fdo-mirrors/NetworkManager
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/NetworkManager/NetworkManager.git synced 2026-01-07 02:00:15 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions

platform: add small backoff time before resync

Browse source 
If the socket's RX buffer is full it's probably because other
process is doing lot of changes very quickly, faster than we
can process them. Let's give the writer a small time to finish:
1. Avoid contending the kernel's RTNL lock, so we don't make
   the whole situation even worse and it can finish earlier.
2. Avoid having to resync again and again due to trying to
   resync while the writer is still doing quick changes, so
   we are unable to catch up yet.

This won't help if this situation takes a long time or is
continuous, but that's unlikely to happen, and if it does,
it's the writer's fault for starving the whole system.

There is no need to progresively increase the backoff time
for the same reason: if this situation takes lot of time,
it's the writer's fault. It's neither a good idea because the whole NM
process will end being sleeping long times, not doing anything at all,
without being able to react when the Netlink messages burst stops.
This commit is contained in:
Íñigo Huguet 2024-08-12 12:04:04 +02:00
parent 926bfab5b5
commit 830dd4ad9c
1 changed files with 16 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
src/libnm-platform/nm-linux-platform.c
View file

@ -361,7 +361,8 @@ struct _ifla_vf_vlan_info {
/*****************************************************************************/
#define RESYNC_RETRIES 50
#define RESYNC_RETRIES 50
#define RESYNC_BACKOFF_SECONDS 1
/*****************************************************************************/
@ -11203,6 +11204,20 @@ event_handler_read_netlink(NMPlatform *platform,
}
_reason;
}));
if (nle == -ENOBUFS) {
/* Netlink notifications are coming faster than what
* we can process them. Backoff a bit so we give some
* time for this burst to finish, and we don't
* contribute to starve the system contending for the
* kernel's RTNL lock.
*/
_LOGI("netlink[%s]: backoff for %d seconds before the resync.",
nmp_netlink_protocol_info(netlink_protocol)->name,
RESYNC_BACKOFF_SECONDS);
sleep(RESYNC_BACKOFF_SECONDS);
}
_netlink_recv_handle(platform, netlink_protocol, FALSE);
delayed_action_wait_for_nl_response_complete_all(
platform,