DSDT - Vanilla Speedstep - Generic Scope (_PR)

[FKA]

thank u mate for response; already knew that

btw: would mind to take a look at

 

I don't think you need the Method (_PCC

 

See > 8.4.4.3 _PPC (Performance Present Capabilities) of the ACPI spec 4.0

 

As your P state values are defined I see no need for this dynamic call ?? Although i could be wrong!

 

D

 

EDIT:

 

If anything it should be returning (0)

            Method (_PPC, 0)
           {
               Return (0)
           }

[valv]

As your P state values are defined I see no need for this dynamic call ?? Although i could be wrong!

Really appreciate u've been taken time to look into that,

concerning _PPC, I think it is needed later on dsdt, cuz' when I was tryin to get rid of it got compiler errs, somethin like _PR.CPU0.PPC

but what about _PTC, do u know how to implement it (it is not present on ssdts), this is the method for throttling control.

u know I've managed to get my cpu running at least p-state. but the problem is it don't throttle to other states and gets very hot.

 

btw: has anyone over here the ssdts from mbp5,4. I know about Tea's on projectosx, I mean the ones he didn't include (cpu0cst..); was asking him already and he hadn't that macbook anymore.

 

btw: has anyone over here the ssdts from mbp5,4. I know about Tea's on projectosx, I mean the ones he didn't include (cpu0cst..); was asking him already and he hadn't that macbook anymore.

[FKA]

I presume you've tried setting to return (0) or as per the example in ACPI spec?

 

Method (_PPC, 0)    // Performance Present Capabilities method 
{ 
} // End of _PPC method

 

EDIT ... sorry I misread your question ... i thought you meant p-states not working !

 

there is a t-state example on page 319 of the ACPIscpec 4.0.

 

does your cpu support t-states ... especially as there is no _PTC in your acpi tables?!

 

Is it not c-states you need to see working _CST?

 

D

[valv]

there is a t-state example on page 319 of the ACPIscpec 4.0.

this could really solve it. not sure where to start from

does your cpu support t-states ... especially as there is no _PTC in your acpi tables?!

there where no t-states on original ssdts

Is it not c-states you need to see working _CST?

mate, I also got c-states working as u can see on IORegEx here

 

 

 

Where did you get "418" for your third C-state package? If that's C2, it's more likely to be "415"

 

mate I got it the linux way (also the everest way, and got the same thing), I copied what stated there.

 

btw: there where to much of c-states, so I cut off just to get c1,c2 and c3 in place

here's the interesting part of the original c-states:

Package (0x04)

{

ResourceTemplate ()

{

Register (SystemIO,

0x08, // Bit Width

0x00, // Bit Offset

0x0000000000000418, // Address

,)

},

 

0x03,

0xA2,

0x64

[FKA]

Possible you've got all you're going to get !?

 

D

[valv]

Possible you've got all you're going to get !?

yeah, but I can't get the freqs to change automatically without voodoopstate

[FKA]

yeah, but I can't get the freqs to change automatically without voodoopstate

 

right so you don't have p-states working!

 

Then i revert back to my previous comments about the Method (_PPC ... !

Also see Beerkex'd's comments on Model ID ...

 

Also try without coolbook and it's related kext ... use pstatechanger instead. I think coolbook bricked my p-states when i tried to use it.

 

D.

[valv]

right so you don't have p-states working!

but as I can understand from what stated on IORegExplorer, the p-states I specified earlier on my dsdt/ssdt got successfully passed to the OS.

I know this as I slightly altered those p-states after using pstate-calculator/changer.

 

Then i revert back to my previous comments about the Method (_PPC ... !

this is really wired as I need this, else I get errors upon compile

..DSDTFiles/dsdt.dsl  1252:                 Store (CUPS, \_SB.CPU0._PPC)
Error    4064 -                              Object does not exist ^  (\_SB.CPU0._PPC)

and what is there on line 1252 is into the _wak method

           If (LEqual (EIST, One))
           {
               Store (Zero, THR0)
               Store (Zero, Q8E0)
               Store (Zero, CUPS)
               Store (CUPS, \_SB.CPU0._PPC)
               If (CMPE)
               {
                   Notify (\_SB.CPU0, 0x80)
                   Notify (\_SB.CPU1, 0x80)
               }
               Else
               {
                   Notify (\_SB.CPU0, 0x80)
               }

there's also 4 more errs: under scope _tz (thermal zone)

Also see Beerkex'd's comments on Model ID ...

