OpenCore General Discussion

[MacNB]

 

36 minutes ago, Andrey1970 said:

 

The UEFI Windows bootloader is located on ESP FAT32.

OC_SCAN_ALLOW_FS_NTFS - it is useless for UEFI Windows.

 

Yes I know it does not work that's why I said "(I did not expect it but I thought I try it)" thinking there's a problem with the new boot policy.

What's your comment on the 4 issues ?

[Andrey1970]

8 minutes ago, insanefrancis said:

I also don't have these in my config.

For Automatic = Yes, need only Generic.

[Andrey1970]

13 minutes ago, MacNB said:

 

 

Yes I know it does not work that's why I said "(I did not expect it but I thought I try it)" thinking there's a problem with the new boot policy.

What's your comment on the 4 issues ?

You shall enable 0x00000200 (bit 9) — OC_SCAN_ALLOW_FS_HFS, allows scanning of HFS file system.

It is necessary for macOS install and Recovery.

[MacNB]

23 minutes ago, insanefrancis said:

 

Do I just run this in terminal?

This is the output.

4D1EDE05-38C7-4A6A-9CC6-4BCCA8B38C14:FirmwareFeatures	w%f5%8f%fd

 

 

 

Your value is FD8FF577 which means FW_FEATURE_SUPPORTS_CSM_LEGACY_MODE bit set.

That means you should not need AdviseWindows=YES.

Hmmm...not sure why you need it.

 

But at least you can now boot Windows from Startup disk 

[MacNB]

24 minutes ago, Andrey1970 said:

You shall enable 0x00000200 (bit 9) — OC_SCAN_ALLOW_FS_HFS, allows scanning of HFS file system.

It is necessary for macOS install and Recovery.

 

Thanks for the tip but that's not the current problem with Windows & EFI partitions.

 

Which means there's a bug with log or code: "Invalid file system policy (200/10F0503) for BB94DA98"

Also, if Bit 9 is necessary for macOS install & Recovery, then I suggest that it should set for Failsafe (by default) and Configuration.pdf changed accordingly.

Currently failsafe = 0x10F0103

I recommend failsafe = 0x10F0303

[arsradu]

Hi guys!

 

Hope you're all doing fantastic!

 

So, I was trying to get a better understanding of how we're supposed to add values like Audio Codec and Audio Out to config. I took a look at the Configuration PDF and I think I finally understood how to set it up.  

 

Still, interestingly enough, in my case at least, I have a boot chime with both Audio Out 0 and 2. I haven't tried the rest of them (from 0 to 4, since I've got 5 output ports, so N-1=5-1=4), but 0 and 2 seemed to work fine. I'm guessing that's probably normal...?

 

Anyway, what I wanted to suggest is if you could maybe update the font of the pdf, because bold text on that font is almost the same as regular which is not good when the bold text is actually what you're looking for. Here's an example of what it looks like in Preview (the default app for PDFs) and Typora.

 

If you can tell me which part is bolded in the first screenshot, without looking at the second one, I'll give you a beer. And not Corona. :))

 

 

And here's from Typora. Same text, just copy-pasted in another app. Literally night and day difference.

 

Edited May 21, 2020 by arsradu

[mhaeuser]

3 hours ago, MacNB said:

With Misc->Entries, if we cannot add an entry that works then what is the point of allowing a Custom entry that is not bootable ? Can you please give an example of "multiple OSes per partition" ?

You nowhere said it was unbootable, only that it was not accepted as default entry from macOS Startup Disk. macOS Startup Disk, as it uses said Apple boot policy, does not have the concept of multiple OSes per partition. It stores only the partition (not file) information into NVRAM and then lets the firmware (Apple EFI, here accounted for by OpenCore) decide which file is bootable on the partition - this happens by an algorithm detailed in the PDF. Without BlessOverride, bootmgfw can never be the determined boot file from this partition node, this is the whole reason we have BlessOverride in the first place, and that is exactly what you are seeing. Previously, we adhered to one OS per partition (enforcing Apple rules and our extensions) and when scanning did not yield any entry for this partition (e.g. BlessOverride does not contain the necessary file), custom entries were chosen by partition matching. Now, we support e.g. arbitrary Boot#### entries and such, hence partition matching is no longer utilised - this is what breaks this behaviour for you. However, that was never "supported/recommended" in the first place, that was just what happened to be the implementation.

