Jump to content


  • Content Count

  • Joined

  • Last visited

About jafd

  • Rank
    InsanelyMac Protégé
  1. jafd

    WhatEverGreen Support Topic

    I have a rather particular use case here. Since QEMU 4.0 (which is just about to be released) and Linux 5.0, virtualized Intel GPUs can now have EDID information, which hopefully removes the last block on the way to have a virtualized macOS with full graphics acceleration on laptops that support this (including Haswell and above Macbooks that run Linux as host, yes!). At least, that's what some topics on Reddit suggested. However, when I tried to launch a QEMU VM with my virtualized Kaby Lake GPU, it didn't work. In fact, the GPU driver crashed. The crash log is linked below. (This is when firmware loading is skipped; enabling firmware results in a boot loop.) From what I have read, the Broadwell drivers had to be patched back in the day to work at all. I guess Kaby Lake/Skylake drivers need a similar patch to, maybe, omit some initialization not needed for a vGPU. I wonder if WhateverGreen could be potentially helpful in doing such patching (I know the actual patches might not be there, but Clover patching of GPU drivers is sort of dead and not considered anymore, thus writing here). Crash log: https://pastebin.com/0ewz5UTs Also might be of interest: the IGPU section is not in the DSDT at all (for obvious reasons: the Q35 definition doesn't have an iGPU). The adapter is recognized by the OS nevertheless. I'm not sure if I need to start from this and copy bits over from the host's DSDT (if that could ever help; a virtual machine doesn't have the same motherboard, and for sure doesn't have the same sort of an embedded controller). What other information should I provide to help you help me?
  2. The options are: 1) you flash a modded BIOS so you can install any WiFi controller, not just those in Lenovo's whitelist 2) you get a macOS-compatible wireless card and swap Intel for it
  3. Nope. Rechecked — X220 and T420 with the same CPU have identical sound chipsets. Norwegian names of inputs and outputs were caused by the vanilla AppleHDA.kext being botched somewhere down the road (I'm enrolled in beta testing, so might be a broken public beta), and Resources/ directory was very incomplete (although the signature was OK from kextutil). I transferred a vanilla one from another laptop and it works okay. Thanks for notes, though. I also have a Thinkpad W510 which is different.
  4. Found it, a broken symlink was the problem. But why are all my input/output names in a Scandinavian language? (Also, it won't switch to speakers/internal mic and is locked to headphones/line input)(seems like a hardware problem)
  5. Can you please list what patches did you use in your DSDT? I'm running your pack successfully on a T420, but on a X220 which has the same CPU the sound won't work. Here's what I see in System Profiler: Screen shot As you can see, there's no "Built-in input" and "Built-in output" which should be normally present under Devices. Also, I'm getting a couple of sound assertions: Sound assertion in AppleHDAEngine at line 17919 Sound assertion in AppleHDADriver at line 2896 Sound assertion in AppleHDADriver at line 201 Sound assertion in AppleHDADriver at line 3575 Sound assertion in AppleHDADriver at line 1084 I'm wondering how to right this wrong...Also, everything else works — it's only sound that won't.
  6. When using the VoodooPS2Controller from your bundle, all of a sudden I have lost my trackpoint and trackpoint buttons. (Note that vanilla VoodooPS2Controller works.) It hurts me because I use the trackpoint and its buttons more than the touchpad. Where do I look to fix it? Upd: never mind, got it back mysteriously after restart. However, it seems like I can lose it sporadically...
  7. Dear threadstarter, I, for one, would buy you a beer every time we meet if you publish some guidelines on porting Linux networking drivers to OSX. The information is rather scarce on the topic of writing NIC drivers at all (there is a nice thread on ArsTechnica, but it applies mostly to version as ancient as 10.2), and documenting roadbumps and mappings from Linux infrastructure to Darwin would be an immense help to those who would like to immerse into writing their own drivers. There is documentation and reading the source, of course, but it means bumping into the same roadblocks everyone else did. (My personal wish is to port somewhere along the line the cfg80211 stack and its drivers; they don't need to be Wi-fi proper, but I'd like them to work at all)
  8. I think your machine is overheating. Do you use, by the way, native AppleIntelCPUPowerManagement.kext? OS X forces itself to idle when it thinks it runs too hot.
  9. Could it be the "idle" process? ;-) On a side note, a communication improvement hint: kernels are those software things, like vmlinuz, ntoskrnl.exe or mach_kernel. Those hardware things in processors are called "cores". ;-) You must be speaking Russian or some similar language.
  10. Sorry if I look rude, but the 10.6.5 kexts sorta don't work. (Acer Aspire One AOA150, 27AE) That means: Installed, repaired permissions, rebooted, had only generic IONDRVSupport and no QE/CI. Rebooted with -f, same thing. Checked with kextutil, it says they are OK, kextstat tells they are loaded. Now to the fun part. Rebooted with -s, loaded both kexts manually, Ctrl-D... And it works. Rebooted, and they don't work. Again. Rebooted with -s, loaded both kexts manually, Ctrl-D... And it works. Well, I was desperate to try to somehow automate the process. So I slapped together a happy launchd job: <?xml version='1.0' encoding='UTF-8'?> <!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd" > <plist version='1.0'> <dict> <key>Label</key><string>org.jafd.intelfb</string> <key>ProgramArguments</key> <array> <string>/sbin/kextload</string> <string>-v</string> <string>/System/Library/Extensions/IOGraphicsFamily.kext</string> <string>/System/Library/Extensions/AppleIntelIntegratedFramebuffer.kext</string> <string>/System/Library/Extensions/AppleIntelGMA950.kext</string> </array> <key>Debug</key><true/> <key>Disabled</key><false/> <key>KeepAlive</key><false/> <key>RunAtLoad</key><true/> <key>OnDemand</key><false/> <key>LaunchOnlyOnce</key><true/> </dict> </plist> But despite all my efforts to make it run as early as possible, I have WindowServer starting up before it kicks in, so that the system is unusable (it has boot screen and a mouse cursor on it; ps ax in a remote shell tells me that everything is loaded, but probably drawn on another framebuffer which I cannot see). To sum it up: impossible to load the kexts automatically, but possible to do that manually; impossible to force-load them in a given timeframe. Maybe someone will point me to a better solution? Ehrm. sudo rm -fR /System/Library/Caches/com.apple.kext.caches/* did the job. (Be careful not to type any extraneous spaces if you don't want to have your system screwed badly.) I think that this directory watching the new kextd sports is far less reliable than simple removal of the cache file we've had back in Leopard days. At least, there were a lot less false negatives.
  11. One moment. I have installed LegacyAGPM.kext and it loads ok, however there's no difference in battery life estimations, as far as I can see. Two hours and that's pretty much all. Does it need DSDT tailored, or am I missing something? I'm on the stock battery. Just two bloody hours. I don't want a Macbook Air, because flash storage, in my very humble opinion, sucks.
  12. jafd

