,7 +2141,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
I40E_RXD_QW1_PTYPE_SHIFT;
/* populate checksum, VLAN, and protocol */
- i40e_process_skb_fields(rx_ring, rx_desc, skb, rx_ptype);
+ i40e_process_skb_fields(rx_ring, rx_desc, skb, rx_ptype, &lpbk);
#ifdef I40E_FCOE
if (unlikely(
@@ -2127,7 +2155,7 @@ static int i40e_clean_rx_irq(struct i40e_ring *rx_ring, int budget)
vlan_tag = (qword & BIT(I40E_RX_DESC_STATUS_L2TAG1P_SHIFT)) ?
le16_to_cpu(rx_desc->wb.qword0.lo_dword.l2tag1) : 0;
- i40e_receive_skb(rx_ring, skb, vlan_tag);
+ i40e_receive_skb(rx_ring, skb, vlan_tag, lpbk);
skb = NULL;
/* update budget accounting */
@@ -2692,6 +2720,27 @@ static int i40e_tso(struct i40e_tx_buffer *first, u8 *hdr_len,
}
/**
+ * i40e_tvsi - set up the target vsi in TX context descriptor
+ * @tx_ring: ptr to the target vsi
+ * @cd_type_cmd_tso_mss: Quad Word 1
+ *
+ * Returns 0
+ **/
+static int i40e_tvsi(struct i40e_vsi *tvsi, u64 *cd_type_cmd_tso_mss)
+{
+ u64 cd_cmd, cd_tvsi;
+
+ cd_cmd = I40E_TX_CTX_DESC_SWTCH_VSI;
+ cd_tvsi = tvsi->id;
+ cd_tvsi = (cd_tvsi << I40E_TXD_CTX_QW1_VSI_SHIFT) &
+ I40E_TXD_CTX_QW1_VSI_MASK;
+ *cd_type_cmd_tso_mss |= (cd_cmd << I40E_TXD_CTX_QW1_CMD_SHIFT) |
+ cd_tvsi;
+
+ return 0;
+}
+
+/**
* i40e_tsyn - set up the tsyn context descriptor
* @tx_ring: ptr to the ring to send
* @skb: ptr to the skb we're sending
@@ -3223,8 +3272,12 @@ static netdev_tx_t i40e_xmit_frame_ring(struct sk_buff *skb,
struct i40e_ring *tx_ring)
{
u64 cd_type_cmd_tso_mss = I40E_TX_DESC_DTYPE_CONTEXT;
+ struct metadata_dst *md_dst = skb_metadata_dst(skb);
u32 cd_tunneling = 0, cd_l2tag2 = 0;
struct i40e_tx_buffer *first;
+ struct i40e_vsi *t_vsi = NULL;
+ struct i40e_vf *t_vf;
+ struct i40e_pf *pf;
u32 td_offset = 0;
u32 tx_flags = 0;
__be16 protocol;
@@ -3276,7 +3329,26 @@ static netdev_tx_t i40e_xmit_frame_ring(struct sk_buff *skb,
else if (protocol == htons(ETH_P_IPV6))
tx_flags |= I40E_TX_FLAGS_IPV6;
- tso = i40e_tso(first, &hdr_len, &cd_type_cmd_tso_mss);
+ /* If skb metadata dst points to a VF id, do a directed transmit to
+ * that VSI. TSO is mutually exclusive with this option. So TSO is not
+ * enabled when doing a directed transmit.
+ */
+ if (md_dst && (md_dst->type == METADATA_HW_PORT_MUX)) {
+ pf = tx_ring->vsi->back;
+ if (md_dst->u.port_info.port_id >= pf->num_alloc_vfs) {
+ WARN_ONCE(1, "Unexpected port_id: %d num_vfs:%d\n",
+ md_dst->u.port_info.port_id,
+ pf->num_alloc_vfs);
+ goto out_drop;
+ }
+ t_vf = &pf->vf[md_dst->u.port_info.port_id];
+ t_vsi = pf->vsi[t_vf->lan_vsi_idx];
+ }
+
+ if (t_vsi)
+ tso = i40e_tvsi(t_vsi, &cd_type_cmd_tso_mss);
+ else
+ tso = i40e_tso(first, &hdr_len, &cd_type_cmd_tso_mss);
if (tso < 0)
goto out_drop;
@@ -3340,3 +3412,27 @@ netdev_tx_t i40e_lan_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
return i40e_xmit_frame_ring(skb, tx_ring);
}
+
+/**
+ * i40e_vfpr_netdev_start_xmit
+ * @skb: send buffer
+ * @netdev: network interface device structure
+ *
+ * Sets skb->dev to PF netdev, VF id in the skb->dst and requeues
+ * skb via dev_queue_xmit()
+ **/
+netdev_tx_t i40e_vfpr_netdev_start_xmit(struct sk_buff *skb,
+ struct net_device *netdev)
+{
+ struct i40e_vfpr_netdev_priv *priv = netdev_priv(netdev);
+ struct i40e_vf *vf = priv->vf;
+ struct i40e_pf *pf = vf->pf;
+ struct i40e_vsi *vsi = pf->vsi[pf->lan_vsi];
+
+ skb_dst_drop(skb);
+ dst_hold(&priv->vfpr_dst->dst);
+ skb_dst_set(skb, &priv->vfpr_dst->dst);
+ skb->dev = vsi->netdev;
+
+ return dev_queue_xmit(skb);
+}
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.h b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
index 3250be7..5e24aef 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.h
@@ -393,6 +393,8 @@ struct i40e_ring_container {
bool i40e_alloc_rx_buffers(struct i40e_ring *rxr, u16 cleaned_count);
netdev_tx_t i40e_lan_xmit_frame(struct sk_buff *skb, struct net_device *netdev);
+netdev_tx_t i40e_vfpr_netdev_start_xmit(struct sk_buff *skb,
+ struct net_device *netdev);
void i40e_clean_tx_ring(struct i40e_ring *tx_ring);
void i40e_clean_rx_ring(struct i40e_ring *rx_ring);
int i40e_setup_tx_descriptors(struct i40e_ring *tx_ring);
diff --git a/drivers/net/ethernet/intel/i40e/i40e_type.h b/drivers/net/ethernet/intel/i40e/i40e_type.h
index 939f9fd..251f57e 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_type.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_type.h
@@ -729,6 +729,9 @@ enum i40e_rx_desc_status_bits {
#define I40E_RXD_QW1_STATUS_TSYNVALID_SHIFT I40E_RX_DESC_STATUS_TSYNVALID_SHIFT
#define I40E_RXD_QW1_STATUS_TSYNVALID_MASK \
BIT_ULL(I40E_RXD_QW1_STATUS_TSYNVALID_SHIFT)
+#define I40E_RXD_QW1_STATUS_LPBK_SHIFT I40E_RX_DESC_STATUS_LPBK_SHIFT
+#define I40E_RXD_QW1_STATUS_LPBK_MASK \
+ BIT_ULL(I40E_RXD_QW1_STATUS_LPBK_SHIFT)
enum i40e_rx_desc_fltstat_values {
I40E_RX_DESC_FLTSTAT_NO_DATA = 0,
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index 275113c..7211fba 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -1060,6 +1060,7 @@ static int i40e_vfpr_netdev_stop(struct net_device *dev)
static const struct net_device_ops i40e_vfpr_netdev_ops = {
.ndo_open = i40e_vfpr_netdev_open,
.ndo_stop = i40e_vfpr_netdev_stop,
+ .ndo_start_xmit = i40e_vfpr_netdev_start_xmit,
};
/**
@@ -1119,6 +1120,10 @@ int i40e_alloc_vfpr_netdev(struct i40e_vf *vf, u16 vf_num)
priv = netdev_priv(vfpr_netdev);
priv->vf = &pf->vf[vf_num];
+ priv->vfpr_dst = metadata_dst_alloc(0, METADATA_HW_PORT_MUX,
+ GFP_KERNEL);
+ priv->vfpr_dst->u.port_info.lower_dev = vsi->netdev;
+ priv->vfpr_dst->u.port_info.port_id = vf->vf_id;
vfpr_netdev->netdev_ops = &i40e_vfpr_netdev_ops;
eth_hw_addr_inherit(vfpr_netdev, vsi->netdev);
@@ -1130,6 +1135,7 @@ int i40e_alloc_vfpr_netdev(struct i40e_vf *vf, u16 vf_num)
if (err) {
dev_err(&pf->pdev->dev, "register_netdev failed for vf: %s\n",
vf->vfpr_netdev->name);
+ dst_release((struct dst_entry *)priv->vfpr_dst);
free_netdev(vfpr_netdev);
return err;
}
@@ -1158,6 +1164,7 @@ int i40e_alloc_vfpr_netdev(struct i40e_vf *vf, u16 vf_num)
**/
void i40e_free_vfpr_netdev(struct i40e_vf *vf)
{
+ struct i40e_vfpr_netdev_priv *priv;
struct i40e_pf *pf = vf->pf;
if (!vf->vfpr_netdev)
@@ -1166,6 +1173,8 @@ void i40e_free_vfpr_netdev(struct i40e_vf *vf)
dev_info(&pf->pdev->dev, "Freeing VF Port representor(%s)\n",
vf->vfpr_netdev->name);
+ priv = netdev_priv(vf->vfpr_netdev);
+ dst_release((struct dst_entry *)priv->vfpr_dst);
unregister_netdev(vf->vfpr_netdev);
free_netdev(vf->vfpr_netdev);
@@ -1940,8 +1949,10 @@ static int i40e_vc_enable_queues_msg(struct i40e_vf *vf, u8 *msg, u16 msglen)
if (i40e_vsi_start_rings(pf->vsi[vf->lan_vsi_idx]))
aq_ret = I40E_ERR_TIMEOUT;
- if ((aq_ret == 0) && vf->vfpr_netdev)
+ if ((aq_ret == 0) && vf->vfpr_netdev) {
+ netif_tx_start_all_queues(vf->vfpr_netdev);
netif_carrier_on(vf->vfpr_netdev);
+ }
error_param:
/* send the response to the VF */
@@ -1987,8 +1998,10 @@ static int i40e_vc_disable_queues_msg(struct i40e_vf *vf, u8 *msg, u16 msglen)
i40e_vsi_stop_rings(pf->vsi[vf->lan_vsi_idx]);
- if ((aq_ret == 0) && vf->vfpr_netdev)
+ if ((aq_ret == 0) && vf->vfpr_netdev) {
+ netif_tx_stop_all_queues(vf->vfpr_netdev);
netif_carrier_off(vf->vfpr_netdev);
+ }
error_param:
/* send the response to the VF */
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h
index 25ce93c..3dea207 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h
@@ -74,6 +74,7 @@ enum i40e_vf_capabilities {
/* VF Port representator netdev private structure */
struct i40e_vfpr_netdev_priv {
+ struct metadata_dst *vfpr_dst;
struct i40e_vf *vf;
};
--
2.5.5
^ permalink raw reply related
* [next-queue v4 PATCH 7/7] i40e: Add support to get switch id and port number for VFPR netdevs
From: Sridhar Samudrala @ 2017-01-12 22:35 UTC (permalink / raw)
To: alexander.h.duyck, john.r.fastabend, anjali.singhai,
jakub.kicinski, davem, scott.d.peterson, gerlitz.or, jiri,
intel-wired-lan, netdev
In-Reply-To: <1484260510-9162-1-git-send-email-sridhar.samudrala@intel.com>
Introduce switchdev_ops to PF and VFPR netdevs to return the switch id
via SWITCHDEV_ATTR_ID_PORT_PARENT_ID attribute.
Also, ndo_get_phys_port_name() support is added to VFPR netdevs to
return the port number.
