Resolved Boot menus for graphics output protocol.(AMD only)see page 13

[cdf]

It's better when people collaborate...

Absolutely. I have a 4k display and will gladly help test things out when my Pulse RX 580 8G arrives next week.

 

[thomasthegps]

The Next Loader has code that says it supports GOP @HiDPI, but I haven't been able to get it to work on my 3,1. I filed an issue, but no one has replied to it yet.

https://github.com/abdyfranco/next-loader/issues/2

However the EmuGOPDxe from next loader is how I got GOP working with CloverUEFI. Clover had support for GOP at the time Jief took the code from Clover to make rEFInd work with GOP, but later code has some of that commented out. However there is still enough GOP code in Clover to make it work with EmuGOPDxe.

I don't have a HiDPI display, just a bunch of older stuff I get by with, but I'm pretty sure @h9826790 still has one, so when he gets the time to do some testing, we may have a better idea how that would work.

Really I've had to do all the heavy lifting in this thread, and put out money to buy hardware to get things working at all. It's better when people collaborate, I just can't debug hardware I don't have, and no one outside of a multi billion dollar corporation can test all the configurations end users might need.

At strangedogs and themacelite we had teams of people helping, and we did what I try to do here, document the process, so others can do what we have done, and hopefully build on it.

As Dave is fond of saying "Make sure brain services is set to on".

I'll soon be back home sometime this week, I have an RX 580 and 4K screen, I'll be able to help with testing.

 

[DearthnVader]

Absolutely. I have a 4k display and will gladly help test things out when my Pulse RX 580 8G arrives next week.

Yes, beta tester welcome and appreciated.

I've asked a few times for someone with an nVidia card and a GOP rom to go to a shell and see if the GOP rom is mapped into ram and or linked, but no one has helped. Even tho I know lots of people use PC Cards in their Mac Pro's and would like to have boot screens for them via the GOP rom that is already on the card.

Show some love people, help me, help everyone.

The more I ask, and the more people don't help, the more hardware I have to buy, and the more inclined I am to just keep my mouth shut about how I get these things to work unless someone is willing to pay for it.

I don't find hacking and making money to be mutually exclusive, and l like to recover the hardware costs it takes me to come up with a useful hack, maybe a bit more so I can come up with the next hack.

However I try and document my hacks, so people that don't want things handed to them in an easy package still have a way of doing for themselves.

I've found over the years of hacking, some people just have more money than time, and even if you explain to them how to do something, they'd rather just pay for it, and the people that read all the documentation and build things from source and debug, normally find things I miss and end up adding to a more perfect solution.

Of course Alfa and Beta testers are part of that process, and are always welcome.

 

[abdyfranco]

The Next Loader has code that says it supports GOP @HiDPI, but I haven't been able to get it to work on my 3,1. I filed an issue, but no one has replied to it yet.

https://github.com/abdyfranco/next-loader/issues/2

However the EmuGOPDxe from next loader is how I got GOP working with CloverUEFI. Clover had support for GOP at the time Jief took the code from Clover to make rEFInd work with GOP, but later code has some of that commented out. However there is still enough GOP code in Clover to make it work with EmuGOPDxe.

I don't have a HiDPI display, just a bunch of older stuff I get by with, but I'm pretty sure @h9826790 still has one, so when he gets the time to do some testing, we may have a better idea how that would work.

Really I've had to do all the heavy lifting in this thread, and put out money to buy hardware to get things working at all. It's better when people collaborate, I just can't debug hardware I don't have, and no one outside of a multi billion dollar corporation can test all the configurations end users might need.

At strangedogs and themacelite we had teams of people helping, and we did what I try to do here, document the process, so others can do what we have done, and hopefully build on it.

As Dave is fond of saying "Make sure brain services is set to on".

I haven't been able to answer or solve the open issues in the Next Loader repository yet, because I haven't had any free time because of my work. I will try to get some free time this weekend to review the bugs and update the base code to the latest version of rEFInd.

