Clover General discussion

[fatala]

your log is wrong both old and new. because your config.plist has dummy option and not clear.

if have problem like you, all clover users has problem. as result, this case include only you.

 

try this with clover r4128. and please upload full log.

once again, this is not clover bug.

 

need to move this discussion to general discussion thread.

Sherlocks, YOU ARE AWESOME!!! It worked! Thank you!

[syscl]

May i try  r4138 if you put it.

Ok, please try this 

r4138.zip

and 

r4139 as well

r4139.zip

 

Thank you,

syscl

[syscl]

Hello, 

 

I've added a new function EnableExtCpuXCPM() for unsupported Cpu to enable XCPM support on r4140. Till now only Broadwell-E/EP has been added, more support will be added later on. 

 

The major problem for me now is that I don't have Broadwell-E/EP platform to test if all the related patches works or not. It's an initial version to enable XCPM on this CPUs.

 

@apanti, since now copious amount of CPUs related patches will be added(Haswell-E, IVY-E, ...), @Sherlocks suggestion may make sense. Can we a boolean key that can disable the execution of EnableExtCpuXCPM? Like a boolean key DisableExtCpuXCPM for people want to disable this feature. We can replace Haswell-E with DisableExtCpuXCPM.

 

Here's a prebuilt version of r4140

r4140.zip

 

Wait for your reports

 

syscl

[Sherlocks]

Hello,

 

I've added a new function EnableExtCpuXCPM() for unsupported Cpu to enable XCPM support on r4140. Till now only Broadwell-E/EP has been added, more support will be added later on.

 

The major problem for me now is that I don't have Broadwell-E/EP platform to test if all the related patches works or not. It's an initial version to enable XCPM on this CPUs.

 

@apanti, since now copious amount of CPUs related patches will be added(Haswell-E, IVY-E, ...), @Sherlocks suggestion may make sense. Can we a boolean key that can disable the execution of EnableExtCpuXCPM? Like a boolean key DisableExtCpuXCPM for people want to disable this feature. We can replace Haswell-E with DisableExtCpuXCPM.

 

Here's a prebuilt version of r4140

r4140.zip

 

Wait for your reports

 

syscl

Now i say last.

Xcpm patches not completed(not perfect).

Low freq and high performance issue that i mentioned It means each kernel patches need to check confirm on each user system.

 

Why should force use this xcpm patch? There is no choice now. It causes to comfuse situtation. This is not best way. Also not clear for stable system. Also cant minimize patches. Never

 

I dont understand it.

 

나의 LG-F800S 의 Tapatalk에서 보냄

[syscl]

Now i say last.

Xcpm patches not completed(not perfect). It means each kernel patches need to check confirm on each user system.

 

Why should force use this xcpm patch? There is no choice now. It causes to comfuse situtation. This is not best way. Also not clear for stable system. Also cant minimize patches. Never

 

I dont understand it.

 

나의 LG-F800S 의 Tapatalk에서 보냄

@Sherlocks there will be an option called DisableExtCpuXCPM to disable all this patches with one boolean key. Will add it.

 

By the way, if new system doesn't contain specific patches, then SearchAndReplace will not replace non-exist pattern which means it will be safe on newer system.. 

 

I will add the boolean key for users who don't want the force patch...But I need finish this patch first..

 

Thank you,

syscl

[Sherlocks]

@Sherlocks there will be an option called DisableExtCpuXCPM to disable all this patches with one boolean key. Will add it.

 

By the way, if new system doesn't contain specific patches, then SearchAndReplace will not replace non-exist pattern which means it will be safe on newer system..

 

I will add the boolean key for users who don't want the force patch...But I need finish this patch first..

 

Thank you,

syscl

Okay.

 

Xcpm(not perfect) force patch - default

-disable xcpm option.

 

Clean kernel status without force patch - default

-just support boolean option like KernelHaswellE.

 

What is best way? I want to know developer and coder think.

 

 

나의 LG-F800S 의 Tapatalk에서 보냄

[syscl]

Okay.