PF: enp5s0f0, VFs: enp5s2,enp5s2f1 VFPRs:enp5s0f0-vf0, enp5s0f0-vf1
# rmmod i40e; modprobe i40e
# devlink dev eswitch set pci/0000:05:00.0 mode switchdev
# echo 2 > /sys/class/net/enp5s0f0/device/sriov_numvfs
# ip -d l show enp5s0f0
32: enp5s0f0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
link/ether 68:05:ca:2e:72:68 brd ff:ff:ff:ff:ff:ff promiscuity 0 addrgenmode eui64 numtxqueues 72 numrxqueues 72 gso_max_size 65536 gso_max_segs 65535 portid 6805ca2e7268 switchid 6805ca2e7268
vf 0 MAC 00:00:00:00:00:00, spoof checking on, link-state disable, trust off
vf 1 MAC 00:00:00:00:00:00, spoof checking on, link-state disable, trust off
# ip -d l show enp5s0f0-vf0
34: enp5s0f0-vf0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
link/ether 68:05:ca:2e:72:68 brd ff:ff:ff:ff:ff:ff promiscuity 0 addrgenmode eui64 numtxqueues 1 numrxqueues 1 gso_max_size 65536 gso_max_segs 65535 portname 0 switchid 6805ca2e7268
# ip -d l show enp5s0f0-vf1
35: enp5s0f0-vf1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
link/ether 68:05:ca:2e:72:68 brd ff:ff:ff:ff:ff:ff promiscuity 0 addrgenmode eui64 numtxqueues 1 numrxqueues 1 gso_max_size 65536 gso_max_segs 65535 portname 1 switchid 6805ca2e7268
Signed-off-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e.h | 1 +
drivers/net/ethernet/intel/i40e/i40e_main.c | 28 +++++++++++++++
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 40 ++++++++++++++++++++++
3 files changed, 69 insertions(+)
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h
index 95cfe46..61720d6 100644
--- a/drivers/net/ethernet/intel/i40e/i40e.h
+++ b/drivers/net/ethernet/intel/i40e/i40e.h
@@ -56,6 +56,7 @@
#include <linux/ptp_clock_kernel.h>
#include <net/devlink.h>
#include <net/dst_metadata.h>
+#include <net/switchdev.h>
#include "i40e_type.h"
#include "i40e_prototype.h"
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index 9212656..cd55665 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -9650,6 +9650,31 @@ static const struct net_device_ops i40e_netdev_ops = {
.ndo_xdp = i40e_xdp,
};
+static int i40e_pf_attr_get(struct net_device *dev, struct switchdev_attr *attr)
+{
+ struct i40e_netdev_priv *np = netdev_priv(dev);
+ struct i40e_vsi *vsi = np->vsi;
+ struct i40e_pf *pf = vsi->back;
+
+ if (pf->eswitch_mode == DEVLINK_ESWITCH_MODE_LEGACY)
+ return -EOPNOTSUPP;
+
+ switch (attr->id) {
+ case SWITCHDEV_ATTR_ID_PORT_PARENT_ID:
+ attr->u.ppid.id_len = ETH_ALEN;
+ ether_addr_copy(attr->u.ppid.id, dev->dev_addr);
+ break;
+ default:
+ return -EOPNOTSUPP;
+ }
+
+ return 0;
+}
+
+static const struct switchdev_ops i40e_pf_switchdev_ops = {
+ .switchdev_port_attr_get = i40e_pf_attr_get,
+};
+
/**
* i40e_config_netdev - Setup the netdev flags
* @vsi: the VSI being configured
@@ -9769,6 +9794,9 @@ static int i40e_config_netdev(struct i40e_vsi *vsi)
#ifdef I40E_FCOE
i40e_fcoe_config_netdev(netdev, vsi);
#endif
+#ifdef CONFIG_NET_SWITCHDEV
+ netdev->switchdev_ops = &i40e_pf_switchdev_ops;
+#endif
/* MTU range: 68 - 9706 */
netdev->min_mtu = ETH_MIN_MTU;
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index 1af8472..465154f 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -1154,6 +1154,22 @@ i40e_vfpr_netdev_get_offload_stats(int attr_id, const struct net_device *dev,
return -EINVAL;
}
+static int
+i40e_vfpr_netdev_get_phys_port_name(struct net_device *dev, char *buf,
+ size_t len)
+{
+ struct i40e_vfpr_netdev_priv *priv = netdev_priv(dev);
+ struct i40e_vf *vf = priv->vf;
+
+ int ret;
+
+ ret = snprintf(buf, len, "%d", vf->vf_id);
+ if (ret >= len)
+ return -EOPNOTSUPP;
+
+ return 0;
+}
+
static const struct net_device_ops i40e_vfpr_netdev_ops = {
.ndo_open = i40e_vfpr_netdev_open,
.ndo_stop = i40e_vfpr_netdev_stop,
@@ -1161,6 +1177,26 @@ static const struct net_device_ops i40e_vfpr_netdev_ops = {
.ndo_get_stats64 = i40e_vfpr_netdev_get_stats64,
.ndo_has_offload_stats = i40e_vfpr_netdev_has_offload_stats,
.ndo_get_offload_stats = i40e_vfpr_netdev_get_offload_stats,
+ .ndo_get_phys_port_name = i40e_vfpr_netdev_get_phys_port_name,
+};
+
+static int i40e_vfpr_attr_get(struct net_device *dev,
+ struct switchdev_attr *attr)
+{
+ switch (attr->id) {
+ case SWITCHDEV_ATTR_ID_PORT_PARENT_ID:
+ attr->u.ppid.id_len = ETH_ALEN;
+ ether_addr_copy(attr->u.ppid.id, dev->dev_addr);
+ break;
+ default:
+ return -EOPNOTSUPP;
+ }
+
+ return 0;
+}
+
+static const struct switchdev_ops i40e_vfpr_switchdev_ops = {
+ .switchdev_port_attr_get = i40e_vfpr_attr_get,
};
/**
@@ -1235,6 +1271,10 @@ int i40e_alloc_vfpr_netdev(struct i40e_vf *vf, u16 vf_num)
vfpr_netdev->netdev_ops = &i40e_vfpr_netdev_ops;
eth_hw_addr_inherit(vfpr_netdev, vsi->netdev);
+#ifdef CONFIG_NET_SWITCHDEV
+ vfpr_netdev->switchdev_ops = &i40e_vfpr_switchdev_ops;
+#endif
+
netif_carrier_off(vfpr_netdev);
netif_tx_disable(vfpr_netdev);
--
2.5.5
^ permalink raw reply related
* [next-queue v4 PATCH 6/7] i40e: Add support for exposing VF port statistics via VFPR netdev on the host.
From: Sridhar Samudrala @ 2017-01-12 22:35 UTC (permalink / raw)
To: alexander.h.duyck, john.r.fastabend, anjali.singhai,
jakub.kicinski, davem, scott.d.peterson, gerlitz.or, jiri,
intel-wired-lan, netdev
In-Reply-To: <1484260510-9162-1-git-send-email-sridhar.samudrala@intel.com>
From: Sridhar Samudrala <sridhar.samudrala@intel.com>
By default stats counted by HW are returned via the original ndo_get_stats64()
api. Stats counted in SW are returned via ndo_get_offload_stats() api.
Small script to demonstrate vfpr stats in switchdev mode.
PF: enp5s0f0, VFs: enp5s2,enp5s2f1 VFPRs:enp5s0f0-vf0, enp5s0f0-vf1
# rmmod i40e; modprobe i40e
# devlink dev eswitch set pci/0000:05:00.0 mode switchdev
# echo 2 > /sys/class/net/enp5s0f0/device/sriov_numvfs
# ip link set enp5s0f0 vf 0 mac 00:11:22:33:44:55
# ip link set enp5s0f0 vf 1 mac 00:11:22:33:44:56
# rmmod i40evf; modprobe i40evf
/* Create 2 namespaces and move the VFs to the corresponding ns */
# ip netns add ns0
# ip link set enp5s2 netns ns0
# ip netns exec ns0 ip addr add 192.168.1.10/24 dev enp5s2
# ip netns exec ns0 ip link set enp5s2 up
# ip netns add ns1
# ip link set enp5s2f1 netns ns1
# ip netns exec ns1 ip addr add 192.168.1.11/24 dev enp5s2f1
# ip netns exec ns1 ip link set enp5s2f1 up
/* bring up pf and vfpr netdevs */
# ip link set enp5s0f0 up
# ip link set enp5s0f0-vf0 up
# ip link set enp5s0f0-vf1 up
/* Create a linux bridge and add vfpr netdevs to it. */
# ip link add vfpr-br type bridge
# ip link set enp5s0f0-vf0 master vfpr-br
# ip link set enp5s0f0-vf1 master vfpr-br
# ip addr add 192.168.1.1/24 dev vfpr-br
# ip link set vfpr-br up
# ip netns exec ns0 ping -c3 192.168.1.11
# ip netns exec ns1 ping -c3 192.168.1.10
# ip netns exec ns0 ip -s l show enp5s2
56: enp5s2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
link/ether 00:11:22:33:44:55 brd ff:ff:ff:ff:ff:ff
RX: bytes packets errors dropped overrun mcast
1468 18 0 0 0 0
TX: bytes packets errors dropped carrier collsns
1398 17 0 0 0 0
# ip -s l show enp5s0f0-vf0
52: enp5s0f0-vf0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel master vfpr-br state UP mode DEFAULT group default qlen 1000
link/ether 68:05:ca:2e:72:68 brd ff:ff:ff:ff:ff:ff
RX: bytes packets errors dropped overrun mcast
1398 17 0 0 0 0
TX: bytes packets errors dropped carrier collsns
1468 18 0 0 0 0
# ip netns exec ns1 ip -s l show enp5s2f1
57: enp5s2f1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
link/ether 00:11:22:33:44:56 brd ff:ff:ff:ff:ff:ff
RX: bytes packets errors dropped overrun mcast
1486 18 0 0 0 0
TX: bytes packets errors dropped carrier collsns
1538 19 0 0 0 0
# ip -s l show enp5s0f0-vf1
53: enp5s0f0-vf1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel master vfpr-br state UP mode DEFAULT group default qlen 1000
link/ether 68:05:ca:2e:72:68 brd ff:ff:ff:ff:ff:ff
RX: bytes packets errors dropped overrun mcast
1538 19 0 0 0 0
TX: bytes packets errors dropped carrier collsns
1486 18 0 0 0 0
Signed-off-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
---
drivers/net/ethernet/intel/i40e/i40e_txrx.c | 44 ++++++++-
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 108 +++++++++++++++++++++
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h | 10 ++
3 files changed, 160 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_txrx.c b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
index f43d1df..d1583ee 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_txrx.c
@@ -1279,6 +1279,32 @@ static bool i40e_alloc_mapped_page(struct i40e_ring *rx_ring,
}
/**
+ * i40e_vfpr_receive_skb
+ * @vf: pointer to VF
+ * @skb: packet to send up
+ *
+ * Update skb dev to vfpr netdev and rx stats.
+ **/
+static void i40e_vfpr_receive_skb(struct i40e_vf *vf, struct sk_buff *skb)
+{
+ struct i40e_vfpr_netdev_priv *priv;
+ struct vfpr_pcpu_stats *vfpr_stats;
+
+ if (!vf->vfpr_netdev)
+ return;
+
+ skb->dev = vf->vfpr_netdev;
+
+ priv = netdev_priv(vf->vfpr_netdev);
+ vfpr_stats = this_cpu_ptr(priv->vfpr_stats);
+
+ u64_stats_update_begin(&vfpr_stats->syncp);
+ vfpr_stats->rx_packets++;
+ vfpr_stats->rx_bytes += skb->len;
+ u64_stats_update_end(&vfpr_stats->syncp);
+}
+
+/**
* i40e_receive_skb - Send a completed packet up the stack
* @rx_ring: rx ring in play
* @skb: packet to send up
@@ -1310,7 +1336,7 @@ static void i40e_receive_skb(struct i40e_ring *rx_ring,
vf = &pf->vf[vf_id];
if (ether_addr_equal(eth->h_source,
vf->default_lan_addr.addr)) {
- skb->dev = vf->vfpr_netdev;
+ i40e_vfpr_receive_skb(vf, skb);
break;
}
}
@@ -3428,11 +3454,25 @@ netdev_tx_t i40e_vfpr_netdev_start_xmit(struct sk_buff *skb,
struct i40e_vf *vf = priv->vf;
struct i40e_pf *pf = vf->pf;
struct i40e_vsi *vsi = pf->vsi[pf->lan_vsi];
+ int ret;
skb_dst_drop(skb);
dst_hold(&priv->vfpr_dst->dst);
skb_dst_set(skb, &priv->vfpr_dst->dst);
skb->dev = vsi->netdev;
- return dev_queue_xmit(skb);
+ ret = dev_queue_xmit(skb);
+ if (likely(ret == NET_XMIT_SUCCESS || ret == NET_XMIT_CN)) {
+ struct vfpr_pcpu_stats *vfpr_stats;
+
+ vfpr_stats = this_cpu_ptr(priv->vfpr_stats);
+ u64_stats_update_begin(&vfpr_stats->syncp);
+ vfpr_stats->tx_packets++;
+ vfpr_stats->tx_bytes += skb->len;
+ u64_stats_update_end(&vfpr_stats->syncp);
+ } else {
+ this_cpu_inc(priv->vfpr_stats->tx_drops);
+ }
+
+ return ret;
}
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
index 7211fba..1af8472 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c
@@ -1057,10 +1057,110 @@ static int i40e_vfpr_netdev_stop(struct net_device *dev)
return 0;
}
+/**
+ * i40e_vfpr_netdev_get_stats64
+ * @dev: network interface device structure
+ * @stats: netlink stats structure
+ *
+ * Fills the hw statistics from the VSI corresponding to the associated VFPR
+ **/
+void
+i40e_vfpr_netdev_get_stats64(struct net_device *netdev,
+ struct rtnl_link_stats64 *stats)
+{
+ struct i40e_vfpr_netdev_priv *priv = netdev_priv(netdev);
+ struct i40e_vf *vf = priv->vf;
+ struct i40e_pf *pf = vf->pf;
+ struct i40e_vsi *vsi;
+ struct i40e_eth_stats *estats;
+
+ vsi = pf->vsi[vf->lan_vsi_idx];
+ i40e_update_stats(vsi);
+
+ estats = &vsi->eth_stats;
+
+ /* TX and RX stats are flipped as we are returning the stats as seen
+ * at the switch port corresponding to the VF.
+ */
+ stats->rx_packets = estats->tx_unicast + estats->tx_multicast +
+ estats->tx_broadcast;
+ stats->tx_packets = estats->rx_unicast + estats->rx_multicast +
+ estats->rx_broadcast;
+ stats->rx_bytes = estats->tx_bytes;
+ stats->tx_bytes = estats->rx_bytes;
+ stats->rx_dropped = estats->tx_discards;
+ stats->tx_dropped = estats->rx_discards;
+}
+
+/**
+ * i40e_vfpr_get_cpu_hit_stats64
+ * @dev: network interface device structure
+ * @stats: netlink stats structure
+ *
+ * stats are filled from the priv structure. correspond to the packets
+ * that are seen by the cpu and sent/received via vfpr netdev.
