Jump to content

Chameleon RC5 mode with mem detection enabled and automatic P-States & C-States generation for native power managment

kozlek

By kozlek,
in Chameleon

 Share
1,214 posts in this topic

Recommended Posts

Slice
Slice
Posted
Posted
Thank you Slice,

I managed to have it working using this modification, 

if (!(long long)mem_base+allocSize<1024*bootInfo->extmem+0x100000)

 

Since this branch is using dynamic library, file Symbols.dylib also changed automatically and need to be copied.

I didn't aware before and was struggling why my modification was not working.

Lesson learned.

Look, please, to my efforts:

http://www.projectosx.com/forum/index.php?showtopic=1106

Link to comment
Share on other sites
Memphiz
Memphiz
Posted
Posted

As i already stated here:

 

http://forum.voodooprojects.org/index.php/topic,1051.0.html

 

This conditional code was wrong. I thought my patch was included in trunk already, but maybe it was dropped again later...

 

Attached is the patch which should get it right for all cases.

 

ciao

 

Memphiz

 

PS: couldn't upload files with extension ".patch" ... so i renamed it to ".txt"

 

hibernatefix.txt

Link to comment
Share on other sites
Slice
Slice
Posted
Posted
As i already stated here:

 

http://forum.voodooprojects.org/index.php/topic,1051.0.html

 

This conditional code was wrong. I thought my patch was included in trunk already, but maybe it was dropped again later...

 

Attached is the patch which should get it right for all cases.

 

ciao

 

Memphiz

 

PS: couldn't upload files with extension ".patch" ... so i renamed it to ".txt"

 

hibernatefix.txt

This is wrong patch as discussed many times.

Link to comment
Share on other sites
Memphiz
Memphiz
Posted
Posted

@ Slice

 

Sometime ago there was a conditional if statement with a negation ("!") at this code. This negation was only applied to the membase because of missing paranthesis. My Patch fixed this by adding the paranthesis so that the whole comparison was negated (as intended). I didn't dig into the semantics here, because the error was obvious.

 

So as it seems there shouldn't be a negation at all. Of course my patch is superfluous than.

 

But it seems that there is still something wrong with the condition else there weren't users reporting the error.

 

keep up the good work

 

Memphiz

 

PS: this statement which seems to work for aikidoka25 is the wrong semantics i ment above

 

if (!(long long)mem_base+allocSize<1024*bootInfo->extmem+0x100000)

 

when compiled with gcc this would do the same as

 

if (1+allocSize<1024*bootInfo->extmem+0x100000)

 

the "1" comes from the bool expression "!(long long)mem_base"

 

see http://en.wikipedia.org/wiki/Operators_in_...ator_precedence for operator precedence here.

Link to comment
Share on other sites
aikidoka25
aikidoka25
Posted
Posted

thank you for the precision memphiz,

 

i tried slice's code above and it works too (hibernate with 3GB RAM), so i think it is better than the meaningless code i used before.

 

cheers

 

thank you for the precision memphiz,

 

i tried slice's code above and it works too (hibernate with 3GB RAM), so i think it is better than the meaningless code i used before.

 

cheers

Link to comment
Share on other sites
mitch_de
mitch_de
Posted
Posted

feature request (ANV has already + works) VID add for OCed system (Pstates gen part)

Would be really nice to have also in the chameleon trunk an key which does add some VID.

like PstatesMoreVID=2

In my case , the autogenerated VIDs (hex values known in dsdt PStates , often between 14..24hex) are for each Pstate little to low = instable cpu or KPs.

 

I think that code must be placed here (acpi_patcher.c) :

p_states[i ].VID = (((maximum.VID > 2) [i]+ PstatesMoreVID [/i];

 

could i try to make an "homemodded (fixed)" add VID by:

p_states[i ].VID = (((maximum.VID > 2) [b]+ 2[/b]; // would this add 2 to the autogen VIDs, so for example 16hex get 18hex ?

Link to comment
Share on other sites
dickhouse
dickhouse
Posted
Posted

Noobs need help. How to get ATI Radeon HD4670 QE/CI working (chameleon rc654) i have tried kalby's branch but got blank screen. If i use the onboard graphics (ATI HD 3200) i can enter osx, but if i use pci-x then blank screen..

i have turn off onboard controller and prefer to use the external one.. have tried all flag combination, -x -f PciRoot=1 but never pass..

 

inject device id into (ati4600controller.kext, atix2000.kext and use custom atisupport & atiframebuffer) no diff.