 

Multiple OSes on a partition refers to the amount of non-alternative (e.g. macOS Recovery can be alternative on FileVault setups) booters on one partition. OpenCore and Windows sharing an ESP does count as "multiple OSes".

[insanefrancis]

1 hour ago, MacNB said:

 

Your value is FD8FF577 which means FW_FEATURE_SUPPORTS_CSM_LEGACY_MODE bit set.

That means you should not need AdviseWindows=YES.

Hmmm...not sure why you need it.

 

But at least you can now boot Windows from Startup disk 

 

I might be wrong but I'm thinking OC sets these values accordingly depending on your partition schemes, whether EFI is first partition or not, when you have AdviseWindows=YES. I don't remember setting any of these values on my config manually. I might try clearing NVRAM to remove these values, set AdviseWindows=NO then check again if OC will set something. I'll probably not, I have it already working. But I may if no one validate or invalidate my theory.

Edited May 21, 2020 by insanefrancis

[arsradu]

 

Thanks, @vit9696  

[MacNB]

3 hours ago, Download-Fritz said:

You nowhere said it was unbootable, only that it was not accepted as default entry from macOS Startup Disk. macOS Startup Disk, as it uses said Apple boot policy, does not have the concept of multiple OSes per partition. It stores only the partition (not file) information into NVRAM and then lets the firmware (Apple EFI, here accounted for by OpenCore) decide which file is bootable on the partition - this happens by an algorithm detailed in the PDF. Without BlessOverride, bootmgfw can never be the determined boot file from this partition node, this is the whole reason we have BlessOverride in the first place, and that is exactly what you are seeing. Previously, we adhered to one OS per partition (enforcing Apple rules and our extensions) and when scanning did not yield any entry for this partition (e.g. BlessOverride does not contain the necessary file), custom entries were chosen by partition matching. Now, we support e.g. arbitrary Boot#### entries and such, hence partition matching is no longer utilised - this is what breaks this behaviour for you. However, that was never "supported/recommended" in the first place, that was just what happened to be the implementation.

 

Multiple OSes on a partition refers to the amount of non-alternative (e.g. macOS Recovery can be alternative on FileVault setups) booters on one partition. OpenCore and Windows sharing an ESP does count as "multiple OSes".

 

Thanks @Download-Fritz for the explanation on how the new boot method is meant to work.

 

To get around the fact that Windows is not accepted as default, I used ScanPolicy of 0x010F0503.

That allows the Windows drive to be selected in Startup Disks and macOS restarts.

On restart the boot picker now displays TWO windows drives: (a) the UN-Blessed one (labeled "Windows") that it finds in the ESP and (b) the Custom Entry I had put in (named "Windows 10 Pro"). BUT the default one selected is the (a) and NOT (b) and when hitting Enter, of course (a) cannot boot. As soon as I provide BlessedOverride, then OC boots Windows perfectly.

Recommendation: now that OC support e.g. arbitrary Boot#### entries and such, hence partition matching is no longer utilised - this is what breaks the Custom Entries behaviour, should be documented in the ScanPolicy section. That is, avoid Custom Entries AND BlessOverride for the same OS.

 

Earlier, I list the problems :

1. With Misc->Entries, if we cannot add an entry that works then what is the point of allowing a Custom entry that is not bootable ? Can you please give an example of "multiple OSes per partition" ?
2. There's another problem with the "by design" boot policy now. I could only get the Windows OS displayed in the boot picker if I enable the bit OC_SCAN_ALLOW_FS_ESP in the ScanPolicy setting. If I unset that bit and instead set OC_SCAN_ALLOW_FS_NTFS, then Windows drive is NOT displayed (I did not expect it but I thought I try it).
3. The side-effect of setting bit OC_SCAN_ALLOW_FS_ESP now also displays EFI harddisk icon (I don't know for which of my 3 drives but I assume OC drive). I am setting scanPolicy value to 0x010F0503.The HideSelf setting has now been removed from OC 0.5.9 is that the problem ?.
4. The OC boot log now shows "Invalid file system policy (200/10F0503) for BB94DA98". Here's a snapshot"