+ **/
+static int
+i40e_vfpr_get_cpu_hit_stats64(const struct net_device *dev,
+ struct rtnl_link_stats64 *stats)
+{
+ struct i40e_vfpr_netdev_priv *priv = netdev_priv(dev);
+ int i;
+
+ for_each_possible_cpu(i) {
+ struct vfpr_pcpu_stats *vfpr_stats;
+ u64 tbytes, tpkts, tdrops, rbytes, rpkts;
+ unsigned int start;
+
+ vfpr_stats = per_cpu_ptr(priv->vfpr_stats, i);
+ do {
+ start = u64_stats_fetch_begin_irq(&vfpr_stats->syncp);
+ tbytes = vfpr_stats->tx_bytes;
+ tpkts = vfpr_stats->tx_packets;
+ tdrops = vfpr_stats->tx_drops;
+ rbytes = vfpr_stats->rx_bytes;
+ rpkts = vfpr_stats->rx_packets;
+ } while (u64_stats_fetch_retry_irq(&vfpr_stats->syncp, start));
+ stats->tx_bytes += tbytes;
+ stats->tx_packets += tpkts;
+ stats->tx_dropped += tdrops;
+ stats->rx_bytes += rbytes;
+ stats->rx_packets += rpkts;
+ }
+
+ return 0;
+}
+
+static bool
+i40e_vfpr_netdev_has_offload_stats(const struct net_device *dev, int attr_id)
+{
+ switch (attr_id) {
+ case IFLA_OFFLOAD_XSTATS_CPU_HIT:
+ return true;
+ }
+
+ return false;
+}
+
+static int
+i40e_vfpr_netdev_get_offload_stats(int attr_id, const struct net_device *dev,
+ void *sp)
+{
+ switch (attr_id) {
+ case IFLA_OFFLOAD_XSTATS_CPU_HIT:
+ return i40e_vfpr_get_cpu_hit_stats64(dev, sp);
+ }
+
+ return -EINVAL;
+}
+
static const struct net_device_ops i40e_vfpr_netdev_ops = {
.ndo_open = i40e_vfpr_netdev_open,
.ndo_stop = i40e_vfpr_netdev_stop,
.ndo_start_xmit = i40e_vfpr_netdev_start_xmit,
+ .ndo_get_stats64 = i40e_vfpr_netdev_get_stats64,
+ .ndo_has_offload_stats = i40e_vfpr_netdev_has_offload_stats,
+ .ndo_get_offload_stats = i40e_vfpr_netdev_get_offload_stats,
};
/**
@@ -1119,6 +1219,13 @@ int i40e_alloc_vfpr_netdev(struct i40e_vf *vf, u16 vf_num)
pf->vf[vf_num].vfpr_netdev = vfpr_netdev;
priv = netdev_priv(vfpr_netdev);
+ priv->vfpr_stats = netdev_alloc_pcpu_stats(struct vfpr_pcpu_stats);
+ if (!priv->vfpr_stats) {
+ dev_err(&pf->pdev->dev, "alloc_pcpu_stats failed for vf:%d\n",
+ vf_num);
+ free_netdev(vfpr_netdev);
+ return -ENOMEM;
+ }
priv->vf = &pf->vf[vf_num];
priv->vfpr_dst = metadata_dst_alloc(0, METADATA_HW_PORT_MUX,
GFP_KERNEL);
@@ -1175,6 +1282,7 @@ void i40e_free_vfpr_netdev(struct i40e_vf *vf)
priv = netdev_priv(vf->vfpr_netdev);
dst_release((struct dst_entry *)priv->vfpr_dst);
+ free_percpu(priv->vfpr_stats);
unregister_netdev(vf->vfpr_netdev);
free_netdev(vf->vfpr_netdev);
diff --git a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h
index 3dea207..52ba9d5 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h
+++ b/drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.h
@@ -72,10 +72,20 @@ enum i40e_vf_capabilities {
I40E_VIRTCHNL_VF_CAP_IWARP,
};
+struct vfpr_pcpu_stats {
+ u64 tx_packets;
+ u64 tx_bytes;
+ u64 tx_drops;
+ u64 rx_packets;
+ u64 rx_bytes;
+ struct u64_stats_sync syncp;
+};
+
/* VF Port representator netdev private structure */
struct i40e_vfpr_netdev_priv {
struct metadata_dst *vfpr_dst;
struct i40e_vf *vf;
+ struct vfpr_pcpu_stats *vfpr_stats;
};
/* VF information structure */
--
2.5.5
^ permalink raw reply related
* RE: Marvell Phy (1510) issue since v4.7 kernel
From: Kwok, WingMan @ 2017-01-12 22:40 UTC (permalink / raw)
To: Russell King - ARM Linux
Cc: Andrew Lunn, Karicheri, Muralidharan, netdev@vger.kernel.org
In-Reply-To: <20170112214140.GM14217@n2100.armlinux.org.uk>
> -----Original Message-----
> From: Russell King - ARM Linux [mailto:linux@armlinux.org.uk]
> Sent: Thursday, January 12, 2017 4:42 PM
> To: Kwok, WingMan
> Cc: Andrew Lunn; Karicheri, Muralidharan; netdev@vger.kernel.org
> Subject: Re: Marvell Phy (1510) issue since v4.7 kernel
>
> On Thu, Jan 12, 2017 at 09:31:27PM +0000, Kwok, WingMan wrote:
> >
> > > > I double checked that ours is actually a 88E1514. According
> > > > to Marvell's brief description, it seems that it does support
> > > > fiber.
> > > >
> > > > Reading page 18 reg 30 returns 6.
> > >
> > > O.K, that is consistent. Looking at the Marvell SDK:
> > >
> > > phy/Include/madApiDefs.h:#define MAD_SUB_88E1510 4 /*
> 88E1510 */
> > > phy/Include/madApiDefs.h:#define MAD_SUB_88E1518 5 /*
> 88E1518 */
> > > phy/Include/madApiDefs.h:#define MAD_SUB_88E1512 6 /*
> > > 88E1512/88E1514 */
> > >
> > > I took another look at the patch adding Fibre support. The
> > > suspend/resume functions are wrong.
> > >
> > > /* Suspend the fiber mode first */
> > > if (!(phydev->supported & SUPPORTED_FIBRE)) {
> > >
> > > The ! should not be there. Does removing them both fix your
> problem?
> > >
> >
> > Hi Andrew,
> >
> > Thanks for taking the time to look into those patches. Yes we notice
> > the error in the suspend/resume functions also.
> >
> > But our problem is caused by the read_status function:
> >
> > if ((phydev->supported & SUPPORTED_FIBRE)) {
> > err = phy_write(phydev, MII_MARVELL_PHY_PAGE,
> MII_M1111_FIBER);
> > if (err < 0)
> > goto error;
> >
> > err = marvell_read_status_page(phydev, MII_M1111_FIBER);
> > if (err < 0)
> > goto error;
> >
> > /* If the fiber link is up, it is the selected and used
> link.
> > * In this case, we need to stay in the fiber page.
> > * Please to be careful about that, avoid to restore Copper
> page
> > * in other functions which could break the behaviour
> > * for some fiber phy like 88E1512.
> > * */
> > if (phydev->link)
> > return 0;
> >
> > which keeps the fiber page if phydev->link is true (for some
> > reason this is the case even though we are not using fiber)
>
> See below.
>
> > However, this causes a problem in kernel reboot because neither
> > the suspend/resume is called to restore the copper page and
> > u-boot marvell phy driver does not support 1510 fiber, which
> > will then result in writing to the wrong phy regs and causes
> > a sgmii auto-nego time out.
>
> So what we need to do is to have a .shutdown function installed - but
> it's not that simple...
>
> .mdiodrv = {
> .driver = {
> .shutdown = mv88e1512_shutdown,
> },
> },
>
> in the appropriate phy_driver array element, and then have:
>
> static void mv88e1512_shutdown(struct device *dev)
> {
> struct phy_device *phydev = to_phy_device(dev);
>
> phy_write(phydev, MII_MARVELL_PHY_PAGE, MII_M1111_COPPER);
> }
>
> which will restore the copper page on non-emergency reboots. For
> emergency reboots, we're out of luck... the real fix would be for
> uboot to ensure that when it detects this PHY, it ensures that the
> right page is selected.
>
> Now, that all said, there's a bug in this code, which is fixed by
> a patch I recently submitted. If you have an 88E151x connected via
> SGMII, then the read_status function incorrectly believes it should
> be reading from the fiber page - when in fact the fiber page is
> reporting the status of the host-side link. So, before trying the
> above modification, please try this patch - it's already been merged
> into the net tree, so should hit -rc soon.
>
> My guess from what you've said above is that your 88E151x is connected
> via SGMII, so you need the patch below rather than trying to install
> a shutdown function.
>
> 8<====
> From: Russell King <rmk+kernel@armlinux.org.uk>
> Subject: [PATCH] net: phy: marvell: fix Marvell 88E1512 used in SGMII
> mode
>
Hi Russell,
Thanks for the explanations. Yes, our 88e1514 is connected to the
host via sgmii and your patch does provide a solution to our problem.
Regards,
WingMan
^ permalink raw reply
* [PATCH] can: Fix kernel panic at security_sock_rcv_skb
From: william.c.roberts @ 2017-01-12 22:40 UTC (permalink / raw)
To: selinux, socketcan, mkl, davem, linux-can, netdev, linux-kernel
Cc: sds, paul, Zhang Yanmin, He, Bo, Liu Shuo A, William Roberts
From: Zhang Yanmin <yanmin.zhang@intel.com>
The patch is for fix the below kernel panic:
BUG: unable to handle kernel NULL pointer dereference at (null)
IP: [<ffffffff81495e25>] selinux_socket_sock_rcv_skb+0x65/0x2a0
Call Trace:
<IRQ>
[<ffffffff81485d8c>] security_sock_rcv_skb+0x4c/0x60
[<ffffffff81d55771>] sk_filter+0x41/0x210
[<ffffffff81d12913>] sock_queue_rcv_skb+0x53/0x3a0
[<ffffffff81f0a2b3>] raw_rcv+0x2a3/0x3c0
[<ffffffff81f06eab>] can_rcv_filter+0x12b/0x370
[<ffffffff81f07af9>] can_receive+0xd9/0x120
[<ffffffff81f07beb>] can_rcv+0xab/0x100
[<ffffffff81d362ac>] __netif_receive_skb_core+0xd8c/0x11f0
[<ffffffff81d36734>] __netif_receive_skb+0x24/0xb0
[<ffffffff81d37f67>] process_backlog+0x127/0x280
[<ffffffff81d36f7b>] net_rx_action+0x33b/0x4f0
[<ffffffff810c88d4>] __do_softirq+0x184/0x440
[<ffffffff81f9e86c>] do_softirq_own_stack+0x1c/0x30
<EOI>
[<ffffffff810c76fb>] do_softirq.part.18+0x3b/0x40
[<ffffffff810c8bed>] do_softirq+0x1d/0x20
[<ffffffff81d30085>] netif_rx_ni+0xe5/0x110
[<ffffffff8199cc87>] slcan_receive_buf+0x507/0x520
[<ffffffff8167ef7c>] flush_to_ldisc+0x21c/0x230
[<ffffffff810e3baf>] process_one_work+0x24f/0x670
[<ffffffff810e44ed>] worker_thread+0x9d/0x6f0
[<ffffffff810e4450>] ? rescuer_thread+0x480/0x480
[<ffffffff810ebafc>] kthread+0x12c/0x150
[<ffffffff81f9ccef>] ret_from_fork+0x3f/0x70
The sk dereferenced in panic has been released. After the rcu_call in
can_rx_unregister, receiver was protected by RCU but inner data was
not, then later sk will be freed while other CPU is still using it.
We need wait here to make sure sk referenced via receiver was safe.
=> security_sk_free
=> sk_destruct
=> __sk_free
=> sk_free
=> raw_release
=> sock_release
=> sock_close
=> __fput
=> ____fput
=> task_work_run
=> exit_to_usermode_loop
=> syscall_return_slowpath
=> int_ret_from_sys_call
Tracked-On: https://jira01.devtools.intel.com/browse/OAM-40528
Signed-off-by: Zhang Yanmin <yanmin.zhang@intel.com>
Signed-off-by: He, Bo <bo.he@intel.com>
Signed-off-by: Liu Shuo A <shuo.a.liu@intel.com>
Signed-off-by: William Roberts <william.c.roberts@intel.com>
---
net/can/af_can.c | 14 ++++++++------
net/can/af_can.h | 1 -
2 files changed, 8 insertions(+), 7 deletions(-)
diff --git a/net/can/af_can.c b/net/can/af_can.c
index 1108079..fcbe971 100644
--- a/net/can/af_can.c
+++ b/net/can/af_can.c
@@ -517,10 +517,8 @@ EXPORT_SYMBOL(can_rx_register);
/*
* can_rx_delete_receiver - rcu callback for single receiver entry removal
*/
-static void can_rx_delete_receiver(struct rcu_head *rp)
+static void can_rx_delete_receiver(struct receiver *r)
{
- struct receiver *r = container_of(rp, struct receiver, rcu);
-
kmem_cache_free(rcv_cache, r);
}
@@ -595,9 +593,13 @@ void can_rx_unregister(struct net_device *dev, canid_t can_id, canid_t mask,
out:
spin_unlock(&can_rcvlists_lock);
- /* schedule the receiver item for deletion */
- if (r)
- call_rcu(&r->rcu, can_rx_delete_receiver);
+ /* synchronize_rcu to wait until a grace period has elapsed, to make
+ * sure all receiver's sk dereferenced by others.