I already resored my model back to mbp5,4 as he suggested, removed the legacy smc plugin (as I was injecting a non exitent model based on the mbp2,2 specs, I know this had the throttle thing)

Also try without coolbook and it's related kext ... use pstatechanger instead. I think coolbook bricked my p-states when i tried to use it.

maybe you'r right on this, as I always kept coolbook to se if it worked, taking over the voodoopstate, cuz' I didn't found any other alternative to look for results (msrtools doesn't work in x64, sysctl gives max freq only, and on x32 with msrtools I can see speedstep is on)

btw: on the FACP table I found this scary line

[037h 0055  1]              P-State Control : 00

and in a mbp2,2 of a friend, this was:

[037h 0055  1]              P-State Control : 80

any thoughts ?!

[FKA]

OK Firstly try removing coolbook and kext and see what you get .. remember you're looking to see frequencies and voltages chaning in pstatechanger voodoomonitor!

 

If that doesn't work then I'm GUESSING the problem is with the PPC method.

As I've said don't remove the method (_PPC) first try to edit it as I satated in post 1353 and or as shown in post 1351

 

Let me know how it goes.

 

D.

[valv]

remember you're looking to see frequencies and voltages chaning in pstatechanger !

As far as I know from last tries, pstatechanger version 1.0.3 (1) would show nothing on the load pane without voodoopstate, and just hangs if I click on the info pane. that's why I started using coolbook (unregistered version, so no way it could touch something) /anyway will give it another try

are u people using another version of pstatechanger than I ?

 

EDIT: nope compile errors:

../DSDTFiles/dsdt.dsl  1915:             If (LNotEqual (CUPS, \_SB.CPU0._PPC ()))
Error    4061 -                          Called method returns no value ^

[FKA]

As far as I know from last tries, pstatechanger version 1.0.3 (1) would show nothing on the load pane without voodoopstate, and just hangs if I click on the info pane. that's why I started using coolbook (unregistered version, so no way it could touch something) /anyway will give it another try

are u people using another version of pstatechanger than I ?

 

Sorry my typo n... i meant voodoomonitor NOT Pstatechanger ... its linked on the first post !

[valv]

yey,

seems to be working

 

Big THANK U to all ya, especially u Formerlyknown

 

To any one trying every thing without succes, I suggest to review your bus speed and max freq with voodoo kernel

(I was wrong to trust cpu-z, it showed 266 although under voodoo kernel log I found it corrected to 280. when I changed the fsb every thing went flawlessly)

 

leo 10.5.8: idles at 25 Degrees

sno 10.6.2(64): idles at 30 Degrees

one more time thank u all; Formerly u'r a

 

btw: Vanilla Speedstep with oem acpi tables {only signatures wre modded}, u do it better ?!

[Silencers]

After updating to 10.6.3 I've noticed that CPU temperature jumped up 20 degrees! I've checked C-states - they were enabled, but MSR Tools showed that CPU frequency is stuck at maximum point (2.33GHz).

 

Then I've checked the difference between my legacy kext (block ACPI_SMC_PlatformPlugin/IOPlatformThermalProfile) and /S/L/E/IOPlatformPluginFamily.kext/Contents/PlugIns/ACPI_SMC_PlatformPlugin.kext/Contents/Resources/MacBookPro2_1.plist which is the one I use as a reference. And the difference was in CtrlLoopArray/0/StepDataDict/SP1 block. It is usually the same across all models, but it was changed in 10.6.3. After update and restart temperature has fallen back to normal.

 

For more info regarding legacy kext for vanilla speed step, please check my guide: http://www.insanelymac.com/forum/index.php?showtopic=211705

[valv]

2~3 C Degrees gained after 10.6.3

everything ok.

Though, even if I change my p-states in SSDT (that's the way I get'em loaded), under coolbook (unregistered version) they don't seem to be changed; to be clear here, speedstepping is still working but always with my old entries for p-states.

any thoughts?

[FKA]

2~3 C Degrees gained after 10.6.3

everything ok.

Though, even if I change my p-states in SSDT (that's the way I get'em loaded), under coolbook (unregistered version) they don't seem to be changed; to be clear here, speedstepping is still working but always with my old entries for p-states.

any thoughts?

 

are you saying your working with your native SSDT tables ?

 

Just FYI ... with my new MSI board I have vanilla speedstep without DSDT edit (dosen't work with MacPro4,1 model id though!).

Have you tested with voodoomonitor to see if you have speedstep is working without and DSDT edit or SSDT table override using your current model Mac ID?

 

D

[juanerson]

2~3 C Degrees gained after 10.6.3

everything ok.

Though, even if I change my p-states in SSDT (that's the way I get'em loaded), under coolbook (unregistered version) they don't seem to be changed; to be clear here, speedstepping is still working but always with my old entries for p-states.

any thoughts?

