Jump to content
586 posts in this topic

Recommended Posts

Clover did no graphics initialization. Pause.... It is BIOS work.

Injection is a set of pairs <key>/<value> and the set is not ordered.

 

I can't make Radeon wake with any Intel injections and initializations. I want another way.

 

Yeah, that's what I meant (not initialized, detect). I also can't get sleep to work, even when settings IGPU to init display first. Strangely though, I still see POST/UEFI/Clover on the display attached to my 280X (even though Clover detects both as if I was using IGPU as primary).

This is a known issue. I think there was a patch posted in either the Clover general discussion/Clover issue reporting threads that initialized the IGPU before the PCIe GPU so you could inject both when PEG is set to primary.

 

Thanks, TRM. I guess you mean this? http://www.insanelymac.com/forum/topic/306156-clover-bugissue-report-and-patch/?p=2173938

Not sure how to apply this.. Looks source code... Will it be included in the next release? I'd really like to give this a shot, no matter if it requires some efforts..

I tried it on the machine in my signature without success because when I select the IGPU as primary graphics the machine doesn't reach S3 anymore.  :(

 

Mieze 

You mean with your H97M System? You did not indicate your Gfx-Card there.. So IGPU works but does not sleep?

 

Doesn't appear to be fixed in 10.11.2b1, for anyone wondering.

 

That is disappointing.. I was hoping it would.

 

Did anybody with a "Mac Edition" Radeon try to use it in his/her Hackintosh? I was wondering what the result of that would be.. I found a guide from Netkas for modding vBIOS for EFI compatibility and reflashing. Didn't dare to reflash, instead I made the modifications with Clover PatchVBios Find&Replace. My config file grew from 4 kb to 180 kb  :blink:. Couldn't check if patches were actually applied, but I got no errors and transition from boot screen to login screen seemed much smoother. Unfortunately, this did not fix wake either  :(

I've connected the drive to the last port and that makes the difference, it's not working every time but on 80-90% wake is successfull.

After i've renamed the SATA devices PRIM to PRT0 and SECD to PRT1 there is no difference on which port the drive is connected.

With an SSD much lower success, about 10-20%

 

Also plugged in an old PCI graphics card and set PCI init first on BIOS an have 100% working wake from sleep, thanks Reno_S for the suggestion.

Did anybody with a "Mac Edition" Radeon try to use it in his/her Hackintosh? I was wondering what the result of that would be.. I found a guide from Netkas for modding vBIOS for EFI compatibility and reflashing. Didn't dare to reflash, instead I made the modifications with Clover PatchVBios Find&Replace. My config file grew from 4 kb to 180 kb  :blink:. Couldn't check if patches were actually applied, but I got no errors and transition from boot screen to login screen seemed much smoother. Unfortunately, this did not fix wake either  :(

 

I already answered that in this thread. There's no use whatsoever for a Mac Edition card in a Hackintosh, the Mac EFI will be ignored since it isn't UEFI compliant. The vBIOS part of the ROM is 1:1 identical to the PC vBIOS.

The Mac EFI doesn't do much more than Clovers AtiInject=True, btw.

 

And as I said, this problem doesn't affect real Macs, no matter if you use a Mac Edition card or a stock PC card.

You mean with your H97M System? You did not indicate your Gfx-Card there.. So IGPU works but does not sleep?

 

No, I was talking about the machine with the Asrock Fatal1ty B85 Killer board and the MSI Radeon 270X Twin Frozr GPU. The other machines are using onboard graphics.

 

Mieze

I tried to remove AMDRadeonX4000.kext from /S/L/E and the system (Radeon R9 270X) woke up from sleep successfully but froze about 30 seconds after that which tells me that there must be something else causing the trouble, probably in the ACPI data structures.

 
Mieze

 

I tried to remove AMDRadeonX4000.kext from /S/L/E and the system (Radeon R9 270X) woke up from sleep successfully but froze about 30 seconds after that which tells me that there must be something else causing the trouble, probably in the ACPI data structures.

 
Mieze

 

Yes, I already had the wake deleting some kexts.

I think the pause/trouble caused by testing ambient light sensor which we have no.

For what its worth, I began seeing this behavior 2 days ago on Yosemite (10.10.5).  I had made no attempt to install/update to El Capitan nor had I installed any other recent system updates.  Sleep had been working fine since installation 2 months ago. 

 

I removed my GeForce video card but the problem persists even with onboard Intel HD Graphics 4000. 

I think the pause/trouble caused by testing ambient light sensor which we have no.

I don't think so. If it was the light sensor, using a MacPro system definition would resolve the problem as there is no ALS on these machines.

