Yet another Lenovo G50-70 install Mavericks thread

[cutterjohn]

If you're getting CMOS resets, make sure you're using config.plist/KernelAndKextPatches/AppleRTC=true

ah hah! That MIGHT be it! *

AND lo and behold a BIOS update from last week that *GASP* IF I leave UEFI boot mode ON I can turn OFF secure boot(in theory), however selecting DISABLED leaves the secure boot status line below still ENABLED!

BIOS is apparently from "insyde" whom I'd never heard of until I specifically checked in hwinfo. They must work super cheap v. AMI or Phoenix...

So now, I have a few more options to play with, so we'll see what happens.

* some of the BIOS funkiness seems to happen IF I go into setting, change some things, e.g. turn OFF net(LAN) booting but the stupid thing still TRIES LAN boot when I exit... I'm going to guess that insyde is bad at writing BIOSes...

[EDIT]
Heh! They didn't update the version # inside of the BIOS setup still says v. 3.7 but obviously it isn't with all of the new options... (It said for win7 support, so I figured that it'd give me the option to kill secure boot, and it did... now if it actually works... BIOS was released last week)
[/EDIT]

[EDIT2]
RTC is already set to "YES" I think it's the kludged together BIOS...
[/EDIT2]

[EDIT3]
The below config plist is the one that I've been using renamed to config.plist of course. Looks OK to me, but I'm still working under the theory of shoddy BIOS...
[/EDIT3]

 

[cutterjohn]

Yep, tinkered around some more last night after updating the BIOS.

Back to around where I started with the HP clover package(minus hp stuff), but looking at the logs, I see packagekit throwing errors 24 and 22, then keychain complained about too many open files and not being able to verify system something(forgot to take a pic or write it down), and there was pread error tossed in there.

Stage 1 where it verifies the disk image on the install partition went fine(back to the last Mavericks install app which I presume would be 10.9.5, but I'll admit to not checking other than it wouldn't make sense for apple not to have the final version up, as I missed individually downloading the install packages after every update). The install just seems to hang once that keychain check can't be done message shows up as after 40m or so still no new messages show in the log which is set to show ALL BTW.

The one that I saved after 10.9.3 complains about the notebook not being a mac when booting installer(USB) partition.

I've got secure boot off now(disabled), UEFI boot, "other os" support, legacy USB, lan boot disabled. I left WiFi enabled as it shouldn't make any difference and makes it easier to boot into win8 w/o having to jump through lenovo hoops to go to BIOS, enable it, reboot...(assuming the flakey BIOS actually saves changes and otherwise doesn't freak).

Install partition: Deleted old "D:" driver partition, shrunk the "C:" partition -> ~500GB empty space -> format exfat -> disk utility -> mac os extended -> install
kexts: fakesmc, genericusb, voodoops2, rt8111
drivers64uefi: delete vbox-hfs, add hfs efi(OS X), aptiofix(not 2 might try that next) and the variable fix one(I forget what it called ATM)

I'll probably also end up trying to get to a console from before starting phase 1 disk image verify to see if disk utility is actually handling the drive properly maybe?

So, I guess the next thing to try is maybe the partition driver, the other optiofix? Or would it be better/easier to just toss in that blank drive that I've mentioned before? (750GB WDD black, too bad it doesn't have msata as I've also got a new 128GB msata drive lying around ATM)

I'm still inclined to generally blame the shoddy BIOS as this notebook did work at one time apparently, and from the messages that I've read on the Lenovo forums it would seem that there have been at least 3 or 4 BIOS versions released already given settings mentioned that do NOT appear in either my original BIOS(didn't write down but old was <bunchaletters>26 new was <same bunchaletters>28 but both just say 3.7 inside of BIOS) nor the "new" BIOS. Additionally from posts in the lenovo forums there are posts of some people having XYZ and other not, e.g. one BIOS apparently had the capability to change ACPI versions.

 

[RehabMan]

Yep, tinkered around some more last night after updating the BIOS.

Back to around where I started with the HP clover package(minus hp stuff), but looking at the logs, I see packagekit throwing errors 24 and 22, then keychain complained about too many open files and not being able to verify system something(forgot to take a pic or write it down), and there was pread error tossed in there.

Stage 1 where it verifies the disk image on the install partition went fine(back to the last Mavericks install app which I presume would be 10.9.5, but I'll admit to not checking other than it wouldn't make sense for apple not to have the final version up, as I missed individually downloading the install packages after every update). The install just seems to hang once that keychain check can't be done message shows up as after 40m or so still no new messages show in the log which is set to show ALL BTW.

