Messages in this thread | | | Subject | Re: [PATCH net-next] dropwatch: Support monitoring of dropped frames | From | Eric Dumazet <> | Date | Tue, 7 Jul 2020 10:33:50 -0700 |
| |
On 7/7/20 10:15 AM, izabela.bakollari@gmail.com wrote: > From: Izabela Bakollari <izabela.bakollari@gmail.com> > > Dropwatch is a utility that monitors dropped frames by having userspace > record them over the dropwatch protocol over a file. This augument > allows live monitoring of dropped frames using tools like tcpdump. > > With this feature, dropwatch allows two additional commands (start and > stop interface) which allows the assignment of a net_device to the > dropwatch protocol. When assinged, dropwatch will clone dropped frames, > and receive them on the assigned interface, allowing tools like tcpdump > to monitor for them. > > With this feature, create a dummy ethernet interface (ip link add dev > dummy0 type dummy), assign it to the dropwatch kernel subsystem, by using > these new commands, and then monitor dropped frames in real time by > running tcpdump -i dummy0. > > Signed-off-by: Izabela Bakollari <izabela.bakollari@gmail.com> > --- > include/uapi/linux/net_dropmon.h | 3 ++ > net/core/drop_monitor.c | 79 +++++++++++++++++++++++++++++++- > 2 files changed, 80 insertions(+), 2 deletions(-) > > diff --git a/include/uapi/linux/net_dropmon.h b/include/uapi/linux/net_dropmon.h > index 67e31f329190..e8e861e03a8a 100644 > --- a/include/uapi/linux/net_dropmon.h > +++ b/include/uapi/linux/net_dropmon.h > @@ -58,6 +58,8 @@ enum { > NET_DM_CMD_CONFIG_NEW, > NET_DM_CMD_STATS_GET, > NET_DM_CMD_STATS_NEW, > + NET_DM_CMD_START_IFC, > + NET_DM_CMD_STOP_IFC, > _NET_DM_CMD_MAX, > }; > > @@ -93,6 +95,7 @@ enum net_dm_attr { > NET_DM_ATTR_SW_DROPS, /* flag */ > NET_DM_ATTR_HW_DROPS, /* flag */ > NET_DM_ATTR_FLOW_ACTION_COOKIE, /* binary */ > + NET_DM_ATTR_IFNAME, /* string */ > > __NET_DM_ATTR_MAX, > NET_DM_ATTR_MAX = __NET_DM_ATTR_MAX - 1 > diff --git a/net/core/drop_monitor.c b/net/core/drop_monitor.c > index 8e33cec9fc4e..8049bff05abd 100644 > --- a/net/core/drop_monitor.c > +++ b/net/core/drop_monitor.c > @@ -30,6 +30,7 @@ > #include <net/genetlink.h> > #include <net/netevent.h> > #include <net/flow_offload.h> > +#include <net/sock.h> > > #include <trace/events/skb.h> > #include <trace/events/napi.h> > @@ -46,6 +47,7 @@ > */ > static int trace_state = TRACE_OFF; > static bool monitor_hw; > +struct net_device *interface; > > /* net_dm_mutex > * > @@ -220,9 +222,8 @@ static void trace_drop_common(struct sk_buff *skb, void *location) > struct per_cpu_dm_data *data; > unsigned long flags; > > - local_irq_save(flags); > + spin_lock_irqsave(&data->lock, flags); > data = this_cpu_ptr(&dm_cpu_data); > - spin_lock(&data->lock);
This change seems unrelated ?
And also buggy, since data is essentially garbage when you call spin_lock_irqsave(&data->lock, flags);
> dskb = data->skb; > > if (!dskb) > @@ -255,6 +256,12 @@ static void trace_drop_common(struct sk_buff *skb, void *location) > > out: > spin_unlock_irqrestore(&data->lock, flags); > + > + if (interface && interface != skb->dev) { > + skb = skb_clone(skb, GFP_ATOMIC);
skb_clone() can return NULL
> + skb->dev = interface; > + netif_receive_skb(skb); > + } > } > > static void trace_kfree_skb_hit(void *ignore, struct sk_buff *skb, void *location) > @@ -1315,6 +1322,63 @@ static int net_dm_cmd_trace(struct sk_buff *skb, > return -EOPNOTSUPP; > } > > +static int net_dm_interface_start(struct net *net, const char *ifname) > +{ > + struct net_device *nd; > + > + nd = dev_get_by_name(net, ifname); > + > + if (nd) { > + interface = nd; > + dev_hold(interface); > + } else { > + return -ENODEV; > + } > + return 0; > +} > +
What happens after this monitoring is started, then the admin does :
rmmod ifb
| |