Linux Mint 13 UEFI not booting --HELP !!! {solved}

[powerhouse]

I'm really having a hard time with this.

I managed to install LM 13 RC but it won't boot. Before the grub menu appears the following message flashes shortly on the screen:

Code: Select all

grubx64.efi
Error: efidisk read error

I went to the command screen and added:

Code: Select all

grub> set pager=0
grub> set debug=all

ESC and selected the kernel to boot. What follows is some output as it processes the config file (insmod lvm etc.) and a long list of mmap EFI memory allocations and it gets stuck at:

Code: Select all

lib/relocator.c:1222: min_addr = 0x0, max_addr = 0xffffffffffffffff, target = 0x88800
lib/relocator.c:433: trying to allocate in 0x0-0xfffffffffffffffff aligned 0x1 size 0x7800
lib/relocator.c:529: Blocking at 0xa786f000-0xde5b9000

Note: Without the set options above I get a black screen without cursor and my keyboard LEDs flash. Only a hard reset makes the PC reboot.

I have the following partitions on my SSD:

/dev/sda1 vfat32 500M /boot/efi - this is the EFI partition
/dev/sda2 ext2 1000M /boot - this is the /boot partition
/dev/mapper/vol1_root 15G LVM partition for /
/dev/mapper/vol1_home 20G LVM partition for /home

During my installation from USB stick I chroot to the /mnt/, installed grub-efi and lvm2 and did the grub-install grub-update.

I also used efibootmgr to update the NVram.

So, in essence:
1. I managed to produce an UEFI bootable USB stick with Linux Mint 13 RC (this was the first hurdle)
2. I installed LM13 using LVM partitions for / and /home (yet another hurdle)
3. My motherboard sees the UEFI bootloader and loads grub, but after selecting the kernel from the grub menu the PC hangs.

Someone here https://bbs.archlinux.org/viewtopic.php?id=134553 wrote that the issue above is related to the grub2 memory allocator (see #17).

Any help is appreciated. This UEFI stuff is a nightmare.

[powerhouse]

Here my /boot/grub/grub.cfg file:

Code: Select all

#
# DO NOT EDIT THIS FILE
#
# It is automatically generated by grub-mkconfig using templates
# from /etc/grub.d and settings from /etc/default/grub
#

### BEGIN /etc/grub.d/00_header ###
if [ -s $prefix/grubenv ]; then
  set have_grubenv=true
  load_env
fi
set default="0"
if [ "${prev_saved_entry}" ]; then
  set saved_entry="${prev_saved_entry}"
  save_env saved_entry
  set prev_saved_entry=
  save_env prev_saved_entry
  set boot_once=true
fi

function savedefault {
  if [ -z "${boot_once}" ]; then
    saved_entry="${chosen}"
    save_env saved_entry
  fi
}

function recordfail {
  set recordfail=1
  if [ -n "${have_grubenv}" ]; then if [ -z "${boot_once}" ]; then save_env recordfail; fi; fi
}

function load_video {
  insmod efi_gop
  insmod efi_uga
  insmod video_bochs
  insmod video_cirrus
}

insmod lvm
insmod part_gpt
insmod ext2
insmod font
set root='(vol1-root)'
search --no-floppy --fs-uuid --set=root bd2b5bb7-b634-4d2b-a84a-cc3f90314ed5
if loadfont /usr/share/grub/unicode.pf2 ; then
  set gfxmode=auto
  load_video
  insmod gfxterm
  insmod part_gpt
  insmod ext2
  set root='(hd0,gpt2)'
  search --no-floppy --fs-uuid --set=root 97c075f0-0f4c-4591-8be7-3f4263d1e554
  set locale_dir=($root)/grub/locale
  set lang=en_US
  insmod gettext
fi
terminal_output gfxterm
if [ "${recordfail}" = 1 ]; then
  set timeout=-1
else
  set timeout=10
fi
### END /etc/grub.d/00_header ###

### BEGIN /etc/grub.d/05_debian_theme ###
set menu_color_normal=white/black
set menu_color_highlight=black/light-gray
### END /etc/grub.d/05_debian_theme ###

### BEGIN /etc/grub.d/06_mint_theme ###
set menu_color_normal=white/black
set menu_color_highlight=white/light-gray
### END /etc/grub.d/06_mint_theme ###

