From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.2.1) Gecko/20010901 Description of problem: I can't get my dual athalon system booting with 2.4.9-31smp Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1.buy an dual athalon system. 2.load 2.4.9-31smp 3.Try and boot it in SMP mode Actual Results: hang. Expected Results: full running system Additional info: Here's the deal. I have purchased 2 dual AMD systems. Motherboard is an ASUS A7M266-D and the CPU's are AMD Athalon MP 1800+. The two have slightly different periferals, but the basic motherboard/CPU/Memory is the same. The first system was purchased 3 months ago, and basically worked great "right out of the box". The second system works great in UP mode, but fails to boot in SMP mode. One thing to note, the first system would not boot in SMP mode unless the smp-athalon kernel was loaded. Both systems have the athalon smp kernel loaded. The last boot messages the problem system prints to the screen while trying to boot into smp mode is the following. CPU1: AMD Athalon(TM) XP 1800+ stepping 02 Total of 2 processors activated (6121.06 Bogomips) Enableing IO-APIC IRQs ... changing IO-APIC physical APIC ID to 2... ok. .. TIMER vector=0x31 pin1=2 pin2=0 testing the IO APIC............. ......... done. Using local APIC timer interupts. Calibrating APIC timer... .... CPU clock speed is 1533.2891 MHz. ..... Host bus clock speed is 266.6590 Mhz. cpu:0, cloks 2666590, slice: 888868 CPU<T0: 2666576, T1: 1777712, D:1, S: 888862, C:2666590> and that's it. What I have done is load as a seperate text file the dmesg output from the working system. You can see that there is a differnece between the two at the hang point. It looks like the working system is able to init the IO-APIC IRQ's since you get that message in the working system just about at the point when the failing system hangs. I've checked the jumper and bios settings on both system and they are identical. I hope you guys have an idea of what is going on here.
Created attachment 50544 [details] dmesg output from working smp athalon system
one other note, I have not tried earlier versions of the 2.4.9 kernel. I don't expect different results but if you guys want me to try that, I can.
Known problem. Ask ASUS for BIOS fixes You can sort of work around the one problem by setting the BIOS to MP1.1, but you may well find the board is still not correctly configured with some card combinations, and is not setting PCI 2.2 compliance. The RH beta fixes that one up, but we can't fix the MP 1.4 table problem. I've got open items with AMD about this board and its problems.
I want to thank Alan Cox for helping me on this problem. Just for the record, the ASUS A7M266-D motherboard's bios has an option to enable or disable MP 1.4. The section is found under the BIOS Boot menu. When I set the MP 1.4 option to disable, the system booted up in smp mode. I don't know what the difference is between MP 1.4 or 1.1, as Alan talks about, if I gain in performance with 1.4 or what, but at least I have 2 CPU's running. Thank you Red Hat and Alan for helping me with this.