   
   07:077 00:005 OCB: Adding fs BBBACE98 (E:0|L:1|P:Success) - PciRoot(0x0)/Pci(0x1,0x1)/Pci(0x0,0x0)/NVMe(0x1,36-39-28-51-29-A7-79-64)/HD(1,GPT,48FD2EC1-C3AE-4458-B959-BDBEED7D2ABE,0x28,0x64000)
   07:079 00:002 OCB: Adding fs BB983B18 (E:0|L:0|P:Success) - PciRoot(0x0)/Pci(0x1F,0x2)/Sata(0x0,0xFFFF,0x0)/HD(1,GPT,B23BFB37-1721-48BA-823F-76152B5DC630,0x28,0x64000)
   07:082 00:002 OCB: Adding fs BB95DB18 (E:0|L:0|P:Success) - PciRoot(0x0)/Pci(0x1F,0x2)/Sata(0x1,0xFFFF,0x0)/HD(1,GPT,A4CE4F1A-43C6-492D-BB16-2DD6EABCE4EF,0x28,0x64000)
   07:083 00:001 OCPI: Located cached partition entries
   07:085 00:001 OCB: Invalid file system policy (200/10F0503) for BB94DA98
   07:087 00:001 OCB: Adding fs BB94DA98 (E:0|L:0|P:Security Violation) - PciRoot(0x0)/Pci(0x1F,0x2)/Sata(0x1,0xFFFF,0x0)/HD(7,GPT,8B191946-A4A6-4CDF-8D18-F0E28B417C68,0x574107A8,0x135F20)
   07:088 00:001 OCPI: Located cached partition entries

    

 

1. above you explained. 2/3. OC now lists EFI drive as bootable now that there's no HideSelf setting. How do we stop EFI drive being listed ?

4. With ScanPolicy of 0x010F0503, the OC log file shows (a warning ?) : "Invalid file system policy (200/10F0503)" @Andrey1970 explained that's because I did not enable the bit 9 (OC_SCAN_ALLOW_FS_HFS) which is needed for macOS Recovery. Recommendation: this should be documented and FailSafe value should be changed to 0x010F0303 (to avoid the error/warning in the OC log).

I now set the ScanPolicy to 0x010F0703 and this "warning" disappears from the OC log.

 

Thanks for all your work and continuous improvements all you Devs are making to this great project. 

[Synapsis]

@vit9696please use a utf8 capable typeface font, if I may suggest a freely available superior fonts:

Input Its superior to other rip-off of it (fontbureau actually knows their s**t)

 

https://input.fontbureau.com/download

https://www.jetbrains.com/de-de/lp/mono

https://fonts.google.com/specimen/IBM+Plex+Mono

 

Jetbrains mono (allthough I dont know if it can be used that way, allthough its a heavily "inspired" by FontBurea Input.

 

For sans-serif typeface (aka copytext), you can use Input Sans which is also quite nice, but I suggest IBM Plex Serif, for all

copy/headlines textes and Input for code listing

 

Please, don't use any half-as*ed fonts like MS Consolas or MS Cascadia, it seems they werer "designed" by people not yet

fully aware the 21st century arrived some time ago and they missed stuff like HIDPI devices, preper hinting and ligatures.

Edited May 21, 2020 by Synapsis
double

[arsradu]

42 minutes ago, Synapsis said:

@vit9696please use a utf8 capable typeface font, if I may suggest a freely available superior fonts:

Input Its superior to other rip-off of it (fontbureau actually knows their s**t)

 

https://input.fontbureau.com/download

https://www.jetbrains.com/de-de/lp/mono

https://fonts.google.com/specimen/IBM+Plex+Mono

 

Jetbrains mono (allthough I dont know if it can be used that way, allthough its a heavily "inspired" by FontBurea Input.

 

For sans-serif typeface (aka copytext), you can use Input Sans which is also quite nice, but I suggest IBM Plex Serif, for all

copy/headlines textes and Input for code listing

 

Please, don't use any half-as*ed fonts like MS Consolas or MS Cascadia, it seems they werer "designed" by people not yet

fully aware the 21st century arrived some time ago and they missed stuff like HIDPI devices, preper hinting and ligatures.

 

Speaking of fonts, I also liked this one from Google Font Directory. Probably less interesting than the IBM Plex (really nice one, by the way). Still very nice and simple. And it seems to have glyphs for other languages, too (including Russian, some special characters for French, I think I saw some Polish ones as well, a few Romanian ones, too, but not all, and probably others, too). So yeah, not bad.

 

You can probably tell I'm not any kind of expert in fonts and stuff like that. :)) I just analyse them mostly on a "visual" basis, and that's about it. I mean, I know the point is to be easy to read. And, you have to have (at least for the language they're meant to be used for) a complete set of glyphs, otherwise you're probably gonna have fonts from different families which is gonna look like garbage, but...specific stuff about them...I leave that to the people with more experience. :))

 

There are a ton of beautiful fonts out there. And the ones from Google if I'm not mistaken are also free to do anything with them, except selling them as your own, of course. Cause that wouldn't be fair to the author(s). But otherwise...yeah, free to use.

