[GUIDE] Installing Snow Leopard/Leopard retail DVD on the GA-E7AUM-DS2H

[blueknight]

So the good news is that Gigabyte still engaged. They are asking if 4GB is detected on POST and/or after the OS boots. I suspect the answer for both is yes but please let me know.

 

Hi dlach,

 

4GB memory is detected in BIOS and after OSX 10.6.1 booting.

I hope the next BIOS will be good at 4GB.

[wintelmac]

I have 8GB RAM, ran memtest in SL both 32bit/64bit kernel. Memalloc stop at 3506! Good news is SL doesn't crash like in Leopard. In Leopard, memtest will need the machine to be reset. SL allows the memtest to complete without crashing.

 

BTW I am testing this on PCIe GeForce 9500 GT.

 

Attach is by DSDT extract from BIOS and patch for RTC + AUDIO

 

32bit kernel

 

malloced meg 3504, res 0xffb00000

malloced meg 3505, res 0xffc00000

malloced meg 3506, res 0xffd00000

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3507 failed

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3508 failed

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3509 failed

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

 

<--- screen break---->

 

malloc meg 8189 failed

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 8190 failed

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 8191 failed

 

 

64bit Kernel

 

malloced meg 3504, res 0xffb00000

malloced meg 3505, res 0xffc00000

malloced meg 3506, res 0xffd00000

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3507 failed

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3508 failed

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3509 failed

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

 

<--- screen break---->

 

malloc meg 8189 failed

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 8190 failed

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 8191 failed

190909_DSDT.zip

[genzai]

I have 8GB RAM, ran memtest in SL both 32bit/64bit kernel. Memalloc stop at 3506! Good news is SL doesn't crash like in Leopard. In Leopard, memtest will need the machine to be reset. SL allows the memtest to complete without crashing.

 

BTW I am testing this on PCIe GeForce 9500 GT.

 

Attach is by DSDT extract from BIOS and patch for RTC + AUDIO

 

32bit kernel

 

malloced meg 3504, res 0xffb00000

malloced meg 3505, res 0xffc00000

malloced meg 3506, res 0xffd00000

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3507 failed

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3508 failed

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3509 failed

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

 

<--- screen break---->

 

malloc meg 8189 failed

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 8190 failed

memtest(184) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 8191 failed

 

 

64bit Kernel

 

malloced meg 3504, res 0xffb00000

malloced meg 3505, res 0xffc00000

malloced meg 3506, res 0xffd00000

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3507 failed

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3508 failed

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 3509 failed

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

 

<--- screen break---->

 

malloc meg 8189 failed

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 8190 failed

memtest(130) malloc: *** mmap(size=1048576) failed (error code=12)

*** error: can't allocate region

*** set a breakpoint in malloc_error_break to debug

malloc meg 8191 failed

 

Well, as far as i know there should be no crash with >=4GB of memory if you are using a seperate video card so unfortunetally that does not tell us much. The Video Corruption / Crash problem is caused when using 4GB or more of memory and also the onboard video.

[wintelmac]

Well, as far as i know there should be no crash with <=4GB of memory if you are using a seperate video card so unfortunetally that does not tell us much. The Video Corruption / Crash problem is caused when using 4GB or more of memory and also the onboard video.

 

I have crash with this card when there is >4GB on previous version of BIOS

[genzai]

I have started delving into the new f4e bios and DSDT.

 

- First, this DSDT has by far the most warnings i have seen in any DSDT during compile. Some of the warnings i don't see common solutions for. I am working through as many of these as i can.

 

- There is one error which is preventing me from doing a proper recompile. I could force recompile but it bothers me that there is an error (again, i can not find a common solution for this one). The error is:

Return (NVH7 ())
Error	4060 -																				  Called method returns no value ^

This seems to be in the onboard video section of the DSDT, hmmmm.

 

- I am looking through Koalala's older dsdt for some hints. It appears that he greatly expanded the onbaord video section of the DSDT. This is also the section with the aforementioned error. I could simply duplicate his additions to this section but it has at least occurred to me this might not be a good idea for a couple of reasons: A) there might be incompatibilities between versions and can we say for certain that these additions are not causing the 4GB error to begin with?

