Chameleon

Issue 276: acpicodec - Unsupported CPU (Intel 3770K)

Reported by Luis Galhardo, Jul 25, 2012

Unsupported CPU
Unsupported CPU: P-States will not be generated !!!

Comment 1 by Evan Lojewski, Jul 25, 2012

Owner: cparm

Comment 2 by armel cadet-petit, Jul 25, 2012

hi, can you post the complete spec of your computer 
(cpu-motherboard-graphiccard-etc)

Comment 3 by armel cadet-petit, Jul 25, 2012

hi, can you post the complete spec of your computer 
(cpu-motherboard-graphiccard-etc)

Comment 4 by Luis Galhardo, Jul 25, 2012

Gigabyte GA-Z77X UD5H, 16GB RAM, GFX Gigabyte 560Ti, CPU Intel 3770K 
3.5Ghz, is it enough or do you need more detailed info ?

Comment 5 by armel cadet-petit, Jul 25, 2012

oh ok, it simply because i have not added support for ivy bridge 
cpus, stay tuned to this thread i will submit soon a beta version

Comment 6 by Luis Galhardo, Jul 25, 2012

Ok, thanks :)

Comment 7 by armel cadet-petit, Jul 28, 2012

hi Luis, and sorry for the delay,
can you test this version:

Comment 8 by Luis Galhardo, Jul 28, 2012

Seems to detect it fine, but CPU only works at 800 and 1600 Mhz -  
MSRDumper PStatesReached: 8 16   (Even using CPU Full Load it stays 
at 800 Mhz)

Attempting to autodetect CPU map from ACPI DSDT; wish me luck
Found 8 processors in ACPI, pmbase : 0x400, cpu_map_count : 8, 
namespace : _PR_
Found processors name : 
CPU0 CPU1 CPU2 CPU3 CPU4 CPU5 CPU6 CPU7

Comment 9 by armel cadet-petit, Jul 28, 2012

and this version ?

ps: can you dump the ssdt generated

Comment 10 by Luis Galhardo, Jul 28, 2012

Now stays at 1600 Mhz with CPU Load goes to 3900 Mhz, it going the 
right way :)

ps. any delays on source, seems that acpicode has slow debug output

Comment 11 by armel cadet-petit, Jul 28, 2012

ok please try this one (i need debug output to see what is happen, 
it will be deactivated on the final version)

Comment 12 by Luis Galhardo, Jul 28, 2012

Still the same, I think but now I got this:

P-States: min 0x10, max 0x23
state with cpu frequency :3500 and ratio :94 will be dropped
state with cpu frequency :3400 and ratio :84 will be dropped
state with cpu frequency :3300 and ratio :74 will be dropped
state with cpu frequency :3200 and ratio :64 will be dropped
state with cpu frequency :3100 and ratio :54 will be dropped
state with cpu frequency :3000 and ratio :44 will be dropped
state with cpu frequency :2900 and ratio :34 will be dropped
state with cpu frequency :2800 and ratio :24 will be dropped
state with cpu frequency :2700 and ratio :14 will be dropped
state with cpu frequency :2600 and ratio :4 will be dropped
state with cpu frequency :2500 and ratio :250 will be dropped
state with cpu frequency :2400 and ratio :240 will be dropped
state with cpu frequency :2300 and ratio :230 will be dropped
state with cpu frequency :2200 and ratio :220 will be dropped
state with cpu frequency :2100 and ratio :210 will be dropped
state with cpu frequency :2000 and ratio :200 will be dropped
No suitable P-states found, P-States will not be generated !!!

Comment 13 by armel cadet-petit, Jul 28, 2012

i think this one should work

Comment 14 by Luis Galhardo, Jul 28, 2012

With previous version I got 2 pstates that give 1600 and 3900 Mhz. 
Now I got 2 pstates that give 1600 and 3500 Mhz.

P-States: min 0x10, max 0x23
state :: frequency :3500 power: 1000000 translatency: 10 bmlatency: 
10 control: 35 status: 1 ratio :36 :: registred !! 
state :: frequency :3400 power: 1000000 translatency: 10 bmlatency: 
10 control: 34 status: 2 ratio :35 :: registred !! 
state :: frequency :3300 power: 960000 translatency: 10 bmlatency: 
10 control: 33 status: 3 ratio :34 :: registred !! 
state :: frequency :3200 power: 883000 translatency: 10 bmlatency: 
10 control: 32 status: 4 ratio :32 :: registred !! 
state :: frequency :3100 power: 845000 translatency: 10 bmlatency: 
10 control: 31 status: 5 ratio :31 :: registred !! 
state :: frequency :3000 power: 808000 translatency: 10 bmlatency: 
10 control: 30 status: 6 ratio :30 :: registred !! 
state :: frequency :2900 power: 737000 translatency: 10 bmlatency: 
10 control: 29 status: 7 ratio :28 :: registred !! 
state :: frequency :2800 power: 702000 translatency: 10 bmlatency: 
10 control: 28 status: 8 ratio :27 :: registred !! 
state :: frequency :2700 power: 635000 translatency: 10 bmlatency: 
10 control: 27 status: 9 ratio :25 :: registred !! 
state :: frequency :2600 power: 602000 translatency: 10 bmlatency: 
10 control: 26 status: 10 ratio :24 :: registred !! 
state :: frequency :2500 power: 570000 translatency: 10 bmlatency: 
10 control: 25 status: 11 ratio :23 :: registred !! 
state :: frequency :2400 power: 508000 translatency: 10 bmlatency: 
10 control: 24 status: 12 ratio :21 :: registred !! 
state :: frequency :2300 power: 478000 translatency: 10 bmlatency: 
10 control: 23 status: 13 ratio :20 :: registred !! 
state :: frequency :2200 power: 448000 translatency: 10 bmlatency: 
10 control: 22 status: 14 ratio :19 :: registred !! 
state :: frequency :2100 power: 391000 translatency: 10 bmlatency: 
10 control: 21 status: 15 ratio :17 :: registred !! 
state :: frequency :2000 power: 363000 translatency: 10 bmlatency: 
10 control: 20 status: 16 ratio :16 :: registred !!

