* Re: [PATCH net-next] net: sock_rps_record_flow() is for connected sockets
From: Eric Dumazet @ 2016-12-07 14:26 UTC (permalink / raw)
To: Paolo Abeni; +Cc: David Miller, netdev, Willem de Bruijn, Tom Herbert
In-Reply-To: <1481097428.5535.12.camel@redhat.com>
On Wed, 2016-12-07 at 08:57 +0100, Paolo Abeni wrote:
> On Tue, 2016-12-06 at 22:47 -0800, Eric Dumazet wrote:
> > On Tue, 2016-12-06 at 19:32 -0800, Eric Dumazet wrote:
> > > A follow up patch will provide a static_key (Jump Label) since most
> > > hosts do not even use RFS.
> >
> > Speaking of static_key, it appears we now have GRO on UDP, and this
> > consumes a considerable amount of cpu cycles.
> >
> > Turning off GRO allows me to get +20 % more packets on my single UDP
> > socket. (1.2 Mpps instead of 1.0 Mpps)
>
> I see also an improvement for single flow tests disabling GRO, but on a
> smaller scale (~5% if I recall correctly).
Was it on a NUMA host ?
My tests are spreading packets on 8 RX queues, 50/50 split on two NUMA
nodes.
^ permalink raw reply
* Re: [PATCH 1/1] NET: usb: cdc_mbim: add quirk for supporting Telit LE922A
From: Bjørn Mork @ 2016-12-07 14:27 UTC (permalink / raw)
To: Daniele Palmas; +Cc: Oliver Neukum, netdev, linux-usb
In-Reply-To: <1481116068-32691-2-git-send-email-dnlplm@gmail.com>
Daniele Palmas <dnlplm@gmail.com> writes:
> Telit LE922A MBIM based composition does not work properly
> with altsetting toggle done in cdc_ncm_bind_common.
>
> This patch adds CDC_MBIM_FLAG_AVOID_ALTSETTING_TOGGLE quirk
> to avoid this procedure that, instead, is mandatory for
> other modems.
>
> Signed-off-by: Daniele Palmas <dnlplm@gmail.com>
Thanks. This looks as nice as we can possibly do for now. I am happy
to add
Reviewed-by: Bjørn Mork <bjorn@mork.no>
This patch also helped me understanding the problem you face even
better: It's not the "set_altsetting 0" that's causing problems, but the
"set_altsetting X" prior to that. And that is certainly not something
that a firmware should expect. I will look at the Huawei modem again
when I get some time and see if there is any way we can possibly avoid
it.
Bjørn
^ permalink raw reply
* Re: [PATCH net-next] net: sock_rps_record_flow() is for connected sockets
From: Eric Dumazet @ 2016-12-07 14:29 UTC (permalink / raw)
To: Paolo Abeni; +Cc: David Miller, netdev, Willem de Bruijn, Tom Herbert
In-Reply-To: <1481097428.5535.12.camel@redhat.com>
On Wed, 2016-12-07 at 08:57 +0100, Paolo Abeni wrote:
> We have some experimental patches to implement GRO for plain UDP
> connected sockets, using frag_list to preserve the individual skb len,
> and deliver the packet to user space individually. With that I got
> ~3mpps with a single queue/user space sink - before the recent udp
> improvements. I would like to present these patches on netdev soon (no
> sooner than next week, anyway).
>
Make sure you handle properly all netfilter helpers :(
Keeping frag_list means you keep one sk_buff per segment, so this really
looks like a legacy UDP server (like a DNS server) wont benefit from
this anyway.
^ permalink raw reply
* Re: [PATCH 3/4] vsock: add pkt cancel capability
From: Peng Tao @ 2016-12-07 14:46 UTC (permalink / raw)
To: Stefan Hajnoczi
Cc: Stefan Hajnoczi, kvm, virtualization, netdev@vger.kernel.org
In-Reply-To: <20161207132206.GC31552@stefanha-x1.localdomain>
On Wed, Dec 7, 2016 at 9:22 PM, Stefan Hajnoczi <stefanha@gmail.com> wrote:
> On Wed, Dec 07, 2016 at 06:00:20PM +0800, Peng Tao wrote:
>> Signed-off-by: Peng Tao <bergwolf@gmail.com>
>> ---
>> net/vmw_vsock/virtio_transport.c | 36 ++++++++++++++++++++++++++++++++++++
>> 1 file changed, 36 insertions(+)
>>
>> diff --git a/net/vmw_vsock/virtio_transport.c b/net/vmw_vsock/virtio_transport.c
>> index 936d7ee..f88b6ed 100644
>> --- a/net/vmw_vsock/virtio_transport.c
>> +++ b/net/vmw_vsock/virtio_transport.c
>> @@ -170,6 +170,41 @@ virtio_transport_send_pkt(struct virtio_vsock_pkt *pkt)
>> return len;
>> }
>>
>> +static int
>> +virtio_transport_cancel_pkt(struct vsock_sock *vsk)
>> +{
>> + struct virtio_vsock *vsock;
>> + struct virtio_vsock_pkt *pkt, *n;
>> + int cnt = 0;
>> + LIST_HEAD(freeme);
>> +
>> + vsock = virtio_vsock_get();
>> + if (!vsock) {
>> + return -ENODEV;
>> + }
>> +
>> + if (pkt->reply)
>
> pkt is uninitialized. I guess this if statement should be deleted, you
> already take care of counting reply packets below.
>
ah, sorry! I forgot to delete it after handling cnt below...
>> + cnt++;
>> +
>> + spin_lock_bh(&vsock->send_pkt_list_lock);
>> + list_for_each_entry_safe(pkt, n, &vsock->send_pkt_list, list) {
>> + if (pkt->vsk != vsk)
>> + continue;
>> + list_move(&pkt->list, &freeme);
>> + }
>> + spin_unlock_bh(&vsock->send_pkt_list_lock);
>> +
>> + list_for_each_entry_safe(pkt, n, &freeme, list) {
>> + if (pkt->reply)
>> + cnt++;
>> + list_del(&pkt->list);
>> + virtio_transport_free_pkt(pkt);
>> + }
>> + atomic_sub(cnt, &vsock->queued_replies);
>
> If we stopped rx because there were too many replies in flight then we
> might be able to resume rx now:
>
> /* Do we now have resources to resume rx processing? */
> if (old_val >= virtqueue_get_vring_size(rx_vq) &&
> new_val < virtqueue_get_vring_size(rx_vq))
> queue_work(virtio_vsock_workqueue, &vsock->rx_work);
Thanks! I totally missed the resume part... I'll send an updated version later.
Cheers,
Tao
^ permalink raw reply
* [net-next] macsec: remove first zero and add attribute name in comments
From: Zhang Shengju @ 2016-12-07 15:02 UTC (permalink / raw)
To: netdev
Remove first zero for add, and use full attribute name in comments.
Signed-off-by: Zhang Shengju <zhangshengju@cmss.chinamobile.com>
---
drivers/net/macsec.c | 25 ++++++++++++-------------
1 file changed, 12 insertions(+), 13 deletions(-)
diff --git a/drivers/net/macsec.c b/drivers/net/macsec.c
index cc00eb0..f83cf66 100644
--- a/drivers/net/macsec.c
+++ b/drivers/net/macsec.c
@@ -3346,19 +3346,18 @@ static struct net *macsec_get_link_net(const struct net_device *dev)
static size_t macsec_get_size(const struct net_device *dev)
{
- return 0 +
- nla_total_size_64bit(8) + /* SCI */
- nla_total_size(1) + /* ICV_LEN */
- nla_total_size_64bit(8) + /* CIPHER_SUITE */
- nla_total_size(4) + /* WINDOW */
- nla_total_size(1) + /* ENCODING_SA */
- nla_total_size(1) + /* ENCRYPT */
- nla_total_size(1) + /* PROTECT */
- nla_total_size(1) + /* INC_SCI */
- nla_total_size(1) + /* ES */
- nla_total_size(1) + /* SCB */
- nla_total_size(1) + /* REPLAY_PROTECT */
- nla_total_size(1) + /* VALIDATION */
+ return nla_total_size_64bit(8) + /* IFLA_MACSEC_SCI */
+ nla_total_size(1) + /* IFLA_MACSEC_ICV_LEN */
+ nla_total_size_64bit(8) + /* IFLA_MACSEC_CIPHER_SUITE */
+ nla_total_size(4) + /* IFLA_MACSEC_WINDOW */
+ nla_total_size(1) + /* IFLA_MACSEC_ENCODING_SA */
+ nla_total_size(1) + /* IFLA_MACSEC_ENCRYPT */
+ nla_total_size(1) + /* IFLA_MACSEC_PROTECT */
+ nla_total_size(1) + /* IFLA_MACSEC_INC_SCI */
+ nla_total_size(1) + /* IFLA_MACSEC_ES */
+ nla_total_size(1) + /* IFLA_MACSEC_SCB */
+ nla_total_size(1) + /* IFLA_MACSEC_REPLAY_PROTECT */
+ nla_total_size(1) + /* IFLA_MACSEC_VALIDATION */
0;
}
--
1.8.3.1
^ permalink raw reply related
* [PATCH v2 2/4] vhost-vsock: add pkt cancel capability
From: Peng Tao @ 2016-12-07 15:14 UTC (permalink / raw)
To: Stefan Hajnoczi; +Cc: kvm, virtualization, netdev, Peng Tao
In-Reply-To: <1481123652-80603-1-git-send-email-bergwolf@gmail.com>
To allow canceling all packets of a connection.
