Hello, I am out of ideas looking for some help!

I have tried every possible exception that I can think of to try and get my ASUS Gryphon Z87 machine working on Server 2008 R2 (Windows 7 kernel of course) with a PCIe SM961. Please let me know if there is anything left to try because I am at a loss. What I was trying to do is simply clone my existing system to this drive and have it magically be bootable and I really don't see what the issue is as I have used every possible obscure solution across the Internet that all of these people post and theorize 'work' (when they actually in real life don't).

Already preinstalled KB2990941 and KB3087873 along with the Samsung NVMe Controller 2.4.7.0

Using GRYPHON-Z87-ASUS-2103.CAP; Modified twice with insertion of NvmExpressDxe_3.ffs, one attempt using MMTool (In the Volume 04) and one using the UEFITool (also under the DXE Driver section).

I have also tried extracting the uncompressed "Nvme.ffs" "NvmeSmm.ffs" and "NVMEINT13.ffs" modules from the Gryphon Z97 just because I figured well hey, they couldn't change THAT much, right? Just FYI, these modules do not work at all.

The NvmExpressDxe_3.ffs - Modified BIOS does display the "PATA SS:" drive in the BIOS so it at least 'partially' works.

Before cloning, tried the "fix_hdc.cmd" from Hirens' in order to wipe out all HDD controllers and put the OS back to IDE mode before cloning. Neither attempts worked. I don't even know if, but this was logical for me to try so that AHCI drivers would not be interfering.

I tried cloning my old system two different ways using Clonezilla sector-by-sector copy and AOMEI, both having the same results.

All attempts were done under manually selecting "PATA SS:" under the "Boot" menu:
- Under CSM enabled and Legacy OPROM First, the SSD is attempted to boot 3 times, the screen stays black but the cursor advances ~10 or so lines to the right and down then restarts at the beginning top left. No additional text is displayed.
- Under CSM enabled and UEFI OPROM First, the SSD is attempted to boot once, the screen stays black and cursor does not advance and then reverts back to the BIOS screen.
- The system will not boot at all with "CSM Disabled" as the BIOS reports that "some bootable devices are imcompatible" and recommends to enable CSM.
- Changing the Secure Boot settings between UEFI and "Other OS" does not give any benefit/detriment either way.

I also tried booting both in Legacy and UEFI modes to my Server 2008 R2 install disk and loading the Samsung NVMe driver (which makes the drive visible) but in either case, the installer reports that it can't install because "Windows cannot boot from this device, please enable the controller in the BIOS".

Also, no, please do not tell me to install Windows 10, obviously I know that it will probably work but I do not want to use it and I never do want to use it. I like my Server 2008 R2. If it can't be done, it can't be done. I just want to raise the attention of this not working so that genius Fernando is aware and maybe would find some solution :)

Inserted the 3 modules only, minus the NvmExpressDxe_3 mod.

No expert here, not using the same board,Using win 7 not server 2008 R2 and installed not cloned but I was reading through your post and noticed a few things different...I used the NvmExpressDxe_3 mod initially and also got a listing for a Pata ssd1 and windows boot manager after I installed this stuff disappeared and instead I got a UEFI Samsung with serial(Samsung PM961) but I first did get a windows boot manager option as the Pata ssd1 would not work as boot option.I then installed(OS) but when reboot time came for install,I believe but not sure I had to choose windows boot manager. But if this is not showing up in your Bios or boot manager then look for a samsung reference. The biggest difference I noticed was when you cloned you disabled the ACHI and I only mention this because that communication protocol has caused me nightmares in the past on windows installs and if you changed the way you initially set it up I could see it hanging and crashing....I left mine set to ACHI enabled and native Ide on ports 4 and 5. No I do not think this will help you but I am pointing it out as maybe a possibilty.....Also if the bios is reporting a PATA SS I would think there would be a another to the drive if not under Windows boot manager or Samsung maybe some obscure reference.

Well, I followed the instructions and made a GPT-based Server 2008 Bootable USB drive with USB3.0 controller drivers, Samsung NVMe driver and KB2990941 and KB3087873 integrated using NTLite but no change from the originally reported problem when completely disabling CSM.

