No graphics / USB / Audio after wake

[Slice]

I have 6870 and sleep/wake worked right after I upgraded to El Capitan while the shutdown doesn't work (always restart instead of shutdown). I then upgraded FakeSMC and move all kexts from /S/L/E/ to Clover/kexts/Others/ this resolved my shutdown (rebooting) problem, but my sleep/wake had the problem described in this thread. I tried to revert back the FakeSMC which make sleep problem even wrose (restart right after wake).

 

I suspect this problem might be related to something changed in FakeSMC, my older one is from Mountain Lion time which worked perfectly until El Capitan released. I don't really have time to try different versions of FakeSMC, but I just want to give my 0.02c that we might looked in the wrong direction before.

It will be very interesting if you can revert to those situation with working sleep/wake. I doubt it will happen.

[banini_jeque]

I have that same line, but it's from yesterday, and I also attempted to wake from sleep a few times today:

10/14/15 11:57:09.000 PM kernel[0]: ATIFramebufferNI::setPowerState(0xe78d9024e0e4068b, 0 -> 1) timed out after 45590 ms

 

Here's a full dump from wake up to hard power off, and the above error does not appear in it:

 

10/15/15 11:36:06.000 PM kernel[0]: Wake reason: GBE EHC1
10/15/15 11:36:06.000 PM syslogd[45]: ASL Sender Statistics
10/15/15 11:36:06.000 PM kernel[0]: RTC: PowerByCalendarDate setting ignored
10/15/15 11:36:06.000 PM kernel[0]: Previous sleep cause: 5
10/15/15 11:36:06.000 PM kernel[0]: Sandbox: mDNSResponder(93) deny(1) mach-lookup com.apple.distributed_notifications@1v3
10/15/15 11:36:06.000 PM kernel[0]: full wake promotion (reason 1) 56 ms
10/15/15 11:36:06.067 PM com.apple.WebKit.Networking[494]: tcp_connection_destination_perform_socket_connect 3312 connectx to 52.0.176.163:80@0 failed: [50] Network is down
10/15/15 11:36:06.300 PM loginwindow[96]: ERROR | -[LWBuiltInScreenLockAuthLion closeAuthAndReset:] | Attempted to remove an observer when not observing
10/15/15 11:36:06.439 PM com.apple.WebKit.WebContent[565]: [23:36:06.439] FigAgglomeratorSetObjectForKey signalled err=-16020 (kFigStringConformerError_ParamErr) (NULL key) at /Library/Caches/com.apple.xbs/Sources/CoreMedia/CoreMedia-1731.15.4/Prototypes/LegibleOutput/FigAgglomerator.c line 92
10/15/15 11:36:06.447 PM com.apple.WebKit.WebContent[565]: [23:36:06.447] FigAgglomeratorSetObjectForKey signalled err=-16020 (kFigStringConformerError_ParamErr) (NULL key) at /Library/Caches/com.apple.xbs/Sources/CoreMedia/CoreMedia-1731.15.4/Prototypes/LegibleOutput/FigAgglomerator.c line 92
10/15/15 11:36:08.000 PM kernel[0]: in6_unlink_ifa: IPv6 address 0x2b3a8d777aa4175 has no prefix
10/15/15 11:36:08.771 PM configd[53]: LINKLOCAL en0: parent has no IP
10/15/15 11:36:08.781 PM configd[53]: network changed: v4(en0-:192.168.9.104) DNS- Proxy-