+ */
+ if (r) {
+ synchronize_rcu();
+ can_rx_delete_receiver(r);
+ }
}
EXPORT_SYMBOL(can_rx_unregister);
diff --git a/net/can/af_can.h b/net/can/af_can.h
index fca0fe9..a0cbf83 100644
--- a/net/can/af_can.h
+++ b/net/can/af_can.h
@@ -50,7 +50,6 @@
struct receiver {
struct hlist_node list;
- struct rcu_head rcu;
canid_t can_id;
canid_t mask;
unsigned long matches;
--
2.7.4
^ permalink raw reply related
* Re: [net-next PATCH 1/3] Revert "icmp: avoid allocating large struct on stack"
From: Cong Wang @ 2017-01-12 22:46 UTC (permalink / raw)
To: David Miller
Cc: Eric Dumazet, Jesper Dangaard Brouer,
Linux Kernel Network Developers
In-Reply-To: <20170110.135421.836122193184579759.davem@davemloft.net>
On Tue, Jan 10, 2017 at 10:54 AM, David Miller <davem@davemloft.net> wrote:
> From: Cong Wang <xiyou.wangcong@gmail.com>
> Date: Tue, 10 Jan 2017 10:44:59 -0800
>> The only countries you hold netdev are Canada, Japan and Spain
>> (to my knowledge). If you check:
>>
>> https://en.wikipedia.org/wiki/Visa_requirements_for_Chinese_citizens#Visa_requirements
>>
>> It is very easy to find out if it is an excuse or a fact.
>
> The conference explciitly offers VISA help for anyone who needs it.
> Many people come to the conference on a VISA we helped them obtain.
I never complain about visa sponsorship or money, this is the last
problem for me to consider.
>
> It is not hard to do. You have put in exactly zero effort in trying
> to solve this problem. If you had simply contacted the conference
> organizers asking for help, you would have gotten all the help you
> needed.
If obtaining a visa were as easy as obtaining a sponsorship, I would
go as many conferences as I know. But the fact is there are already
too many hassles to obtain even _one_ visa, not to mention I need
to obtain _two_ visas to go to Canada (or Japan, Spain) and return
to US. I'd be very happy to attend it if it were held in US, but this never
happens.
^ permalink raw reply
* Re: [PATCH v2] tcp: fix tcp_fastopen unaligned access complaints on sparc
From: Eric Dumazet @ 2017-01-12 22:49 UTC (permalink / raw)
To: Shannon Nelson; +Cc: netdev, davem, sparclinux, linux-kernel, rob.gardner
In-Reply-To: <1484259898-30719-1-git-send-email-shannon.nelson@oracle.com>
On Thu, 2017-01-12 at 14:24 -0800, Shannon Nelson wrote:
> Fix up a data alignment issue on sparc by swapping the order
> of the cookie byte array field with the length field in
> struct tcp_fastopen_cookie, and making it a proper union
> to clean up the typecasting.
>
> This addresses log complaints like these:
> log_unaligned: 113 callbacks suppressed
> Kernel unaligned access at TPC[976490] tcp_try_fastopen+0x2d0/0x360
> Kernel unaligned access at TPC[9764ac] tcp_try_fastopen+0x2ec/0x360
> Kernel unaligned access at TPC[9764c8] tcp_try_fastopen+0x308/0x360
> Kernel unaligned access at TPC[9764e4] tcp_try_fastopen+0x324/0x360
> Kernel unaligned access at TPC[976490] tcp_try_fastopen+0x2d0/0x360
>
> Cc: Eric Dumazet <eric.dumazet@gmail.com>
> Signed-off-by: Shannon Nelson <shannon.nelson@oracle.com>
> ---
> v2: Use Eric's suggestion for a union in the struct
Acked-by: Eric Dumazet <edumazet@google.com>
Thanks for fixing this !
^ permalink raw reply
* RE: Marvell Phy (1510) issue since v4.7 kernel
From: Kwok, WingMan @ 2017-01-12 22:49 UTC (permalink / raw)
To: Andrew Lunn
Cc: rmk+kernel@arm.linux.org.uk, Karicheri, Muralidharan,
netdev@vger.kernel.org
In-Reply-To: <20170112214941.GE10203@lunn.ch>
Hi Andrew,
> -----Original Message-----
> From: Andrew Lunn [mailto:andrew@lunn.ch]
> Sent: Thursday, January 12, 2017 4:50 PM
> To: Kwok, WingMan
> Cc: rmk+kernel@arm.linux.org.uk; Karicheri, Muralidharan;
> netdev@vger.kernel.org
> Subject: Re: Marvell Phy (1510) issue since v4.7 kernel
>
> > But our problem is caused by the read_status function:
> >
> > if ((phydev->supported & SUPPORTED_FIBRE)) {
> > err = phy_write(phydev, MII_MARVELL_PHY_PAGE,
> MII_M1111_FIBER);
> > if (err < 0)
> > goto error;
> >
> > err = marvell_read_status_page(phydev, MII_M1111_FIBER);
> > if (err < 0)
> > goto error;
> >
> > /* If the fiber link is up, it is the selected and used
> link.
> > * In this case, we need to stay in the fiber page.
> > * Please to be careful about that, avoid to restore Copper
> page
> > * in other functions which could break the behaviour
> > * for some fiber phy like 88E1512.
> > * */
> > if (phydev->link)
> > return 0;
> >
> > which keeps the fiber page if phydev->link is true (for some
> > reason this is the case even though we are not using fiber)
>
> How are you using the PHY. What phy-mode do you have set? Do you
> happen to be using it as an RGMII to SERDES/SGMII bridge? This is what
> Russell King is doing, i think.
>
our 88e1514 is connected to the host via sgmii.
> Have you tried the patch Russell submitted recently.
>
> Author: Russell King <rmk+kernel@armlinux.org.uk>
> Date: Tue Jan 10 23:13:45 2017 +0000
>
> net: phy: marvell: fix Marvell 88E1512 used in SGMII mode
>
> When an Marvell 88E1512 PHY is connected to a nic in SGMII mode,
> the
> fiber page is used for the SGMII host-side connection. The PHY
> driver
> notices that SUPPORTED_FIBRE is set, so it tries reading the fiber
> page
> for the link status, and ends up reading the MAC-side status
> instead of
> the outgoing (copper) link. This leads to incorrect results
> reported
> via ethtool.
>
> If the PHY is connected via SGMII to the host, ignore the fiber
> page.
> However, continue to allow the existing power management code to
> suspend and resume the fiber page.
>
Thanks for pointer. It does fix the problem.
> > However, this causes a problem in kernel reboot because neither
> > the suspend/resume is called to restore the copper page and
> > u-boot marvell phy driver does not support 1510 fiber, which
> > will then result in writing to the wrong phy regs and causes
> > a sgmii auto-nego time out.
>
> This is still a u-boot bug. It should not assume the PHY is in a sane
> state. It should reset it and configure it as needed. So far, i don't
> think you have reported any issues with Linux usage of the PHY. There
> clearly are bugs, but your real problem is u-boot.
>
Yes. Agree.
> > In addition to fixing the ! in suspend/resume, my suggestion
> > would be to change also the read_status function to
> > always restore the copper page after doing the fiber stuffs:
>
> Nope. This is done deliberately, as the comment suggests:
>
> > /* If the fiber link is up, it is the selected and used
> link.
> > * In this case, we need to stay in the fiber page.
> > * Please to be careful about that, avoid to restore Copper
> page
> > * in other functions which could break the behaviour
> > * for some fiber phy like 88E1512.
> > * */
> > if (phydev->link)
> > return 0;
>
> The point is, the phylib will continue polling the PHY registers,
> reading them. If the FIBRE is up, we want to read the FIBRE values,
> not the copper.
>
Thanks for the explanations.
> > Another issue is that, as of now, FIBER is enabled regardless
> > of the specific 88e151x. But I believe there is 88e151x chip(s)
> > that does not support fiber. Should fiber be enabled only for
> > those that do support fiber?
>
> Yes, we should look at register 30, page 18 any set SUPPORTED_FIBRE
> based on that.
>
Thanks for the suggestions.
> Andrew
Just want to know if there is already a patch in the net tree
fixing the incorrect ! in the suspend/resume functions also?
WingMan
^ permalink raw reply
* Re: [PATCH net-next v2 05/10] drivers: base: Add device_find_class()
From: Florian Fainelli @ 2017-01-12 22:50 UTC (permalink / raw)
To: David Miller
Cc: netdev, jason, andrew, sebastian.hesselbarth, gregory.clement,
linux, vivien.didelot, linux-arm-kernel, linux-kernel, gregkh
In-Reply-To: <20170112.162135.441956368122992032.davem@davemloft.net>
On 01/12/2017 01:21 PM, David Miller wrote:
> From: Florian Fainelli <f.fainelli@gmail.com>
> Date: Wed, 11 Jan 2017 19:41:16 -0800
>
>> Add a helper function to lookup a device reference given a class name.
>> This is a preliminary patch to remove adhoc code from net/dsa/dsa.c and
>> make it more generic.
>>
>> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
>> ---
>> drivers/base/core.c | 19 +++++++++++++++++++
>> include/linux/device.h | 1 +
>> 2 files changed, 20 insertions(+)
>>
>> diff --git a/drivers/base/core.c b/drivers/base/core.c
>> index 020ea7f05520..3dd6047c10d8 100644
>> --- a/drivers/base/core.c
>> +++ b/drivers/base/core.c
>> @@ -2065,6 +2065,25 @@ struct device *device_find_child(struct device *parent, void *data,
>> }
>> EXPORT_SYMBOL_GPL(device_find_child);
>>
>> +static int dev_is_class(struct device *dev, void *class)
>
> I know you are just moving code, but this class argumnet is a string
> and thus should be "char *" or even "const char *".
Well, this is really so that we don't need to cast the arguments passed
to device_find_child(), which takes a void *data as well. If we made
that a const char *class, we'd get warnings that look like these:
drivers/base/core.c: In function 'device_find_class':
drivers/base/core.c:2083:2: warning: passing argument 2 of
'device_find_child' discards 'const' qualifier from pointer target type
[enabled by default]
return device_find_child(parent, class, dev_is_class);
^
drivers/base/core.c:2050:16: note: expected 'void *' but argument is of
type 'const char *'
struct device *device_find_child(struct device *parent, void *data,
^
drivers/base/core.c:2083:2: warning: passing argument 3 of
'device_find_child' from incompatible pointer type [enabled by default]
return device_find_child(parent, class, dev_is_class);
^
drivers/base/core.c:2050:16: note: expected 'int (*)(struct device *,
void *)' but argument is of type 'int (*)(struct device *, const char *)'
struct device *device_find_child(struct device *parent, void *data,
^
--
Florian
^ permalink raw reply
* Re: [PATCH] ARM: dts: am57xx-beagle-x15: implement errata "Ethernet RGMII2 Limited to 10/100 Mbps"
From: Tony Lindgren @ 2017-01-12 22:53 UTC (permalink / raw)
To: Grygorii Strashko; +Cc: Mugunthan V N, linux-omap, Sekhar Nori, netdev
In-Reply-To: <20170112171429.8106-1-grygorii.strashko@ti.com>
* Grygorii Strashko <grygorii.strashko@ti.com> [170112 09:15]:
> According to errata i880 description the speed of Ethernet port 1 on AM572x
> SoCs rev 1.1 shuld be limited to 10/100Mbps, because RGMII2 Switching
> Characteristics are not compatible with 1000 Mbps operation [1].
> The issue is fixed with Rev 2.0 silicon.
>
> Hence, rework Beagle-X15 and Begale-X15-revb1 to use phy-handle instead of
> phy_id and apply corresponding limitation to the Ethernet Phy 1.
>
> [1] http://www.ti.com/lit/er/sprz429j/sprz429j.pdf
Applying into omap-for-v4.11/dt thanks.
Tony
^ permalink raw reply
* [�PATCH net-next] ARM: dts: vf610-zii-dev: add EEPROM entry to Rev B
From: Vivien Didelot @ 2017-01-12 22:56 UTC (permalink / raw)
To: netdev
Cc: linux-kernel, kernel, David S. Miller, Florian Fainelli,
Andrew Lunn, cphealy, Vivien Didelot
The ZII Dev Rev B board has EEPROMs hanging the 88E6352 Ethernet switch
chips. Add an "eeprom-length" property to allow access from ethtool.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
---
arch/arm/boot/dts/vf610-zii-dev-rev-b.dts | 2 ++
1 file changed, 2 insertions(+)
diff --git a/arch/arm/boot/dts/vf610-zii-dev-rev-b.dts b/arch/arm/boot/dts/vf610-zii-dev-rev-b.dts
index 958b4c42d320..b60d3d03f58c 100644
--- a/arch/arm/boot/dts/vf610-zii-dev-rev-b.dts
+++ b/arch/arm/boot/dts/vf610-zii-dev-rev-b.dts
@@ -98,6 +98,7 @@
interrupts = <27 IRQ_TYPE_LEVEL_LOW>;
interrupt-controller;
#interrupt-cells = <2>;
+ eeprom-length = <512>;
ports {
#address-cells = <1>;
@@ -181,6 +182,7 @@
interrupts = <26 IRQ_TYPE_LEVEL_LOW>;
interrupt-controller;
#interrupt-cells = <2>;
+ eeprom-length = <512>;
ports {
#address-cells = <1>;
--
2.11.0
^ permalink raw reply related
* [PATCH net-next] net: dsa: mv88e6xxx: add EEPROM support to 6390
From: Vivien Didelot @ 2017-01-12 23:07 UTC (permalink / raw)
To: netdev
Cc: linux-kernel, kernel, David S. Miller, Florian Fainelli,
Andrew Lunn, cphealy, Nikita Yushchenko, Vivien Didelot
The Marvell 6352 chip has a 8-bit address/16-bit data EEPROM access.
