Jump to content

No Wake From Sleep - Discussion / Fixes thread


219 posts in this topic

Recommended Posts

Ok, another theory, from grepping thru my IOReg -l output....

 

The onboard GMA950 I have is in there, matching as 'display'. I tried removing all GMA-related kexts but its still in the ioreg.

 

Could the problem be that the onboard video is getting the wakeup call instead of the PCI-E card?

 

Does anyone else with the problem NOT have onboard video?

 

I will remove my card and see what happens with onboard, used to work in .6 ...

 

 

-Damn, you could be onto something - sleep works fine w/ 950 - and there was me feeling inadequate with an EIST-less CPU. Bloody 950, oh how I hate thee... a pox on all your (4) pipelines.

Edited by consolation

lol... interesting that it works for you. my current install it doesnt. then again, i installed with myzar 10.4.5, then jas 10.4.6 then apple 10.4.8 combo then mifki kernel + paulicat smbios, then semthex kernel and now paulicat sleep kernel ;)

 

maybe things are a little messed up in there...

i'm also going to try a few things suggested by carpao, like killing the loginwindow or systemuiserver from the ssh session.

Munky what I did was this:

 

- IDE Hard Drive

- Integrated and PCIe 7900 GT CO eVGA Nvidia card

- D945GNTL Motherboard

 

Sleep worked for me in that configuration I'm gonna do more tests and see why It worked for me on an IDE hard drive.

 

By the way Safe Sleep (Hibernate) doesn't works on hackintoshes I don't know why probably is something with EFI.

Munky what I did was this:

 

- IDE Hard Drive

- Integrated and PCIe 7900 GT CO eVGA Nvidia card

- D945GNTL Motherboard

 

Sleep worked for me in that configuration I'm gonna do more tests and see why It worked for me on an IDE hard drive.

 

By the way Safe Sleep (Hibernate) doesn't works on hackintoshes I don't know why probably is something with EFI.

 

safe sleep works with the 8.4.1 kernel :) not hackintosh related.

i'm also going to try a few things suggested by carpao, like killing the loginwindow or systemuiserver from the ssh session.

 

remember:

killing loginwindow has immediate effect but corresponds to make a signoff, so you are prompted again for login and your previous session is destroyed (open applications killed)

 

killing systemuiserver has (if any, ie.e it didnt work always) not immediate effect... in my casae some time after more or less 60 seconds it presentes the login to resume after the sleep...

Sleep works on my hackintosh, takes about 30 seconds for it to start responding again, mouse needs re-plugging though >:thumbsdown_anim:

Now that *is* interesting. I wonder whats missing / broken from the 8.8.1 kernel which breaks safe sleep?

 

Oh, the results of tonights experiments: nothing worked, basically. The SetDisplay thing hung when looking for the display mode, and killing various things had no discernible effect - nothing magically reawakened the video.

 

I also tried to compile libx86 and a program I found called boot-radeon.c which was written by someone who had the same problem with a radeon card under linux (ie no video on s3 resume), but i couldnt get the libx86 thing to compile. SourceForge reckons its for all UNIXes, including BSD, but i suspect im missing linux-specific headers or something.

 

I have been doing quite a lot of digging into the matter, and plenty folks in the linux world have similar problems, esp with radeon cards.

 

I will dump here some links which may be of interest / use. Earlier today I found a link with mention of a specific VESA call which should reinit the video bios at a low level, but i've lost it (i know, im a fool). I will try to find that again and post it.

 

http://developer.osdl.org/dev/robustmutexe...power/video.txt

hey munky,

 

I have 10.4.7 installed on a usb drive.. Just tried the sleep with that install..

 

works perfectly.. whole things shuts off, power down, hibernates... when i press the power button,

 

it resumes the system and wakes up...

 

Is there any log that would show what is happening exatly when the computer is hibernating..

 

i played around with the console, couldnt find anything..

 

Let me know what I can do to help.

It had everything working for me with the old kernel. But I had to tinker with 10.4.5 (8.4.1 kernel) to make it work. OS 10.4.3 worked right off the bat. everything worked. Sleep, safe sleep and all that. I can't remember if it was on an IDE or an SATA but I had an SATA drive in at the time. I think it was installed on an SATA drive (OSX) so that SATA talk seems silly. I have an x1600xt with sleep working too.

 

KIKO do you have sleep working on the new kernel?

