Firmware test

I’ve been trying to get 11.0, and now 11.1 installed and stable on my machine for about a week an a half. I think I’ve reinstalled about a dozen times between the two versions. I keep running into weird problems like:

  • video problem (fixed by setting it to run in VESA)
  • hangs when trying to install various packaged (usually GNOME)
  • hangs on reboot
  • reboot to black screen
  • fail on reboot with various error messages (last time was unable to access /var)
  • unable to access NVIDIA repositories
  • etc.

I found out that my 11.0 download didn’t match the MD5 checksum, so that accounts for some of it. 11.1 does match and I ran the disk verification which passed. I ran memtest and that passes as long as I run it. I tried the firmware test and I got a lot of fail messages, but I don’t know what any of them mean. Can anybody give me any insight into these, or point me to a page with info (I did search and came up with nothing useful)?

[FAIL] OS/2 memory hole test
	F		The memory map has a memory hole between 15Mb and 16Mb
[FAIL] EDD Boot disk hinting
	F		Boot device 0x80 does not support EDD
[FAIL] DMI information check
	F	No SMBIOS nor DMI entry point found.
[FAIL] HPET configuration test
	F	Failed to locate HPET base
[WARN] CPU frequency scaling tests (1-2 mins)
	W	Frequency scaling not supported
[INFO] SVM Virtualization extensions
		Processor does not support Virtualization extensions
[INFO] Processor microcode update
		Kernel too old; the kernel does not export microcode version
		Kernel too old; the kernel does not export microcode version
[INFO] (experimental) DMA Remapping (Vt-d) test
		No DMAR ACPI table found.
[INFO] Battery tests
		No battery information present
[PASS] Validate assigned PCI resources
[PASS] MTRR validation
[PASS] PCI Express MaxReadReq tuning
[PASS] Fan tests
		Fan FAM status is:  on
[PASS] EBDA region
		EBDA region is correctly reserved in the E820 table.
[PASS] (experimental) APIC Edge/Level check
[PASS] General ACPI information
[PASS] DSDT AML verification
		Tested table DSDT.dsl
[PASS] SUM duplicate test
		Tested _SUM ids, successfully found no duplicates

I don’t know if it matters, but here’s my hardware info (I’m installing the 64-bit version of 11.1):
Asus A8N-SLI Deluxe motherboard
AMD Athlon 64 3200+
Corsair 1Gb dual-channel
NVIDIA GeForce 7300 video
Turtle Beach Santa Cruz
(I’ve disabled serial, parallel ports on mobo as well as onboard sound. I’m using the nforce4 onboard ethernet).

[FAIL] OS/2 memory hole test
F The memory map has a memory hole between 15Mb and 16Mb

I just looked at a BIOS, which has this as an option, disable it if you can, it may be just called OS/2 compatability or something, rather than memory hole.

[FAIL] EDD Boot disk hinting
F Boot device 0x80 does not support EDD
[FAIL] DMI information check
F No SMBIOS nor DMI entry point found.
[FAIL] HPET configuration test
F Failed to locate HPET base
[WARN] CPU frequency scaling tests (1-2 mins)
W Frequency scaling not supported

EDD, DMI & HPET - no worries. They dont’ have to work, you may get some warnings, or stern messages from these modules in the boot log though if you look.

CPU freq - fine, you can pass CPUFreq=no on the boot line, to prevent it trying. It’s for CPUs which can alter speed depending on how idle they are to save power.

[INFO] SVM Virtualization extensions
Processor does not support Virtualization extensions
[INFO] Processor microcode update
Kernel too old; the kernel does not export microcode version
Kernel too old; the kernel does not export microcode version
[INFO] (experimental) DMA Remapping (Vt-d) test
No DMAR ACPI table found.
[INFO] Battery tests
No battery information present

VIrtualistion and microcode update, you shouldn’t need. The microcode update, can help with buggy CPUs with bad steppings, it’s a field maintenance issue, to avoid repeats of the Pentium bugs for example. (Fdiv and F00F)

The DMA Remapping (experimental) is a newish feature, the Battery info, won’t matter unless you’re on a lap top.

[PASS] Validate assigned PCI resources
[PASS] MTRR validation
[PASS] PCI Express MaxReadReq tuning
[PASS] Fan tests
Fan FAM status is: on
[PASS] EBDA region
EBDA region is correctly reserved in the E820 table.
[PASS] (experimental) APIC Edge/Level check
[PASS] General ACPI information
[PASS] DSDT AML verification
Tested table DSDT.dsl
[PASS] SUM duplicate test
Tested _SUM ids, successfully found no duplicates

Your Firmware looks fine to me.