    10 Things Linux Does Better Than Windows

    I have a swap file on my ext4 partition. Really makes no difference (and I usually try to not swap processes out). The only problem with this is that making hibernation (s2disk) work is, say, far from obvious (compared to usual hibernation to a swap partition, which works out of the box). Yes, and it's usually clean-wipe, or voodoo registry editing. Guess that none is user friendly: the first one, because none of average Joe Users does backups, and the second one, because editing the registry is far from intuitive and documented. According to far too many MSCEs out there (I have some reliable sources in that field), easy to troubleshoot things are easy and hard ones are impossible. Ask London Stock Exchange, they switched over for a reason, I guess. From which I can only conclude that you do quite a poor job at maintaining all three of them... When I use Windows, even without an antivirus package, I just somehow manage to not infect it with malware (though the last time when I did it profesionally was in 2008). I still don't know what I did wrong. I'm telling that installing from repositories is far less troublesome than googling and point-and-click installing. Ever heard about the DLL hell? There's a lot of respectable software package which carry redistributable libraries (for compatibility reasons, of course) which they install in some system-wide location. Guess what happens if you install two packages each carrying different versions of the same redist, each tied to its particular version. Right, you find your system screwed. One piece of software works but another is broken. And you have no certainty in DLL load order, no LD_LIBRARY_PATH, no chroot, no nothing to prevent the screwup — except maybe a virtual machine which is an overkill most of the time. Guess what, even googling software up for windows is not so trouble free. That being because if the software in question is Windows-native/Windows-only/proprietary, I have trouble tracking down the genuine homepage — the first page of search results is invariably polluted with softpedia/cnet/you-name-it-software-portals which I simply don't trust enough to get software from. I think of that in the Slashdot way. There is effectively no censorship, you just set your threshold to -1 and view everything uncut. If one is too lazy to drop down a select box and then push a button, calling censorship instead, just let him die slowly, his laziness makes him really deserve that.
  13. jafd

