Jump to content

[HOWTO] Boot Macs with Intel Chipset in AHCI Win7 Vista XP Linux

ludacrisvp

By ludacrisvp,
in Boot Camp

 Share
423 posts in this topic

Recommended Posts

Euclid1
Euclid1
Posted
Posted

>MV=0

This value is RO (read only) so you cannot write it. It's probably set to 1 for legacy controllers that do not support AHCI.

 

>0x60 does switch to AHCI

 

0x60 does successfully change the controller to AHCI. I have verified this. It can also be verified using the RWEverything tool:

http://jacky5488.myweb.hinet.net/

 

1. Start Windows with IDE

2. Install the tool

3. Run the tool, select the IDE controller in the interface

4. Now go to the command line (in the tool)

5. And write the word 0x60 to 0x90 (check the help file for the command)

 

 

The view then updates and you can see that the device id has changed and the mode as well.

 

Also the reason that Windows cannot find the boot device (blue screen error 0x0007b) is a strong indication that it has changed to AHCI (otherwise it would find the boot device).

 

I also did a CriticalDeviceDatabase injection into the Windows registry:

Windows Registry Editor Version 5.00

 

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\CriticalDeviceDatabase\pci#ven_8086&dev_1c03]

"Service"="msahci"

"ClassGUID"="{4d36e96a-e325-11ce-bfc1-08002be10318}"

 

This tells Windows to use the msahci controller if it finds a new device with vendor/deviceid of the controller we have. But this didn't make a change either.

Link to comment
Share on other sites
ludacrisvp
ludacrisvp
Posted
  • Author
Posted

From the inf file for the sandybridge drivers from intel show that it uses the same driver for both of the following device ids.

 

[strings]

INTEL="Intel"

;CougarPoint AHCI

PCI\VEN_8086&DEV_1C02.DeviceDesc="Intel® 6 Series/C200 Series Chipset Family 6 Port SATA AHCI Controller - 1C02"

PCI\VEN_8086&DEV_1C03.DeviceDesc="Intel® 6 Series/C200 Series Chipset Family 6 Port SATA AHCI Controller - 1C03"

 

I would say that a forced driver install has the best chance of success.

Link to comment
Share on other sites
Euclid1
Euclid1
Posted
Posted

The msahci.sys is what the Intel inf "driver" tells Windows to use:

cougahci.inf

 

[iNTEL_HDC]

;CougarPoint AHCI

%PCI\VEN_8086&DEV_1C02.DeviceDesc%=intelahci,PCI\VEN_8086&DEV_1C02

%PCI\VEN_8086&DEV_1C03.DeviceDesc%=intelahci,PCI\VEN_8086&DEV_1C03

 

 

[iNTEL_HDC.NTamd64]

;CougarPoint AHCI

%PCI\VEN_8086&DEV_1C02.DeviceDesc%=intelahci,PCI\VEN_8086&DEV_1C02

%PCI\VEN_8086&DEV_1C03.DeviceDesc%=intelahci,PCI\VEN_8086&DEV_1C03

 

 

;**********************************************************

; Intel AHCI Controller

 

[intelahci]

Include=mshdc.inf

Needs=msahci_Inst

 

[intelahci.Services]

Include=mshdc.inf

Needs=msahci_Inst.Services

Link to comment
Share on other sites
Euclid1
Euclid1
Posted
Posted

The drivers are 100% correct (see cougahci.inf). I also have another Sandy Bridge system here with the exact same controller (deviceid/revision) and it boots with the msahci.sys driver.

 

I believe you have a pre 2011 MBP, can you run some simple tests? Basically what I need to know is:

 

What registers change when the controller is manually switched from 0x20 to 0x40. This can be seen using the RWEverything tool:

http://jacky5488.myweb.hinet.net/

 

The best would be if you can boot Windows with the IDE controller.

 

Or alternatively, boot with a Ubuntu CD into Grub2 and read the pci registers with lspci and then change the controller to AHCI using setpci and read the pci registers out again.

 

Please let me know if that would be possible and where I send/post the instructions for the tests?

 

Thanks.

Link to comment
Share on other sites
ludacrisvp
ludacrisvp
Posted
  • Author
Posted

I had a MacBook pro 4,1 that was previous to the unibody MacBook pro. I returned it after 2 weeks due to screen warping into a smiley face. I got a real Mac pro 2008 instead. Both new from apple online and the Mac pro was cheaper than the MacBook pro 17" that I had.

Link to comment
Share on other sites
Euclid1
Euclid1
Posted
Posted

Last update is as following.