Comment 15 by armel cadet-petit, Jul 28, 2012

humm, can you send me the ssdt file that you usually use

Comment 16 by Luis Galhardo, Jul 28, 2012

I don't use and ssdt

Comment 17 by Luis Galhardo, Jul 28, 2012

*any

Comment 18 by Luis Galhardo, Jul 29, 2012

I noticed this on log:

DSMOS has arrived
WARNING: IOPlatformPluginUtil : getCPUIDInfo: this is an unknown CPU 
model 0x3a
  -- power management may be incomplete or unsupported

Can it be because of this to not have linear speedstep ?

Comment 19 by armel cadet-petit, Jul 29, 2012

ok, can you boot without the acpicodec module, and dump the ssdt

Comment 20 by Luis Galhardo, Jul 29, 2012

ok, here it is

Comment 21 by armel cadet-petit, Jul 29, 2012

hi, the weird thing is that i don't see any pss method in any ssdt 
tables that you sent me, 

can you give me your boot.plist file ?

Comment 22 by Luis Galhardo, Jul 29, 2012

Here it is:

Comment 23 by armel cadet-petit, Jul 29, 2012

Oh, i forgot with acpipatcher there is to ssdt files generated , one 
for cst methods (the ssdt_without_acpicodec.dsl that you uploaded 
today) , and one for the pas methods, can you send me the second one

Comment 24 by Luis Galhardo, Jul 29, 2012

Tables extracted with DSDT Simple Editor

ssdt-3 has this:

"  Name (VERS, "Core i7 3.9 GHz Maximum Clock SSDT based 
on Greggen\'s at tonymacx86.com 2012-05-16")"

I don't know why !!! I used an ssdt from there a couple of days ago 
but I'm not using one anymore.

Comment 25 by Luis Galhardo, Jul 29, 2012

Rest of the files:

Comment 26 by Luis Galhardo, Jul 29, 2012

And the logs:

Comment 27 by armel cadet-petit, Jul 30, 2012

can you test this version

Comment 28 by Luis Galhardo, Jul 30, 2012

I can only test it in about 3 or 4 hours from now, I will report 
when I get home

Comment 29 by Luis Galhardo, Jul 30, 2012

Just tested the last version,and I get 2 pstates with 1600 and 3900 
Mhz which is better that the last result of 1600/3500 Mhz.

Comment 30 by armel cadet-petit, Jul 30, 2012

Can you dumping again your ssdt table, bdmesg, etc ,
but i am really surprised That your ssdt(genereted with acpicodec) 
dont contain any pss method, i not at home so i cant do anything for 
now, but the next debug version Will dump the ssdt directly in your 
bdmesg right after acpicodec generated it

Comment 31 by Luis Galhardo, Jul 30, 2012

error here: 

Building _CST Method
Error: SSDT Buffer Length exceeded available space 
Error: Failed to build SSDT table

Comment 32 by Luis Galhardo, Jul 30, 2012

Logs:

Comment 33 by armel cadet-petit, Aug 1, 2012

can you boot with this version, and dump the ssdt, bdmseg, etc ...

Comment 34 by armel cadet-petit, Aug 1, 2012

can you boot with this version, and dump the ssdt, bdmseg, etc ...

Comment 35 by Luis Galhardo, Aug 1, 2012

here it goes:

Comment 36 by armel cadet-petit, Aug 1, 2012

and this one ?
ps: if you have problems to boot, set dropssdt to NO

Comment 37 by Luis Galhardo, Aug 1, 2012

Well, I got the never ending boot, and dropssdt doesn't help.

Comment 38 by armel cadet-petit, Aug 2, 2012

i was meaning DropSSDT not dropssdt

Comment 39 by Luis Galhardo, Aug 2, 2012

I did use "DropSSDT=No" but it didn't work. Maybe because 
my system is running in raid 0, I don't know.... fyi, boot problem 
was fixed, I have a usb pen with boot loader

Comment 40 by armel cadet-petit, Aug 2, 2012

for many reason it's better to not boot from raid (or at least when 
you test a new bootloader),

this version is basically the same that allowed you to boot, if 
acpicodec fail, it will pause to let you take a screenshot, then 
just press any key to continue the boot process

Comment 41 by armel cadet-petit, Aug 2, 2012

Status: AwaitingInformation

Comment 42 by Luis Galhardo, Aug 2, 2012

This starts dumping and dumping and dumping data onscreen, I waited 
for about 30 minutes and still the same. Is this supposed to end ?