The Marvell 6390 chip has a 16-bit address/8-bit data EEPROM access.
This patch implements the 8-bit data EEPROM access in the mv88e6xxx
driver and adds its support to chips of the 6390 family.
Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
---
drivers/net/dsa/mv88e6xxx/chip.c | 14 ++++++
drivers/net/dsa/mv88e6xxx/global2.c | 93 ++++++++++++++++++++++++++++++++++-
drivers/net/dsa/mv88e6xxx/global2.h | 21 ++++++++
drivers/net/dsa/mv88e6xxx/mv88e6xxx.h | 1 +
4 files changed, 128 insertions(+), 1 deletion(-)
diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c
index eea8e0176e33..987b2dbbd35a 100644
--- a/drivers/net/dsa/mv88e6xxx/chip.c
+++ b/drivers/net/dsa/mv88e6xxx/chip.c
@@ -3453,6 +3453,8 @@ static const struct mv88e6xxx_ops mv88e6185_ops = {
static const struct mv88e6xxx_ops mv88e6190_ops = {
/* MV88E6XXX_FAMILY_6390 */
+ .get_eeprom = mv88e6xxx_g2_get_eeprom8,
+ .set_eeprom = mv88e6xxx_g2_set_eeprom8,
.set_switch_mac = mv88e6xxx_g2_set_switch_mac,
.phy_read = mv88e6xxx_g2_smi_phy_read,
.phy_write = mv88e6xxx_g2_smi_phy_write,
@@ -3478,6 +3480,8 @@ static const struct mv88e6xxx_ops mv88e6190_ops = {
static const struct mv88e6xxx_ops mv88e6190x_ops = {
/* MV88E6XXX_FAMILY_6390 */
+ .get_eeprom = mv88e6xxx_g2_get_eeprom8,
+ .set_eeprom = mv88e6xxx_g2_set_eeprom8,
.set_switch_mac = mv88e6xxx_g2_set_switch_mac,
.phy_read = mv88e6xxx_g2_smi_phy_read,
.phy_write = mv88e6xxx_g2_smi_phy_write,
@@ -3503,6 +3507,8 @@ static const struct mv88e6xxx_ops mv88e6190x_ops = {
static const struct mv88e6xxx_ops mv88e6191_ops = {
/* MV88E6XXX_FAMILY_6390 */
+ .get_eeprom = mv88e6xxx_g2_get_eeprom8,
+ .set_eeprom = mv88e6xxx_g2_set_eeprom8,
.set_switch_mac = mv88e6xxx_g2_set_switch_mac,
.phy_read = mv88e6xxx_g2_smi_phy_read,
.phy_write = mv88e6xxx_g2_smi_phy_write,
@@ -3556,6 +3562,8 @@ static const struct mv88e6xxx_ops mv88e6240_ops = {
static const struct mv88e6xxx_ops mv88e6290_ops = {
/* MV88E6XXX_FAMILY_6390 */
+ .get_eeprom = mv88e6xxx_g2_get_eeprom8,
+ .set_eeprom = mv88e6xxx_g2_set_eeprom8,
.set_switch_mac = mv88e6xxx_g2_set_switch_mac,
.phy_read = mv88e6xxx_g2_smi_phy_read,
.phy_write = mv88e6xxx_g2_smi_phy_write,
@@ -3714,6 +3722,8 @@ static const struct mv88e6xxx_ops mv88e6352_ops = {
static const struct mv88e6xxx_ops mv88e6390_ops = {
/* MV88E6XXX_FAMILY_6390 */
+ .get_eeprom = mv88e6xxx_g2_get_eeprom8,
+ .set_eeprom = mv88e6xxx_g2_set_eeprom8,
.set_switch_mac = mv88e6xxx_g2_set_switch_mac,
.phy_read = mv88e6xxx_g2_smi_phy_read,
.phy_write = mv88e6xxx_g2_smi_phy_write,
@@ -3741,6 +3751,8 @@ static const struct mv88e6xxx_ops mv88e6390_ops = {
static const struct mv88e6xxx_ops mv88e6390x_ops = {
/* MV88E6XXX_FAMILY_6390 */
+ .get_eeprom = mv88e6xxx_g2_get_eeprom8,
+ .set_eeprom = mv88e6xxx_g2_set_eeprom8,
.set_switch_mac = mv88e6xxx_g2_set_switch_mac,
.phy_read = mv88e6xxx_g2_smi_phy_read,
.phy_write = mv88e6xxx_g2_smi_phy_write,
@@ -3768,6 +3780,8 @@ static const struct mv88e6xxx_ops mv88e6390x_ops = {
static const struct mv88e6xxx_ops mv88e6391_ops = {
/* MV88E6XXX_FAMILY_6390 */
+ .get_eeprom = mv88e6xxx_g2_get_eeprom8,
+ .set_eeprom = mv88e6xxx_g2_set_eeprom8,
.set_switch_mac = mv88e6xxx_g2_set_switch_mac,
.phy_read = mv88e6xxx_g2_smi_phy_read,
.phy_write = mv88e6xxx_g2_smi_phy_write,
diff --git a/drivers/net/dsa/mv88e6xxx/global2.c b/drivers/net/dsa/mv88e6xxx/global2.c
index 3e77071949ab..ead2e265c9ef 100644
--- a/drivers/net/dsa/mv88e6xxx/global2.c
+++ b/drivers/net/dsa/mv88e6xxx/global2.c
@@ -218,7 +218,8 @@ static int mv88e6xxx_g2_clear_pot(struct mv88e6xxx_chip *chip)
}
/* Offset 0x14: EEPROM Command
- * Offset 0x15: EEPROM Data
+ * Offset 0x15: EEPROM Data (for 16-bit data access)
+ * Offset 0x15: EEPROM Addr (for 8-bit data access)
*/
static int mv88e6xxx_g2_eeprom_wait(struct mv88e6xxx_chip *chip)
@@ -239,6 +240,50 @@ static int mv88e6xxx_g2_eeprom_cmd(struct mv88e6xxx_chip *chip, u16 cmd)
return mv88e6xxx_g2_eeprom_wait(chip);
}
+static int mv88e6xxx_g2_eeprom_read8(struct mv88e6xxx_chip *chip,
+ u16 addr, u8 *data)
+{
+ u16 cmd = GLOBAL2_EEPROM_CMD_OP_READ;
+ int err;
+
+ err = mv88e6xxx_g2_eeprom_wait(chip);
+ if (err)
+ return err;
+
+ err = mv88e6xxx_g2_write(chip, GLOBAL2_EEPROM_ADDR, addr);
+ if (err)
+ return err;
+
+ err = mv88e6xxx_g2_eeprom_cmd(chip, cmd);
+ if (err)
+ return err;
+
+ err = mv88e6xxx_g2_read(chip, GLOBAL2_EEPROM_CMD, &cmd);
+ if (err)
+ return err;
+
+ *data = cmd & 0xff;
+
+ return 0;
+}
+
+static int mv88e6xxx_g2_eeprom_write8(struct mv88e6xxx_chip *chip,
+ u16 addr, u8 data)
+{
+ u16 cmd = GLOBAL2_EEPROM_CMD_OP_WRITE | GLOBAL2_EEPROM_CMD_WRITE_EN;
+ int err;
+
+ err = mv88e6xxx_g2_eeprom_wait(chip);
+ if (err)
+ return err;
+
+ err = mv88e6xxx_g2_write(chip, GLOBAL2_EEPROM_ADDR, addr);
+ if (err)
+ return err;
+
+ return mv88e6xxx_g2_eeprom_cmd(chip, cmd | data);
+}
+
static int mv88e6xxx_g2_eeprom_read16(struct mv88e6xxx_chip *chip,
u8 addr, u16 *data)
{
@@ -273,6 +318,52 @@ static int mv88e6xxx_g2_eeprom_write16(struct mv88e6xxx_chip *chip,
return mv88e6xxx_g2_eeprom_cmd(chip, cmd);
}
+int mv88e6xxx_g2_get_eeprom8(struct mv88e6xxx_chip *chip,
+ struct ethtool_eeprom *eeprom, u8 *data)
+{
+ unsigned int offset = eeprom->offset;
+ unsigned int len = eeprom->len;
+ int err;
+
+ eeprom->len = 0;
+
+ while (len) {
+ err = mv88e6xxx_g2_eeprom_read8(chip, offset, data);
+ if (err)
+ return err;
+
+ eeprom->len++;
+ offset++;
+ data++;
+ len--;
+ }
+
+ return 0;
+}
+
+int mv88e6xxx_g2_set_eeprom8(struct mv88e6xxx_chip *chip,
+ struct ethtool_eeprom *eeprom, u8 *data)
+{
+ unsigned int offset = eeprom->offset;
+ unsigned int len = eeprom->len;
+ int err;
+
+ eeprom->len = 0;
+
+ while (len) {
+ err = mv88e6xxx_g2_eeprom_write8(chip, offset, *data);
+ if (err)
+ return err;
+
+ eeprom->len++;
+ offset++;
+ data++;
+ len--;
+ }
+
+ return 0;
+}
+
int mv88e6xxx_g2_get_eeprom16(struct mv88e6xxx_chip *chip,
struct ethtool_eeprom *eeprom, u8 *data)
{
diff --git a/drivers/net/dsa/mv88e6xxx/global2.h b/drivers/net/dsa/mv88e6xxx/global2.h
index 9aefb7d8b0ad..2918f22388f7 100644
--- a/drivers/net/dsa/mv88e6xxx/global2.h
+++ b/drivers/net/dsa/mv88e6xxx/global2.h
@@ -28,10 +28,17 @@ int mv88e6xxx_g2_smi_phy_read(struct mv88e6xxx_chip *chip, int addr, int reg,
int mv88e6xxx_g2_smi_phy_write(struct mv88e6xxx_chip *chip, int addr, int reg,
u16 val);
int mv88e6xxx_g2_set_switch_mac(struct mv88e6xxx_chip *chip, u8 *addr);
+
+int mv88e6xxx_g2_get_eeprom8(struct mv88e6xxx_chip *chip,
+ struct ethtool_eeprom *eeprom, u8 *data);
+int mv88e6xxx_g2_set_eeprom8(struct mv88e6xxx_chip *chip,
+ struct ethtool_eeprom *eeprom, u8 *data);
+
int mv88e6xxx_g2_get_eeprom16(struct mv88e6xxx_chip *chip,
struct ethtool_eeprom *eeprom, u8 *data);
int mv88e6xxx_g2_set_eeprom16(struct mv88e6xxx_chip *chip,
struct ethtool_eeprom *eeprom, u8 *data);
+
int mv88e6xxx_g2_setup(struct mv88e6xxx_chip *chip);
int mv88e6xxx_g2_irq_setup(struct mv88e6xxx_chip *chip);
void mv88e6xxx_g2_irq_free(struct mv88e6xxx_chip *chip);
@@ -67,6 +74,20 @@ static inline int mv88e6xxx_g2_set_switch_mac(struct mv88e6xxx_chip *chip,
return -EOPNOTSUPP;
}
+static inline int mv88e6xxx_g2_get_eeprom8(struct mv88e6xxx_chip *chip,
+ struct ethtool_eeprom *eeprom,
+ u8 *data)
+{
+ return -EOPNOTSUPP;
+}
+
+static inline int mv88e6xxx_g2_set_eeprom8(struct mv88e6xxx_chip *chip,
+ struct ethtool_eeprom *eeprom,
+ u8 *data)
+{
+ return -EOPNOTSUPP;
+}
+
static inline int mv88e6xxx_g2_get_eeprom16(struct mv88e6xxx_chip *chip,
struct ethtool_eeprom *eeprom,
u8 *data)
diff --git a/drivers/net/dsa/mv88e6xxx/mv88e6xxx.h b/drivers/net/dsa/mv88e6xxx/mv88e6xxx.h
index a224d66dafd9..466cfdadb7bd 100644
--- a/drivers/net/dsa/mv88e6xxx/mv88e6xxx.h
+++ b/drivers/net/dsa/mv88e6xxx/mv88e6xxx.h
@@ -382,6 +382,7 @@
#define GLOBAL2_EEPROM_CMD_WRITE_EN BIT(10)
#define GLOBAL2_EEPROM_CMD_ADDR_MASK 0xff
#define GLOBAL2_EEPROM_DATA 0x15
+#define GLOBAL2_EEPROM_ADDR 0x15 /* 6390 */
#define GLOBAL2_PTP_AVB_OP 0x16
#define GLOBAL2_PTP_AVB_DATA 0x17
#define GLOBAL2_SMI_PHY_CMD 0x18
--
2.11.0
^ permalink raw reply related
* Re: [PATCH] xen-netfront: Fix Rx stall during network stress and OOM
From: Vineeth Remanan Pillai @ 2017-01-12 23:09 UTC (permalink / raw)
To: David Miller
Cc: boris.ostrovsky, jgross, xen-devel, netdev, linux-kernel, kamatam,
aliguori
In-Reply-To: <20170112.151706.1389870722624942785.davem@davemloft.net>
On 01/12/2017 12:17 PM, David Miller wrote:
> From: Vineeth Remanan Pillai <vineethp@amazon.com>
> Date: Wed, 11 Jan 2017 23:17:17 +0000
>
>> @@ -1054,7 +1059,11 @@ static int xennet_poll(struct napi_struct *napi, int budget)
>> napi_complete(napi);
>>
>> RING_FINAL_CHECK_FOR_RESPONSES(&queue->rx, more_to_do);
>> - if (more_to_do)
>> +
>> + /* If there is more work to do or could not allocate
>> + * rx buffers, re-enable polling.