 

- Installed Ubuntu 10.10 from the Live CD

- Booted into Grub 1.98 (had to select Windows in the refi menu and not Linux)

- Went to the command line "c"

- setpci -d 8086:1c01 90.b=60 (Also verified that if this is not done, that IDE mode is used in Ubuntu)

- Esc

- Started Ubuntu

 

Result

- AHCI loads fine and this is what lspci shows:

 

00:1f.2 SATA controller: Intel Corporation Cougar Point 6 port SATA AHCI Controller (rev 05) (prog-if 01 [AHCI 1.0])

Subsystem: Intel Corporation Device 7270

Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+

Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-

Latency: 0

Interrupt: pin B routed to IRQ 46

Region 0: I/O ports at 3148

Region 1: I/O ports at 315c

Region 2: I/O ports at 3140

Region 3: I/O ports at 3158

Region 4: I/O ports at 3060

Region 5: Memory at 8fa00000 (32-bit, non-prefetchable)

Capabilities: [80] MSI: Enable+ Count=1/1 Maskable- 64bit-

Address: fee1100c Data: 4199

Capabilities: [70] Power Management version 3

Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot+,D3cold-)

Status: D0 NoSoftRst+ PME-Enable- DSel=0 DScale=0 PME-

Capabilities: [a8] SATA HBA v1.0 BAR4 Offset=00000004

Capabilities: [b0] PCI Advanced Features

AFCap: TP+ FLR+

AFCtrl: FLR-

AFStatus: TP-

Kernel driver in use: ahci

Kernel modules: ahci

00: 86 80 03 1c 07 04 b0 02 05 01 06 01 00 00 00 00

10: 49 31 00 00 5d 31 00 00 41 31 00 00 59 31 00 00

20: 61 30 00 00 00 00 a0 8f 00 00 00 00 86 80 70 72

30: 00 00 00 00 80 00 00 00 00 00 00 00 0b 02 00 00

40: 00 80 00 80 00 00 00 00 00 00 00 00 00 00 00 00

50: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

60: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

70: 01 a8 03 40 08 00 00 00 00 00 00 00 00 00 00 00

80: 05 70 01 00 0c 10 e1 fe 99 41 00 00 00 00 00 00

90: 60 3a 05 85 83 01 00 3a 08 42 5c 01 00 00 00 00

a0: e0 00 00 00 39 0f 00 00 12 b0 10 00 48 00 00 00

b0: 13 00 06 03 00 00 00 00 00 00 00 00 00 00 00 00

c0: 00 00 00 00 07 00 00 00 00 00 00 00 00 00 00 00

d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

f0: 00 00 00 00 00 00 00 00 87 0f 05 08 00 00 00 00

 

Tried the same on Windows and BSOD as expected. So apparently the Linux kernel can handle everything without any problems.

 

Next step is to compare the registers on a PC when BIOS is set to IDE and then switched to AHCI.

 

Update 1

 

There does not seem to be anything unexpected in the PCI registers. I have created a compilation of all dumps I have collected from MBP 2011 and Dell XPS 17x2: https://www.rapidshare.com/#!download|4...umps.rar|72.654

 

The readme.txt with a description is inside.

 

 

Update 2

 

I have found some Intel (Intel® RST Driver Files) beta drivers for the "Intel® Mobile Express Chipset SATA AHCI Controller" controller (non RAID) from here:

http://www.station-drivers.com/page/intel%20raid.htm

 

and manually installed them:

- Device Manager

- Update Drivers

- Search for drivers on the computer, Have Disk

- Selected iaAHCI.inf (the iaStor.inf is for the RAID controllers)

- Select "Intel® Mobile Express Chipset SATA AHCI Controller"

- Driver installs and reboots

- In the Grub loader, set 90.b to 60

- Windows does not BSOD with 7b, so I think it actually finds a boot device

- But when iaStore.sys continues, I get a BSOD DRIVER_IRQL_NOT_LESS_OR_EQUAL caused by iaStor.sys.

 

I tried the following driver versions:

10.5.0.1022

10.5.0.1015

 

and the official WHQ ones from here:

http://downloadcenter.intel.com/Detail_Des...p;DwnldID=19601

 

Update 3

 

Windows Vista/7 MBR to enable AHCI. Tested on MBP 2011 17". If you do not have the exact same hardware configuration as the MBP 2011 17" it probably won't work.

 

https://www.rapidshare.com/#!download|4...edcode.bin|0.44

 

To patch the MBR:

 

  1. Copy the patchedcode.bin to a USB stick
  2. Boot with MacOSX Setup CD (or Ubuntu Live CD or anything else where you can write the MBR to the disk)
  3. In the first screen select your language, in the 2nd screen, open a Terminal through the menu at the top.
  4. Go to your USB stick: cd /Volumes/USBSTICKNAME
  5. IMPORTANT: Make a backup of your mbr first: dd if=/dev/disk0 of=backup.bin bs=512 count=1
  6. Unmount all mapped drives from your disk0: umount /dev/disk0s1, umount /dev/disk0s2, etc.
  7. Write the new mbr: dd if=patchedcode.bin of=/dev/disk0 bs=440 count=1
  8. Type: reboot now

If for some reason it should not work for you, you can restore the MBR as following:

dd if=backup.bin of=/dev/disk0 bs=440 count=1

 

It is assumed that you already have set the msahci service start type to 0 in the Windows operating system.

