Bug 5282
Summary: | C1 not detected / usage not counted on NForce2 | ||
---|---|---|---|
Product: | ACPI | Reporter: | Prakash Punnoor (prakash) |
Component: | Power-Processor | Assignee: | Venkatesh Pallipadi (venki) |
Status: | REJECTED INVALID | ||
Severity: | normal | CC: | acpi-bugzilla |
Priority: | P2 | ||
Hardware: | i386 | ||
OS: | Linux | ||
Kernel Version: | 2.6.14-rc2 | Subsystem: | |
Regression: | --- | Bisected commit-id: |
Description
Prakash Punnoor
2005-09-21 01:08:11 UTC
i expect that 2.6.11 was before the original issue was fixed. was this working in 2.6.12 or 2.6.13 on this box? AFAIK no, both 2.6.12 and .13 were showing the same behaviour. *C1: type[C1] promotion[--] demotion[--] latency[000] usage[00000000] Says C1 is detected and ACPI idle routine is active. Issue is that the PCu is not going idle for some reason. Do you see system response and performance normal? Anything suspicious in /proc/interrupts ? Aargh, I seem to have resolved it. Seems like user error: I had the option idle=halt as kernel parameter. I took it away and now it works. What I am wondering though, with the old patch in the old bug report 2635, the C1 counting did work with idle=halt. I wonder whether something in the Documentation is missing, as I can only find idle=poll and idle=halt documented... BTW, I didn't notice anything strange when C1 wasn't working. Yes. That part has been changed recently. The logic is, by default acpi_processor_idle() will be used for idling and /proc/acpi interface will report the idle counts. But, if user chooses any specific idle routines, then acpi idle routine will not be used and hence counts wont appear in the /proc/acpi interface. As it is working without idle=halt, I am closing the bug. |