In part, I paused the project for lack of beta testers and people who would like to collaborate, but seeing this thread it seems that several people are willing to collaborate.
[doublepost=1534877130][/doublepost]

Yes, beta tester welcome and appreciated.

I've asked a few times for someone with an nVidia card and a GOP rom to go to a shell and see if the GOP rom is mapped into ram and or linked, but no one has helped. Even tho I know lots of people use PC Cards in their Mac Pro's and would like to have boot screens for them via the GOP rom that is already on the card.

Show some love people, help me, help everyone.

The more I ask, and the more people don't help, the more hardware I have to buy, and the more inclined I am to just keep my mouth shut about how I get these things to work unless someone is willing to pay for it.

I don't find hacking and making money to be mutually exclusive, and l like to recover the hardware costs it takes me to come up with a useful hack, maybe a bit more so I can come up with the next hack.

However I try and document my hacks, so people that don't want things handed to them in an easy package still have a way of doing for themselves.

I've found over the years of hacking, some people just have more money than time, and even if you explain to them how to do something, they'd rather just pay for it, and the people that read all the documentation and build things from source and debug, normally find things I miss and end up adding to a more perfect solution.

Of course Alfa and Beta testers are part of that process, and are always welcome.

It's exactly the same problem I found myself developing Next Loader, it requires money to buy hardware to test, and only a few people decided to collaborate as beta testers.

 

[DearthnVader]

I haven't been able to answer or solve the open issues in the Next Loader repository yet, because I haven't had any free time because of my work. I will try to get some free time this weekend to review the bugs and update the base code to the latest version of rEFInd.

In part, I paused the project for lack of beta testers and people who would like to collaborate, but seeing this thread it seems that several people are willing to collaborate.
[doublepost=1534877130][/doublepost]
It's exactly the same problem I found myself developing Next Loader, it requires money to buy hardware to test, and only a few people decided to collaborate as beta testers.

ThanKs Abdy, I wasn't sure if you were still working on the project.. I really like it a lot, and I'm sure I'd like to pick you brain to see what we can come up with.

Sadly, it just won't run on my Mac, so I can't offer much feedback, and I haven't looked very deep into the code to find the source I my error, just too many hacks and so little time.
[doublepost=1534878501][/doublepost]

I doubt apple would pull the MP5,1 this late in the game especially considering the BootROM updates apple have made for the 5,1.


the G3 beige thing IIRC was also over OpenGL acceleration which the Rage Pro GPU supported in Mac OS 8.x/9.x but did not in OS X that pissed off a few people to say the least.

and no OS X 10.x DP ever supported any Pre G3 Macintosh the only versions of OS X that do are the 1.x Sever versions. and there where proper retail releases of OS X 1.x as OS X Server 1.x, so no Beige machine ever got dropped during a DP cycle.

the only PCI Beige machine with HDD/CDROM Booting issues like what you describe was the Power Macintosh 4400 where one of its ATA channels could not be used in OpenFirmware (at least as it stands, maybe if someone developed a proper nvramrc script patch for the 4400 it could of, the 4400 was very much a red-headed step child, you can read about my antics with mine here https://forums.macrumors.com/threads/success-with-the-power-macintosh-4400.2077767/ )

all other Macs where fine at an OpenFirmware level AFAIK. (thanks in part to them having nice decent NVRAMRC patches)

Thanks for the correction, sometime we tend to remember things different than they actually were. More the way we want to remember them, this has been shown in studies to be true, ask three people how they remember a crime they all witnessed, and you'll get three different accounts, even if it just happened.

Sad for people on death row do to "Eye Witness" testimony.

 

[Ludacrisvp]

Yes, beta tester welcome and appreciated.

I've asked a few times for someone with an nVidia card and a GOP rom to go to a shell and see if the GOP rom is mapped into ram and or linked, but no one has helped. Even tho I know lots of people use PC Cards in their Mac Pro's and would like to have boot screens for them via the GOP rom that is already on the card.