 

I am eager to get this working and so i will be trying some actual tests as much as i can. But it is looking as though working through some of these issues may take a lot of time and testing and i am hoping some others on this thread will be able to help test and lend ideas based on these findings.

 

Thanks,

g\

[dlach]

I have started delving into the new f4e bios and DSDT.

 

- First, this DSDT has by far the most warnings i have seen in any DSDT during compile. Some of the warnings i don't see common solutions for. I am working through as many of these as i can.

 

- There is one error which is preventing me from doing a proper recompile. I could force recompile but it bothers me that there is an error (again, i can not find a common solution for this one). The error is:

Return (NVH7 ())
 Error	4060 -																				  Called method returns no value ^

This seems to be in the onboard video section of the DSDT, hmmmm.

 

- I am looking through Koalala's older dsdt for some hints. It appears that he greatly expanded the onbaord video section of the DSDT. This is also the section with the aforementioned error. I could simply duplicate his additions to this section but it has at least occurred to me this might not be a good idea for a couple of reasons: A) there might be incompatibilities between versions and can we say for certain that these additions are not causing the 4GB error to begin with?

 

I am eager to get this working and so i will be trying some actual tests as much as i can. But it is looking as though working through some of these issues may take a lot of time and testing and i am hoping some others on this thread will be able to help test and lend ideas based on these findings.

 

Thanks,

g\

 

The video ram problem was experienced before Koalala started patching. That line comes from the original bios and existing in the new bios as well. Also, Blueknight has patched that particular line is his SL dsdt and it has not fixed the problem, but then maybe the patch is incorrect.

 

I am not surprised that the gigabyte fixes aren't working since the dsdt in the bios they sent is clearly not ready to be interpreted by OSX. I have been working on merging koalala's dsdt.dsl and the one extracted from the new gigabyte BIOS which I ran through Koalala's patcher. I got one that I think has not quite the minimum changes from koalala's (plus the rtc cmos fix and bluenight's NVH7 fix (which I wish he would explain). I patched the new bios with the DSDT and my 10.5.8 ide install won't boot past the OSX spinning circle (the screen grays out and the "you have to shutdown" message comes up), however the same BIOS does let me boot my SL DVD via a chameleon bootCD and I am trying an install right now.

 

I also intend to try to get the dsdt diffs down to what look like the minimum memory map changes and try that with 10.5.8.

 

Will posted the results.

[blueknight]

I am so sorry that I do not understand dsdt.dsl at all.

 

I only read dsdt.dsl and find that

Method (NVH7, 0, NotSerialized)
				{
					Store (0x966A, ^^^SBUS.SMT2)
				}

 

and

 

If (CondRefOf (NVH7, Local7))
				{
					  Return (NVH7 ())
				 }

 

I think that method "NVH7" does only storing 0x966A to anywhere.

and iasl says that NVH7 has no return.

 

So there is no need to call method "NVH7".

Let's store 0x966A directly.

 

Then I modfy like this,

 

If (CondRefOf (NVH7, Local7))
				{
					  Store (0x966A, ^^^SBUS.SMT2)
				 }

[dlach]

I am so sorry that I do not understand dsdt.dsl at all.

 

I only read dsdt.dsl and find that

Method (NVH7, 0, NotSerialized)
				 {
					 Store (0x966A, ^^^SBUS.SMT2)
				 }

 

and

 

If (CondRefOf (NVH7, Local7))
				 {
					   Return (NVH7 ())
				  }

 

I think that method "NVH7" does only storing 0x966A to anywhere.

and iasl says that NVH7 has no return.

 

So there is no need to call method "NVH7".

Let's store 0x966A directly.

 

Then I modfy like this,

 

If (CondRefOf (NVH7, Local7))
				 {
					   Store (0x966A, ^^^SBUS.SMT2)
				  }

 

Thanks Blueknight. I am still workiing on the dsdt merge and want to work with genzai on getting rid of the errors, especially the _WAK error since it is how the bios responds to wake from sleep.