The one that I saved after 10.9.3 complains about the notebook not being a mac when booting installer(USB) partition.

I've got secure boot off now(disabled), UEFI boot, "other os" support, legacy USB, lan boot disabled. I left WiFi enabled as it shouldn't make any difference and makes it easier to boot into win8 w/o having to jump through lenovo hoops to go to BIOS, enable it, reboot...(assuming the flakey BIOS actually saves changes and otherwise doesn't freak).

Install partition: Deleted old "D:" driver partition, shrunk the "C:" partition -> ~500GB empty space -> format exfat -> disk utility -> mac os extended -> install
kexts: fakesmc, genericusb, voodoops2, rt8111
drivers64uefi: delete vbox-hfs, add hfs efi(OS X), aptiofix(not 2 might try that next) and the variable fix one(I forget what it called ATM)

I'll probably also end up trying to get to a console from before starting phase 1 disk image verify to see if disk utility is actually handling the drive properly maybe?

So, I guess the next thing to try is maybe the partition driver, the other optiofix? Or would it be better/easier to just toss in that blank drive that I've mentioned before? (750GB WDD black, too bad it doesn't have msata as I've also got a new 128GB msata drive lying around ATM)

I'm still inclined to generally blame the shoddy BIOS as this notebook did work at one time apparently, and from the messages that I've read on the Lenovo forums it would seem that there have been at least 3 or 4 BIOS versions released already given settings mentioned that do NOT appear in either my original BIOS(didn't write down but old was <bunchaletters>26 new was <same bunchaletters>28 but both just say 3.7 inside of BIOS) nor the "new" BIOS. Additionally from posts in the lenovo forums there are posts of some people having XYZ and other not, e.g. one BIOS apparently had the capability to change ACPI versions.

See link post #2.

 

[cutterjohn]

See link post #2.

If you follow the breadcrumb trail, I believe what he is referring to is that they discovered that installing OSX to an existing GPT drive that the usually installed M$ reserved partition breaks OSX all over the place(mostly disk utility, and most likely the entire disk subsystem I'd hazard) so the "fix" appears to be to remove said reserved partition.

Also mentioned were EFI parts <100MB were apparently also problematical for OSX.

http://www.tonymacx86.com/multi-booting/133940-mavericks-windows-8-same-drive-without-erasing.html

I'll update this when I get around to trying this, maybe over the weekend or IOW the "BIOS"(they're not really any longer, more like picoOSes) MAY not be as much of a problem as expected, although as usual Lenovo's "BIOS" seems to leave a great deal to be desired in functionality let alone options.

I suspect that that IS the problem as I suspect that Disk Utility APPEARING to cleanly format the designated install part doesn't really work, partially evidenced by win8 diskmgmt still thinks that it's a NTFS/exfat part even though Disk Utility has theoretically reformatted it as Mac OS Extended(which of course winders shouldn't cotton to).

[EDIT]
Hopefully this eventually saves some people from wasting their time pointlessly wading through a bunch of disjointed wandering threads...
[/EDIT]

 

[RehabMan]

If you follow the breadcrumb trail, I believe what he is referring to is that they discovered that installing OSX to an existing GPT drive that the usually installed M$ reserved partition breaks OSX all over the place(mostly disk utility, and most likely the entire disk subsystem I'd hazard) so the "fix" appears to be to remove said reserved partition.

Also mentioned were EFI parts <100MB were apparently also problematical for OSX.

http://www.tonymacx86.com/multi-booting/133940-mavericks-windows-8-same-drive-without-erasing.html

I'll update this when I get around to trying this, maybe over the weekend or IOW the "BIOS"(they're not really any longer, more like picoOSes) MAY not be as much of a problem as expected, although as usual Lenovo's "BIOS" seems to leave a great deal to be desired in functionality let alone options.

I suspect that that IS the problem as I suspect that Disk Utility APPEARING to cleanly format the designated install part doesn't really work, partially evidenced by win8 diskmgmt still thinks that it's a NTFS/exfat part even though Disk Utility has theoretically reformatted it as Mac OS Extended(which of course winders shouldn't cotton to).

[EDIT]
Hopefully this eventually saves some people from wasting their time pointlessly wading through a bunch of disjointed wandering threads...
[/EDIT]

LOL...

I would have never guessed your rambling (no offense) in post #12 was referring to the inability to create and/or format a partition in OS X Disk Utility.

 

[cutterjohn]

LOL...

I would have never guessed your rambling (no offense) in post #12 was referring to the inability to create and/or format a partition in OS X Disk Utility.