Show some love people, help me, help everyone.

The more I ask, and the more people don't help, the more hardware I have to buy, and the more inclined I am to just keep my mouth shut about how I get these things to work unless someone is willing to pay for it.

I don't find hacking and making money to be mutually exclusive, and l like to recover the hardware costs it takes me to come up with a useful hack, maybe a bit more so I can come up with the next hack.

However I try and document my hacks, so people that don't want things handed to them in an easy package still have a way of doing for themselves.

I've found over the years of hacking, some people just have more money than time, and even if you explain to them how to do something, they'd rather just pay for it, and the people that read all the documentation and build things from source and debug, normally find things I miss and end up adding to a more perfect solution.

Of course Alfa and Beta testers are part of that process, and are always welcome.

If I had any relevant hardware I’d help.
But I’d doubt that HD5770/5450 and GTX650/8800GT in a 3,1 have any relevance these days.

 

[h9826790]

There are few other .efi files inside the refind folder. So, only the REFIND.efi is the correct one, right?

To test it, I only need to

1) mount a EFI partiton
2) copy the whole refind folder into the EFI folder of the EFI partiton
3) bless the REFIND.efi file correctly
4) reboot

And this should be done with my original Sapphire PULSE RX580 ROM

Is that correct?

 

[thomasthegps]

Yes, beta tester welcome and appreciated.

I've asked a few times for someone with an nVidia card and a GOP rom to go to a shell and see if the GOP rom is mapped into ram and or linked, but no one has helped. Even tho I know lots of people use PC Cards in their Mac Pro's and would like to have boot screens for them via the GOP rom that is already on the card.

Show some love people, help me, help everyone.

The more I ask, and the more people don't help, the more hardware I have to buy, and the more inclined I am to just keep my mouth shut about how I get these things to work unless someone is willing to pay for it.

I don't find hacking and making money to be mutually exclusive, and l like to recover the hardware costs it takes me to come up with a useful hack, maybe a bit more so I can come up with the next hack.

However I try and document my hacks, so people that don't want things handed to them in an easy package still have a way of doing for themselves.

I've found over the years of hacking, some people just have more money than time, and even if you explain to them how to do something, they'd rather just pay for it, and the people that read all the documentation and build things from source and debug, normally find things I miss and end up adding to a more perfect solution.

Of course Alfa and Beta testers are part of that process, and are always welcome.

I have a GTX 960 laying around, I can try stuff out. To be honest I was completely unaware that this thread existed... Sometime ago I had installed reFind on my Mac Pro trying to manually change the pci speed in the EFI console to 5GT/sec as detailed on a netkas post to no avail. The OP was using refit I think instead of refind, maybe thats why it didn't work...

 

[h9826790]

Anyway, Apple has reply to my “no boot screen” bug report (for the Sapphire PUSLE RX580).

In fact, the reply came within one hour after I submit the bug report. It breaks my “quickest reply” record.

I am sure they are looking into this issue now. May be at quite high priority as well.

My bug report about the no HEVC decoding except in QT Player also got reply within 10hours (again, for the Sapphire PULSE RX580).

It seems Apple really willing to “fix” the issue which is Sapphire PULSE RX580 related now.

For those who want the boot screen, please go to bugreport.apple.com and submit a bug report.

 

[thomasthegps]

Anyway, Apple has reply to my “no boot screen” bug report (for the Sapphire PUSLE RX580).

In fact, the reply came within one hour after I submit the bug report. It breaks my “quickest reply” record.

I am sure they are looking into this issue now. May be at quite high priority as well.

My bug report about the no HEVC decoding except in QT Player also got reply within 10hours (again, for the Sapphire PULSE RX580).

It seems Apple really willing to “fix” the issue which is Sapphire PULSE RX580 related now.

For those who want the boot screen, please go to bugreport.apple.com and submit a bug report.

well it looks like we might get bootscreens !

