Created attachment 23915 [details] dmesg from 2.6.31-1-686 Hi So far I have only experienced it when I build eclipse in a chroot'ed system. When I just leave the computer to do its "normal" job (e.g. hosting git a couple of git repositories) it does not happen. ~Niels
Created attachment 23916 [details] Other info about the kernel. Hi I have attached some other information that my distribution also collects on a bug report in case it is important. This bug (as I forgot to mention in the original message) is a forward of the Debian bug 554272 [1]. ~Niels [1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=554272
Created attachment 23917 [details] dmesg from 2.6.30-2-686 I have attached the log from 2.6.30-2-686 where I originally experienced a problem (The log refers to it as a "dereference of NULL pointer" though), in case it helps you locate the issue. The log here contains two separate "dereferences" from two separate build attempts. ~Niels
Is this in any way reproducible? I've never seen a NULL pointer dereference in generic_file_aio_read (which actually is common code used by most filesystems before)
Hi Personally I do find it strange that my other machine do not show any signs of this issue. My affected machine did however reliably (often, not always) reproduce it (or the NULL pointer issue with the previous kernel). Both my machines uses xfs, though my unaffected machine also have an lvm/lvm2 layer beneath the xfs. Since I did not use my affected machine for anything, I decided to take it offline last week. Though I do not mind booting it up and see if I can still reproduce the issue, if you want that - also let me know if there is anything besides the dmesg log you want for debugging. ~Niels
Hi, I have decommission the affected machine and replaced it with a far better machine for my purpose. My new machine has not suffered this issue or anything like it. I still have the old machine lying around if you need extra information; but I am willing to write this off as a "possible hardware defect" and "call it a day". ~Niels
Christoph et al: any idea idea how to track down whether this is a hardware problem or a symptom of some particular filesystem corruption? Niels: am I correct in guessing you've already tried memtest86+ and badblocks?
Hi, I cannot remember what I tested (or didn't test) about 3 years ago, so it is entirely possible I forgot to check one of those. I cannot reproduce it on any system I am currently using. So to be honest, I am okay with this being closed this as "unreproducible" ... also partly because I am too lazy to reinstall an OS on it. :) ~Niels