Signed-off-by: Peng Tao <bergwolf@gmail.com>
---
drivers/vhost/vsock.c | 41 +++++++++++++++++++++++++++++++++++++++++
include/linux/virtio_vsock.h | 3 +++
2 files changed, 44 insertions(+)
diff --git a/drivers/vhost/vsock.c b/drivers/vhost/vsock.c
index a504e2e0..d01e4a4 100644
--- a/drivers/vhost/vsock.c
+++ b/drivers/vhost/vsock.c
@@ -218,6 +218,46 @@ vhost_transport_send_pkt(struct virtio_vsock_pkt *pkt)
return len;
}
+static int
+vhost_transport_cancel_pkt(struct vsock_sock *vsk)
+{
+ struct vhost_vsock *vsock;
+ struct virtio_vsock_pkt *pkt, *n;
+ int cnt = 0;
+ LIST_HEAD(freeme);
+
+ /* Find the vhost_vsock according to guest context id */
+ vsock = vhost_vsock_get(vsk->remote_addr.svm_cid);
+ if (!vsock)
+ return -ENODEV;
+
+ spin_lock_bh(&vsock->send_pkt_list_lock);
+ list_for_each_entry_safe(pkt, n, &vsock->send_pkt_list, list) {
+ if (pkt->vsk != vsk)
+ continue;
+ list_move(&pkt->list, &freeme);
+ }
+ spin_unlock_bh(&vsock->send_pkt_list_lock);
+
+ list_for_each_entry_safe(pkt, n, &freeme, list) {
+ if (pkt->reply)
+ cnt++;
+ list_del(&pkt->list);
+ virtio_transport_free_pkt(pkt);
+ }
+
+ if (cnt) {
+ struct vhost_virtqueue *tx_vq = &vsock->vqs[VSOCK_VQ_TX];
+ int new_cnt;
+
+ new_cnt = atomic_sub_return(cnt, &vsock->queued_replies);
+ if (new_cnt + cnt >= tx_vq->num && new_cnt < tx_vq->num)
+ vhost_poll_queue(&tx_vq->poll);
+ }
+
+ return 0;
+}
+
static struct virtio_vsock_pkt *
vhost_vsock_alloc_pkt(struct vhost_virtqueue *vq,
unsigned int out, unsigned int in)
@@ -698,6 +738,7 @@ static struct virtio_transport vhost_transport = {
},
.send_pkt = vhost_transport_send_pkt,
+ .cancel_pkt = vhost_transport_cancel_pkt,
};
static int __init vhost_vsock_init(void)
diff --git a/include/linux/virtio_vsock.h b/include/linux/virtio_vsock.h
index 6dd3242..b92e88d 100644
--- a/include/linux/virtio_vsock.h
+++ b/include/linux/virtio_vsock.h
@@ -72,6 +72,9 @@ struct virtio_transport {
/* Takes ownership of the packet */
int (*send_pkt)(struct virtio_vsock_pkt *pkt);
+
+ /* Cancel packets belonging the same vsock */
+ int (*cancel_pkt)(struct vsock_sock *vsk);
};
ssize_t
--
2.7.4
^ permalink raw reply related
* [PATCH v2 3/4] vsock: add pkt cancel capability
From: Peng Tao @ 2016-12-07 15:14 UTC (permalink / raw)
To: Stefan Hajnoczi; +Cc: kvm, virtualization, netdev, Peng Tao
In-Reply-To: <1481123652-80603-1-git-send-email-bergwolf@gmail.com>
Signed-off-by: Peng Tao <bergwolf@gmail.com>
---
net/vmw_vsock/virtio_transport.c | 42 ++++++++++++++++++++++++++++++++++++++++
1 file changed, 42 insertions(+)
diff --git a/net/vmw_vsock/virtio_transport.c b/net/vmw_vsock/virtio_transport.c
index 936d7ee..a5f3833 100644
--- a/net/vmw_vsock/virtio_transport.c
+++ b/net/vmw_vsock/virtio_transport.c
@@ -170,6 +170,47 @@ virtio_transport_send_pkt(struct virtio_vsock_pkt *pkt)
return len;
}
+static int
+virtio_transport_cancel_pkt(struct vsock_sock *vsk)
+{
+ struct virtio_vsock *vsock;
+ struct virtio_vsock_pkt *pkt, *n;
+ int cnt = 0;
+ LIST_HEAD(freeme);
+
+ vsock = virtio_vsock_get();
+ if (!vsock) {
+ return -ENODEV;
+ }
+
+ spin_lock_bh(&vsock->send_pkt_list_lock);
+ list_for_each_entry_safe(pkt, n, &vsock->send_pkt_list, list) {
+ if (pkt->vsk != vsk)
+ continue;
+ list_move(&pkt->list, &freeme);
+ }
+ spin_unlock_bh(&vsock->send_pkt_list_lock);
+
+ list_for_each_entry_safe(pkt, n, &freeme, list) {
+ if (pkt->reply)
+ cnt++;
+ list_del(&pkt->list);
+ virtio_transport_free_pkt(pkt);
+ }
+
+ if (cnt) {
+ struct virtqueue *rx_vq = vsock->vqs[VSOCK_VQ_RX];
+ int new_cnt;
+
+ new_cnt = atomic_sub_return(cnt, &vsock->queued_replies);
+ if (new_cnt + cnt >= virtqueue_get_vring_size(rx_vq) &&
+ new_cnt < virtqueue_get_vring_size(rx_vq))
+ queue_work(virtio_vsock_workqueue, &vsock->rx_work);
+ }
+
+ return 0;
+}
+
static void virtio_vsock_rx_fill(struct virtio_vsock *vsock)
{
int buf_len = VIRTIO_VSOCK_DEFAULT_RX_BUF_SIZE;
@@ -453,6 +494,7 @@ static struct virtio_transport virtio_transport = {
},
.send_pkt = virtio_transport_send_pkt,
+ .cancel_pkt = virtio_transport_cancel_pkt,
};
static int virtio_vsock_probe(struct virtio_device *vdev)
--
2.7.4
^ permalink raw reply related
* [PATCH v2 4/4] vsock: cancel packets when failing to connect
From: Peng Tao @ 2016-12-07 15:14 UTC (permalink / raw)
To: Stefan Hajnoczi; +Cc: kvm, virtualization, netdev, Peng Tao
In-Reply-To: <1481123652-80603-1-git-send-email-bergwolf@gmail.com>
Otherwise we'll leave the packets queued until releasing vsock device.
