[Solved]Dual Athlon-MP on AMD760 MP - fs corruption with IDE

[tarwin]

Hello,

i bought an dual socket A board to be my gentoo package crunsher but i can't even install it properly... (or so i thought till today afternoon. (read on))

So i got this board, Tyan MP Thunder with two Athlon MP 1.8+ and a Samsung IDE-Drive. I want to install gentoo on it. No problem, i installed gentoo lots of times, i think and start installing it. Install (Stage3) runs smooth, but after the first reboot: FS Errors. I tried fsck, it can fix it. After reboot: More errors. I tried reiserfs and ext3: after every reboot i got "bad leafs" or broken inodes... After the third rebootor so, the fs is totaly wasted.

I thought ACPI is the problem, so i disabled it. Didn't work. Then i disabled DMA. Didn't work. I build in a promise IDE-Controller i had arround and installed over it. Didn't work. Today i tried an old 4,3 GB SCSI drive on an Adaptec SCSI HBA. Did work! No errors. I dont get it!? There are lots of people out there who use this board as linux box, and, i forgot to mention earlier, Windows 2k runs perfectly (as perfect as it can "work" ) !!!!

I tried copying/installing the working system from the SCSI drive on IDE but failed as usual with a corrupt fs.

I tried kernel 2.4.20-r7 and 2.6.8-r4. No difference.

In the dmesg is on thing: "I/O APIC: AMD Errata 22 may be present. In the event of instability try booting with the "noapic" option."
So i booted with noapic, the only thing that changed was that the first cpu has now all interrupts (what is exactly what the option should cause).

Here some other highlights from the dmesg:

mtrr: v2.0 (20020519)
mtrr: your CPUs had inconsistent fixed MTRR settings
mtrr: probably your BIOS does not setup all CPUs.
mtrr: corrected configuration.
ACPI: Subsystem revision 20040326
ACPI-0352: *** Error: Looking up [Z00Q] in namespace, AE_NOT_FOUND
search_node dff3a520 start_node dff3a520 return_node 00000000
ACPI-1133: *** Error: Method execution failed [\_SB_.PCI0.ISA_.SIO_.COM1._STA] (Node dff3a520), AE_NOT_FOUND
ACPI-0352: *** Error: Looking up [Z00Q] in namespace, AE_NOT_FOUND
search_node dff3a3a0 start_node dff3a3a0 return_node 00000000
ACPI-1133: *** Error: Method execution failed [\_SB_.PCI0.ISA_.SIO_.COM2._STA] (Node dff3a3a0), AE_NOT_FOUND
ACPI-0352: *** Error: Looking up [Z00Q] in namespace, AE_NOT_FOUND
search_node dff3a1a0 start_node dff3a1a0 return_node 00000000
ACPI-1133: *** Error: Method execution failed [\_SB_.PCI0.ISA_.SIO_.LPT_._STA] (Node dff3a1a0), AE_NOT_FOUND
ACPI: Interpreter enabled
ACPI: Using IOAPIC for interrupt routing
ACPI: PCI Root Bridge [PCI0] (00:00)
PCI: Probing PCI hardware (bus 00)
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.AGP_._PRT]
ACPI: PCI Interrupt Link [LNKA] (IRQs 3 5 10 11) *9
ACPI: PCI Interrupt Link [LNKB] (IRQs 3 5 10 11) *0, disabled.
ACPI: PCI Interrupt Link [LNKC] (IRQs 3 5 *10 11)
ACPI: PCI Interrupt Link [LNKD] (IRQs 3 5 10 *11)
ACPI-0352: *** Error: Looking up [Z00Q] in namespace, AE_NOT_FOUND
search_node dff3a520 start_node dff3a520 return_node 00000000
ACPI-1133: *** Error: Method execution failed [\_SB_.PCI0.ISA_.SIO_.COM1._STA] (Node dff3a520), AE_NOT_FOUND
ACPI-0352: *** Error: Looking up [Z00Q] in namespace, AE_NOT_FOUND
search_node dff3a3a0 start_node dff3a3a0 return_node 00000000
ACPI-1133: *** Error: Method execution failed [\_SB_.PCI0.ISA_.SIO_.COM2._STA] (Node dff3a3a0), AE_NOT_FOUND
ACPI-0352: *** Error: Looking up [Z00Q] in namespace, AE_NOT_FOUND
search_node dff3a1a0 start_node dff3a1a0 return_node 00000000
ACPI-1133: *** Error: Method execution failed [\_SB_.PCI0.ISA_.SIO_.LPT_._STA] (Node dff3a1a0), AE_NOT_FOUND
Linux Plug and Play Support v0.97 (c) Adam Belay