>> + */
>> + if (more_to_do || err != 0)
>> napi_schedule(napi);
> Just polling endlessly in a loop retrying the SKB allocation over and over
> again until it succeeds is not very nice behavior.
>
> You already have that refill timer, so please use that to retry instead
> of wasting cpu cycles looping in NAPI poll.
Thanks Dave for the inputs.
On further look, I think I can fix it much simpler by correcting the
test condition
for minimum slots for pushing requests. Existing test is like this:
<snip>
/* Not enough requests? Try again later. */
if (req_prod - queue->rx.rsp_cons < NET_RX_SLOTS_MIN) {
mod_timer(&queue->rx_refill_timer, jiffies + (HZ/10));
return;
}
</snip>
Actually the above check counts more than the newly created request slots
as it counts from rsp_cons. The actual count should be the difference
between
new req_prod and old req_prod(in the queue). If skbs cannot be created,
this
count remains small and hence we would schedule the timer. So the fix
could be:
/* Not enough requests? Try again later. */
- if (req_prod - queue->rx.rsp_cons < NET_RX_SLOTS_MIN) {
+ if (req_prod - queue->rx.sring->req_prod < NET_RX_SLOTS_MIN) {
I have done some initial testing to verify the fix. Will send out v2
patch after couple
more round of testing.
Thanks,
Vineeth
^ permalink raw reply
* Re: [PATCH] can: Fix kernel panic at security_sock_rcv_skb
From: Eric Dumazet @ 2017-01-12 23:10 UTC (permalink / raw)
To: william.c.roberts
Cc: selinux, socketcan, mkl, davem, linux-can, netdev, linux-kernel,
sds, paul, Zhang Yanmin, He, Bo, Liu Shuo A
In-Reply-To: <1484260855-15844-1-git-send-email-william.c.roberts@intel.com>
On Thu, 2017-01-12 at 14:40 -0800, william.c.roberts@intel.com wrote:
> From: Zhang Yanmin <yanmin.zhang@intel.com>
>
> The patch is for fix the below kernel panic:
> BUG: unable to handle kernel NULL pointer dereference at (null)
> IP: [<ffffffff81495e25>] selinux_socket_sock_rcv_skb+0x65/0x2a0
Same patch was sent earlier, and we gave a feedback on it.
Adding synchronize_rcu() calls is a step backward.
https://patchwork.ozlabs.org/patch/714446/
^ permalink raw reply
* Re: [PATCH net-next] net/mlx5e: Support bpf_xdp_adjust_head()
From: Martin KaFai Lau @ 2017-01-12 23:25 UTC (permalink / raw)
To: Saeed Mahameed
Cc: Linux Netdev List, Saeed Mahameed, Tariq Toukan, Kernel Team
In-Reply-To: <CALzJLG-fj3LHEuZ78tRS8OL-x0rPRgKnsoxtJFon4Fw8-WK1AA@mail.gmail.com>
On Thu, Jan 12, 2017 at 03:10:30PM +0200, Saeed Mahameed wrote:
> On Thu, Jan 12, 2017 at 4:09 AM, Martin KaFai Lau <kafai@fb.com> wrote:
> > This patch adds bpf_xdp_adjust_head() support to mlx5e.
>
> Hi Martin, Thanks for the patch !
>
> you can find some comments below.
>
> >
> > 1. rx_headroom is added to struct mlx5e_rq. It uses
> > an existing 4 byte hole in the struct.
> > 2. The adjusted data length is checked against
> > MLX5E_XDP_MIN_INLINE and MLX5E_SW2HW_MTU(rq->netdev->mtu).
> > 3. The macro MLX5E_SW2HW_MTU is moved from en_main.c to en.h.
> > MLX5E_HW2SW_MTU is also moved to en.h for symmetric reason
> > but it is not a must.
> >
> > Signed-off-by: Martin KaFai Lau <kafai@fb.com>
> > ---
> > drivers/net/ethernet/mellanox/mlx5/core/en.h | 4 ++
> > drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 18 +++----
> > drivers/net/ethernet/mellanox/mlx5/core/en_rx.c | 63 ++++++++++++++---------
> > 3 files changed, 51 insertions(+), 34 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
> > index a473cea10c16..0d9dd860a295 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
> > @@ -51,6 +51,9 @@
> >
> > #define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v)
> >
> > +#define MLX5E_HW2SW_MTU(hwmtu) ((hwmtu) - (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN))
> > +#define MLX5E_SW2HW_MTU(swmtu) ((swmtu) + (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN))
> > +
> > #define MLX5E_MAX_NUM_TC 8
> >
> > #define MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE 0x6
> > @@ -369,6 +372,7 @@ struct mlx5e_rq {
> >
> > unsigned long state;
> > int ix;
> > + u16 rx_headroom;
> >
> > struct mlx5e_rx_am am; /* Adaptive Moderation */
> > struct bpf_prog *xdp_prog;
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
> > index f74ba73c55c7..aba3691e0919 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
> > @@ -343,9 +343,6 @@ static void mlx5e_disable_async_events(struct mlx5e_priv *priv)
> > synchronize_irq(mlx5_get_msix_vec(priv->mdev, MLX5_EQ_VEC_ASYNC));
> > }
> >
> > -#define MLX5E_HW2SW_MTU(hwmtu) (hwmtu - (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN))
> > -#define MLX5E_SW2HW_MTU(swmtu) (swmtu + (ETH_HLEN + VLAN_HLEN + ETH_FCS_LEN))
> > -
> > static inline int mlx5e_get_wqe_mtt_sz(void)
> > {
> > /* UMR copies MTTs in units of MLX5_UMR_MTT_ALIGNMENT bytes.
> > @@ -534,9 +531,13 @@ static int mlx5e_create_rq(struct mlx5e_channel *c,
> > goto err_rq_wq_destroy;
> > }
> >
> > - rq->buff.map_dir = DMA_FROM_DEVICE;
> > - if (rq->xdp_prog)
> > + if (rq->xdp_prog) {
> > rq->buff.map_dir = DMA_BIDIRECTIONAL;
> > + rq->rx_headroom = XDP_PACKET_HEADROOM;
> > + } else {
> > + rq->buff.map_dir = DMA_FROM_DEVICE;
> > + rq->rx_headroom = MLX5_RX_HEADROOM;
> > + }
> >
> > switch (priv->params.rq_wq_type) {
> > case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
> > @@ -586,7 +587,7 @@ static int mlx5e_create_rq(struct mlx5e_channel *c,
> > byte_count = rq->buff.wqe_sz;
> >
> > /* calc the required page order */
> > - frag_sz = MLX5_RX_HEADROOM +
> > + frag_sz = rq->rx_headroom +
> > byte_count /* packet data */ +
> > SKB_DATA_ALIGN(sizeof(struct skb_shared_info));
> > frag_sz = SKB_DATA_ALIGN(frag_sz);
> > @@ -3153,11 +3154,6 @@ static int mlx5e_xdp_set(struct net_device *netdev, struct bpf_prog *prog)
> > bool reset, was_opened;
> > int i;
> >
> > - if (prog && prog->xdp_adjust_head) {
> > - netdev_err(netdev, "Does not support bpf_xdp_adjust_head()\n");
> > - return -EOPNOTSUPP;
> > - }
> > -
> > mutex_lock(&priv->state_lock);
> >
> > if ((netdev->features & NETIF_F_LRO) && prog) {
> > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
> > index 0e2fb3ed1790..914e00132e08 100644
> > --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
> > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
> > @@ -264,7 +264,7 @@ int mlx5e_alloc_rx_wqe(struct mlx5e_rq *rq, struct mlx5e_rx_wqe *wqe, u16 ix)
> > if (unlikely(mlx5e_page_alloc_mapped(rq, di)))
> > return -ENOMEM;
> >
> > - wqe->data.addr = cpu_to_be64(di->addr + MLX5_RX_HEADROOM);
> > + wqe->data.addr = cpu_to_be64(di->addr + rq->rx_headroom);
> > return 0;
> > }
> >
> > @@ -646,8 +646,7 @@ static inline void mlx5e_xmit_xdp_doorbell(struct mlx5e_sq *sq)
> >
> > static inline void mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq,
> > struct mlx5e_dma_info *di,
> > - unsigned int data_offset,
> > - int len)
> > + const struct xdp_buff *xdp)
> > {
> > struct mlx5e_sq *sq = &rq->channel->xdp_sq;
> > struct mlx5_wq_cyc *wq = &sq->wq;
> > @@ -659,9 +658,17 @@ static inline void mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq,
> > struct mlx5_wqe_eth_seg *eseg = &wqe->eth;
> > struct mlx5_wqe_data_seg *dseg;
> >
> > + ptrdiff_t data_offset = xdp->data - xdp->data_hard_start;
> > dma_addr_t dma_addr = di->addr + data_offset + MLX5E_XDP_MIN_INLINE;
> > - unsigned int dma_len = len - MLX5E_XDP_MIN_INLINE;
> > - void *data = page_address(di->page) + data_offset;
> > + unsigned int dma_len = xdp->data_end - xdp->data;
> > +
> > + if (unlikely(dma_len < MLX5E_XDP_MIN_INLINE ||
>
> I don't think this can happen, MLX5E_XDP_MIN_INLINE is 18 bytes and
> should not get bigger in the future,
> Also i don't think it is possible for XDP prog to xmit packets smaller
> than 18 bytes, let's remove this
The bpf_xdp_adjust_head() only checks for a min of ETH_HLEN which is 14.
Hence, <18 bytes is still possible here.
>
> > + MLX5E_SW2HW_MTU(rq->netdev->mtu) < dma_len)) {
> > + rq->stats.xdp_drop++;
> > + mlx5e_page_release(rq, di, true);
> > + return;
> > + }
> > + dma_len -= MLX5E_XDP_MIN_INLINE;
>
> Move this to after the below sanity check, it is better to separate
> logic from pre-conditions
Will do.
>
> >
> > if (unlikely(!mlx5e_sq_has_room_for(sq, MLX5E_XDP_TX_WQEBBS))) {
> > if (sq->db.xdp.doorbell) {
> > @@ -680,7 +687,7 @@ static inline void mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq,
> > memset(wqe, 0, sizeof(*wqe));
> >
> > /* copy the inline part */
> > - memcpy(eseg->inline_hdr_start, data, MLX5E_XDP_MIN_INLINE);
> > + memcpy(eseg->inline_hdr_start, xdp->data, MLX5E_XDP_MIN_INLINE);
> > eseg->inline_hdr_sz = cpu_to_be16(MLX5E_XDP_MIN_INLINE);
> >
> > dseg = (struct mlx5_wqe_data_seg *)cseg + (MLX5E_XDP_TX_DS_COUNT - 1);
> > @@ -706,22 +713,16 @@ static inline void mlx5e_xmit_xdp_frame(struct mlx5e_rq *rq,
> > static inline bool mlx5e_xdp_handle(struct mlx5e_rq *rq,
> > const struct bpf_prog *prog,
> > struct mlx5e_dma_info *di,
> > - void *data, u16 len)
> > + struct xdp_buff *xdp)
> > {
> > - struct xdp_buff xdp;
> > u32 act;
> >
> > - if (!prog)
> > - return false;
> > -
> > - xdp.data = data;
> > - xdp.data_end = xdp.data + len;
> > - act = bpf_prog_run_xdp(prog, &xdp);
> > + act = bpf_prog_run_xdp(prog, xdp);
> > switch (act) {
> > case XDP_PASS:
> > return false;
> > case XDP_TX:
> > - mlx5e_xmit_xdp_frame(rq, di, MLX5_RX_HEADROOM, len);
> > + mlx5e_xmit_xdp_frame(rq, di, xdp);
> > return true;
> > default:
> > bpf_warn_invalid_xdp_action(act);
> > @@ -737,18 +738,19 @@ static inline
> > struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
> > u16 wqe_counter, u32 cqe_bcnt)
> > {
> > + const struct bpf_prog *xdp_prog;
> > struct mlx5e_dma_info *di;
> > struct sk_buff *skb;
> > void *va, *data;
> > - bool consumed;
> > + u16 rx_headroom = rq->rx_headroom;
> >
> > di = &rq->dma_info[wqe_counter];
> > va = page_address(di->page);
> > - data = va + MLX5_RX_HEADROOM;
> > + data = va + rx_headroom;
> >
> > dma_sync_single_range_for_cpu(rq->pdev,
> > di->addr,
> > - MLX5_RX_HEADROOM,
> > + rx_headroom,
> > rq->buff.wqe_sz,
> > DMA_FROM_DEVICE);
> > prefetch(data);
> > @@ -760,11 +762,26 @@ struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
> > }
> >
> > rcu_read_lock();
> > - consumed = mlx5e_xdp_handle(rq, READ_ONCE(rq->xdp_prog), di, data,
> > - cqe_bcnt);
> > + xdp_prog = READ_ONCE(rq->xdp_prog);
> > + if (xdp_prog) {
> > + struct xdp_buff xdp;
> > + bool consumed;
> > +
> > + xdp.data = data;
> > + xdp.data_end = xdp.data + cqe_bcnt;
> > + xdp.data_hard_start = va;
> > +
> > + consumed = mlx5e_xdp_handle(rq, xdp_prog, di, &xdp);
> > +
> > + if (consumed) {
> > + rcu_read_unlock();
> > + return NULL; /* page/packet was consumed by XDP */
> > + }
> > +
> > + rx_headroom = xdp.data - xdp.data_hard_start;
> > + cqe_bcnt = xdp.data_end - xdp.data;
> > + }
>
> This whole new logic belongs to mlx5e_xdp_handle, I would like to keep
> xdp related code in one place.