E.g., if guest is slow to start up, resulting ETIMEDOUT on connect, guest
will get the connect requests from failed host sockets.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Peng Tao <bergwolf@gmail.com>
---
include/linux/virtio_vsock.h | 7 +++++++
net/vmw_vsock/af_vsock.c | 7 +++++++
net/vmw_vsock/virtio_transport_common.c | 7 -------
3 files changed, 14 insertions(+), 7 deletions(-)
diff --git a/include/linux/virtio_vsock.h b/include/linux/virtio_vsock.h
index b92e88d..ff6850a 100644
--- a/include/linux/virtio_vsock.h
+++ b/include/linux/virtio_vsock.h
@@ -156,4 +156,11 @@ void virtio_transport_inc_tx_pkt(struct virtio_vsock_sock *vvs, struct virtio_vs
u32 virtio_transport_get_credit(struct virtio_vsock_sock *vvs, u32 wanted);
void virtio_transport_put_credit(struct virtio_vsock_sock *vvs, u32 credit);
+static inline const struct virtio_transport *virtio_transport_get_ops(void)
+{
+ const struct vsock_transport *t = vsock_core_get_transport();
+
+ return container_of(t, struct virtio_transport, transport);
+}
+
#endif /* _LINUX_VIRTIO_VSOCK_H */
diff --git a/net/vmw_vsock/af_vsock.c b/net/vmw_vsock/af_vsock.c
index 8a398b3..ebb50d6 100644
--- a/net/vmw_vsock/af_vsock.c
+++ b/net/vmw_vsock/af_vsock.c
@@ -104,6 +104,7 @@
#include <linux/unistd.h>
#include <linux/wait.h>
#include <linux/workqueue.h>
+#include <linux/virtio_vsock.h>
#include <net/sock.h>
#include <net/af_vsock.h>
@@ -1105,6 +1106,7 @@ static void vsock_connect_timeout(struct work_struct *work)
{
struct sock *sk;
struct vsock_sock *vsk;
+ int cancel = 0;
vsk = container_of(work, struct vsock_sock, dwork.work);
sk = sk_vsock(vsk);
@@ -1115,8 +1117,11 @@ static void vsock_connect_timeout(struct work_struct *work)
sk->sk_state = SS_UNCONNECTED;
sk->sk_err = ETIMEDOUT;
sk->sk_error_report(sk);
+ cancel = 1;
}
release_sock(sk);
+ if (cancel)
+ virtio_transport_get_ops()->cancel_pkt(vsk);
sock_put(sk);
}
@@ -1223,11 +1228,13 @@ static int vsock_stream_connect(struct socket *sock, struct sockaddr *addr,
err = sock_intr_errno(timeout);
sk->sk_state = SS_UNCONNECTED;
sock->state = SS_UNCONNECTED;
+ virtio_transport_get_ops()->cancel_pkt(vsk);
goto out_wait;
} else if (timeout == 0) {
err = -ETIMEDOUT;
sk->sk_state = SS_UNCONNECTED;
sock->state = SS_UNCONNECTED;
+ virtio_transport_get_ops()->cancel_pkt(vsk);
goto out_wait;
}
diff --git a/net/vmw_vsock/virtio_transport_common.c b/net/vmw_vsock/virtio_transport_common.c
index cc1eeb5..72c5dff 100644
--- a/net/vmw_vsock/virtio_transport_common.c
+++ b/net/vmw_vsock/virtio_transport_common.c
@@ -25,13 +25,6 @@
/* How long to wait for graceful shutdown of a connection */
#define VSOCK_CLOSE_TIMEOUT (8 * HZ)
-static const struct virtio_transport *virtio_transport_get_ops(void)
-{
- const struct vsock_transport *t = vsock_core_get_transport();
-
- return container_of(t, struct virtio_transport, transport);
-}
-
struct virtio_vsock_pkt *
virtio_transport_alloc_pkt(struct virtio_vsock_pkt_info *info,
size_t len,
--
2.7.4
^ permalink raw reply related
* [PATCH v2 0/4] vsock: cancel connect packets when failing to connect
From: Peng Tao @ 2016-12-07 15:14 UTC (permalink / raw)
To: Stefan Hajnoczi; +Cc: kvm, virtualization, netdev, Peng Tao
Currently, if a connect call fails on a signal or timeout (e.g., guest is still
in the process of starting up), we'll just return to caller and leave the connect
packet queued and they are sent even though the connection is considered a failure,
which can confuse applications with unwanted false connect attempt.
The patchset enables vsock (both host and guest) to cancel queued packets when
a connect attempt is considered to fail.
v2 changelog:
- fix queued_replies counting and resume tx/rx when necessary
Peng Tao (4):
vsock: track pkt owner vsock
vhost-vsock: add pkt cancel capability
vsock: add pkt cancel capability
vsock: cancel packets when failing to connect
drivers/vhost/vsock.c | 41 ++++++++++++++++++++++++++++++++
include/linux/virtio_vsock.h | 12 ++++++++++
net/vmw_vsock/af_vsock.c | 7 ++++++
net/vmw_vsock/virtio_transport.c | 42 +++++++++++++++++++++++++++++++++
net/vmw_vsock/virtio_transport_common.c | 14 +++++------
5 files changed, 109 insertions(+), 7 deletions(-)
--
2.7.4
^ permalink raw reply
* [PATCH v2 1/4] vsock: track pkt owner vsock
From: Peng Tao @ 2016-12-07 15:14 UTC (permalink / raw)
To: Stefan Hajnoczi; +Cc: kvm, virtualization, netdev, Peng Tao
In-Reply-To: <1481123652-80603-1-git-send-email-bergwolf@gmail.com>
So that we can cancel a queued pkt later if necessary.
Signed-off-by: Peng Tao <bergwolf@gmail.com>
---
include/linux/virtio_vsock.h | 2 ++
net/vmw_vsock/virtio_transport_common.c | 7 +++++++
2 files changed, 9 insertions(+)
diff --git a/include/linux/virtio_vsock.h b/include/linux/virtio_vsock.h
index 9638bfe..6dd3242 100644
--- a/include/linux/virtio_vsock.h
+++ b/include/linux/virtio_vsock.h
@@ -48,6 +48,7 @@ struct virtio_vsock_pkt {
struct virtio_vsock_hdr hdr;
struct work_struct work;
struct list_head list;
+ struct vsock_sock *vsk;
void *buf;
u32 len;
u32 off;
@@ -56,6 +57,7 @@ struct virtio_vsock_pkt {
struct virtio_vsock_pkt_info {
u32 remote_cid, remote_port;
+ struct vsock_sock *vsk;
struct msghdr *msg;
u32 pkt_len;
u16 type;
diff --git a/net/vmw_vsock/virtio_transport_common.c b/net/vmw_vsock/virtio_transport_common.c
index a53b3a1..cc1eeb5 100644
--- a/net/vmw_vsock/virtio_transport_common.c
+++ b/net/vmw_vsock/virtio_transport_common.c
@@ -57,6 +57,7 @@ virtio_transport_alloc_pkt(struct virtio_vsock_pkt_info *info,
pkt->len = len;
pkt->hdr.len = cpu_to_le32(len);
pkt->reply = info->reply;
+ pkt->vsk = info->vsk;
if (info->msg && len > 0) {
pkt->buf = kmalloc(len, GFP_KERNEL);
@@ -180,6 +181,7 @@ static int virtio_transport_send_credit_update(struct vsock_sock *vsk,
struct virtio_vsock_pkt_info info = {
.op = VIRTIO_VSOCK_OP_CREDIT_UPDATE,
.type = type,
+ .vsk = vsk,
};
return virtio_transport_send_pkt_info(vsk, &info);
@@ -519,6 +521,7 @@ int virtio_transport_connect(struct vsock_sock *vsk)
struct virtio_vsock_pkt_info info = {
.op = VIRTIO_VSOCK_OP_REQUEST,
.type = VIRTIO_VSOCK_TYPE_STREAM,
+ .vsk = vsk,
};
return virtio_transport_send_pkt_info(vsk, &info);
@@ -534,6 +537,7 @@ int virtio_transport_shutdown(struct vsock_sock *vsk, int mode)
VIRTIO_VSOCK_SHUTDOWN_RCV : 0) |
(mode & SEND_SHUTDOWN ?
VIRTIO_VSOCK_SHUTDOWN_SEND : 0),
+ .vsk = vsk,
};
return virtio_transport_send_pkt_info(vsk, &info);
@@ -560,6 +564,7 @@ virtio_transport_stream_enqueue(struct vsock_sock *vsk,
.type = VIRTIO_VSOCK_TYPE_STREAM,
.msg = msg,
.pkt_len = len,
+ .vsk = vsk,
};
return virtio_transport_send_pkt_info(vsk, &info);
@@ -581,6 +586,7 @@ static int virtio_transport_reset(struct vsock_sock *vsk,
.op = VIRTIO_VSOCK_OP_RST,
.type = VIRTIO_VSOCK_TYPE_STREAM,
.reply = !!pkt,
+ .vsk = vsk,
};
/* Send RST only if the original pkt is not a RST pkt */
@@ -826,6 +832,7 @@ virtio_transport_send_response(struct vsock_sock *vsk,
.remote_cid = le32_to_cpu(pkt->hdr.src_cid),
.remote_port = le32_to_cpu(pkt->hdr.src_port),
.reply = true,
+ .vsk = vsk,
};
return virtio_transport_send_pkt_info(vsk, &info);
--
2.7.4
^ permalink raw reply related
* 4.9.0-rc8: tg3 dead after resume
From: Billy Shuman @ 2016-12-07 15:20 UTC (permalink / raw)
To: netdev
After resume on 4.9.0-rc8 tg3 is dead.
In logs I see:
kernel: tg3 0000:44:00.0: phy probe failed, err -19
kernel: tg3 0000:44:00.0: Problem fetching invariants of chip, aborting
rmmod and modprobe does not fix the problem only a reboot resolves the issue.
Billy
^ permalink raw reply
* [RFC PATCH net-next v3 2/2] macb: Enable 1588 support in SAMA5Dx platforms.
From: Andrei Pistirica @ 2016-12-07 18:21 UTC (permalink / raw)
To: netdev, linux-kernel, linux-arm-kernel, davem, nicolas.ferre,
harinikatakamlinux, harini.katakam
Cc: punnaia, michals, anirudh, boris.brezillon, alexandre.belloni,
tbultel, richardcochran, rafalo, Andrei Pistirica
In-Reply-To: <1481134912-2243-1-git-send-email-andrei.pistirica@microchip.com>
This patch does the following:
- Enable HW time stamp for the following platforms: SAMA5D2, SAMA5D3 and
SAMA5D4.
- HW time stamp capabilities are advertised via ethtool and macb ioctl is
updated accordingly.
- HW time stamp on the PTP Ethernet packets are received using the
SO_TIMESTAMPING API. Where timers are obtained from the PTP event/peer
registers.
Note: Patch on net-next, on December 7th.
Signed-off-by: Andrei Pistirica <andrei.pistirica@microchip.com>
---
Patch history:
Version 1:
Integration with SAMA5D2 only. This feature wasn't tested on any
other platform that might use cadence/gem.
