Kris Kennaway
2006-Mar-13 08:40 UTC
Problems with hw.acpi.cpu.cx_lowest=C2 (Re: cvs commit: src/etc/defaults rc.conf)
On Sun, Jan 29, 2006 at 01:30:07AM -0500, Kris Kennaway wrote:> On Sat, Jan 28, 2006 at 10:17:47PM -0800, Nate Lawson wrote: > > Kris Kennaway wrote: > > >On Sun, Jan 29, 2006 at 01:06:54AM -0500, Kris Kennaway wrote: > > > > > >>On Sun, Jan 29, 2006 at 05:51:58AM +0000, Nate Lawson wrote: > > >> > > >>>njl 2006-01-29 05:51:58 UTC > > >>> > > >>> FreeBSD src repository > > >>> > > >>> Modified files: > > >>> etc/defaults rc.conf > > >>> Log: > > >>> Enable the lowest Cx state by default. This will save power and we have > > >>> had enough testing of acpi_cpu to know this is stable now. > > >> > > >>On my desktop system (running RELENG_6 though), setting > > >>hw.acpi.cpu.cx_lowest=C0 causes atrocious performance. Is it broken > > >>in 6.x? > > > > > >C2, sorry. > > > > Ah, C0 should be disallowed already I thought (try it). > > > > As for C2, I MFCd a patch to acpi_cpu.c in November that should prevent > > this (1.57.2.1). Do you get a printf on console? > > This might be it: turns out I never rebooted to the newer kernel I > built earlier this month, and I am still running a kernel from Nov 6. > I'll get back to you.I finally managed to test this in RELENG_6. The system performance is not obviously bad, but sound playback is distorted (e.g. the 'bell' in KDE is much higher pitched than it should be, and on the console it is low pitched and lasting about 3 seconds). Nothing is logged on console. There may be other problems that I didn't notice right away, but the sound problems were enough to make me turn it off again. Kris -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://lists.freebsd.org/pipermail/freebsd-stable/attachments/20060313/a5d781e2/attachment.pgp
Nate Lawson
2006-Mar-13 17:04 UTC
Problems with hw.acpi.cpu.cx_lowest=C2 (Re: cvs commit: src/etc/defaults rc.conf)
Kris Kennaway wrote:> On Sun, Jan 29, 2006 at 01:30:07AM -0500, Kris Kennaway wrote: >> On Sat, Jan 28, 2006 at 10:17:47PM -0800, Nate Lawson wrote: >>> Kris Kennaway wrote: >>>> On Sun, Jan 29, 2006 at 01:06:54AM -0500, Kris Kennaway wrote: >>>> >>>>> On Sun, Jan 29, 2006 at 05:51:58AM +0000, Nate Lawson wrote: >>>>> >>>>>> njl 2006-01-29 05:51:58 UTC >>>>>> >>>>>> FreeBSD src repository >>>>>> >>>>>> Modified files: >>>>>> etc/defaults rc.conf >>>>>> Log: >>>>>> Enable the lowest Cx state by default. This will save power and we have >>>>>> had enough testing of acpi_cpu to know this is stable now. >>>>> On my desktop system (running RELENG_6 though), setting >>>>> hw.acpi.cpu.cx_lowest=C0 causes atrocious performance. Is it broken >>>>> in 6.x? >>>> C2, sorry. >>> Ah, C0 should be disallowed already I thought (try it). >>> >>> As for C2, I MFCd a patch to acpi_cpu.c in November that should prevent >>> this (1.57.2.1). Do you get a printf on console? >> This might be it: turns out I never rebooted to the newer kernel I >> built earlier this month, and I am still running a kernel from Nov 6. >> I'll get back to you. > > I finally managed to test this in RELENG_6. The system performance is > not obviously bad, but sound playback is distorted (e.g. the 'bell' in > KDE is much higher pitched than it should be, and on the console it is > low pitched and lasting about 3 seconds). Nothing is logged on > console. > > There may be other problems that I didn't notice right away, but the > sound problems were enough to make me turn it off again.Are you sure that's the cause? (Does setting cx_lowest to C1 fix it?) Can you send the output of sysctl hw.acpi so we can see how often each cx type is being run? -- Nate