Clover General discussion

[chris1111]

Did something change between r3280 → r3292 in how Clover interacts with FakeSMC? HWMonitor normally looks like this:

 

 

 

After upgrading to r3292, many of these readings were missing (CPU Frequency, Memory Modules, CPU Package Graphics…). Also had readings for items I clearly shouldn't have on a desktop, such as Battery. Mainboard Proximity was renamed to Northbridge Proximity. And so on.

 

I neglected to take a screenshot of this before downgrading to r3280, which fixed it.

No Issue here about the Latest HWSensors and Clover r3292

 

[oSxFr33k]

FYI.  If you have Clover SE installed and then try to install a Non Special Edition, the boot stalls near the end of the boot process (Ethernet/Wifi/BT) no Desktop.  This has happened to me more than once in both Yosemite and El Capitan and I could not catch the pattern until just now.

 

Maybe wiping EFI then installing may work but upgrading does not for sure!!

[chris1111]

FYI.  If you have Clover SE installed and then try to install a Non Special Edition, the boot stalls near the end of the boot process (Ethernet/Wifi/BT) no Desktop.  This has happened to me more than once in both Yosemite and El Capitan and I could not catch the pattern until just now.

 

Maybe wiping EFI then installing may work but upgrading does not for sure!!

I think you need to read more about Clover my freind

SE its a same Package, the only difference is the config.plist and FakeSMC.kext

already on EFI /CLOVER/kexts

 

I run Clover Package Source Forge on my both Pc, Laptop with no issue

same with SE 

[Slice]

I think you need to read more about Clover my freind

SE its a same Package, the only difference is the config.plist and FakeSMC.kext

already on EFI /CLOVER/kexts

 

I run Clover Package Source Forge on my both Pc, Laptop with no issue

same with SE 

It may be other FakeSMC

[oSxFr33k]

I think you need to read more about Clover my freind

SE its a same Package, the only difference is the config.plist and FakeSMC.kext

already on EFI /CLOVER/kexts

 

I run Clover Package Source Forge on my both Pc, Laptop with no issue

same with SE 

 

 

I want to agree with you and I know they are the same but in my case every single time I install non SE over SE it does not boot period!!  I can replicate this every single time.

[thy1225]

anyone know how to set "only search for Sata0 drivers" ??

My setup boot ridiculously slow between the bios and the clover GUI (took about 20secs) and I'm trying to make it faster

I read the clover wiki saying if I set "only search for Sata0 Drivers" can make boot time 9 seconds difference

anyone know how to do that?

 

here's the clover wiki page i mentioned.  (under the "Fast" topic)

[oSxFr33k]

It may be other FakeSMC

 

Yes I am thinking it could be this.  I can replicate this every time.

[Dr. Hurt]

anyone know how to set "only search for Sata0 drivers" ??[/size]

My setup boot ridiculously slow between the bios and the clover GUI (took about 20secs) and I'm trying to make it faster[/size]

I read the clover wiki saying if I set "only search for Sata0 Drivers" can make boot time 9 seconds difference[/size]

anyone know how to do that?

 

here's the clover wiki page i mentioned.  (under the "Fast" topic)

Select BiosBlockIO instead of SATA when installing clover. 

[thy1225]

Select BiosBlockIO instead of SATA when installing clover. 

 

Thanks Dr. Hurt, I just tried, but still not feeling any significant speedup.

The GUI still took about 12~15s to show up, do I also have to make any other changes in the clover or BIOS?

[smolderas]

Thanks Dr. Hurt, I just tried, but still not feeling any significant speedup.

The GUI still took about 12~15s to show up, do I also have to make any other changes in the clover or BIOS?

I think you need to build clover with the option "--only-sata0".