 

I will post a detailed writeup later.

Link to comment
Share on other sites
jacoch
jacoch
Posted
Posted

GREAT JOB! Tried on my MBP 15" 2011. Works fine. How did you get this working MBR?

 

EDIT

I sometime get a BSOD (0x0000007B) when starting MBP with a second HD in place of super drive. With Intel drivers 10.1.0.1008. Will try beta drivers above to see if it fixes the problem.

Link to comment
Share on other sites
ilkevinli
ilkevinli
Posted
Posted

Thanks so much for posting this and for your time researching this. This method worked great on my 13" 2011 MacBook Pro i5.

 

 

 

 

 

Last update is as following.

 

- Installed Ubuntu 10.10 from the Live CD

- Booted into Grub 1.98 (had to select Windows in the refi menu and not Linux)

- Went to the command line "c"

- setpci -d 8086:1c01 90.b=60 (Also verified that if this is not done, that IDE mode is used in Ubuntu)

- Esc

- Started Ubuntu

 

Result

- AHCI loads fine and this is what lspci shows:

 

 

 

Tried the same on Windows and BSOD as expected. So apparently the Linux kernel can handle everything without any problems.

 

Next step is to compare the registers on a PC when BIOS is set to IDE and then switched to AHCI.

 

Update 1

 

There does not seem to be anything unexpected in the PCI registers. I have created a compilation of all dumps I have collected from MBP 2011 and Dell XPS 17x2: https://www.rapidshare.com/#!download|4...umps.rar|72.654

 

The readme.txt with a description is inside.

 

 

Update 2

 

I have found some Intel (Intel® RST Driver Files) beta drivers for the "Intel® Mobile Express Chipset SATA AHCI Controller" controller (non RAID) from here:

http://www.station-drivers.com/page/intel%20raid.htm

 

and manually installed them:

- Device Manager

- Update Drivers

- Search for drivers on the computer, Have Disk

- Selected iaAHCI.inf (the iaStor.inf is for the RAID controllers)

- Select "Intel® Mobile Express Chipset SATA AHCI Controller"

- Driver installs and reboots

- In the Grub loader, set 90.b to 60

- Windows does not BSOD with 7b, so I think it actually finds a boot device

- But when iaStore.sys continues, I get a BSOD DRIVER_IRQL_NOT_LESS_OR_EQUAL caused by iaStor.sys.

 

I tried the following driver versions:

10.5.0.1022

10.5.0.1015

 

and the official WHQ ones from here:

http://downloadcenter.intel.com/Detail_Des...p;DwnldID=19601

 

Update 3

 

Windows Vista/7 MBR to enable AHCI. Tested on MBP 2011 17". If you do not have the exact same hardware configuration as the MBP 2011 17" it probably won't work.

 

https://www.rapidshare.com/#!download|4...edcode.bin|0.44

 

To patch the MBR:

 

  1. Copy the patchedcode.bin to a USB stick
  2. Boot with MacOSX Setup CD (or Ubuntu Live CD or anything else where you can write the MBR to the disk)
  3. In the first screen select your language, in the 2nd screen, open a Terminal through the menu at the top.
  4. Go to your USB stick: cd /Volumes/USBSTICKNAME
  5. IMPORTANT: Make a backup of your mbr first: dd if=/dev/disk0 of=backup.bin bs=512 count=1
  6. Unmount all mapped drives from your disk0: umount /dev/disk0s1, umount /dev/disk0s2, etc.
  7. Write the new mbr: dd if=patchedcode.bin of=/dev/disk0 bs=440 count=1
  8. Type: reboot now

If for some reason it should not work for you, you can restore the MBR as following:

dd if=backup.bin of=/dev/disk0 bs=440 count=1

 

It is assumed that you already have set the msahci service start type to 0 in the Windows operating system.

 

I will post a detailed writeup later.

Link to comment
Share on other sites
Shade69
Shade69
Posted
Posted

Thanks for your hard working ! (ludascrisvp & Euclid1) :D

I follow this thread since March.

Works perfectly on a MacbookPro 2011" 15 Corei7 (6750M) (Win7 / GUID Partition - Winclone)

I use the Apple SSD 120GO in the superdrive bay for bootcamp.

 

Euclid1, what's the difference between previous MBR registry code and now ? (For personal g33k knowledge) :P

Link to comment
Share on other sites
jacoch
jacoch
Posted
Posted
Nice! I am happy to finally see that you were able to figure it out. Like Shade69 mentioned, what was different between the original method and what worked? Was it simply using the new drivers you found?

