Re: acpi C2/3 problems?

Nate Lawson wrote:

>> Just to be sure, you're not claiming defaults/rc.conf on 6.x has the
>> setting for Cx to LOW.  I never MFCd that and just double-checked now.
>> 
>> You're saying that you had your own settings for that in rc.conf, and
>> that it stopped working correctly.

That is correct.
In /etc/rc.conf I had
###performance_cx_lowest="LOW"
###economy_cx_lowest="LOW"

which are now commented out as shown.


>> When did the regression occur exactly?

Cannot say exactly. Since its my laptop I dont keep full logs.
>From some root mails (from when I leave it on overnite) I see

>> FreeBSD 5.4-RELEASE #8: Fri Sep 16 09:50:58 EST 2005
>> FreeBSD 5.4-STABLE #0: Tue Oct  4 22:28:44 EST 2005
>> FreeBSD 5.4-STABLE #1: Sat Jan 14 10:37:35 EST 2006
>> FreeBSD 5.5-PRERELEASE #2: Thu Mar  2 20:45:25 EST 2006
>> FreeBSD 6.1-PRERELEASE #1: Fri Mar  3 07:04:00 EST 2006

root  wheel  7331390 Mar  2 19:59 checkouts.cvs:RELENG_5
root  wheel  8117041 Mar  2 22:05 checkouts.cvs:RELENG_6_0_0_RELEASE

I finally decided to upgrade on Thu Mar 2 so went to RELENG_5 as
intermediate step before going to 6-Release then to RELENG_6.
I did not notice anything stange on RELENG_5 or 6-release but I only ran
them  briefly before upgrading to RELENG_6. And while machine was busy
doing buildworlds/rebuilding ports it appeared to be ok.
It was definitely ok on 5.4-STABLE.

>> Nothing has changed in the acpi code in that area for years so
>> I'm surprised if it broke for you between 5.4 and 6.1.  I'm suspicious
>> that another device is remapping those IO ports or something.  Could you
>> send the output of devinfo -r on the broken and non-broken versions (say
>> use a 5.4 livecd or something)?

Here is the diff between the two outputs (which also are attached):

diff 5.4/devinfo-r 6.1/devinfo-r 
58,59c58,59
<     acpi_cmbat0
<     acpi_cmbat1
---
>     battery0
>     battery1
104a105,109
>         ehci0
>             I/O memory addresses:
>                 0xffa80800-0xffa80bff
>           usb4
>             uhub4
111a117,126
>             iwi0
>                 I/O memory addresses:
>                     0xdfbff000-0xdfbfffff
>         pcm0
>             I/O ports:
>                 0xec40-0xec7f
>                 0xed00-0xedff
>             I/O memory addresses:
>                 0xdffffd00-0xdffffdff
>                 0xdffffe00-0xdfffffff
133a149,150
>             ad0
>               subdisk0
136a154
>             acd0
166a185,191
>     pci_link0
>     pci_link1
>     pci_link2
>     pci_link3
>     pci_link4
>     pci_link5
>     pci_link6

thanks
--
tonym