http://bugzilla.kernel.org/show_bug.cgi?id=14771 Summary: "ondemand" never raises frequency on an Intel Core2 Due (T9900) in a recent Dell E6500 Product: Power Management Version: 2.5 Kernel Version: 2.6.32 Platform: All OS/Version: Linux Tree: Mainline Status: NEW Severity: normal Priority: P1 Component: cpufreq AssignedTo: cpufreq@xxxxxxxxxxxxxxx ReportedBy: vapier@xxxxxxxxxx Regression: No i just got a new laptop (Dell E6500 w/T9900 and bios A18) and was in the process of migrating from the old one (Dell D820 w/T2500). since it's a fresh machine, i dont have any references for when things were working. but my old machine works fine (largely same kernel settings, and same kernel version). only significant difference is that the new one is 64bit while the old one is 32bit. i have not disabled Intel SpeedStep in the BIOS. i tried disabling various kconfig options to see if it would make a difference (virtualization, hpet, tickless, cpufreq tables), but alas, it does not. 2.6.31.6 and 2.6.32 behave the same way. by default, i have it boot up with the performance governor. then i run: # cd /sys/devices/system/cpu # grep -r . cpu?/cpufreq/ cpu0/cpufreq/cpuinfo_min_freq:800000 cpu0/cpufreq/cpuinfo_max_freq:3068000 cpu0/cpufreq/cpuinfo_transition_latency:10000 cpu0/cpufreq/scaling_min_freq:800000 cpu0/cpufreq/scaling_max_freq:3068000 cpu0/cpufreq/affected_cpus:0 cpu0/cpufreq/related_cpus:0 1 cpu0/cpufreq/scaling_governor:performance cpu0/cpufreq/scaling_driver:acpi-cpufreq cpu0/cpufreq/scaling_available_governors:conservative ondemand powersave performance cpu0/cpufreq/scaling_setspeed:<unsupported> cpu0/cpufreq/scaling_available_frequencies:3068000 3067000 2134000 1600000 800000 cpu0/cpufreq/cpuinfo_cur_freq:3067000 cpu0/cpufreq/scaling_cur_freq:3068000 cpu1/cpufreq/cpuinfo_min_freq:800000 cpu1/cpufreq/cpuinfo_max_freq:3068000 cpu1/cpufreq/cpuinfo_transition_latency:10000 cpu1/cpufreq/scaling_min_freq:800000 cpu1/cpufreq/scaling_max_freq:3068000 cpu1/cpufreq/affected_cpus:1 cpu1/cpufreq/related_cpus:0 1 cpu1/cpufreq/scaling_governor:performance cpu1/cpufreq/scaling_driver:acpi-cpufreq cpu1/cpufreq/scaling_available_governors:conservative ondemand powersave performance cpu1/cpufreq/scaling_setspeed:<unsupported> cpu1/cpufreq/scaling_available_frequencies:3068000 3067000 2134000 1600000 800000 cpu1/cpufreq/cpuinfo_cur_freq:3067000 cpu1/cpufreq/scaling_cur_freq:3068000 then i enable the ondemand governor: # cpufreq-set -g ondemand -c 0 # cpufreq-set -g ondemand -c 1 then i check the current frequency: # grep -r . cpu?/cpufreq/*_cur_freq cpu0/cpufreq/cpuinfo_cur_freq:800000 cpu0/cpufreq/scaling_cur_freq:800000 cpu1/cpufreq/cpuinfo_cur_freq:800000 cpu1/cpufreq/scaling_cur_freq:800000 this is fine; i'm currently idling. then i throw cpuburn at the system and wait a little bit (so the load goes up to ~4.0): # burnP6& burnP6& burnP6& burnP6& (and top does show 4 burnP6's with each using 50% of the system) the frequency never increases: # uptime 19:38:29 up 23 min, 2 users, load average: 4.13, 3.90, 2.89 # grep -r . cpu?/cpufreq/*_cur_freq cpu0/cpufreq/cpuinfo_cur_freq:800000 cpu0/cpufreq/scaling_cur_freq:800000 cpu1/cpufreq/cpuinfo_cur_freq:800000 cpu1/cpufreq/scaling_cur_freq:800000 same behavior can be seen when recompiling my kernel with `make -j4` -- the frequency stays at 800MHz and it takes much longer (obviously) i noticed that if have the governor set to performance and at the max freq, then start loading the cpu (with 4 x burnP6), then set the governor to ondemand, the freq will stay at 3.0GHz until i kill the burnP6's. then the freq falls again and doesnt come back up. if i force the "up_threshold" down to like 70 (80+ didnt help), then the speed starts bouncing up to the max 3.0GHz as i'd expect. the problem here is that it means it'll also bounce to the max when doing simpler things like `ps aux` that largely dont need the higher speeds. -- Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug. -- To unsubscribe from this list: send the line "unsubscribe cpufreq" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