 

Edit: I could optimize clover to boot to OS X just in 0,75 seconds, at least it is what bdmesg reports:

...
0:701  0:000  stringlength = 5294
0:701  0:000  CurrentMode: Width=1920 Height=1080
0:701  0:000  Beginning FSInjection
FSInjectionInstall ...
- Our FSI_SIMPLE_FILE_SYSTEM_PROTOCOL installed on handle: CE831C98
FSInjectionInstall ...
- Our FSI_SIMPLE_FILE_SYSTEM_PROTOCOL installed on handle: CE831C98
0:702  0:000  Preparing kexts injection for arch=x86_64 from EFI\CLOVER\kexts\Other
0:704  0:002  Preparing kexts injection for arch=x86_64 from EFI\CLOVER\kexts\10.11
0:704  0:000    Extra kext: EFI\CLOVER\kexts\10.11\FakeSMC.kext
0:710  0:005      Extra PlugIn kext: EFI\CLOVER\kexts\10.11\FakeSMC.kext\Contents\PlugIns\ACPISensors.kext
0:713  0:002      Extra PlugIn kext: EFI\CLOVER\kexts\10.11\FakeSMC.kext\Contents\PlugIns\CPUSensors.kext
0:715  0:002      Extra PlugIn kext: EFI\CLOVER\kexts\10.11\FakeSMC.kext\Contents\PlugIns\GPUSensors.kext
0:717  0:001      Extra PlugIn kext: EFI\CLOVER\kexts\10.11\FakeSMC.kext\Contents\PlugIns\LPCSensors.kext
0:722  0:005    Extra kext: EFI\CLOVER\kexts\10.11\RealtekRTL81xx.kext
0:729  0:006    Extra kext: EFI\CLOVER\kexts\10.11\realtekALC.kext
0:734  0:004  Removed efi-boot-device-data variable: Not Found
0:734  0:000  SetScreenResolution: 1920x1080 - already set
0:734  0:000  Custom boot is disabled
0:734  0:000  Closing log

[Slice]

 

I see that the kernel patch for haswell-E (10.9.5 10.10 10.10.1) has too many information to patch, because its no necesary two patch for these cpu. I clean the patch, I don't if anyone want to merge or use. Anyway I leave here.

edk2/Clover/rEFIt_UEFI/Platform/kernel_patcher.c edk2/Clover/rEFIt_UEFI/Platform/kernel_patcher.c 
--- edk2/Clover/rEFIt_UEFI/Platform/kernel_patcher.c
+++ edk2/Clover/rEFIt_UEFI/Platform/kernel_patcher.c
@@ -802,19 +802,7 @@
     if (Bytes[Index] == 0x74 && Bytes[Index + 1] == 0x11 && Bytes[Index + 2] == 0x83 && Bytes[Index + 3] == 0xF8 && Bytes[Index + 4] == 0x3C) {
       Bytes[Index + 4] = 0x3F;
 
-      DBG("Found Haswell-E pattern #1; patched.\n");
-
-      if (PatchApplied) {
-        break;
-      }
-
-      PatchApplied = TRUE;
-    }
-    
-    if (Bytes[Index] == 0xEB && Bytes[Index + 1] == 0x0A && Bytes[Index + 2] == 0x83 && Bytes[Index + 3] == 0xF8 && Bytes[Index + 4] == 0x3A) {
-      Bytes[Index + 4] = 0x3F;
-
-      DBG("Found Haswell-E pattern #2; patched.\n");
+      DBG("Found Haswell-E pattern; patched.\n");
 
       if (PatchApplied) {
         break;

I think I understand you. Committed to 3293.

[arsradu]

Hey guys.

 

I know this is a cosmetic thing, but it looks like it might be an issue with Clover.

 

There is a rounding issue when it comes to the detected amount of VRAM and the CPU speed. They are both less by exactly 1 (MB for the VRAM, GHz for the CPU), compared to what they should be. If you forcefully inject the amount of VRAM in device-properties for example, it will show you 2048 MB for the video card. And I suppose there is a way to force the CPU to show 3.2 GHz instead of 3.19, as well. But that's not the point. The point is for the automatic detection to work on its own. And it doesn't seem to be the case here. Please, correct me if I'm wrong.

 

Below you can see a comparison between what I got on my computer, running Clover, and what it is on a real Mac (in this case it's an iMac 27" mid 2010, running Yosemite). 

 

 

 

Also, any idea how to get rid of this "ghost" volume in Clover UI? I made a mistake and launched the installer for OS X El Capitan from another partition, instead of running it from Applications. And although the installer app is neither in Applications, nor on the initial partition (since I removed it from both places), I have no idea where is this still coming from.