Oh. Anyways... for the time being I've wussed out and sacrificed the scorpio black 750GB to the greater good of a quicker OS X install.

So, I've got 10.10.1 running instead of Mavericks(figured WTH), and am now down to DSDT SSDT audio etc. patching. The RT USB ac1200 is working with the RT genero drivers.

Plan now will be to have the 750GB OS X install drive reside where the ODD used to be(ordered hdd caddy), OEM hdd back to it's original spot -> as mentioned just need to start fixing up the onerous details.

Time permitting, I'll eventually mess around with removing the M$ reserved part and see how that goes, but I'm in q&d get it done today mode w/o f'ing around any longer. Assuming that I end up with time to muck around with removing the reserved part, I'll list the steps along with install replacing the original post.

Now back to digging up DSDT/SSDT info.

Clovergrowerpro got version bumped again, 3037 IIRC still builds, even built it on the g50/yosemite. (God I've got to find a compatible(or can be forced to) minipcie this ginormous USB dongle is annoying...)

[EDIT]
So, we're at:
Fastest way to get a basic OS X up and running is to either blow away the OEM drive with windows or swap that drive for a nice blank drive and install OS X per clover installer methods.

NOTE: IF looking at one of the ones with phase 1/phase 2 for me ONLY the last OS X mavericks install.app from the store had two phases. Phase 1 was merely verifying the install drive image contained within the "app" and phase 2 being the actual installing of Mavericks to the target disk/part. The downloaded Yosemite install "app" skips the image verification and just goes straight to installing, so no phase 2 required.

I only used aptiofix and uefi variables fixes and it seems to be working fine for the g50-70.

Installing to an existing win8(probably win7 too) disk will require two things IF you dig through the install OS X to an existing win8(7) installed disk:
1. UEFI part >100MB (was already 260MB on my OEM setup, so that wasn't my problem)

2. The Microsoft Reserved part(looks to be used for bitlocker) needs to be removed OR even IF Disk Utility LOOKS like it's working, it doesn't. i.e. it can't format your target part

The reserved part needs to be removed. There are instructions buried in the aforementioned thread, but should I get excited I'll try an install to OEM drive and list all the steps, barring the EFI resize, which you'll have to trust the install to existing windows drive thread for.

3. The target part need to be formatted in windows (compmgmt) or Disk Utility WILL complain and not do anything anyways

Now on to DSDT/SSDT patching on my squeaky clean drive. Let the new headaches begin!
[/EDIT]

 

[cutterjohn]

alrighty dumped DSDT and all 7 SSDTs.

decompiled dsdt.aml -> dsdt.dsl

maciasl -> dsdt.dsl

added your laptop repo(http://raw.github.com/RehabMan/Laptop-DSDT-Patch/master)

compile dsdt.dsl -> 17 syntax error with acpi v4

compile dsdt.dsl -> 17 syntax errors with v5 set(I think that it's v5 as one of the lenovo thread(lenovo forum) mentioned a BIOS option to switch acpi to v4 which is no longer present...

syntax errors: no clue
e.g.

Store (REG6, Local1)
Store (0x00F0F000, REG6)
Store (\_GPE.MMTB (Local2, \_GPE.OSUP (Local2)), Store (Local1, REG6))

The error is the last line. Not sure why. It has two args UNLESS is it supposed to be:

Store(\_GPE.MMTB (Local2), \_GPE.OSUP (Local2))
Store(Local1, REG6)

the guide that I'm looking at is pretty worthless about these syntax errors as their examples are the most trivial case
Error for this one is:

7117, 6126, syntax error, unexpected PARSEOP_STORE

and then I have ones like:

If (CondRefOf (\_SB.PCI0.XHC.PS0X))
                {
                    PS0X
                }

the error maciasl shows

10402, 6126, syntax error, unexpected '}', expecting '('

I have a bunch like these and they all have CondRefOf in common, same basic pattern:

If(CondRefOf (\_SB.PCI0.XHC.<XYZ>))
{
       <PQR>
}

I haven't even gotten any farther than this yet as the guide recommends(obviously) getting a cleanly compiling .dsl and THEN starting to patch, which is pretty obvious.

So, am I missing something obvious here?

[EDIT]
For context, here's an entire method(?) that's throwing the same syntax error as above, and I still don't see anything wrong...