 

[bookemdano]

Anyway, Apple has reply to my “no boot screen” bug report (for the Sapphire PUSLE RX580).

In fact, the reply came within one hour after I submit the bug report. It breaks my “quickest reply” record.

I am sure they are looking into this issue now. May be at quite high priority as well.

My bug report about the no HEVC decoding except in QT Player also got reply within 10hours (again, for the Sapphire PULSE RX580).

It seems Apple really willing to “fix” the issue which is Sapphire PULSE RX580 related now.

For those who want the boot screen, please go to bugreport.apple.com and submit a bug report.

Encouraging! What do the replies say? I've only used Feedback Assistant to report a few bugs and I don't see any reply there, but maybe Apple doesn't reply to bugs submitted that way. Is bugreport.apple.com the best way to go?

 

[h9826790]

Encouraging! What do the replies say? I've only used Feedback Assistant to report a few bugs and I don't see any reply there, but maybe Apple doesn't reply to bugs submitted that way. Is bugreport.apple.com the best way to go?

I never receive any reply in feedback assistant. But most of my bug reports in bugreport.apple.com does. I really believe that's the best way to report a bug (or give suggestion).

They gave me the "standard" reply, ask me to submit the sysdiagnoses. And I just explained to them it's impossible because the boot screen issue happens even before macOS start to load.

I have this similar conversation with the Apple engineer regarding the "cannot update firmware with the RX580", they also ask for sysdiagnoses, I told them it's impossible and explain the why. Then they fixed it after 2 weeks.

I don't think the boot screen will really come after another 2 weeks, but it seems they really working on it.

They gave my no reply about the HEVC encoding bug report yet, which I submitted before the "boot screen bug report". So, I believe this boot screen issue is at a higher priority.

I never expect that they will reply my bug report in just 52min.

 

[crjackson2134]

I’ve gotten questions and diagnostic instructions based on bug reports filed with the feedback assistant, but it comes to my email. Otherwise I don’t usually get much bounce back communications.

They did fix one bug I filed against safari and were nice enough to tell me they addressed the bug based on my report, and to test further to ensure it was corrected. It was...

The web interface reporter has been more responsive for me.

 

[DearthnVader]

There are few other .efi files inside the refind folder. So, only the REFIND.efi is the correct one, right?

To test it, I only need to

1) mount a EFI partiton
2) copy the whole refind folder into the EFI folder of the EFI partiton
3) bless the REFIND.efi file correctly
4) reboot

And this should be done with my original Sapphire PULSE RX580 ROM

Is that correct?

That sounds about correct.

 

[MisterAndrew]

Anyway, Apple has reply to my “no boot screen” bug report (for the Sapphire PUSLE RX580).

In fact, the reply came within one hour after I submit the bug report. It breaks my “quickest reply” record.

I am sure they are looking into this issue now. May be at quite high priority as well.

My bug report about the no HEVC decoding except in QT Player also got reply within 10hours (again, for the Sapphire PULSE RX580).

It seems Apple really willing to “fix” the issue which is Sapphire PULSE RX580 related now.

For those who want the boot screen, please go to bugreport.apple.com and submit a bug report.

It does appear making new AMD cards work properly with the Mac Pro is a priority right now since they have been responding to bug reports for things like the fans and boot screen, and also fixed the link speed. I bet we’ll have boot screens in time for the GM.

 

[tsialex]

It does appear making new AMD cards work properly with the Mac Pro is a priority right now since they have been responding to bug reports for things like the fans and boot screen, and also fixed the link speed. I bet we’ll have boot screens in time for the GM.

I just want the SMC update…

 

[MisterAndrew]

I just want the SMC update…

What will the SMC update provide?

 

[DearthnVader]

I have a GTX 960 laying around, I can try stuff out. To be honest I was completely unaware that this thread existed... Sometime ago I had installed reFind on my Mac Pro trying to manually change the pci speed in the EFI console to 5GT/sec as detailed on a netkas post to no avail. The OP was using refit I think instead of refind, maybe thats why it didn't work...

