Messages in this thread | | | From | Dominik Karall <> | Subject | Re: [patch] preempt-trace.patch (mono preempt-trace) | Date | Fri, 5 Aug 2005 21:23:45 +0200 |
| |
On Friday 05 August 2005 20:46, Hugh Dickins wrote: > On Fri, 5 Aug 2005, Dominik Karall wrote: > > On Friday 05 August 2005 17:22, Ingo Molnar wrote: > > > please enable CONFIG_FRAME_POINTERS! > > > > I'm sorry, but I think I can't enable CONFIG_FRAME_POINTERS. > > Depends on: DEBUG_KERNEL && (X86 && !X86_64 || CRIS || M68K || M68KNOMMU > > || FRV || UML) > > > > Seems to be disabled for x86_64. > > It is disabled for x86_64, but not for any very good reason (beyond > reducing the test matrix). I work with CONFIG_FRAME_POINTERS on x86_64 > with no trouble, just add in the patch below, make oldconfig, choose > frame pointers and rebuild). But I can't guarantee it'll actually > reveal the info Ingo and all are longing to see.
With FRAME_POINTERS enabled:
BUG: mono[3193] exited with nonzero preempt_count 1! --------------------------- | preempt count: 00000001 ] | 1 level deep critical section nesting: ---------------------------------------- .. [<ffffffff80400a46>] .... _spin_lock+0x16/0x80 .....[<ffffffff801ed30c>] .. ( <= sys_semtimedop+0x28c/0x7c0)
hth, let me know!
dominik [unhandled content-type:application/pgp-signature] | |