### BEGIN /etc/grub.d/10_linux ###
function gfxmode {
	set gfxpayload="$1"
	if [ "$1" = "keep" ]; then
		set vt_handoff=vt.handoff=7
	else
		set vt_handoff=
	fi
}
if [ ${recordfail} != 1 ]; then
  if [ -e ${prefix}/gfxblacklist.txt ]; then
    if hwmatch ${prefix}/gfxblacklist.txt 3; then
      if [ ${match} = 0 ]; then
        set linux_gfx_mode=keep
      else
        set linux_gfx_mode=text
      fi
    else
      set linux_gfx_mode=text
    fi
  else
    set linux_gfx_mode=keep
  fi
else
  set linux_gfx_mode=text
fi
export linux_gfx_mode
if [ "$linux_gfx_mode" != "text" ]; then load_video; fi
menuentry 'Linux Mint 13 MATE 64-bit, 3.2.0-23-generic (/dev/sda2)' --class linuxmint --class gnu-linux --class gnu --class os {
	recordfail
	gfxmode $linux_gfx_mode
	insmod gzio
	insmod part_gpt
	insmod ext2
	set root='(hd0,gpt2)'
	search --no-floppy --fs-uuid --set=root 97c075f0-0f4c-4591-8be7-3f4263d1e554
	linux	/vmlinuz-3.2.0-23-generic root=/dev/mapper/vol1-root ro   quiet splash $vt_handoff
	initrd	/initrd.img-3.2.0-23-generic
}
menuentry 'Linux Mint 13 MATE 64-bit, 3.2.0-23-generic (/dev/sda2) -- recovery mode' --class linuxmint --class gnu-linux --class gnu --class os {
	recordfail
	insmod gzio
	insmod part_gpt
	insmod ext2
	set root='(hd0,gpt2)'
	search --no-floppy --fs-uuid --set=root 97c075f0-0f4c-4591-8be7-3f4263d1e554
	echo	'Loading Linux 3.2.0-23-generic ...'
	linux	/vmlinuz-3.2.0-23-generic root=/dev/mapper/vol1-root ro recovery nomodeset 
	echo	'Loading initial ramdisk ...'
	initrd	/initrd.img-3.2.0-23-generic
}
### END /etc/grub.d/10_linux ###

### BEGIN /etc/grub.d/10_lupin ###
### END /etc/grub.d/10_lupin ###

### BEGIN /etc/grub.d/20_linux_xen ###
### END /etc/grub.d/20_linux_xen ###

### BEGIN /etc/grub.d/20_memtest86+ ###
menuentry "Memory test (memtest86+)" {
	insmod part_gpt
	insmod ext2
	set root='(hd0,gpt2)'
	search --no-floppy --fs-uuid --set=root 97c075f0-0f4c-4591-8be7-3f4263d1e554
	linux16	/memtest86+.bin
}
menuentry "Memory test (memtest86+, serial console 115200)" {
	insmod part_gpt
	insmod ext2
	set root='(hd0,gpt2)'
	search --no-floppy --fs-uuid --set=root 97c075f0-0f4c-4591-8be7-3f4263d1e554
	linux16	/memtest86+.bin console=ttyS0,115200n8
}
### END /etc/grub.d/20_memtest86+ ###

### BEGIN /etc/grub.d/30_os-prober ###
if [ "x${timeout}" != "x-1" ]; then
  if keystatus; then
    if keystatus --shift; then
      set timeout=-1
    else
      set timeout=0
    fi
  else
    if sleep --interruptible 3 ; then
      set timeout=0
    fi
  fi
fi
### END /etc/grub.d/30_os-prober ###

### BEGIN /etc/grub.d/40_custom ###
# This file provides an easy way to add custom menu entries.  Simply type the
# menu entries you want to add after this comment.  Be careful not to change
# the 'exec tail' line above.
### END /etc/grub.d/40_custom ###

### BEGIN /etc/grub.d/41_custom ###
if [ -f  $prefix/custom.cfg ]; then
  source $prefix/custom.cfg;
fi
### END /etc/grub.d/41_custom ###

Here the output of blkid (running a live USB stick):

