From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 4.0; WAN00) Description of problem: Kernel dumps CPU state on CRT with message: Unable to handle kernel paging request at virthal address xxxxxx How reproducible: Always Steps to Reproduce: 1. Start RH 6.1 2. Wait several hours. 3. Actual Results: Symptoms: After running for several hours (random 1 hour up to 2 days) I get the following dump on the screen (this occurs most times while doing no work, just idling). The dumps get more frequent until the computer finally locks up: Unable to handle kernel paging request at virtual address 41c1b880 current->tss.cr3 = 04948000, %cr3 = 04948000 *pde = 00000000 Oops: 0002 CPU: 0 EIP: 0010:[<c0130b61>] EFLAGS: 00010206 eax: c1d598c0 ebx: c1c1b760 ecx: c1cb780 edx: 41c1b880 esi: 00000000 edi: 080aa338 ebp: bffffa40 esp: c2609f9c ds: 0018 es: 0018 ss: 0018 Process bash (pid: 7156, process nr: 64, stackpage=c2609000) <stackdump goes here> <call trace> Code: 89 02 8d 43 18 39 43 18 0f 85 95 00 00 00 8b 43 28 8b 53 2c ------ That's it! I get this paging problem only in 6.1. I tried upgrading to the latest 6.1 kernel, and it does this even more often. 7.1 kernel will not deal with the Adaptec, 7.0 and 6.2 crash on loading with scsi errors. This hardware runs win98/winNT just fine, no glitches at all. Expected Results: Should run flawlessly under 6.1, 6.2 7.0 or 7.1 Additional info: Hardware info: Intel LX chipset on motherboard (P2BLX) Intel 333 mhz Celeron cpu, 256mb ram Adaptec 2940U scsi card with 3 IBM 9.1 gb drives (narrow scsi buss) Toshiba scsi CD drive, HP scsi tape drive RH6.1 installed on drive 0. /boot 23mb, /swap 256mb, / 8.1gb
I assume the errata'd kernel is 2.2.19, could you check this ? You can most likely install 7.1 by passing it the "noapic" option to work around some motherboard/biosbugs.
Kernel version is 2.2.12-20
Sorry, guys. This turned out to be a hardware problem. After extensive testing, I have found that the problem is related to a defective Programmabe Interrupt Controller on the motherboard, when the board warms up over a certain temperature, it begins to mess up interrupt routing. I have replaced the board and all is well with version 6.1 now. Cannot load 7.1 no matter what options I pass to the kernel. Those problems are better described in other existing bug reports. Thanks, Del