[PATCH 1/1] ftrace: adding other non-leaving .text sections

[PATCH 1/1] ftrace: adding other non-leaving .text sections

[Liming Wang]

Please review this patch, and I don't know whether new sections will conflict 
with sched tracer or others. 
This patch is against Steven tree's tip/devel latest commit.

Impact: improve recordmcount.pl

Besides .text section, there are three .text sections that won't
be freed after kernel booting. They are: .sched.text, .spinlock.text
and .kprobes.text, which contain functions we can trace. But the last
section ".kprobes.text" is particular, which has been marked as "notrace",
we ignore it. Thus we add other two sections.

Signed-off-by: Liming Wang <liming.wang@windriver.com>
---
 scripts/recordmcount.pl |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl
index 0197e2f..0b1dc9f 100755
--- a/scripts/recordmcount.pl
+++ b/scripts/recordmcount.pl
@@ -112,6 +112,8 @@ my ($arch, $bits, $objdump, $objcopy, $cc,
 # Acceptable sections to record.
 my %text_sections = (
      ".text" => 1,
+     ".sched.text" => 1,
+     ".spinlock.text" => 1,
 );
 
 $objdump = "objdump" if ((length $objdump) == 0);
-- 
1.6.0.3

Re: [PATCH 1/1] ftrace: adding other non-leaving .text sections

[Ingo Molnar]

* Liming Wang <liming.wang@windriver.com> wrote:

> Please review this patch, and I don't know whether new sections will conflict 
> with sched tracer or others. 
> This patch is against Steven tree's tip/devel latest commit.
> 
> Impact: improve recordmcount.pl
> 
> Besides .text section, there are three .text sections that won't
> be freed after kernel booting. They are: .sched.text, .spinlock.text
> and .kprobes.text, which contain functions we can trace. But the last
> section ".kprobes.text" is particular, which has been marked as "notrace",
> we ignore it. Thus we add other two sections.
> 
> Signed-off-by: Liming Wang <liming.wang@windriver.com>
> ---
>  scripts/recordmcount.pl |    2 ++
>  1 files changed, 2 insertions(+), 0 deletions(-)
> 
> diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl
> index 0197e2f..0b1dc9f 100755
> --- a/scripts/recordmcount.pl
> +++ b/scripts/recordmcount.pl
> @@ -112,6 +112,8 @@ my ($arch, $bits, $objdump, $objcopy, $cc,
>  # Acceptable sections to record.
>  my %text_sections = (
>       ".text" => 1,
> +     ".sched.text" => 1,
> +     ".spinlock.text" => 1,
>  );

make sense - applied to tip/tracing/ftrace, thanks!

	Ingo

Re: [PATCH 1/1] ftrace: adding other non-leaving .text sections

[Steven Rostedt]

On Wed, 26 Nov 2008, Liming Wang wrote:

> Please review this patch, and I don't know whether new sections will conflict 
> with sched tracer or others. 
> This patch is against Steven tree's tip/devel latest commit.
> 
> Impact: improve recordmcount.pl
> 
> Besides .text section, there are three .text sections that won't
> be freed after kernel booting. They are: .sched.text, .spinlock.text
> and .kprobes.text, which contain functions we can trace. But the last
> section ".kprobes.text" is particular, which has been marked as "notrace",
> we ignore it. Thus we add other two sections.

Yep, only those sections that we know will not disappear are OK. I was
conservative with only using ".text" but I knew there were other sections 
that were OK, so I used the hash array to make it easy to add others.

Thanks!

Acked-by: Steven Rostedt <srostedt@redhat.com>

-- Steve

> 
> Signed-off-by: Liming Wang <liming.wang@windriver.com>
> ---
>  scripts/recordmcount.pl |    2 ++
>  1 files changed, 2 insertions(+), 0 deletions(-)
> 
> diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl
> index 0197e2f..0b1dc9f 100755
> --- a/scripts/recordmcount.pl
> +++ b/scripts/recordmcount.pl
> @@ -112,6 +112,8 @@ my ($arch, $bits, $objdump, $objcopy, $cc,
>  # Acceptable sections to record.
>  my %text_sections = (
>       ".text" => 1,
> +     ".sched.text" => 1,
> +     ".spinlock.text" => 1,
>  );
>  
>  $objdump = "objdump" if ((length $objdump) == 0);
> -- 
> 1.6.0.3
> 
> 
> 

Re: [PATCH 1/1] ftrace: adding other non-leaving .text sections

[Paul Mundt]

On Tue, Nov 25, 2008 at 10:03:35PM -0500, Steven Rostedt wrote:
> 
> On Wed, 26 Nov 2008, Liming Wang wrote:
> 
> > Please review this patch, and I don't know whether new sections will conflict 
> > with sched tracer or others. 
> > This patch is against Steven tree's tip/devel latest commit.
> > 
> > Impact: improve recordmcount.pl
> > 
> > Besides .text section, there are three .text sections that won't
> > be freed after kernel booting. They are: .sched.text, .spinlock.text
> > and .kprobes.text, which contain functions we can trace. But the last
> > section ".kprobes.text" is particular, which has been marked as "notrace",
> > we ignore it. Thus we add other two sections.
> 
> Yep, only those sections that we know will not disappear are OK. I was
> conservative with only using ".text" but I knew there were other sections 
> that were OK, so I used the hash array to make it easy to add others.
> 
Does your .text cover .text.head by the way?