 

Xcpm(not perfect) force patch - default

-disable xcpm option.

 

Clean kernel status without force patch - default

-just support boolean option like KernelHaswellE.

 

What is best way? I want to know developer and coder think.

 

 

나의 LG-F800S 의 Tapatalk에서 보냄

boolean key: EnableExtCpuXCPM vs DisableExtCpuXCPM... Or, make it more general, EnableExtCpuPM vs DisableExtCpuPM(for SandyBridge-E support)...

 

syscl

[PMheart]

boolean key: EnableExtCpuXCPM vs DisableExtCpuXCPM... Or, make it more general, EnableExtCpuPM vs DisableExtCpuPM(for SandyBridge-E support)...

 

syscl

SNB-E never has XCPM support, as far as I know. Unless you want to patch AICPUPM as well like what @stinga11 did. As for name, I prefer "Enablexxxx", and in my humble opinion, no bin-patch should be applied automatically.

By the way, why do you need FakeCPUID after correcting _cpuid_set_info(), I guess only one of them is really needed. (And I prefer the latter)

 

 

Okay.

 

Xcpm(not perfect) force patch - default

-disable xcpm option.

 

Clean kernel status without force patch - default

-just support boolean option like KernelHaswellE.

 

What is best way? I want to know developer and coder think.

 

 

나의 LG-F800S 의 Tapatalk에서 보냄

 

Yes, that's also what I want.

 

 

Now i say last.

Xcpm patches not completed(not perfect).

Low freq and high performance issue that i mentioned It means each kernel patches need to check confirm on each user system.

 

Why should force use this xcpm patch? There is no choice now. It causes to comfuse situtation. This is not best way. Also not clear for stable system. Also cant minimize patches. Never

 

I dont understand it.

 

나의 LG-F800S 의 Tapatalk에서 보냄

 

Yes indeed. As for "performance fix" patch, some users may want lowered frequencies whereas others want highest performance, which is hard to determine...

I agree. No changes should be mandatory, when they are just not obligatory...

[danif]

Hello, 

 

I've added a new function EnableExtCpuXCPM() for unsupported Cpu to enable XCPM support on r4140. Till now only Broadwell-E/EP has been added, more support will be added later on. 

 

The major problem for me now is that I don't have Broadwell-E/EP platform to test if all the related patches works or not. It's an initial version to enable XCPM on this CPUs.

 

@apanti, since now copious amount of CPUs related patches will be added(Haswell-E, IVY-E, ...), @Sherlocks suggestion may make sense. Can we a boolean key that can disable the execution of EnableExtCpuXCPM? Like a boolean key DisableExtCpuXCPM for people want to disable this feature. We can replace Haswell-E with DisableExtCpuXCPM.

 

Here's a prebuilt version of r4140

r4140.zip

 

Wait for your reports

 

syscl

 

Hi, I have a Broadwell-E cpu and I can help with testing if you guide me. I have alredy compiled r4141 and I'm attaching config.plist, AppleIntelInfo, bootlog and maclog output (by the way is there any option for maclog to only log the last boot?)

 

 

XPCM is not enabled. should I add a new boolean key in config.plist?

 

Running 10.12.6

 

Thanks.

Broadwell-E test1 .zip

[syscl]

Hi, I have a Broadwell-E cpu and I can help with testing if you guide me. I have alredy compiled r4141 and I'm attaching config.plist, AppleIntelInfo, bootlog and maclog output (by the way is there any option for maclog to only log the last boot?)

 

 

XPCM is not enabled. should I add a new boolean key in config.plist?

 

Running 10.12.6

 

Thanks.

Hi, really appreciate your report. Could you give me some clear guide which patches require for Broadwell-E to enable XCPM on 10.12.6? I don't have computer to test, but I can help if you provide some information.

 

Thank you,

syscl

[danif]

Hi, really appreciate your report. Could you give me some clear guide which patches require for Broadwell-E to enable XCPM on 10.12.6? I don't have computer to test, but I can help if you provide some information.