Patch is not completely ported to the very latest version of net-next,
and it will be after review.
Version 2 modifications:
- add PTP caps for SAMA5D2/3/4 platforms
- and cosmetic changes
Version 3 modifications:
- add support for sama5D2/3/4 platforms using GEM-PTP interface.
drivers/net/ethernet/cadence/macb.c | 165 ++++++++++++++++++++++++++++++++++--
1 file changed, 160 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/cadence/macb.c b/drivers/net/ethernet/cadence/macb.c
index 538544a..e2d4983 100644
--- a/drivers/net/ethernet/cadence/macb.c
+++ b/drivers/net/ethernet/cadence/macb.c
@@ -714,6 +714,8 @@ static void macb_tx_interrupt(struct macb_queue *queue)
/* First, update TX stats if needed */
if (skb) {
+ gem_ptp_do_txstamp(bp, skb);
+
netdev_vdbg(bp->dev, "skb %u (data %p) TX complete\n",
macb_tx_ring_wrap(bp, tail),
skb->data);
@@ -878,6 +880,8 @@ static int gem_rx(struct macb *bp, int budget)
GEM_BFEXT(RX_CSUM, ctrl) & GEM_RX_CSUM_CHECKED_MASK)
skb->ip_summed = CHECKSUM_UNNECESSARY;
+ gem_ptp_do_rxstamp(bp, skb);
+
bp->stats.rx_packets++;
bp->stats.rx_bytes += skb->len;
@@ -2080,6 +2084,9 @@ static int macb_open(struct net_device *dev)
netif_tx_start_all_queues(dev);
+ if (bp->ptp_info)
+ bp->ptp_info->ptp_init(dev);
+
return 0;
}
@@ -2101,6 +2108,9 @@ static int macb_close(struct net_device *dev)
macb_free_consistent(bp);
+ if (bp->ptp_info)
+ bp->ptp_info->ptp_remove(dev);
+
return 0;
}
@@ -2374,6 +2384,130 @@ static int macb_set_ringparam(struct net_device *netdev,
return 0;
}
+#ifdef CONFIG_MACB_USE_HWSTAMP
+static unsigned int gem_get_tsu_rate(struct macb *bp)
+{
+ /* Note: TSU rate is hardwired to PCLK. */
+ return clk_get_rate(bp->pclk);
+}
+
+static int gem_get_ts_info(struct net_device *dev,
+ struct ethtool_ts_info *info)
+{
+ struct macb *bp = netdev_priv(dev);
+
+ ethtool_op_get_ts_info(dev, info);
+ info->so_timestamping =
+ SOF_TIMESTAMPING_TX_SOFTWARE |
+ SOF_TIMESTAMPING_RX_SOFTWARE |
+ SOF_TIMESTAMPING_SOFTWARE |
+ SOF_TIMESTAMPING_TX_HARDWARE |
+ SOF_TIMESTAMPING_RX_HARDWARE |
+ SOF_TIMESTAMPING_RAW_HARDWARE;
+ info->phc_index = -1;
+
+ if (bp->ptp_clock)
+ info->phc_index = ptp_clock_index(bp->ptp_clock);
+
+ return 0;
+}
+
+static int gem_hwtst_set(struct net_device *netdev,
+ struct ifreq *ifr, int cmd)
+{
+ struct hwtstamp_config config;
+ struct macb *priv = netdev_priv(netdev);
+ u32 regval;
+
+ netdev_vdbg(netdev, "macb_hwtstamp_ioctl\n");
+
+ if (copy_from_user(&config, ifr->ifr_data, sizeof(config)))
+ return -EFAULT;
+
+ /* reserved for future extensions */
+ if (config.flags)
+ return -EINVAL;
+
+ switch (config.tx_type) {
+ case HWTSTAMP_TX_OFF:
+ priv->hwts_tx_en = false;
+ break;
+ case HWTSTAMP_TX_ON:
+ priv->hwts_tx_en = true;
+ break;
+ default:
+ return -ERANGE;
+ }
+
+ switch (config.rx_filter) {
+ case HWTSTAMP_FILTER_NONE:
+ if (priv->hwts_rx_en)
+ priv->hwts_rx_en = 0;
+ break;
+ case HWTSTAMP_FILTER_PTP_V1_L4_EVENT:
+ case HWTSTAMP_FILTER_PTP_V2_L4_EVENT:
+ case HWTSTAMP_FILTER_PTP_V2_L2_EVENT:
+ case HWTSTAMP_FILTER_ALL:
+ case HWTSTAMP_FILTER_PTP_V1_L4_SYNC:
+ case HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ:
+ case HWTSTAMP_FILTER_PTP_V2_L2_SYNC:
+ case HWTSTAMP_FILTER_PTP_V2_L4_SYNC:
+ case HWTSTAMP_FILTER_PTP_V2_L2_DELAY_REQ:
+ case HWTSTAMP_FILTER_PTP_V2_L4_DELAY_REQ:
+ case HWTSTAMP_FILTER_PTP_V2_EVENT:
+ case HWTSTAMP_FILTER_PTP_V2_SYNC:
+ case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ:
+ config.rx_filter = HWTSTAMP_FILTER_ALL;
+ regval = macb_readl(priv, NCR);
+ macb_writel(priv, NCR, (regval | MACB_BIT(SRTSM)));
+
+ if (!priv->hwts_rx_en)
+ priv->hwts_rx_en = 1;
+ break;
+ default:
+ return -ERANGE;
+ }
+
+ return copy_to_user(ifr->ifr_data, &config, sizeof(config)) ?
+ -EFAULT : 0;
+}
+
+static int gem_hwtst_get(struct net_device *netdev,
+ struct ifreq *ifr)
+{
+ struct hwtstamp_config config;
+ struct macb *priv = netdev_priv(netdev);
+
+ config.flags = 0;
+ config.tx_type = priv->hwts_tx_en ? HWTSTAMP_TX_ON : HWTSTAMP_TX_OFF;
+ config.rx_filter = (priv->hwts_rx_en ?
+ HWTSTAMP_FILTER_ALL : HWTSTAMP_FILTER_NONE);
+
+ return copy_to_user(ifr->ifr_data, &config, sizeof(config)) ?
+ -EFAULT : 0;
+}
+
+static struct macb_ptp_info gem_ptp_info = {
+ .ptp_init = gem_ptp_init,
+ .ptp_remove = gem_ptp_remove,
+ .get_tsu_rate = gem_get_tsu_rate,
+ .get_ts_info = gem_get_ts_info,
+ .hwtst_get = gem_hwtst_get,
+ .hwtst_set = gem_hwtst_set,
+};
+#endif
+
+static int macb_get_ts_info(struct net_device *netdev,
+ struct ethtool_ts_info *info)
+{
+ struct macb *bp = netdev_priv(netdev);
+
+ if (bp->ptp_info)
+ return bp->ptp_info->get_ts_info(netdev, info);
+
+ return ethtool_op_get_ts_info(netdev, info);
+}
+
static const struct ethtool_ops macb_ethtool_ops = {
.get_regs_len = macb_get_regs_len,
.get_regs = macb_get_regs,
@@ -2391,7 +2525,7 @@ static const struct ethtool_ops gem_ethtool_ops = {
.get_regs_len = macb_get_regs_len,
.get_regs = macb_get_regs,
.get_link = ethtool_op_get_link,
- .get_ts_info = ethtool_op_get_ts_info,
+ .get_ts_info = macb_get_ts_info,
.get_ethtool_stats = gem_get_ethtool_stats,
.get_strings = gem_get_ethtool_strings,
.get_sset_count = gem_get_sset_count,
@@ -2404,6 +2538,7 @@ static const struct ethtool_ops gem_ethtool_ops = {
static int macb_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
{
struct phy_device *phydev = dev->phydev;
+ struct macb *bp = netdev_priv(dev);
if (!netif_running(dev))
return -EINVAL;
@@ -2411,7 +2546,20 @@ static int macb_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
if (!phydev)
return -ENODEV;
- return phy_mii_ioctl(phydev, rq, cmd);
+ switch (cmd) {
+ case SIOCSHWTSTAMP:
+ if (bp->ptp_info)
+ return bp->ptp_info->hwtst_set(dev, rq, cmd);
+
+ return -EOPNOTSUPP;
+ case SIOCGHWTSTAMP:
+ if (bp->ptp_info)
+ return bp->ptp_info->hwtst_get(dev, rq);
+
+ return -EOPNOTSUPP;
+ default:
+ return phy_mii_ioctl(phydev, rq, cmd);
+ }
}
static int macb_set_features(struct net_device *netdev,
@@ -2485,6 +2633,12 @@ static void macb_configure_caps(struct macb *bp,
dcfg = gem_readl(bp, DCFG2);
if ((dcfg & (GEM_BIT(RX_PKT_BUFF) | GEM_BIT(TX_PKT_BUFF))) == 0)
bp->caps |= MACB_CAPS_FIFO_MODE;
+
+ /* iff HWSTAMP is configure and gem has the capability */
+#ifdef CONFIG_MACB_USE_HWSTAMP
+ if (gem_has_ptp(bp))
+ bp->ptp_info = &gem_ptp_info;
+#endif
}
dev_dbg(&bp->pdev->dev, "Cadence caps 0x%08x\n", bp->caps);
@@ -3041,7 +3195,7 @@ static const struct macb_config pc302gem_config = {
};
static const struct macb_config sama5d2_config = {
- .caps = MACB_CAPS_USRIO_DEFAULT_IS_MII_GMII,
+ .caps = MACB_CAPS_USRIO_DEFAULT_IS_MII_GMII | MACB_CAPS_GEM_HAS_PTP,
.dma_burst_length = 16,
.clk_init = macb_clk_init,
.init = macb_init,
@@ -3049,14 +3203,15 @@ static const struct macb_config sama5d2_config = {
static const struct macb_config sama5d3_config = {
.caps = MACB_CAPS_SG_DISABLED | MACB_CAPS_GIGABIT_MODE_AVAILABLE
- | MACB_CAPS_USRIO_DEFAULT_IS_MII_GMII,
+ | MACB_CAPS_USRIO_DEFAULT_IS_MII_GMII
+ | MACB_CAPS_GEM_HAS_PTP,
.dma_burst_length = 16,
.clk_init = macb_clk_init,
.init = macb_init,
};
static const struct macb_config sama5d4_config = {
- .caps = MACB_CAPS_USRIO_DEFAULT_IS_MII_GMII,
+ .caps = MACB_CAPS_USRIO_DEFAULT_IS_MII_GMII | MACB_CAPS_GEM_HAS_PTP,
.dma_burst_length = 4,
.clk_init = macb_clk_init,
.init = macb_init,
--
2.7.4
^ permalink raw reply related
* Re: [PATCH net-next] net: sock_rps_record_flow() is for connected sockets
From: David Miller @ 2016-12-07 15:47 UTC (permalink / raw)
To: eric.dumazet; +Cc: pabeni, netdev, willemb
In-Reply-To: <1481081570.18162.626.camel@edumazet-glaptop3.roam.corp.google.com>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Tue, 06 Dec 2016 19:32:50 -0800
> From: Eric Dumazet <edumazet@google.com>
>
> Paolo noticed a cache line miss in UDP recvmsg() to access
> sk_rxhash, sharing a cache line with sk_drops.
