Jump to content

d2a

Members
  • Content Count

    29
  • Joined

  • Last visited

  1. awesome - thanks for the clarification and your time.
  2. So just so I understand: You have updated your Mobo firmware to f8? You successfully installed Clover UEFI with this board? That's great news if so - gives me hope I can get there. I'll give it another go tomorrow. Thanks for the quick reply
  3. So I'm finally joining the grown-ups and trying to set up a new(ish) hack. In the past I've used legacy Clover with my old Gigabyte board and thought I knew what I was doing. But I've been trying to get a UEFI Clover-based install to work and it's going nowhere. My new system (listed as SysB in sig) is a Gigabyte Z97X-UD3H-BK rev1.1 with firmware F6, an i7 4790K CPU and 4x4GB Corsair 1866MHz. I have the firmware settings set as per various guides and thought this should be a pretty painless transition... I have successfully run the installer for El Cap several times, but can't get the machine to boot reliably - it seems to change behaviour each boot. Sometimes boot hangs at the Gigabyte logo after choosing boot volume via F12, sometimes it boots to clover but displays no bootable drives, sometimes displays ElCap SSD in Clover but wont boot from it with kernel cache error, and occasionally it will boot all the way. What I'd like to know is: Should I update the mobo firmware to the latest version? This board has two more recent Bios software downloads, but they wont allow me to drop back to the version currently installed. I'm nervous to update unless this is likely to be the culprit. Is this just a UEFI issue? If the simple answer is just go legacy mode, maybe I should just avoid the issue... FInally, does anyone else have a working clover install on this board they could share with me? I think I've got it set up as it should be but the strange inconsistent boot behaviour makes me wonder... Thanks
  4. Cool - thanks again for the pointer to that thread. I'll pick it up with them.
  5. So after hacking around the script some more, it seems to have issues on my system because: the first test it uses to determine the filesystem file -sb /dev/rdisk0s1 returns something like this on my drive x86 boot sector, code offset 0x63, OEM-ID "BSD 4.4", Media descriptor 0xf8, heads 32, hidden sectors 40, sectors 409600 (volumes > 32 MB) , FAT (32 bit), sectors/FAT 3152, reserved3 0x800000, serial number 0x30691edf, label: "EFI " which does not match any of the cases in the test. It expects "DOS/MBR boot sector," or "DOS/MBR boot sector " or "Macintosh HFS Extend" Secondly, once it fails on this test, it goes on to attempt to mount the partition using msdos hfs exfat in turn. but further down the function, it relies on the ${RETURN_VAL} == 0 test to confirm successfully getting the mount point, but this is reset by each loop of the mount attempts. so if the FS is msdos as in my case, the RETURN_VAL is set to 0, then reset to 1 by the hfs and exfat mount attempts.
  6. So a quick test of the script run under terminal. I commented out a few lines here and there, but the basic flow seems to be: first run it reports: Found ESP on disk0s1**** disk0s1 Get mount point failed - all EFI partitions on attached drives are mounted and visible on Desktop. .metadata_never_index file is written to / if run a second time (with EFI partitions still mounted) it reports: Found ESP on disk0s1 with /Volumes/ESP-01/EFI/CLOVER. This correctly checks the EFI, spots that there's already a never index file and ends. So perhaps there's an issue in not mounting the partitions fast enough before checking for the mount point. But there's certainly a logic problem, in that if no EFI mount point is found, it defaults to root.
  7. OK - you were right about the Clover RC script - 80.save_nvram_plist.local - I commented out the contents of the function dealing with .metadata_never_index, removed the .metadata_never_index from the /, enabled spotlight, and rebooted. File was not replaced, and indexing survived the reboot. Phew! SO next question is why the script incorrectly identifies all the partitions of my drive as EFI and inserts the never_index file? I'll take a look over the script and see what i can deduce, but thanks for taking the time and for the pointer to the shutdown script. If I learn more I'll post back here and possibly let the clover folks know too.
  8. A ha - could be on to something. I'm on Clover 4035 according to the log. I have an EFI partition on the system drive, which does have the .metadata_never_index file in place. In fact, all partitions on this drive (recovery partition, EFI, OS) have a never index file. Is it possible i've misconfigured Clover? maybe i'll try a clean config file...
  9. Nope - can't enable spotlight under a new user either. Added System (boot OS partition) to privacy list in Spotlight, close prefs, and reopen and it's marked as private, reboot and it's gone. Can't help feeling this may be a deeper partition level problem. the drive initially refused to show up on the desktop (unlike all other drives and partitions) or in the sidebar of the finder, after update. Could still be browsed fine, but i had to chflag the volume to get it to show up. Perhaps there's a flag or property in the partition table that's messing this up? The .metadata_never_index file is added at logoff as far as i can tell, rather than startup...
  10. Thanks. sure - i think it might be a permissions issue but i've repaired them a few times, both from System and using the Recovery partition, without improvement. i've tried disabling, enabling, adding to privacy in spotlight and then removing, without luck. i've seen similar problems posted elsewhere so i don't believe it's limited to osx86 One thought: I had to run the yosemite installer app using the root user as it failed for the normal user. perhaps this has corrupted the permissions? any thoughts on how to diagnose that?
  11. So i'm happily chugging along with Yosemite - I know, it's old, so am I - and all's well. Basically a clover-based install on an old GA-P35C board, that was originally SL, then Mav, and finally updated to Yosemite. No clean installs - all updates. But: When I log out or reboot, the startup partition becomes unindexable and search is disabled. sudo mdutil -s / /: Indexing and searching disabled. If i renable search: sudo mdutil -i on / The old index becomes useable for search, but the drive is still not indexable. I have to remove the metadata_never_index file from the root of the partition: sudo rm /.metadata_never_index OK - all good right? But as soon as I log off or reboot, the partition has the never index file back in place and searching is disabled. So what have I got to do to the drive to get it back? Is there a secret boot flag (clover?) or disk partition flag that's been accidentally set? I did have issues initially getting the drive to show up in the finder, and found it had a hidden flag applied - could that be related? Really painful and I'm sure it's a corner case but if anyone out there has any inspiration I'd be grateful - thanks.
  12. d2a

