[Devel] [PATCH RHEL7 COMMIT] ms/drop_monitor: Reset per-CPU data before starting to trace

Konstantin Khorenko khorenko at virtuozzo.com
Thu Jul 28 22:28:45 MSK 2022


The commit is pushed to "branch-rh7-3.10.0-1160.66.1.vz7.188.x-ovz" and will appear at https://src.openvz.org/scm/ovz/vzkernel.git
after rh7-3.10.0-1160.66.1.vz7.188.6
------>
commit b48f8e3b8ff20e8055a9d5d27fc7050f23837608
Author: Ido Schimmel <idosch at mellanox.com>
Date:   Mon Jul 4 21:51:59 2022 +0300

    ms/drop_monitor: Reset per-CPU data before starting to trace
    
    The function reset_per_cpu_data() allocates and prepares a new skb for
    the summary netlink alert message ('NET_DM_CMD_ALERT'). The new skb is
    stored in the per-CPU 'data' variable and the old is returned.
    
    The function is invoked during module initialization and from the
    workqueue, before an alert is sent. This means that it is possible to
    receive an alert with stale data, if we stopped tracing when the
    hysteresis timer ('data->send_timer') was pending.
    
    Instead of invoking the function during module initialization, invoke it
    just before we start tracing and ensure we get a fresh skb.
    
    This also allows us to remove the calls to initialize the timer and the
    work item from the module initialization path, since both could have
    been triggered by the error paths of reset_per_cpu_data().
    
    Signed-off-by: Ido Schimmel <idosch at mellanox.com>
    Signed-off-by: David S. Miller <davem at davemloft.net>
    (cherry picked from commit 44075f563793880d17d1a2f3a17e06835b157d07)
    
    Signed-off-by: Alexander Mikhalitsyn <alexander.mikhalitsyn at virtuozzo.com>
    
    =================
    Patchset description:
    drop_monitor: Add packet alert mode
    
    Just port and adapt packet alert mode feature for RHEL7 kernel.
    
    https://jira.sw.ru/browse/PSBM-140937
    
    Ported-by: Alexander Mikhalitsyn <alexander.mikhalitsyn at virtuozzo.com>
---
 net/core/drop_monitor.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/net/core/drop_monitor.c b/net/core/drop_monitor.c
index 41b558c589253..eaa1a0c7a0284 100644
--- a/net/core/drop_monitor.c
+++ b/net/core/drop_monitor.c
@@ -248,9 +248,16 @@ static int net_dm_trace_on_set(void)
 
 	for_each_possible_cpu(cpu) {
 		struct per_cpu_dm_data *data = &per_cpu(dm_cpu_data, cpu);
+		struct sk_buff *skb;
 
 		INIT_WORK(&data->dm_alert_work, send_dm_alert);
 		timer_setup(&data->send_timer, sched_send_work, 0);
+		/* Allocate a new per-CPU skb for the summary alert message and
+		 * free the old one which might contain stale data from
+		 * previous tracing.
+		 */
+		skb = reset_per_cpu_data(data);
+		consume_skb(skb);
 	}
 
 	rc = register_trace_kfree_skb(trace_kfree_skb_hit, NULL);
@@ -463,10 +470,7 @@ static int __init init_net_drop_monitor(void)
 
 	for_each_possible_cpu(cpu) {
 		data = &per_cpu(dm_cpu_data, cpu);
-		INIT_WORK(&data->dm_alert_work, send_dm_alert);
-		timer_setup(&data->send_timer, sched_send_work, 0);
 		spin_lock_init(&data->lock);
-		reset_per_cpu_data(data);
 	}
 
 


More information about the Devel mailing list