Code: Select all

/dev/loop0: TYPE="squashfs"
/dev/sda1: UUID="91E9-8D10" TYPE="vfat" 
/dev/sda2: UUID="97c075f0-0f4c-4591-8be7-3f4263d1e554" TYPE="ext2" 
/dev/sda3: UUID="Dh1Xhg-EBuW-Ax53-TmN8-hN0o-oFlx-lFhZj8" TYPE="LVM2_member" 
/dev/sdb1: UUID="2K8dz0-XJXp-Dytx-rUN6-ZS14-Cq0A-aUnqJM" TYPE="LVM2_member" 
/dev/sdc1: UUID="3EED-0D84" TYPE="vfat" 
/dev/mapper/vol1-root: UUID="bd2b5bb7-b634-4d2b-a84a-cc3f90314ed5" TYPE="ext4" 
/dev/mapper/vol1-home: UUID="dac3b695-c06c-45d3-bbba-294cf3d9547c" TYPE="ext4" 
/dev/mapper/vol2-photos: LABEL="photos" UUID="e3f37639-c90a-449b-9232-f2965d432599" TYPE="ext4"

[srs5694]

I've never seen that error before, and I've not studied the GRUB code; however, it sounds as if this is either a GRUB memory management bug or a sign of damaged RAM. To rule out the second possibility, I suggest running some memory diagnostics. Several such tools are available. I don't recall if Mint's installer ships with one, but some Linux distributions do, selectable as an option alongside the various installation options (run from the boot loader). You might need to boot the installation medium in BIOS mode, rather than in EFI mode, to get access to these utilities.

If this is a sign of a GRUB bug, then upgrading GRUB might fix it, but you might need to compile GRUB from source code. The Ubuntu UEFI Booting page describes how to do this. An option that's more likely to work is to switch to a different boot loader entirely. My recommendation would be either the Linux kernel's EFI stub loader (which requires a 3.3.0 or later kernel) or ELILO. See my EFI Boot Loaders page for more information on both of these, and others. If you're dual-booting with Windows, using either of those boot loaders will require another boot manager to transfer control to Windows. The boot manager built into your firmware might be adequate, but if not, my rEFInd is probably the way to go. (GRUB Legacy and GRUB 2 can play this role, too, but of course if you're going down this path because of a GRUB 2 bug, that one's out of the running.) In fact, if you use the kernel's EFI stub loader, rEFInd can be useful for selecting which kernel to boot and what options to use (enabling single-user mode or the like).

[powerhouse]

Thanks Rod (that is your name?)!

That's what I've found too in another forum - grub memory management issue.

I haven't checked my RAM yet, but I will try to do that. Booting live USB works without issues at all, and since I managed to install LMDE in the past I didn't think it's related to RAM. I will try to rule that out and report back.

[powerhouse]

I forgot, I also installed rEFInd. The problem is that LM13 comes with a 3.2.0 kernel and I don't know how to use rEFInd to boot the kernel directly.

Since this is most likely a grub issue, I would be glad to find an alternative that works in my installation (with LVM root partition !!!). Any ideas? I read your website and you suggested ELILO, but I can't see how to make this work with LVM. I simply haven't found anything about loading the LVM module via ELILO.

[powerhouse]

Just for the fun of it, or to see if it is any better, I installed the final LM 13 release. Same error, same problem.

This is going nowhere.

Is there anyone out there who successfully installed LM13 on an UEFI disk with grub2 (grub-efi)? Perhaps also using LVM for /?

LM13 is using grub 1.99 and I read somewhere (see link above) that this version may have a memory problem. I checked my RAM in the BIOS and it reports OK. Also, everything works fine from a live USB stick boot (also UEFI).

The Ubuntu forum suggests to compile grub2 with UEFI from source - they can't be serious. If grub UEFI support is in such bad shape, I strongly suggest to find a better alternative. Unfortunately grub2 has become the default bootloader in almost all Linux distribution - I can't get it.

Sorry, I had to get 3 weeks frustration off the chest.

[srs5694]

The Ubuntu forum suggests to compile grub2 with UEFI from source - they can't be serious. If grub UEFI support is in such bad shape, I strongly suggest to find a better alternative. Unfortunately grub2 has become the default bootloader in almost all Linux distribution - I can't get it.

My own opinion of GRUB 2 is pretty low. IMHO, GRUB 2 tries to be too many things to too many people, and in the end this creates a lot of problems, particularly when moving away from the core function of booting Linux on BIOS-based computers. Of relevance for your problem, GRUB 2's EFI support is unreliable. (OTOH, some people have no problems with it. This seems to reflect the poor state of EFI consistency -- there are a lot of bugs in EFI, and they vary from one implementation to another.) Unfortunately, EFI is still pretty "bleeding-edge" stuff. Too many companies and organizations have put off working on it for too long, so now that it's appearing on large numbers of computers, the problems that should have been ironed out five years ago are appearing on end users' systems.

