Re: [PATCH v7 2/2] ARM: ftrace: Add MODULE_PLTS support

From: Qais Yousef <qais.yousef@arm.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: Alexander Sverdlin <alexander.sverdlin@nokia.com>,
	Ingo Molnar <mingo@redhat.com>,
	Russell King <linux@armlinux.org.uk>,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, Ard Biesheuvel <ardb@kernel.org>,
	Linus Walleij <linus.walleij@linaro.org>,
	Florian Fainelli <f.fainelli@gmail.com>
Subject: Re: [PATCH v7 2/2] ARM: ftrace: Add MODULE_PLTS support
Date: Mon, 22 Mar 2021 16:32:48 +0000	[thread overview]
Message-ID: <20210322163248.id7qplbk6och6kuw@e107158-lin> (raw)
In-Reply-To: <20210322110106.2bed3d50@gandalf.local.home>

On 03/22/21 11:01, Steven Rostedt wrote:
> On Sun, 21 Mar 2021 19:06:11 +0000
> Qais Yousef <qais.yousef@arm.com> wrote:
> 
> >  #ifdef CONFIG_DYNAMIC_FTRACE
> >  struct dyn_arch_ftrace {
> > -#ifdef CONFIG_ARM_MODULE_PLTS
> >  	struct module *mod;
> > -#endif
> >  };
> >  
> 
> I know you want to reduce the "ifdefery", but please note that the
> dyn_arch_ftrace is defined once for every function that can be traced. If
> you have 40,000 functions that can be traced, that pointer is created
> 40,000 times. Thus, you really only want fields in the struct
> dyn_arch_ftrace if you really need them, otherwise, that's a lot of memory
> that is wasted.

Yes you're right. I was a bit optimistic on CONFIG_DYNAMIC_FTRACE will imply
CONFIG_ARM_MODULE_PLTS is enabled too.

It only has an impact on reducing ifdefery when calling

	ftrace_call_replace_mod(rec->arch.mod, ...)

Should be easy to wrap rec->arch.mod with its own accessor that will return
NULL if !CONFIG_ARM_MODULE_PLTS or just ifdef the functions.

Up to Alexander to pick what he prefers :-)

Thanks

--
Qais Yousef