 

kexts in E\E: AHCIPortInjector, ElliotForceLegacyRTC, Ev0Reboot, IOAHCIBlockStorageInjector, SleepEnabler

 

did i miss something! please any guide/help would be appreciated..

 

Sorry dp, just cant solve it by my self

Link to comment
Share on other sites
valv
valv
Posted
Posted
feature request (ANV has already + works) VID add for OCed system (Pstates gen part)

Would be really nice to have also in the chameleon trunk an key which does add some VID.

like PstatesMoreVID=2

In my case , the autogenerated VIDs (hex values known in dsdt PStates , often between 14..24hex) are for each Pstate little to low = instable cpu or KPs.

 

I think that code must be placed here (acpi_patcher.c) :

p_states[i ].VID = (((maximum.VID << 2) - (vidstep * u)) >> 2) [i]+ PstatesMoreVID [/i];

 

could i try to make an "homemodded (fixed)" add VID by:

p_states[i ].VID = (((maximum.VID << 2) - (vidstep * u)) >> 2) [b]+ 2[/b]; // would this add 2 to the autogen VIDs, so for example 16hex get 18hex ?

ANV refers to andy; not the same :blink:

Other than that, this topic seems to be related to mozo's branch! For requests related to trunk ask Zef.

Link to comment
Share on other sites
Slice
Slice
Posted
Posted
ANV refers to andy; not the same ;)

Other than that, this topic seems to be related to mozo's branch! For requests related to trunk ask Zef.

Don't think so. The question is about p-states so it is for mozo.

Link to comment
Share on other sites
aikidoka25
aikidoka25
Posted
Posted

To meklort and others who use this branch,

 

I compiled the latest revision (i have 660) and during boot time there is a message

"Unable to bind symbol _acpi10_p" and it repeated if i pressed enter

 

After that the booter complained about unsupported cpu and P-States not generated

 

Continuing, the process will stop at line

mig_table_max_displ = 73

 

My unit is a HP Mini 311, with Atom processor, that's why I use this particular branch.

 

FYI: the unit is able to boot with revision 653

 

Any insight why the booter failed?

Link to comment
Share on other sites
valv
valv
Posted
Posted
Don't think so. The question is about p-states so it is for mozo.
yeah but he was asking to merge it to the trunk, and trunk is not mozo's branch. correct me if am wrong.
Link to comment
Share on other sites
Slice
Slice
Posted
Posted
yeah but he was asking to merge it to the trunk, and trunk is not mozo's branch. correct me if am wrong.

You are wrong. Latest mozo's patches already in trunc rather then in mozo's branch.

 

aikidoka25, check me.

Link to comment
Share on other sites
valv
valv
Posted
Posted
You are wrong. Latest mozo's patches already in trunc rather then in mozo's branch.

 

aikidoka25, check me.

That's because he decided to do so!

And believe me, Trunk is in noway one's dev branch. That's how we 'r organized. Everyone has his own branch. When something should stick, it is merged into trunk (collective branch so to say). For the sake of keeping things clean-and-working, most of us don't merge stuff directly into trunk (not before an collective/admin's decision). And I don't think mozo (with my respect to him) is admin (as of now).

Sorry to bring this up here, but mitch_de was requesting (mozo ?) to bring stuff from my branch into trunk. I was telling him to ask Zef (as admin), to avoid me/mozo/member_x to mess with the trunk without other members acceptance.

Don't get me wrong mate ;)

Link to comment
Share on other sites
mitch_de
mitch_de
Posted
Posted

Yep, valv is right what i wanted to say.

So i must ask the chameleon head devs to add this feature in the main trunk ( for my opionen really no mans alone code!!!) which is already an mix of many devs code .

THANKS.

PS: I asked first here, because in the past i nearly never get any answer when i used that chameleon forum.

Link to comment
Share on other sites
Slice
Slice
Posted
Posted
Yep, valv is right what i wanted to say.

So i must ask the chameleon head devs to add this feature in the main trunk ( for my opionen really no mans alone code!!!) which is already an mix of many devs code .

THANKS.

I want to say that p-states feature is under control of mozo. If he accepts your proposition he can propose to admin to include the patch into trunc. Other chameleons members will not consider is it good or not.

PS: I asked first here, because in the past i nearly never get any answer when i used that chameleon forum.

That's why I am not there.

There is no Team. There are some people making own versions of Chameleon.

What is the trunc? This is a stable featureless version. It is obsolete (?!).