 

Thank you,

syscl

 

I had this 4 patches in config.plist when booting clover 4070:

 

10.12.x, _cpuid_set_info() Fix for Broadwell-E © Pike R. Alpha

83c0e9 -> 83c0e1

 

10.12.6, Enable XCPM on Broadwell-E © Pike R. Alpha

8d43c483 f822 -> 8d43bc83 f822

 

10.12.x, Reboot fix #1 © Pike R. Alpha

554889e5 41574156 41554154 53504189d64189f7 4889fb45 85ff0f84 -> c39089e5 41574156 41554154 53504189d64189f7 4889fb45 85ff0f84

 

10.12.x, Reboot fix #2 © Pike R. Alpha

005de908 0000000f 1f840000 000000554889e541 -> 005de908 0000000f 1f840000 000000c39089e541

[syscl]

I had this 4 patches in config.plist when booting clover 4070:

 

10.12.x, _cpuid_set_info() Fix for Broadwell-E © Pike R. Alpha

83c0e9 -> 83c0e1

 

10.12.6, Enable XCPM on Broadwell-E © Pike R. Alpha

8d43c483 f822 -> 8d43bc83 f822

 

10.12.x, Reboot fix #1 © Pike R. Alpha

554889e5 41574156 41554154 53504189d64189f7 4889fb45 85ff0f84 -> c39089e5 41574156 41554154 53504189d64189f7 4889fb45 85ff0f84

 

10.12.x, Reboot fix #2 © Pike R. Alpha

005de908 0000000f 1f840000 000000554889e541 -> 005de908 0000000f 1f840000 000000c39089e541

Thank you for your information. 

 

I will try to add this patches manually. Then please try newer version once Broadwell-E parts gets update

 

syscl

[Loloflat6]

r4039 works in Maverick

[syscl]

@Loloflat6 @danif

 

Please try this version r4143

r4143.zip

 

I added reboot fix #2 for Broadwell-E/EP.

 

For Celeron/Pentium is stable with new Clover

 

Thank you in advance,

syscl

[PMheart]

I had this 4 patches in config.plist when booting clover 4070:

 

10.12.x, _cpuid_set_info() Fix for Broadwell-E © Pike R. Alpha

83c0e9 -> 83c0e1

 

10.12.6, Enable XCPM on Broadwell-E © Pike R. Alpha

8d43c483 f822 -> 8d43bc83 f822

 

10.12.x, Reboot fix #1 © Pike R. Alpha

554889e5 41574156 41554154 53504189d64189f7 4889fb45 85ff0f84 -> c39089e5 41574156 41554154 53504189d64189f7 4889fb45 85ff0f84

 

10.12.x, Reboot fix #2 © Pike R. Alpha

005de908 0000000f 1f840000 000000554889e541 -> 005de908 0000000f 1f840000 000000c39089e541

Hi, actually only one of the two "Reboot fix" is needed, they are just the same patch. The whole one looks like this: (from 10.12.6 16G29)

00005DE9 08000000 0F1F8400 00000000 554889E5 41574156 41554154 53504189 D64189F7 4889FB45 85FF0F84

 

And you change the 5548 in red to c390.

So @syscl please remove either of them, thanks in advance.

[gujiangjiang]

Hlad to see @syscl works hard on xcpm and hwp and i prefered use xcpm or hwp by default.

 

If someone dont like xcpm or hwp we can add one option like DisableExtCPUXCPM or DisableHWP to use.

 

Because the real mac use hwp by default on laptop since skylake and now the hwp is more perfect than the old way so i prefer to use this by default.

 

Thanks syscl.

[danif]

Hi, actually only one of the two "Reboot fix" is needed, they are just the same patch. The whole one looks like this: (from 10.12.6 16G29)

00005DE9 08000000 0F1F8400 00000000 554889E5 41574156 41554154 53504189 D64189F7 4889FB45 85FF0F84

 

And you change the 5548 in red to c390.

So @syscl please remove either of them, thanks in advance.

 

Thanks, you are right, I can boot with only one of them.

