<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000000"> Sergey Tikhonov wrote: <blockquote cite="mid425E936B.4050108@solvo.ru" type="cite"> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> <title></title> Hello, Jay Estabrook wrote: <blockquote cite="mid20050413092749.A6687@linux04.mro.cpqcorp.net" type="cite"> <pre wrap="">On Wed, Apr 13, 2005 at 10:20:52AM +0400, Sergey Tikhonov wrote: </pre> <blockquote type="cite"> <pre wrap="">Could be the following message in /var/log/messages: kernel: pci: passed pte write cache snoop test kernel: pci: failed valid tag invalid pte reload test (mcheck; workaround enabled) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ kernel: pci: tbia workaround enabled give an explanation of the problem I have with SRM on my LX164 machine? </pre> </blockquote> <pre wrap=""> Somewhat unlikely, as that's a message that one sees on any LX; it indicates a weakness of the PYXIS core-logic chipset used there (it also shows up on MIATAs). </pre> </blockquote> I tried to disconnect SCSI controller (actually I just disconnected scsi disk) and try to boot from ide disks. The SRM loaded without problems. Linux started to complain about dma problems and lost interrupt during ide initialization and hang on initialization of ncr825 adapter. Could it shed some lights on why SRM could crash when scsi disk is connected and srm is trying to reset scsi bus? Same problem when it tries to access floppy. I am supprised how different srm and alphabios in terms of initializing and handling of the same hardware. May be there is some undocumented command that I could use to switch off srm turbo mode? :) Thank you, PS. The memory I am using is: <ul> <li>Manufactured to the highest standards by <a href="http://www.msc-ge.com/e/wir/index.html">MSC </a> </li> <li>Hyundai RAM chips </li> <li>SPD - allows the BIOS to detect the memory type </li> <li>Gold Pins - high quality </li> <li>3.3V </li> <li>Double Sided - 18 chips (9 per side) </li> <li>Unbuffered memory (not registered) </li> <li>16MB x 72 organisation</li> </ul> as required by docs. </blockquote> I do not know that this is the problem, but a few years ago we tried some MSC RAM for UP2000+ and could not get it to work. They claimed it was made to the specs stated. Perhaps so, but then Alphas can be a bit fussy about RAM. With our best regards, Maurice W. Hilarius Telephone: 01-780-456-9771 Hard Data Ltd. FAX: 01-780-456-9772 11060 - 166 Avenue <a class="moz-txt-link-abbreviated" href="mailto:email:maurice@harddata.com">email:maurice@harddata.com</a> Edmonton, AB, Canada <a class="moz-txt-link-freetext" href="http://www.harddata.com/">http://www.harddata.com/</a> T5X 1Y3 This email, message, and content, should be considered confidential, and is the copyrighted property of Hard Data Ltd., unless stated otherwise. </body> </html>