 

Mieze

What else the difference between Mac and Hack?

Maybe some missing device-properties? https://pikeralpha.wordpress.com/2015/10/24/device-properties-used-in-the-new-imac171/

 

Also from what I read in this thread, the wake is possible with legacy boot and not possible on UEFI. MacPros are waking properly with PC cards, what device-properties do they have? 

What else the difference between Mac and Hack?

The ACPI data structures. As of now I focused on the wakeup code but the problem might also be located in the sleep routines. It's obvious that Apple has reworked much of the code which deals with the ACPI stuff and it wouldn't be surprising if that's the reason. Many things that worked for years, are causing trouble now. Take the USB subsystem for example.

 

Mieze

Edited by Mieze

...the problem might also be located in the sleep routines.

I think I can confirm this. I compared the logs during sleep and wake for failing and successful sleep/wakes (with IGPU active). Comparing both logs line by line, I found the problem begins with this line:

 

"WindowServer[163]: no sleep images for WillPowerOffWithImages"

 

Found a thread where also a genuine Mac Pro produced the same message with the same result.. Not sure if it can be compared directly, but here, the problem was apparently related to USB

https://discussions.apple.com/thread/7265175

 

So, it seems like the problem is with the sleep routines, where no sleep image can be created (and thus can't wake from one).

 

Reno

I can't exclude that we do something wrong in DSDT or DeviceProperties but I surely claim that the problem elsewhere because we tested with applelife.ru guys mostly all possible variants without success.

 

I have reports from native iMac12,2 before and after wake. I found nothing new.

 

I'll be glad to be wrong.

AMD graphics not wakes from sleep in ElCapitan. This is a task for research.

not true my r7 250 6870 7970 290x and 390x all wake just fine without issue audio works usb works if there is a issue it's not with everyone

not true my r7 250 6870 7970 290x and 390x all wake just fine without issue audio works usb works if there is a issue it's not with everyone

Have you read the whole thread? We know they wake in "non-primary" configurations. But not with PCIe as primary.. How are your configurations?

Have you read the whole thread? We know they wake in "non-primary" configurations. But not with PCIe as primary.. How are your configurations?

pci-e slot 0 first slot on my board boot as main card always 

I've got a new idea. I'm using system definition iMac15,1 and inject the proper platform ID for this system in the DSDT. On boot I get the following kernel message:

Oct 29 00:56:35 Lucy kernel[0]: IGPU: Not usable

Another difference is that my IORegistry entries of the IGPU differ from those of the real iMac15,1, in particular I've got AppleUpstreamUserClient, which is something you only have on "real" display adapters, attached to the device but the real iMac15,1 doesn't have it. I checked and found out that is is also true in 10.10.5, although it doesn't cause any problems on Yosemite.

 

When I completely disable the IGPU via _STA in the DSDT, the message changes to:

Oct 29 15:52:46 Lucy kernel[0]: display: Not usable

Injecting a platform ID with > 0 connectors makes the message disappear, but of course doesn't solve the problem.

 

Well, what if the graphics subsystem is still attached to the IGPU in any way even if the PCIe graphics card is set as the primary adapter so that it gets confused on wakeup because it tries to wake up the IGPU no matter what you selected? Basically it boils down to a communication problem between BIOS/DSDT and the OS with regard to the primary display adapter. This would also explain why using the IGPU as the primary display seems to solve the problem for some users? And it would explain the strange IORegistry entry!

 

Mieze

Edited by Mieze

I'm joining the group of people who are struggeling to get wake working with a Radeon 7770 in El Capitan.

 

I can report that when waking with the Radeon 7770 I have black screen. After power off and boot I can find following message in the syslog as already reported in earlier posts:

kernel[0]: AMDFramebufferSI::setPowerState(0x9c671b04c8e912f5, 0 -> 1) times out after 45914 ms

 

When booting with the onboard HD4000 sleep/wake works w/o problems.

I've tried the fix as described by setting the iGPU as primary but no success. When booting with the iGPU as primary my system even doesnt load the Clover boot screen, I only have a flashing cursor. This also happens when booting from the Clover USB installer so it might be a specific error to my system.

 

I'm looking forward to help with fixing this issue

I enabled embedded graphic in BIOS, made FakeID=0x01068086 and got new result

29.10.15 21:40:28,000 kernel[0]: ATIFramebufferNI::setPowerState(0x260361a46e0ba681, 0 -> 2) timed out after 45377 ms
29.10.15 21:40:28,000 kernel[0]: AppleIntelFramebuffer::setPowerState(0x260361a46dc47681, 0 -> 2) timed out after 45384 ms

Both adapters not wake!

  • Like 1
×
×
  • Create New...