 

My new hackintosh doesn't even support Restart.

How do I fix that? I have an Intel D945GCL board. The board I had all that stuff working on before was a D945PSN.

thers some light at the end of the natit tunnel....

 

fresh install of jas amd/intell sse3/sse2 v2 and my x1300pro.

 

nothing selected in video options so im running latest apple atidrivers.

added my new revision of NatitX1300 (not yet released).

now every single ati driver is intitated corectly, including Caretta.kext in ATINDRV.kext.

 

now hit the sleep button in the apple menu...

sleep for a minute. click mouse button and back to desktop i am :thumbsup_anim:

 

Semthex beta 9 not paulicat kernel

intel 915gux mobo and Asus eax1300PRO graphics.

 

-DiaboliK

plenty folks in the linux world have similar problems, esp with radeon cards.

I haven't had too many problems with sleep on an x800xt, only a couple of distros ever gave me trouble :thumbsup_anim: I don't ever use hibernate, so I cant comment on that.

 

On my lappy with 950, I have problems with sleep with just about any distro that will install. Its 50/50 if it will come back on or not.

hello again munky!

 

i got another sleep and waker!

Sapphire X1600PRO 512mb Pci-e

used an NatitX1600 based on the new unrealesed X1300 that sleeps and wakes. :D

Still no luck on pure dvi output or extended monitors tho.... ^_^ (still working on it)

 

Screenie:

post-14143-1173613114_thumb.jpg

 

same jas amd/intel sse2/sse3 v2 disk. semthex kernel.

 

Looks like its time to make munky a new NatitX1800.kext! :)

 

-DiaboliK

DiaboliK has made some nice progress and has posted a sleepable X1300 Natit kext on his helpful files site..

 

He's fixing some bugs for the X1600 version, and I am currently testing an X1800 version (first build didnt work, but we're investigating).

DiaboliK's theory is that if you make sure all the kexts have their dependencies satisfied, sleep starts to work. He's had success with an X1300 and X1600 so far.

 

The X1800 kext he built for me didnt work, but it did make my machine load ATITVOut.kext, which I hadnt seen before. :)

 

Fingers crossed...

On the topic of making sure all kexts have their dependencies satisfied, is there anyway to do that without recompiling extensions?

 

Most of my extensions are looking for various versions of IOKit or Libkern other than the ones actually installed on my system, Apple System profiler shows tns of unsatisfied dependencies.

 

Back to the main topic:

 

After much tweaking my computer now will sleep fine, and will come back from sleep upon tapping the keyboard or moving the mouse. The system is not locked up at all after sleeping, however all USB devices stop working. If you unplug anything (for example the mouse) it will seem to not be powered when plugging it back in (doesn't light up.)

 

I can tell that the OS itself is fine, because when it looses my bluetooth connection (D-Link USB adapter) it freaks out and throws up an error message on the display.

 

This message has always popped up when booting with -v:

 

OHCI: USB will be unloaded across system sleep.

 

My BIOS is configured to S3 level sleep, my hardware is below, and I have the following kernel as revealed by uname -a:

 

Darwin Crackintosh.local 8.8.1 Darwin Kernel Version 8.8.1: Sat Dec 9 22:18:27 AZOT 2006; semthex:/nebukadnezar/BUILD/obj/RELEASE_I386 i386 i386

DiaboliKs new X1800 kext works as well as ever, but still no wake from sleep. I checked thru some of the system logs, and found these:

 

Mar 12 20:20:16 Zaphod kernel[0]: Started CPU 01

Mar 12 20:20:16 Zaphod kernel[0]: System Wake

 

 

Mar 12 20:20:26 Zaphod kernel[0]: in_delmulti - ignorning invalid inm (0x35d63d)

 

 

Mar 12 20:20:40 Zaphod diskarbitrationd[68]: SystemUIServer [253]:24835 not responding.

 

 

Mar 12 20:20:41 Zaphod kernel[0]: jnl: journal start/end pointers reset! (jnl 0x24e6c98; s 0x1a1800 e 0x1a1800)

 

 

I tried killing and restarting SystemUIServer, which made no difference other than i could no longer connect via VNC. (I could before killing it, but i just got a black window).

 

Any ideas?

A more exact readout of the error (this isn't saved in a log I can find, had to copy it down by hand)

 

USBF: 34.486 AppleUSBOHCI[Ox1814000]::start OHCI controller will be unloaded across sleep

USB caused wake event (EHCI)

 

The numbers in red are different every time the system boots.