Edited May 21, 2020 by arsradu

[pkdesign]

Source Code Pro is another great font. Lots of variation.

[SavageAUS]

Can I get some help with a verbose message I get when loading OC. 

ocs: No schema for RequestBootVarFallback at 4 index

RequestBootVarFallback is set to false in config. 

 

And can anyone point me in the direction for setting my ethernet as built-in. Solved.

Config attached for checking please

config.plist

Edited May 22, 2020 by SavageAUS

[Matgen84]

3 hours ago, SavageAUS said:

Can I get some help with a verbose message I get when loading OC. 

ocs: No schema for RequestBootVarFallback at 4 index

RequestBootVarFallback is set to false in config. 

 

And can anyone point me in the direction for setting my ethernet as built-in. Solved.

Config attached for checking please

config.plist

 

If you use OpenCore 0.5.9 up to date: see in changelog.

• Removed RequestBootVarFallback due to numerous bugs

So delete it in your config.plist

[SavageAUS]

1 hour ago, Matgen84 said:

 
If you use OpenCore 0.5.9 up to date: see in changelog.

• Removed RequestBootVarFallback due to numerous bugs

So delete it in your config.plist

Annoying text gone now, thanks.

With clover I used voodoops2controller and my brightness keys worked. With OC and same kext brightness keys not working.


Sent from my iPhone using Tapatalk

Edited May 22, 2020 by SavageAUS

[vit9696]

@Synapsis,@arsradu, 7-bit ASCII font usage is intentional for security reasons.

[SavageAUS]

Update:

Adding windows into bless override solved my windows booting issue.

I recorded screen in slomo to capture it.

The error i was getting was

System BootOrder not found, Initializing defaults.

Reset System

Now windows boots fine, although first boot into i had to fix microsoft account problem, hopefully next boots into windows it doesn't occur. (Which it didn’t)

Edited May 22, 2020 by SavageAUS

[SavageAUS]

On 11/12/2019 at 4:19 AM, jl4c said:

I am not sure of this, as an example, you can take the method with which Rehabman fix instant wake on usb ports, where the SSDT returns to the original code (XPRW) in DSDT.

DefinitionBlock ("", "SSDT", 2, "HACK", "GPRW", 0x00000000)
{
    External (XPRW, MethodObj)    // 2 Arguments
    Method (GPRW, 2, NotSerialized)
    {
        If ((0x0D == Arg0))
        {
            Return (Package (0x02)
            {
                0x0D, 
                Zero
            })
        }
        Return (XPRW (Arg0, Arg1))
    }
}

 

Did you manage to get your SSDT to work in OC for mapping and unmapping brightness hotkeys? If so can you share it with me please. 

[arsradu]

Hi guys!

 

Quick question: is it normal to have Block for Kernel section, when pretty much everything else has been renamed to Delete?

 

I've got no problem with that. I just like consistency. :))

 

So, if that's intended, no problem. If not, maybe that needs to be renamed, too...?

[mhaeuser]

41 minutes ago, arsradu said:

I've got no problem with that. I just like consistency. :))

Well, that was the whole point of the rename, "delete" and "block" have different semantics (one-time vs persistent).

[arsradu]

10 minutes ago, Download-Fritz said:

Well, that was the whole point of the rename, "delete" and "block" have different semantics (one-time vs persistent).

 

Oh, I see...  So it is intended. Cool! Thank you!

[MorenoAv]

Hi guys,

Quick question, in geek bench why is the motherboard and bios recognized as Acidanthera OPenCore 0.5.9?

Thanks

 

[vandroiy2012]

14 minutes ago, MorenoAv said:

Hi guys,

Quick question, in geek bench why is the motherboard and bios recognized as Acidanthera OPenCore 0.5.9?

Thanks

 

 

Because we care about Apple. They don't need hackintosh reports. You can disable this feature if you want with SpoofVendor=false, but it is not recommended. 

We don't bother them, they don't bother us. 

 

P.S. More info in Configuration.pdf

Edited May 22, 2020 by vandroiy2012

[MorenoAv]

Hi @vandroiy2012, 

Thanks for your answer, it was my curiosity only, and let it stay as it is, Apple have to much work already, earning too much money...