    10 Things Linux Does Better Than Windows

    I have a single-partition Linux setup. What in the world am I doing wrong? This should be modded -1 Troll, as the specifics are not mentioned. I can as well say that in Linux customizations rarely differ in their difficulty, but in the Windows, easy things are irrelevant to me (I don't give a cr-p about wallpapers 'cause I don't see them most of the time) and relevant things are next to impossible (say, multiple workspaces, custom daemons as services and so on). Except that if it spills some indecipherable hexadecimal {censored} on you out of nowhere and bluescreens itself, you go routinely to clean-wipe and reinstall. For all 17 years. My 9 years of Linux usage experience taught me to troubleshoot to the actual root cause and fixing it without reinstalling and losing data, and without black voodoo magic. As in, it manages to crash more often? Other than things that make it malfunction and which plague users 90% of their time? The only thing Linux is missing from OS X in its current state, is damn desktop consistency. Single GUI, single behavior, a single set of shortcuts which works everywhere. Other than that, it's the same to me except Linux is much more administration-friendly. In Linux, as well as OS X, you have both ways out of the box, right here at your fingertips. Linux: apt-get install yourpackage If no package, find it on sf.net, wget, untar, configure, make, make install. Several commands and you're done, can be made into one for greater pleasure. While the machine is busy with its thing you can make some decent coffee. Windows: no package management at all, so go to the step 2. Find the software installer, point and click, point, click, click, click, click, click, click, reboot, click, click, click, OH MY GOD WHAT IS IT THAT MY PC BECAME SO SLOW????!!! Worst of all, you cannot make a coffee, because you must watch for stupid prompts to click "OK" and "Next". Because the machine is usually obedient even to the most foolish requests. Machines must do, humans must think. At least I can get my work with Linux's command line. Command line in Windows is reminiscent to me of 80s: no saveable history, no sensible scripting, lame autocompletion, no job control, lame redirection. Those people haven't ever tried to work without a keyboard, with mouse alone. And without middle button to select-paste.
  14. Just a quick note, “Invalid frame pointer” does not have anything to do with framebuffer, graphics or anything like that. It has to do with the way C (or C++) functions get called and, to be specific, with the way arguments are passed to them. Just in case someone reads that and suggests that video drivers must be the culprit.
  15. jafd

    Acer Aspire One A150

    There are some problems I'm having with my Acer Aspire One A150 and Snow Leopard. The BIOS version I'm using is 3309. The DSDT was pulled, I believe, from some install guide on here, and now I wonder if it is broken. <rant>There are many “advices” on this forum how to “fix” the DSDT, in one way or another, but these involve some black magic voodoo nearly all the time. The same situation I observe on mobile forums where members offer to unlock handsets, but while the mobile guys are doing it for a price, it seems like on the OSX86 scene people either got their DSDT problems fixed by sheer luck while randomly poking around, or that some of them have more profound ACPI knowledge but don't want to share it because of their hubris or whatnot.</rant> What I basically want is some pointers, apart from the specifications found on http://acpi.info/ which I could use to repair my DSDT, if no one has done it before me which I think is the case. This is because I want to 1) have knowledge of how the damn thing works instead of poking around blind, 2) share my findings with the general public, again, knowing what I'm talking about. This way the knowledge is not only useful, it will also be reusable which is what we all want. Now to the specifics. The brightness problem. While Fn+Left and Right Arrows do adjust screen brightness, the display preferences pane has a brightness slider but it doesn't work. I would like it to work, because otherwise I cannot calibrate my display the way it should be (OSX thinks it adjusts the brightness while it doesn't). Other than that, I believe that brightness could be controlled more smoothly than when using the BIOS inbuild table with 5 rigid steps or so, when the least bright isn't dim enough for me. The sleep problem. The machine just won't go to sleep. It shuts off the display, but then hangs and I can't say why (because the displays shut off, and should there be kernel panics, I can't see them). The power LED is green instead of blinking orange. The hibernation problem. Actually the same as sleep. The machine shuts displays off and hangs, the sleepimage is never dumped to disk. Attached is the DSDT my machine currently runs with, if you have some hints at where I should look, I would appreciate that, and kextstat output, should it be that some KEXT is known to cause any of the abovementioned problems. And I would like you all to please refrain from “Me too” kind of posts which add nothing to the discussions, nor are welcome the “bump” posts with nothing useful to say. If the thread dies off, it's just that “nobody knows” and that is a valid answer. Then, all I'm left with is the ACPI specification which gives no hints as to what bugs a vendor (Acer & Insyde in this case) might incorporate. I would like to add that Linux has no problems with both sleep and hibernation, it might be due to the fact that kernel actually knows how to power manage each peripheral and is not reliant on DSDT so much as OSX is. DSDT.aml.gz kextstat.txt.gz