 

One thing you could try if you have some time is to boot the new bios on you 10.5 partition without a dsdt.aml, and then run your DSDT patcher tool to try to create a new DSDT.

[Tensuke]

Wow, thanks so much dlach and everyone else getting this to work so well! Color me VERY impressed.

[dlach]

Well the good news is that I have installed Snow Leopard to an all SATA setup using a BootCD and successfully updated to 10.6.1.

 

The bad news is that my patched bios with the merge from Gigabytes bios and Koalala's bios didn't fix memory issue.

 

Also sleep is broken. I may go back to just patching Koalala's bios to fix the CMOS reset and see if that does any better. I think Scoobydu reported everything was working so may I broke something with this merge.

 

When I get a little further I'll update the guide. I think Signal64 from the other thread may take interest and since he did a lot of the fixing of the original DSDT maybe he'll spot something.

 

Gigabyte is apparently still interested in helping us so if we can come up with something to help them out it would be great.

[THud33]

Hey dlach,

I tried your suggestions, but obviously I put the CDrom as first boot device.

I did get past the apple screen, but now when I pick use english language, it pops up ands says mac osx can not be installed on this machine. Any help here?

[dlach]

Hey dlach,

I tried your suggestions, but obviously I put the CDrom as first boot device.

I did get past the apple screen, but now when I pick use english language, it pops up ands says mac osx can not be installed on this machine. Any help here?

 

As soon as you selected a language? What exactly is the sequence? Did you try to partition your hard drive. Tell us your complete configuration cpu, memory, disks. Have you installed OSX before?

 

The way it works is you pick the language and then it asks you to click a license agreement and then it says "continue". But before you continue you have to create a GUID partition as in the guide.

[THud33]

As soon as you selected a language? What exactly is the sequence? Did you try to partition your hard drive. Tell us your complete configuration cpu, memory, disks. Have you installed OSX before?

 

The way it works is you pick the language and then it asks you to click a license agreement and then it says "continue". But before you continue you have to create a GUID partition as in the guide.

 

Yeah I figured it out. The retail dvd my uncle gave me might have been a one time install one, not the family pack one. So i just went ahead and substituted the dvd for a dvd of iatkos v7 and got it to install very easily and i still followed along with your guide till the end. thanks everyone, especially dlach!

[jon55]

In case you are interested in here I leave the bios F3A with slic 2.1 for

Windows7 and foe of roisoft DSDT that put here:

http://www.applesana.es/foro/14/17760/4/gigabyte-ga-

E7AUM-DS2H-clone-imac-2009-A.html:

Have included HPET + rtc + GFX + Hdef I put the fix

the CMOS. I have not tried the dream and is not corrected

the problem of the 2 gb. With these kext bios and what

I am working with sound.

Bios:

http://www.megaupload.com/?d=FI8EFZ41

Kext:

http://www.megaupload.com/?d=BP9B7BEZ

Sorry for my English I am Spanish.

[olegischinski]

Thanks Blueknight. I am still workiing on the dsdt merge and want to work with genzai on getting rid of the errors, especially the _WAK error since it is how the bios responds to wake from sleep.

 

One thing you could try if you have some time is to boot the new bios on you 10.5 partition without a dsdt.aml, and then run your DSDT patcher tool to try to create a new DSDT.

 

hi,

 

try to help, hope it is usefull. I've booted with the new bios into a 10.5.8 installation without a dsdt.aml and run the dsdt_patcher by fassl. Here is the unmodified result.

dsdt.dsl.zip

[dlach]

jon55: thanks for contributing, don't apologize for your english, it is much better than my Spanish ;^)

 

 

olegischinski: thanks that is a big help

 

[genzai]

@dlach,

 

I guess i cant send files with PM.

Here is my dsdt based on f4e with the NVH7 error. There are only 3 warnings now out of almost 40 in the original dsdt extraction.

Thanks,

g\

 

dsdt.dsl.zip

[dlach]

@dlach,

 

I guess i cant send files with PM.

