dave schrieb: > Can you remember if you were running a 2.6.26 kernel at the time? A lot of > the reports mentioned 2.6.26 RT, and included the firefox error lines.
...which may or may not be of significance. On the usual desktop machine, firefox is the largest program in memory which almost everyone will have running after booting up the system. Running memtest is of course a good idea, just to exclude this possible and rather common problem (of some memory unit being broken). Besides that, you could try to observe the messages. Are they repeated? Any correlations with what you are doing? E.g. if you start open office or use ardour (or any other program with large footprint), do you get new messages? Btw, I also had those messages sometimes in past on a machine with ubutustudio (but I am running now a 2.6.22-RT kernel). Just to explain: the messages indicate that something with memory management goes wrong. The kernel found a memory page with inconsistent status flags. This page belongs to the firefox process. Now, this may be just accidental (and because firefox typically owns one of the largest memory chunks, it got a high probability to suffer from any sporadic problems). Or, it might be a bug within firefox. Or, it might be a problem with the kernel and/or realtime patch, which happens to be unearthed by something firefox does. You could try to rule out those options, by not starting firefox, and by booting with another kernel. If, for example, you don't start firefox, but instead start open office (which is similar fat), or ardour, while remaining on your realtime kernel, and you still get the message (now with another process), than you could rule out firefox as a cause. I saw several systems with defective hardware, which produced a very low rate of corrupted bits on every transmitted data. If, for example (as it was in the case of one of my past desktop machines), there is a probability that the system flips one bit per 50MB of data transmitted from HD, and this only when it's hot (in Summer), then you'll get all sorts of unspecific crashes, warnings, sometimes corrupted checksums and never find any cause. *In such a case*, re-installing any piece of software may happen to behave like it fixed a problem. But note, this is the sole reason why re-installing may change something. Please remember, this isn't Windows here. Usually, people always waste a lot of time with re-booting, closing all windows, re-installing software, re-installing the whole OS, baking kernels on a random base without knowing what they do. When you google for any problem, usually you'll find a lot of advice telling you to do such crap and a lot of people claiming they fixed stuff this way. I want to give you the advice never to hear or to follow such recipes, unless you are completely desperate and void of any real idea how to tackle the problem. If just "doing" something makes you feel better, then please, re-install all of your OS five times every week. But, if you want to solve a problem in a rational manner, then try to isolate the cause, apply *only one single change a time* and then re-evaluate if the problem persists, use the help of real experts, i.e. people who act based on knowledge, not based on assumptions and opinions. Cheers, Hermann _______________________________________________ 64studio-users mailing list [email protected] http://lists.64studio.com/mailman/listinfo/64studio-users