Re: [PATCH 1/1] ftrace: adding other non-leaving .text sections

[walimis]

On Wed, Nov 26, 2008 at 12:04:46PM +0900, Paul Mundt wrote:
>On Tue, Nov 25, 2008 at 10:03:35PM -0500, Steven Rostedt wrote:
>> 
>> On Wed, 26 Nov 2008, Liming Wang wrote:
>> 
>> > Please review this patch, and I don't know whether new sections will conflict 
>> > with sched tracer or others. 
>> > This patch is against Steven tree's tip/devel latest commit.
>> > 
>> > Impact: improve recordmcount.pl
>> > 
>> > Besides .text section, there are three .text sections that won't
>> > be freed after kernel booting. They are: .sched.text, .spinlock.text
>> > and .kprobes.text, which contain functions we can trace. But the last
>> > section ".kprobes.text" is particular, which has been marked as "notrace",
>> > we ignore it. Thus we add other two sections.
>> 
>> Yep, only those sections that we know will not disappear are OK. I was
>> conservative with only using ".text" but I knew there were other sections 
>> that were OK, so I used the hash array to make it easy to add others.
>> 
>Does your .text cover .text.head by the way?
No, and not needed. Because .text.head is of booting code, to which
function tracer can't trace.

walimis
>--
>To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html
>Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 1/1] ftrace: adding other non-leaving .text sections

[Paul Mundt]

On Wed, Nov 26, 2008 at 11:14:38AM +0800, walimis wrote:
> On Wed, Nov 26, 2008 at 12:04:46PM +0900, Paul Mundt wrote:
> >On Tue, Nov 25, 2008 at 10:03:35PM -0500, Steven Rostedt wrote:
> >> 
> >> On Wed, 26 Nov 2008, Liming Wang wrote:
> >> 
> >> > Please review this patch, and I don't know whether new sections will conflict 
> >> > with sched tracer or others. 
> >> > This patch is against Steven tree's tip/devel latest commit.
> >> > 
> >> > Impact: improve recordmcount.pl
> >> > 
> >> > Besides .text section, there are three .text sections that won't
> >> > be freed after kernel booting. They are: .sched.text, .spinlock.text
> >> > and .kprobes.text, which contain functions we can trace. But the last
> >> > section ".kprobes.text" is particular, which has been marked as "notrace",
> >> > we ignore it. Thus we add other two sections.
> >> 
> >> Yep, only those sections that we know will not disappear are OK. I was
> >> conservative with only using ".text" but I knew there were other sections 
> >> that were OK, so I used the hash array to make it easy to add others.
> >> 
> >Does your .text cover .text.head by the way?
> No, and not needed. Because .text.head is of booting code, to which
> function tracer can't trace.
> 
In the case of CPU hotplug it can be entered at any time. I don't see why
it can't be traced from the boot CPU at least.

Re: [PATCH 1/1] ftrace: adding other non-leaving .text sections

[walimis]

On Wed, Nov 26, 2008 at 12:18:50PM +0900, Paul Mundt wrote:
>On Wed, Nov 26, 2008 at 11:14:38AM +0800, walimis wrote:
>> On Wed, Nov 26, 2008 at 12:04:46PM +0900, Paul Mundt wrote:
>> >On Tue, Nov 25, 2008 at 10:03:35PM -0500, Steven Rostedt wrote:
>> >> 
>> >> On Wed, 26 Nov 2008, Liming Wang wrote:
>> >> 
>> >> > Please review this patch, and I don't know whether new sections will conflict 
>> >> > with sched tracer or others. 
>> >> > This patch is against Steven tree's tip/devel latest commit.
>> >> > 
>> >> > Impact: improve recordmcount.pl
>> >> > 
>> >> > Besides .text section, there are three .text sections that won't
>> >> > be freed after kernel booting. They are: .sched.text, .spinlock.text
>> >> > and .kprobes.text, which contain functions we can trace. But the last
>> >> > section ".kprobes.text" is particular, which has been marked as "notrace",
>> >> > we ignore it. Thus we add other two sections.
>> >> 
>> >> Yep, only those sections that we know will not disappear are OK. I was
>> >> conservative with only using ".text" but I knew there were other sections 
>> >> that were OK, so I used the hash array to make it easy to add others.
>> >> 
>> >Does your .text cover .text.head by the way?
>> No, and not needed. Because .text.head is of booting code, to which
>> function tracer can't trace.
>> 
>In the case of CPU hotplug it can be entered at any time. I don't see why
>it can't be traced from the boot CPU at least.
OK, I see. Need more investigating on it.
Thanks.

Re: [PATCH 1/1] ftrace: adding other non-leaving .text sections

[Steven Rostedt]

On Wed, 26 Nov 2008, Paul Mundt wrote:

> On Tue, Nov 25, 2008 at 10:03:35PM -0500, Steven Rostedt wrote:
> > 
> > On Wed, 26 Nov 2008, Liming Wang wrote:
> > 
> > > Please review this patch, and I don't know whether new sections will conflict 
> > > with sched tracer or others. 
> > > This patch is against Steven tree's tip/devel latest commit.
> > > 
> > > Impact: improve recordmcount.pl
> > > 
> > > Besides .text section, there are three .text sections that won't
> > > be freed after kernel booting. They are: .sched.text, .spinlock.text
> > > and .kprobes.text, which contain functions we can trace. But the last
> > > section ".kprobes.text" is particular, which has been marked as "notrace",
> > > we ignore it. Thus we add other two sections.
> > 
> > Yep, only those sections that we know will not disappear are OK. I was
> > conservative with only using ".text" but I knew there were other sections 
> > that were OK, so I used the hash array to make it easy to add others.
> > 
> Does your .text cover .text.head by the way?

No, you can add it if it is safe to trace, and it will not ever disappear.

-- Steve