10/15/15 11:36:08.822 PM com.apple.WebKit.WebContent[565]: [23:36:08.822] HTTPRequest figHttpRequestDidFailCallback: Network error: domain=kCFErrorDomainCFNetwork, code=-1009, Error Domain=kCFErrorDomainCFNetwork Code=-1009 "The Internet connection appears to be offline." UserInfo={NSErrorFailingURLStringKey=http://cdnapi.kaltura.com/p/1807801/sp/180780100/playManifest/entryId/1_20imy7k0/flavorIds/1_9dk2ml77,1_79byy4st,1_7y1n7rbg,1_q9so16ir/format/applehttp/protocol/http/a.m3u8?referrer=aHR0cDovL3d3dy5hdXRvbW9iaWxlbWFnLmNvbQ==&ks=djJ8MTgwNzgwMXy4iT-UelMxo6W00J8LLGPB6Y47WjWut_9NKqnKMl3XXZu2bQ1zdWTdU0uNSZJ1uf06yiuveI12mevqlWNzYV9UImpHem-xEZYXQiLWK8RL_g==&playSessionId=fcc41fb0-fd5d-cc9b-f7b8-9e04f50dd753&clientTag=html5:v2.35.6&preferredBitrate=1600&uiConfId=30359752, NSErrorFailingURLKey=http://cdnapi.kaltura.com/p/1807801/sp/180780100/playManifest/entryId/1_20imy7k0/flavorIds/1_9dk2ml77,1_79byy4st,1_7y1n7rbg,1_q9so16ir/format/applehttp/protocol/http/a.m3u8?referrer=aHR0cDovL3d3dy5hdXRvbW9iaWxlbWFnLmNvbQ==&ks=djJ8MTgwNzgwMXy4iT-UelMxo6W00J8LLGPB6Y47WjWut_9NKqnKMl3XXZu2bQ1zdWTdU0uNSZJ1uf06yiuveI12mevqlWNzYV9UImpHem-xEZYXQiLWK8RL_g==&playSessionId=fcc41fb0-fd5d-cc9b-f7b8-9e04f50dd753&clientTag=html5:v2.35.6&preferredBitrate=1600&uiConfId=30359752, _kCFStreamErrorCodeKey=8, _kCFStreamErrorDomainKey=12, NSLocalizedDescription=The Internet connection appears to be offline.} for http://cdnapi.kaltura.com/p/1807801/sp/180780100/playManifest/entryId/1_20imy7k0/flavorIds/1_9dk2ml77,1_79byy4st,1_7y1n7rbg,1_q9so16ir/format/applehttp/protocol/http/a.m3u8?referrer=aHR0cDovL3d3dy5hdXRvbW9iaWxlbWFnLmNvbQ==&ks=djJ8MTgwNzgwMXy4iT-UelMxo6W00J8LLGPB6Y47WjWut_9NKqnKMl3XXZu2bQ1zdWTdU0uNSZJ1uf06yiuveI12mevqlWNzYV9UImpHem-xEZYXQiLWK8RL_g==&playSessionId=fcc41fb0-fd5d-cc9b-f7b8-9e04f50dd753&clientTag=html5:v2.35.6&preferredBitrate=1600&uiConfId=30359752
10/15/15 11:36:08.853 PM configd[53]: network changed: v4(en0+:192.168.9.104) DNS+ Proxy+ SMB
10/15/15 11:36:08.907 PM mDNSResponder[93]: mDNS_RegisterInterface: Frequent transitions for interface en0 (192.168.9.104)
10/15/15 11:36:10.093 PM configd[53]: network changed: v4(en0-:192.168.9.104) DNS- Proxy-
10/15/15 11:36:10.584 PM netbiosd[401]: network_reachability_changed : network is not reachable, netbiosd is shutting down
10/15/15 11:36:10.586 PM com.apple.WebKit.WebContent[565]: [23:36:10.586] <<<< CRABS >>>> crabsFlumeHostUnavailable: [0x7f976690bda0] Byte flume reports host unavailable.
10/15/15 11:36:10.587 PM symptomsd[232]: __73-[NetworkAnalyticsEngine observeValueForKeyPath:ofObject:change:context:]_block_invoke unexpected switch value 2
10/15/15 11:36:15.722 PM configd[53]: network changed: v4(en0+:192.168.9.100) DNS+ Proxy+ SMB
10/15/15 11:36:16.206 PM symptomsd[232]: __73-[NetworkAnalyticsEngine observeValueForKeyPath:ofObject:change:context:]_block_invoke unexpected switch value 2
10/15/15 11:36:16.207 PM com.apple.WebKit.WebContent[565]: [23:36:16.207] <<<< CRABS >>>> crabsFlumeHostAvailable: [0x7f976690bda0] Byte flume reports host available again.
10/15/15 11:36:16.207 PM com.apple.WebKit.WebContent[565]: [23:36:16.207] segPumpSetCurrentAlternate signalled err=-12640 (kFigBytePumpError_ParamErr) (no such alternate) at /Library/Caches/com.apple.xbs/Sources/CoreMedia/CoreMedia-1731.15.4/Prototypes/FigByteStreamPrototypes/FigBytePumpHTTPSegment.c line 14541
10/15/15 11:36:16.303 PM loginwindow[96]: CoreAnimation: warning, deleted thread with uncommitted CATransaction; set CA_DEBUG_TRANSACTIONS=1 in environment to log backtraces.

