[LU-10807] ksocknal_reaper() jitter on b2_10 Created: 12/Mar/18  Updated: 08/Nov/19  Resolved: 08/Nov/19

Status: Resolved
Project: Lustre
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Bug Priority: Minor
Reporter: John Hammond Assignee: John Hammond
Resolution: Duplicate Votes: 0
Labels: None

Issue Links:
Related
is related to LU-10707 TCP eth routed LNet traffic broken Resolved
Severity: 3
Rank (Obsolete): 9223372036854775807

 Description   

On b2_10, ksocknal_reaper() needs the following fix which landed to master as part of LU-9019 lnd: remove remaining cfs_time wrappers:

diff --git a/lnet/klnds/socklnd/socklnd_cb.c b/lnet/klnds/socklnd/socklnd_cb.c
index ceb8254..078638f 100644
--- a/lnet/klnds/socklnd/socklnd_cb.c
+++ b/lnet/klnds/socklnd/socklnd_cb.c
@@ -2627,7 +2627,7 @@ int ksocknal_reaper(void *arg)
                if (!ksocknal_data.ksnd_shuttingdown &&
                    list_empty(&ksocknal_data.ksnd_deathrow_conns) &&
                    list_empty(&ksocknal_data.ksnd_zombie_conns))
-                       schedule_timeout(cfs_duration_sec(timeout));
+                       schedule_timeout(cfs_time_seconds(timeout));
 
                set_current_state(TASK_RUNNING);
                remove_wait_queue(&ksocknal_data.ksnd_reaper_waitq, &wait);

Without it the ksocknal_reaper thread will be using 1% of a CPU at all times.



 Comments   
Comment by Gerrit Updater [ 13/Apr/18 ]

John L. Hammond (john.hammond@intel.com) uploaded a new patch: https://review.whamcloud.com/31992
Subject: LU-10807 lnet: correct schedule timeout in ksocknal_reaper()
Project: fs/lustre-release
Branch: b2_10
Current Patch Set: 1
Commit: 766ab5c778063671e780963896477f8105945748

Comment by Andreas Dilger [ 08/Nov/19 ]

Fixed by patch https://review.whamcloud.com/32015 "LU-10707 ksocklnd: revert back to jiffies".

Generated at Sat Feb 10 02:38:21 UTC 2024 using Jira 9.4.14#940014-sha1:734e6822bbf0d45eff9af51f82432957f73aa32c.