 

Uninstall coolbook 2.16 and erase the coolbook.app from Applications, then do a clean installation from coolbook.dmg --- placed to Applications and run it... done! You should see the p-states changes in frequencies and voltages. Seems this happens in every system update... and I am sure in every FID-VID SSDT modification for me.

[valv]

are you saying your working with your native SSDT tables ?

yep

Have you tested with voodoomonitor to see if you have speedstep is working without and DSDT edit or SSDT table override using your current model Mac ID?

already looked into voodoomonitor (without coolbook / cpu-x showing max freq) and it shows speedstep is working nicely without editing DSDT;

also vanilla SSDT shows speedstep. but not without SSDT override (extracting vanilla tables, changing memory adresses caught by using VK, and loadin'em back with AnV's branch of Chameleon).

 

As I said on an earlier post, it seems to be a problem with the fsb value. Somehow, as soon as I changed it from 266 to 280 Mhz on smbios.plist, I ended with every thing working; eventhough still seeing same freqs as old fsb value.

Having tried with Macbook4,1 first and that also worked oob / then switched back to mbp5,4 as it narrows my specs (look into my sig) and that also worked. Only got to leagcy my smc plugin.

Uninstall coolbook 2.16 and erase the coolbook.app from Applications, then do a clean installation from coolbook.dmg --- placed to Applications and run it... done! You should see the p-states changes in frequencies and voltages. Seems this happens in every system update... and I am sure in every FID-VID SSDT modification for me.

ok thank u juanerson, am gonna try this out and tell u back

[valv]

here we go again,

on SSDT-2 I've got _PSS & SPSS showing like this (before any edits)

Then, I edit my _PSS section using p-states calculator to end up with these

 

BTW: On VoodooMonitor I get 11 p-states whatever I change

with those values my lowest p-state is shown here

that means it doesn't take count of the last 3 p-states

 

Regarding FSB, my smbios.plist I inject those values

<key>SMexternalclock</key>
<string>280</string>
<key>SMmaximalclock</key>
<string>2527</string>

this is confirmed into System Profiler, as u can see here

But again VoodooPower shows those values

what the heck is this? why is voodooMonitor not showing the new FSB value I injected?

 

Juanerson,

I made exactly what u've been suggesting, and noway: I get always the same values back.

Something abnormal thoe: coolbook shows jumps between first and last p-state (for me those are: 1596Mhz/0.9375V & 2394Mhz/1.0625V)

[FKA]

what the heck is this? why is voodooMonitor not showing the new FSB value I injected?

 

possibly voodoomonitor is reading the FSB correctly rather than taking your SMBIOS injected value.

[valv]

possibly voodoomonitor is reading the FSB correctly rather than taking your SMBIOS injected value.

so how-comes to speedstep being only working when the FSB is (280Mhz) different from correct value (266Mhz) ?

If voodooMonitor is like you said readin' the FSB correctly, it wouldn't do better than kernel. so why this little change is making all the difference in my case ?

 

Formerly,

I know there is a part of truth in your answer, because even coolbook is showing the same as voodooMonitor.

but I cannot understand why speedstep happens to work correctly only when injecting a different value of FSB into smbios.plist.

[FKA]

Formerly,

I know there is a part of truth in your answer, because even coolbook is showing the same as voodooMonitor.

but I cannot understand why speedstep happens to work correctly only when injecting a different value of FSB into smbios.plist.

 

That I dont know .. it is odd!

[sephira]

Hi Guys, having a problem in that when I use pstatechanger to view my cpu states both the mV and FID columns only contain 0's. Also the VID's are 0d,0c,0b,0a,09 etc....

 

I am assuming something is going wrong....

 

Can anyone help me out?

 

My Processor is always running on max. Using MacBookPro5,5 smbios config.

 

Thanks very much!

[valv]

Hi Guys, having a problem in that when I use pstatechanger to view my cpu states both the mV and FID columns only contain 0's. Also the VID's are 0d,0c,0b,0a,09 etc....

 

I am assuming something is going wrong....

 

Can anyone help me out?

 

My Processor is always running on max. Using MacBookPro5,5 smbios config.

 

Thanks very much!

Are you using a legacy Apple_smc_plugin kext ? what did u put into ?

Could u post an ioreg file of your hack ?

[sephira]

Are you using a legacy Apple_smc_plugin kext ? what did u put into ?

Could u post an ioreg file of your hack ?

 

Nope no legacy kext that i am aware of.

 

I am running 10.6.3 is that likely to be a problem?

[valv]

Nope no legacy kext that i am aware of.

 

I am running 10.6.3 is that likely to be a problem?

Not really,

I suggest u take out the voodooPstate kext if u already have vanilla p-states, if not try to edit the Acpi key inside info.plist to Yes