Messages in this thread | | | Date | Thu, 6 Nov 2008 07:30:12 +0100 | From | Ingo Molnar <> | Subject | Re: [patch] sched: fix single-depth wchan output |
| |
* Ken Chen <kenchen@google.com> wrote:
> On Wed, Nov 5, 2008 at 10:11 PM, Ingo Molnar <mingo@elte.hu> wrote: > > > hm, this adds overhead - and the thing is that WCHAN is rather > > uninformative to begin with (because it's a single dimension), so > > we should phase it out, not expand it. > > > > How about adding a /proc/<PID>/stacktrace file that gives us the > > stack trace of any task in the system? That would be useful for a > > number of other purposes as well, and about 100 times more useful > > than wchan. (often it would be more useful than sysrq-t dumps) > > Sure, my main motivation is to remove frame pointer generation. > x86_64 unconditionally adds fp for kernel/sched.c right now. I'm > all for phasing out wchan if people don't think there is value in > it.
are you interested in adding /proc/<PID>/stacktrace? If yes then we could remove fp generation for 64-bit right now and add your stacktrace patch when you are done with it.
Generally we want frame pointers for high quality backtraces and trouble-shooting. The small cost is almost always worth paying and most distros enable framepointers for that reason. On 32-bit a no-framepointers kernel image has less register pressure, but on 64-bit there's little reason to not enable them.
Ingo
| |