NVIDIA Chipset DMA/SATA HD Corruption Revisited, CHIPSET FAILS WITH IDE/SATA AGAIN Options

[gadLinux]

Hi all!

I recently updated my computer to a AMD dualcore with an ASUS mobo A8N-SLI and worked well for a while...
Then the hell came...

I detected a HD disk problem with the motherboard:

---------------------------
Sep 12 20:14:21 localhost kernel: ata3: status=0x51 { DriveReady SeekComplete Error }
Sep 12 20:14:21 localhost kernel: ata3: error=0x84 { DriveStatusError BadCRC }
Sep 12 20:14:21 localhost kernel: ata3: status=0x51 { DriveReady SeekComplete Error }
Sep 12 20:14:21 localhost kernel: ata3: error=0x84 { DriveStatusError BadCRC }
Sep 12 20:14:21 localhost kernel: ata3: status=0x51 { DriveReady SeekComplete Error }
Sep 12 20:14:21 localhost kernel: ata3: error=0x84 { DriveStatusError BadCRC }
Sep 12 20:14:21 localhost kernel: ata3: status=0x51 { DriveReady SeekComplete Error }
Sep 12 20:14:21 localhost kernel: ata3: error=0x84 { DriveStatusError BadCRC }
Sep 12 20:14:21 localhost kernel: sd 2:0:0:0: SCSI error: return code = 0x8000002
Sep 12 20:14:21 localhost kernel: sdb: Current: sense key=0xb
Sep 12 20:14:21 localhost kernel: ASC=0x47 ASCQ=0x0
--------------------------

I lost a lot of information. The disk failed to start and cannot boot into the system. I had to reinstall the OS.

I thought that this was a disk problem or the disk crashed. So I brought two new disks and tried to set it into a Raid-1 mode.

I started my gentoo live cd 2006.1 and prepared to do the raid array with dodmraid at boot so I can use the nvidia Raid BIOS facility.

Every went right until my mkfs.ext3 command. One of the disk started to do the same errors as before...
Damn! Something should be wrong...

I restarted the computer and start the proccess from the begining.

The same error repeated with the second disk at other point of the mkfs command.

Then I thought that something serious occurs and tried to set the disk into raid without nvidia BIOS facilities. And everythink when ok until some point of the copy of files...

Same error appeared.

I wrote ASUS, both disk manufacturers (Maxtor (second ones), and Seagate). The only answers I found was to upgrade BIOS (like always), install new drivers, and try hdutilies with the disks.

Everything of this was done with the same results.

Exausted, I returned the motherboard to the store I brought it. And I decided to try another MOBO but the only available with my requeriments were nvidia based. I thought that it's a chipset problem so brought a A8N-SLI Premium and a Thermaltake 640W Power Supply because it has another chip for the SATA interface (SIL one).

This way I can check what vendor told me about missing power supply into the motherboard (I previously brought a 500W one) but just to be sure...

Result=289,64� more spent in the computer...

Well I attached my RAID array to the SIL controller the configuration of my system goes as follow:

IDE Interface:
Sep 19 09:32:04 localhost kernel: hda: IC35L080AVVA07-0, ATA DISK drive
Sep 19 09:32:04 localhost kernel: hdb: PIONEER DVD-RW DVR-108, ATAPI CD/DVD-ROM drive
Sep 19 09:32:04 localhost kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
Sep 19 09:32:04 localhost kernel: hdc: ST3160021A, ATA DISK drive
Sep 19 09:32:04 localhost kernel: hdd: PLEXTOR CD-R PX-W4824A, ATAPI CD/DVD-ROM drive

Sata:
One Barracuda 7200.8 200GB (Not currently connected, this was the first disk failed. Now it works properly...)

SIL:
Sep 19 09:32:04 localhost kernel: scsi3 : sata_sil
Sep 19 09:32:04 localhost kernel: Vendor: ATA Model: Maxtor 6V200E0 Rev: VA11
Sep 19 09:32:04 localhost kernel: Type: Direct-Access ANSI SCSI revision: 05
Sep 19 09:32:04 localhost kernel: Vendor: ATA Model: Maxtor 6V200E0 Rev: VA11
Sep 19 09:32:04 localhost kernel: Type: Direct-Access ANSI SCSI revision: 05


This does a total of 5 hard drives, one CD RW, and one DVD RW.

Well it worked like a charm for 5 days but today it break away. AGAIN!!!!

First time the system hanged into the disk and memory information screen at boot.
Rebooted and happened the same.

Rebooted again and set the boot disk to the CD. It booted linux x64 and tried fsck on the disk. The disk was clean so I tried with the -f (force) setting. It started but:

The ide disk hdc failed with dma error unfortunately I have not the error to post here, but seemed to be almost the same as above but with DMA data transfers...

I used hdparm to deactivate dma transfers. Result: Worked well for the check.

Rebooted again and the system hanged.

I entered the BIOS deactivated all dma stuff and rebooted again.

The system failed on Linux boot.

Rebooted again with flags at boot "ide=nodma"

Now it's working. slow, but working.

I have to say that. These problems has nothing to do with disks. The disks are working perfectly. A surface scan says that. No problems on files, no problems on filesystem.

NVIDA, ASUS and other manufacturers, please:

I REALLY DON'T UNDESTAND WHAT'S GOING ON WITH THESE MOBOS
WE ARE SPENDING A LOT OF TIME AND MONEY TO JUST RUN OUR O.S.
I SWITCHED AMD AND NVIDIA SOME TIME AGO BUT ME AND OTHERS SURELY WILL SWITCH AGAIN TO INTEL AND OTHER CHIPSET MANUFACTURES IF THESE PROBLEMS DOES NOT GET ATTENTION.

I SAW PEOPLE SAYING PROBLEMS ABOUT DATA CORRUPTION, DISK FAILLURES AND OTHERS

THIS IS SERIOUS!!!

AN ANSWER LIKE, PLEASE UPDATE YOUR DRIVERS IS NOT SOLUTION!!!

WE WANT TO KNOW UNDER WHAT CIRCUMSTANCES THIS HAPPENS, WHY AND HOW TO GET AROUND THE PROBLEM.

IF THE PROBLEM HAS NOT SOLUTION JUST SAY: UPGRADE TO "THIS" MOTHERBOARD, THIS WORKS FLAWLESSLY


Thank you for your support