On Fri, 28 Feb 2014 10:07:29 -0500 (EST)
Vince Weaver <vincent.wea...@maine.edu> wrote:
> On Fri, 28 Feb 2014, Steven Rostedt wrote:
> 199.900696: function: __module_address
> ...
> 199.900705: function: __kernel_text_address
> 199.900809: kernel_stack: <stack trace>
> => perf_callchain (ffffffff810d35a2)
> => perf_prepare_sample (ffffffff810cfae3)
> => __perf_event_overflow (ffffffff810d02f4)
> => perf_swevent_overflow (ffffffff810d04e3)
> => perf_swevent_event (ffffffff810d0574)
> => perf_tp_event (ffffffff810d070c)
> => perf_trace_x86_exceptions (ffffffff810341b6)
> => trace_do_page_fault (ffffffff81537702)
> => trace_page_fault (ffffffff81534772)
Thank you!!! You just found the bug :-)
The bug was caused by:
commit 25c74b10bacead867478480170083f69cfc0db48
x86, trace: Register exception handler to trace IDT
With this code:
dotraplinkage void __kprobes
trace_do_page_fault(struct pt_regs *regs, unsigned long error_code)
{
enum ctx_state prev_state;
prev_state = exception_enter();
trace_page_fault_entries(regs, error_code);
__do_page_fault(regs, error_code);
exception_exit(prev_state);
}
The trace_page_fault_entries() which is called before the cr2 is saved
can fault by perf doing a userspace stack trace. But the cr2 is not
restored when calling __do_page_fault() and that gets the wrong cr2.
Below is a patch that should fix this. Please remove all other patches
and try this out.
Thanks,
-- Steve
> 199.900810: function: perf_output_begin
> 199.900810: function: __do_page_fault
> 199.900810: function: __perf_sw_event
> 199.900810: function: perf_swevent_get_recursion_context
> 199.900811: function: down_read_trylock
> 199.900811: function: _cond_resched
> 199.900811: function: find_vma
> 199.900811: function: bad_area
> 199.900812: function: up_read
> 199.900812: function: __bad_area_nosemaphore
> 199.900812: function: is_prefetch
> 199.900812: function: convert_ip_to_linear
> 199.900813: function: unhandled_signal
> 199.900813: function: __printk_ratelimit
> 199.900813: function: _raw_spin_trylock
> 199.900813: function: _raw_spin_unlock_irqrestore
> 199.900814: function: printk
> 199.900814: function: vprintk_emit
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 6dea040..66b636d 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -1271,9 +1271,15 @@ dotraplinkage void __kprobes
trace_do_page_fault(struct pt_regs *regs, unsigned long error_code)
{
enum ctx_state prev_state;
+ unsigned long cr2;
prev_state = exception_enter();
+ /* The trace might fault, save the cr2 register */
+ cr2 = read_cr2();
trace_page_fault_entries(regs, error_code);
+ /* Put back the original cr2 if needed */
+ if (cr2 != read_cr2())
+ write_cr2(cr2);
__do_page_fault(regs, error_code);
exception_exit(prev_state);
}
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
