Re: [RFC PATCH v1 3/4] arm64: Detect FTRACE cases that make the stack trace unreliable

[Madhavan T. Venkataraman]

On 4/1/21 9:27 AM, Mark Brown wrote:
> On Tue, Mar 30, 2021 at 02:09:54PM -0500, madvenka@xxxxxxxxxxxxxxxxxxx wrote:
> 
>> +	 * FTRACE trampolines.
>> +	 */
>> +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_REGS
>> +	{ (unsigned long) &ftrace_graph_call, 0 },
>> +#ifdef CONFIG_FUNCTION_GRAPH_TRACER
>> +	{ (unsigned long) ftrace_graph_caller, 0 },
>> +	{ (unsigned long) return_to_handler, 0 },
>> +#endif
>> +#endif
> 
> It's weird that we take the address of ftrace_graph_call but not the
> other functions - we should be consistent or explain why.  It'd probably
> also look nicer to not nest the ifdefs, the dependencies in Kconfig will
> ensure we only get things when we should.
> 

I have explained it in the comment in the FTRACE trampoline right above
ftrace_graph_call().

        /*
         * The only call in the FTRACE trampoline code is above. The above
         * instruction is patched to call a tracer function. Its return
         * address is below (ftrace_graph_call). In a stack trace taken from
         * a tracer function, ftrace_graph_call() will show. The unwinder
         * checks this for reliable stack trace. Please see the comments
         * in stacktrace.c. If another call is added in the FTRACE
         * trampoline code, the special_functions[] array in stacktrace.c
         * must be updated.
         */

I also noticed that I have to fix something here. The label ftrace_graph_call
is defined like this:


#ifdef CONFIG_FUNCTION_GRAPH_TRACER
SYM_INNER_LABEL(ftrace_graph_call, SYM_L_GLOBAL) // ftrace_graph_caller();
        nop                             // If enabled, this will be replaced
                                        // "b ftrace_graph_caller"
#endif

So, it is only defined if CONFIG_FUNCTION_GRAPH_TRACER is defined. I can address
this as well as your comment by defining another label whose name is more meaningful
to our use:


+SYM_INNER_LABEL(ftrace_trampoline, SYM_L_GLOBAL) // checked by the unwinder
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
SYM_INNER_LABEL(ftrace_graph_call, SYM_L_GLOBAL) // ftrace_graph_caller();
        nop                             // If enabled, this will be replaced
                                        // "b ftrace_graph_caller"
#endif

Is this acceptable?

Thanks.

Madhavan