On Wed, 2007-04-25 at 14:54 +0300, )\(@sS wrote:
im sorry i understood nothing of your explanation... :) thats due to my lack of any knowledge on the matter.. i will get ksymoops but could you break down what you mean by back traces and what by 'bogus'
Oopses basically happen for two reasons: software and hardware failure. Hardware failures tend to cause mystic and nondeterministic behaviour, crashing at random locations. On the consistent backtraces I meant that several of the Oopses that you posted had the same backtrace (e.g. the same crash location), which indicates that it's more likely a software related problem. The listing I gave you, is the direct x86 dissassembly of the code found at the crash location, this is reported as a sequence of bytes in the Oops output, right after the "Code:" label. I said it was bogus as it made no sense. This is not valid code, so either something corrupted the instruction pointer (by jumping to an invalid function pointer), or something overwrote the stack. I checked the oopses again, don't go to LKML, your kernel is tainted (e.g. it has a proprietary module loaded, they will not help you). -- Bazsi