 

I remember I had this issue before and there was something under the El Capitan partition. But if it's there, I couldn't find it. Any idea? It's not a big issue. But it's annoying when you don't know what's causing it.

 

[TheRacerMaster]

Hey guys.

 

I know this is a cosmetic thing, but it looks like it might be an issue with Clover.

 

There is a rounding issue when it comes to the detected amount of VRAM and the CPU speed. They are both less by exactly 1 (MB for the VRAM, GHz for the CPU), compared to what they should be. If you forcefully inject the amount of VRAM in device-properties for example, it will show you 2048 MB for the video card. And I suppose there is a way to force the CPU to show 3.2 GHz instead of 3.19, as well. But that's not the point. The point is for the automatic detection to work on its own. And it doesn't seem to be the case here. Please, correct me if I'm wrong.

 

Below you can see a comparison between what I got on my computer, running Clover, and what it is on a real Mac (in this case it's an iMac 27" mid 2010, running Yosemite). 

 

Clover.png RealMac.png

 

Also, any idea how to get rid of this "ghost" volume in Clover UI? I made a mistake and launched the installer for OS X El Capitan from another partition, instead of running it from Applications. And although the installer app is neither in Applications, nor on the initial partition (since I removed it from both places), I have no idea where is this still coming from.

 

I remember I had this issue before and there was something under the El Capitan partition. But if it's there, I couldn't find it. Any idea? It's not a big issue. But it's annoying when you don't know what's causing it.

 

CloverUI.png

 

The 2047MB thing is caused the NVIDIA drivers IIRC. You can see in bdmesg how Clover detects your hardware (post it here if you want to help solve the issue). Also, look for .IABootFiles (and some file, don't remember exactly ATM) in the root of your drive and delete them if they're present.

[Andres ZeroCross]

Hey guys.

 

I know this is a cosmetic thing, but it looks like it might be an issue with Clover.

 

There is a rounding issue when it comes to the detected amount of VRAM and the CPU speed. They are both less by exactly 1 (MB for the VRAM, GHz for the CPU), compared to what they should be. If you forcefully inject the amount of VRAM in device-properties for example, it will show you 2048 MB for the video card. And I suppose there is a way to force the CPU to show 3.2 GHz instead of 3.19, as well. But that's not the point. The point is for the automatic detection to work on its own. And it doesn't seem to be the case here. Please, correct me if I'm wrong.

 

Below you can see a comparison between what I got on my computer, running Clover, and what it is on a real Mac (in this case it's an iMac 27" mid 2010, running Yosemite). 

 

Clover.png RealMac.png

 

Also, any idea how to get rid of this "ghost" volume in Clover UI? I made a mistake and launched the installer for OS X El Capitan from another partition, instead of running it from Applications. And although the installer app is neither in Applications, nor on the initial partition (since I removed it from both places), I have no idea where is this still coming from.

 

I remember I had this issue before and there was something under the El Capitan partition. But if it's there, I couldn't find it. Any idea? It's not a big issue. But it's annoying when you don't know what's causing it.

 

CloverUI.png

Maybe this article related to your issue Arsradu https://pikeralpha.wordpress.com/2015/09/27/how-to-replace-boot-efi-with-mine/

[arsradu]

The 2047MB thing is caused the NVIDIA drivers IIRC. You can see in bdmesg how Clover detects your hardware (post it here if you want to help solve the issue). Also, look for .IABootFiles (and some file, don't remember exactly ATM) in the root of your drive and delete them if they're present.

 

Uninstalling the Nvidia drivers seems to have fixed the Graphics issue (VRAM is now 2048). Still, weird that this happens with Nvidia drivers installed but not applied.

I extracted the content of bdmesg into a txt for debugging of the other issue.

 

And there doesn't seem to be any .IABootFiles (I looked for hidden files too) anywhere on my computer. I'll keep looking.