 

 

 

I also have these two lines a couple of times:

10/15/15 12:32:45.000 AM kernel[0]: AMDRadeonX3000_AMDAccelDevice: IOUserClient outputCount count mismatch

10/15/15 9:29:40.005 AM WindowServer[214]: WSMachineIsAMD: true

Not sure why it would be RadeonX3000 when I have a 6870, but I don't know all that much about those cards. Hope this helps!

[Slice]

Yes, Radeon 6870 is controlled by RadeonX3000.kext.

This is interesting observation.

I may also say that if we disable acceleration by removing some AMD kexts then sleep/wake will work. Looks like they replaced plain framebuffer with accelerated one.

Everyone here. Look, please, OpenGL Extension viewer, how your device is named.

if HD6670 - good

if Turks Prototype - bad

[theandy94]

 

I have that same line, but it's from yesterday, and I also attempted to wake from sleep a few times today:

10/14/15 11:57:09.000 PM kernel[0]: ATIFramebufferNI::setPowerState(0xe78d9024e0e4068b, 0 -> 1) timed out after 45590 ms

 

Here's a full dump from wake up to hard power off, and the above error does not appear in it...

If you convert the time counted in ms (milliseconds) to seconds you can see, that it takes about 45 seconds for setPowerState to time out.

Since you rebooted after just 10 seconds there is no chance for it to time out.

 

EDIT:

 

Everyone here. Look, please, OpenGL Extension viewer, how your device is named.

if HD6670 - good

if Turks Prototype - bad

[Slice]

If you convert the time counted in ms (milliseconds) to seconds you can see, that it takes about 45 seconds for setPowerState to time out.

Since you rebooted after just 10 seconds there is no chance for it to time out.

 

EDIT:

Screenshot.png

Yes, Prototype - this is bad.

Can you make any additional injection, for example by FakePCIID.kext (rehabman) for the better identification?

[rayap]

Rolled back to 10.10.5 AMD* and IOAcc*, sleep/wake works.

[Slice]

Rolled back to 10.10.5 AMD* and IOAcc*, sleep/wake works.

I checked this way but I got lags and freezes.

[wcpemm]

I'm having a similar problem, but I'm not running an AMD. My machine will sleep and wake on USB input, but after a long sleep the USBs all seem to shut off after it gets the initial wake signal. Plugging and unplugging the mouse/keyboard (which are in 2.0 ports) doesn't work, and I have to restart.

 

Tried to attach my config.plist in case it was relevant but I guess my account is too new, it won't allow me to.

[holyfield]

Maybe this article is helpful?

 

http://www.idelta.info/archives/some-hidden-changes-in-os-x-el-capitan/

[banini_jeque]

If you convert the time counted in ms (milliseconds) to seconds you can see, that it takes about 45 seconds for setPowerState to time out.

Since you rebooted after just 10 seconds there is no chance for it to time out.

 

 

Haha, I just realized that this morning. Silly me.

[Slice]

Maybe this article is helpful?

 

http://www.idelta.info/archives/some-hidden-changes-in-os-x-el-capitan/

Yes, this article will be helpful for someone wanted to have hibernation. It tell how to set mode 29!

I am not successful in hibernation with any flags.