I mucked about in the efi shell with those commands too, trying to get 8GT/s working on my Qemu virtual machine for High Sierra with the RX580, no love for me either, tho Mojave PB's macOS runtime drivers enable it

I'd be really interested in the results of the "drivers" command from an EFI Shell with the GTX 960 installed, it maybe some time before I get my hands on another nVidia card with a GOP firmware on it.

I posted a EFI Shell and some basic instructions on how to bless/boot and use it.

At this time, no one has reported any nVidia card as working with GOP enabled boot loaders in a Mac Pro, so two cards are required for testing, one with a Apple compatible UGA EFI ROM, and an nVidia PC Card with GOP.

 

[tsialex]

If we have a new SMC version, we could update 2009 single processors to 2012 and have total parity, maybe even duals firmware parity.

Trays will be interchangeable between years.
[doublepost=1534908373][/doublepost]

one with a Apple compatible UGA EFI ROM, and an nVidia PC Card with GOP.

I think I have both, eVGA GTX 680 Mac Edition and eVGA GTX 680 FTW+ 4GB. But I have to much projects at the moment, people will eat me alive if I stop the BootROM reconstruction one.

 

[MriX]

Anyway, Apple has reply to my “no boot screen” bug report (for the Sapphire PUSLE RX580).

In fact, the reply came within one hour after I submit the bug report. It breaks my “quickest reply” record.

I am sure they are looking into this issue now. May be at quite high priority as well.

My bug report about the no HEVC decoding except in QT Player also got reply within 10hours (again, for the Sapphire PULSE RX580).

It seems Apple really willing to “fix” the issue which is Sapphire PULSE RX580 related now.

For those who want the boot screen, please go to bugreport.apple.com and submit a bug report.

I have opened 2 bugs about the Bootscreen Problem and never got a reply for 10 days.

 

[bookemdano]

I wonder... since h9826790 has submitted previous bug reports that were acted upon if his subsequent bug reports get flagged for closer attention than ones coming from users who haven't submitted lots of previous reports. It wouldn't surprise me if they have such a system that ranks bug reports from trusted/reliable people higher than the rest of us.

Anyway, we should all be happy that *someone* is getting a response from Apple, even if it's not us. Because we will all benefit from any fixes. We should all continue to report everything though (and re-report with each beta) because the more people who report it, the better chance it gets noticed and perhaps fixed.

 

[MriX]

I wonder... since h9826790 has submitted previous bug reports that were acted upon if his subsequent bug reports get flagged for closer attention than ones coming from users who haven't submitted lots of previous reports. It wouldn't surprise me if they have such a system that ranks bug reports from trusted/reliable people higher than the rest of us.

Anyway, we should all be happy that *someone* is getting a response from Apple, even if it's not us. Because we will all benefit from any fixes. We should all continue to report everything though (and re-report with each beta) because the more people who report it, the better chance it gets noticed and perhaps fixed.

That could be totally possible. Maybe the also sort for paid developer accounts? I’m using only the public beta.

 

[h9826790]

I have opened 2 bugs about the Bootscreen Problem and never got a reply for 10 days.

That could be totally possible. Maybe the also sort for paid developer accounts? I’m using only the public beta.

I am also only a normal beta tester, no paid developer account. That's why I use PB, but not DP. But I do wrote bug reports for quite a few years already, and always try to provide as much data / details as possible.

When I wrote this boot screen issue report. I emphasised the terms "Apple recommended Sapphire PULSE RX580", may be that make the difference.

Anyway, at least one bug report is addressed. The hope still there.
[doublepost=1534922120][/doublepost]

That sounds about correct.

Have a sudden duty changed. So had few hours for some tests.

