Jump to content

Clover General discussion

ErmaC

By ErmaC,
in Clover

 Share
29,866 posts in this topic

Recommended Posts

RehabMan
RehabMan
Posted
Posted

Any conclusion here? HWP is turned off definitely after sleep. :(

 

It seems PC UEFI/BIOS make CPU initialization not as we want for macOS drivers. New kext needed.

 

I just tested on my Skylake Intel NUC6i7KYK.

I'm using iMac17,1 with a modified plist in X86PlatformPlugin (one that enables HWP).

I used AppleIntelInfo.kext to determine HWP enablement before and after sleep.

My tests confirm that HWP is enabled after sleep.

 

My procedure:

- enable HWP (using an X86PlatformPlugin "symlink" injector kext with modified Resources/board-id plist)

- use kextutil to load AppleIntelInfo.kext prior to sleep

- check for HWP enablement in AppleIntelInfo output

- unload AppleIntelInfo.kext with kextunload

- rm /tmp/AppleIntelInfo.dat (just to make sure)

- sleep/wake

- use kextutil to load AppleIntelInfo.kext prior to sleep

- check for HWP enablement in AppleIntelInfo output

  • Like 1
Link to comment
Share on other sites
RehabMan
RehabMan
Posted
Posted

I saw iMac17,1, iMac18,x and even iMac19,1 (the coming iMac Pro) have no "hwp" power feature in their FreqeuncyVectors, unlike MacBook(Pro).

Maybe just like @Slice said, the mechanism is a little bit different...

 

It is why I added HWP to my iMac17,1 X86PlatformPlugin injector kext.

  • Like 2
Link to comment
Share on other sites
MattsCreative
MattsCreative
Posted
Posted

 

hey wanted to tell you about a issue with 10.13 for my 290x 

 

so i use a bios mod to use multi monitors and have sleep work oob 

it worked in 10.12 and to 10.10 

the issue is when i try and boot with the bios mod on it never happens anymore so did anything change gfx wise with 10.13 and clover because ozmosis can still boot bios mod way

Link to comment
Share on other sites
PMheart
PMheart
Posted
Posted

It is why I added HWP to my iMac17,1 X86PlatformPlugin injector kext.

Yes... I've also done the same thing a little bit earlier. (http://www.insanelymac.com/forum/topic/320517-pre-release-macos-sierra-10124/page-5?do=findComment&comment=2368367)

But, has Apple really enabled HWP for their real iMac? Unfortunately I don't have such models and therefore I can't check it. IMO, it's very likely that, Apple doesn't enable HWP for iMac.

Link to comment
Share on other sites
RehabMan
RehabMan
Posted
Posted

Yes... I've also done the same thing a little bit earlier. (http://www.insanelymac.com/forum/topic/320517-pre-release-macos-sierra-10124/page-5?do=findComment&comment=2368367)

But, has Apple really enabled HWP for their real iMac? Unfortunately I don't have such models and therefore I can't check it. IMO, it's very likely that, Apple doesn't enable HWP for iMac.

 

iMac17,1 is not enabled by Apple provided plists.

I haven't checked the other new iMac plists.

You can check easily with freqVectorsEdit.sh.

Link to comment
Share on other sites
PMheart
PMheart
Posted
Posted

iMac17,1 is not enabled by Apple provided plists.

I haven't checked the other new iMac plists.

You can check easily with freqVectorsEdit.sh.

Well, I've checked it earlier today, the same on iMac18,x and iMac19,1, no HWP support in their FrequencyVectors.

So thing is. Why Apple don't enable HWP for iMac...

Link to comment
Share on other sites
RehabMan
RehabMan
Posted
Posted

Well, I've checked it earlier today, the same on iMac18,x and iMac19,1, no HWP support in their FrequencyVectors.

So thing is. Why Apple don't enable HWP for iMac...

They probably find it can affect performance negatively on desktops or has no positive effect at all...  I have seen the same...

Point is though... easy to add HWP to your chosen SMBIOS... and then it will be restored across sleep/wake cycle.

Link to comment
Share on other sites
PMheart
PMheart
Posted
Posted

They probably find it can affect performance negatively on desktops or has no positive effect at all...  I have seen the same...

Point is though... easy to add HWP to your chosen SMBIOS... and then it will be restored across sleep/wake cycle.

In this case, I think it's better for us to follow Apple, not enabling HWP for now on desktops.

Ah yes... Just adding related power features.

It won't be too complex I guess, with the help of some kexts that help us write corresponding MSRs.

 

(I don't own any Skylake+ desktops to test though...)

Without speedstep and without HWP I have 3300MHz only and T=30 maximum. So yes, I wish not reduce performance for the aim of CPU cooling.

IMO it must bring negative effects to performance without SpeedStep. HWP can be just a small part of PM, it doesn't really matter. :)

  • Like 1
Link to comment
Share on other sites
RehabMan
RehabMan
Posted
Posted

In this case, I think it's better for us to follow Apple, not enabling HWP for now on desktops.

Ah yes... Just adding related power features.

It won't be too complex I guess, with the help of some kexts that help us write corresponding MSRs.

 

(I don't own any Skylake+ desktops to test though...)

IMO it must bring negative effects to performance without SpeedStep. HWP can be just a small part of PM, it doesn't really matter. :)

 

No special kext necessary for HWP enabling (including after sleep/wake).  The code is already in the kernel/X86PlatformPlugin.