 

@Loloflat6 @danif

 

Please try this version r4143

r4143.zip

 

I added reboot fix #2 for Broadwell-E/EP.

 

For Celeron/Pentium is stable with new Clover

 

Thank you in advance,

syscl

 

Thanks, as @PMheart noted, this wasn't the problem. I'm attaching new logs with new and old clover info. Let me know if you need any other logs

Broadwell-E test2.zip

[Loloflat6]

@Loloflat6 @danif

 

Please try this version r4143

r4143.zip

 

I added reboot fix #2 for Broadwell-E/EP.

 

For Celeron/Pentium is stable with new Clover

 

Thank you in advance,

syscl

 For me it's ok :

- High Sierra 10.13. beta (17A315i)

- Sierra 10.12.6

- Maverick 10.9.5

 

In verbose mode i can see at the beginning your CPU process patch running

 

Here my preboot log

Loloflat6 preboot.zip

[syscl]

Thanks, you are right, I can boot with only one of them.

 

 

Thanks, as @PMheart noted, this wasn't the problem. I'm attaching new logs with new and old clover info. Let me know if you need any other logs

So now, you can boot with latest Clover right?

 

Does xcpm enable automatically? Thank you.

 

syscl

 For me it's ok :

- High Sierra 10.13. beta (17A315i)

- Sierra 10.12.6

- Maverick 10.9.5

 

In verbose mode i can see at the beginning your CPU process patch running

 

Here my preboot log

Glad to see that.

 

Tested OK on 10.11 on my system.

 

 

Thank you guys, 

syscl

[danif]

So now, you can boot with latest Clover right?

 

Does xcpm enable automatically? Thank you.

 

syscl

 

 

No sorry, it doesn't enable xcpm, that's the problem.

 

I can and could boot ok with r4141 and r4143 without any kernel patch in config.plist but it doesn't enable xcpm it tries to enable AppleIntelCPUPowerManagement mode.

 

You can see it in my latest logs file name: "systemLatestClover.log" with my previous clover revision (i think it was 4076) it enables xcpm, obviously with config.plist kernel patches. You can check it in "systemOldClover.log".

 

Sorry for my english.

[syscl]

No sorry, it doesn't enable xcpm, that's the problem.

 

I can and could boot ok with r4141 and r4143 without any kernel patch in config.plist but it doesn't enable xcpm it tries to enable AppleIntelCPUPowerManagement mode.

 

You can see it in my latest logs file name: "systemLatestClover.log" with my previous clover revision (i think it was 4076) it enables xcpm, obviously with config.plist kernel patches. You can check it in "systemOldClover.log".

 

Sorry for my english.

Seems I know the reason, you don't have FakeCPUID right?

 

I see stinga11 use FakeCPUID in his config.plist, while you remove it? 

 

Please try this one:

r4143-w-patch.zip

 

I have some questions: can Broadwell-E remove FakeCPUID on: 10.10? 10.11? 10.12+? Thank you in advance,

 

syscl

 

[Loloflat6]

May i try 4143 on my i3 4150 but later because of changing my smbios on my config.plist , i will repport this night .

[danif]

Problem solved in r4144, now xcpm is enabled automatically without kernel patches in config.plist

 

Thanks syscl

[apianti]

I don't care, do whatever, this conversation is going nowhere. Instead of suggesting to make the patches better so they work since THEY ARE NEEDED TO BOOT. You are all arguing about whether they should be able to be turned off, for what purpose? If the patch is not performed you cannot boot, this only applies to specific CPUs, these CPUs should be detected, all the information to determine need is there. In my opinion adding a boolean key makes it actually worse user experience because people are going to turn it on when they don't need it and wonder why they can't boot, or not turn it on when they do need it. And the most qualified person to solve an issue is a developer, I've never seen any random person just come on here and be like if you search for this binary data and replace it with this then this will work. You need the set of skills of a developer to be able to debug something.

 

EDIT: And now you've gone and irritated me. 

[Loloflat6]

Well done r4143 in my i3 4150 build  but easy because this CPU is supported.