Jump to content
Welcome to InsanelyMac Forum

Register now to gain access to all of our features. Once registered and logged in, you will be able to contribute to this site by submitting your own content or replying to existing content. You'll be able to customize your profile, receive reputation points as a reward for submitting content, while also communicating with other members via your own private inbox, plus much more! This message will be removed once you have signed in.


  • Content count

  • Joined

  • Last visited

About pista7

  • Rank
    InsanelyMac Protégé

Profile Information

  • Gender
  1. xnu-1228 / Boot Arguments

    I'm afraid you won't be able to use the Debug object - the AppleACPIPlatform is apparently compiled without debug support. If you look at ACPICA sources (the reference ACPI implementation), the trick should be done by setting acpi_level=2 acpi_layer=0x80. However, this doesn't work with AppleACPIPlatform.
  2. 10.5.8 deep sleep problem?

    I have explained back when I've posted the first version of this extension, that all builds are tightly coupled to a particular version of the kernel (if you look in the Info.plist, you can see three lines with version number, either 9.8.0 or 10.0.0 near the end). Thanks to this, a mismatched kernel won't even be able to load the kext, which is good, because it would panic sooner or later. This one is for Snow Leopard (kernel version 10.0.0)
  3. 10.5.8 deep sleep problem?

    I'll compile the SleepEnabler.kext for 64-bit once I have SL running - it's not possible to build a 64-bit kext in Leopard. Maybe someone else with working Snow Leo could try to recompile it as 64-bit - attaching sources. source.zip
  4. 10.5.8 deep sleep problem?

    Hi, if anybody would like to try a version for Snow Leopard, check this out and let me know what happens - it's complete guesswork from SL's mach_kernel disassembly, so anything might happen (including it will just work . However, there may be more changes in SL's power management, so even if it works as in 10.5.8, it might not help it go to sleep. sleep_sl.zip
  5. 10.5.8 deep sleep problem?

    Technically, there are only two modifications required: - create a configuration/modify existing one to define KERNEL_VERSION as 10.0.0 - create a new folder Headers_10.0.0 (like the existing Headers_9.7.0 and Headers_9.8.0), and copy there the files pmCPU.h and cpu_topology.h from the 10.0.0 kernel However, the latter is a bit of a problem, since the SL kernel isn't outsourced yet - you'd have to use an available version, and guess: a) the PM_DISPATCH_VERSION constant (instant panic on mismatch) - probably best to start from 16 a move upwards (better stop around one thousand if it doesn't work )) the layout of the two structures (pmDispatch and pmCallbacks) - the best case would be if any fields were only appended to these and no fields were removed - in this case, it should be safe to simply add a random number of padding bytes at the end of both, and it should work I'll definitely look into it once I get SL up and running (it somehow insists on KP-ing all the time on boot, first on fsck_hfs, and when I suppressed it, it crashes on 'ReportCrash'... funny, isn't it? if you have any suggestions, throw them at me!)
  6. 10.5.8 deep sleep problem?

    S1 - it works as advertised power on suspend S3 - the part affected by SleepEnabler.kext is the proper go-to-sleep (i.e. without hang or reboot) - if you get as far as waking the system up again after sleep (even if only the fans start spinning), it has done it's job. There are tens or even hundreds hardware and configuration-specific issues that may prevent your system from waking properly, and I'm no expert on these, having built only a handful (and quite similar) systems. However, I'd suspect a bugged DSDT, or a problematic (not necessarily unfixable) GFX/Sound/LAN card to be the problem here.
  7. 10.5.8 deep sleep problem?

    It's normal for this message to appear when you use Disabler or a similar kext to suppress AppleIntelCPUPowerManagement. It has no direct relation to SleepEnabler. Which brings me to a fact, that many people here seem to attribute various positive and/or negative effects to SleepEnabler.kext. However, the kext really does nothing that could have such effects (for those who can read source code, just look at it and you'll see) - it simply provides 'fake' or 'empty' implementation for callbacks normally implemented by AppleIntelCPUPowerManagement - which is why the two can't live together, but they also never need to - if you have AICPM running, you definitely don't need SleepEnabler.kext and vice versa. Or, put differently - 10.5.8 kernel + SleepEnabler = roughly 10.5.7 kernel All other effects on hibernation/safe sleep/wakeup/wireless/display/Idon'tknowwhatelse can be only attributed to other changes in 10.5.8, or for example the fact that many people modified their DSDTs to be able to run AICPM before SleepEnabler has been written, or pretty much anything else in the universe except SleepEnabler.kext
  8. 10.5.8 deep sleep problem?

    Hi, if you could do a regular sleep without SleepEnabler, it means that you have either AppleIntelCPUPowerManagement running, or are not running 10.5.8 at all (your signature says 10.5.7) SleepEnabler is only for 10.5.8, and only does a single thing normally done by AICPM.kext, required for a piece of kernel code (new in 10.5.8) to pass during the go-to-sleep procedure. It should have no direct effect on deep sleep/various wakeup problems/anything else sometimes mentioned in this thread. However, there are definitely other changes in 10.5.8 (new AppleACPIPlatform among others), which can be responsible for such problems. Since these components, unlike the kernel itself, are not open source, there is no feasible way to determine what has been changed and how to fix it.
  9. 10.5.8 deep sleep problem?