Anyhow, you can use ELILO. Set it up as described here. Don't worry about LVM; you'll need to copy the Linux kernel and initial RAM disk to the ESP in any event (at least for the simplest setup), so there's nothing special about configuring it for LVM. You'll just need to pass an LVM device identifier as the "root=" option. If you want to make for more flexibility moving forward, you could set up a Linux /boot partition (preferably using ReiserFS, ext2fs, or ext3fs on it) outside of the LVM. That way, you can use an EFI filesystem driver to give the EFI access to the /boot partition, and ELILO will be able to load the kernel from there. In the future, or if you compile your own kernel or can find a precompiled 3.3.x or 3.4.x kernel for Mint, you could use the EFI stub loader directly from the /boot partition.

[powerhouse]

Thanks for the quick reply. I will try ELILO then and see how it works. I've nothing to loose. I'll report back with the results.

I'm a bit worried about it since I need to install Xen kernel and Nvidia proprietary driver which is also modifying the kernel. Hope to be able to make that bootable.

[powerhouse]

Rod, another question regarding ELILO. Somewhere I read that ELILO only takes compressed kernel files, ie vmlinuz or bzImage... Is this correct?

Since LM 13 gives me the regular kernel I would have to produce the compressed kernel files. Is that a simple tar or whatever procedure?

Last not least, how do I specify a LVM root directory in the elilo.conf file? Just root=/dev/mapper/vol1-root ?

Or is the meaning of root actually the /boot directory where the modules etc. reside?

How does ELILO find the LVM module and others to load when they are usually present at /boot/grub/...?

It somehow perplexes me how this can work.

[srs5694]

Rod, another question regarding ELILO. Somewhere I read that ELILO only takes compressed kernel files, ie vmlinuz or bzImage... Is this correct?

I'd not heard that, but....

Since LM 13 gives me the regular kernel I would have to produce the compressed kernel files. Is that a simple tar or whatever procedure?

I haven't yet installed LM13, but I suspect you're mistaken about this. Every distribution I've ever used, including LM12, has delivered compressed kernels.

Last not least, how do I specify a LVM root directory in the elilo.conf file? Just root=/dev/mapper/vol1-root ?

Yes, although of course the precise value will depend on how your LVM is configured.

Or is the meaning of root actually the /boot directory where the modules etc. reside?

Not for the "root=" line in elilo.conf; that refers to the Linux root (/) filesystem.

How does ELILO find the LVM module and others to load when they are usually present at /boot/grub/...?

Necessary kernel modules go in the initial RAM disk file, which you must also copy to the ESP and reference on the elilo.conf "initrd=" line. Those files do not reside in /boot/grub; that directory holds GRUB's drivers, configuration files, and so on. If you're using ELILO, those files become 100% irrelevant.

It somehow perplexes me how this can work.

Whether booting via BIOS or EFI, and no matter what boot loader is used, the boot process involves the boot loader doing three critical things:

1. Loading the kernel file from hard disk to memory.
2. Loading the initial RAM disk file from hard disk to memory.
3. Executing the kernel, including passing it any options it requires.

ELILO does the first two tasks by using EFI functions for reading files from filesystems, hence the need to put those two files on the ESP (or on some other partition that the EFI can read). GRUB includes its own filesystem drivers, and so can load those files from a Linux partition even if the EFI can't read that partition; but this increases GRUB's complexity and means that it needs support files in /boot/grub. Otherwise what these two boot loaders do is pretty similar, at least when considered broadly.

[powerhouse]

Thanks for the in-depth explanations - I'm learning every day. I hope to give ELILO a try later today.

[powerhouse]

Ok, I installed ELINO from the LM13 repository (while in chroot environment with all partitions mounted). During the install a menu appeared with the options "yes" or "no" to automatically install ELILO in its right place. I chose yes but the autoinstaller quit saying no /etc/elilo.conf file found.

Looking at the script files that were installed I saw that this ELILO package was geared for Ubuntu, for example it would install in (perhaps even expect) the /boot/efi/EFI/ubuntu folder.

So I manually created the elilo.conf file based on srs... (Rod) website and copied it into the /boot/efi/EFI/ubuntu folder, together with all the kernel-related files at /boot, specifically vmlinuz... and the initrd.img... files.

