Bug 14734
Summary: | BUG: unable to handle kernel NULL pointer dereference at 0000000000000808 | ||
---|---|---|---|
Product: | File System | Reporter: | stmcdermott |
Component: | ext3 | Assignee: | fs_ext3 (fs_ext3) |
Status: | CLOSED UNREPRODUCIBLE | ||
Severity: | normal | CC: | alan, jack, mkatiyar, ruan.zhengwang |
Priority: | P1 | ||
Hardware: | All | ||
OS: | Linux | ||
Kernel Version: | 2.6.31.6 | Subsystem: | |
Regression: | No | Bisected commit-id: |
Description
stmcdermott
2009-12-04 15:39:28 UTC
On Fri, Dec 04, 2009 at 03:39:30PM +0000, bugzilla-daemon@bugzilla.kernel.org wrote: > > Just let me know what else I should post in addition to the following. I've > gotten this bug pretty consistently, on 2.6. Hmm.... there's a strnage fsnotify_clear_marks_by_inode() on the stack trace which doesn't seem to make any sense and which I can't place. Are you using the fsnotify feature for some reason (i.e., some new-fangled proprietnary "linux anti-virus" package or some such?). If so, does the problem go away if you stop using said package and not use the fsnotify interface? Also, do you have to do anything special to trigger it? Is the "rm" in the process handle an "rm -rf", or some such? - Ted I believe this is a faulty memory: Looking at the disassembly we most probably oopsed while doing page_cache_get_speculative(page) where page was 0000000000000800. So it's a single bit flip from NULL in which case we'd just go on with another page. Could you try running memtest for some time? I'll execute a memtest tonight. thanks |