 

Not really. As mentionned above, I first tried with official Intel drivers. And it worked fine (if super drive is in its slot). So he changed something in modified MBR. Regarding drivers, with beta ones, I don't get BSOD anymore when a HD is installed in place of super drive.

Link to comment
Share on other sites
Euclid1
Euclid1
Posted
Posted

The ABAR address needs to be set as well. The trick was to find a correct address. The setpci commands are as following:

 

setpci -d 8086:1c01 90.b=60

setpci -d 8086:1c03 24.l=8fa00000

 

This is basically what the MBR patch does. I moved the code to a different place than the original patch. The original patch had the code running multiple times. This was most likely not desired. Now the code runs right after the A20 initialization and before the TPM init.

 

;------- A20 init ---------
...

seg000:00E2 loc_E2:							   ; CODE XREF: seg000:00C9j
seg000:00E2				 jmp	 loc_163
seg000:00E5; ---------------------------------------------------------------------------
seg000:00E5

; ------- TPM init---------
; ....

; and this is contoller setup

seg000:0163 loc_163:							  ; CODE XREF: seg000:loc_E2j
seg000:0163				 mov	 eax, 8000FA90h
seg000:0169				 mov	 dx, 0CF8h
seg000:016C				 out	 dx, eax
seg000:016E				 mov	 ax, 60h; '`'
seg000:0171				 mov	 dx, 0CFCh
seg000:0174				 out	 dx, ax
seg000:0175				 mov	 eax, 8000FA24h
seg000:017B				 mov	 dx, 0CF8h
seg000:017E				 out	 dx, eax
seg000:0180				 mov	 eax, 8FA00000h
seg000:0186				 mov	 dx, 0CFCh
seg000:0189				 out	 dx, eax
seg000:018B				 mov	 ax, 0BB00h
seg000:018E				 jmp	 loc_E5

Link to comment
Share on other sites
gabiz_ro
gabiz_ro
Posted
Posted

Was not enough only to switch controller to AHCI by rewriting offset 90h with 60 value.

There is need also to set ABAR so chipset can have memory space.

After switching from IDE to AHCI chipset registers was left set to I/O ports and needed to be rewrite offset 24h to set base address for memory space.

 

:) posted at same time

 

Bad new is that if Apple really don't want you tu use it on AHCI on next update that memory mapped region could be allocated to something else ,that will result in error.

 

But since Euclid1 know that linux can start in AHCI and alocate itself what free region find,then patch can be readjusted.

Link to comment
Share on other sites
Euclid1
Euclid1
Posted
Posted
But since Euclid1 know that linux can start in AHCI and alocate itself what free region find,then patch can be readjusted.

 

Correct. I figured out that Ubuntu 10.10 (BIOS mode) was able to use the AHCI controller even though the ABAR (BAR5) was not correct. So my guess is that the Linux kernel automatically fixes the problem by setting the BAR5 to a free region. If the PCI configuration changes (maybe due to a firmware upgrade in the future) a new ABAR address needs to be found by booting into GRUB to switch the controller to AHCI and then continuing with a Linux kernel. And then the base address can be extracted from a linux terminal with: setpci -d 8086:1c03 24.l

Link to comment
Share on other sites
Euclid1
Euclid1
Posted
Posted
Have you contacted johnsock about getting this added into the GUI application that runs in osx to mod the mbr easier?

 

I tried to contact him before but never heard anything back. But he is free to use/incorporate the new MBR patch in/into his tool.

Link to comment
Share on other sites
ludacrisvp
ludacrisvp
Posted
  • Author
Posted

How does your modded mbr compare to a normal one when viewing it as hex it seems that the first 350 bytes or so seem to be like any other mbr but about the last 80 or so bytes (up to the 440 mark) seem to be different the script is setup to write hex string to mbr so it's as simple as modifying the hex string from the script to match what changes you have made. I'm not sure how to make changes to his app right bow but the script is easy enough to change.

Link to comment
Share on other sites
Shade69
Shade69
Posted
Posted

I found a critical bug with this parallel desktop 6 (VM = use the Bootcamp partition mode) and the patch. It seems that Parallel modify at the beginning the mbr for his own ahci activation. So if you want to use at least the win7 partition you've to re-apply the patch everytime you use Parallel... :)

Euclid1 do you know if the problem could be fixed?

Thanks. :)

Link to comment
Share on other sites
ludacrisvp
ludacrisvp
Posted
  • Author
Posted

Chances are the only fix there would be to modify parallels to not write anything to the mbr or make it write the new code instead. You might also consider switching over to vmware fusion instead and see if they cause the same issue. I'm personally a fan of vmware.

 

Edit: updated first post with directions for euclid1's mbr patch.

Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...