>
> sk_drops might be heavily incremented by cpus handling a flood targeting
> this socket.
>
> We might place sk_drops on a separate cache line, but lets try
> to avoid wasting 64 bytes per socket just for this, since we have
> other bottlenecks to take care of.
>
> sock_rps_record_flow() should only access sk_rxhash for connected
> flows.
>
> Testing sk_state for TCP_ESTABLISHED covers most of the cases for
> connected sockets, for a zero cost, since system calls using
> sock_rps_record_flow() also access sk->sk_prot which is on the
> same cache line.
>
> A follow up patch will provide a static_key (Jump Label) since most
> hosts do not even use RFS.
>
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> Reported-by: Paolo Abeni <pabeni@redhat.com>
Applied, thanks Eric.
^ permalink raw reply
* Re: [net-next][PATCH v2 10/18] RDS: IB: track and log active side endpoint in connection
From: David Miller @ 2016-12-07 15:53 UTC (permalink / raw)
To: santosh.shilimkar; +Cc: netdev, linux-kernel
In-Reply-To: <1481083316-11648-11-git-send-email-santosh.shilimkar@oracle.com>
From: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Date: Tue, 6 Dec 2016 20:01:48 -0800
> @@ -181,6 +181,9 @@ struct rds_ib_connection {
>
> /* Batched completions */
> unsigned int i_unsignaled_wrs;
> +
> + /* Endpoint role in connection */
> + int i_active_side;
> };
>
...
> @@ -685,6 +686,7 @@ int rds_ib_cm_initiate_connect(struct rdma_cm_id *cm_id)
> if (ic->i_cm_id == cm_id)
> ret = 0;
> }
> + ic->i_active_side = true;
> return ret;
> }
>
> @@ -859,6 +861,7 @@ void rds_ib_conn_path_shutdown(struct rds_conn_path *cp)
> ic->i_sends = NULL;
> vfree(ic->i_recvs);
> ic->i_recvs = NULL;
> + ic->i_active_side = false;
> }
>
> int rds_ib_conn_alloc(struct rds_connection *conn, gfp_t gfp)
Use type "bool", not "int", for true/false values.
^ permalink raw reply
* Re: [net-next][PATCH v2 18/18] RDS: IB: add missing connection cache usage info
From: David Miller @ 2016-12-07 15:55 UTC (permalink / raw)
To: santosh.shilimkar; +Cc: netdev, linux-kernel
In-Reply-To: <1481083316-11648-19-git-send-email-santosh.shilimkar@oracle.com>
From: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Date: Tue, 6 Dec 2016 20:01:56 -0800
> rds-tools already support it.
>
> Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
> ---
> include/uapi/linux/rds.h | 1 +
> net/rds/ib.c | 1 +
> 2 files changed, 2 insertions(+)
>
> diff --git a/include/uapi/linux/rds.h b/include/uapi/linux/rds.h
> index 3833113..410ae3c 100644
> --- a/include/uapi/linux/rds.h
> +++ b/include/uapi/linux/rds.h
> @@ -183,6 +183,7 @@ struct rds_info_rdma_connection {
> uint32_t max_send_sge;
> uint32_t rdma_mr_max;
> uint32_t rdma_mr_size;
> + uint32_t cache_allocs;
> };
>
> /* RDS message Receive Path Latency points */
What level of compatability exists here? If we run an old tool on a new
kernel, or a new tool on an old kernel, does it work properly?
^ permalink raw reply
* Re: [PATCH net-next v2 0/7] bnxt_en: Add interface to support RDMA driver.
From: David Miller @ 2016-12-07 15:59 UTC (permalink / raw)
To: michael.chan-dY08KVG/lbpWk0Htik3J/w
Cc: netdev-u79uwXL29TY76Z2rM5mHXA,
selvin.xavier-dY08KVG/lbpWk0Htik3J/w,
somnath.kotur-dY08KVG/lbpWk0Htik3J/w,
dledford-H+wXaHxf7aLQT0dZR+AlfA,
linux-rdma-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <1481088381-30411-1-git-send-email-michael.chan-dY08KVG/lbpWk0Htik3J/w@public.gmane.org>
From: Michael Chan <michael.chan-dY08KVG/lbpWk0Htik3J/w@public.gmane.org>
Date: Wed, 7 Dec 2016 00:26:14 -0500
> This series adds an interface to support a brand new RDMA driver bnxt_re.
> The first step is to re-arrange some code so that pci_enable_msix() can
> be called during pci probe. The purpose is to allow the RDMA driver to
> initialize and stay initialized whether the netdev is up or down.
>
> Then we make some changes to VF resource allocation so that there is
> enough resources to support RDMA.
>
> Finally the last patch adds a simple interface to allow the RDMA driver to
> probe and register itself with any bnxt_en devices that support RDMA.
> Once registered, the RDMA driver can request MSIX, send fw messages, and
> receive some notifications.
>
> v2: Fixed kbuild test robot warnings.
>
> David, please consider this series for net-next. Thanks.
Series applied, thanks.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH net-next] net: sock_rps_record_flow() is for connected sockets
From: Eric Dumazet @ 2016-12-07 15:59 UTC (permalink / raw)
To: Paolo Abeni; +Cc: David Miller, netdev, Willem de Bruijn, Tom Herbert
In-Reply-To: <1481120956.4930.7.camel@edumazet-glaptop3.roam.corp.google.com>
On Wed, 2016-12-07 at 06:29 -0800, Eric Dumazet wrote:
> Keeping frag_list means you keep one sk_buff per segment, so this really
> looks like a legacy UDP server (like a DNS server) wont benefit from
> this anyway.
I played with the idea of preparing the skb for minimal overhead for the
process doing udp_recvmsg().
If socket is under pressure, softirq handler(s) can try to pull in
skb->head the payload of the packet if it fits.
Meaning the softirq handler can free/reuse the page fragment
immediately, instead of letting udp_recvmsg() do this hundreds of usec
later.
( Sort of copybreak, but without reallocating skb->head )
Gains :
- We reduce skb->truesize and thus can store more packets per SO_RCVBUF
KB
- We avoid a cache line misses at copyout() time and consume_skb() time,
and avoid one put_page() with potential alien freeing on NUMA hosts.
^ permalink raw reply
* Re: [PATCH v5 00/13] net: ethernet: ti: cpts: update and fixes
From: David Miller @ 2016-12-07 16:14 UTC (permalink / raw)
To: grygorii.strashko
Cc: netdev, mugunthanvnm, richardcochran, nsekhar, linux-kernel,
linux-omap, devicetree, m-karicheri2, w-kwok2, tglx
In-Reply-To: <20161207000045.28333-1-grygorii.strashko@ti.com>
From: Grygorii Strashko <grygorii.strashko@ti.com>
Date: Tue, 6 Dec 2016 18:00:32 -0600
> It is preparation series intended to clean up and optimize TI CPTS driver to
> facilitate further integration with other TI's SoCs like Keystone 2.
