[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150619084149.1c625bf8@grimm.local.home>
Date: Fri, 19 Jun 2015 08:41:49 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: Torsten Duwe <duwe@....de>
Cc: Denis Kirjanov <kda@...ux-powerpc.org>,
Michael Ellerman <mpe@...erman.id.au>,
Jiri Kosina <jkosina@...e.cz>, linuxppc-dev@...ts.ozlabs.org,
linux-kernel@...r.kernel.org
Subject: Re: [RESEND PATCH 0/4] ppc64 LE ABI v2 ftrace-with-regs
implementation
On Fri, 19 Jun 2015 14:17:03 +0200
Torsten Duwe <duwe@....de> wrote:
> On Fri, Jun 19, 2015 at 10:15:44AM +0300, Denis Kirjanov wrote:
> >
> > Hi Torsten,
> >
> > could you provide a description how you did test that.
>
> Vanilla kernel, those patches on top, built as described to
> Steven in the previous mail (.config attached there, built
> natively on ppc64le ABIv2, mentioned tool versions)
>
Great that you got it working on your machine. But you need to do it in
a way that you don't break builds of existing machines, with existing
(still supported) compilers.
-- Steve
> make vmlinux, make modules, build initrd
> qemu-system-ppc64 -no-reboot -enable-kvm -m 20480 -kernel vmlinux
> -initrd initrd.gz -append console=hvc0 -nographic -vga none
> inside:
> cd /sys/kernel/debug/tracing
> mkdir instances/Foo
> cd instances/Foo
> echo function > current_tracer
>
> current_tracer holds "nop" or "function"; the transition is accompanied
> by heavy binary patching. When the function tracer is active, almost all
> calls go through ftrace_caller. This should be enough for a start; the
> proper stack frame layout details for the tracer functions still remain
> to be checked...
>
> Torsten
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists