2022-02-09 12:42:41

by Menglong Dong

[permalink] [raw]
Subject: [PATCH v8 net-next] net: drop_monitor: support drop reason

From: Menglong Dong <[email protected]>

In the commit c504e5c2f964 ("net: skb: introduce kfree_skb_reason()")
drop reason is introduced to the tracepoint of kfree_skb. Therefore,
drop_monitor is able to report the drop reason to users by netlink.

The drop reasons are reported as string to users, which is exactly
the same as what we do when reporting it to ftrace.

Signed-off-by: Menglong Dong <[email protected]>
---
v8:
- pass drop reason to net_dm_packet_report_size()
- move drop reason validation to net_dm_packet_trace_kfree_skb_hit()

v7:
- take the size of NET_DM_ATTR_REASON into accounting in
net_dm_packet_report_size()
- let compiler define the size of drop_reasons

v6:
- check the range of drop reason in net_dm_packet_report_fill()

v5:
- check if drop reason larger than SKB_DROP_REASON_MAX

v4:
- report drop reasons as string

v3:
- referring to cb->reason and cb->pc directly in
net_dm_packet_report_fill()

v2:
- get a pointer to struct net_dm_skb_cb instead of local var for
each field
---
include/uapi/linux/net_dropmon.h | 1 +
net/core/drop_monitor.c | 41 +++++++++++++++++++++++++++-----
2 files changed, 36 insertions(+), 6 deletions(-)

diff --git a/include/uapi/linux/net_dropmon.h b/include/uapi/linux/net_dropmon.h
index 66048cc5d7b3..1bbea8f0681e 100644
--- a/include/uapi/linux/net_dropmon.h
+++ b/include/uapi/linux/net_dropmon.h
@@ -93,6 +93,7 @@ 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_REASON, /* 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 7b288a121a41..4641126b8a20 100644
--- a/net/core/drop_monitor.c
+++ b/net/core/drop_monitor.c
@@ -48,6 +48,19 @@
static int trace_state = TRACE_OFF;
static bool monitor_hw;

+#undef EM
+#undef EMe
+
+#define EM(a, b) [a] = #b,
+#define EMe(a, b) [a] = #b
+
+/* drop_reasons is used to translate 'enum skb_drop_reason' to string,
+ * which is reported to user space.
+ */
+static const char * const drop_reasons[] = {
+ TRACE_SKB_DROP_REASON
+};
+
/* net_dm_mutex
*
* An overall lock guarding every operation coming from userspace.
@@ -126,6 +139,7 @@ struct net_dm_skb_cb {
struct devlink_trap_metadata *hw_metadata;
void *pc;
};
+ enum skb_drop_reason reason;
};

#define NET_DM_SKB_CB(__skb) ((struct net_dm_skb_cb *)&((__skb)->cb[0]))
@@ -498,6 +512,7 @@ static void net_dm_packet_trace_kfree_skb_hit(void *ignore,
{
ktime_t tstamp = ktime_get_real();
struct per_cpu_dm_data *data;
+ struct net_dm_skb_cb *cb;
struct sk_buff *nskb;
unsigned long flags;

@@ -508,7 +523,11 @@ static void net_dm_packet_trace_kfree_skb_hit(void *ignore,
if (!nskb)
return;

- NET_DM_SKB_CB(nskb)->pc = location;
+ if ((unsigned int)reason >= SKB_DROP_REASON_MAX)
+ reason = SKB_DROP_REASON_NOT_SPECIFIED;
+ cb = NET_DM_SKB_CB(nskb);
+ cb->reason = reason;
+ cb->pc = location;
/* Override the timestamp because we care about the time when the
* packet was dropped.
*/
@@ -553,7 +572,8 @@ static size_t net_dm_in_port_size(void)

#define NET_DM_MAX_SYMBOL_LEN 40

-static size_t net_dm_packet_report_size(size_t payload_len)
+static size_t net_dm_packet_report_size(size_t payload_len,
+ enum skb_drop_reason reason)
{
size_t size;

@@ -574,6 +594,8 @@ static size_t net_dm_packet_report_size(size_t payload_len)
nla_total_size(sizeof(u32)) +
/* NET_DM_ATTR_PROTO */
nla_total_size(sizeof(u16)) +
+ /* NET_DM_ATTR_REASON */
+ nla_total_size(strlen(drop_reasons[reason]) + 1) +
/* NET_DM_ATTR_PAYLOAD */
nla_total_size(payload_len);
}
@@ -606,7 +628,7 @@ static int net_dm_packet_report_in_port_put(struct sk_buff *msg, int ifindex,
static int net_dm_packet_report_fill(struct sk_buff *msg, struct sk_buff *skb,
size_t payload_len)
{
- u64 pc = (u64)(uintptr_t) NET_DM_SKB_CB(skb)->pc;
+ struct net_dm_skb_cb *cb = NET_DM_SKB_CB(skb);
char buf[NET_DM_MAX_SYMBOL_LEN];
struct nlattr *attr;
void *hdr;
@@ -620,10 +642,15 @@ static int net_dm_packet_report_fill(struct sk_buff *msg, struct sk_buff *skb,
if (nla_put_u16(msg, NET_DM_ATTR_ORIGIN, NET_DM_ORIGIN_SW))
goto nla_put_failure;

