On Mon, 2013-09-16 at 23:13 -0700, Brian Silverman wrote: > Package: src:linux > Version: 3.2.46-1+deb7u1 > Severity: normal > > I was working on some custom mutex code (implemented using futexes), and > it wasn't working, so I started it up under GDB, waited until > it died, and then tried reverse stepping back to where it did something > wrong. I then got a kernel oops. > > Here's exactly what I did: > In GDB, I set up a breakpoint in the thread which dies before the > point at which it dies, `run`, `record`, `cont`, and then (after it > crashed) `reverse-next` (might have been `reverse-step`). My X11 server > then went down and dropped me back at a virtual terminal with a kernel > backtrace on it. > > I know that there is a bug in the way that the code > uses futexes, but it shouldn't lead to a kernel oops... Indeed. > I'm attaching > both the program that causes this problem (it's compiled for amd64 with > -m32) and the kernel backtrace. [...] The attachments didn't arrive; please try again. Ben. -- Ben Hutchings The two most common things in the universe are hydrogen and stupidity.
Attachment:
signature.asc
Description: This is a digitally signed message part