Messages in this thread | | | Date | Fri, 5 Jul 2013 12:05:31 +0200 | From | Ingo Molnar <> | Subject | Re: scheduling while atomic & hang. |
| |
* Frederic Weisbecker <fweisbec@gmail.com> wrote:
> On Fri, Jul 05, 2013 at 08:51:13AM +0200, Ingo Molnar wrote: > > > > * H. Peter Anvin <hpa@zytor.com> wrote: > > > > > On 07/03/2013 07:49 PM, Linus Torvalds wrote: > > > >> [<ffffffff816f42bf>] __schedule+0x94f/0x9c0 > > > >> [<ffffffff816f487e>] schedule_user+0x2e/0x70 > > > >> [<ffffffff816f6de4>] retint_careful+0x12/0x2e > > > > > > This call trace does indeed indicate that we took a hardware > > > interrupt which caused a reschedule. It doesn't necessarily have to > > > be a quantum expiration interrupt, it could just as easily be an I/O > > > completion interrupt. > > > > Too bad the kernel stack frame at this point is so small [we are just > > before return to userspace], so we don't really get our usual trace of > > recent activities via the '?' entries ... > > > > It would be nice to have a full-kernel-stack backtrace printout as an > > option, which prints entries 'above' the current RSP. > > > > I've Cc:-ed Frederic, maybe he has interest in implementing that? > > Yeah that sounds interesting. I may give it a try.
Which would be visible clearly in the output, via some clear delineation from the 'real' portion of the stack trace - a separator line or so?
Thanks,
Ingo
| |