acpi_video on SONY VAIO PCG-Z1 [was: acpi_video users needed]

[takawata at jp.freebsd.org]

In message <1094031813.903.13.camel at localhost>, Vladimir Grebenschikov wrote:
>On Wed, 2004-09-01 at 12:43 +0400, Toxa wrote:
>> On Wed, Sep 01, 2004 at 11:38:57AM +0900, takawata at jp.freebsd.org wrote:
>> 
>> > I don't imagine why. But any other method in SNC may do something.
>> > I modified the driver so that it exports more methods.
>> > http://www.init-main.com/acpi_snc2.tar.gz
>> 
>> > Testers wanted.
>> 
>> sony vaio pcg-v505bx...
>
>sony vzio pcg-z1aw
>
>almost same:
>
># sysctl dev.acpi_snc
>dev.acpi_snc.0.brightness: 98
>dev.acpi_snc.0.ctr: 0
>dev.acpi_snc.0.pcr: 0
>dev.acpi_snc.0.cmi: -1044294180
>dev.acpi_snc.0.wdp: 1281
>dev.acpi_snc.0.cdp: 1
>dev.acpi_snc.0.%desc: Sony notebook controller
>dev.acpi_snc.0.%driver: acpi_snc
>dev.acpi_snc.0.%location: handle=\_SB_.PCI0.LPCB.SNC_
>dev.acpi_snc.0.%pnpinfo: _HID=SNY5001 _UID=0
>dev.acpi_snc.0.%parent: acpi0
># 
>
>dmesg message appears:
>    ACPI-1303: *** Error: Method execution failed [\
>\_SB_.PCI0.LPCB.SNC_.GCMI] (Node 0xc1a3a6a0), AE_AML_UNINITIALIZED_ARG
>
>
>on each 
># sysctl dev.acpi_snc.0.cmi

This value should not exported in this way, it seems.
Please comment out the entry in the array for defining sysctl value.

>value of 
>dev.acpi_snc.0.brightness
>change real brightness of screen but by very strange law:
>
># i=1; while [ $i -le 100 ]; do echo "try $i"; sysctl
>dev.acpi_snc.0.brightness=$i; echo -n 'read value '; setbrightness --
>read; i=$(($i+1)); done
>read value 248

(snip)

>Real display brightness changed according "read value %d", so while this
>operations screen brightness was jumping.

setbritness is dangerous with this driver, because it use same register
without locking.
How about real brightness change?