>
> move the xdp_buff initialization back to there and keep the xdp_prog
> check in mlx5e_xdp_handle;
> + xdp_prog = READ_ONCE(rq->xdp_prog);
> + if (!xdp_prog)
> + return false
>
> you can remove "const struct bpf_prog *prog" parameter from
> mlx5e_xdp_handle and take it directly from rq.
>
> if you need va for xdp_buff you can pass it as a paramter to
> mlx5e_xdp_handle as well:
> mlx5e_xdp_handle(rq, di, va, data, cqe_bcnt);
> Make sense ?
I moved them because xdp.data could be adjusted which then
rx_headroom and cqe_bcnt have to be adjusted accordingly
in skb_from_cqe() also.
I understand your point. After another quick thought,
the adjusted xdp.data is the only one that we want in skb_from_cqe().
I will try to make mlx5e_xdp_handle() to return the adjusted xdp.data
instead of bool.
Thanks,
--Martin
>
> > rcu_read_unlock();
> > - if (consumed)
> > - return NULL; /* page/packet was consumed by XDP */
> >
> > skb = build_skb(va, RQ_PAGE_SIZE(rq));
> > if (unlikely(!skb)) {
> > @@ -777,7 +794,7 @@ struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
> > page_ref_inc(di->page);
> > mlx5e_page_release(rq, di, true);
> >
> > - skb_reserve(skb, MLX5_RX_HEADROOM);
> > + skb_reserve(skb, rx_headroom);
> > skb_put(skb, cqe_bcnt);
> >
> > return skb;
> > --
> > 2.5.1
> >
^ permalink raw reply
* Re: [net PATCH 5/5] virtio_net: XDP support for adjust_head
From: John Fastabend @ 2017-01-12 23:26 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: jasowang, john.r.fastabend, netdev, alexei.starovoitov, daniel
In-Reply-To: <20170113002221-mutt-send-email-mst@kernel.org>
On 17-01-12 02:22 PM, Michael S. Tsirkin wrote:
> On Thu, Jan 12, 2017 at 01:45:19PM -0800, John Fastabend wrote:
>> Add support for XDP adjust head by allocating a 256B header region
>> that XDP programs can grow into. This is only enabled when a XDP
>> program is loaded.
>>
>> In order to ensure that we do not have to unwind queue headroom push
>> queue setup below bpf_prog_add. It reads better to do a prog ref
>> unwind vs another queue setup call.
>>
>> At the moment this code must do a full reset to ensure old buffers
>> without headroom on program add or with headroom on program removal
>> are not used incorrectly in the datapath. Ideally we would only
>> have to disable/enable the RX queues being updated but there is no
>> API to do this at the moment in virtio so use the big hammer. In
>> practice it is likely not that big of a problem as this will only
>> happen when XDP is enabled/disabled changing programs does not
>> require the reset. There is some risk that the driver may either
>> have an allocation failure or for some reason fail to correctly
>> negotiate with the underlying backend in this case the driver will
>> be left uninitialized. I have not seen this ever happen on my test
>> systems and for what its worth this same failure case can occur
>> from probe and other contexts in virtio framework.
>
> Could you explain about this a bit more?
> Thanks!
>
Sure. There are two existing paths and this patch adds a third one
where the driver basically goes through this reset path. First one
is on probe the other one is on the freeze/restore path.
The virtnet_freeze() path eventually free's the memory for rq/sq
(receive queues and send queues).
virtnet_freeze()
...
remove_vq_common()
...
virtnet_dev_vqs()
vdev->config->del_vqs()
virtnet_free_queues <- this does a kfree
On virtnet_restore() path we then have to reallocate and reneg with
backend.
virtnet_retore()
...
init_vqs()
...
virtnet_alloc_queues() <- alloc sq/rq
virtnet_find_vqs()
(allocates callbacks/names/vqs)
So the above allocs could fail and leave the device in a FAILED
state. This can happen today on probe or freeze/restore paths and
after this patch possibly on XDP load. Although as noted I have not
seen it happen in any of the above cases.
Second failure mode could happen if virtio_finalize_features() fails.
This seems unlikely because in order to probe successfully we had to
finalize the features successfully earlier. But it could I guess happen
based on return codes. Again never seen this actually happen. This is
called in probe case, freeze/restore case, and XDP now as well.
Does that help? Also I need to send a v2 to fix a spelling mistake and
to convert a 'unsigned' to 'unsigned int' per checkpatch warning. Always
better to run checkpatch before submitting vs after.
Thanks,
John
^ permalink raw reply
* Re: [Patch net] atm: remove an unnecessary loop
From: Francois Romieu @ 2017-01-13 0:07 UTC (permalink / raw)
To: Cong Wang; +Cc: netdev, mhocko, 3chas3, andreyknvl
In-Reply-To: <1484197322-958-1-git-send-email-xiyou.wangcong@gmail.com>
Cong Wang <xiyou.wangcong@gmail.com> :
[...]
> diff --git a/net/atm/common.c b/net/atm/common.c
> index a3ca922..7ec3bbc 100644
> --- a/net/atm/common.c
> +++ b/net/atm/common.c
> @@ -72,10 +72,11 @@ static struct sk_buff *alloc_tx(struct atm_vcc *vcc, unsigned int size)
> sk_wmem_alloc_get(sk), size, sk->sk_sndbuf);
> return NULL;
> }
> - while (!(skb = alloc_skb(size, GFP_KERNEL)))
> - schedule();
> - pr_debug("%d += %d\n", sk_wmem_alloc_get(sk), skb->truesize);
> - atomic_add(skb->truesize, &sk->sk_wmem_alloc);
> + skb = alloc_skb(size, GFP_KERNEL);
> + if (skb) {
> + pr_debug("%d += %d\n", sk_wmem_alloc_get(sk), skb->truesize);
> + atomic_add(skb->truesize, &sk->sk_wmem_alloc);
> + }
> return skb;
> }
Were alloc_skb moved one level up in the call stack, there would be
no need to use the new wait api in the subsequent page, thus easing
pre 3.19 longterm kernel maintenance (at least those on korg page).
But it tastes a tad bit too masochistic.
--
Ueimor
^ permalink raw reply
* Re: [Patch net] atm: remove an unnecessary loop
From: Cong Wang @ 2017-01-13 0:14 UTC (permalink / raw)
To: Francois Romieu
Cc: Linux Kernel Network Developers, Michal Hocko, Chas Williams,
Andrey Konovalov
In-Reply-To: <20170113000700.GA1482@electric-eye.fr.zoreil.com>
On Thu, Jan 12, 2017 at 4:07 PM, Francois Romieu <romieu@fr.zoreil.com> wrote:
> Cong Wang <xiyou.wangcong@gmail.com> :
> [...]
>> diff --git a/net/atm/common.c b/net/atm/common.c
>> index a3ca922..7ec3bbc 100644
>> --- a/net/atm/common.c
>> +++ b/net/atm/common.c
>> @@ -72,10 +72,11 @@ static struct sk_buff *alloc_tx(struct atm_vcc *vcc, unsigned int size)
>> sk_wmem_alloc_get(sk), size, sk->sk_sndbuf);
>> return NULL;
>> }
>> - while (!(skb = alloc_skb(size, GFP_KERNEL)))
>> - schedule();
>> - pr_debug("%d += %d\n", sk_wmem_alloc_get(sk), skb->truesize);
>> - atomic_add(skb->truesize, &sk->sk_wmem_alloc);
>> + skb = alloc_skb(size, GFP_KERNEL);
>> + if (skb) {
>> + pr_debug("%d += %d\n", sk_wmem_alloc_get(sk), skb->truesize);
>> + atomic_add(skb->truesize, &sk->sk_wmem_alloc);
>> + }
>> return skb;
>> }
>
> Were alloc_skb moved one level up in the call stack, there would be
> no need to use the new wait api in the subsequent page, thus easing
> pre 3.19 longterm kernel maintenance (at least those on korg page).
alloc_skb(GFP_KERNEL) itself is sleeping, so the new wait api is still
needed.
^ permalink raw reply
* [PATCH net-next] liquidio: use fallback for selecting txq
From: Felix Manlunas @ 2017-01-13 0:18 UTC (permalink / raw)
To: davem; +Cc: netdev, raghu.vatsavayi, derek.chickles, satananda.burla
From: Satanand Burla <satananda.burla@cavium.com>
Remove assignment to ndo_select_queue so that fallback is used for
selecting txq. Also remove the now-useless function that used to be
assigned to ndo_select_queue.
Signed-off-by: Satanand Burla <satananda.burla@cavium.com>
Signed-off-by: Felix Manlunas <felix.manlunas@cavium.com>
Signed-off-by: Derek Chickles <derek.chickles@cavium.com>
---
drivers/net/ethernet/cavium/liquidio/lio_main.c | 20 --------------------
drivers/net/ethernet/cavium/liquidio/lio_vf_main.c | 21 ---------------------
2 files changed, 41 deletions(-)
diff --git a/drivers/net/ethernet/cavium/liquidio/lio_main.c b/drivers/net/ethernet/cavium/liquidio/lio_main.c
index cc825d5..2b89ec2 100644
--- a/drivers/net/ethernet/cavium/liquidio/lio_main.c
+++ b/drivers/net/ethernet/cavium/liquidio/lio_main.c
@@ -2223,25 +2223,6 @@ static void if_cfg_callback(struct octeon_device *oct,
wake_up_interruptible(&ctx->wc);
}
-/**
- * \brief Select queue based on hash
- * @param dev Net device
- * @param skb sk_buff structure
- * @returns selected queue number
- */
-static u16 select_q(struct net_device *dev, struct sk_buff *skb,
- void *accel_priv __attribute__((unused)),
- select_queue_fallback_t fallback __attribute__((unused)))
-{
- u32 qindex = 0;
- struct lio *lio;
-
- lio = GET_LIO(dev);
- qindex = skb_tx_hash(dev, skb);
-
- return (u16)(qindex % (lio->linfo.num_txpciq));
-}
-
/** Routine to push packets arriving on Octeon interface upto network layer.
* @param oct_id - octeon device id.
* @param skbuff - skbuff struct to be passed to network layer.
@@ -3755,7 +3736,6 @@ static const struct net_device_ops lionetdevops = {
.ndo_set_vf_vlan = liquidio_set_vf_vlan,
.ndo_get_vf_config = liquidio_get_vf_config,
.ndo_set_vf_link_state = liquidio_set_vf_link_state,
- .ndo_select_queue = select_q
};
/** \brief Entry point for the liquidio module
diff --git a/drivers/net/ethernet/cavium/liquidio/lio_vf_main.c b/drivers/net/ethernet/cavium/liquidio/lio_vf_main.c
index ad2e72d..19d88fb 100644
--- a/drivers/net/ethernet/cavium/liquidio/lio_vf_main.c
+++ b/drivers/net/ethernet/cavium/liquidio/lio_vf_main.c
@@ -1455,26 +1455,6 @@ static void if_cfg_callback(struct octeon_device *oct,
wake_up_interruptible(&ctx->wc);
}
-/**
- * \brief Select queue based on hash
- * @param dev Net device
- * @param skb sk_buff structure
- * @returns selected queue number
- */
-static u16 select_q(struct net_device *dev, struct sk_buff *skb,
- void *accel_priv __attribute__((unused)),
- select_queue_fallback_t fallback __attribute__((unused)))
-{
- struct lio *lio;
- u32 qindex;
-
- lio = GET_LIO(dev);
-
- qindex = skb_tx_hash(dev, skb);
-
- return (u16)(qindex % (lio->linfo.num_txpciq));
-}
-
/** Routine to push packets arriving on Octeon interface upto network layer.
* @param oct_id - octeon device id.
* @param skbuff - skbuff struct to be passed to network layer.