 

Edit: output of system.log when sleeping as follows:

 

Mar 12 14:04:15 Crackintosh launchd: Server 0 in bootstrap 1103 uid 0: "/usr/sbin/lookupd"[150]: exited abnormally: Hangup
Mar 12 14:04:15 Crackintosh configd[55]: posting notification com.apple.system.config.network_change
Mar 12 14:04:15 Crackintosh lookupd[259]: lookupd (version 369.6) starting - Mon Mar 12 14:04:15 2007
Mar 12 14:04:18 Crackintosh kernel[0]: IOBluetoothHCIController::terminateWL .. done
Mar 12 14:04:20 Crackintosh kernel[0]: System Doze
Mar 12 14:04:21 Crackintosh kernel[0]: IOBluetoothHCIController::start Idle Timer Stopped
Mar 12 14:04:21 Crackintosh kernel[0]: Registering For 802.11 Events
Mar 12 14:04:21 Crackintosh kernel[0]: [HCIController][setupHardware] AFH Is Supported
Mar 12 14:04:21 Crackintosh configd[55]: posting notification com.apple.system.config.network_change
Mar 12 14:04:25 Crackintosh kernel[0]: IOBluetoothHCIController::restartShutdownWL this is a wake from sleep
Mar 12 14:04:29 Crackintosh kernel[0]: AppleDP83816Ethernet: phyWaitForAutoNegotiation TIMEOUT
Mar 12 14:04:32 Crackintosh lookupd[263]: lookupd (version 369.6) starting - Mon Mar 12 14:04:32 2007
Mar 12 14:04:32 Crackintosh pppd[265]: pppd 2.4.2 (Apple version 233-13) started by root, uid 0
Mar 12 14:04:32 Crackintosh launchd: Server 4a3b in bootstrap 1103 uid 0: "/usr/sbin/lookupd"[263]: exited abnormally: Hangup
Mar 12 14:04:32 Crackintosh pppd[265]: local  IP address 10.64.64.64
Mar 12 14:04:32 Crackintosh pppd[265]: remote IP address 10.112.112.112
Mar 12 14:04:32 Crackintosh configd[55]: posting notification com.apple.system.config.network_change
Mar 12 14:04:32 Crackintosh lookupd[266]: lookupd (version 369.6) starting - Mon Mar 12 14:04:32 2007
Mar 12 14:04:32 Crackintosh configd[55]: posting notification com.apple.system.config.network_change
Mar 12 14:04:32 Crackintosh lookupd[267]: lookupd (version 369.6) starting - Mon Mar 12 14:04:32 2007
Mar 12 14:05:12 Crackintosh pppd[265]: Failed to open /dev/cu.Bluetooth-Modem: Resource busy

Edited by Spart

Spart - my machine's log shows a whole heap of USB problems on wake. It keeps saying it will have to rebuild the root hub, which it eventually does. My USB devices would *appear* to be fine after wake (my mouse has lit up). Though I cant cd to /Volumes/cPod (which is my partner's iPod nano)... interesting....

 

EDIT: ive just ripped out all my USB devices and am going to boot. i have a ps2 keyboard and i can VNC in from my MBP... :D

 

EDIT2: lol : http://guides.macrumors.com/SystemUIServer

Right, some more findings:

 

1) Mar 12 22:24:03 Zaphod kernel[0]: ATIRadeon::wait_for_rb_space: Overflowed block waiting for FIFO space. Have 0, need 5. RBBM_STATUS 0x90010140

 

Followed by a big ASIC Hang Log. (256 entries of 0xdeadbeef, followed by some real data - more info in attachment)

 

2) After this happened, I tried to reboot, and my machine does a thing which it does from time to time - it boots with no video, and the GPU fan on 100%. The difference is that this time I VNCd in. System Profiler showed Unknown ATI and no displays connected!? WTF? *I HADNT CHANGED ANYTHING*. This *must* be some recovery mode the GPU is going into after the hang.

 

So i'm now convinced that the GPU is hanging on S3 resume, and that this is the cause of my problems. I'm going to do a clean JaS 10.4.8 install without selecting any video patches, then apply DiaboliK's new X1800 kext. (This is the install route he used).

x1800crash.txt

×
×
  • Create New...