...
Series applied.
^ permalink raw reply
* [PATCH] sh_eth: add wake-on-lan support via magic packet
From: Niklas Söderlund @ 2016-12-07 16:28 UTC (permalink / raw)
To: Sergei Shtylyov, Simon Horman, netdev, linux-renesas-soc
Cc: Niklas Söderlund
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
---
drivers/net/ethernet/renesas/sh_eth.c | 120 +++++++++++++++++++++++++++++++---
drivers/net/ethernet/renesas/sh_eth.h | 4 ++
2 files changed, 116 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/renesas/sh_eth.c b/drivers/net/ethernet/renesas/sh_eth.c
index 05b0dc5..3974046 100644
--- a/drivers/net/ethernet/renesas/sh_eth.c
+++ b/drivers/net/ethernet/renesas/sh_eth.c
@@ -624,7 +624,7 @@ static struct sh_eth_cpu_data r8a779x_data = {
.register_type = SH_ETH_REG_FAST_RCAR,
- .ecsr_value = ECSR_PSRTO | ECSR_LCHNG | ECSR_ICD,
+ .ecsr_value = ECSR_PSRTO | ECSR_LCHNG | ECSR_ICD | ECSR_MPD,
.ecsipr_value = ECSIPR_PSRTOIP | ECSIPR_LCHNGIP | ECSIPR_ICDIP,
.eesipr_value = 0x01ff009f,
@@ -641,6 +641,7 @@ static struct sh_eth_cpu_data r8a779x_data = {
.tpauser = 1,
.hw_swap = 1,
.rmiimode = 1,
+ .magic = 1,
};
#endif /* CONFIG_OF */
@@ -1657,6 +1658,10 @@ static irqreturn_t sh_eth_interrupt(int irq, void *netdev)
goto out;
if (!likely(mdp->irq_enabled)) {
+ /* Handle MagicPacket interrupt */
+ if (sh_eth_read(ndev, ECSR) & ECSR_MPD)
+ sh_eth_modify(ndev, ECSR, 0, ECSR_MPD);
+
sh_eth_write(ndev, 0, EESIPR);
goto out;
}
@@ -2199,6 +2204,33 @@ static int sh_eth_set_ringparam(struct net_device *ndev,
return 0;
}
+static void sh_eth_get_wol(struct net_device *ndev, struct ethtool_wolinfo *wol)
+{
+ struct sh_eth_private *mdp = netdev_priv(ndev);
+
+ wol->supported = 0;
+ wol->wolopts = 0;
+
+ if (mdp->cd->magic && mdp->clk) {
+ wol->supported = WAKE_MAGIC;
+ wol->wolopts = mdp->wol_enabled ? WAKE_MAGIC : 0;
+ }
+}
+
+static int sh_eth_set_wol(struct net_device *ndev, struct ethtool_wolinfo *wol)
+{
+ struct sh_eth_private *mdp = netdev_priv(ndev);
+
+ if (!mdp->cd->magic || !mdp->clk || wol->wolopts & ~WAKE_MAGIC)
+ return -EOPNOTSUPP;
+
+ mdp->wol_enabled = wol->wolopts & WAKE_MAGIC;
+
+ device_set_wakeup_enable(&mdp->pdev->dev, mdp->wol_enabled);
+
+ return 0;
+}
+
static const struct ethtool_ops sh_eth_ethtool_ops = {
.get_regs_len = sh_eth_get_regs_len,
.get_regs = sh_eth_get_regs,
@@ -2213,6 +2245,8 @@ static const struct ethtool_ops sh_eth_ethtool_ops = {
.set_ringparam = sh_eth_set_ringparam,
.get_link_ksettings = sh_eth_get_link_ksettings,
.set_link_ksettings = sh_eth_set_link_ksettings,
+ .get_wol = sh_eth_get_wol,
+ .set_wol = sh_eth_set_wol,
};
/* network device open function */
@@ -3017,6 +3051,11 @@ static int sh_eth_drv_probe(struct platform_device *pdev)
goto out_release;
}
+ /* Get clock, if not found that's OK but Wake-On-Lan is unavailable */
+ mdp->clk = devm_clk_get(&pdev->dev, NULL);
+ if (IS_ERR(mdp->clk))
+ mdp->clk = NULL;
+
ndev->base_addr = res->start;
spin_lock_init(&mdp->lock);
@@ -3111,6 +3150,10 @@ static int sh_eth_drv_probe(struct platform_device *pdev)
if (ret)
goto out_napi_del;
+ mdp->wol_enabled = false;
+ if (mdp->cd->magic && mdp->clk)
+ device_set_wakeup_capable(&pdev->dev, 1);
+
/* print device information */
netdev_info(ndev, "Base address at 0x%x, %pM, IRQ %d.\n",
(u32)ndev->base_addr, ndev->dev_addr, ndev->irq);
@@ -3150,15 +3193,71 @@ static int sh_eth_drv_remove(struct platform_device *pdev)
#ifdef CONFIG_PM
#ifdef CONFIG_PM_SLEEP
+static int sh_eth_wol_setup(struct net_device *ndev)
+{
+ struct sh_eth_private *mdp = netdev_priv(ndev);
+
+ /* Only allow ECI interrupts */
+ mdp->irq_enabled = false;
+ synchronize_irq(ndev->irq);
+ napi_disable(&mdp->napi);
+ sh_eth_write(ndev, DMAC_M_ECI, EESIPR);
+
+ /* Enable ECI MagicPacket interrupt */
+ sh_eth_write(ndev, ECSIPR_MPDIP, ECSIPR);
+
+ /* Enable MagicPacket */
+ sh_eth_modify(ndev, ECMR, 0, ECMR_PMDE);
+
+ /* Increased clock usage so device won't be suspended */
+ clk_enable(mdp->clk);
+
+ return enable_irq_wake(ndev->irq);
+}
+
+static int sh_eth_wol_restore(struct net_device *ndev)
+{
+ struct sh_eth_private *mdp = netdev_priv(ndev);
+ int ret;
+
+ napi_enable(&mdp->napi);
+
+ /* Disable MagicPacket */
+ sh_eth_modify(ndev, ECMR, ECMR_PMDE, 0);
+
+ /* The device need to be reset to restore MagicPacket logic
+ * for next wakeup. If we close and open the device it will
+ * both be reset and all registers restored. This is what
+ * happens during suspend and resume without WoL enabled.
+ */
+ ret = sh_eth_close(ndev);
+ if (ret < 0)
+ return ret;
+ ret = sh_eth_open(ndev);
+ if (ret < 0)
+ return ret;
+
+ /* Restore clock usage count */
+ clk_disable(mdp->clk);
+
+ return disable_irq_wake(ndev->irq);
+}
+
static int sh_eth_suspend(struct device *dev)
{
struct net_device *ndev = dev_get_drvdata(dev);
+ struct sh_eth_private *mdp = netdev_priv(ndev);
int ret = 0;
- if (netif_running(ndev)) {
- netif_device_detach(ndev);
+ if (!netif_running(ndev))
+ return 0;
+
+ netif_device_detach(ndev);
+
+ if (mdp->wol_enabled)
+ ret = sh_eth_wol_setup(ndev);
+ else
ret = sh_eth_close(ndev);
- }
return ret;
}
@@ -3166,14 +3265,19 @@ static int sh_eth_suspend(struct device *dev)
static int sh_eth_resume(struct device *dev)
{
struct net_device *ndev = dev_get_drvdata(dev);
+ struct sh_eth_private *mdp = netdev_priv(ndev);
int ret = 0;
- if (netif_running(ndev)) {
+ if (!netif_running(ndev))
+ return 0;
+
+ if (mdp->wol_enabled)
+ ret = sh_eth_wol_restore(ndev);
+ else
ret = sh_eth_open(ndev);
- if (ret < 0)
- return ret;
+
+ if (!ret)
netif_device_attach(ndev);
- }
return ret;
}
diff --git a/drivers/net/ethernet/renesas/sh_eth.h b/drivers/net/ethernet/renesas/sh_eth.h
index d050f37..26c6620 100644
--- a/drivers/net/ethernet/renesas/sh_eth.h
+++ b/drivers/net/ethernet/renesas/sh_eth.h
@@ -493,6 +493,7 @@ struct sh_eth_cpu_data {
unsigned shift_rd0:1; /* shift Rx descriptor word 0 right by 16 */
unsigned rmiimode:1; /* EtherC has RMIIMODE register */
unsigned rtrate:1; /* EtherC has RTRATE register */
+ unsigned magic:1; /* EtherC have PMDE in ECMR and MPDIP in ECSIPR */
};
struct sh_eth_private {
@@ -529,6 +530,9 @@ struct sh_eth_private {
unsigned no_ether_link:1;
unsigned ether_link_active_low:1;
unsigned is_opened:1;
+
+ bool wol_enabled;
+ struct clk *clk;
};
static inline void sh_eth_soft_swap(char *src, int len)
--
2.10.2
^ permalink raw reply related
* Re: [net-next][PATCH v2 10/18] RDS: IB: track and log active side endpoint in connection
From: Santosh Shilimkar @ 2016-12-07 16:29 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-kernel
In-Reply-To: <20161207.105318.2143064665652853543.davem@davemloft.net>
On 12/7/2016 7:53 AM, David Miller wrote:
> From: Santosh Shilimkar <santosh.shilimkar@oracle.com>
> Date: Tue, 6 Dec 2016 20:01:48 -0800
>
>> @@ -181,6 +181,9 @@ struct rds_ib_connection {
>>
>> /* Batched completions */
>> unsigned int i_unsignaled_wrs;
>> +
>> + /* Endpoint role in connection */
>> + int i_active_side;
>> };
>>
> ...