What I did

1) insert a USB thumb drive into the cMP front USB 2.0 port. That's a known good bootable USB drive.
2) format the drive to HFS+ (let Disk Utility to re-create the required EFI partition)
3) run

diskutil list

to check the USB thumb driver disk number

4) mount the EFI partition

sudo mkdir /Volumes/EFI

then

sudo mount -t msdos /dev/diskxs1 /Volumes/EFI

with corrected disk number

5) create a EFI folder inside the EFI partition in Finder
5) copy the refind folder to the EFI folder inside the EFI partition
6) bless the REFIND.efi

sudo bless --mount /Volumes/EFI --setBoot -nextonly --file /Volumes/EFI/efi/refind/REFIND.efi

I add the next only argument to avoid possible further trouble

Then

I tested threes different scenario

A) Modded RX580 ROM (GPU clock, voltage pointer, VRAM voltage) -> bless REFIND.efi
B) Original RX580 ROM -> bless REFIND.efi
c) Original RX580 ROM -> bless refind_GOPx64.efi

All three scenario has no boot screen with just the RX580 installed in slot 1.

rEFInd log shows the following for all 3 tests

Obviously case C is not good because I tried to bless the refind_GOPx64.efi but not REFIND.efi

Case A and B has ScanVolumes, but nothing to display.

I am 100% sure Case B using the original RX580 ROM that I dumped from the card.

 

[DearthnVader]

I am also only a normal beta tester, no paid developer account. That's why I use PB, but not DP. But I do wrote bug reports for quite a few years already, and always try to provide as much data / details as possible.

When I wrote this boot screen issue report. I emphasised the terms "Apple recommended Sapphire PULSE RX580", may be that make the difference.

Anyway, at least one bug report is addressed. The hope still there.
[doublepost=1534922120][/doublepost]

Have a sudden duty changed. So had few hours for some tests.

What I did

1) insert a USB thumb drive into the cMP front USB 2.0 port. That's a known good bootable USB drive.
2) format the drive to HFS+ (let Disk Utility to re-create the required EFI partition)
3) run

diskutil list

to check the USB thumb driver disk number

4) mount the EFI partition

sudo mkdir /Volumes/EFI

then

sudo mount -t msdos /dev/diskxs1 /Volumes/EFI

with corrected disk number

5) create a EFI folder inside the EFI partition in Finder
5) copy the refind folder to the EFI folder inside the EFI partition
6) bless the REFIND.efi

sudo bless --mount /Volumes/EFI --setBoot -nextonly --file /Volumes/EFI/efi/refind/REFIND.efi

I add the next only argument to avoid possible further trouble

Then

I tested threes different scenario

A) Modded RX580 ROM (GPU clock, voltage pointer, VRAM voltage) -> bless REFIND.efi
B) Original RX580 ROM -> bless REFIND.efi
c) Original RX580 ROM -> bless refind_GOPx64.efi

All three scenario has no boot screen with just the RX580 installed in slot 1.

rEFInd log shows the following for all 3 tests
View attachment 777172
Obviously case C is not good because I tried to bless the refind_GOPx64.efi but not REFIND.efi

Case A and B has ScanVolumes, but nothing to display.

I am 100% sure Case B using the original RX580 ROM that I dumped from the card.

Not sure the trouble, but if that is the entire log, then it looks like it's stalling scanning the discs. See my full log from https://forums.macrumors.com/thread...-output-protocol.2089181/page-4#post-26367972

It maybe because rEFInd doesn't like to be booted from USB?

 

[h9826790]

Not sure the trouble, but if that is the entire log, then it looks like it's stalling scanning the discs. See my full log from https://forums.macrumors.com/thread...-output-protocol.2089181/page-4#post-26367972

It maybe because rEFInd doesn't like to be booted from USB?

Possible, I try it again from an internal HDD.
[doublepost=1534926164][/doublepost]Tested, same result for internal SATA II connected 8TB WD Red HDD (Verified bootable, it's my daily clone backup boot drive)