 

Nevermind, found them. Thank you!  Since I started the installation from the Tools partition, the .IABootFiles folder was present there.

bdmesg.txt

[TheRacerMaster]

Maybe this article related to your issue Arsradu https://pikeralpha.wordpress.com/2015/09/27/how-to-replace-boot-efi-with-mine/

 

Nothing to do with this issue, or Hacks in general. This is for replacing boot.efi with Pike's (amazing work BTW, it's amazing how well a 2006 Mac Pro runs 10.11) for unsupported real Macs with 32-bit EFI (Mavericks & above need 64-bit, tiamo & Pike have made their own versions of boot.efi that thunk 32-bit EFI calls to 64-bit).

Uninstalling the Nvidia drivers seems to have fixed the Graphics issue (VRAM is now 2048). Still, weird that this happens with Nvidia drivers installed but not applied.

I extracted the content of bdmesg into a txt for debugging of the other issue.

 

This is your issue:

6:928  0:000  Finally: Bus=99768kHz CPU=3193MHz

For whatever reason, your base clock (BCLK) is 99.768 MHz. Multiply this by your main CPU multiplier (which, in this case, is 32x), and you get 3.193576 GHz, which is rounded to 3.19 GHz. Try setting the base clock in your BIOS to 100.05, as suggested by Piker-Alpha. This should give the correct 3.20 GHz frequency.

[arsradu]

Nothing to do with this issue, or Hacks in general. This is for replacing boot.efi with Pike's (amazing work BTW, it's amazing how well a 2006 Mac Pro runs 10.11) for unsupported real Macs with 32-bit EFI (Mavericks & above need 64-bit, tiamo & Pike have made their own versions of boot.efi that thunk 32-bit EFI calls to 64-bit).

 

 

This is your issue:

6:928  0:000  Finally: Bus=99768kHz CPU=3193MHz

For whatever reason, your base clock (BCLK) is 99.768 MHz. Multiply this by your main CPU multiplier (which, in this case, is 32x), and you get 3.193576 GHz, which is rounded to 3.19 GHz. Try setting the base clock in your BIOS to 100.05, as suggested by Piker-Alpha. This should give the correct 3.20 GHz frequency.

Right. It is indeed set to 99.78-79MHz in EUFI Bios. But I couldn't find a way to change that. Is there? Can I do that on a non-k cpu? And why is the frequency still 3.2 GHz in Windows? )

 

Also, I do remember not having this issue with older versions of Clover in early stages of Yosemite (don't remember which version I was running back then). Really old one, for sure.

So yes, a rounding issue. But it doesn't happen with Windows, and it didn't happen with older (really old) versions of Clover.

[Pike R. Alpha]

 

Right. It is indeed set to 99.78-79MHz in EUFI Bios. But I couldn't find a way to change that. Is there? Can I do that on a non-k cpu? And why is the frequency still 3.2 GHz in Windows? )

 

Also, I do remember not having this issue with older versions of Clover in early stages of Yosemite (don't remember which version I was running back then). Really old one, for sure.

So yes, a rounding issue. But it doesn't happen with Windows, and it didn't happen with older (really old) versions of Clover.

Hi,

 

I may be wrong, but wasn't there a config setting to correct this? Bus Frequency or something like that?

[arsradu]

  Hi,

 

I may be wrong, but wasn't there a config setting to correct this? Bus Frequency or something like that?

Hi Pike 

 

If you're talking about this setting here

then yes, there is.

 

BUT, setting that to 100.5 (KHz) results in the following misbehaviors:

1. unbootable system (system hangs on first stage boot)

2. Clover UI and first stage booting displayed at higher than normal resolution (1600x1200 when it should be 1440x900)

3. THIS beautiful blue screen reminding me of the early ages of Windows XP.

 

You can input 3200 as your CPU frequency in MHz, and that will force it to display 3.2 GHz instead of 3.19. 

 

But, as I said, I'm more interested to understand why doesn't it do that, on its own, rather than me tweaking/forcing it to display a certain value.

[AsusFreak]

  Hi,

 

I may be wrong, but wasn't there a config setting to correct this? Bus Frequency or something like that?

Yep, something like this fixes Clover mis-reporting...