>> @@ -685,6 +686,7 @@ int rds_ib_cm_initiate_connect(struct rdma_cm_id *cm_id)
>> if (ic->i_cm_id == cm_id)
>> ret = 0;
>> }
>> + ic->i_active_side = true;
>> return ret;
>> }
>>
>> @@ -859,6 +861,7 @@ void rds_ib_conn_path_shutdown(struct rds_conn_path *cp)
>> ic->i_sends = NULL;
>> vfree(ic->i_recvs);
>> ic->i_recvs = NULL;
>> + ic->i_active_side = false;
>> }
>>
>> int rds_ib_conn_alloc(struct rds_connection *conn, gfp_t gfp)
>
> Use type "bool", not "int", for true/false values.
>
OK. Will fix it.
Regards,
Santosh
^ permalink raw reply
* [PATCH net-next] net: rfs: add a jump label
From: Eric Dumazet @ 2016-12-07 16:29 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Paolo Abeni
From: Eric Dumazet <edumazet@google.com>
RFS is not commonly used, so add a jump label to avoid some conditionals
in fast path.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Paolo Abeni <pabeni@redhat.com>
---
include/linux/netdevice.h | 1 +
include/net/sock.h | 25 ++++++++++++++-----------
net/core/dev.c | 2 ++
net/core/sysctl_net_core.c | 5 ++++-
4 files changed, 21 insertions(+), 12 deletions(-)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 1ff5ea6e12214db818c2cfa8a9b8ed5cbddc307c..994f7423a74bd622884c3b646f4123d28697b8ad 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -192,6 +192,7 @@ struct net_device_stats {
#ifdef CONFIG_RPS
#include <linux/static_key.h>
extern struct static_key rps_needed;
+extern struct static_key rfs_needed;
#endif
struct neighbour;
diff --git a/include/net/sock.h b/include/net/sock.h
index 1749e38d03014558ac882b5d1fb37b11ac5e6705..2729e77950b762e2f246646ef80f6cc6c75d7a5d 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h
@@ -913,17 +913,20 @@ static inline void sock_rps_record_flow_hash(__u32 hash)
static inline void sock_rps_record_flow(const struct sock *sk)
{
#ifdef CONFIG_RPS
- /* Reading sk->sk_rxhash might incur an expensive cache line miss.
- *
- * TCP_ESTABLISHED does cover almost all states where RFS
- * might be useful, and is cheaper [1] than testing :
- * IPv4: inet_sk(sk)->inet_daddr
- * IPv6: ipv6_addr_any(&sk->sk_v6_daddr)
- * OR an additional socket flag
- * [1] : sk_state and sk_prot are in the same cache line.
- */
- if (sk->sk_state == TCP_ESTABLISHED)
- sock_rps_record_flow_hash(sk->sk_rxhash);
+ if (static_key_false(&rfs_needed)) {
+ /* Reading sk->sk_rxhash might incur an expensive cache line
+ * miss.
+ *
+ * TCP_ESTABLISHED does cover almost all states where RFS
+ * might be useful, and is cheaper [1] than testing :
+ * IPv4: inet_sk(sk)->inet_daddr
+ * IPv6: ipv6_addr_any(&sk->sk_v6_daddr)
+ * OR an additional socket flag
+ * [1] : sk_state and sk_prot are in the same cache line.
+ */
+ if (sk->sk_state == TCP_ESTABLISHED)
+ sock_rps_record_flow_hash(sk->sk_rxhash);
+ }
#endif
}
diff --git a/net/core/dev.c b/net/core/dev.c
index bffb5253e77867b1d6a0ada7cc99f4605e03ad28..1d33ce03365f1e10996ad5274e86bf351a526284 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -3447,6 +3447,8 @@ EXPORT_SYMBOL(rps_cpu_mask);
struct static_key rps_needed __read_mostly;
EXPORT_SYMBOL(rps_needed);
+struct static_key rfs_needed __read_mostly;
+EXPORT_SYMBOL(rfs_needed);
static struct rps_dev_flow *
set_rps_cpu(struct net_device *dev, struct sk_buff *skb,
diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c
index 0df2aa6525308a365d89f57f6da76d57a24238f0..2a46e4009f62d8c2ac8949789ae9626b0c016a11 100644
--- a/net/core/sysctl_net_core.c
+++ b/net/core/sysctl_net_core.c
@@ -79,10 +79,13 @@ static int rps_sock_flow_sysctl(struct ctl_table *table, int write,
if (sock_table != orig_sock_table) {
rcu_assign_pointer(rps_sock_flow_table, sock_table);
- if (sock_table)
+ if (sock_table) {
static_key_slow_inc(&rps_needed);
+ static_key_slow_inc(&rfs_needed);
+ }
if (orig_sock_table) {
static_key_slow_dec(&rps_needed);
+ static_key_slow_dec(&rfs_needed);
synchronize_rcu();
vfree(orig_sock_table);
}
^ permalink raw reply related
* Re: [PATCH v3 net-next 1/4] bpf: xdp: Allow head adjustment in XDP prog
From: Alexei Starovoitov @ 2016-12-07 16:37 UTC (permalink / raw)
To: Jakub Kicinski
Cc: Daniel Borkmann, Martin KaFai Lau, netdev, Alexei Starovoitov,
Brenden Blanco, David Miller, Jesper Dangaard Brouer,
John Fastabend, Saeed Mahameed, Tariq Toukan, Kernel Team
In-Reply-To: <20161207114112.6ad86da3@jkicinski-Precision-T1700>
On Wed, Dec 07, 2016 at 11:41:12AM +0000, Jakub Kicinski wrote:
> > I see nothing wrong if this is exposed/made visible in the usual way through
> > ethtool -k as well. I guess at least that would be the expected way to query
> > for such driver capabilities.
>
> +1 on exposing this to user space. Whether via ethtool -k or a
> separate XDP-specific netlink message is mostly a question of whether
> we expect the need to expose more complex capabilities than bits.
I'm very much against using NETIF_F_ flags and exposing this to user space.
I see this xdp feature flag as temporary workaround until all drivers
support adjust_head() helper. It is very much a fundamental feature for xdp.
Without being able to add/remove headers the usability of xdp becomes very limited.
If you guys dont like extra ndo_xdp command, I'd suggest to do
"if (prog->xdp_adjust_head)" check in the driver and if driver doesn't
support it yet, just fail XDP_SETUP_PROG command.
imo that will be more flexible interface, since in the future drivers
can fail on different combination of features and simple boolean flag
unlikely to serve us for long time.
^ permalink raw reply
* Re: [PATCH net-next v2] dsa:mv88e6xxx: dispose irq mapping for chip->irq
From: Volodymyr Bendiuga @ 2016-12-07 16:40 UTC (permalink / raw)
To: Andrew Lunn
Cc: vivien.didelot-4ysUXcep3aM1wj+D4I0NRVaTQe2KTcn/,
f.fainelli-Re5JQEeQqe8AvxtiuMwx3w, netdev-u79uwXL29TY76Z2rM5mHXA,
volodymyr.bendiuga-Re5JQEeQqe8AvxtiuMwx3w, Rob Herring,
devicetree-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20161207141144.GJ18817-g2DYL2Zd6BY@public.gmane.org>
Yes, most of the users of of_irq_get() do not use irq_dispose_mapping().
But some of them do (some irq chips), and I believe the correct way of
doing this is to
dispose irq mapping, as the description for this function says that it
unmaps
the irq, which is mapped by of_irq_parse_and_map(). Not disposing irq
might not make
any affect on most drivers, but some, that get EPROBE_DEFER error do
need to dispose.
This is what I get when I run the code.
of_irq_put() could be implemented, and it would be a wrapper for
irq_dispose_mapping()
as I can see it. Should I do it this way?
On 2016-12-07 15:11, Andrew Lunn wrote:
> On Wed, Dec 07, 2016 at 11:35:07AM +0100, Volodymyr Bendiuga wrote:
>> Signed-off-by: Volodymyr Bendiuga <volodymyr.bendiuga-qeDNsGSBLoYwFerOooGFRg@public.gmane.org>
> You need some text in the Change log. Say why this change is needed,
> etc.
>
> Looking through other users of of_irq_get(), i don't see any disposing
> of the mapping. It is not obvious you need to do this. The name does
> not give any hint.
>
> Maybe we should add an of_irq_put() which would clean up the mapping?