I used efibootmgr to add the new ELILO boot loader to the system.

Booting the PC and selecting the BIOS boot manager I found the ELILO entry and ran it. I got a very short message of something missing (not really sure) but the boot... of ELILO appeared. I pressed enter and - I suddenly got the grub boot menu. Trying to boot from that menu led to the usual black screen of grub and the PC hang.

I'm wondering if I ever get this PC work with Linux.

[powerhouse]

I tried rebooting with ELILO several times to see the error message - I actually filmed it to be able to read it.

It printed efidisk: read error

Then I used rEFInd and selected the ELILO boot loader which gave me finally a clear message:

Starting elilo.efi

ELILO boot: .....
elilo.c(line 79) :Kernel file not found vmlinuz-3.2.0-23-generic
Error: Load error returned from elilo.efi

Here are the files I have in /boot/eli/ELI/ubuntu:
abi-3.2.0-23-generic
config-3.2.0-23-generic
elilo.conf
elilo.efi
initrd.img-3.2.0-23-generic
System.map-3.2.0-23-generic
vmlinuz-3.2.0-23-generic

I'll copy them again using cp -a and not, as I did, the drag and drop. Hope it makes a difference.

Here is the elilo.conf file:

Code: Select all

prompt
timeout=50
default=LM13
#chooser=textmenu

image=vmlinuz-3.2.0-23-generic
        label=LM13
        initrd=initrd.img-3.2.0-23-generic
        read-only
        root=/dev/mapper/vol1-root
        append=""

[powerhouse]

I copied the files again using cp -a - makes no difference at all.

I am at a loss.

Perhaps my method of trying to install LM13 directly onto LVM doesn't work.

Oh well, I will have to try yet another way of getting this to work. I will remove LVM and install the system on partitions - the straight way. Hope that this gives me a bootable system.

[powerhouse]

So, I booted LM13 from USB stick (in UEFI mode), wiped my SSD clean and created a GPT type disk. I then ran the installer and partitioned:

/dev/sda1 EFI 512MB
/dev/sda2 ext2 1024MB as /boot
/dev/sda3 ext4 15GB as /
/dev/sda4 ext4 20GB as /home

Having 32GB RAM I didn't see a need for a swap partition at this point, particularly not on an SSD. I reserved 8GB on a regular HDD for swap to be included later on.

I let the installer continue to install and rebooted.

1st time around it booted into my new LM13, but complained about a CRC error in a theme picture file. When the desktop appeared the Mint Menu was missing. Tried to add it to the panel but it gave me an error (some OA... error). The other Linux menu options worked, though.

I continued to update the system, installed the latest kernel from the repository, added Xen Hypervisor. When rebooting into the latest kernel, it hang. When rebooting into the Xen hypervisor - dito. Only the original kernel would boot. Still with the CRC error and the missing Mint Menu.

When installing the other kernels I noticed some US-UTF... not present error.

I decided to reinstall again. This time it didn't even get to the grub loader - it complained about characters not in the right order or so (that was the weirdest message I ever saw on Linux).

I downloaded again LM13 (perhaps the new download includes some fixes, or perhaps my live USB stick had errors) and created a new live USB stick, just to make sure it's not a problem with my original live USB stick.

Will report how this works.

[eanfrid]

I didn't mess with UEFI yet but,

When installing the other kernels I noticed some US-UTF... not present error.

My 2 cents: I think this is only a cosmetic "error". I get the same kind of message with my own locale (fr_FR) while installing a new kernel and updating grub. No further consequences

[srs5694]

A CRC error is troubling. That suggests a problem with your memory or storage. I recommend you run hardware diagnostics to test both of them, such as a memory-testing program and a SMART utility for your disks.

[powerhouse]

I managed to install LM13 the regular way - taking the whole disk, with UEFI /boot/efi (I believe). It boots OK but after some updates the troubles start again.

I tried to run the memtest at boot, but it gives me an error Linux16 not found.

I then installed memtester and ran that on most of the memory (32Gig). To speed up the process I ran 10 instances (this way it uses 10 CPU threads). I ran that three times, with reboot in between.

The first run showed an error at a specific memory address. The second was clean, the third one showed again an error at another address, but when I stopped the instance and restarted it it was clean.

