latest -git: [x86/oprofile] BUG: using smp_processor_id() in preemptible

[Vegard Nossum]

Hi,

Just got this on latest -git (+ unrelated build fix):

BUG: using smp_processor_id() in preemptible [00000000] code: oprofiled/4133
APIC error on CPU1: 00(40)
APIC error on CPU0: 00(40)
APIC error on CPU1: 40(40)
APIC error on CPU0: 40(40)
caller is get_stagger+0x9/0x30
Pid: 4133, comm: oprofiled Not tainted 2.6.27-rc3-00415-g122c9e0 #9
 [<c037228e>] debug_smp_processor_id+0xce/0xd0
 [<c0583d19>] get_stagger+0x9/0x30
 [<c05843de>] p4_fill_in_addresses+0x1e/0x3a0
 [<c058313d>] nmi_setup+0xcd/0x190
 [<c05811fa>] oprofile_setup+0x3a/0xc0
 [<c05820a6>] event_buffer_open+0x56/0x80
 [<c01a22c4>] __dentry_open+0xf4/0x1f0
 [<c01a2407>] nameidata_to_filp+0x47/0x60
 [<c0582050>] ? event_buffer_open+0x0/0x80
 [<c01ae1bc>] do_filp_open+0x18c/0x6d0
 [<c015626d>] ? put_lock_stats+0xd/0x30
 [<c01b883c>] ? alloc_fd+0xdc/0x100
 [<c0371bb6>] ? _raw_spin_unlock+0x46/0x80
 [<c066af57>] ? _spin_unlock+0x27/0x50
 [<c01a208b>] do_sys_open+0x4b/0xe0
 [<c0362da4>] ? trace_hardirqs_on_thunk+0xc/0x10
 [<c01a2189>] sys_open+0x29/0x40
 [<c01040db>] sysenter_do_call+0x12/0x3f
 =======================

I didn't want to simply wrap the code in
preempt_enable()/preempt_disable(), because that feels too much like
papering over a bigger design error.

I think oprofile is in big trouble SMP-wise, also indicated by Mike
Travis's comments in commit d18d00f5dbcd1a95811617e9812cf0560bd465ee:
"The existing code passed a reference to cpu 0's instance of struct
op_msrs to model->shutdown, whilst the other functions are passed a
reference to <this cpu's> instance of a struct op_msrs.  This seemed
to be a bug to me even though as long as cpu 0 and <this cpu> are of
the same type it would have the same effect...?"

Any ideas on how to best solve this?


Vegard

-- 
"The animistic metaphor of the bug that maliciously sneaked in while
the programmer was not looking is intellectually dishonest as it
disguises that the error is the programmer's own creation."
	-- E. W. Dijkstra, EWD1036
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/