Yeah I don't know at this point, I think my motherboard might be screwed up or the Server 2008 R2 install is incompatible with the board's UEFI system and I don't expect an answer, it is just disappointing. The board itself does not support full secure UEFI mode when booting into a USB-based or DVD-based Server 2008 R2 install for whatever reason, maybe the BIOS and/or board is just crappy and nobody bothered to make it 100% right. During the Windows 7 colored logo 'coming together', the machine reboots and reports:

"The current BIOS settings do not fully support the Windows UEFI mode."
"The CSM (Compatibility Support Module) settings have been changed for better compatibility.

I already tried removing every component one by one, literally all except for the PCIe SM961 including the videocard (Used the onboard iGPU) producing the same error.
I know it is not the NvmExpressDxe_3.ffs module as I reflashed the BIOS with stock GRYPHON-Z87-ASUS-2103.CAP and got the same problem.

So I basically give up unless anyone has any possible reason why this error is happening or maybe knows if a previous or differently modified BIOS would be more compatible. I know it's unrelated to this topic. I assume the answer is simply that Server 2008 R2 doesn't support this particular board's full secure UEFI mode (and maybe others as well). Probably Windows 10 does but I have no means to try it and I do not want to anyway.

I will just stick with the PCIe SSD for use as a secondary program drive in Server 2008 R2 which it indeed does work with/without the BIOS mod (and KB2990941 and KB3087873 installed, of course). Thanks for all your help.

"https://social.technet.microsoft.com/Forums/en-US/a6679550-a343-463c-9301-1bf9706ad9af/howto-install-2008-r2-on-uefi-systems-using-gpt-disk-using-mdt-2010-with-update-1?forum=mdt" ......Don't know if you have already been here but if not might want to check it out. Also I was reading through your post and following your method and would suggest a few other ways....First I would use the USB 2.0 as the USB 3.0 is not native to the older Windows and seems to be a pain. I would use your NTlite to inject the hotfix's but then I would take that image and try MSI Win7 SmartTool because like you pointed out the Server 2008 R2 has that win 7 kernel so it might pick it up? if so then I would use it to create another ISO with the NVME drivers added there and then I would take that image to rufus and create your USB 2.0 GPT with UEFI install drive. The one thing I don't really have knowledge of is the CSM setting in Bios because I do not have that specific setting in my Bios but if your disabling it and things are not working maybe you should not disable it. these are just suggestions and I am trying to learn and maybe possibly even give something back along the way. This is Fernando's Bus and I am just a passenger :-)

Whoa thanks Harry! I didn't actually look too much into the Windows side of things, thank you. That post has basically exactly the issue in KB975535, in fact literally THE error and there is KB979374 too!

I will work on this for awhile, you have gave me new hope.

Hello, Fernando.

We have an interesting discussion on overclockers.ru about mention of the Variant c)

We have a several general questions:

- Is there ability of "cross-module" updating: Can user change an implemented MVMe module by manufacturer to the Variant c) module w/o having problems ?
What mean exactly: "which natively do support NVMe" ?

- What are the benefits of GUID header self-sufficiency ?

- Which are an essential steps, user can do to Define -> Search -> Update of the already existent module using UEFITool ?
Is "above 'Volume free space'" the right place for insertion in that case also? (But not the place of exisent module?)

I guess, first, user have to Search existent module in the "Text" column by "NvmExpressDxe" line.
Then delete the old one and insert the new one above 'Volume free space'. Right?

This info will help much to provide nice FAQ on the overclockers.ru forum for all users. Thank you!

Kohina definitely would like to know the results of how it turns out for you, I too am not a win 10 fan but I think eventually they will force us there. Look forward to hearing your positive results.

Sorry, I mean c) = "pure" (my bad)
But the question about "cross-module" updating of "pure" over "existent NVMe by manufacturer" is our main question.
Is here any problems being expected?

What are the benefits of: "without any GUID header" technically?
I guess, GUID doesn't matter, if I have already modded DxE Driver inside the BIOS and going to update "PURE" NVMe module using UEFITool within the *.efi body only:
>GUID... -> PE32 image section -> RMB -> Replace body... -> NvmExpressDxe_3.efi
The real important thing is updating 3rd-to-3rd Party Driver to newer/older version (permitted), not Manufacture-to-3rd Party Driver (restricted).
Right?

Zitat von Fernando im Beitrag #2860

---

Already existing DXE Driver modules have to be updated at their location.

---