Agree probably best to avoid HWP on desktop (even though it works), as Apple is not using it.  But it is still interesting to experiment with it...

  • Like 1
Link to comment
Share on other sites
mashny
mashny
Posted
Posted

Same here.

 

Shouldn't this version removed ASAP? r4094 works fine.

 

You got it to upgrade with r4094? I couldn't find 94 but I just tried 97 but no luck. Care to provide the download link for 4094?

 

Edit : Just tried 4094. No dice

Link to comment
Share on other sites
schot
schot
Posted
Posted

Do you have EFI backup folder on main partition?. These older settings in config.plist may stop new settings in the EFI partition to take hold. Both EFI and EFI backup config.plist should match.

 

If you have 2 disks and both have clover bootloaders installed, only one bootloader will be detected usually the last one used for booting or from the last booted volume. Try disconnecting the drive you don't want the clover bootloader to load from. Start computer wait for clover screen then attach other drive, reboot and your selected clover bootloader should boot either drive with the config.plist you want to use.

I don't think is is clear what I mean.

 

 

I have 1 SSD with Sierra, High Sierra and Windows. One EFI partition on that same disk with clover.

I have a config.plist, config_sierra.plist, and config_HS.plist in EFI/Clover

I have config.plist file is a duplicate of config_sierra.plist, I have a AMD RX480 that needs kext patches in Sierra.

But... not in High Sierra, also using some patches and fakeID's that are not needed for High Sierra.

So when clover bootmenu comes up, I go to Options -> config -> and choose my config_HS from the list.

My bootargs are changed (npci=0x3000 to 0x2000) rest is the same, but that is also in the config files. When I look at the kext patches I see all patches from config.plist file and not from config_HS.plist which I actually selected, which actually should be there because I selected another config file.

I know when testing Sierra in DP state this was not there! It worked!

I don't think this has anything to do with a EFI-backups on my Sierra partition, because it loads the EFI partition.

By the way, I only load UEFI boots so no legacy!

 

If I could program I would definitely look into this but sorry I don't know nothing about programming ;-)

Also don't know when this broke.... I only know it was working before in r30XX but I don't know till which version, normally I do not need it only now.

I use now refi shell to switch the config names and reboot or boot into the latest state and replace it but it should be great if some of you smart guys and girls can fix this.

Link to comment
Share on other sites
apianti
apianti
Posted
Posted

So....... Slight set back on version 3. My laptop and phone were stolen, and my laptop had all of my coding projects on it. I hadn't pushed any changes in a while and now I'm basically rewriting all the code since what's in the repo was really far behind - but I guess that's what I get for not keeping the repo up to date. Also since I don't have my laptop anymore, I have a much more limited amount of time to code on my desktop. For the love of god, if anyone wants to help me out, please, please, please, do, I don't even have a phone right now.

Link to comment
Share on other sites
Cyberdevs
Cyberdevs
Posted
Posted

hello clover users.

 

has anyone nehalem cpu?

 

ex i7 870 or i5 750.

or gulftown etc

 

i need clover log. if has these cpu, please upload clover log here. thanks in advance

Sorry I don't have any Nehalem processors  :(

Link to comment
Share on other sites
savvas
savvas
Posted
Posted

No clue what they are talking about, but yes, ApfsJumpStart is what loada APFS.efi... it's a small driver btw.

EDIT: Yep, he's on track. ;)

I loaded apfs.efi using efi_disk_io_protocol.  But what about FV2 encrypted system partition? 

I have no idea how Apple load apfs.efi from FV2 encrypted partition using block io..

Link to comment
Share on other sites
crusher
crusher
Posted
Posted

I loaded apfs.efi using efi_disk_io_protocol.  But what about FV2 encrypted system partition? 

I have no idea how Apple load apfs.efi from FV2 encrypted partition using block io..

On my real Mac work fv2...

Link to comment
Share on other sites
mhaeuser
mhaeuser
Posted
Posted

I loaded apfs.efi using efi_disk_io_protocol.  But what about FV2 encrypted system partition? 

I have no idea how Apple load apfs.efi from FV2 encrypted partition using block io..

I don't know how Jumpstart works (I have 0 experience/knowledge with/about FSs), maybe the Jumpstart record is unencrypted? no idea :)

Did you reverse Jumpstart?

Link to comment
Share on other sites
savvas
savvas
Posted
Posted

I don't know how Jumpstart works (I have 0 experience/knowledge with/about FSs), maybe the Jumpstart record is unencrypted? no idea :)

Did you reverse Jumpstart?

Yep. 

Jump start uses ApplePartitionInfo proto to determine information to do jump, so thats why it doesn't work properly "OOB" due to Clover doesn't have such proto.

Seems that apfs.efi driver embedded inside  container, not partition.

Check that code portion

APFSContainerEFIEmbed(wholeDiskBSD, (const char *)CFDataGetBytePtr(apfsDriverData), CFDataGetLength(apfsDriverData));

 

/*!

 * @define kBL_PATH_I386_APFS_EFI
 * @discussion APFS driver for Darwin x86 on EFI-based systems
 */
#define kBL_PATH_I386_APFS_EFI "/usr/standalone/i386/apfs.efi"
  • Like 3
Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...