[banini_jeque]

In OpenGL Driver Monitor, mine says AMDRadeonX3000GLDriver: AMD Radeon Barts XT Prototype, so yeah, still prototype which I assume isn't what will work for us right now, right Slice?

[Slice]

In OpenGL Driver Monitor, mine says AMDRadeonX3000GLDriver: AMD Radeon Barts XT Prototype, so yeah, still prototype which I assume isn't what will work for us right now, right Slice?

It works but without awake.

[Stiffy]

prototype, bad

[Stiffy]

update to 10.11.1 Beta (15B38b), still no wake :(

[Fl0r!an]

Made a short experiment, can confirm that this issue is 100% Hackintosh only:

 

- MacPro 3,1 with R9 280 (PC vBIOS => no injection) => wakes perfectly fine

- MacPro 3,1 with R9 280 (Mac EFI => injection) => wakes perfectly fine

- Hackintosh with R9 280 (no injection) => doesn't wake

- Hackintosh with R9 280 (Injection = Hamachi [same as Mac EFI]) => black screen on boot

 

 

I've attached ioreg dumps, maybe it's useful to someone (will look into them later myself).

Hack_AutoInit.ioreg.txt

Mac_AutoInit.ioreg.txt

Mac_Inject.ioreg.txt

[jihu]

 

 

Edit: in my case I was able to fix the sleep issue by disabling ATI injection in clover and reverting to the default Radeon frame buffer. Sleep functions normally now. 

I can confirm the above method works for my (power colour) Radeon 5770 .  

[Matty___]

I can confirm the above method works for my (power colour) Radeon 5770 .  

When you run openGl viewer. How does it display your card? Does it say prototype as does with most 6xxx cards?

[ijhu]

When you run openGl viewer. How does it display your card? Does it say prototype as does with most 6xxx cards?

in OpenGL viewer, it displays Radeon 5770.  (Not Prototype)

[Slice]

I can confirm the above method works for my (power colour) Radeon 5770 .  

It is known that Radeon 5xxx works with sleep/wake. The problem is with 6000+

[ijhu]

It is known that Radeon 5xxx works with sleep/wake. The problem is with 6000+

I just want to share my info on this thread regarding sleep/wake only works if I remove ATI injection from clover. 

[KonstantinosNks]

I just want to share my info on this thread regarding sleep/wake only works if I remove ATI injection from clover. 

I tried the same way to my r9-270 and even without specified framebuffer (DSDT Injection) my card is recognized as AMD Radeon 7xxxx 2048mb (only dvi works)  -no Wake

With framebuffer injection on dsdt my card recognised as AMD r9 270 2048mb (HDMI/DVI Works) - no Wake

[Slice]

I just want to share my info on this thread regarding sleep/wake only works if I remove ATI injection from clover. 

Your info is useless and may led to wrong conclusion.

With Radeon 6000+ there are no ways to make awake. 

With or without injection.

With or without connector patches.

With or without DSDT edit.

With Clover or with Chameleon.

We checked numerous ideas.

 

I can accept ideas about making new kext. With what?

About binary patch of AMD kexts. There was some attempts but no success.

About SMC written into SMRAM.  

[Mikland]

I don't really get why people still make update of El Capitan, it was said already, you should do New Clean Install, other way, plenty of issues,,including, kernel panic that shutdown your Mac,,when is working normally,,after all,,i've trie all versions of El Capitan and i still dislike it,,even if apple says is faster, disk utility is a real {censored},,once you have 2 partitions and you want to repartiton your HD or SSD, making only 1 partition, you find a death end, no way to do it,,,

ohh,, i forgot,, is because of thirt party stuff,,for information,,lol

[Monq]

I don't really get why people still make update of El Capitan, it was said already, you should do New Clean Install, other way, plenty of issues,,including, kernel panic that shutdown your Mac,,when is working normally,,after all,,i've trie all versions of El Capitan and i still dislike it,,even if apple says is faster, disk utility is a real {censored},,once you have 2 partitions and you want to repartiton your HD or SSD, making only 1 partition, you find a death end, no way to do it,,,

People want to use the last version, especially taking into account how Yosemite was a bit of a disappointment.