- if (nla_put_u64_64bit(msg, NET_DM_ATTR_PC, pc, NET_DM_ATTR_PAD))
+ if (nla_put_u64_64bit(msg, NET_DM_ATTR_PC, (u64)(uintptr_t)cb->pc,
+ NET_DM_ATTR_PAD))
+ goto nla_put_failure;
+
+ if (nla_put_string(msg, NET_DM_ATTR_REASON,
+ drop_reasons[cb->reason]))
goto nla_put_failure;

- snprintf(buf, sizeof(buf), "%pS", NET_DM_SKB_CB(skb)->pc);
+ snprintf(buf, sizeof(buf), "%pS", cb->pc);
if (nla_put_string(msg, NET_DM_ATTR_SYMBOL, buf))
goto nla_put_failure;

@@ -679,7 +706,9 @@ static void net_dm_packet_report(struct sk_buff *skb)
if (net_dm_trunc_len)
payload_len = min_t(size_t, net_dm_trunc_len, payload_len);

- msg = nlmsg_new(net_dm_packet_report_size(payload_len), GFP_KERNEL);
+ msg = nlmsg_new(net_dm_packet_report_size(payload_len,
+ NET_DM_SKB_CB(skb)->reason),
+ GFP_KERNEL);
if (!msg)
goto out;

--
2.34.1



2022-02-09 19:10:15

by Ido Schimmel

[permalink] [raw]
Subject: Re: [PATCH v8 net-next] net: drop_monitor: support drop reason

On Wed, Feb 09, 2022 at 02:08:38PM +0800, [email protected] wrote:
> From: Menglong Dong <[email protected]>
>
> In the commit c504e5c2f964 ("net: skb: introduce kfree_skb_reason()")
> drop reason is introduced to the tracepoint of kfree_skb. Therefore,
> drop_monitor is able to report the drop reason to users by netlink.
>
> The drop reasons are reported as string to users, which is exactly
> the same as what we do when reporting it to ftrace.
>
> Signed-off-by: Menglong Dong <[email protected]>

Reviewed-by: Ido Schimmel <[email protected]>

2022-02-09 23:27:43

by David Ahern

[permalink] [raw]
Subject: Re: [PATCH v8 net-next] net: drop_monitor: support drop reason

On 2/8/22 10:08 PM, [email protected] wrote:
> From: Menglong Dong <[email protected]>
>
> In the commit c504e5c2f964 ("net: skb: introduce kfree_skb_reason()")
> drop reason is introduced to the tracepoint of kfree_skb. Therefore,
> drop_monitor is able to report the drop reason to users by netlink.
>
> The drop reasons are reported as string to users, which is exactly
> the same as what we do when reporting it to ftrace.
>
> Signed-off-by: Menglong Dong <[email protected]>
> ---
> v8:
> - pass drop reason to net_dm_packet_report_size()
> - move drop reason validation to net_dm_packet_trace_kfree_skb_hit()
>
> v7:
> - take the size of NET_DM_ATTR_REASON into accounting in
> net_dm_packet_report_size()
> - let compiler define the size of drop_reasons
>
> v6:
> - check the range of drop reason in net_dm_packet_report_fill()
>
> v5:
> - check if drop reason larger than SKB_DROP_REASON_MAX
>
> v4:
> - report drop reasons as string
>
> v3:
> - referring to cb->reason and cb->pc directly in
> net_dm_packet_report_fill()
>
> v2:
> - get a pointer to struct net_dm_skb_cb instead of local var for
> each field
> ---
> include/uapi/linux/net_dropmon.h | 1 +
> net/core/drop_monitor.c | 41 +++++++++++++++++++++++++++-----
> 2 files changed, 36 insertions(+), 6 deletions(-)
>

Reviewed-by: David Ahern <[email protected]>



2022-02-10 01:31:22

by patchwork-bot+netdevbpf

[permalink] [raw]
Subject: Re: [PATCH v8 net-next] net: drop_monitor: support drop reason

Hello:

This patch was applied to netdev/net-next.git (master)
by Jakub Kicinski <[email protected]>:

On Wed, 9 Feb 2022 14:08:38 +0800 you wrote:
> From: Menglong Dong <[email protected]>
>
> In the commit c504e5c2f964 ("net: skb: introduce kfree_skb_reason()")
> drop reason is introduced to the tracepoint of kfree_skb. Therefore,
> drop_monitor is able to report the drop reason to users by netlink.
>
> The drop reasons are reported as string to users, which is exactly
> the same as what we do when reporting it to ftrace.
>
> [...]

Here is the summary with links:
- [v8,net-next] net: drop_monitor: support drop reason
https://git.kernel.org/netdev/net-next/c/5cad527d5ffa

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html