I don't know what to make of that, but considering the erratic behavior with synaptic segfaulting from time to time I suspect there may be a hardware issue. I downloaded memtest86 USB - both regular and server edition - but the USB stick won't boot, just hangs.

I consider turning in the PC for repair. Unfortunately I built it myself from components (that I checked for compatibility via the vendor websites, particularly the ASUS website for the motherboard). I also don't have a Windows 7 Professional edition to test that, at least not yet. I wanted to make Linux and Xen work and then install Win7 as a domU.

Ah, I also got a drive error on one of my reboots (fixed by fsck). So there could also be an issue with the SSD.

This is the first time ever that I experienced so many issues at once. What perplexes me is that the live USB boots and runs smoothly, without the slightest issue. The troubles only start once I install the system on SSD.

My last option, before turning in the PC, is to install on my regular hard drive to see if this is a SSD problem. Before doing that I will try once more to get memtest to work at boot from USB or CD.

The motherboard, an Asus Sabertooth X79, is at the default settings except VT-d enabled and some monitoring tweaks for the CPU fans (they run at very low rpm, ~400rpm, and cause an error at boot). The motherboard has also the latest firmware.

Any suggestions?

[srs5694]

Running multiple instances of a memory tester strikes me as questionable; however, I'm not an expert on this area, so maybe it's OK or even beneficial. I'd try to get a low-level (pre-boot) memory tester to run and just let it go for a while. See if you can find some firmware option to disable EFI-mode booting or enable legacy (BIOS-mode) booting. That might enable you to launch a BIOS-based memory tester.

Removing the SSD from the equation by installing directly to a regular hard disk is also worth trying. You could also see if there are any errors on a SMART test of the drive. Another thing to check along these lines is your SATA cable; a bad cable can cause no end of problems.

[powerhouse]

I finally got memtest86+ installed and boot from USB stick. I ran it for nearly 3 hours, half through the second pass - no errors whatsoever (a disadvantage of having 32Gig mem). So I assume the memory is OK.

I also ran smartctl to see the SSD status and any errors or relocations. It output looked fine - no errors, no relocates, and life time at 100.

Though by then LM13 booted via UEFI and was more or less problem free, I dared not to continue my journey to install Xen etc. My previous attempts left me with an unusable system.

So I got the Ubuntu 12.04 alternate AMD64+mac CD image and prepared a USB stick. It didn't boot via UEFI (couldn't find any UEFI stuff on the stick), so my first install failed with fatal grub error.

The second time I wiped the SSD clean and created a new MBR partition table. I then installed Ubuntu with /boot ext2, and the rest of the drive as a LVM volume group with / and /home as separate logical volumes. It installed fine and booted.

Then Firefox crashed. I saw there were some 170 updates, so I had the updates installed in the hope it might fix the Firefox issue. But the update exited halfway through with an error and a hint to use apt-get to fix the issues. The usual apt-get clean and apt-get update didn't do much good. Ubuntu seems to display similar issues as with LM13.

In between these installation attempts I ran Ubuntu 12.04 desktop from a live USB stick. As with LM13, it worked just fine.

The only thing left to check is the SSD. Has anyone had similar difficulties with an SSD? Any suggestions on how to identify a faulty SSD? I use a Sandisk Extreme 120GB SSD.

I hate to use my regular HD for the OS just to see if it's an SSD issue.

For completeness, here is my hardware:

Asus Sabertooth X79 motherboard with latest BIOS firmware
Intel i7 3930K with C2 stepping and VT-d support
PNY Quatro 600 Nvidia graphics adapter
32 GB 1600 memory (can't remember the brand, but it was approved by Asus for this motherboard)
Sandisk Extreme 120G SSD
Western Digital Green 2TB HD
Seasonic 650W Gold power supply
Corsair 500 chassis and a huge dual-fan CPU cooler (can't remember the name, but it's designed for overclocking)

All motherboard settings are at default, except VT-d enabled and some bit protection disabled (I changed it at one point since this setting requires an OS that supports it - Windows does, so do some Linux kernels/distributions - I didn't want to take a chance and disabled it).

I checked the Internet before I purchased the hardware and saw that the motherboard was listed as Ubuntu 12.04 and even 11... compatible. My drives are connected to the Intel SATA ports, not the additional Marvel SATA ports (I somewhere found some recommendation for doing so). When it boots, boot-up time can be measured in seconds (around 25s from pressing the button to a graphical desktop).