@@ -2717,7 +2697,6 @@ static const struct net_device_ops lionetdevops = {
.ndo_set_features = liquidio_set_features,
.ndo_udp_tunnel_add = liquidio_add_vxlan_port,
.ndo_udp_tunnel_del = liquidio_del_vxlan_port,
- .ndo_select_queue = select_q,
};
static int lio_nic_info(struct octeon_recv_info *recv_info, void *buf)
^ permalink raw reply related
* [PATCH net] openvswitch: maintain correct checksum state in conntrack actions
From: Lance Richardson @ 2017-01-13 0:33 UTC (permalink / raw)
To: netdev, dev
When executing conntrack actions on skbuffs with checksum mode
CHECKSUM_COMPLETE, the checksum must be updated to account for
header pushes and pulls. Otherwise we get "hw csum failure"
logs similar to this (ICMP packet received on geneve tunnel
via ixgbe NIC):
[ 405.740065] genev_sys_6081: hw csum failure
[ 405.740106] CPU: 3 PID: 0 Comm: swapper/3 Tainted: G I 4.10.0-rc3+ #1
[ 405.740108] Call Trace:
[ 405.740110] <IRQ>
[ 405.740113] dump_stack+0x63/0x87
[ 405.740116] netdev_rx_csum_fault+0x3a/0x40
[ 405.740118] __skb_checksum_complete+0xcf/0xe0
[ 405.740120] nf_ip_checksum+0xc8/0xf0
[ 405.740124] icmp_error+0x1de/0x351 [nf_conntrack_ipv4]
[ 405.740132] nf_conntrack_in+0xe1/0x550 [nf_conntrack]
[ 405.740137] ? find_bucket.isra.2+0x62/0x70 [openvswitch]
[ 405.740143] __ovs_ct_lookup+0x95/0x980 [openvswitch]
[ 405.740145] ? netif_rx_internal+0x44/0x110
[ 405.740149] ovs_ct_execute+0x147/0x4b0 [openvswitch]
[ 405.740153] do_execute_actions+0x22e/0xa70 [openvswitch]
[ 405.740157] ovs_execute_actions+0x40/0x120 [openvswitch]
[ 405.740161] ovs_dp_process_packet+0x84/0x120 [openvswitch]
[ 405.740166] ovs_vport_receive+0x73/0xd0 [openvswitch]
[ 405.740168] ? udp_rcv+0x1a/0x20
[ 405.740170] ? ip_local_deliver_finish+0x93/0x1e0
[ 405.740172] ? ip_local_deliver+0x6f/0xe0
[ 405.740174] ? ip_rcv_finish+0x3a0/0x3a0
[ 405.740176] ? ip_rcv_finish+0xdb/0x3a0
[ 405.740177] ? ip_rcv+0x2a7/0x400
[ 405.740180] ? __netif_receive_skb_core+0x970/0xa00
[ 405.740185] netdev_frame_hook+0xd3/0x160 [openvswitch]
[ 405.740187] __netif_receive_skb_core+0x1dc/0xa00
[ 405.740194] ? ixgbe_clean_rx_irq+0x46d/0xa20 [ixgbe]
[ 405.740197] __netif_receive_skb+0x18/0x60
[ 405.740199] netif_receive_skb_internal+0x40/0xb0
[ 405.740201] napi_gro_receive+0xcd/0x120
[ 405.740204] gro_cell_poll+0x57/0x80 [geneve]
[ 405.740206] net_rx_action+0x260/0x3c0
[ 405.740209] __do_softirq+0xc9/0x28c
[ 405.740211] irq_exit+0xd9/0xf0
[ 405.740213] do_IRQ+0x51/0xd0
[ 405.740215] common_interrupt+0x93/0x93
Fixes: 7f8a436eaa2c ("openvswitch: Add conntrack action")
Signed-off-by: Lance Richardson <lrichard@redhat.com>
---
net/openvswitch/conntrack.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/net/openvswitch/conntrack.c b/net/openvswitch/conntrack.c
index 6b78bab..54253ea 100644
--- a/net/openvswitch/conntrack.c
+++ b/net/openvswitch/conntrack.c
@@ -514,7 +514,7 @@ static int ovs_ct_nat_execute(struct sk_buff *skb, struct nf_conn *ct,
int hooknum, nh_off, err = NF_ACCEPT;
nh_off = skb_network_offset(skb);
- skb_pull(skb, nh_off);
+ skb_pull_rcsum(skb, nh_off);
/* See HOOK2MANIP(). */
if (maniptype == NF_NAT_MANIP_SRC)
@@ -579,6 +579,7 @@ static int ovs_ct_nat_execute(struct sk_buff *skb, struct nf_conn *ct,
err = nf_nat_packet(ct, ctinfo, hooknum, skb);
push:
skb_push(skb, nh_off);
+ skb_postpush_rcsum(skb, skb->data, nh_off);
return err;
}
@@ -886,7 +887,7 @@ int ovs_ct_execute(struct net *net, struct sk_buff *skb,
/* The conntrack module expects to be working at L3. */
nh_ofs = skb_network_offset(skb);
- skb_pull(skb, nh_ofs);
+ skb_pull_rcsum(skb, nh_ofs);
if (key->ip.frag != OVS_FRAG_TYPE_NONE) {
err = handle_fragments(net, key, info->zone.id, skb);
@@ -900,6 +901,7 @@ int ovs_ct_execute(struct net *net, struct sk_buff *skb,
err = ovs_ct_lookup(net, key, info, skb);
skb_push(skb, nh_ofs);
+ skb_postpush_rcsum(skb, skb->data, nh_ofs);
if (err)
kfree_skb(skb);
return err;
--
1.8.3.1
^ permalink raw reply related
* [net PATCH v2 0/5] virtio_net XDP fixes and adjust_header support
From: John Fastabend @ 2017-01-13 0:34 UTC (permalink / raw)
To: jasowang, mst
Cc: john.r.fastabend, netdev, john.fastabend, alexei.starovoitov,
daniel
This has a fix to handle small buffer free logic correctly and then
also adds adjust head support.
I pushed adjust head at net (even though its rc3) to avoid having
to push another exception case into virtio_net to catch if the
program uses adjust_head and then block it. If there are any strong
objections to this we can push it at net-next and use a patch from
Jakub to add the exception handling but then user space has to deal
with it either via try/fail logic or via kernel version checks. Granted
we already have some cases that need to be configured to enable XDP
but I don't see any reason to have yet another one when we can fix it
now vs delaying a kernel version.
v2: fix spelling error, convert unsigned -> unsigned int
---
John Fastabend (5):
virtio_net: use dev_kfree_skb for small buffer XDP receive
net: virtio: wrap rtnl_lock in test for calling with lock already held
virtio_net: factor out xdp handler for readability
virtio_net: remove duplicate queue pair binding in XDP
virtio_net: XDP support for adjust_head
drivers/net/virtio_net.c | 251 ++++++++++++++++++++++++++++++++--------------
drivers/virtio/virtio.c | 9 +-
include/linux/virtio.h | 3 +
3 files changed, 183 insertions(+), 80 deletions(-)
^ permalink raw reply
* [net PATCH v2 1/5] virtio_net: use dev_kfree_skb for small buffer XDP receive
From: John Fastabend @ 2017-01-13 0:35 UTC (permalink / raw)
To: jasowang, mst
Cc: john.r.fastabend, netdev, john.fastabend, alexei.starovoitov,
daniel
In-Reply-To: <20170113003412.4167.12250.stgit@john-Precision-Tower-5810>
In the small buffer case during driver unload we currently use
put_page instead of dev_kfree_skb. Resolve this by adding a check
for virtnet mode when checking XDP queue type. Also name the
function so that the code reads correctly to match the additional
check.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
---
drivers/net/virtio_net.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 4a10500..d97bb71 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -1890,8 +1890,12 @@ static void free_receive_page_frags(struct virtnet_info *vi)
put_page(vi->rq[i].alloc_frag.page);
}
-static bool is_xdp_queue(struct virtnet_info *vi, int q)
+static bool is_xdp_raw_buffer_queue(struct virtnet_info *vi, int q)
{
+ /* For small receive mode always use kfree_skb variants */
+ if (!vi->mergeable_rx_bufs)
+ return false;
+
if (q < (vi->curr_queue_pairs - vi->xdp_queue_pairs))
return false;
else if (q < vi->curr_queue_pairs)
@@ -1908,7 +1912,7 @@ static void free_unused_bufs(struct virtnet_info *vi)
for (i = 0; i < vi->max_queue_pairs; i++) {
struct virtqueue *vq = vi->sq[i].vq;
while ((buf = virtqueue_detach_unused_buf(vq)) != NULL) {
- if (!is_xdp_queue(vi, i))
+ if (!is_xdp_raw_buffer_queue(vi, i))
dev_kfree_skb(buf);
else
put_page(virt_to_head_page(buf));
^ permalink raw reply related
* [net PATCH v2 2/5] net: virtio: wrap rtnl_lock in test for calling with lock already held
From: John Fastabend @ 2017-01-13 0:35 UTC (permalink / raw)
To: jasowang, mst
Cc: john.r.fastabend, netdev, john.fastabend, alexei.starovoitov,
daniel
In-Reply-To: <20170113003412.4167.12250.stgit@john-Precision-Tower-5810>
For XDP use case and to allow ethtool reset tests it is useful to be
able to use reset routines from contexts where rtnl lock is already
held.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
---
drivers/net/virtio_net.c | 16 +++++++++-------
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index d97bb71..43cb2e0 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -1864,12 +1864,13 @@ static void virtnet_free_queues(struct virtnet_info *vi)
kfree(vi->sq);
}
-static void free_receive_bufs(struct virtnet_info *vi)
+static void free_receive_bufs(struct virtnet_info *vi, bool need_lock)
{
struct bpf_prog *old_prog;
int i;
- rtnl_lock();
+ if (need_lock)
+ rtnl_lock();
for (i = 0; i < vi->max_queue_pairs; i++) {
while (vi->rq[i].pages)
__free_pages(get_a_page(&vi->rq[i], GFP_KERNEL), 0);
@@ -1879,7 +1880,8 @@ static void free_receive_bufs(struct virtnet_info *vi)
if (old_prog)
bpf_prog_put(old_prog);
}
- rtnl_unlock();
+ if (need_lock)
+ rtnl_unlock();
}
static void free_receive_page_frags(struct virtnet_info *vi)
@@ -2351,14 +2353,14 @@ static int virtnet_probe(struct virtio_device *vdev)
return err;
}
-static void remove_vq_common(struct virtnet_info *vi)
+static void remove_vq_common(struct virtnet_info *vi, bool lock)
{
vi->vdev->config->reset(vi->vdev);
/* Free unused buffers in both send and recv, if any. */
free_unused_bufs(vi);
- free_receive_bufs(vi);
+ free_receive_bufs(vi, lock);
free_receive_page_frags(vi);
@@ -2376,7 +2378,7 @@ static void virtnet_remove(struct virtio_device *vdev)
unregister_netdev(vi->dev);
- remove_vq_common(vi);
+ remove_vq_common(vi, true);
free_percpu(vi->stats);
free_netdev(vi->dev);
@@ -2401,7 +2403,7 @@ static int virtnet_freeze(struct virtio_device *vdev)
napi_disable(&vi->rq[i].napi);
}
- remove_vq_common(vi);
+ remove_vq_common(vi, true);
return 0;
}
^ permalink raw reply related
* Re: [PATCH v2 7/7] uapi: export all headers under uapi directories
From: Jeff Epler @ 2017-01-13 1:04 UTC (permalink / raw)
To: Nicolas Dichtel
Cc: linux-mips, alsa-devel, linux-ia64, linux-doc, airlied,
linux-fbdev, dri-devel, linux-mtd, sparclinux, linux-arch,
linux-s390, linux-am33-list, linux-c6x-dev, linux-rdma,
linux-hexagon, linux-sh, Christoph Hellwig, coreteam, fcoe-devel,
xen-devel, linux-snps-arc, linux-media, uclinux-h8-devel,
linux-xtensa, arnd, linux-kbuild, adi-buildroot-devel, linux-raid,
linux-m68k, mmarek, linux-metag
In-Reply-To: <9d68af8a-a609-d7b1-58a9-f1155313b077@6wind.com>
On Thu, Jan 12, 2017 at 05:32:09PM +0100, Nicolas Dichtel wrote:
> What I was trying to say is that I export those directories like other are.
> Removing those files is not related to that series.
Perhaps the correct solution is to only copy files matching "*.h" to
reduce the risk of copying files incidentally created by kbuild but
which shouldn't be installed as uapi headers.
jeff
^ permalink raw reply
* [PATCH] i40e: Invoke softirqs after napi_reschedule
From: Benjamin Poirier @ 2017-01-13 1:04 UTC (permalink / raw)
To: Jeff Kirsher; +Cc: intel-wired-lan, netdev
The following message is logged from time to time when using i40e:
NOHZ: local_softirq_pending 08
i40e may schedule napi from a workqueue. Afterwards, softirqs are not run
in a deterministic time frame. The problem is the same as what was
described in commit ec13ee80145c ("virtio_net: invoke softirqs after
__napi_schedule") and this patch applies the same fix to i40e.
Signed-off-by: Benjamin Poirier <bpoirier@suse.com>
---
drivers/net/ethernet/intel/i40e/i40e_main.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c
index ad4cf63..d65488c 100644
--- a/drivers/net/ethernet/intel/i40e/i40e_main.c
+++ b/drivers/net/ethernet/intel/i40e/i40e_main.c
@@ -4621,8 +4621,10 @@ static void i40e_detect_recover_hung_queue(int q_idx, struct i40e_vsi *vsi)
*/
if ((!tx_pending_hw) && i40e_get_tx_pending(tx_ring, true) &&
(!(val & I40E_PFINT_DYN_CTLN_INTENA_MASK))) {
+ local_bh_disable();
if (napi_reschedule(&tx_ring->q_vector->napi))
tx_ring->tx_stats.tx_lost_interrupt++;
+ local_bh_enable();
}
}
--
2.10.2
^ 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