>
> Andrew
>
>
>> ---
>> drivers/net/dsa/mv88e6xxx/chip.c | 6 +++++-
>> 1 file changed, 5 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c
>> index 173ea97..0c3271d 100644
>> --- a/drivers/net/dsa/mv88e6xxx/chip.c
>> +++ b/drivers/net/dsa/mv88e6xxx/chip.c
>> @@ -4483,7 +4483,7 @@ static int mv88e6xxx_probe(struct mdio_device *mdiodev)
>> mutex_unlock(&chip->reg_lock);
>>
>> if (err)
>> - goto out;
>> + goto out_dispose;
>>
>> if (mv88e6xxx_has(chip, MV88E6XXX_FLAG_G2_INT)) {
>> err = mv88e6xxx_g2_irq_setup(chip);
>> @@ -4513,6 +4513,9 @@ static int mv88e6xxx_probe(struct mdio_device *mdiodev)
>> mv88e6xxx_g1_irq_free(chip);
>> mutex_unlock(&chip->reg_lock);
>> }
>> +
>> +out_dispose:
>> + irq_dispose_mapping(chip->irq);
>> out:
>> return err;
>> }
>> @@ -4530,6 +4533,7 @@ static void mv88e6xxx_remove(struct mdio_device *mdiodev)
>> if (mv88e6xxx_has(chip, MV88E6XXX_FLAG_G2_INT))
>> mv88e6xxx_g2_irq_free(chip);
>> mv88e6xxx_g1_irq_free(chip);
>> + irq_dispose_mapping(chip->irq);
>> }
>> }
>>
>> --
>> 2.7.4
>>
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [net-next][PATCH v2 18/18] RDS: IB: add missing connection cache usage info
From: Santosh Shilimkar @ 2016-12-07 16:44 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-kernel
In-Reply-To: <20161207.105531.1347292514118544125.davem@davemloft.net>
On 12/7/2016 7:55 AM, David Miller wrote:
> From: Santosh Shilimkar <santosh.shilimkar@oracle.com>
> Date: Tue, 6 Dec 2016 20:01:56 -0800
>
>> rds-tools already support it.
>>
>> Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
>> ---
>> include/uapi/linux/rds.h | 1 +
>> net/rds/ib.c | 1 +
>> 2 files changed, 2 insertions(+)
>>
>> diff --git a/include/uapi/linux/rds.h b/include/uapi/linux/rds.h
>> index 3833113..410ae3c 100644
>> --- a/include/uapi/linux/rds.h
>> +++ b/include/uapi/linux/rds.h
>> @@ -183,6 +183,7 @@ struct rds_info_rdma_connection {
>> uint32_t max_send_sge;
>> uint32_t rdma_mr_max;
>> uint32_t rdma_mr_size;
>> + uint32_t cache_allocs;
>> };
>>
>> /* RDS message Receive Path Latency points */
>
> What level of compatability exists here? If we run an old tool on a new
> kernel, or a new tool on an old kernel, does it work properly?
>
Tools repo carries a copy of the header and thats how the old tool and
new tools have been running with older/newer kernels. There are few more
bits left before I can start using directly kernel header for newer
tools.
Moreover this particular parameter is only used for verbose mode which
isn't used in default options.
Regards,
Santosh
^ permalink raw reply
* Re: ixgbe Port cannot load, "failed to register GSI"
From: Alexander Duyck @ 2016-12-07 16:48 UTC (permalink / raw)
To: Ben Greear; +Cc: e1000-devel@lists.sourceforge.net, netdev, intel-wired-lan
In-Reply-To: <60fc28f5-9780-dc04-2721-043412d23d31@candelatech.com>
On Tue, Dec 6, 2016 at 3:22 PM, Ben Greear <greearb@candelatech.com> wrote:
> We put 3 10-g dual-port ixgbe NICs and 4 4-port I350 NICs in a 2U rackmount,
> and one of the ixgbe ports
> fails to come up. This previously worked before reboot, so maybe it is a
> race somehow. Kernel is 4.4.11+,
> but not hacks to ixgbe or I350 drivers.
>
> Anyone know if there is some sort of way to make this work reliably?
>
> dmesg | grep ixgbe
>
> [ 5.803307] ixgbe: Intel(R) 10 Gigabit PCI Express Network Driver -
> version 4.2.1-k
> [ 5.803309] ixgbe: Copyright (c) 1999-2015 Intel Corporation.
> [ 5.952119] ixgbe 0000:04:00.0: Multiqueue Enabled: Rx Queue count = 8,
> Tx Queue count = 8
> [ 5.952245] ixgbe 0000:04:00.0: PCI Express bandwidth of 32GT/s available
> [ 5.952246] ixgbe 0000:04:00.0: (Speed:5.0GT/s, Width: x8, Encoding
> Loss:20%)
> [ 5.952328] ixgbe 0000:04:00.0: MAC: 2, PHY: 15, SFP+: 5, PBA No:
> FFFFFF-0FF
> [ 5.952330] ixgbe 0000:04:00.0: 00:e0:ed:77:09:16
> [ 5.954004] ixgbe 0000:04:00.0: Intel(R) 10 Gigabit Network Connection
> [ 6.102346] ixgbe 0000:04:00.1: Multiqueue Enabled: Rx Queue count = 8,
> Tx Queue count = 8
> [ 6.102475] ixgbe 0000:04:00.1: PCI Express bandwidth of 32GT/s available
> [ 6.102478] ixgbe 0000:04:00.1: (Speed:5.0GT/s, Width: x8, Encoding
> Loss:20%)
> [ 6.102562] ixgbe 0000:04:00.1: MAC: 2, PHY: 15, SFP+: 6, PBA No:
> FFFFFF-0FF
> [ 6.102564] ixgbe 0000:04:00.1: 00:e0:ed:77:09:17
> [ 6.104869] ixgbe 0000:04:00.1: Intel(R) 10 Gigabit Network Connection
> [ 6.253429] ixgbe 0000:05:00.0: Multiqueue Enabled: Rx Queue count = 8,
> Tx Queue count = 8
> [ 6.253558] ixgbe 0000:05:00.0: PCI Express bandwidth of 32GT/s available
> [ 6.253560] ixgbe 0000:05:00.0: (Speed:5.0GT/s, Width: x8, Encoding
> Loss:20%)
> [ 6.253644] ixgbe 0000:05:00.0: MAC: 2, PHY: 15, SFP+: 5, PBA No:
> FFFFFF-0FF
> [ 6.253646] ixgbe 0000:05:00.0: 00:e0:ed:79:06:50
> [ 6.255855] ixgbe 0000:05:00.0: Intel(R) 10 Gigabit Network Connection
> [ 6.404128] ixgbe 0000:05:00.1: Multiqueue Enabled: Rx Queue count = 8,
> Tx Queue count = 8
> [ 6.404254] ixgbe 0000:05:00.1: PCI Express bandwidth of 32GT/s available
> [ 6.404255] ixgbe 0000:05:00.1: (Speed:5.0GT/s, Width: x8, Encoding
> Loss:20%)
> [ 6.404337] ixgbe 0000:05:00.1: MAC: 2, PHY: 15, SFP+: 6, PBA No:
> FFFFFF-0FF
> [ 6.404339] ixgbe 0000:05:00.1: 00:e0:ed:79:06:51
> [ 6.405914] ixgbe 0000:05:00.1: Intel(R) 10 Gigabit Network Connection
> [ 6.554373] ixgbe 0000:06:00.0: Multiqueue Enabled: Rx Queue count = 8,
> Tx Queue count = 8
> [ 6.554501] ixgbe 0000:06:00.0: PCI Express bandwidth of 32GT/s available
> [ 6.554504] ixgbe 0000:06:00.0: (Speed:5.0GT/s, Width: x8, Encoding
> Loss:20%)
> [ 6.554588] ixgbe 0000:06:00.0: MAC: 2, PHY: 15, SFP+: 5, PBA No:
> FFFFFF-0FF
> [ 6.554590] ixgbe 0000:06:00.0: 00:e0:ed:79:06:56
> [ 6.556994] ixgbe 0000:06:00.0: Intel(R) 10 Gigabit Network Connection
> [ 6.557160] ixgbe 0000:06:00.1: PCI INT B: failed to register GSI
> [ 6.557169] ixgbe: probe of 0000:06:00.1 failed with error -28
>
> Thanks,
> Ben
> --
> Ben Greear <greearb@candelatech.com>
> Candela Technologies Inc http://www.candelatech.com
>
I'm adding e1000-devel and intel-wired-lan as they would be a better
place for an ixgbe question.
It might be useful if you could also provide an "lspci -vvv" for the
system, or at least for the network ports so that we can verify if
enough resources are present in terms of memory to allocate memory for
MMIO on all the devices. Also any info you could provide on the
platform itself would be useful as it is possible what you are seeing
is a resource constraint as the GSI error seems to indicate ACPI is
having some sort of issue.
- Alex
------------------------------------------------------------------------------
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today.http://sdm.link/xeonphi
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel® Ethernet, visit http://communities.intel.com/community/wired
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox