Thanks, I’ll check it out. Although I’ve never had trouble with Windows 7/10 and AHCI, but that was on ICH8-M or higher. There must be something special that Windows is requiring. Unfortunately I’m not sure if I have access to any other ICH7 systems to compare the chipset configuration. Most of the stuff I have at work are Dell systems. No rush on the /proc/iomem. It’s just a one-line change in the DSDT. I think I’m going to revert back to an IDE BIOS and check out if suspend/resume works. Then I’ll know that I need to do something in the ACPI wake method. I tried applying AHCI again but that was worse than not doing anything. UPDATE Suspend and resume work in IDE mode under CentOS 7...not so much in AHCI. I briefly saw an error pop up in the console about trying to reset the SATA bus, but the drive was not responding. Unfortunately too many times doing this and the XFS filesystem got corrupted. I believe I do have a few Dell D820's lying around at work...will have to see if any still function.
I see - I read your update. What do you plan to do with the Dell D820? One more day delay: I was bitten by some strange grey striped insect today and have an allergic reaction - arm swollen like a ball & itching like hell. I took some anti-histamine, can do nothing but crush now and hope I'm fit again tomorrow to report back.
I'm not sure anymore. From what I've read, Dell didn't enable AHCI in the D820 either...was hoping that they did and I could peak at the SATA registers. Hope you feel better.
Thanks. Am feeling pretty sick. Still went to work which took me to the edge. Can't do anything anymore right now but go to sleep. Maybe tomorrow
Am back up and running (after lots of sleep). Attached you find the /proc/iomem file from my Toshi P100 booted in 32bit mode (because I run both, Mac OS 10.7.5 and Windows 10 in 32bit mode, too) with an old Parted Magic version from 2013 included in the latest System Rescue CD (I guess that was the latest freely available edition of PM before they went retail). Hope it's helpful. Let me know in case it doesn't contain the data you need, and I can download a more recent Linux distro.
Just a quick update. I have suspend/resume working in CentOS 7 in AHCI mode!!! I reviewed your /proc/iomem file and was able to find a free memory region that we have in common. Windows 7 & 10 installers do not see any hard drives when in SATA AHCI mode. It does see the SATA AHCI controller though, so it's definitely running my DSDT code. Microsoft Standard AHCI ==> no drives detected Intel MSM v7.0.0 ==> no drives detected Intel RST v9.6.0 ==> BSOD Intel RST v10.1.0 ==> BSOD Intel RST v11.2.0 ==> BSOD Intel RSTe v11.7.4 ==> BSOD I'm trying some other register probing, but I think I'm close! Do you know what the latest and pain-free way to try Mac OS on the P105? I'd like to see if that "just works" like Linux does.
Amazing - yes, sounds as if you're close! Mac OS must be 10.7.5 Lion. I did not find a website where you could download a hackintosh image. I'm at the end of the road here with very "thin" upload speed, so can't help with providing via upload. Best chance might be some via some old torrent some people hopefully still share, but I don't have any link or torrent right now :/
Just checking in. I haven't gotten around to downloading macOS Lion to a bootable USB stick. I'm still trying to figure out the Windows issue. When I open a command prompt during the Windows install, it does show the SATA controller is in AHCI mode, but no disks are detected (both the installer and diskpart). I'm looking at some SATA initialization code that is mentioned but it's kind of voodoo magic without any kind of a guide.
yee-ha, check it out, @tqhoang : not even any extra kext (kernel extension aka "driver") required - stock OS X! You rock ! I'll do the Windows reg edit for AHCI tomorrow to see if it boots, and also run benchmarks in OS X with both, IDE and AHCI. I'll let you know then. Again, a huge THANK YOU Let me know if you're still interested in OS X Lion, albeit ancient.
Wow, notable difference! Geekbench Mac 32bit: Code: Single-Core Multi-Core AHCI: Integer Score 609 1135 Floating Point Score 609 1151 Memory Score 560 586 Geekbench Score 599 1031 IDE: Integer Score 593 874 Floating Point Score 601 874 Memory Score 528 554 Geekbench Score 583 810 What's more important is that TRIM indeed is only active in AHCI mode - no TRIM in IDE comp. mode because in Mac OS, TRIM is handled by the IOAHCIBlockStorage.kext. So for SSD health, AHCI is indispensable. Geekbench Mac 64bit: Code: Single-Core Multi-Core AHCI: Integer Score 678 999 Floating Point Score 625 944 Memory Score 524 543 Geekbench Score 626 885 IDE: Integer Score 680 1005 Floating Point Score 627 913 Memory Score 516 547 Geekbench Score 626 876 Geekbench 3.4.1 in Mac OS X Lion 10.7.5 - results in one table. "Winners" in each group and consequently summed up/ indicated as either AHCI or IDE in 'red'. If two values in a group were very close to each other then I marked them both as 'winners in red'. PHP: Scores/ Single-Core Multi-Core MODE 32/64 bit 32/64 bitInteger 'AHCI' 609 /'678' '1135'/ 999 IDE 593 /'680' 874 /1005Floating Point 'AHCI' 609 /'625' '1151'/ 944 IDE 601 /'627' 874 / 913Memory 'AHCI' '560'/ 524 '586'/ 543 IDE 528 / 516 554 / 547Geekbench 'AHCI' '599'/ 626 '1031'/ 885 IDE 583 / 626 810 / 876 Clear result as expected: AHCI over IDE! Despite of 64bit CPU, run Mac OS in x86 base mode but with the possibility for 64bit Apps to run in 64bit mode.
That's great to see that MacOS works too! I did a basic benchmark under CentOS Linux 7 with dd. IIRC it was a slight improvement in hard disk throughput (<10%), but the main thing like you said is the capability of TRIM for SSD's. I'm still not sure why Windows just doesn't see the drives. I verified the controller is changed into the AHCI mode by going to a command prompt and running the Windows build of setpci. I'm going through the Intel SATA spec document now and coding some things to try. No ETA obviously, but I definitely haven't given up.