Bug 232704 - [regresion] top -SH wrongly shows CPU/WCPU on fully-loaded system
Summary: [regresion] top -SH wrongly shows CPU/WCPU on fully-loaded system
Status: New
Alias: None
Product: Base System
Classification: Unclassified
Component: bin (show other bugs)
Version: CURRENT
Hardware: Any Any
: --- Affects Only Me
Assignee: freebsd-bugs mailing list
URL:
Keywords: regression
Depends on:
Blocks:
 
Reported: 2018-10-26 10:21 UTC by Lev A. Serebryakov
Modified: 2018-10-26 10:22 UTC (History)
1 user (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Lev A. Serebryakov freebsd_committer 2018-10-26 10:21:35 UTC
top -SH on 13-CURRENT (r339662) doesn't show NIC interrupts.
Here is screenshot under high network load, It could be seen, that 50% (one core of dual-core machine) is consumed by system, but there is no process/thread which consumes CPU:

# top -HSIzts1

last pid:  3261;  load averages:  0.60,  0.26,  0.23                                                                                                                                                                 up 0+12:55:34  13:17:05
117 threads:   3 running, 101 sleeping, 13 waiting
CPU:  0.0% user,  0.0% nice, 50.0% system,  0.0% interrupt, 50.0% idle
Mem: 4448K Active, 15M Inact, 198M Wired, 20M Buf, 3716M Free
Swap:

  PID USERNAME    PRI NICE   SIZE    RES STATE    C   TIME    WCPU COMMAND
    0 root        -76    -      0  2816K -        1   0:28   0.08% kernel{if_config_tqg_0}
   11 root        -60    -      0  1664K WAIT     1   0:23   0.07% intr{swi4: clock (0)}
 3169 lev          20    0    17M  7740K select   1   0:00   0.02% sshd
  698 root         20    0    18M    18M select   1   0:03   0.01% ntpd{ntpd}
    6 root        -16    -      0   128K -        1   0:02   0.00% rand_harvestq
   16 root        -16    -      0   384K psleep   1   0:02   0.00% pagedaemon{dom0}
   18 root        -16    -      0   384K qsleep   1   0:01   0.00% bufdaemon{bufdaemon}
  795 root        -16    -      0   128K nfscl    1   0:00   0.00% nfscl
  754 root         52    0    10M  2696K rpcsvc   1   0:00   0.00% nfscbd{nfscbd: master}
   19 root         16    -      0   128K syncer   1   0:01   0.00% syncer
   16 root        -16    -      0   384K umarcl   1   0:00   0.00% pagedaemon{uma}
   20 root        -16    -      0   128K vlruwt   1   0:00   0.00% vnlru
   18 root        -16    -      0   384K -        1   0:00   0.00% bufdaemon{bufspacedaemon-0}
   18 root        -16    -      0   384K -        1   0:00   0.00% bufdaemon{bufspacedaemon-1}

If load  is lower (one core is not consumed), CPU consumption is attributed to "kernel{if_config_tqg_0}" and/or "kernel{if_config_tqg_1}".

I can not reproduce it on 11.2-STABLE.