As for you install problems, are you overclocking CPU? If you haven’t run Linux before, it may be stressing the CPU by keeping it busier than Windows, and over-heating may result.

I had an NVidia install and it came up with the ‘nv’ opensource driver, which was fine for 2D work, my card is no longer supported by NVidia so 3D driver, is pretty irrelevant, unless I want to attempt compiling and porting the source interface part myself against a modern kernel… I can’t see Nouveau worrying about the card either.

Apart from trying safe settings and an install without the sound card, I’m not sure what else to suggest.

May be if you have IDE/PATA, rather than SATA around, trying to see if “lsmod | grep pata” turns up anything, by CNTRL-ALT-F2 -ing when the installer has been started. If so, attempt an install with the older IDE/ATA rather than experimental libata/pata driver by “brokenmodules=pata_xxx” depending on what lsmod showed.

Memory Hole in the BIOS is disabled. Weird. Anyway, thanks for validating my firmware test! :slight_smile:

I am not overclocking at all. I used to do that, but then I came to the conclusion that the loss of stability and work involved wasn’t justified by the performance increases.

I’m not having a problem installing as a result of not being able to connect to the NVIDIA repositories. Sorry, I guess I kind of gave that impression. After the install, when I’m trying to get my video drivers set up I’ve been unable to connect.

I haven’t tried it without the sound card. I may give that a shot, though the system I am using now has the same sound card and I’ve never had a problem with Suse on it.

I don’t follow you on the hard drive suggestion. I understand that by using a PATA drive it could identify possible problems with the drive/controller/drivers. However, why would I use the PATA drivers when I want to run SATA drives on this box?

I have made some progress toward figuring out my stability issues. Apparently, it may have something to do with my BIOS settings. Part of my earlier problems were fixed when I set my DVD-ROM drive to “LARGE” (as opposed to “OFF” as it was set, or “LBA”). Everything else for that drive and the SATA drive are set to “AUTO”. Any suggestions on what I should set my “Access Mode” to for my SATA drive? I’ve got it on “AUTO” now. IIRC, I don’t want to use HDD SMART Monitoring either. Is that correct? I’ve got it disabled now.

While I was playing with all these BIOS settings, I noticed that my memory timings were all set to “AUTO” and that is apparently not very stable with this Corsair TWINX-1024 3200C2 memory on this mobo. I’ve had problems with it before. I fixed the timings in the BIOS.

Now though, I’m getting the following every time I try to install:

Probing EDD (edd=off to disable)... ok

crc error

 -- System halted

So I thought maybe my DVD got messed up. I ran the media verification and got a kernel panic. Here’s what I got on screen (1024x768 jpg):
http://www.yippee38.net/images/Temp/error.jpg

That doesn’t look good. :frowning:

I updated my BIOS with the latest version (the version that was on the system was very old), and my problems seem to have evaporated.

Yep, the obsession with CPU speed in the PC press is incorrect. In real use, the CPU is probably spending most of it’s time waiting on memory access, disk transfers and user input.

I don’t follow you on the hard drive suggestion. I understand that by using a PATA drive it could identify possible problems with the drive/controller/drivers. However, why would I use the PATA drivers when I want to run SATA drives on this box?

You wouldn’t but you were asking a general question about stability and performance of 11.1. The info you gave, did not mean I could rule out PATA controller in your system, without doing additional research.

It’s more obvious to you what hardware you have, than to me, from the description you gave :slight_smile:

Anyway well done for sleuthing it down to the BIOS update, I wouldn’t expect that as the kernel doesn’t use the BIOS once it’s running, though I guess the ACPI code does get interpreted, and the hardware may benefit from better intitialisation. BIOS updates have got less pleasant, with them using Windows flash program, and under Vista I lost one Mobo (fortunately under support and replaced) due to Winflash stalling, and leaving Vista without it’s all important license key. That meant Winflash refused to run to fix, and furthermore Vista then intervened with a pop up, force rebooting the system, which was catastrophic, as no BIOS present to bring it back up!

For some reason, this motherboard and memory combo seem to be very tricky. If the memory timings aren’t just right, the system goes crazy. It’s still not perfect, but I’m working with the memory manufacturer now to get the right settings.

Friends don’t let friends use Vista. I just made a boot CD using Dr-DOS.

Quick additional thoughts . . .

The disk access mode setting should probably be LBA. Auto works with most bios’s, i.e., Auto should choose LBA. This is 48-bit LBA and required for (what several years ago were) large drives (i.e., >137GB).

You may get around that edd error (this is the kernel trying to read the video bios, and usually means that the bios is older or the implementation flakey), add this to the boot options:


edd=off