Here is my dsdt based on f4e with the NVH7 error. There are only 3 warnings now out of almost 40 in the original dsdt extraction.

Thanks,

g\

 

dsdt.dsl.zip

 

Genzai,

 

Here is the aml and patched bios. I think the reason you couldn't compile it was the absolute path in the definition block. Anyone else who wants to play with these be forewarned, they are an intermediate step genzai wants to make before we try to merge in the bits that make the dsdt and bios work well with OSX.

genzai_dsdt.aml.zip

e7aumd2h_.genzai.f4e.zip

[olegischinski]

Hi,

 

tested this BIOS and dsdt.aml without success on my 10.5.8-Installation. Boot -v comes until loading the AppleIntelCPUPM. (no HPET-Configuration or so - this is enabled in BIOS)

Same message while booting with this BIOS without dsdt.aml.

 

Under the "old" circumstances I'm using netkas' fakesmc.kext.

Hope, this test was helpfull.

[dlach]

Hi,

 

tested this BIOS and dsdt.aml without success on my 10.5.8-Installation. Boot -v comes until loading the AppleIntelCPUPM. (no HPET-Configuration or so - this is enabled in BIOS)

Same message while booting with this BIOS without dsdt.aml.

 

Under the "old" circumstances I'm using netkas' fakesmc.kext.

Hope, this test was helpfull.

 

olegischinski - thanks for the test. We haven't put the entries into that dsdt/bios yet for HPET so if you want to try it some more use intelCPUPMdisabler.kext.

IntelCPUPMDisabler.kext.zip

[adidasprince]

THANK YOU dlach !!! GREATE JOB !!!

 

Looking forward to your SL tutorial ;-)

[olegischinski]

ok, here is the result.

 

starting-point:

- BIOS-flash

- IntelCPUPMDisabler instead of fakesmc in /E/E

- dsdt.aml in /E

 

Booting stops at the point, when the loginwindow should start. No kernel panic, it just stops. No difference, if there is a dsdt.aml or not.

 

Last Log entries:

krb5kdc[76] setsockopt(IPV6_V6ONLY,1) worked

kextd[10] writing kernel link data to /var/run/mach.sym

[dlach]

ok, here is the result.

 

starting-point:

- BIOS-flash

- IntelCPUPMDisabler instead of fakesmc in /E/E

- dsdt.aml in /E

 

Booting stops at the point, when the loginwindow should start. No kernel panic, it just stops. No difference, if there is a dsdt.aml or not.

 

Last Log entries:

krb5kdc[76] setsockopt(IPV6_V6ONLY,1) worked

kextd[10] writing kernel link data to /var/run/mach.sym

Thanks! This is interesting. Did you see any suspicious errors in the log? Kexts that failed to load etc?

[olegischinski]

Thanks! This is interesting. Did you see any suspicious errors in the log? Kexts that failed to load etc?

 

No suspicious errors, only as usual under "old" circumstances

 

- kernel IGPU: family specific matching fails

- kernel Sound assertion "0 == pathMap_aDriverInstance" failed in "/SourceCache/AppleHDA/AppleHDA-171.1.2/AppleHDA/AppleHDADriver.cpp" at line 1459 goto Exit

- kernel Sound assertion "0 != createAudioEngines ( fPathMap_aDriverInstance )" failed in "/SourceCache/AppleHDA/AppleHDA-171.1.2/AppleHDA/AppleHDADriver.cpp" at line 172 goto Exit

 

And I detected a new one, don't know since when it's in the log. It didn't disappear after switch back to the old BIOS and kext.

kernel ALF ALERT: sockwall_cntl_updaterules ctl_enqueuedata rts err 55

I'm thinking it has nothing to do with this test.

 

EDIT:

ALF disappeared after restoring a TM-Backup made before playing with the Test-BIOS

[dlach]

Well now Gigabyte is asking questions that don't give me much hope that they'll fix it. They want to know if we've run the memory test on Windows under the new bios and if so does the full 4 GB of ram show up. Anyone have a windows boot they can test this on?