 

    <key>CPU</key>

    <dict>

        <key>BusSpeedkHz</key>

        <integer>100000</integer>

    </dict>

[AsusFreak]

The biggest problem here is user dependency on "Clover Configuator". Nice tool but not a Clover dev supported tool. Suggest you research problems here:  http://clover-wiki.zetam.org/Home, http://www.insanelymac.com/forum/topic/304530-clover-change-explanations/, and http://www.insanelymac.com/forum/topic/284656-clover-general-discussion/page-434
 

[arsradu]

Yep, something like this fixes Clover mis-reporting...

 

    <key>CPU</key>

    <dict>

        <key>BusSpeedkHz</key>

        <integer>100000</integer>

    </dict>

 

Ok, please, ignore my previous post. I set the bus speed in MHz, instead of KHz. My bad.

 

Yeah, setting it to 100,000 KHz or setting the CPU frequency to 3200 MHz will show 3.2 GHz in About This Mac.

 

But...why do we have to do this manually? I don't remember messing with this stuff before. I might be wrong though.

 

Update: apparently setting that to 100,000 KHz affects performance. In YouTube, for example, the audio starts stuttering after about a few minutes of playback (with Safari and Flash player installed).

[zxv]

No Issue here about the Latest HWSensors and Clover r3292

Are you running 10.11.1 (15B38b)? I assume this is something related to the kernelPM patch introduced in r3292.

 

For whatever reason, your base clock (BCLK) is 99.768 MHz. Multiply this by your main CPU multiplier (which, in this case, is 32x), and you get 3.193576 GHz, which is rounded to 3.19 GHz. Try setting the base clock in your BIOS to 100.05, as suggested by Piker-Alpha. This should give the correct 3.20 GHz frequency.

Yes. I've been setting my BCLK in BIOS to 100.1 (smallest decimal allowed) for awhile now to correct this behavior.

 

This isn't a Clover issue, and I'd be skeptical of using the BusSpeedkHz setting to resolve it.

[arsradu]

Are you running 10.11.1 (15B38b)? I assume this is something related to the kernelPM patch introduced in r3292.

 

 

Yes. I've been setting my BCLK in BIOS to 100.1 (smallest decimal allowed) for awhile now to correct this behavior.

 

This isn't a Clover issue, and I'd be skeptical of using the BusSpeedkHz setting to resolve it.

How do I set that value? Can it be set on non-k cpus?

If it's not a Clover issue, then where is the issue? If you say CPU, then why, with the same CPU and motherboard, I get correct/appropriate readings in Windows (haven't tried Ubuntu yet) and (althogh I can't remember the exact version number) with that older version of Clover? Was it a bug then? Is it a bug now? This is what I'm trying to understand.

 

And yes, as I said, changing the bus speed in config leads to some issues im.YouTube. If you only set the frequency (3200 in my case), the issue is gone and the displayed value is also 3.2 GHz.

 

I'll try to run bdmesg on that real imac and see what kind of bus speeds do i get there...

[Slice]

Hi arsradu,

as well as other discussing the non-round cpu clock.

There are settings in Clover config.plist

	<key>CPU</key>
	<dict>
		<key>#BusSpeedkHz</key>
		<integer>133330</integer>
		<key>#FrequencyMHz</key>
		<integer>3140</integer>

BusSpeedkHz will set Bus Speed and tell to kernel FSBFrequency.

	if (DTLookupEntry(0, "/efi/platform", &entry) != kSuccess) {
		kprintf("EFI_FSB_frequency: didn't find /efi/platform\n");
		return 0;
	}
	if (DTGetProperty(entry,FSB_Frequency_prop,&value,&size) != kSuccess) {
		kprintf("EFI_FSB_frequency: property %s not found\n",
			FSB_Frequency_prop);
		return 0;
	}

Clover calculated value for me 99790. DO NOT ROUND IT!!! It is correct value. It is more correct value then 100000!

It is checked and proved. Clover calculate the value with good precision, much more precise then you think.

Changing the value will affect Clock, Sound and other system behavior.

 

FrequencyMHz is a value to be written into DMI and affect only on AboutMac with System Profiler to show CPU Speed. It is safe to write here all you want.

 

 

PS. CloverConfigurator tends to write something here. This is bad!

Open the final config.plist with TextEditor and erase the lines with BusSpeedkHz to be not having this value.

Or place # at the key as in my example.