Red Hat Bugzilla – Bug 37503
[lmsensors] Unresponsive system after "kernel BUG at memory.c:358!"
Last modified: 2007-04-18 12:32:50 EDT
While running sensors (lm_sensors) on a via686* and rebuilding the kernel SRPM, this error was thrown to the console. The comments within the sources make it seem non-fatal and having once-in-a-long-while frequency, but I hit it with < 15 mins. uptime. I then halted the both programs and restarted them. A third vc that I was logged into would no longer start programs and although init caught my ctrl-alt-del and started sending TERM signals, CROND ran its 10 minute jobs twice before I finally cut power. I saw similiar behavior once during the beta while doing the same thing, but I was away for hours while this was left running and could not check the logs. Reproducible: Sometimes Steps to Reproduce: 1. Run a CPU intensive task (suggest "while true ; do rpm --rebuild --clean kernel.src.rpm ; done") 2. Run 'watch sensors|tail' at the same time 3. Wait. Actual Results: Ocassional hard lockups. Expected Results: Frequent hard lockups. Hardware was: Athlon 1GHz (10x100) at rated clock, Abit KT7-RAID m/b 512MB of PC133 RAM with CAS3 setting output: === kernel BUG at memory.c:358! invalid operand: 0000 CPU: 0 EIP: 0010:[<c0121d5e>] EFLAGS: 00010282 eax: 0000001c ebx: d000de60 ecx: 00000008 edx: 00000000 esi: 00000000 edi: df46f500 ebp: 00000000 esp: dabd5ef0 ds: 0018 es: 0018 ss: 0018 Process cc1 (pid: 12737, stackpage=dabd5000) Stack: c020cbbb c020cd7d 00000166 00000001 00153000 00000001 00153000 d5c62408 00000000 00000000 d5c62000 00000001 dabd4000 dabd4000 d9e036a0 d9e036a0 00030002 cfe57320 c0123ff3 c1893c14 d9e036a0 d9e036a0 00000292 d000de60 Call Trace: [<c020cbbb>] [<c020cd7d>] [<c0123ff3>] [<c0124698>] [<c0114b86>] [<c01188d9>] [<c01243db>] [<c0112f30>] [<c010901b>] Code: 0f 0b 83 c4 0c 8d b6 00 00 00 00 8d bc 27 00 00 00 00 8b 44 ===
The machine runs fine if you don't have the lmsensors module loaded ?
No, I haven't been able to reproduce that exact error. I ran into the following two errors while trying to reproduce it, but I suspect a partially hardware related cause at this point. Unable to handle kernel paging request at virtual address 7564705d printing eip: c01339b8 pgd entry cf86a754: 0000000000000000 pmd entry cf86a754: 0000000000000000 ... pmd not present! Oops: 0000 CPU: 0 EIP: 0010:[<c01339b8>] EFLAGS: 00010206 eax: 75646f6d ebx: d5b831a0 ecx: cfbc0320 edx: 00001000 esi: 00001000 edi: 00000000 ebp: 00001000 esp: cf86df90 ds: 0018 es: 0018 ss: 0018 Process cc1 (pid: 4280, stackpage=cf86d000) Stack: cf86dfbc dea5f640 cfb28000 dea5f640 cf86c000 00000000 cf86c000 00000006 cf86c000 401508e0 4015e000 bfffe648 c010901b 00000000 4015e000 00001000 401508e0 4015e000 bfffe648 00000003 0000002b 0000002b 00000003 40101f44 Call Trace: [<c010901b>] Code: f6 80 f0 00 00 00 01 74 0a 6a 01 50 e8 c7 50 01 00 5f 5d 89 kernel BUG at page_alloc.c:90! invalid operand: 0000 CPU: 0 EIP: 0010:[<c012d0ea>] EFLAGS: 00010282 eax: 0000001f ebx: 00000000 ecx: 00000005 edx: 00000000 esi: 00000000 edi: c10cc384 ebp: 00000000 esp: ce1b3ea8 ds: 0018 es: 0018 ss: 0018 Process rm (pid: 23050, stackpage=ce1b3000) Stack: c020ea7b c020ec89 0000005a c10cc384 c0135b93 c10cc384 00000000 c10cc384 c10cc384 c10cc384 c10cc384 00000000 c0124d6c c10cc384 00000000 c19e3a00 c285b1a0 000001a0 c285b00c ce1b3f0c cc913108 00000000 bfffed68 c0124e2d Call Trace: [<c020ea7b>] [<c020ec89>] [<c0135b93>] [<c0124d6c>] [<c0124e2d>] [<c 0147b57>] [<c0158b4f>] [<c0158b61>] [<c014643c>] [<c013fb6c>] [<c013fc39>] [<c010901b>] Code: 0f 0b 83 c4 0c 8b 47 18 83 e0 20 74 16 6a 5c 68 89 ec 20 c0
Closing. Seems to match the VIA chipset flaw that 2.4.9 and later work around. If it stil occurs feel free to re-open