....

udf: registering filesystem
BIOS failed to enable PCI standards compliance, fixing this error.
I/O APIC: AMD Errata #22 may be present. In the event of instability try
: booting with the "noapic" option.

.....

AMD7411: IDE controller at PCI slot 0000:00:07.1
AMD7411: chipset revision 1
AMD7411: not 100% native mode: will probe irqs later
AMD7411: 0000:00:07.1 (rev 01) UDMA100 controller
ide0: BM-DMA at 0xf000-0xf007, BIOS settings: hda:DMA, hdb:pio
ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:pio, hdd:pio


While browsing throu the dmesg output i've desided to build a kernel without acpi, mtrr and with a generic ide driver and try this out.

I'm really frustated with this problem... I'm sitting in front of it for 3 weeks now (no not full time ) and am ready to kick the box out of the window. Today, a better-than-nothing sollution came with the SCSI drive, but i need more space than 4.3 GB and i'm afraid to access my normal IDE data disks with an os that eats partitions...

Please help me!

Tarwin

[rush_ad]

fast system. (just wanted to let you know)

[tarwin]

Ok i found a solution.
I have to take out the ide-contoller kernel-driver for AMD and plug my promise controller in.
Than i can use ide without errors. If i use the promise with the AMD drivers in the kernel, the filesystem gets wasted...

Really strange. But what the heck... it works!!!

Yay gentoo! (And yes... its fast! )

[pharaoh]

That's weird...I have a Thunder K7 and I get most of your same errors but never had a problem with them. Glad to hear you're up and running though!
_________________
RYZEN 5 3600 Matisse (Zen 2) 6-Core 3.6 GHz Socket AM4 65W
ASRock B550M PRO4
Crucial Ballistix 3200 MHz DDR4 DRAM 16GB
EVGA GeForce GTX 1060 6GB

[trapine]

It's interesting that someone else has had problems with the MP boards. I consistantly had issues with IDE drives plugged into the onboard controllers. I eventually put in a SATA card which has (thus far) elliminated my problems.

[GentooNewbie]

What type of mouse are you using? I have 2 MSI K7D Master-L boards (MPX chipset), one with 1.4Ghz and the other with 2.0Ghz. I struggled with filesystem corruption with this board for a long time. Until I stopped using a USB mouse and started using a PS/2 mouse. I have been rock solid and stable on the machines ever since.

[trapine]

I'm using an MS explorer mouse - USB->PS/2 adapter to a KVM switch.
I wonder if I should disable USB on the motherboard.

[GentooNewbie]

I can still use the onboard USB for joysticks with out a problem. Just have to have a mouse plugged into the PS/2 port. You *should* be OK with you setup. I too am using a MS explorer mouse - USB->PS/2 adapter. Can't hurt to try plugging your mouse directly into the PS/2 port and skip the switch.

[GentooNewbie]

http://www.canopus22.demon.co.uk/links/dualamd.htm

AMD 766 Peripheral Bus Controller Revision Guide & Errata Section 2.22 on Page 25 of this 228KB pdf document says:


Multiprocessor System May Hang While in FULL APIC Mode and IOAPIC Interrupt is Masked

Note: No failure has been observed when using a PS/2 mouse.
Suggested Workaround: None.
Resolution Status: No fix planned.