Lindows OS 3.0 X-Windows issues with Athlon /AGP GART

--------------------------------------------------
Item 1)

I tried installing LOS 3.0 after a fresh download & burn to CD on my main PC "cromwell5B" with the following hardware:
1. a plug-in dedicated DMA100 8.3 MB HDD,
2. 768 MB PC133 SDRAM,
3. Asus A7A266-E board, Ali M1647 chipset,
4. Athlon 1800+ cpu,
5. Primary display is on a Inno3D Tornado TNT2 (nVidia Riva TNT2 Model 64 chipset) 32 MB AGP display card into an AOC Spectrum A770 Vlr 17" VDU,
6. Secondary display on a 2nd identical VDU on a 16 MB STB Velocity PCI card (nVidia Riva TNT chipset),
7. cable broadband through a Netgear FA311 NIC,
8. Ethernet LAN to two other PCs via a Realtek 8139 NIC, and 10/100 switch.

This booted off the CD OK, but using "Takeover Install" gave "fatal error - code 30 - reboot". My AMIBIOSes provide for AUTO, or USER in which case it offers MODE on/off. I tried all possible variations in BIOS to do with the IDE/ATA/ATAPI drives but no change.

However using the "Advanced Install" and selecting a partition previously used by LOS 2.1, it completed the installation, and rebooted! The bad news is that it gave me the tty1 login only. I logged in as root, and tried "startx". As for LOS 2.0 & 2.1, I got the bottom of pages of setup report, ending with:

Build Operating System: Linux 2.4.19 i686 [ELF]
Module loader present
(WW) NV: no matching Device section for instance (BUS ID PCI:11:0) found
(II) calling GIxExtensionInit

fatal server error
caught signal 11
aborting
see log in "/var/log/XFree86.0.log"
using config file "/etc/X11/XF86Config-4"
report problems to debian-X@lists.debian.org
----------------------------------------------------
Item 2)

I then repeated the exercises on another older PC "cromwell5c" with the following hardware:
1.GA-5AA v.1.2 m/board,
2.AMD K62-500, 3.192 MB SDRAM,
4.plugin 15 GB ATA 66 HDD,
5.Nvidia TNT2 Mod64 32 MB AGP display card,
6.17" AOC 7vlr monitor,
7.Epson SCP750 USB printer,
8.Realtek 8139 NIC.

The takeover install again gave "fatal error - code 30 - reboot" as before. I couldn't progress at all with the Advanced Install.

The PCs in Item1 & Item2 both have AMD processors and use AGP nVidia display cards.

Both these PCs run LOS SPX 1.1.1, Mandrake 9, and Windoze 98/ME/2K/XP without this problem , so it comes down to the Lindows OS distribution.

Has anyone else has reported a similar set of problems privately? I have not seen any references to this issue in the Lindows Insiders forums.

--------------------------------------------------------
Item 3.

I tried the same LOS 3.0 CD using the same hard drive in a friends PC which had a Celeron 1GB cpu, 256 MB RAM, Intel main board, and PCI display card. This installed without any problems and gave a working KDE Desktop in about 15 minutes. This seems to clear the drive & CD of having any errors.

I then transferred this drive back to my PC "cromwell5b", and it booted straight into the Lindows desktop. I configured both LAN and broadband network cards (with some difficulty as eth01 was not recognised initially) and rebooted to pick these up. When it rebooted I only got the tty1 login again, and it refused to run X-Windows.
---------------------------------------------
It seems quite clear to me that this is a bug in LOS recent distributions i.e. 2.1 -> 3.0, as SPX 1.1.1 loaded OK.

I have found several references on the Net that there is an Athlon/AGP issue in the Debian Linux kernel relating to, e.g. see:
http://www.cs.helsinki.fi/linux/linux-kernel/2002-03/0711.html

To quote:

AMD's educated guess is that these Athlon/AGP stability problems have to do with speculative writes by the CPU and how they can cause inadvertent trashing of AGP memory if pages are mapped with indiscretion by the OS and drivers.

-- AMD explanation follows ---

This note documents a subtle problem that AMD has seen in the field.

The operating system has created a 4MB translation that has attribute
bits that allow it to be cacheable. GART also contains translations to
part of the underlying physical memory of this 4MB translation.

This situation is fundamentally illegal because GART is non-coherent and all translations that the processor could use to access the AGP memory must, therefore, be non-cacheable. Although we have seen no intentional access to the AGP memory by the processor via the 4MB cacheable translation we have seen legitimate, speculative, accesses performed by the processor.

The problem that has been experienced is caused by a speculative store
instruction that is not ultimately executed. The logical address of the
store is through the 4MB translation to physical memory also translated
by GART and used by the AGP processor.

The effect of the store is to write-allocate a cache line in the data
cache and fill that cache line with data from the underlying physical
memory. Because the line was write-allocated it is subsequently written
back to physical memory even though the bits have not been changed by
the processor. This write-back occurs when the cache-line is
re-allocated based upon replacement policy and is far removed in time
from the point at which the bits were read.

Between the time of the read and the time of the write, the AGP
processor has modified the bits in physical memory and the bits in the
data cache are stale. This is happens because GART, being non-coherent,
does not snoop the processor caches for modified data.

When the cache-line eviction occurs the stale data written to physical
memory has fatal side effects.

Our conclusion is that the operating system is creating coherency
problems within the system by creating cacheable translation to AGP
GART-mapped physical memory.

-- end of AMD explanation --

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.426 / Virus Database: 239 - Release Date: 02/12/2002