    If you're talking about the SleepEnabler.kext, then definitely not - the kext only compensates for a modification in the 10.5.8 kernel. If the 'this idea' is something else, let me know One way SleepEnabler could affect power management is if you have AppleIntelCPUPowerManagement.kext running and working (I'm not sure if it works for anyone - it doesn't for me, so I don't really care ) - this is because if SleepEnabler.kext will replace AICPM.kext as the PM kext (there can be only one registered in the kernel) if it loads second. However, it would have a nasty side effect of crashing the system rather quickly, in most cases The other case could be an older version of VoodooPower.kext that would hook itself as the PM kext - this is just my speculation, I didn't investigate. The last version (1.2.3) doesn't do this. However, there is a declaration of the pmKextRegister function that is used to register the kext for PM in VoodooPower sources, though it's unused in the last version. That's why I guess it might have been used previously.
  10. 10.5.8 deep sleep problem?

    Errrr... I can see no reason why you'd use the SleepEnabler.kext if you have AppleIntelCPUPowerManagement running. The only thing SpeedEnabler.kext does is that it provides (empty) implementation for a callback required by the 9.8.0 kernel to sleep. If you have AppleIntelCPUPM running, it already provides a (probably correct) implementation of this callback (which is why it was used as a workaround for this issue in the first place). Running both causes them to fight for the callback registration, and since there is no check in the kernel preventing duplicate registration, the last kext to register wins - yielding an unstable system.
  11. 10.5.8 deep sleep problem?

    Ok, here's the autostarting version - hopefully I haven't broken anything (blind-coding again ) Make really really sure you test that it loads successfully before putting it to an Extensions directory, as it would most likely make your system unbootable if it panics or something (it gets loaded even before you get to the single-user prompt with -s, so in case it blows up, you'll have to boot from a DVD/different partition). Also, I can't say what would happen if both AppleIntelCPUPowerManagement.kext and this on would be loaded at the same time - most likely nothing, but better avoid trying this at home Make sure you have disabler.kext or a similar kext up and running. sleep.zip source.zip
  12. 10.5.8 deep sleep problem?

    No, it won't. It's built specifically for 10.5.8 (i.e. the 9.8.0 kernel), and it shouldn't even load on any other kernel version, as it would most likely KP immediately.
  13. 10.5.8 deep sleep problem?

    Yessssss!!!! Not only have I managed to revive my desktop (I had a great idea to turn HPET from 32-bit to 64-bit mode while testing SL, and it resulted in Leo being completely unusable , but sleep works for me now!!!! For those interested, there was a small bug in the kext I posted previously - I didn't let the kernel know that PM has been "initialized", so it still didn't use the provided callbacks. This version should work correctly (it works for me, I'm typing this on a freshly waked computer The sources are also attached (all 5 lines of them ) Update: Since I forgot that non-IOKit kexts are not loaded automatically at startup, it won't help you to move this kext to an Extensions folder - I will fix this tomorrow (hopefully) - for the time being, simply start it manually and as long as sleep works, you should be fine Just remember to reload the kext after reboot (or use launchd or whatever) Edit: downloads removed - the last version is here: http://www.insanelymac.com/forum/index.php...t&p=1224830
  14. 10.5.8 deep sleep problem?

    Visible KP? that's at least a bit better than hanging after the display is turned off Could you try to take a picture of the KP screen with debug turned on? (enabled with something like debug=0x144 I believe, in kernel arguments)
  15. 10.5.8 deep sleep problem?

    Ok, here it is. However, since I've managed to completely destabilize my desktop (running vanilla 10.5.8) yesterday, I've had to blind-code this on my Pentium-M notebook, on which I still run the voodoo 10.5.7 kernel, so it's completely untested (except it is loadable without the important part on 10.5.7 when built for 10.5.7 - the version attached is hopefully built to run on 10.5.8, and including the important part). Testing suggestions (lines with $ are supposed to be entered into Terminal.app with current directory set to wherever SleepEnabler.kext is extracted 1. extract the attached kext somewhere else than /Extra/Extensions or /S/L/E so you won't make your system unbootable in case something goes awfully wrong 2. $ sudo chown -R 0:0 SleepEnabler.kext 3. $ sync (just to be safe) 4. $ sudo kextload -t SleepEnabler.kext if the kext loads successfully, try if sleep works, and if it does, move it to a more permanent location using any tool/procedure you are accustomed to if it doesn't: 1. if the kextload fails, post here the error messages 2. if it panics, try to provide a readable screenshot/photo of the panic output (I believe you have to enable the panic output somehow - search the forums I'm too tired to look for it right now. happy testing (and, hopefully, sleeping...) Edit: download removed, the final (working) version is here: http://www.insanelymac.com/forum/index.php...t&p=1224830