Link to comment
Share on other sites
valv
valv
Posted
Posted
I want to say that p-states feature is under control of mozo.
No. Just go back in time to see: new cpus support feature (that c/p-states relies on it). It wasn't from mozo. And I never complained why things got merged to trunk without asking me. Instead I was glad :) And in no means I'd feel like some portion of code is under my control. This is sense-less.

Plus c/p-states weren't mozo's own invention: ported from SuperHai's efforts, that was inspired from other open-source OS's. U know that! I don't understand why people pertain to negatively criticize the team.

Sorry but that's not constructive.

The team is there, and with or without your consent, trunk is not obsolete, thanks to team members like meklort, kabyl, mozo... And forks are advantageous, as they serve the testing matters.

Greetz,

Link to comment
Share on other sites
shuvro
shuvro
Posted
Posted

Hi

Can any one tall why i am gating (boot1:/done1_)

After i install Chameleon RC5 ?

This is my partition

 

/dev/disk0

#: TYPE NAME SIZE IDENTIFIER

0: FDisk_partition_scheme *160.0 GB disk0

1: Windows_NTFS ACER 83.9 GB disk0s1

2: Apple_HFS MacOSX 31.8 GB disk0s2

3: DOS_FAT_32 DATA 44.3 GB disk0s3

 

Every time i boot it show this verbose log then boot:

boot0: done

boot1:/done1_

 

Can anyone help me to fix this pls

 

shuvro

Link to comment
Share on other sites
meklort
meklort
Posted
Posted
To meklort and others who use this branch,

 

I compiled the latest revision (i have 660) and during boot time there is a message

"Unable to bind symbol _acpi10_p" and it repeated if i pressed enter

 

I must have removed something (that symbol) that I shouldn't have in my branch. I'll take a look tonight. Also note that the topic isn't about my branch but about trunk.

Link to comment
Share on other sites
Time2Retire
Time2Retire
Posted
Posted

I think that having spin-offs of Chameleon is good for as long as Chameleon gets something back from it. I mean if it is interested enough for Chameleon is another thing, but I just found out (just one example) that kDefaultCachePath in boot.c is set wrong, apparently ever since the introduction of OS X 10.6 and as a result of it we never get to use the Kernel Cache.

Link to comment
Share on other sites
Slice
Slice
Posted
Posted
No. Just go back in time to see: new cpus support feature (that c/p-states relies on it). It wasn't from mozo. And I never complained why things got merged to trunk without asking me. Instead I was glad :P And in no means I'd feel like some portion of code is under my control. This is sense-less.

Plus c/p-states weren't mozo's own invention: ported from SuperHai's efforts, that was inspired from other open-source OS's. U know that! I don't understand why people pertain to negatively criticize the team.

Sorry but that's not constructive.

The team is there, and with or without your consent, trunk is not obsolete, thanks to team members like meklort, kabyl, mozo... And forks are advantageous, as they serve the testing matters.

Greetz,

:thumbsup_anim: P-states invented by Intel corporation. But in Chameleon the feature appears namely due to mozo. I know better then you because I was present ab ovo.

 

I think that having spin-offs of Chameleon is good for as long as Chameleon gets something back from it. I mean if it is interested enough for Chameleon is another thing, but I just found out (just one example) that kDefaultCachePath in boot.c is set wrong, apparently ever since the introduction of OS X 10.6 and as a result of it we never get to use the Kernel Cache.

I know.

/*
* Default path to kernel cache file
*/
//Slice - first one for Leopard
#define kDefaultCachePath "/System/Library/Caches/com.apple.kernelcaches/kernelcache"
#define kDefaultCachePathSnow "/System/Library/Caches/com.apple.kext.caches/Startup/kernelcache"
#define kDefaultCachePathTiger "/System/Library/Extensions.kextcache"

I just wonder why Chameleon team do not want to take this into account for a half a year.

Link to comment
Share on other sites
valv
valv
Posted
Posted
I just wonder why Chameleon team do not want to take this into account for a half a year.
am giving it a look :D
:P P-states invented by Intel corporation. But in Chameleon the feature appears namely due to mozo. I know better then you because I was present ab ovo.
;) yeah u are right! what I did not follow u in, was why should some pretend a feature is under control of one guy as long as it got merged into trunk. I don't see things so.

to end up with this crazy subject, the code on the trunk is for everyone. the only code that is being under mozo's control is on his branch. same thing goes to me and the other members. don't take me wrong mate, trunk is not for own and/or friends ego; it is there for stability.

sorry about that.

Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...