Comment 43 by armel cadet-petit, Aug 2, 2012

the good new is that if it dump something this mean the ssdt have 
been successfully generated, "the bad new" but it true 
with any chameleon release, is that we don't have a proper mtrr code 
for core I cpus, that why render can be very slow on this platforms,

here is exactly the same version without the ssdt dump(please don't 
forget to dump the ssdt and the bdmesg)

Comment 44 by Luis Galhardo, Aug 2, 2012

It's taking forever to extract SSDT-1 and even SSDT it's huge !

There are some _CST errors on dmesg as well as LPC error I think

Comment 45 by armel cadet-petit, Aug 2, 2012

something bizarre happened, i added some debug messages that i don't 
see here,
i re-up the good package, please be sure to replace acpicodec.dylib 
by the new one

Comment 46 by Luis Galhardo, Aug 2, 2012

Some errors on dmesg:

ACPI_SMC_PlatformPlugin::pushCPU_CSTData - _CST did not return a 
package
[AGPM Controller] build GPUDict by Vendor10deDevice1200
ACPI_SMC_PlatformPlugin::pushCPU_CSTData - _CST did not return a 
package
ACPI_SMC_PlatformPlugin::registerLPCDriver - WARNING - LPC device 
initialization failed: C-state power management not initialized

But now it was possible to extract ssdt data

Comment 47 by armel cadet-petit, Aug 2, 2012

we still have a bug with the ssdt generation (10 meg ???), but cst 
and pss method are here, can you try to add this in your boot.plist

and what's about your pstates?

Comment 48 by armel cadet-petit, Aug 2, 2012

oups, can you try to add this in your boot.plist:

<key>ACPIDropTables</key>
	<dict>
		<key>SSDT</key>
		<string></string>		
	</dict>

Comment 49 by Luis Galhardo, Aug 2, 2012

2 pstates 1600 and 3500 Mhz

Add what to my boot.plist ?!

Comment 50 by Luis Galhardo, Aug 2, 2012

oh, now I see what to add :)

Comment 51 by armel cadet-petit, Aug 3, 2012

so what's happen when you put it in your boot.plist ?? 

ps: can you also try GenerateCStates to NO and delete EnableSpeedStep

Comment 52 by Luis Galhardo, Aug 3, 2012

It was better this way, I think... less errors on dmesg

But I got this one:

"ACPI_SMC_PlatformPlugin::pushCPU_CSTData - _CST evaluation 
failed"

Comment 53 by armel cadet-petit, Aug 12, 2012

hi, can you send me your DSDT

Comment 54 by Luis Galhardo, Aug 12, 2012

here it is, but I don't use a dsdt.aml with this board

Comment 55 by armel cadet-petit, Aug 12, 2012

can you test this ssdt,

1) you must drop the oem ssdt with this for acpicodec:

<key>ACPIDropTables</key>
	<dict>
		<key>SSDT</key>
		<string></string>		
	</dict>
 
or with DropSSDT = Yes without the acpicodec module,

2) set GenerateCStates = No and GeneratePStates = No (with or 
without the acpicodec module)

Comment 56 by Luis Galhardo, Aug 13, 2012

System reboots this way

Comment 57 by armel cadet-petit, Aug 13, 2012

sorry, but you surely have done something wrong, in the worst case 
you should have a kernel panic,

i can't help you for this
Labels: Priority:Low Priority:Medium

Comment 58 by Luis Galhardo, Aug 13, 2012

I've tested it with no acpicodec module

On org.chameleon.Boot.plist, I've changed

- GenerateCStates = No and GeneratePStates = No 

DropSSDT I already had it set to Yes

and replaced the ssdt.aml with the one you send me.

I don't think I did something wrong, these are simple steps and 
since GenerateCStates and GeneratePStates = No don't reboot my 
system, them it seems that the fault was on this new ssdt.aml

Comment 59 by armel cadet-petit, Aug 13, 2012

maybe you should try with acpicodec

Comment 60 by Luis Galhardo, Aug 13, 2012

Yeah,  I'll do that later. Now it's too late and I have to go to 
work in a few hours. Thanks

Comment 61 by armel cadet-petit, Aug 15, 2012

have you tested with acpicodec ?

Comment 62 by Luis Galhardo, Aug 15, 2012

no yet, I´m having problems booting from my other hdd, and I don't 
want to mess with my raid 0 again.

Comment 63 by Cosmosis Jones, Oct 29, 2012

No Movement
Status: Invalid

Created: 11 years 7 months ago by Luis Galhardo

Updated: 11 years 4 months ago

Status: Invalid

Owner: armel cadet-petit

Followed by: 2 persons

Labels:
Type:Defect
Priority:Low