Comment 8 for bug 44912

Hmm... it doesn't help. max_cstate was 8 by default, setting it to 1 seems to have no effect. (XOrg cpu usage still sits around 10% even with nothing but a top update every 2 secs) But something is screwy with the kernel's cpu usage reporting: I was trying to find out if the figures could be trusted and look:

This is on a Pentium M 1.6. Note the two bash instances at the top - they
are both running while true ; do true ; done loops.

This is a steady state....

Now /proc/cpuinfo shows only 1 cpu.... what is going on? Is this some new super new version of hyperthreading that has more instances?

top - 20:57:37 up 10:20, 5 users, load average: 1.59, 0.85, 0.57
Tasks: 138 total, 3 running, 135 sleeping, 0 stopped, 0 zombie
Cpu(s): 99.0% us, 0.4% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.4% hi, 0.3% si
Mem: 1294416k total, 696156k used, 598260k free, 135464k buffers
Swap: 979924k total, 0k used, 979924k free, 264580k cached

  PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
26176 brian 25 0 4536 2136 1324 R 99.9 0.2 9:10.68 bash
26190 brian 25 0 4540 2140 1324 R 99.9 0.2 6:45.28 bash
 4653 root 15 0 40384 30m 9384 S 6.0 2.4 507:28.82 Xorg
26167 root 16 0 2196 1132 856 R 3.7 0.1 0:03.59 top
 4582 brian 15 0 50616 22m 11m S 3.3 1.7 23:48.13 gnome-terminal
25528 brian 15 0 143m 77m 20m S 3.3 6.1 13:49.78 firefox-bin
 5655 brian 15 0 49716 29m 7516 S 2.3 2.3 16:02.64 gnome-netstatus

Hmm... this is silly. Now I've got 10 running, all showing 99.9% cpu.

When did the semantics of the cpu usage column change, or is this a bug?