    DSDT fixes for Gigabyte boards

    before i start a big thank you to all who've contributed so much to the gigabyte users in this thread - this stuff is rocket science and it's great to watch it happen. i've recently begun to sort out my dsdt after basically cludging along with efi strings, extra kexts and so on. my machine status is now: GA-P35C-DS3R rev.2.0 - bios f10 - E6750 - nvidia 7800GT 256MB - SATA HD x 4 - Pioneeer IDE DVDWriter - TI Firewire PCI SL 10.6.4 Vanilla DSDT fixed for 889a audio, vanilla speedstep and nvidia 7800GT graphics. EDIT1: Added dsdt fix for native onboard (Realtek 8111b) NIC so I could remove IONetworkingFamily and Realtek kexts from /Extra but although mostly functional, it does not seem to enable bonjour promiscuous mode, so reverted to kext approach for now. /Cham/Extra/Extensions/ contains: fakesmc.kext IONetworkingFamily.kext - patched to enable bonjour JMicron36xATA.kext - needed for IDE DVD-writer support LegacyHDA.kext - needed for 889a audio RealtekR1000SL.kext - needed to enable bonjour MacPro3,1 in SMBIOS Machine shuts down, restarts and sleeps. Will not wake from sleep with mouse (MS optical via Apple USB keyboard hub) or keyboard (white Apple USB) but mouse does have power (optical light still active). Wakes perfectly with powerbutton if you're quick, but if displays (twin CRTs via VGA to DVI adapters) have gone into standby mode (after 20 secs. or so), on wake mouse movement is reversed and display resolution is wrong. sleep continues to "work" in this way but can only get resolution and mouse back with restart. Console messages include: 05/07/2010 09:37:04 kernel The USB device Hub in Apple Extended USB Keyboard (Port 2 of Hub at 0x3a000000) may have caused a wake by issuing a remote wakeup 05/07/2010 09:37:04 kernel The USB device Microsoft Optical Mouse with Tilt Wheel (Port 1 of Hub at 0x3a200000) may have caused a wake by issuing a remote wakeup 05/07/2010 09:37:04 kernel The USB device Apple Extended USB Keyboard (Port 3 of Hub at 0x3a200000) may have caused a wake by issuing a remote wakeup 05/07/2010 09:43:13 Finder[199] _NXGetScreenRect: error getting display bounds (1001) 05/07/2010 09:43:13 Finder[199] kCGErrorIllegalArgument: CGSGetDisplayBounds (display 5b81c5c0) 05/07/2010 09:45:07 kernel IOHIDSystem::relativePointerEventGated: Unwanted Direction Change Y: oldDy=1 dy=-73 this may just be my GPU's problem but it'd be good to hear any thoughts... EDIT2: Just did a few experiments with swapping cables around etc and now have (fingers crossed) working sleep after display standby. main point seemed to be that one CRT did not receive the wake signal and remained in standby, but after a round of cable swaps and finally a few Detect Displays clicks to wake it up, it is now behaving normally after wake. still no wake from mouse or keyboard though... these are my current DSDT and IORegDump files ioregdump.txt.zip dsdt.zip
  13. Hi there - i have the same bluetooth adapter and have so far failed to get it working. can you share in more detail how you got it working?
  14. Just echo everyone else's thoughts - great guide! my specs are in my sig - was running kalyway 10.5 up to 10.5.3 since ~Jan 08 with a pretty hacked together set of kexts, but it worked ok, apart from sleep. recently got a bit antsy and decided to update to 10.5.6 and started to have problems. chameleon bootloader and using efi strings for video was great but video was glitchy and system started to feel very unstable. so followed guide v4.4, vanilla, with no step 3 as i was using 10.5.6 retail. at step 6 (add efi strings using efistudio) i used the most recent version with 7800GT strings in it, and added display, ethernet and HDEF strings to boot.plist. reboot and video is perfect, audio is detected in system profiler and working, time machine working, bonjour etc. restored my user via time machine and i was pretty much set. however: sleep doesn't work. displays turn off but when i resume screens return but system is unresponsive, mouse is jerky and clicks not registered. tried the checking "Restart automatically after a power failure" box with no change. anyone with same specs have this working? maybe a bios setting i haven't tried... scheduled power down works, but scheduled start up or wake does not. this did work on my old kalyway install. displays - i run dual CRTs via DVI to vga adapters. at startup the second display is often not brought out of standby and only powers up on boot into OSX. the displays aren't detected and default to 800x600. this was also an issue with the kalyway install and hitting detect displays usually fixes things fine, but now the primary display is always detected as VGA Display, unlike the secondary which has correct model and resolutions. Not a big deal but would be good to find a fix... possibly related to 3. machine is now shown as a P35C-DS3R and has an iMac icon - is there a way of getting it back to being a mac pro? i know it's purely cosmetic but it's in a nice aluminium case with two crts connected - it'll have an inferiority complex! any thoughts gratefully received - i love this box and it's been great to get it almost vanilla but the closer to perfection it gets the more the little things bug me
×