                Method (_PS3, 0, Serialized)
                {
                    ADBG ("WiFi1 Enter D3")
                    If (CondRefOf (\_SB.PCI0.SDHC.WI01.PS3X))
                    {
                        PS3X
                    }
                }

I used this one because it's short enough to not be annoying in it's entirety, some of these are quite lengthy...[/EDIT]

 

[RehabMan]

alrighty dumped DSDT and all 7 SSDTs.

decompiled dsdt.aml -> dsdt.dsl

maciasl -> dsdt.dsl

added your laptop repo(http://raw.github.com/RehabMan/Laptop-DSDT-Patch/master)

compile dsdt.dsl -> 17 syntax error with acpi v4

compile dsdt.dsl -> 17 syntax errors with v5 set(I think that it's v5 as one of the lenovo thread(lenovo forum) mentioned a BIOS option to switch acpi to v4 which is no longer present...

syntax errors: no clue
e.g.

Store (REG6, Local1)
Store (0x00F0F000, REG6)
Store (\_GPE.MMTB (Local2, \_GPE.OSUP (Local2)), Store (Local1, REG6))

The error is the last line. Not sure why. It has two args UNLESS is it supposed to be:

Store(\_GPE.MMTB (Local2), \_GPE.OSUP (Local2))
Store(Local1, REG6)

the guide that I'm looking at is pretty worthless about these syntax errors as their examples are the most trivial case
Error for this one is:

7117, 6126, syntax error, unexpected PARSEOP_STORE

and then I have ones like:

If (CondRefOf (\_SB.PCI0.XHC.PS0X))
                {
                    PS0X
                }

the error maciasl shows

10402, 6126, syntax error, unexpected '}', expecting '('

I have a bunch like these and they all have CondRefOf in common, same basic pattern:

If(CondRefOf (\_SB.PCI0.XHC.<XYZ>))
{
       <PQR>
}

I haven't even gotten any farther than this yet as the guide recommends(obviously) getting a cleanly compiling .dsl and THEN starting to patch, which is pretty obvious.

So, am I missing something obvious here?

It seems to me you're not disassembling correctly...

You will get a better disassembly if you disassemble all at once...

You will need to extract all DSDT/SSDT from Linux. They are available in /sys/firmware/acpi/tables and /sys/firmware/acpi/tables/dynamic. Place them on USB or otherwise transfer to OS X.

It is not necessary to install Linux. Simply run it from USB: http://www.ubuntu.com/download/desktop/create-a-usb-stick-on-windows

In Linux Terminal:

# substitute DEST with the mountpoint of a FAT32 formatted USB stick
sudo cp -R /sys/firmware/acpi/tables DEST

Place all SSDT/DSDT in a single directory (rename them to *.aml as you move them) and use a recent build of iasl to disassemble:
https://bitbucket.org/RehabMan/acpica/downloads

cd "to directory where you placed all SSDT/DSDT"
iasl -da -dl *.aml

Then work with the resulting *.dsl. You'll find you have less errors to deal with.

 

[cutterjohn]

rehab: already did all of that. As of matter of fact I extracted the dsdt & ssdts by two methods(backup my clover config and installed 14.04LTS Ubuntu on there to do the below):
method 1:
manually cp'ed them from /sys/firmware/acpi/tables & tables/dynamic

method 2:
ran acpidump and acpiextract IIRC which I did just to make sure that I didn't miss any ssdts

So, I'm pretty satisfied that I have good DSDT/SSDT.

Also, checking things, I just realized that I made a BIG mistake. The Errors above are from the DSDT that MacIASL sucks out when it starts up. (I REALLY wish that it would NOT do that...)

OK. I had just been doing:
rename DSDT -> DSDT.aml, SSDTx -> SSDTx.aml where x=1-7
iasl -da *.aml

OK adding in the -dl switch gets me down to two errors:

7121, 6126, syntax error, unexpected PARSEOP_STORE

for this same line again:

                Store (REG6, Local1)
                Store (0x00F0F000, REG6)
                Store (\_GPE.MMTB (Local2, \_GPE.OSUP (Local2)), Store (Local1, REG6))

that last line is the one where maciasl is claiming the error and my fix would be the same as proposed above maybe?

                Store (REG6, Local1)
                Store (0x00F0F000, REG6)
                Store (\_GPE.MMTB (Local2), \_GPE.OSUP (Local2)), 
                Store (Local1, REG6)

I only have TWO errors now, and they're both like this. So:

7121, 6126, syntax error, unexpected PARSEOP_STORE
7613, 6126, syntax error, unexpected PARSEOP_STORE

[EDIT]
I'm giving up on this for tonight as I'm practically going crosseyed looking at this mess, and I haven't even gotten to the patching fun yet...

Thanks
[/EDIT]

 

[RehabMan]

There is a patch in my repo for those errors. Hint: the fix is to remove them.